Investigation of the Performance of the
New Orleans Flood Protection Systems
in Hurricane Katrina on August 29, 2005
Volume I: Main Text and Executive Summary
by
R. B. Seed, R. G. Bea, R. I. Abdelmalak, A. G. Athanasopoulos, G. P. Boutwell, J. D. Bray,
J.L. Briaud, C. Cheung, D. CobosRoa, J. CohenWaeber, B. D. Collins, L. Ehrensing, D. Farber,
M. Hanemann, L. F. Harder, K. S. Inkabi, A. M. Kammerer, D. Karadeniz, R.E. Kayen, R. E. S. Moss, J. Nicks,
S. Nimmala, J. M. Pestana, J. Porter, K. Rhee, M. F. Riemer, K. Roberts, J. D. Rogers, R. Storesund,
A. V. Govindasamy, X. VeraGrunauer, J. E. Wartman, C. M. Watkins, E. Wenk Jr., and S. C. Yim
Final Report
July 31, 2006
New Orleans Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
This project was supported, in part, by the National Science
Foundation under Grants No. CMS0413327 and CMS0611632.
Any opinions, findings, and conclusions or recommendations
expressed in this report are those of the author(s) and do not
necessarily reflect the views of the Foundation.
This report contains the observations and findings of an investigation
by an independent team of professional engineers and researchers with a wide
array of expertise. The materials contained herein are the observations
and professional opinions of these individuals, and do not necessarily reflect
the opinions or endorsement of any other group or agency.
Note: Cover Image from http://www.photolibrary.fema.gov/photodata/original/15022.jpg
i
New Orleans Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Source: www.noaa.gov
This report is dedicated to the people of the greater New Orleans region;
to those that perished, to those that lost friends and loved ones,
and to those that lost their homes, their businesses, their place of work,
and their community.
New Orleans has now been flooded by hurricanes six times
over the past century; in 1915, 1940, 1947, 1965, 1969 and 2005.
It must be our goal that it not be allowed to happen again.
ii
New Orleans Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Table of Contents
EXECUTIVE SUMMARY ……………………………………………………. xix
The Investigation Team ……………………………..…………………………… xxvi
Acknowledgements …………………………………...………………………..… xxix
VOLUME I: MAIN TEXT AND EXECUTIVE SUMMARY
PART I – INTRODUCTION:
Chapter 1: Introduction and Overview
1.1 Introduction …………………………………………………………….
11
1.2 Initial PostEvent Field Investigations………………………………….
11
1.3 Current Studies and Investigations ……………………………………
12
1.4 Organization of This Report ...………………………………………… 13
1.5 Elevation Datum ………………………………………………………
15
1.6 References ……………………………………………………………… 15
PART II – TECHNICAL STUDIES:
Chapter 2: Overview of Hurricane Katrina and its Aftermath
2.1 Hurricane Katrina ……..………………………………………………..
21
2.2 Overview of the New Orleans Flood Protection Systems ……………... 21
2.3 Overview of Flood Protection System Performance During
Hurricane Katrina …………………………………………………..
23
2.3.1
Storm Surge During Hurricane Katrina ……………………..... 23
2.3.2
Overview of the Performance of the Regional
Flood Protection System ……………………………..
2.3.3
25
Brief Comments on the Consequences of the Flooding
of New Orleans ……………………………………… 211
2.4 References ……………………………………………………………..
213
Chapter 3: Geology of the New Orleans Region
3.1 General Overview of the Geology of New Orleans …..………...……… 31
iii
New Orleans Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
3.1.1 Introduction ………………………………….…………………. 31
3.1.2 Evolution of the Mississippi Delta beneath New Orleans ……… 31
3.1.3 Pine Island Beach Trend ………………………….…….……... 33
3.1.4 Interdistributary Zones ………………………………….……..
34
3.1.5 Paludal Environments ……………………………….………… 35
3.1.5.1 Marshes ……………………………………….……… 35
3.1.5.2 Swamps ……………………………………….……… 36
3.1.5.3 Lacustrine Deposits …………………………..………
38
3.1.6 Recognition Keys for Depositional Environments ……..……… 39
3.1.7 Holocene Geology of New Orleans ……………………..……..
39
3.1.8 Faulting and Seismic Conditions ………………..……….…….. 311
3.2 Geologic Conditions at 17th Street Canal Breach …...…………….……. 311
3.2.1 Introduction …………………………………………….………. 311
3.2.2 Interpretation of Geology from Auger borings ………..……….. 311
3.2.3 Interpretation of Data from CPT Soundings …...……….……… 314
3.3 Geologic Conditions at London Avenue Canal (North) Breach ….……. 315
3.3.1 Introduction ………………….…………………………..……... 315
3.3.2 Geology Beneath the Levees ……………………………..……. 315
3.4 Geologic Conditions at London Avenue (South) Canal Breach …….…. 316
3.4.1 Introduction …………………………………………..………… 316
3.4.2 Geology Beneath the Levees …………………………..………. 316
3.5 Geologic Conditions along the Inner Harbor Navigation Canal ……….
317
3.5.1 Introduction …………………………………..………………… 317
3.5.2 Geology ……………………………………..………………….. 317
3.6 Paleontology and Age Dating …………………………..……………… 318
3.6.1 Introduction ………………………………..…………………… 318
3.6.2 Palynology ………………………………..…...………………... 318
3.6.3 Foraminifera …………………………….……………………… 318
3.6.4 Carbon 14 Dating …………………….………………………… 319
3.7 Mechanisms of Ground Settlement and Land Loss in Greater
New Orleans …………………………….………………...……….. 319
3.7.1 Settlement Measurements …….………………………………… 319
3.7.2 Tectonic Subsidence ……………………….…………………… 319
3.7.3 Lystric Growth Faults ………………….……………………….. 319
3.7.4 Compaction of Surficial Organic Swamp and
iv
New Orleans Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Marsh Deposit …………………………………….…………. 320
3.7.5 Structural Surcharging …………………………………….…….. 321
3.7.6 Extraction of Oil, Gas, and Water ……………………….……..
321
3.7.7 Coastal Land Loss …………………………………….………..
322
3.7.8 Negative Impact of Ground Settlement on Storm Surge …….… 322
3.7.9 Conclusions about Ground Settlement ……………….…….…..
323
3.8 References ……………………………………………………………..
323
Chapter 4: History of the New Orleans Flood Protection System ………….. 41
4.1 Origins of Lower New Orleans …………………….………………….. 41
4.2 Mississippi River Floods ……………………………………………….. 42
4.2.1 Mississippi River is the High Ground ……………………….….. 42
4.2.2 Flooding from the Mississippi River ……………………………. 42
4.3 The Mississippi River and Tributaries Project 19311972 ……...………. 46
4.3.1 Dimensions of Navigation Channels Maintained by the
Corps of Engineers on the Lower Mississippi River ………… 48
4.4 Flooding of the New Orleans Area by Hurricanes ……………………… 49
4.5 Flooding of New Orleans Caused by Intense Rain Storms …..………… 412
4.6 New Orleans Drainage Canals …………………………………………. 413
4.7 City Adopts Aggressive Drainage System ……………….……………… 416
4.7.1 PreKatrina Conditions and Maintenance by the S&WB ………. 419
4.7.2 Damage to S&WB Facilities and Capabilities Caused by
Hurricane Katrina and Rita ………..………………….……... 419
4.7.2 Reclamation of the MidCity Lowlands (early 1900s) …………. 420
4.7.3 1915 Flood Triggers Heightening of Drainage Canal Levees …… 420
4.7.4 The Lakefront Improvement Project (192634) ………………… 421
4.7.5 Second Generation of Heightening Drainage Canal
Levee Embankments (1947) …………………...……….......... 422
4.7.6 Federal Involvement with the City Drainage
Canals (1955 – present) ……………..……………………….
422
4.7.7 Hurricane Katrina Strikes New Orleans – August 2005 ………
423
4.8 Commercial Navigation Corridors ……………………………...………. 424
4.8.1 Inner Harbor Navigation Canal/Industrial Canal ……………...
424
4.8.2 Flooding Problems Around the IHNC ………………………...
426
4.8.3 Intracoastal Waterway …………………………………………
426
4.8.4 Mississippi River Gulf Outlet ………………………………….
427
v
New Orleans Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
4.9 Influence of Elevation Datums on New Orleans
Flood Protection System …………………………………… 428
4.9.1 Introduction …………………………………………………….. 428
4.9.2 17th Street Outfall Canal ……………………………………….. 429
4.9.3 London Avenue Outfall Canal …………………………………. 429
4.9.4 Orleans Outfall Canal ……………………..…………………… 429
4.9.5 Inner Harbor Navigation Canal – East Levee ………………….. 429
4.9.6 Inability to Apply Universal Corrections for Elevation Datums .. 430
4.10 Names of New Orleans Neighborhoods …………..…………………… 430
4.11 References ……………………………………………………………… 430
Chapter 5: The Lower Mississippi Region and Plaquemines Parish
5.1 Overview ………………………………………….…………………….. 51
5.2 Point a la Hache ………………………...………………………………. 52
5.3 Erosion Studies …..……………..……………………………………… 53
5.4 Summary ………………………………….……………………………. 53
5.5 References ……………………………………………………...………. 54
Chapter 6: The St. Bernard Parish and Lower Ninth Ward Protected Area
6.1 Introduction ……………………………………………………………. 61
6.2 The Northeast Frontage Levee …………………………………………. 61
6.3 The Two large Breaches on the East Bank of the IHNC at the Lower
Ninth Ward …………………………..…………..….……………. 65
6.3.1
The IHNC East Bank (South) Breach at the Lower Ninth Ward
66
6.3.2
The IHNC East Bank (North) Breach at the Lower Ninth Ward
613
6.3.3
Summary ……………………………………………………… 614
6.4 Summary and Findings ……………………………….………………… 615
6.5 References ……………………………………………………………… 616
Chapter 7: The New Orleans East Protected Area
7.1 Introduction………………………………………………………..……… 71
7.2 New Orleans East Hurricane Protection System ……….………………. 71
7.3 Performance of the New Orleans Hurricane Protection System
In Hurricane Katrina …………………………………………………. 72
7.3.1 Overview ………………………………………………………… 72
7.3.2 Chronology of Events in the New Orleans East
Protected Area ………………………………………………… 72
vi
New Orleans Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
7.3.3 Damage to Levee System Frontages ……………………………. 73
7.3.3.1 GIWW Frontage (Citrus Back and New Orleans
East Back Levees ……………………………………… 73
7.3.3.2 IHNC Frontage (IHNC East Levee) ……………………… 75
7.3.3.2 Lake Pontchartrain (New Orleans Lakefront, Citrus
Lakefront and New Orleans East Lakefront Levees)
and East Side Frontages (New Orleans East Levee) ….. 75
7.4 Summary of Findings for New Orleans Protected Area ………………… 75
7.5 References ………………………………………………………………. 76
Chapter 8: The Orleans East Bank (Downtown) and Canal District Protected Area
8.1 Overview ……………………………...………………………………… 81
8.2 Performance of the Flood Protection System Along the West
Bank of the Inner Harbor Navigation Channel (IHNC) …………….
83
8.2.1 An Early Breach at About 4:45 am ……………………………..
83
8.2.2 The CSX Railroad Breach ……………………………………… 84
8.2.3 Breaches and Distressed Sections at the Port of New Orleans …. 85
8.2.3.1 Breach at Rail Yard Behind the Port of New Orleans ….. 86
8.2.3.2 Erosional Distress at Floodgate Structure Behind
the Port of New Orleans ……………………………… 87
8.2.3.3 Two Adjacent Erosional Embankment Breaches at the
North End of the Port of New Orleans ………………. 88
8.2.4 Summary and Findings …………………………………………. 88
8.3 The Canal District Failures …..……………………..………...…………. 810
8.3.1 Introduction ……………………………….……………………. 810
8.3.2 The Lining of the Drainage Canals …………………………….. 811
8.3.3 The E99 Sheetpile Wall Test Section ………………………….. 812
8.3.4 Field Tests for Assessment of Underseepage Risk at the Canals .. 813
8.3.5 Water Levels Within the Canals During Hurricane Katrina ……... 814
8.3.6 The Orleans Canal ……………………………………………….. 815
8.3.7 The 17th Street Canal …………………………………………….. 817
8.3.7.1 The Breach on the East Bank …………………………… 817
8.3.7.2 Distressed Section on the West Bank …………………… 831
8.3.8 The Breach Near the South End of London Avenue Canal ……… 832
8.3.9 The Breach and Distressed Sections Near the North End
of the London Avenue Canal ……………………………………. 835
8.3.10 Summary and Findings …………………………………………. 839
vii
New Orleans Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
8.4 References ……………………………………………………………….. 842
Chapter 9:
Overtopping-Induced Erosion Studies
9.1 Erodibility: A Definition ……………………………………..………… 91
9.2 Erosion Process ………………………………………………………… 91
9.3 Velocity vs. Shear Stress ………………………………………..……… 91
9.4 Erosion Threshold and Erosion Categories ……………………..……… 92
9.5 Erodibility of CoarseGrained Soils ……………………….…………… 92
9.6 Erodibility of FineGrained Soils …………………………….………… 94
9.7 Erodibility and Correlation to Soil Properties ….……………………….. 96
9.8 The EFA: Erosion Function Apparatus ………………..……………….. 97
9.9 Some Existing Knowledge on Levee Erosion ………………………….. 99
9.9.1 Current Considerations in Design ………………………………. 99
9.9.2 Failure Mechanism ……………………………………...……… 99
9.9.3 Numerical Modeling ………………………………………….… 910
9.9.4 Laboratory Tests ……………………………………….………... 910
9.9.5 Field Tests ………………………………………………………. 911
9.9.6 Factors Influencing Resistance to Overtopping ………………… 912
9.9.7 Influence of Grass Cover on Surface Erosion ………….………. 913
9.10 Soil and Water Samples Used for Erosion Tests ………………….…... 914
9.11 Erosion Function Apparatus (EFA) Test Results ……………………… 916
9.11.1 Sample Preparation …………………………………………… 916
9.11.2 Sample EFA Test Results ……………………………………... 916
9.11.3 Summary Erosion Chart …………………………………...…… 917
9.11.4 Influence of Compaction on Erodibility ……………………….. 917
9.11.5 Influences of Water Salinity on Erodibility ……………………. 918
9.12 Index Properties of the Samples Tested in the EFA …………………… 918
9.13 Levee Overtopping and Erosion Failure Guideline Chart ……………... 918
9.14 Summary ……………………………………………………………….. 919
9.15 References ………………………………………………………..……. 919
Chapter 10:
Earthen Levee Evaluation
10.1 Overview ………………………………………………..…………….. 101
10.2 Levee Failure Mechanisms …………………………………………… 101
10.1.1 Structural Causes ……………………………………………… 102
10.1.2 Causes due to Hydraulic Forces ……………………………… 102
viii
New Orleans Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
10.1.3 Causes Involving Surface Degradation ………………………. 103
10.3 Design Standards ……………………………………………………... 104
10.3.1 United States Army Corps of Engineers Design Standards …… 104
10.3.1.1 Primary Design Procedure …………………………… 105
10.3.1.2 Material Selection ……………………………………. 106
10.3.1.3 Required Levee Soil Compaction ……………………. 106
10.3.1.4 Embankment Geometry ……………………………… 107
10.3.1.5 Identified Failure Modes …………………………….. 107
10.3.1.5 Erosion Susceptibility ………………………………... 107
10.3.2 United States Federal Emergency Management Agency
Design Standards ………………………………………….. 109
10.3.2.1 Freeboard …………………………………………….. 109
10.3.2.2 Closures ………………………………………………. 109
10.3.2.3 Embankment Protection ……………………………… 109
10.3.2.4 Embankment and Foundation Stability ……………….. 109
10.3.2.5 Settlement …………………………………………..
1010
10.3.2.6 Interior Drainage …………………………………..
1010
10.3.2.7 Other Design Criteria ………………………………..
1010
10.3.2.8 Other FEMA Requirements ………………………..
1011
10.6 Storm Surge and Wave Action During Hurricane Katrina …………
1011
10.7 Field Reconnaissance and Levee Condition Mapping ………………..
1011
10.7.1 Location 1 – Lakefront Airport ……………………………….
1012
10.7.2 Location 2 – Jahncke Pump Station Outfall …………………..
1013
10.7.3 Location 3 – Eastern Perimeter of New Orleans East …………
1014
10.7.4 Location 4 – Southeast Corner of New Orleans East ………….
1014
10.7.5 Location 5 – Entergy Michoud Generating Plant ……………… 1015
10.7.6 Location 6 – ICWW/MRGO Southern Levee ………………….. 1015
10.7.7 Location 7 – Bayou Bienvenue Control Structure ……………
1016
10.7.8 Location 8 – Mississippi River Gulf Outlet ……………………
1017
10.7.9 Location 9 – Bayou Dupre Control Structure ………………….
1018
10.7.10 Location 10 – St. Bernard Parish Interior Levee ……………... 1019
10.7.11 Summary of Observed Performance Factors …………………. 1020
10.8 Erosion Evaluation …………………………………………………….. 1022
10.9 Establishment of Design Criteria and Acceptability Performance …….
1026
10.9.1 USACE Risk Management Approach ………………………… 1026
ix
New Orleans Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
10.9.2 Other RiskBased Approaches …………………………………
1028
10.10 Conclusions …………………………………………………………… 1029
10.11 References …………………………………………………………….
1030
Chapter 11: Summary of Engineering Lessons
11.1 Introduction ……………………………………………………………
111
11.2 Overarching Strategic Issues …………………………………………..
111
11.2.1
Targeted Levels of Safety and Reliability …………………… 111
11.2.2 Funding and Resources ………………………………………..
112
11.3 Principal Engineering Findings and Lessons ………………………….
114
11.3.1 Introduction and Overview …………………………………… 114
11.3.2 Plaquemines Parish ……………………………………………
115
11.3.3 The East Flank; New Orleans East and the
St. Bernard/Lower Ninth Ward Protected Areas …………... 115
11.3.4 The Central Region; the IHNC and the GIWW/MRGO
Channel Frontages ………………………………………….. 119
11.3.5 The Lake Pontchartrain Frontage, and the Drainage Canals ….. 1114
11.4 References …………………………………………………………
1122
PART III – ORGANIZATIONAL AND INSTITUTIONAL ISSUES:
Chapter 12: Organized for Failure
12.1 Introduction …………………………………………………….…..….. 122
12.2 Purposes ……………………………………………………….………. 122
12.3 Failure of the NOFDS ………………………………………….……… 122
12.4 Extrinsic Factors …………………………….………………….……… 124
12.5 Intrinsic Factors ………………………………………………………… 1210
12.5.1 Standard Project Hurricane ……………………….…………… 1211
12.5.2 Failure Modes and Safety Factors ………………….…………. 1213
12.6 LifeCycle Development of Flaws ……………………….……….…… 1216
12.7 Findings – Looking Back ……………………………………………… 1217
12.8 References …………………………………………………………….. 1219
Chapter 13: Organized for Success
13.1 How Safe is Safe Enough ………………………..…………………….. 133
13.1.1 The Engineering Response to “How Safe is Safe?” …………… 135
x
New Orleans Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
13.1.2 Insights from Addressing These Issues …………………….….. 136
13.2 Maximizing How Safe is Safe Enough in the
U.S. Army Corps of Engineers (Context) …………………… 137
13.2.1 The Office of the President, the Congress, and the Corps …….. 137
13.2.2 Additional External Interstices for the Corps ……………......... 139
13.2.3 The Corps’ Internal Interstices ………………………………… 1311
13.4 Preventing the Next Katrina …………………………………………… 1311
13.5. Reengineering the USACE …………………………………………... 1312
13.5.1 Rebuilding the USACE Capacity …………………………….. 1313
13.5.2 Restructuring the Federal/State Relationship
in Flood Defense …………………………………………… 1313
13.5.3 Developing a National Flood Defense Authority ……………… 1314
13.5.4 Creating Effective Disaster Planning ………………………..... 1314
13.5.4.1 Creating a National Disaster Advisory
Office in the White House …………………….. 1315
13.5.4.2 Creating a Catastrophic Risk Office in Congress….. 1315
13.5.4.3 Making FEMA an HRO …………………………… 1316
13.6 Recommendations – Organizing for Success …………………………. 1316
13.7 References …………………………………………………………….. 1317
Chapter 14: Engineering for Success
14.1 Introduction …………………………………………………………… 141
14.2 Engineering Considerations …………………………………………… 143
14.2.1 Physical Facilities ………………………………………………. 143
14.3 Engineering Criteria and Guidelines ………………………………….. 1410
14.4 References …………………………………………………………….. 1411
PART IV – SUMMARY AND FINDINGS
Chapter 15: Findings and Recommendations
15.1 Overview ………………………………………………………….…… . 151
15.2 Performance of the Regional Flood Defense System
During Hurricane Katrina ……………………………………………… 151
15.3 Engineering Issues ……………………………………………………… 155
15.4 Looking Back – Organized for Failure …………………………………. 158
xi
New Orleans Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
15.5 Looking Forward – Organizing for Success ……………………………. 1510
15.5.1 Strategic and Engineering System Issues ……………………... 1510
15.5.2 Technology Delivery System Developments –
Organizing for Success ………………………………………… 1512
15.6 Conclusion ……………………………………………………………… 1513
xii
New Orleans Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
VOLUME II: APPENDICES
APPENDIX A: TERRESTRIAL LIDAR IMAGERY OF NEW ORLEANS
LEVEES AFFECTED BY HURRICANE KATRINA
A.1 Introduction ……………………………………………………………... A1
A.2 Methodology ……………………………………………………………. A1
A.3 Georeferencing of LIDAR survey data ………………………………….. A3
A.4 Processing of LIDAR Imagery ………………………………………….. A4
A.5 Data Coverage: LIDAR scan sites at Levee Breaks within the
New Orleans Area …………………………………….……………… A4
A.6 Analysis Examples of Levee Deformation Using LIDAR Data ………… A4
A.7 Summary ……..…………………………………………………………. A6
A.8 References ………………………………………………………………. A6
APPENDIX B: BORING LOGS ………………………………………………. B-1
APPENDIX C: CPT LOGS ……………..……………………………………… C-1
APPENDIX D: STE LABORATORY TESTING …………………..………. D-1
APPENDIX E: U.C. BERKELEY LABORATORY TESTING AND
ILIT IN-SITU FIELD VANE SHEAR TESTING ..……… E-1
APPENDIX F: LOOKING BACK
F.1 Synopsis of History of the New Orleans Flood Defense
System 1965 – 2005 ……………………………………………….. F1
F.2 Learning from Failures ………………………….……………………….. F7
F.2.1 Engineered Systems ……………………………………………. F7
F.2.2 Causes of Failures …………………………...…………………. F8
F.2.3 Magnitude of Failures ………………………………………….. F9
F.2.4 Breaching Defenses ……………………………………………… F9
F.2.5 Knowledge Challenges …………………………………………. F10
F.2.6 Organizational Malfunctions ……………………………………. F11
F.2.7 Engineering Challenges ………………………………………… F12
F.2.8 Initiating, Contributing, Compounding Events ………………… F13
F.2.9 High and Low Reliability Organizations: The NASA Columbia
Accident Investigation ………………………………………. F14
xiii
New Orleans Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
F.2.10 High Reliability Organizations ……………………………….. F14
F.2.11 Low Reliability Organizations ……………………………….. F17
F.2.12 Columbia Accident Investigation Board Findings …………… F17
F.2.13 Summary ……………………………………………………… F21
F.3 Quotations from Key Reports and Papers ……………………………….. F22
F.3.1 Townsend, F.F (2006). The Federal Response to Hurricane
Katrina, Lessons Learned, Report to the President of the United
States, The White House, Washington, D.C., February ………… F22
F.3.2 Select Bipartisan Committee to Investigate the Preparation for
and Response to Hurricane Katrina, 2006.A Failure of Initiative,
U.S. Government Printing Office, Washington, D.C. ………….. F23
F.3.3 Report of the Committee on Homeland Security and Governmental
Affairs. Hurricane Katrina, A Nation Still Unprepared, United States
Senate, Washington, D.C., May 2006. …………………………. F28
F.3.4 American Society of Civil Engineers External Review Panel
(ERP). Letter to LTG Carl Strock, Chief of U.S. Army Corps of
Engineers, February 20, 2006. ………………………………… F39
F.3.5 Committee on New Orleans Regional Hurricane Protection Projects,
National Academy of Engineering and the National Research Council,
2006. Report to The Honorable John Paul Woodley, Assistant
Secretary of the Army, Civil Works, Washington, D.C. February.. F41
F.3.6 U.S. Government Accountability Office, Army Corps of Engineers
History of the Lake Pontchartrain and Vicinity Hurricane Protection
Project, Statement of Anu Mittal, Direction Natural Resources and
Environment, Testimony Before the Committee on Environment and
Public Works, U.S. Senate, November 9, 2005; also Testimony before
the Subcommittee on Energy and Water Development,
Committee on Appropriations, House of
Representatives, September 28, 2005 ………………………… F42
F.3.7 U.S. General Accounting Office, Improved Planning Needed
by the Corps of Engineers to Resolve Environmental, Technical,
and Financial Issues on the Lake Pontchartrain Hurricane
Protection Project, Report to the Secretary of the Army,
August 17, 1982. ………………………………………………. F43
F.3.8 Houck, O. (2006). “Can We Save New Orleans?” Tulane
Environmental Law Journal, Vol 19, Issue 1, 168, New Orleans,
Louisiana. ……………………………………………………… F44
F.3.9 Member Scholars of the Center for Progressive Reform (2005).
An Unnatural Disaster: The Aftermath of Hurricane Katrina, Center
for Progressive Reform Publication, CPR Publication #512,
September. ……………………………………………………… F51
xiv
New Orleans Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
F.3.10 Braun, S. and Vartabedian, R. (2005). “The Politics of Flood
Control,” Los Angeles Times, December 25. ………………… F55
F.3.11 Vartabedian, R. and Braun, S. (2006). “Fatal Flaws: Why
the Walls Tumbled in New Orleans,” Los Angeles Times,
January 17. ……………………………………………………. F61
F.3.12 Irons, L. (2005). “Hurricane Katrina as a Predictable Surprise,”
Homeland Security Affairs, Vol. I, Issue 2, Article 7,
http://hsaj.org/hsa ...................................................................... F66
F.3.13 Congressional Research Service Report for Congress (2005).
Protecting New Orleans: From Hurricane Barriers to Floodwalls,
N.T. Carter, Washington, D.C., December. ………………….. F69
F.3.14 Congressional Research Service Report for Congress (2005).
Flood Risk Management: Federal Role in Infrastructure, N. T.
Carter, October, Washington, D.C. ……………………………. F72
F.3.15 Office of Management and Budget (2006). Agency Scorecards,
Washington, D.C., …………………………………………….. F76
F.3.16 Senator Susan Collins and Senator Joseph Liberman, Senate
Homeland Security Committee Holds Hurricane Katrina Hearing
to Examine Levees in New Orleans, Press Release,
November 2, 2005. …………………………………………… F76
F.3.17 Senator Susan Collins (2005). “Hurricane Katrina: Who’s in
Charge of the New Orleans Levees?” Hearing Statement Before
Homeland Security and Governmental Affairs Committee,
December 15, Washington, D.C. ……………………………… F77
F.3.18 Herman Leonard and Arnold Howitt (2006). “Katrina as Prelude:
Preparing for and Responding to Future KatrinaClass
Disturbances in the United States,” Testimony U.S. Senate Homeland
Security and Governmental Affairs Committee, Washington,
D.C., March 8. …………………………………………………. F77
F.3.19 Congressional Research Service (2003). Army Corps of Engineers
Civil Works Program: Issues for Congress, Issue Brief for
Congress, N. Carter and P. Sheikh, Washington, D.C., May 21.. F78
F.3.20 U.S. General Accounting Office (2003). Corps of Engineers
Improved Analysis of Costs and Benefits Needed for Sacramento
Flood Protection Project, Report to Congressional Requesters,
GAO0430, Washington, D.C., October ……………………… F80
F.3.21 Heinzerling, L. and Ackerman, F. (2002). “Pricing the Priceless:
CostBenefit Analysis of Environmental Protection,” Georgetown
Environmental Law and Policy Institute, University
Law Center …………………………………………………….. F82
F.4 References ………………………………………………………………. F83
xv
New Orleans Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
APPENDIX G: LOOKING FORWARD
G.1 High Reliability Organization: The USN Nuclear Propulsion
Program ……………………………………………………………….. G2
G.1.1 The USN Nuclear Propulsion Program ………………………… G2
G.1.2 Personnel and Recruitment Retention ………………………….. G4
G.1.3 Engineering Assumptions ………………………………………. G6
G.1.4 Conclusion ………………………………………………………. G6
G.2 Findings from Other Studies: Organizing for Success
G.2.1 Report of the Committee on Homeland Security and Governmental
Affairs (2006). Hurricane Katrina, A Nation Still Unprepared,
United States Senate, Washington, D.C., May ……………………….. G7
G.2.2 Senator Susan Collins (2006). “Opening Statement,” Committee on
Homeland Security and Governmental Affairs, Hurricane Katrina:
Recommendations for Reform,” Washington, D.C., March 8 ………… G10
G.2.3 Newt Gingrich (2006). “Why New Orleans Needs Saving.”
Time Magazine, March 6. ……………………………………………… G10
G.2.4 Houck, O. (2006). “Can We Save New Orleans?” Tulane
Environmental Law Journal, Vol. 19, Issue 1, 168, New Orleans
Louisiana ………………………………………………………………. G11
G.2.5 Netherlands Water Partnership (2005). Dutch Expertise, Water
Management & Flood Control, Delft, The Netherlands,
November. ……………………………………………………………. G17
G.2.6 Interagency Floodplain Management Review Committee (1994).
Sharing the Challenge: Floodplain Management into the 21st
Century, Report to Administration Floodplain Management
Task Force, Washington, D.C. June. ………………………………….. G18
G.2.7 Input from Citizens of the Greater New Orleans Area:
Levees.Org. ……………………………………………………………. G19
G.2.8 Congressional Research Service (2005). Aging Infrastructure:
Dam Safety Report, Report for Congress, K. Powers, Washington,
D.C., September 29. ……………………………………………
G20
G.2.9 Sparks, R.E., (2006). “Rethinking, Then Rebuilding New Orleans,”
Issues in Science and Technology, National Academy Press,
Winter 2006, p 3339, Washington, D.C. …………………………….. G20
G.2.10 Curole, W. (2005). Comprehensive Hurricane Protection Plan
Guidelines, General Manager, South Lafourche Levee District
Presentation to French Quarter Citizens Group,
November 2005. ……………………………………………………… G23
G.2.11 Lopez, J. (2005). The Multiple Lines of Defense Strategy to
Sustain Louisiana’s Coast. Report to Lake Pontchartrain
xvi
New Orleans Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Basin Foundation, New Orleans. ……………………………………… G23
G.2.12 Committee on the Restoration and Protection of Coastal
Louisiana (2006). Drawing Louisiana’s New Map, Ocean Studies
Board, National Research Council, The National Academies Press,
Washington, D.C. ……………………………………………………… G27
G.2.13 Working Group for PostHurricane Planning for the Louisiana Coast,
A New Framework for Planning the Future of Coastal Louisiana
after the Hurricanes of 2005, University of Maryland Center for
Environmental Science, Cambridge, January 26, 2006. ………………. G33
G.3 References ……………………………………………………………...
G41
APPENDIX H: HOW SAFE IS SAFE? Coping with Mother Nature, Human
Nature and Technology’s Unintended Consequences
H.1 Preface ………………………………………………………………..
H1
H.2 Introduction ……………………………………………………………
H2
H.2.1 How Safe is Safe? ………………………………………………. H2
H.2.2 Risk Analysis as a Survival Skill ……………………………….. H7
H.2.3 Tradeoffs Between Risks, Cost of Mitigation,
and Performance ……………………………………………….. H11
H.2.4 Voluntary versus Involuntary Risk ……………………………… H13
H.2.5 Coping with Threats to Life, Liberty, Property,
and the Environment ……………………………………….. H14
H.3 Government’s Responsibility for Security ……………………………
H17
H.3.1 Risk Management: Our Constitution, Public Policy
and our Culture …………………………………………………. H17
H.3.2 Resolution by Political Power and Political Will ……………….. H19
H.3.3 The President and Congress: Needs for Advice
and Counsel ……………………………………………………... H21
H.4 Technology and Its Side Effects ………………………………………
H21
H.4.1 Beyond Technique, Technology as a Social Process …………… H21
H.4.2 Technology’s Unintended Consequences ……………………… H24
H.4.3 What You Can’t Model You Can’t Manage ……………………. H27
H.4.4 OverDesign as a Safety Margin ……………………………….. H28
H.5 Bed Rock Values in Public Policy ……………………………………
H31
H.5.1 The Rainbow of Stakeholders ………………………………….. H31
H.5.2 Conflict Management to Balance Benefits and Costs ………….. H33
H.5.3 Tensions Between Industry and Government …………………... H35
H.6 The Ethics of Informed Consent ………………………………………
xvii
H37
New Orleans Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
H.6.1 The Role of Media in Exposing Risks ………………………….. H37
H.6.2 The Power of Informed Consent ……………………………….. H40
H.7 Lessons from the Past …………………………………………………
H42
H.7.1 The Exxon Valdez as a Metaphor for System Failure ………….. H42
H.7.2 Deficits of Foresight, Vigilance, Contingency
Resources, Political Will and Trust …………………………….. H45
H.8 Thinking About The Future …………………………………………...
H.8.1 Evaluating Social Choice by Outcomes for the
Children …………………………………………………….
H48
H48
H.8.2 Foresight as an Imperative in Risk Management ………………. H49
H.8.3 Pathologies for the Short Run ………………………………….. H51
H.8.4 Early Warning of Close Encounters ……………………………. H53
H.9 The Anatomy of Risk – A Summary …………………………………… H54
H.9.1 Applying These Concepts to Katrina …………………………… H56
APPENDIX I: EROSION TEST RESULTS ON
NEW ORLEANS LEVEE SAMPLES
I.1 The EFA: Erosion Function Apparatus ………………………………… I1
I.1.1 EFA test procedures ……………………………………………… I1
I.1.2 EFA test data reduction…………………………………………… I1
I.2 Soil and Water Samples Used for Erosion Tests ………………………… I3
I.3 Erosion Function Apparatus (EFA) Test Results ………………………… I5
I.3.1 Sample preparation ………………………………………………. I5
EFA Test Results …………………………………………………………….. I7
xviii
New Orleans Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
EXECUTIVE SUMMARY
This report presents the results of an investigation of the performance of the New
Orleans regional flood protection system during and after Hurricane Katrina, which
struck the New Orleans region on August 29, 2005. This event resulted in the single
most costly catastrophic failure of an engineered system in history. Current damage
estimates at the time of this writing are on the order of $100 to $200 billion in the greater
New Orleans area, and the official death count in New Orleans and southern Louisiana at
the time of this writing stands at 1,293, with an additional 306 deaths in nearby southern
Mississippi. An additional approximately 300 people are currently still listed as
“missing”; it is expected that some of these missing were temporarily lost in the shuffle
of the regional evacuation, but some of these are expected to have been carried out into
the swamps and the Gulf of Mexico by the storm’s floodwaters, and some are expected to
be recovered in the ongoing sifting through the debris of wrecked homes and businesses,
so the current overall regional death count of 1,599 is expected to continue to rise a bit
further. More than 450,000 people were initially displaced by this catastrophe, and at the
time of this writing more than 200,000 residents of the greater New Orleans metropolitan
area continue to be displaced from their homes by the floodwater damages from this
storm event.
This investigation has targeted three main questions as follow: (1) What
happened?, (2) Why?, and (3) What types of changes are necessary to prevent recurrence
of a disaster of this scale again in the future?
To address these questions, this investigation has involved: (1) an initial field
reconnaissance, forensic study and data gathering effort performed quickly after the
arrival of Hurricanes Katrina (August 29, 2005) and Rita (September 24, 2005), (2) a
review of the history of the regional flood protection system and its development, (3) a
review of the challenging regional geology, (4) detailed studies of the events during
Hurricanes Katrina and Rita, as well as the causes and mechanisms of the principal
failures, (4) studies of the organizational and institutional issues affecting the
performance of the flood protection system, (5) observations regarding the emergency
repair and ongoing interim levee reconstruction efforts, and (6) development of findings
and preliminary recommendations regarding changes that appear warranted in order to
prevent recurrence of this type of catastrophe in the future.
In the end, it is concluded that many things went wrong with the New Orleans
flood protection system during Hurricane Katrina, and that the resulting catastrophe had
it roots in three main causes: (1) a major natural disaster (the Hurricane itself), (2) the
poor performance of the flood protection system, due to localized engineering failures,
questionable judgments, errors, etc. involved in the detailed design, construction,
operation and maintenance of the system, and (3) more global “organizational” and
institutional problems associated with the governmental and local organizations
responsible for the design, construction, operation, maintenance and funding of the
overall flood protection system.
xix
New Orleans Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
After eight months of detailed study, a much clearer picture has now emerged
regarding the causes and mechanisms of this catastrophe. Many of the findings of this
study represent a different view of key elements of this event than has been publicly
presented to date.
Hurricane Katrina was a large hurricane, and its arrival at New Orleans
represented the root cause of a natural disaster. This disaster grew to a full blown
catastrophe, however, principally due to the massive and repeated failure of the regional
flood protection system and the consequent flooding of approximately 85% of the greater
metropolitan area of New Orleans.
As Hurricane Katrina initially approached the coast, the resulting storm surge and
waves rose over the levees protecting much of a narrow strip of land on both sides of the
lower Mississippi River extending from the southern edge of New Orleans to the Gulf of
Mexico. Most of this narrow protected zone, Plaquemines Parish, was massively
inundated by the waters of the Gulf.
The eye of the storm next proceeded to the north, on a path that would take it just
slightly to the east of New Orleans.
Hurricane Katrina has been widely reported to have overwhelmed the eastern side
of the New Orleans flood protection system with storm surge and wave loading that
exceeded the levels used for design of the system in that area. That is a true statement,
but it is also an incomplete view. The storm surge and wave loading at the eastern flank
of the New Orleans flood protection system was not vastly greater than design levels, and
the carnage that resulted owed much to the inadequacies of the system as it existed at the
time of Katrina’s arrival. Some overtopping of levees along the eastern flank of the
system (along the northeastern frontage of the St. Bernard and Ninth Ward protected
basin, and at the southeast corner of the New Orleans East protected basin), and also in
central areas (along the GIWW channel and the IHNC channel) was inevitable given the
design levels authorized by Congress and the surge levels produced in these areas by the
actual storm. It does not follow, however, that this overtopping had to result in
catastrophic failures and breaching of major portions of the levees protecting these areas,
nor the ensuing catastrophic flooding of these populous areas.
The northeast flank of the St. Bernard/Ninth Ward basin’s protecting “ring” of
levees and floodwalls was incomplete at the time of Katrina’s arrival. The critical 11
mile long levee section fronting “Lake” Borgne (which is actually a Bay, connected
directly to the Gulf of Mexico) was being constructed in stages, and funding
appropriation for the final stage had long been requested by the U.S. Army Corps of
Engineers (USACE), but this did not arrive before Katrina struck; as a result large
portions of this critical levee frontage were several feet below final design grade. In
addition, an unfortunate decision had been made to use local dredge spoils from the
excavation of the adjacent MRGO channel for construction of major portions of the
xx
New Orleans Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
levees along this frontage. The result was that major portions of these levees were
comprised of highly erodeable sand and lightweight shell sand fill.
When the storm surge arrived, massive portions of these levees eroded
catastrophically and the storm surge passed through this frontage while still on the rise,
crossed an open swamp area that should have safely absorbed most of the overtopping
flow from the outer levees (if they had not catastrophically eroded), and it then crossed
easily over a secondary levee of lesser height that had not been intended to face a storm
surge largely undiminished by the minimal interference of the too rapidly eroded outer
levees fronting Lake Borgne. The resulting carnage in St. Bernard Parish was
devastating, as the storm surge rapidly filled the protected basin to an elevation of
approximately +12 feet above sea level; deeply inundating even neighborhoods with
ground elevations well above sea level in this area.
The storm surge swelled waters of Lake Borgne also passed over and then
through a length of levees at the southeast corner of the New Orleans East protected
basin. Here too, the levees fronting Lake Borgne had been constructed primarily using
materials dredged from the excavation of an adjacent channel (the GIWW channel), and
these levees also contained major volumes of highly erodeable sands and lightweight
shell sands. These levees were also massively eroded, and produced the principal source
of flooding that eventually inundated the New Orleans East protected area. Here again
there was an area of undeveloped swampland behind the outer levees that might have
absorbed the brunt of any overtopping flow, and a secondary levee of lesser height was in
place behind this swampland that might then have prevented catastrophic flooding of the
populous areas of New Orleans East. This secondary levee was not able to resist the
massive flows resulting from the catastrophic erosion of the highly erodeable section of
the Lake Borgne frontage levee, however, and the floodwaters passed over the secondary
levee and began the filling of the New Orleans East protected basin.
The catastrophic erosion of these two critical levee frontages need not have
occurred. These frontages could instead have been constructed using well compacted
clay fill with good resistance to erosion, and they could have been further armored in
anticipation of the storm surge and wave loading from Lake Borgne. The levee at the
northeast edge of St. Bernard Parish could have been completed in a more timely manner.
The result would have been some overtopping, but not catastrophic erosion and
uncontrolled breaching of these critical frontages. Some flooding and damage would
have been expected, but it need not have been catastrophic.
The storm surge swollen waters of Lake Borgne next passed laterally along the
eastwest trending GIWW/MRGO channel to its intersection at a “T” with the north
south oriented IHNC channel, overtopping levees along both banks to a limited degree.
This produced an additional breach of a composite earthen levee and concrete floodwall
section along the southern edge of New Orleans East, adding additional uncontrolled
inflow to this protected basin. This failure could have been prevented at little
incremental cost if erosion protection (e.g. a concrete splash pad, or similar) had been
emplaced along the back side of the concrete floodwall at the levee crest, but the USACE
xxi
New Orleans Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
felt that this was precluded by Federal rules and regulations regarding authorized levels
of protection.
The surge next raised the water levels within the IHNC channel, and produced a
number of failures on both the east and west banks. Two major failures occurred on the
east side of the IHNC, at the west edge of the Ninth Ward. Overtopping occurred at both
of these locations, but this was not the principal cause of either of these failures. Both
failures were principally due to underseepage flows that passed beneath the sheetpile
curtains supporting the concrete floodwalls at the crests of the levees. Like many
sections of the flood protection system, these sheetpiles were too shallow to adequately
cut off, and thus reduce, these underseepage flows. The result was two massive breaches
that devastated the adjacent Ninth Ward neighborhood, and then pushed east to meet with
the floodwaters already rapidly approaching from the east from St. Bernard Parish as a
result of the earlier catastrophic erosion of the Lake Borgne frontage levees.
Several additional breaches also occurred farther north on the east side of the
IHNC fronting the west side of New Orleans East, but these were relatively small
features and they just added further to the uncontrolled flows that were now progressively
filling this protected basin. These breaches occurred mainly at junctures between
adjoining, dissimilar levee and floodwall sections, and represented good examples of
widespread failure to adequately engineer these “transitions” between sections of the
regional flood protection system.
Several breaches occurred on the west side of the IHNC, and these represented the
first failures to admit uncontrolled floodwaters into the main metropolitan (downtown)
protected area of New Orleans. These features did not scour and erode a path below sea
level, however, so they admitted floodwaters for a number of hours and then these
inflows ceased as the storm surge in the IHNC eventually subsided. Only 10% to 20% of
the floodwaters that eventually inundated a majority of the main (downtown) New
Orleans protected basin entered through these features.
These failures and breaches on the west side of the IHNC all appear to have been
preventable. One failure was the result of overtopping of an Iwall, with the overtopping
flow then eroding a trench in the earthen levee crest at the inboard side of the floodwall.
This removal of lateral support unbraced the floodwall, and it was pushed over laterally
by the water pressures from the storm surge on the outboard side. Here again the
installation of erosional protection (e.g. concrete splash pads or similar) might have
prevented the failure.
The other failures in this area occurred at “transitions” between disparate levee
and floodwall sections, and/or at sections where unsuitable and highly erodible
lightweight shell sand fills had been used to construct levee embankments. Here, again,
these failures were as much the result of design choices and/or engineering and oversight
issues as the storm surge itself.
xxii
New Orleans Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
As the eye of the hurricane next passed to the northeast of New Orleans, the
counterclockwise swirl of the storm winds produced a storm surge against the southern
edge of Lake Pontchartrain. This produced additional temporary overtopping of a long
section of levee and floodwall at the west end of the lakefront levees of New Orleans
east, behind the old airport, adding further to the flows that were progressively filling this
protected basin.
The surge against the southern edge of Lake Pontchartrain also elevated the water
levels within three drainage canals at the northern edge of the main metropolitan
(downtown) New Orleans protected basin, and this would produce the final, and most
damaging, failures and flooding of the overall event.
The three drainage canals should not have been accessible to the storm surge.
The USACE had tried for many years to obtain authorization to install floodgates at the
north ends of the three drainage canals that could be closed to prevent storm surges from
raising the water levels within the canals. That would have been the superior technical
solution. Dysfunctional interaction between the local Levee Board (who were
responsible for levees and floodwalls, etc.) and the local Water and Sewerage Board
(who were responsible for pumping water from the city via the drainage canals)
prevented the installation of these gates, however, and as a result many miles of the sides
of these three canals had instead to be lined with levees and floodwalls.
The lining of these canals with levees topped with concrete floodwalls was
rendered very challenging due to (a) the difficult local geology of the foundation soils,
and (b) the narrow right of way (or available “footprint”) for these levees. As a result of
the decision not to install the floodgates, the three canals represented potentially
vulnerable “daggers” pointed at the heart of the main metropolitan New Orleans
protected basin. Three major breaches would occur on these canals; two on the London
Avenue Canal and one on the 17th Street Canal. All three of these breaches eroded and
scoured rapidly to well below sea level, and these three major breaches were the source
of approximately 80% of the floodwaters that then flowed into the main (downtown)
protected basin over the next three days, finally equilibrating with the still slightly
elevated waters of Lake Pontchartrain on Thursday, September 1.
The central canal of the three, the Orleans Canal, did not suffer breaching, but a
section of floodwall topping the earthen levee approximately 300 feet in length near the
south end of the canal had been left incomplete, again as a result of dysfunctional
interaction between the local levee board and the water and sewerage board. This
effectively reduced the level of protection for this canal from about +12 to +13 feet above
sea level (the height of the tops of the floodwalls lining the many miles of the canal) to an
elevation of about +6 to +7 feet above sea level (the height of the earthen levee crest
along the 300 foot length where the floodwall that should have topped this levee was
omitted). As a result of the missing floodwall section, flow passed through this “hole”
and began filling the heart of the main New Orleans protected basin. This flow
eventually ceased as the storm surge subsided, and so was locally damaging but not
catastrophic.
xxiii
New Orleans Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
The three breaches on the 17th Street and London Avenue canals were
catastrophic. None of these failures were the result of overtopping; surge levels in all
three drainage canals were well below the design levels, and well below the tops of the
floodwalls. Two of these breaches were the result of stability failures of the foundation
soils underlying the earthen levees and their floodwalls, and the third was the result of
underseepage passing beneath the sheetpile curtain and resultant catastrophic erosion near
the inboard toe of the levee that eventually undermined the levee and floodwall.
A large number of engineering errors and poor judgements contributed to these
three catastrophic design failures, as detailed in Chapter 8. In addition, a number of these
same problems appear to be somewhat pervasive, and call into question the integrity and
reliability of other sections of the flood protection system that did not fail during this
event. Indeed, additional levee and floodwall sections appear to have been potentially
heading towards failure when they were “saved” by the occurrence of the three large
breaches (which rapidly drew down the canal water levels and thus reduced the loading
on nearby levee and floodwall sections.)
The New Orleans regional flood protection system failed at many locations during
Hurricane Katrina, and by many different modes and mechanisms. This unacceptable
performance was to a large degree the result of more global underlying “organizational”
and institutional problems associated with the governmental and local organizations
jointly responsible for the design, construction, operation, and maintenance of the flood
protection system, including provision of timely funding and other critical resources.
Our findings to date indicate that no one group or organization had a monopoly on
responsibility for the catastrophic failure of this regional flood protection system. Many
groups, organizations and even individuals had a hand in the numerous failures and
shortcomings that proved so catastrophic on August 29th. It is a complex situation,
without simple answers.
It is not without answers and potential solutions, however, just not simple ones.
There is a need to change the process by which these types of large and critical protective
systems are created and maintained. It will not be feasible to provide an assured level of
protection for this large metropolitan region without first making significant changes in
the organizational structure and interactions of the national and more local governmental
bodies and agencies jointly responsible for this effort. Significant changes are also
needed in the engineering approaches and procedures used for many aspects of this work,
and there is a need for interactive and independent expert technical oversight and review
as well. In numerous cases, it appears that such review would have likely caught and
challenged errors and poor judgements (both in engineering, and in policy and funding)
that led to failures during Hurricane Katrina.
Simply updating engineering procedures and design manuals will not provide the
needed level of assurance of safety of the population and properties of this major
metropolitan region. Design procedures and standards employed for many elements of
xxiv
New Orleans Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
the flood protection system can be traced back to initial development and use for design
and construction of levees intended for protection of largely unpopulated agrarian land,
not a major urban region. Design levels of safety and reliability were nowhere near those
generally used for major dams; largely because dams are considered to pose a potential
risk to large populations. There are few U.S. dams that pose risk to populations as large
as the greater New Orleans region, however, and it is one of the recommendations of this
study that standards and policies much like those used for “dams” should be adopted for
levee systems protecting such regions.
Simply addressing engineering design standards and procedures is unlikely to be
sufficient to provide a suitably reliable level of protection. There is also a need to resolve
dysfunctional relationships between federal and more local government, and the federal
and local agencies responsible for the actual design, construction and maintenance of
such flood protection systems. Some of these groups need to enhance their technical
capabilities; a longterm expense that would clearly represent a prudent investment at
both the national and local level, given the stakes as demonstrated by the losses in this
recent event. Steady commitment and reliable funding, shorter design and construction
timeframes, clear lines of authority and responsibility, and improved overall coordination
of disparate system elements and functions are all needed as well.
And there is some urgency to all of this. The greater New Orleans regional flood
protection system was significantly upgraded in response to flooding produced by
Hurricane Betsy in 1965. The improved flood protection system was intended to be
completed in 2017, fully 52 years after Betsy’s calamitous passage. The system was
incomplete when Katrina arrived. As a nation, we must manage to dedicate the resources
necessary to complete projects with such clear and obvious ramifications for public safety
in a more timely manner.
New Orleans has now been flooded by hurricanes six times over the past century;
in 1915, 1940, 1947, 1965, 1969 and 2005. It should not be allowed to happen again.
xxv
New Orleans Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
THE INVESTIGATION TEAM
The University of California at Berkeley led Independent Levee Investigation
Team (ILIT) grew through the course of this investigation, and eventually numbered 35
very dedicated and accomplished individuals.
The team included a large number of leading experts across a diverse range of
fields. Team members came from six states, and they came from universities, private
engineering firms, and state and federal agencies.
As a group, the investigation team had very impressive prior experience with
forensic studies of major disasters and catastrophes. For example, the team members had
previously investigated 12 major earthquakes and 8 major hurricanes (both domestic and
foreign), 14 dam failures, more than a dozen levee failures, numerous landslides, one
tsunami, the pivotal Kettleman Hills waste landfill failure, the Challenger and Columbia
space shuttle disasters, the Exxon Valdez tanker disaster, and a number of major offshore
pipeline and oil platform failures. They are well experienced with the carnage and
disarray of disasters, and with the unforgettable smell of death. They are also well
experienced at the delicate and deliberate art and science of piecing their way through the
devastation, carefully and professionally, and figuring out what had happened, and why;
the art and science of engineering forensics.
The calibre of these assembled experts is such that we could never possibly have
afforded to hire them. Instead, excepting a handful of graduate research students who
worked for very low wages, these world class experts all volunteered, and they worked
pro bono (for free.) They did this for the intellectual challenge, for the camaraderie of a
very special group of accomplished colleagues, for the chance to make a positive
difference, because it was important, and most importantly because it was the right and
necessary thing to do.
The pages that follow list the names and affiliations of the members of the
Independent Levee Investigation Team. I have had the opportunity to work on a number
of investigations of major catastrophes and disasters, but I have never worked with a finer
group. They are all heroes in my book.
Dr. Raymond B. Seed
Head, ILIT
xxvi
New Orleans Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
The Independent Levee Investigation Team
Remon I. Abdelmalak, Ph.D., Graduate Researcher, Zachry Department of Civil Engineering,
Texas A&M University, TX
Adda G. Athanasopoulos, P.E., Ph.D. Student, Department of Civil and Environmental
Engineering, University of California, Berkeley, CA
Robert G. Bea, Ph.D., P.E., G.E., S.E., Professor, Department of Civil and Environmental
Engineering, University of California, Berkeley, CA
Gordon P. Boutwell, Jr., Ph.D., P.E., President, Soil Testing Engineers, Inc., Baton Rouge &
New Orleans, LA
Jonathan D. Bray, Ph.D., P.E., Professor, Department of Civil and Environmental Engineering,
University of California, Berkeley, CA
Jean-Louis Briaud, Ph.D., P.E., Professor and Holder of the Buchanan Chair, Zachry
Department of Civil Engineering, Texas A&M University, TX
Carmen Cheung, Graduate Researcher, Department of Civil and Environmental Engineering,
University of California, Berkeley, CA
Diego Cobos-Roa, P.E., Graduate Researcher, Department of Civil and Environmental
Engineering, University of California, Berkeley, CA
Julien Cohen-Waeber, Graduate Researcher, Department of Civil and Environmental Engineer
ing, University of California, Berkeley, CA
Brian D. Collins, Ph.D., P.E., Research Civil Engineer, United States Geological Survey, Menlo
Park, CA
Luke Ehrensing, P.E., President, Thigpen Construction, New Orleans, LA.
Dan A. Farber, J.D., Sho Sato Professor of Law, University of California, Berkeley, CA
W. Michael Hanneman, Ph.D., Chancellor's Professor, Department of Agricultural & Resource
Economics and Goldman School of Public Policy, University of California, Berkeley, CA
Leslie F. Harder. Jr., Ph.D. P.E., Acting Deputy Director for Public Safety, California
Department of Water Resources, Sacramento, CA
Kofi Inkabi, Graduate Researcher, Department of Civil and Environmental Engineering,
University of California, Berkeley, CA
Anne M. Kammerer, Ph.D., P.E., Senior Risk Consultant, Arup, San Francisco, CA
Deniz Karadeniz, Ph.D., Candidate, Department of Geological Sciences and Engineering,
University of MissouriRolla, MO
Robert E. Kayen, Ph.D., P.E., Research Scientist, United States Geological Survey, Menlo Park,
CA
Robb E. S. Moss, Ph.D., P.E., Assistant Professor of Civil Engineering, California Polytechnic
Institute and State University, San Luis Obispo, CA
Jennifer Nicks, Graduate Researcher, Zachry Department of Civil Engineering, Texas A&M
University, TX
xxvii
New Orleans Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Seshu Nimala, Graduate Researcher, Civil, Construction, and Environmental Engineering,
Oregon State University, OR
Juan M. Pestana, Ph.D., P.E., Associate Professor, Department of Civil and Environmental
Engineering, University of California, Berkeley, CA
Jim Porter, C.E.T., Soil Testing Engineers, Inc., Baton Rouge & New Orleans, LA
Keunyong Rhee, Graduate Researcher, Zachry Department of Civil Engineering, Texas A&M
University, TX
Michael F. Riemer, Ph.D., Associate Adjunct Professor, Civil and Environmental Engineering,
University of California, Berkeley, CA
Karlene Roberts, Ph.D., Haas School of Business, University of California, Berkeley, CA
J. David Rogers, Ph.D., P.E., R.G., Hasselmann Chair in Geological Engineering, University of
MissouriRolla, MO
Raymond B. Seed, Ph.D., Professor, Department of Civil and Environmental Engineering,
University of California, Berkeley, CA
Rune Storesund, P.E., Consulting Engineer, Storesund Consulting, Albany, CA
Anand V Govindsamy, Graduate Researcher, Zachry Department of Civil Engineering, Texas
A&M University. TX
Xavier Vera-Grunauer, P.E., CVA Consulting group, Guayaquil, Ecuador
Joseph Wartman, Ph.D., P.E., Assistant Professor, Civil, Architectural and Environmental
Engineering, Drexel University, Philadelphia, PA
Conor M. Watkins, Graduate Researcher, Department of Geological Sciences and Engineering,
University of MissouriRolla, MO
Ed Wenk, Jr., D. Eng., Emeritus Professor of Engineering, Public Administration and Social
Management of Technology, University of Washington at Seattle, WA
Solomon C. Yim, Ph.D., P.E., Professor, Civil, Construction, and Environmental Engineering,
Oregon State University, OR
xxviii
New Orleans Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
ACKNOWLEDGEMENTS
This report would not have been possible without the generous help of many individuals
and organizations.
This project was supported, in part, by the National Science Foundation (NSF) under
Grants No. CMS0413327 and CMS0611632, and additional support was provided by the Center
for Information Technology Research in the service of Society (CITRIS) at the University of
California at Berkeley. This support is gratefully acknowledged.
The authors also wish to express their gratitude to the U.S. Army Corps of Engineers
(USACE) for their considerable assistance with numerous elements of this work. Their field
investigation team from the Engineer Research Development Center (ERDC) in Vicksburg hosted
and assisted our own field investigation team in the critical early days of late September and early
October. The USACE has also posted massive amounts of background documents on their
website, and this has been an invaluable resource. The USACE, and the Interagency Performance
Evaluation Team (IPET) have graciously shared much of their field and laboratory data, and we
have done the same. This positive sharing and collaboration helps everyone by providing the best
possible basis for study and analysis of this event.
We are also deeply grateful to the honorable men and women of the USACE who have
taken extra measures to help to provide additional documents, data and insight. Many of these
prefer not to be named, but their dedication to service of the greater public good in this difficult
situation has been admirable.
We are deeply grateful to the members of the State of Louisiana’s independent
investigation team, Team Louisiana, for their tremendous efforts and dogged persistence under
very difficult circumstances, and for their generous mutual sharing of data and insights
throughout this investigation. This team consists of Dr. Ivor Van Heerden, Dr. Paul Kemp and
Dr. Hassan Mashriqui (all from the Louisiana State University Hurricane Research Center), Billy
Prochaska and Dr. Lou Cappozzoli (both local geotechnical consultants), and Art Theis (retired
head of the Louisiana Department of Public Works.) The people of Louisiana, and the nation,
owe these gentlemen a great debt as their persistent efforts have, time and again, produced critical
data and insights that would not otherwise have been available.
We are also grateful to the members of the field investigation team of the American
Society of Civil Engineers, who jointly formed a combined team with ours in the urgent initial
postevent field studies when it was of vital importance to gather all possible data and
observations while (fully necessary) emergency repair operations were already damaging and
burying critical evidence. This was a very strong field forensics team, and their collaboration
both in the field and in the subsequent preparation of an initial Preliminary Report which was
issued in early November of 2005, was of great value.
Finally we are deeply grateful to the many others who will remain anonymous, but who
have assisted by providing information, data, background history and other information that
might otherwise not have been available.
A great many people gave generously of themselves, their time, and their expertise to
assist these studies. It was important, and we are profoundly grateful.
xxix
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
CHAPTER ONE: INTRODUCTION AND OVERVIEW
1.1 Introduction
This report presents the results of an investigation of the performance of the New
Orleans regional flood protection system during and after Hurricane Katrina, which struck the
New Orleans region on August 29, 2005. This event resulted in the single most costly
catastrophic failure of an engineered system in history. Current damage estimates at the time
of this writing are on the order of $100 to $200 billion in the greater New Orleans area, and
the official death count in New Orleans and southern Louisiana at the time of this writing
stands at 1,293 with an additional 306 deaths in nearby southern Mississippi. An additional
approximately 300 people are currently still listed as “missing”; it is expected that some of
these missing were temporarily lost in the shuffle of the regional evacuation, but some of
these are expected to have been carried out into the swamps and the Gulf of Mexico by the
storm’s floodwaters, and some are expected to be recovered in the still ongoing sifting
through the debris of wrecked homes and businesses, so the current overall regional death
count of 1,599 is expected to continue to rise a bit further. More than 450,000 people were
initially displaced by this catastrophe, and at the time of this writing, more than 200,000
residents of the greater New Orleans metropolitan area continue to be displaced from their
homes by the floodwater damages from this storm event.
This investigation targets three main questions as follow: (1) What happened? (2)
Why? and (3) What types of changes are necessary to prevent recurrence of a disaster of this
scale again in the future?
To address these questions, this investigation has involved: (1) an initial field
reconnaissance, forensic study and data gathering effort performed quickly after the arrival of
Hurricanes Katrina (August 29, 2005) and Rita (September 24, 2005), (2) a review of the
history of the regional flood protection system and its development, (3) a review of the
challenging regional geology, (4) detailed studies of the events during Hurricanes Katrina and
Rita, as well as the causes and mechanisms of the principal failures, and studies of sections
that performed successfully as well, (5) studies of the organizational and institutional issues
affecting the performance of the flood protection system, (6) observations regarding the
emergency repair and ongoing interim levee reconstruction efforts, and (7) development of
findings and preliminary recommendations regarding changes that appear warranted in order
to prevent recurrence of this type of catastrophe in the future.
1.2 Initial Post-Event Field Investigations
A critical early stage of this investigation was the initial field investigations performed
by collaborating teams of engineers and scientists in the wake of the passage of Hurricane
Katrina, to study performance of the regional flood protection system and the resulting
flooding that occurred in the New Orleans area. The principal focus of these efforts was to
capture perishable data and observations related to the performance of flood protection system
before they were lost to ongoing emergency response and repair operations.
11
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Several independent investigation teams jointly pooled their efforts in order to capture
as much data as possible in the precious timeframe available. The two principal participating
teams were from the University of California at Berkeley (UC Berkeley) which included a
number of colleagues from other firms and institutions, and a team from the American Society
of Civil Engineers (ASCE) organized by its GeoInstitute and by its Coasts, Oceans, Ports,
and Rivers Institute. A team from Louisiana State University’s Hurricane Research Center
(LSU/HRC) also accompanied the field investigation teams during their first week of
investigations. These teams were accompanied and assisted in the field by members of the
U.S. Army Corps of Engineers (USACE) levee investigation team from the Engineer
Research and Development Center (ERDC). All of these investigative teams shared data and
findings freely and openly, and the mutual pooling of talents and expertise greatly benefited
all as it enabled the field teams to gather more data in the critical days available.
These initial field investigations occurred over a span of approximately three weeks,
from September 26 through October 15, 2005, and the preliminary observations and findings
were presented in a report jointly authored by the UC Berkeleyled field investigation team
and the ASCE field investigation team (Seed, et al.; November 15, 2005.)
1.3 Current Studies and Investigations
Subsequent to these initial field investigations, three main investigations have been
carried forward. The largest of these is the U.S. Army Corps of Engineers’ own internal
investigation, the Interagency Performance Evaluation Team (IPET) study. The IPET study is
by far the largest of the three investigations, and has a budget of approximately $20 million.
The American Society of Civil Engineers (ASCE) has been hired, for an additional $2
million, to form a review panel (called the External Review Panel, E.R.P.) to review the
results of the IPET studies. This ASCE review panel works and consults closely with the
IPET studies and is focused specifically on reviewing the IPET investigation efforts, data and
findings. The National Research Council (NRC) has also been hired, by the Department of
Defense, to provide an additional review of the IPET studies after the ASCE’s E.R.P.
completes its task. This NRC review panel has announced its intention of reviewing input
from all investigation teams and efforts as part of this task.
The IPET study is narrowly focused and constrained in its first year to consideration
and study of only “what happened” in a strictly physical sense; it is specifically not to address
underlying faults or to assign “responsibility” in its initial studies (Final Draft Report due to
the ASCE review panel on May 15, 2006, and Final Report due on June 1, 2006), but rather to
wait and study “organizational issues”, “human factors”, etc. during the following year.
The second investigation team moving forward is Team Louisiana, representing the
interests of the State of Louisiana in performing an investigation independent of the USACE.
Team Louisiana is led by Dr. Ivor Van Heerden, and its core is formed by a number of his
colleagues from the Louisiana State University (LSU) Hurricane Research Center
(LSU/HRC), with additional members from a number of local engineering consulting firms
and state organizations. Team Louisiana does not have the massive funding or manpower of
the IPET team, but they are strongly motivated and have worked very hard and well given
12
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
their logistical limitations and the difficult situation of the region (which has directly affected
some of the team’s members, as well as many of their friends and colleagues.)
The third investigation team moving forward is our own UC Berkeleyled Independent
Levee Investigation Team (ILIT). Our budget is also not as large as that of the IPET study,
and currently stands at approximately $350,000. We have, however, managed to assemble a
team of 37 outstanding engineers and researchers. Pages “xxv” through “xxvi” describe the
team. As a group, the conjugate forensic experience in prior investigations of numerous
major engineering and natural disasters is very impressive. This is an amazingly strong team,
and we could never possibly have afforded to hire them within our small budget. These
leading experts have, instead, volunteered to work for free (pro bono), and our budget is thus
devoted instead towards covering travel expenses, field borings and sampling, and laboratory
testing, etc. We have elected to decline proffered offers of additional funding, as it appears
important that our investigation team maintain its demonstrable independence and neutrality
in these studies.
1.4 Organization of this Report
This report presents the results of studies directed towards answering three main sets
of questions as follow:
1. What happened? What events transpired during Hurricane Katrina and during its
aftermath? How did the regional flood protection system perform? What were the
successes, and what were the shortcomings and failures? What mechanisms and
forces, etc., led to these performances?
2. Why did this happen? What were the underlying issues that led to the observed
performance of the system elements? What were the influences of regional and local
geology? How did the history of the evolution of the flood protection system
contribute to its performance? What were the design assumptions, engineering studies
and analyses, etc., and what effect did these have on the performance of the system
elements? What overarching organizational, institutional, political and funding issues
may have played a role?
3. What can be done to ensure that a similar catastrophe does not recur in the
future? This report presents preliminary findings and recommendations regarding
changes in organization of the overall governmental/institutional “system” responsible
for the conception, design, construction, operation and maintenance of the complex
regional flood defense system, as well as the making of political decisions regarding
levels of protection to be provided, and the provision of funding to support the
creation and operation/maintenance of such a system. This report also presents
preliminary findings and recommendations regarding a number of focused areas for
improvement of the conceptual design, analysis and engineering design, and
construction and maintenance of such a system.
In the end, it is concluded that many things went wrong with the New Orleans flood
protection system during Hurricane Katrina, and that the resulting catastrophe had it roots in
13
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
three main causes: (1) a major natural disaster (the Hurricane itself), (2) the poor performance
of the flood protection system, due to localized engineering failures, questionable judgements,
errors, etc. involved in the detailed design, construction, operation and maintenance of the
system, and (3) more global “organizational” and institutional problems associated with the
governmental and local organizations responsible for the design, construction, operation,
maintenance and funding of the overall flood protection system.
Chapter 2 presents an overview of the principal events that occurred during and after
the arrival of Hurricane Katrina in the New Orleans area, with emphasis on the storm surge
and wave loadings, and the resulting performance of the regional flood protection system.
Chapter 3 presents a summary overview of the challenging regional and local geology
that so strongly affects the difficulties associated with the creation of regional flood protection
systems, and their performance as well.
Chapter 4 presents a review of the history of the development of the New Orleans
regional flood protection system. It is a truism of levees and flood protection that the fabric
and history of a given region is usually closely interwoven within the fabric of the levees and
flood protection systems that are created in that region.
Chapters 5 through 8 present the results of studies and analyses of the performance of
the four main leveeprotected areas principally affected by Hurricane Katrina. These chapters
present overviews of the performance of the flood protection system in each of the four areas,
of the flooding that occurred within each of these areas, and detailed analyses of the
performance of critical subelements of the system within each area. These analyses include
an investigation of the causes of critical failures, and the apparent reasons for these including
both engineering/construction types of issues as well as organizational/institutional issues.
These chapters also present observations, recommendations and findings related to some of
the emergency posthurricane repair and reconstruction efforts.
Chapter 9 presents the results of studies of issues associated with overtopping erosion
and scour; a key phenomenon involved in both the successful and unsuccessful performances
of numerous critical levee and floodwall sections throughout the region.
Chapter 10 briefly addresses a series of “other issues”, including a brief overview of
design standards, observations regarding a number of recurrent issues that appear to be
problematic throughout the regional flood protection system, performance assessment with
regard to erosion and erodeability of placed fills, a brief overview of the performance of the
pumping systems that “unwater” the protected areas of these studies, and observations and
comments regarding the initial emergency levee and floodwall breach repair efforts, and the
ongoing interim repair and reconstruction efforts, at a number of locations.
Chapter 11 presents a summary review of the engineering issues addressed in Chapters
2 through 10, and recommendations for changes in engineering and design practices to
address these.
14
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Chapters 12 through 14 examine a number of organizational and institutional issues
that affected the performance of the regional flood protection systems during Hurricane
Katrina. They also address recommendations for moving forward; recommendations for a
number of changes to ensure that we never again have to study a catastrophe of this type and
scale in southern Louisiana.
Chapter 12 begins with a review of background and history pertaining to these types
of issues. Chapter 13 then presents a review and examination of critical organizational,
institutional, political and funding issues that directly affected the performance of the New
Orleans regional flood protection system, and also some of the posthurricane repair and
reconstruction efforts. These organizational/institutional issues had a dominant impact on the
overall performance of the regional flood protection systems, and many of the problems that
led to the catastrophic flooding of much of the greater New Orleans region can be traced
directly (at least in large part) to these types of underlying issues.
Chapter 14 presents preliminary recommendations for changes that can and should be
made in moving forward, in order to ensure that a catastrophe of this scale is never repeated in
the future. The New Orleans regional flood protection system did not perform well in
Hurricane Katrina. We can do better. This chapter presents recommendations for changes in
specific engineering analysis and design procedures, conceptual design features and
approaches, specific system elements, etc. This chapter also presents recommendations
regarding changes in the overall system of governmental bodies, governmental agencies,
outsourced (private sector) engineering and construction, local oversight agencies, and the
regulations and procedures involved in the overall conception, design, construction, operation
and maintenance of complex and regionally massive systems protecting vital public safety for
populous regions such as this.
Finally, Chapter 15 presents a summary overview of these studies, and of the principal
findings and recommendations.
1.5 Elevation Datum
There are a number of datums that have been and continue to be used for elevation
references throughout the New Orleans Region. A good discussion of these is presented in
the IPET Interim Report No. 2 (IPET; April 1, 2006). The situation is further confused as
some regional benchmarks, which were considered stable, have recently been found to have
instead subsided, so that elevations based on these require correction. In this present report,
all elevations are stated in terms of local Mean Sea Level (MSL), which corresponds
approximately to the NAVD88 (2004.65) datum. [This NAVD88 (2004.65) datum is currently
thought to be within approximately 3inches of Mean Sea Level in the New Orleans area.]
All elevations in this report have been resolved, as best we were able with the information
available, to this MSL (or approximately NAVD88; 2004.65) datum.
1.6 References
Seed, R. B., et al., “Preliminary Report on the Performance of the New Orleans Levee
Systems in Hurricane Katrina on August 29, 2005”, Report No. UCB/CITRIS – 05/01,
November 17, 2005.
15
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
CHAPTER TWO: OVERVIEW OF HURRICANE KATRINA
AND ITS AFTERMATH
2.1
Hurricane Katrina
The path of Hurricane Katrina’s eye is shown in Figures 2.1 and 2.2. Hurricane Katrina
crossed the Florida peninsula on August 25, 2005 as a Category 1 hurricane. It then entered the
Gulf of Mexico, where it gathered energy from the warm Gulf waters, producing a hurricane that
eventually reached Category 5 status on Sunday, August 28, shortly before making its second
mainland landfall just to the east of New Orleans early on Monday, August 29, as shown in
Figures 2.1 and 2.2. The Hurricane had weakened to a Category 4 level prior to landfall on the
morning of August 29, and it weakened further as it came ashore.
Because the eye of this hurricane passed just slightly to the east of New Orleans, the
hurricane imposed unusually severe wind loads and storm surges (and waves) on the New
Orleans region and its flood protection systems.
2.2
Overview of the New Orleans Flood Protection Systems
Figure 2.3 shows the main study region. The City of New Orleans is largely situated
between the Mississippi River, which passes along the southern edge of the main portion of the
city, and Lake Pontchartrain, which fronts the city to the north. Lake Borgne lies to the east,
separated from developed areas by open swampland. “Lake” Borgne is not really a lake at all;
instead it is a bay as it is directly connected to the waters of the Gulf of Mexico. To the southeast
of the city, the Mississippi River bends to the south and flows out through its delta into the Gulf
of Mexico.
The flood protection system that protects the New Orleans region is organized as a series
of protected basins or “protected areas”, each protected by its own perimeter levee system, and
these are “unwatered” by pumps.
As shown in Figures 2.4 and 2.5, there are four main protected areas that comprise the
New Orleans flood protection system of interest. A number of additional leveeprotected units
also exist in this area, but the focus of these current studies is the four main protected areas shown
in Figures 2.4 and 2.5. These were largely constructed under the supervision of the U.S. Army
Corps of Engineers, to provide improved flood protection in the wake of the devastating flooding
caused by Hurricane Betsy in 1965.
Figures 2.4 and 2.5 show the locations of most of the levee breaches and severely
distressed (but nonbreached, or only partially breached) levee sections covered by these studies.
Levee breaches are shown with solid blue stars, and distressed sections as well as minor or partial
breaches are indicated by red stars. The original base maps, and many of the stars, were
graciously provided by the USACE (2005), and a number of additional blue and red stars have
been added to the map in Figure 2.4 as a result of the studies reported herein. The yellow stars
shown in these figures correspond to deliberate breaches made after Hurricane Katrina, to
facilitate draining the flooded areas after the storm.
21
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
The pink shading in Figures 2.4 and 2.5 shows developed areas that were flooded, and the
areas shaded with blue crosshatching indicate undeveloped swamp land that was flooded. The
deeper blue shading (near the east end of New Orleans East) denotes areas that still remained to
be unwatered as late as September 28, 2005. As shown in these figures, approximately 85% of
the metropolitan area of New Orleans was flooded during this event.
As shown in Figure 2.4, the Orleans East Bank (Metro Orleans) section is one
contiguously protected section. This protected unit contains the downtown district, the French
Quarter, the Garden District, and the “Canal” District. The northern edge of this protected area is
fronted by Lake Pontchartrain on the north, and the Mississippi River passes along its southern
edge. The Inner Harbor Navigation Canal (also locally known as “the Industrial Canal”) passes
along the east flank of this protected section, separating the Orleans East Bank protected section
from New Orleans East (to the northeast) and from the Lower Ninth Ward and St. Bernard Parish
(directly to the east.) Three large drainage canals extend into the Orleans East Bank protected
section from Lake Pontchartrain to the north, for the purpose of conveying water pumped north
into the lake by large pump stations within the city. These canals, from west to east, are the 17th
Street Canal, the Orleans Canal, and the London Avenue Canal.
A second protected section surrounds and protects New Orleans East, as shown in Figure
2.4. This protected section fronts Lake Pontchartrain along its north edge, and the Inner Harbor
Navigation Canal (IHNC) along its west flank. The southern edge is fronted by the Mississippi
River Gulf Outlet channel (MRGO) which coexists with the Gulf Intracoastal Waterway
(GIWW) along this stretch. The eastern portion of this protected section is currently largely
undeveloped swampland, contained within the protective levee ring. The east flank of this
protected section is fronted by additional swampland, and Lake Borgne is located slightly to the
southeast.
The third main protected section contains both the Lower Ninth Ward and St. Bernard
Parish, as shown in Figure 2.4. This protected section is also fronted by the Inner Harbor
Navigation Canal on its west flank, and has the MRGO/GIWW channel along its northern edge.
At the northeastern corner, the MRGO bends to the south (away from the GIWW channel) and
fronts the boundary of this protected area along the northeastern edge. Open swampland occurs
to the south and southeast. Lake Borgne occurs to the east, separated from this protected section
by the MRGO channel and by a narrow strip of undeveloped marshland. The main urban areas
occur within the southern and western portions of this protected area. The fairly densely
populated Lower Ninth Ward is located at the west end, and St. Bernard Parish along
approximately the southern half of the rest of this protected area. The northeastern portion of this
protected section is undeveloped marshy wetland, as indicated in Figure 2.4. A secondary levee,
operated and maintained by local levee boards, separates the undeveloped marshlands of the
northeastern portions of this protected area from the Ninth Ward and St. Bernard Parish urban
areas.
The fourth main protected area is a narrow, protected strip along the lower reaches of the
Mississippi River heading south from St. Bernard Parish to the mouth of the river at the Gulf of
Mexico, as shown in Figure 2.5. This protected strip, with “river” levees fronting the Mississippi
River and a second, parallel set of “storm” levees facing away from the river forming a protected
corridor less than a mile wide, serves to protect a number of small communities as well as utilities
and pipelines. This protected corridor also provides protected access for workers, supplies and
22
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
gas and oil pipelines servicing the large offshore oil fields out in the Gulf of Mexico. This will be
referred to in this report as “the Plaquemines Parish” levee protected zone.
The current perimeter levee and floodwall defense systems for these four protected areas
were largely designed and constructed under the supervision of the U.S. Army Corps of
Engineers in the wake of the catastrophic flooding caused by Hurricane Betsy of 1965. These
flood protection improvements typically involved either new levee construction, or raising
existing levee defenses and/or adding new floodwalls, to provide storm flood protection for
higher elevations of storm surge waters (and waves) at locations throughout the region.
2.3 Overview of Flood Protection System Performance During Hurricane Katrina
2.3.1
Storm Surge During Hurricane Katrina
The regional flood protection system had been designed to safely withstand the storm
surges and waves associated with the Standard Project Hurricane, which was intended to
represent a scenario roughly “typical” of a rapidly moving Category 3 hurricane passing close to
the New Orleans metropolitan region. Chapter 12 (Section 12.5.1) presents a more detailed
discussion of the “Standard Project Hurricane”, and the criteria for which the regional flood
protection system was designed. In simple terms, the system was intended to have been designed
to safely withstand storm surge levels (plus waves) to specified elevations at various locations, as
shown in Figures 2.6 and 2.7.
In general, the “Standard Project Hurricane” provided for design to safely withstand storm
surge rises (plus waves) to prescribed elevations at various locations throughout the system. The
levels selected correspond generally to the storm surge level (mean peak storm surge water
elevation, without waves) associated with the “Standard Project Hurricane” conditions plus an
additional allowance for most (but not always all) of expected additional wave runup.
As shown in Figures 2.6 and 2.7, this resulted in a targeted protection level of about
elevation +17 feet to +19 feet (MSL), or 17 to 19 feet above Mean Sea Level, at the eastern flank
of the system, and + 13.5 feet to +18 feet (MSL) along much of the southern edge of Lake
Pontchartrain. The storm surge levels within the various drainage canals and navigational
channels varied, and the storm surge levels for design were typically on the order of Elev. + 14
feet to + 16 feet (MSL) along the GIWW and IHNC channels, and Elev. + 12.5 feet to + 14.5 feet
(MSL) along the 17th Street, Orleans, and London Avenue Canals in the “Canal District”. There
is some minor confusion as to the most recent “Standard Project Hurricane”, and the most recent
storm surge design levels at some locations; the values indicated in Figure 2.6 are an
interpretation by the Government Accountability Office (GAO, 2006) based in part on initial
research by the staff of the New Orleans Times Picayune, and the values shown in Figure 2.7
have been added to this figure by our team, and are our own current best interpretation.
The situation is further clouded a bit, as the actual targeted levee and floodwall heights
along a given section also varied slightly as a function of waterside topography, obstacles and
vegetation, levee geometry, orientation and potential wind fetch (distance of potential wind travel
across the top of open water), etc. as these would affect the potential runup heights of storm
waves. Variations for these types of issues were typically minor, on the order of two feet or less.
23
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
There is, however, no “typical” hurricane, nor associated storm surge, and the actual wind,
wave and storm surge loadings imposed at any location within the overall flood protection system
during an actual hurricane are a function of location relative to the storm, wind speed and
direction, orientation of levees, local bodies of water, channel configurations, offshore contours,
vegetative cover, etc. These loadings vary over time, as the storm moves progressively through
the region.
Figures 2.8 and 2.9 show plots of storm surge levels resulting from numerical modeling
simulations performed by the LSU Hurricane Research Center, for two different points in time
during Hurricane Katrina, based on analyses of the storm track, wind speeds, regional topography
and local conditions (marsh growth, soil stiffness, offshore contours, etc.) (Louisiana State
University Hurricane Center, 2005.) The water levels shown in Figures 2.8 and 2.9 were
predicted using a regionally calibrated numerical model, and the results shown in Figure 2.8
represent a point in time when the eye of the hurricane was first approaching the coast from the
Gulf of Mexico, and those shown in Figure 2.9 correspond to a time when the eye of the storm
was passing slightly to the east of New Orleans. These calculations are part of an overall single
analysis of storm surge levels throughout the region, and throughout the continuous period of
time as the storm approached and then passed through the region. Based on actual field
observations and measurements of maximum storm surge levels at more than 100 locations
throughout the region, this global analysis of storm surge levels is expected to be accurate
(relative to surge levels that actually occurred) within approximately ± 15% at all locations of
interest for these current studies (IPET, 2006.)
Predicted and actual storm surge heights varied over time, at different locations, and the
water levels shown in Figures 2.8 and 2.9 do not represent predictions of the peak storm surges
noted at all locations. Instead, these images show calculated conditions at two interesting points
in time when: (a) [Fig. 2.8] the initial large surge was being driven up against the coast of the
Gulf of Mexico in the New Orleans region by the approaching storm, and (b) [Fig. 2.9] at a
particularly critical moment when a large storm surge had first “inflated” (raised the level of)
Lake Borgne, then the locally prevailing westward swirl of the counterclockwise hurricane winds
threw the risen waters of Lake Borgne westward over the adjacent levees protecting eastern
flanks of the New Orleans East and St. Bernard/Lower Ninth Ward protected areas, as shown
schematically in Figure 2.11.
These types of storm surge modeling calculations are being performed by a number of
research and investigation teams, and are constantly being calibrated and updated based on actual
field measurements of high water marks, etc. The USACE’s IPET investigation team are
devoting significant effort to these types of hydrodynamic analytical “hindcasts”, and the IPET
back analyses provided to date to our UC Berkeleyled ILIT study team are in good agreement
with the storm surge predictions shown in Figures 2.8 and 2.9 at most locations of interest for
these studies (IPET; Draft Final Report, June 1, 2006).
Figure 2.10 shows an aggregate summary of the calculated peak storm surges, at any point
in time during Hurricane Katrina, based on similar calculations performed by the IPET study
(IPET; March, 2006). These calculations are very similar to those developed by the Louisiana
investigation team, and both the IPET and Team Louisiana analyses will be used as a partial basis
for estimation of storm surge levels and wave conditions in these current studies. The maximum
flood stages calculated (predicted) by the two sets of analyses are generally in good agreement at
24
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
most points of interest. Agreement regarding storm waves is also generally good, but the
differences between the two sets of predicted storm waves are a bit more significant at a few
locations of interest. Discussions of the IPET and Team Louisiana hydrodynamic storm surge
and storm wave calculations will be presented, in more detail, at locations of interest in the
chapters that follow.
It should be noted that a number of different datums have been used as elevation
references throughout the historic development of the New Orleans regional levee systems, and
this situation is further complicated by ongoing subsidence in the region. This investigation has
elected to resolve these differences between different datums, and to refer to all elevations in this
report (as consistently as possible) in terms of elevation with respect to the NAVD88 (2004.65)
datum; approximately “mean sea level” in the region. This particular version of the NAVD88
datum is currently thought to be within about 3inches of Mean Sea Level (MSL) in the New
Orleans region. For a more indepth discussion of differences between the various datums used in
the greater New Orleans region, please see IPET Interim Report No. 2 (IPET; March, 2006).
2.3.2
Overview of the Performance of the Regional Flood Protection System
Hurricane Katrina, as expected, produced a large onshore storm surge from the Gulf of
Mexico. As shown in Figures 2.8 through 2.10 this produced significant overtopping of storm
levees along the lower Mississippi River reaches in the Plaquemines Parish area, and numerous
levee breaches occurred in this area, as shown previously in Figure 2.5. In simple terms, the
“storm” levees of Plaquemines Parish were largely overwhelmed by the large storm surge; they
were overtopped by the storm surge and by the large storm waves that accompanied the average
rise (storm surge) in water levels. Fortunately, the Plaquemines Parish protected corridor is only
sparsely populated, and the local inhabitants were acutely aware of the risk that they faced so that
evacuation in advance of the storm was unusually complete.
Plaquemines Parish was largely inundated by the massive storm surge and the numerous
resulting levee breaches. Most breaches appear to have been primarily the result of overtopping
and erosion, and it is interesting to note that these breaches occurred mainly in the “storm” levees,
while the “river” levees often better withstood the storm surge (and waves) without catastrophic
erosion. The devastation within Plaquemines parish produced by this flooding was very severe,
as described in Chapter 5. By approximately 7:00 a.m. on the morning of Monday, September
29, most of Plaquemines Parish was under water.
A more detailed discussion of the performance of the flood protection systems in the
Plaquemines Parish area is presented in Chapter 5.
As the storm surge began to raise the water levels throughout the New Orleans region, it
began to raise the water levels within the GIWW, MRGO and IHNC channels. As the water level
within the IHNC began to rise, the first “breach” within the metropolitan New Orleans region
(north of Plaquemines Parish) occurred at about 5:00 a.m. somewhere along the IHNC. This was
evidenced by a pronounced, and shortlived, decrease in the rate of water level rise at two gage
stations along the IHNC at this point in time. There are several breaches along this section of the
IHNC that might have accounted for this observed water level gage behavior, and this is
discussed in Chapter 8. This was a “noncatastrophic” failure; although the breach eroded and
became enlarged by the flow, the “lip” of the breach remained above sea level. As a result,
25
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
although water flowed for a while into the protected area, this flow later stopped as the storm
surge subsequently subsided. Simple calculations, based on flood stages and breach sequences
and dimensions, suggest that less than 5% of the water that eventually flowed into the main
Orleans East Bank (downtown) protected zone entered through this breach.
The large onshore storm surge also raised water levels within Lake Borgne (which is
directly connected to the Gulf.) Lake Borgne rose up, and outgrew its normal banks. As the
storm then passed to the east of New Orleans, the prevailing counterclockwise swirl of the storm
winds drove the waters of Lake Borgne as a large storm surge to the west, against the eastern
flank of the regional flood protection systems as shown schematically in Figure 2.11. This
produced a storm surge estimated at approximately +16 to +18 feet (MSL), as shown in Figures
2.9 and 2.10.
This storm surge level exceeded the crest heights of the levees along a nearly 11mile long
stretch of the northeastern edge of the St. Bernard/Lower Ninth Ward protected area. The levees
along this frontage were intended to be built to provide protection to a level of approximately
+17.5 feet (MSL), but at the time of Hurricane Katrina many of the levees along this frontage had
crest elevations approximately 2 to 4 feet lower than that. This was because the levees along this
frontage had not yet been completed. These were “virgin” levees, being constructed on swampy
foundation soils that had not previously had significant levees before. Accordingly, the swampy
shallow foundation soils were both weak and compressible, and the levees were being constructed
in stages to allow time for consolidation and settlement of the foundations soils. This process
also allowed time for the drying of the very wet locally excavated soils used for some portions of
the levee embankment fills, and also for increases in strength of the underlying foundation soils
as they compressed under the weights of the growing levees.
Construction of the first phase of the levees along this frontage began in the late 1960’s.
The last major work in this area prior to Katrina had been the construction of the third phase, in
199495. Since that time, the USACE had been waiting for Congressional appropriation of the
funds necessary to construct the final stage (to the full design height, with allowance for
anticipated future settlements.) Now it is too late.
In addition to the levees along this frontage being well below design grade, the manner of
construction and the materials used were nontypical of most other USACE levees in the region.
Ordinarily, the USACE requires the use of “cohesive” (clayey) soils to create an embankment fill
that is both strong and relatively resistant to erosion. The levees along the “MRGO” frontage at
the northeast edge of the St. Bernard Parish/Ninth Ward protected area were instead “sand core”
levees (USACE, 1966). These levees were constructed using locally available soils, including
dredge spoils from the excavation of the adjacent MRGO channel.
This is a region with predominantly marshy deposits, consisting largely of organic soils
and soft paludal swamp clays with very high water contents. Beneath these generally poor
surficial soils, the most common materials occurring at shallow, relatively accessible depths tend
to be predominantly sandy soils that are highly erodeable and generally unsuitable for levee
embankment fill. A decision was made, however, to attempt to use the locally available soils
rather than importing higher quality soil fill materials. The USACE Design Memorandum
describing this design refers to these as “sand core” levees (USACE, 1966).
26
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
The levees along this MRGO frontage section (along the northeastern edge of the St.
Bernard protected area) were, in the end, constructed using large volumes of the spoil material
excavated during the dredging of the adjacent MRGO shipping channel, and they contained
unusually large quantities of highly erodeable sandy soils. In addition, some of the more cohesive
(clayey) soils were too wet to be compacted effectively, and some sections of the embankments
remained wet and soft for many years after construction. Chapter 6 presents a more detailed
discussion of the erodeability of the levee embankments along the MRGO frontage. In simple
terms, these levees were unusually massively erodeable, and this (combined with their lack of
crest height) caused them to be unusually rapidly eroded as the storm surge from Lake Borgne
approached and passed over, and through, these levees.
Based on analytical storm surge analyses and analytical “hindcasts” performed by various
investigation teams, as well as eyewitness reports and timings of flooding and damages in St.
Bernard Parish and the Ninth Ward, it is estimated that the storm surge passed over and through
the MRGO levee frontage between approximately 6:00 to 7:00 a.m. The storm surge along the
northeastern frontage of the St. Bernard Parish protected area peaked at approximately 7:30 to
8:00 a.m. (see Figure 2.9.) By the time the storm surge peaked along this important frontage,
however, the unfinished “sand core” levees fronting Lake Borgne had been massively eroded and
the brunt of the storm surge passed over and through the levees and raced across the undeveloped
swamplands shown in Figure 2.11 towards the developed areas of St. Bernard Parish.
This is illustrated schematically in Figure 2.11. The levees along this frontage were so
badly eroded, and so rapidly, that they did little to impede the passage of the storm surge which
then crossed the roughly 7 to 10 miles of open swamp and reached the secondary levee that
separates the northern (undeveloped) swampy section of this protected area from the populated
southern section.
The secondary levee had not been intended to face the full fury of a storm surge of this
magnitude; it had been assumed that the MRGO frontage levees would absorb much of the energy
and provide more resistance. Accordingly, the storm surge passed over the secondary levee
(which had lesser typical crest heights of only + 7.5 feet to + 10 feet, MSL) and washed into the
populated regions of St. Bernard Parish. A number of minor breaches were produced by the
overtopping (and erosion) of this secondary levee, but it is interesting to note that although this
secondary levee must have been massively overtopped along much of its length, relatively little
erosion damage resulted. The secondary levee was properly constructed, using compacted clayey
soils, and the resulting levee embankment generally performed well with regard to resisting
erosion. It was not, however, tall enough to restrain the massive overtopping from the storm
surge which had passed so easily through the MRGO frontage levees.
The resulting carnage in St. Bernard Parish was devastating. A wall of water raced over
the secondary levee; pushing homes laterally (Figure 2.16), flipping cars like toys and leaving
them leaning against buildings, and driving large shrimp boats deep into the heart of residential
neighborhoods (see Chapter 6.) The flooding of St. Bernard Parish was unexpectedly rapid. The
peak depth of flooding in St. Bernard Parish was also unexpectedly deep because the floodwaters
were pushed by the still rising storm surge (rather than having to flow more slowly, over time,
through more finite breaches as the storm surge subsided; as occurred in most other parts of the
greater New Orleans area) so that the top of the floodwaters at their peak within the developed
areas were at an elevation well above mean sea level (approximately Elev. +12 feet, MSL.)
27
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Indeed, after the storm surge subsided, “notches” were excavated through a number of local
levees to let floodwaters drain under gravity loading from the significantly “plus mean sea level”
flooding entrapped in some areas.
Figure 2.12 shows a plot of the locations where dead bodies were retrieved after the
disaster as of December 2005. This map shows locations for only approximately 960 of the
approximately 1,296 official deaths (to date) in the greater New Orleans area, but this map serves
well to show the general distribution of deaths attributed to the flooding produced by this event.
As shown in Figure 2.12, approximately 30% of these deaths occurred in St. Bernard Parish. In
addition to those who perished, considerable damage was done to many thousands of homes and
businesses in this area (see Chapter 6.)
The same storm surge from Lake Borgne that topped and eroded the levees along the
“MRGO” frontage also pushed westward over the southeastern corner of the New Orleans East
protected section, as shown in Figures 2.9 through 2.11, and this produced overtopping and a
number of breaches, as shown previously in Figure 2.4. This was a principal source of the
catastrophic flooding that subsequently made its way across the local undeveloped swamplands
and into the populated areas of New Orleans East. Like the MRGO levee frontage discussed
above, large portions of this levee frontage section had been constructed using materials
excavated from the adjacent shipping channel (in this case the GIWW channel), and large
portions of the levee were comprised of highly erodeable sandy and lightweight shell sand fill.
This storm surge from Lake Borgne also passed westward into a Vshaped “funnel” as it
entered the shared GIWW/MRGO channel that separates the St. Bernard and New Orleans East
protected areas, and this in turn resulted in an elevated surge of water that passed westward along
the waterway to its juncture (at a “T”) with the IHNC channel, overtopping a number of levees
and floodwalls on both the north and south sides of this eastwest trending channel and producing
levee distress and several breaches (as shown in Figures 2.4 and 2.11.) After reaching the “T”
intersection with the IHNC channel, the surge then passed to the north and south (from the “T”)
along the IHNC channel, periodically overtopping many (but not all) of the sections of levees and
floodwalls lining the east and west sides of the IHNC, and causing a number breaches as shown
in Figures 2.4 and 2.11. By about 6:45 to 7:00 a.m. overtopping (by up to as much as 1 to 2 feet
at it’s peak at most locations) was occurring along a number of levee and floodwall sections
lining the IHNC channel. This overtopping did not occur at all locations, and was only of limited
duration (typically several hours or less) where it did occur.
A pair of major breaches occurred at the west end of the Lower Ninth Ward as this
overtopping occurred along the IHNC, and the larger of these two breaches is shown (roughly
seven weeks later, after construction of an interim repair embankment just outside the breach) in
Figure 2.13. A large barge passed in through this breach, and can be seen in the rear of the
photo. It is worth noting the tremendous scourinduced damage to the homes immediately
inboard of this massive breach; most of the homes in Figure 2.13 were washed off of their
foundations and transported laterally (often in pieces) by the inrushing floodwaters. A more
detailed examination of the two large breaches at the west end of the Ninth Ward is presented in
Chapter 6; Sections 6.4 and 6.5. The large breaches at the west end of the Lower Ninth Ward
appear to have occurred by approximately 7:45 a.m. (Louisiana State University Hurricane
Center, 2006.)
28
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Like St. Bernard Parish, the breaches at the west end of the Lower Ninth Ward occurred
before the storm surge peaked (at about 8:30 a.m. in the IHNC channel), so the Lower Ninth
Ward was flooded to a level well above mean sea level before the storm surge subsequently
subsided. This neighborhood, which had ground surface elevations of generally between about 3
to 6 feet (MSL) was flooded to elevations of up to as much as 10 to 12 feet above sea level. The
resulting carnage, in terms of both loss of life (as shown in Figure 2.12) and destruction of homes
and businesses was considerable, as the flooding rose above the tops of many of the onestory
homes in this densely packed neighborhood.
The protected area of New Orleans East, directly to the north of the St. Bernard
Parish/Ninth Ward protected area, had been breached at its southeastern corner by the initial
storm surge and lateral rush from Lake Borgne (as shown schematically in Figure 2.11) by about
6:00 to 7:00 a.m., though the resulting breaches were confined to several locations so that the
inflowing waters began to make their way across the undeveloped swamplands of the eastern
portion of this protected area and timing is thus difficult to pin down with exactitude. The storm
surge then passed laterally along the GIWW/MRGO eastwest channel and produced another
finite breach on the north side of this channel and several additional distressed sections. This
breach added to the sources of water beginning to flow into this protected area.
The surge that passed west along the GIWW/MRGO eastwest channel then pushed north
along the IHNC, and produced several additional breaches and distressed sections, of varying
severity, along the IHNC frontage as shown in Figure 2.4. These, too, added to the flow into the
protected area of New Orleans East.
The lateral storm surge that passed westward along the eastwest trending GIWW/MRGO
channel between New Orleans East and St. Bernard Parish also attacked the west side of the
IHNC channel, at the eastern edge of the main Orleans East Bank (downtown New Orleans)
protected area. This produced three additional breaches along this frontage, as shown in Figures
2.4 and 2.11. Floodwaters began to flow into the main New Orleans metropolitan (downtown)
protected area through these breaches between approximately 7:00 to 8:30 a.m. Although three
of these breaches were relatively significant, all three breaches along this frontage failed to scour
to significant depths. As a result, all three either had “lips” with lowest elevations above mean
sea level, or there were points along the path from the IHNC to the breach that were above mean
sea level. Accordingly, although all three breaches allowed some flow of water into the main
Orleans East Bank (downtown) protected area, they allowed only limited flow and this flow
stopped as the storm surge subsequently subsided. It would be the subsequent breaches in the
drainage canals, to the northwest (along the edge of Lake Pontchartrain) that would prove to be
devastating for this main (downtown) protected area.
As the hurricane then passed northwards to the east of New Orleans, the counterclockwise
direction of the storm winds also produced a wellpredicted storm surge southwards towards the
south shore of Lake Pontchartrain. The lake level rose, but mainly stayed below the crests of
most of the lakefront levees. The lake rose approximately to the tops of the lakefront levees at a
number of locations, especially along the shoreline of New Orleans East, and there was moderate
overtopping (or at least storm wave splashover) and some resulting erosion on the crests and
inboard faces of some lakefront levee sections along the Lake frontage. Significant overtopping
occurred over a long section of concrete floodwall near the west end of the New Orleans East
protected area lakefront (behind the Old Lakefront Airport), where the floodwall appears to have
29
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
been inexplicably lower than the adjacent earthen levee sections. This, too, added to the flow into
the New Orleans East protected area, which was now continuing to fill with water even as the
original storm surges subsided.
Farther to the west, the storm surge along the Pontchartrain lakefront (which peaked at
about 9:00 to 9:30 a.m. at an elevation of about +10 feet, MSL) did not produce water levels
sufficiently high as to overtop the crests of the concrete floodwalls atop the earthen levees lining
the three drainage canals that extend from just north of downtown to Lake Pontchartrain; the 17th
Street Canal, the Orleans Canal, and the London Avenue Canal. Three major breaches occurred
along these canals, however, and these produced significant flooding of large areas within the
Orleans East Bank protected area (as shown in Figure 2.4.) Figure 2.13 shows military
helicopters lowering oversized bags of gravel into the levee breach on the east side of the 17th
Street Canal, near the north end of the canal. Note that the flood waters have equilibrated, and
that there is no net flow through the breach at the time of this photo.
The first breach along the drainage canals occurred near the south end of the London
Avenue canal, between about 7:00 to 8:00 a.m. The second breach occurred near the north end of
the London Avenue canal, and the best current estimates of the timing of this breach are between
about 7:30 to 8:30 a.m. The third major breach occurred near the north end of the 17th Street
canal. The main breach here occurred between about 9:00 to 9:15 a.m., but this may have been
preceded by earlier visually observable distress at this same location. All three of these breaches
rapidly scoured to depths well below mean sea level, so they continued to transmit water into the
main Orleans East Bank (downtown) protected area after the storm surges subsided. A more
detailed discussion and analyses of these catastrophic drainage canal breaches are presented in
Chapter 8.
The resulting flooding of the main Orleans East Bank (Downtown) protected area was
catastrophic, and resulted in at least 588 of the approximately 1,293 deaths attributed (to date) to
the flooding of New Orleans by this event. Contributions to this flooding came from the
overtopping and breaches along the IHNC channel at the east side of this protected area, but the
majority of the flooding came from the three catastrophic failures along the drainage canals at the
northern portion of this protected area.
In addition, one of the drainage canals (the Orleans Canal) had not yet been fully “sealed”
at its southern end, so that floodwaters flowed freely into New Orleans during the storm surge
through this unfinished drainage canal. A section of levee and floodwall approximately 200 feet
in length had been omitted at the southern end of this drainage canal, so that despite the expense
of constructing nearly 5 miles of levees and floodwalls lining the rest of this canal, as the
floodwaters rose along the southern edge of lake Pontchartrain, the floodwaters did not rise fully
within the Orleans canal; instead they simply flowed freely into downtown New Orleans.
Chapters 4 through 8 present a more detailed discussion of the performance of the flood
protection systems nominally intended to protect the main Orleans East Bank area, and studies of
the major failures and near failures within this critical area.
By approximately 9:30 a.m. the principal levee failures had occurred, and most of New
Orleans was rapidly flooding.
2 10
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
2.3.3
Brief Comments on the Consequences of the Flooding of New Orleans
The consequences of the flooding of major portions of all four leveeprotected areas of
New Orleans were catastrophic. Approximately 85% of the metropolitan area of greater New
Orleans was flooded, as shown in Figures 2.4 and 2.5. In Figure 2.4, the flooded areas are shown
in pink, and those that remained still to be “unwatered” as late as September 28th are shown in
darker blue. The blue crosshatched areas were open, undeveloped swamplands, and these were
also flooded but were not counted in determining the 85% flooding figure.
Large developed areas within all of the four main “protected areas” were flooded, and
most remained inundated for two to three weeks before levee breaches could be repaired and the
waters fully pumped out.
Figure 2.15 shows the approximate depth of flooding that remained on September 2nd,
four days after Hurricane Katrina, in the St. Bernard Parish and Lower Ninth Ward protected
area, based on an estimated surface water elevation of approximately +5 ft. (MSL) at that time.
This is a significantly lower flood level than the estimated peak flooding to an elevation of up to
+10 to 12 feet above mean sea level during the actual hurricane. The undeveloped swampland to
the north of the populated areas can be seen in this Figure to also still be flooded on September
2nd, but the flood depths are not indicated.
Figure 2.16 shows the approximate depth of flooding that remained on September 2nd,
again four days after the hurricane, in the New Orleans East protected area. As this protected area
filled slowly during and after the hurricane, and as it was “unwatered” relatively slowly over the
days and weeks that followed, this represents nearly the full depth of flooding in this area.
Figure 2.17 shows the approximate depth of flooding of the main Orleans East Bank
(downtown) protected area on September 2nd. Like the New Orleans East protected area, this
large protected “basin” filled relatively slowly over time. By September 2nd, the breaches had not
yet all been closed by emergency repairs, so the depths of flooding in Figure 2.17 represent the
nearly the full depth of flooding at its worst in this area.
Neighborhoods that were inundated exhibit stark evidence of this catastrophic flooding.
Water marks, resembling oversized bathtub rings, line the sides of buildings and cars in these
stricken neighborhoods, as shown in Figure 2.18. Household and commercial chemicals and
solvents, as well as gasoline, mixed with the salty floodwaters in many neighborhoods, and at the
time of this investigation’s first field visits shortly after the event the paint on cars below the
watermarks on adjacent buildings had been severely damaged, and bushes and shrubs were
browned below the watermarks, but often starkly green above. Driving through neighborhoods
that had been flooded, there was often the impression that one was viewing a television screen
where the color of the picture was somehow distorted or altered below a horizontal line; the level
at which the floodwaters had been ponded. The devastation in these neighborhoods, and its
lateral extent across many miles of developed neighborhoods, was stunning even to the many
experienced members of our forensic teams that had seen numerous devastating earthquakes, tidal
waves, and other major disasters.
Close to major breaches, the hydraulic forces of the inflowing floodwaters often had
devastating effect on the communities. Figure 2.13 shows the devastation immediately inboard
2 11
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
from the large breach at the west end of the Ninth Ward site after the area had been unwatered.
Note the numerous empty slabs where homes had been stripped away and scattered, mostly in
pieces, across a large area.
Figure 2.19 shows another aspect of the flooding. This photograph shows a region within
St. Bernard Parish in which some of the homes were transported from their original locations by
the floodwaters, and then deposited in new locations. Figure 2.20 shows a number of homes in
the Plaquemines Parish polder that were carried across the narrow polder (from left to right in this
photograph) as the west side (left side of photo) “hurricane levee” or back levee was breached,
and were then deposited on the crest of the Mississippi River levee. The water side slope face of
the Mississippi River levee is clearly shown in this photograph, as evinced by the concrete slope
face protection on the outboard side of the riverfront levee in the right foreground of the figure.
Figures 2.18 through 2.25 show examples of the devastation that occurred within the
stricken flooded areas. The spray painted markings on the sides of the buildings in these areas
were left by search and rescue teams, and they denote a number of important findings within each
dwelling, including toxic contamination, etc. The most important numbers are those centered at
the base of the large “X”, as these denote the number of dead bodies found within the building.
In most cases this number was “0”, as for example in Figures 2.18 and 2.22. But this was not
always the case. Figure 2.24 shows the outside of a dwelling in the Ninth Ward with a “3”
beneath the X, indicating three deaths within. This was a housing unit, and the wheelchair ramp
from the front door is askew at the bottom of the photograph. Figure 2.25 shows the muddy
devastation, and a wheelchair, within this flooded structure.
Figure 2.26 gives another sense of perspective regarding the terrible and pervasive
devastation wreaked by the flooding of large urbanized areas. This photo shows the flooding of
an area of New Orleans East, but it could just as well be any of a number of large areas of New
Orleans. Figure 2.27 gives a similar sense of perspective. In this photo, the flooded Lower Ninth
Ward is in the foreground, and virtually every neighborhood shown (including those in the far
background behind the tall downtown buildings) is flooded, excepting only the small area
occupied by the tall buildings of the downtown area.
At the time of the writing of this report, the death toll from the flooding of New Orleans
has risen to 1,293. It is expected to continue to climb a bit higher as some of those currently
listed as “missing” will likely have been drawn out into the swamps and the Gulf by the
floodwaters. Loss projections continue to evolve, but estimates of overall losses have now
climbed to the $100 to $ 200 billion range for the metropolitan New Orleans region.
The members of this investigation team extend their hearts and their deepest condolences
to those who were devastated by Hurricane Katrina, and by the flooding of most of New Orleans.
The suffering and losses of those most intimately involved are almost beyond comprehension. It
must be the goal and objective of all of us that a catastrophe of this sort never be allowed to
happen again.
2 12
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
2.4 References
Interagency Performance Evaluation Task Force, (2006), “Performance Evaluation, Status and
Interim Results, Report 2 of a Series, Performance Evaluation of the New Orleans and
Southeast Louisiana Hurricane Protection System,” Interim Report No. 2, March 10,
2006.
Interagency Performance Evaluation Task Force, (2006), “Performance Evaluation of the New
Orleans and Southeast Louisiana Hurricane Protection System ”, Draft Final Report, June
1, 2006.
Louisiana State University Hurricane Center, (2005), “Hurricane Katrina Advisory #22,”
available from: http://hurricane.lsu.edu/floodprediction/katrina22/; date accessed:
November 10, 2005.
United States Army Corps of Engineers, (1967), “Lake Pontchartrain, LA, and Vicinity,
Chalmette Area Plan, Hurricane Protection Levee First Lift, Sta. 594+00 – Sta. 770+00
(Not Continuous),” File No. H824100, May 11, 1967.
United States Army Corps of Engineers, (1966), “Lake Pontchartrain, LA. and Vicinity,
Chalmette Area Plan, Design Memorandum No. 3, General Design,” November, 1966.
2 13
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
New Orleans
Source: http://flhurricane.com/googlemap
Figure 2.1: Location of New Orleans, and map of the path of the eye of Hurricane Katrina.
2 14
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
New Orleans
New Orleans
Source: Mashriqui, 2006
Figure 2.2: Traced path of the eye of Hurricane Katrina at landfall in the New Orleans area.
2 15
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Principal
Study Area
Lake Pontchartrain
Lake Borgne
Mississippi River
Source: ESRI North American Thematic Basemap, ArcGIS 9.0
Figure 2.3: The greater New Orleans region levee and flood protection system Study Area.
2 16
Independent Levee
Investigation Team
2 17
Figure 2.4: Map showing principal features of the main flood protection rings or “protected areas” in the New Orleans area.
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Source: Modified after USACE, 2005
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Source: Modified after USACE, 2005
Figure 2.5: Map showing the levee protected areas along the lower reaches of the
Mississippi River (in the Plaquemines Parish Area.)
2 18
2 19
Figure 2.6: Map showing design flood stage elevations throughout the New Orleans region.
Source: Graphic by Emmet Mayer III/emayer@timespicayune.com (2005)
Independent Levee
Investigation Team
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
2 20
Figure 2.7: Map showing the design flood stage levels for selected locations in the New Orleans Area.
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Source: Modified after USACE, 2005
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Source: http://hurricane.lsu.edu/floodprediction/
Figure 2.8: Calculated storm surge against the coast at about 7:30 am (CDT), August 29, 2006.
2 21
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Source: http://hurricane.lsu.edu/floodprediction/
Figure 2.9: Map of calculated storm surge levels, at time when the eye of the storm passed close to
the east of New Orleans at about 8:30 am (CDT).
2 22
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
S
Source: IPET Interim Report No. 2; April, 2006
Figure 2.10: Map showing calculated aggregate maximum storm surge levels (maximum
values at any point in time).
2 23
Independent Levee
Investigation Team
2 24
Figure 2.11: Storm surge overtopping the eastern flank of the regional flood protection system at the northeast edge of the St.
Bernard Parish and Ninth Ward protected areas.
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Source: Modified after USACE, 2005
Independent Levee
Investigation Team
2 25
Figure 2.12: Map showing locations of confirmed deaths (as of December 2005) as a result of Hurricane Katrina.
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Source: Times Picayune (2005)
2 26
Figure 2.13: Oblique view of the (south) levee break at the Inner Harbor Navigation Canal into the lower Ninth Ward.
Photograph by Les Harder
Independent Levee
Investigation Team
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photo courtesy of the U.S. Army Corps of Engineers
Figure 2.14: Initial closure of the large breach at the north end of the 17th Street Canal.
2 27
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Source: LSU Hurricane Center, 2006
Figure 2.15: Depth of flooding of New Orleans East on September 2nd (4 days after
Hurricane Katrina)
Source: LSU Hurricane Center, 2006
Figure 2.16: Depth of flooding of St. Bernard Parish and the Lower Ninth Ward on Sept.
2nd (4 days after Hurricane Katrina).
2 28
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Source: LSU Hurricane Center, 2006
Figure 2.17: Depth of flooding of the Orleans East Bank (Downtown) protected area on
September 2nd (4 days after Hurricane Katrina).
2 29
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph by Rune Storesund
Figure 2.18: High water marks remain on structures after temporary levee repairs
have been completed and flood waters have been pumped out.
Photograph by Les Harder
Figure 2.19: Flooded neighborhood in St. Bernard Parish, showing homes floated off
their foundations and transported by floodwaters.
2 30
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
.
Photograph by Les Harder
Figure 2.20: Homes in Plaquemines Parish carried from left to right in photo and strewn
across the crown of the Mississippi Riverfront levee.
Photograph by Rune Storesund
Figure 2.21: Damage to a residential neighborhood in the 17th Street Canal
area due to flooding.
2 31
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph by Rune Storesund
Figure 2.22: Search and rescue markings on a residence in the Canal District.
Photograph by Rune Storesund
Figure 2.23: Another view of flooding damage in the Canal District.
2 32
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph by Les Harder
Figure 2.24: Search and rescue team markings on a building in the lower Ninth
Ward where three inhabitants died.
Photograph by Les Harder
Figure 2.25: View inside structure shown previously in Figure 2.21.
2 33
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photo Courtesy of http://www.wwltv.com/sharedcontent/breakingnews/slideshow/083005_dmnkatrina/7.html
Figure 2.26: Neighborhood in New Orleans East fully flooded.
Photo courtesy of http://www.wwltv.com/sharedcontent/breakingnews/slideshow/083005_dmnkatrina/7.html
Figure 2.27: View of the City of New Orleans at the peak of the flooding.
2 34
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
CHAPTER THREE: GEOLOGY OF THE NEW ORLEANS REGION
3.1
General Overview of the Geology of New Orleans
3.1.1 Introduction
Hurricane Katrina brought devastation to New Orleans and the surrounding Gulf
Coast Region during late August 2005. Although there was wind damage in New Orleans,
most of the devastation was caused by flooding after the levee system adjacent to Lake
Pontchartrain, Lake Borgne and Inner Harbor areas of the city systematically failed. The
storm surge fed by winds from Hurricane Katrina moved into Lake Pontchartrain from the
Gulf of Mexico through Lake Borgne, backing up water into the drainage and navigation
canals serving New Orleans. The storm surge overwhelmed levees surrounding these
engineered works, flooding approximately 80% of New Orleans.
Although some levees/levee walls were overtopped by the storm surge, the London
Avenue and 17th Street drainage canal walls were not overtopped. They appear to have
suffered foundation failures when water rose no higher than about 4 to 5 feet below the crest
of the flood walls. This occurrence has led investigators to carefully investigate and
characterize the foundation conditions beneath the levees that failed. A partnership between
the U.S. Geological Survey’s MidContinent Geologic Science Center and the University of
Missouri – Rolla, both located in Rolla, MO, was established in the days immediately after
the disaster to make a field reconnaissance to record perishable data. This engineering
geology team was subsequently absorbed into the forensic investigation team from the
University of California, Berkeley, funded by the National Science Foundation.
The team has taken multiple trips to the devastated areas. During these trips team
members collected physical data on the levee failures, much of which was subsequently
destroyed or covered by emergency repair operations on the levees. Our team also logged a
series of subsurface exploratory borings to characterize the geological conditions present in
and around the levee failure sites.
3.1.2 Evolution of the Mississippi Delta beneath New Orleans
The Mississippi River drains approximately 41% of the Continental United States, a
land area of 1.2 million mi2 (3.2 million km2). The great majority of its bed load is deposited
as subaerial sediment on a well developed flood plain upstream of Baton Rouge, as opposed
to subaqueous deposits in the Gulf of Mexico. The Mississippi Delta has been lain down by
an intricate system of distributary channels; that periodically overflow into shallow swamps
and marshes lying between the channels (Figure 3.1, upper). The modern delta extends more
or less from the presentday position of Baton Rouge (on the Mississippi River) and Krotz
Springs (on the Atchafalaya River). The major depositional lobes are shown in Figure 3.1
(lower).
Between 12,000 to 6,000 years ago sea level rose dramatically as the climate changed
and became warmer, entering the present interglacial period, which geologists term the
31
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Holocene Epoch (last 11,000 years). During this interim, sea level rose approximately 350
feet, causing the Gulf of Mexico to retreat into southeastern Louisiana inundating vast tracts
of coastline. By 7,000 years ago sea level had risen to within about 30 feet of its present
level. By 6,000 years ago the Gulf had risen to within 10 to 15 feet of its present level.
The modern Mississippi Delta is a system of distributary channels that have deposited
large quantities of sediment over the past 6,000 to 7,000 years (Figure 3.1 –upper). Six major
depositional lobes, or coalescing zones of deposition, have been identified, as presented in
Figure 3.1 (lower). In southeastern Louisiana deltaic sedimentation did not begin until just
the last 5,000 years (Saucier, 1994). Four of these emanate from the modern Mississippi
River and two from the Atchafalaya River, where the sediments reach their greatest thickness.
The St. Bernard Delta extending beneath Lake Borgne, Chandeleur and Breton Sounds to the
Chandeleur and Breton Shoals was likely deposited between 600 and 4,700 years ago. The
50+ miles of the modern PlaqueminesBalize Delta downstream of New Orleans has all been
deposited in just the last 800 to 1,000 years (Darut et al. (2005).
During this same period (last 7,000 years) the Mississippi River has advanced its
mouth approximately 200 river miles into the Gulf of Mexico. The emplacement of jetties at
the river’s mouth in the late 1870’s served to accelerate the seaward extension of the main
distributary passes (utilized as shipping channels) to an average advance of about 70 meters
per year, or about six times the historic rate (Coleman, 1988; Gould, 1970). The combination
of channel extension and sea level rise has served to flatten the grade of the river and its
adjoining flood plains, diminishing the mean grain size of the river’s bed load, causing it to
deposit increasing fine grained sediments. Channel sands are laterally restricted to the main
stem channel of the Mississippi River, or major distributary channels, or “passes”, like the
MetairieGentilly Ridge. The vast majority of the coastal lowland is infilled with silt, clay,
peat, and organic matter.
Geologic sections through the Mississippi Embayment show that an enormous
thickness of sediment has been deposited in southern Louisiana (Figure 3.2). During the
Quaternary Period, or Ice Ages, (11,000 to 1.6 million years ago) the proto Mississippi River
conveyed a significantly greater volume of water on a much steeper hydraulic grade. This
allowed large quantities of graveliferous deposits beneath what is now New Orleans, reaching
thicknesses of up to 3600 feet (Figures 3.2 and 3.3). These stiff undifferentiated Pleistocene
sands and gravels generally lie between 40 and 150 feet beneath New Orleans, and much
shallower beneath Lake Pontchartrain and Lake Borgne (as one approaches the Pleistocene
outcrop along the North Shore of Lake Pontchartrain).
Just south of the Louisiana coast, the Mississippi River sediments reach thicknesses of
30,000 feet or more. The enormous weight of this sediment mass has caused the earth’s crust
to sag in this area, resulting in a structure known as the Gulf Geosyncline (Figure 3.2). Flow
of mantle material from below the Gulf Geosyncline is causing an uplift along about the
latitude of Wiggins, MS. This is one cause of subsidence in South Louisiana (discussed in
Section 3.7.2).
Figure 3.4 presents a generalized geologic map of the New Orleans area, highlighting
the salient depositional features. Depth contours on the upper Pleistocene age (late Wisconsin
glacial stage) horizons are shown in red. Sea level was about 100 feet lower than present
32
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
about 9000 years ago, so the 100 ft contour represents the approximate shoreline of the Gulf
at that time, just south of the current Mississippi River channel. Figure 3.5 presents a more
detailed view of the dissected late Wisconsin stage erosional surface beneath New Orleans.
This system emanates from the Lake Pontchartrain depression and reaches depths of 150 feet
below sea level where it is truncated by the modern channel of the Mississippi River, which is
not as deeply incised. A veneer of interdistributary deltaic deposits covers this older surface
and is widely recognized for having spawned differential settlement of the cover materials
where variations in thickness are severe, such as the Garden District.
3.1.3 Pine Island Beach Trend
Relict beach deposits emanating from the Pearl River are shown in stippled yellow on
Figure 3.4. Saucier (1963) named these relic beaches the Pine Island and Miltons Island
beach trends. These sands emanate from the Pearl River between Louisiana and Mississippi,
to the northeast. The Miltons Island Beach Trend lies beneath the north shore of Lake
Pontchartrain, while the Pine Island Beach Trend runs northeasterly, beneath the Lakeview
and Gentilly neighborhoods of New Orleans up to the Rogolets. The Pine Island Beach Trend
is believed to have been deposited when sea level had almost risen to its present level, about
4500 years ago. At that juncture, the rate of sea level rise began to slow and there was an
unusually large amount of sand being deposited near the ancient shoreline by the Pearl River,
which was spread westerly by longshore drift, in a long linear sand shoal, which soon
emerged into a beach ridge along a northeastsouthwest trend (Saucier, 1963). The
subsequent development of accretion ridges indicate that shoreline retreat halted and the
beach prograded southwestward, into what is now the Gentilly and Lakeview areas. By
about 5,000 years ago, the beach has risen sufficiently to form a true barrier spit anchored to
the mainland near the present Rigolets, with a large lagoon forming on its northern side (what
is now Lake Pontchartrain, which occupies an area of 635 mi2).
Sometime after this spit formed, distributaries of the Mississippi River (shown as
yellow bands on Figure 3.4) began depositing deltaic sediments seaward of the beach trend,
isolating it from the Gulf of Mexico. The Pine Island Beach Trend was subsequently
surrounded and buried by sediment and the Pine Island sands have subsided 25 to 45 feet over
the past 5,000 years (assuming it once stood 5 to 10 feet above sea level). The distribution of
the Pine Island Beach Trend across lower New Orleans is shown in Figure 3.6. The Pine
Island sands reach thicknesses of more than 40 feet in the Gentilly area, but diminish towards
the Lakeview area, pinching out near the New Orleans/Jefferson Parish boundary (close to the
17th Street Canal breach). The Pine Island beach sands created a natural border that helped
form the southern shoreline of Lake Pontchartrain, along with deposition by the Mississippi
River near its present course. Lake Pontchartrain was not sealed off entirely until about 3,000
years ago, by deposition in the St. Bernard’s Deltaic lobe (Kolb, Smith, and Silva, 1975).
The Pine Island Beach Trend peters out beneath Jefferson Parish, as shown in Figures 3.4 and
3.6.
3.1.4
Interdistributary Zones
Most of New Orleans’ residential areas lie within what is called an interdistributary
zone, underlain by lacustrine, swamp, and marsh deposits, shown schematically in Figure 3.7.
33
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
This low lying area rests on a relatively thin deltaic plain, filled with marsh, swamp, and
lacustrine sediments. The drainage canals were originally constructed between 183378 on
interdistributary embayments, which are underlain by fat clays deposited in a quiet water, or
paludal, environment (Kolb and Van Lopik, 1958).
Interdistributary sediments are deposited in low lying areas between modern
distributory channels and old deltas of the Mississippi River, shown schematically in Figure
3.8. The low angle bifurcation of distributary streams promotes troughlike deposits that
widen towards the gulfward. Sediment charged water spilling over natural channel levees
tends to drop its coarse sediment closest to the channel (e.g. Metairie and Gentilly Ridges)
while the finest sediment settles out in shallow basins between the distrubutaries. Fine
grained sediment can also be carried into the interdistributary basins through crevassesplays
well upstream, which find their way into low lying areas downstream. Storms can blow
sedimentladen waters back upstream into basins, while hurricanes can dump sedimentladen
waters onshore, though these may be deposited in a temporarily brackish environment.
Considerable thickness of interdistributary clays can be deposited as the delta builds
seaward. Kolb and Van Lopik (1958) noted that interdistributary clays often grade downward
into prodelta clays and upward into richly organic clays of swamp or marsh deposits. The
demarcation between clays deposited in these respective environments is often indistinct.
True swamp or marsh deposits only initiate when the water depth shallows sufficiently to
support vegetation (e.g. cypress swamp or grassy marsh). The interdistributary zone is
typified by organic clays, with about 60% by volume being inorganic fat clays, and 10% or
less being silt (usually in thin, hardly discernable stringers). Kolb and Van Lopik (1958)
reported cohesive strengths of interdistributary clays as ordinarily being something between
100 and 400 psf. These strengths, of course, depend also on the past effective overburden
pressure.
Careful logging is required to identify the depositional boundary between
interdistributary (marsh and swamp) and prodelta clays (Figure 3.9). The silt and fine sand
fractions in interdistributary materials are usually paperthin partings. Prodelta clays are
typified by a massive, homogeneous appearance with no visible planes or partings.
Geologically recent interdistributary clays, like those in lower New Orleans, also tend to
exhibit underconsolidation, because they were deposited so recently. Interdistributary clays in
vicinity of South Pass (45 miles downstream of New Orleans) exhibit little increase in
strengths to depths of as much as 375 ft. This is because these materials were deposited
rapidly, during the past 600 to 1,000 years, and insufficient time has passed to allow for
normal consolidation, given the low drainage characteristics of the units. This phenomenon
was noted and analyzed for offshore clays by Terzaghi (1956). The older prodelta clays
underlying recent interdistributary clays tend to exhibit almost linear increase of density and
strength with depth, because these materials were deposited very slowly.
So, the
environment of deposition greatly impacts soil strength.
3.1.5
Paludal environments
Paludal environments on the Mississippi River deltaic plain are characterized by
organic to highly organic sediments deposited in swamps and marshes. Paludal environments
34
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
are typified halfland and halfwater, with water depths seldom exceeding two feet above
mean gulf level. 90% of New Orleans is covered by swamp or marsh deposits (excluding
filled areas). Lacustrine (lake) and tidal channel deposits can be complexly intermingled with
swamp and marsh deposits.
3.1.5.1 Marshes
More than half of the New Orleans area was once covered by marshes, essentially flat
areas where the only vegetation is grasses and sedges. Tufts of marsh grass often grow with
mud or open water between them. When these expanses are dry, locals often refer to them as
“prairies.” As the marshes subside, grasses become increasingly sensitive to increasing
salinity. As grasses requiring fresh water die out, these zones transition into a myriad of
small lakes, eventually becoming connected to an intricate network of intertidal channels that
rise and fall with diurnal tides. These are often noted on older maps as “brackish” or “sea
marshes” to discern them from adjoining fresh water swamps and marshes (Figure 3.9).
Marsh deposits in New Orleans are typically comprised of organic materials in
varying degrees of decomposition. These include peats, organic oozes, and humus formed as
marsh plants die and are covered by water. Because the land is sinking, subaerial oxidation is
limited, decay being largely fomented by anerobic bacteria. In stagnant water thick deposits
consisting almost entirely of organic debris are commonplace. The low relative density of
these materials and flooded nature provides insufficient effective stress to cause
consolidation. As a consequence, the coastal marsh surface tends to “build down,” as new
vegetation springs up each year at a nearconstant elevation, while the land continues to
subside. In areas bereft of inorganic sediment, thick sequences of organic peat will
accumulate, with low relative density. If the vegetation cannot keep pace with subsidence,
marine waters will inundate the coastal marsh zone, as noted in the 1849 map in Figure 3.10.
Peats are the most common variety of marsh deposits in New Orleans. They usually
consist of brown to black fibrous or felty masses of partially decomposed vegetative matter.
Materials noted on many of the older boring logs as “muck” or “swamp muck” are usually
detrital organic particles transported by marsh drainage or decomposed vegetative matter.
These mucks are watery oozes that exhibit very low shear strength and cannot support any
appreciable weight.
Inorganic sediments may also accumulate in marshes, depending on the nearness of a
sediment source(s). Common examples are sedimentladen marine waters and muddy
fluvatile waters. Brackish marsh deposits interfinger with fresh water deposits along the
southern shore of Lake Pontchartrain, but dominate the shoreline around Lake Borgne.
Floating marsh materials underlie much of the zone along old watercourses, like Bayou St.
John and Bayou des Chapitoulas. Kolb and Van Lopik (1958) delineated four principal types
of marsh deposits in New Orleans:
1.
Fresh water marsh consists of a vegetative mat underlain by clays and organic
clays. Fresh water marshes generally form as a band along the landward border of established
marshes and in those areas repeatedly subjected to fresh water inundation. In most instances
an upper mat of roots and plant parts at least 12 inches thick overlies fairly soft organic clays,
35
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
which become firmer and less organic with depth. Peat layers are often discontinuous and
their organic content is usually between 20 and 50%.
2. Floating marsh or flotant is a vegetative mat underlain by organic ooze. This is
sometimes referred to as a “floating fresh marsh” or “floating threecornered grass marsh.”
The vegetative mat is typically between 4 and 14 inches thick, floating on 3 to 15 ft of finely
divided muck or organic ooze, grading into clay with depth. The ooze often consolidates with
depth and grades into a black organic clay or peat layer.
3. Brackishfresh water marsh sequence consists of a vegetative mat underlain by peat.
The upper mat of roots and recent marsh vegetation is typically 4 to 8 inches thick and
underlain by 1 to 10 ft of coarse to medium textured fibrous peat. This layer is often
underlain by a fairly firm, bluegrey clay and silty clay with thick lenses of dark grey clays
and silty clays with high organic contents. The great majority of marsh deposits in New
Orleans are of this type, with a very high peat and humus content, easily revealed by
gravimetric water content and/or dry bulk density values.
4. Salinebrackish water marsh is identified by a vegetative mat underlain by clays.
These are sometimes termed “drained salt marshes” on older maps. The typical sequence
consists of a mat of roots, stems, and leaves from 2 to 8 inches thick, underlain by a fairly
firm bluegrey clay containing roots and plant parts. Tiny organic flakes and particles are
disseminated through the clay horizon. The clays tend to become less organic and firmer with
depth. The saline to brackish water marsh occupies a belt ½ to 8 miles wide flanking the
present day shoreline, along the coast.
The strengths of marsh deposits are generally quite low, depending on their water
content. Embankments have been placed on vegetative mats underlain by ooze, supporting
as much as 2 or 3 psi of loading, provided it is uniformly applied over reasonable distances,
carefully (Kolb and Van Lopik, 1958). Field observations of sloped levees founded on such
materials indicate failure at heights of around 6 feet, which exert pressures close to those cited
above.
3.1.5.2 Swamps
Before development, swamps in the New Orleans were easily distinguishable from
marshes because of the dense growth of cypress trees. All of the pre1900 maps make
reference to extensive cypress marshes in lower New Orleans, between the French Quarter
and Lake Pontchartrain (Figure 3.11). Encountering cypress wood in boreholes or excavations
is generally indicative of a swamp environment. These cypress swamps thrived in 2 to 6 feet
of water, but cannot regenerate unless new influx of sediment is deposited in the swamp,
reducing the water depth. Brackish water intrusion can also cause flocculation of clay and
premature die out of the cypress trees.
Two layers of cypress swamp deposits are recognized to extend over large tracts of
New Orleans (WPALA, 1937). The upper layer is the historic swamp occupying the original
ground surface where infilling has occurred since the founding of the city in 1718; and the
second; is a pervasive layer of cypress tree stumps that lies 20 to 30 feet below the ground
surface, around 25 ft MGL (Mean Gulf Level). This older cypress forest was undoubtedly
36
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
killed off and buried in a significant prehistoric flood event, fomented by considerable
deposition of inorganic sediment. This sudden influx of sediment may have come from a
crevassesplay along the Mississippi River upstream of New Orleans, as in most of the
damaging floods that befell the city prior to 1849.
There are two principal types of swamps in the New Orleans area, inland swamps and
mangrove swamps. Inland swamps typically occupy poorly drained areas enclosed by higher
ground; either natural levee ridges (like Metairie Ridge) or, much older (Pleistocene age)
Prairie Terraces. These basins receive fresh water from overflow of adjacent channels during
late spring and early summer runoff. The trees growing in inland swamps are very sensitive
to increases in salinity, even for shortlived periods. Continued subsidence allows eventual
encroachment of saline water, gradually transforming the swamp to a grassy marsh. The
relative age of the tree dieoff is readily seen in the form of countless dead tree trunks,
followed by stumps, which become buried in the marsh that supersedes the swamp. As a
consequence, a thin veneer of marsh deposits often overlies extensive sequences of woody
swamp deposits. The converse is true in areas experiencing high levels of sedimentation,
such as those along the historic Mississippi and Atchafalaya River channels, where old
brackish water marshes are buried by more recent fresh water swamp deposits. Swamp
deposits typically contain logs, stumps, and arboreal root systems, which are highly
permeable and conductive to seepage.
Mangrove swamps are the variety that thrives in salt water, with the two principal
varieties being black and honey mangrove. Mangrove swamps are found along the distal
islands of the Mississippi Delta, such as Timbalier, Freemason North, and the Chandeleur
Islands, well offshore. Mangrove swamps also fringe the St. Bernard Marsh, Breton and
Chandeleur Sounds, often rooting themselves on submerged natural levees. Mangrove
swamps can reach heights of 20 to 25 feet in Plaquemines Parish. A typical soil column in a
mangrove swamp consists of a thin layer of soft black organic silty clay with interlocking root
zone that averages 5 to 12 inches thick. Tubelike roots usually extend a few inches above the
ground surface. Thicknesses of five feet or more are common. Where they grow on sandy
barrier beaches, the mangrove swamps thrive on the leeward side, where silts and clays
intermingle with washover sands off the windward side, usually mixed with shells.
Surficial swamp deposits provide the least favorable foundations for structures and
manmade improvements, like streets and buried utilities. Kolb and Saucier (1982) noted
that the amount of structural damage in New Orleans was almost directly proportional to the
thickness of surficial organic deposits (swamps and marshes). This peaty surface layer
reaches thicknesses of up to 16 ft, as shown in Figure 3.12. Most of this foundation distress is
attributable to differential settlement engendered by recent dewatering (discussed in Section
3.7.4).
3.1.5.3 Lacustrine Deposits
Lacustrine deposits are also deposited in a paludal environment of deltaic plains. This
sequence most often occurs as marshes deteriorate (from lack of sediment) or subside (or
both). These lakes vary in size, from a few feet in diameter to the largest, Lake Salvador (a
few miles southwest of New Orleans), which measures 6 by 13 miles. Lake Pontchartrain (25
37
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
x 40 miles) is much larger, but is not a true marshland lake. The depths of these lakes vary
from as little as 1.5 feet to about 8 feet (Lake Pontchartrain and Lake Borgne average 15 and
10 feet deep, respectively).
Small inland lakes within the marsh environment usually evolve from subsidence and
erosion from wind shear and hurricane tides. Waves set up a winnowing action which
concentrates the coarser material into the deepest portion of the lake. These lakes are
generally quite shallow, often only a foot or two deep, even though up to a mile long. They
are simply waterfilled depressions on the underlying marsh, often identified in sampling by
fine grained oozes overlying peats and organic clays of the marsh that preceded the transition
to lake. The ooze become increasingly cohesive with age and depth, but is generally
restricted to only 1 to 3 feet in thickness in small inland lakes.
Transitional lakes are those that become larger and more numerous closer to the
actively retreating shoreline of the delta. These lake waters are free to move with the tides
and currents affecting the open water of adjacent bays and sounds. Fines are often winnowed
from the beds of these lakes and moved seaward, leaving behind silts and fine sands.
Sediments in these lakes are transitional between inland lakes and the largely inorganic silty
and sandy materials flooring bays and sounds.
Large inland lakes are the only lacustrine bodies where significant volumes of
sediment are deposited. Principal examples would be the western side of Lake Borgne, Lake
Pontchartrain, and Lake Maurepas, among others. Lacustrine clays form a significant portion
of the upper 20 to 30 feet of the deltaic plain surrounding New Orleans. Lake Pontchartrain
appears to have been a marine water body prior to the deposition of the Metairie Ridge
distributary channel, which formed its southern shoreline, sealing it off from the Gulf. The
central and western floor of Lake Pontchartrain is covered by clays, but the northern, eastern
and southern shores are covered by silts and sands, likely due to the choppy waveagitated
floor of the shallow lake. Deeper in the sediment sequence oyster shells are encountered,
testifying that saline conditions once existed when the lake was open to the ocean. The
dominant type of mollusk within Lake Pontchartrain today is the clam Rangia cuneata, which
favors brackish water. Dredging for shells was common in Lake Pontchartrain until the late
1970’s.
During Hurricanes Katrina and Rita in 2005, wind shear removed extensive tracts of
marsh cover, creating 118 square miles of new water surface in the delta. Fortyone square
miles of shearexpanded pools were added to the Breton Sound Basin within Plaquemines
Parish. This was more erosion and land loss than had occurred during the previous 50 years
combined (Map USGSNWRC 2006110049).
3.1.6 Recognition Keys for Depositional Environments
Marsh deposits are typified by fibrous peats; from three principal environments: 1)
fresh water marshes; 2) floating marsh – roots and grass sitting on an ooze of fresh water; and
3) saltwater marshes along the coast. The New Orleans marsh tends to be grassy marsh on a
flat area that is “building down,” underlain by soft organic clays. Low strength smectite clays
38
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
tend to flocculate during brackish water intrusions, most commonly triggered by hurricanes
making landfall in the proximate area.
Typical recognition keys for depositional environments have been summarized as follows.
•
•
•
•
•
•
Cypress wood = fresh water swamp
Fibrous peaty materials = marshes
Fat Clays with organics; usually lacustrine. A pure fat clay has high water content
(w/c) and consistency of peanut butter
Interdistributary clays; paludal environments; lakes Silt lenses when water is shallow
and influenced by wind swept waves
Lean clays CL Liquid Limit (LL) <50, silty and w/c <60%
Fat clays CH Liquid Limit (LL) >50, no silt and w/c >70%
Abandoned meanders result in complex mixtures of channel sands, fat clay, lean clay,
fibrous peat, and cypress swamp materials, which can be nearly impossible to correlate
linearly between boreholes. The New Orleans District of the Corps of Engineers has
historically employed 3inch diameter Steel Shelby tubes and 5inch diameter piston sampler,
referring to samples recovered from the 5inch sampler as their “undisturbed samples.” These
are useful for characterizing the depositional environment of the soils. The larger diameter
“undisturbed” samples are usually identified on boring logs and cross sections in the New
Orleans District Design Memoranda by the modifier “U” for “undisturbed” samples (e.g.
Boring prefixes XU, UMPX, MUEX, MUGX, and MUWX).
3.1.7 Holocene Geology of New Orleans
The surficial geology of the New Orleans area is shown in Figure 3.13. The
Mississippi River levees form the high ground, underlain by sands (shown as bright yellow in
Figure 3.13). The old cypress swamps (shown in green) and grassy marshlands (shown in
brown) occupied the low lying areas. The Midtown area between the Mississippi and
Metairie Ridge was an enclosed depression (shown in green) known as a “levee flank
depression” (Russell, 1967). The much older Pleistocene age Prairie formation (shown in
ochre) lies north of Lake Pontchartrain. This unit dips down beneath the city and is generally
encountered at depths greater than 40 feet between the city (described previously).
The levee backslope and former swamplands north of Metairie Ridge are underlain by
four principal stratigraphic units, shown in Figure 3.14. The surface is covered by a thin
veneer of recent fill, generally a few inches to several few feet thick, depending on location.
This is underlain by peaty swamp and marsh deposits, which are highly organic and
susceptible to consolidation. Entire cypress trunks are commonly encountered in exploratory
borings, as shown in Figure 3.15. This unit contains two levels of old cypress swamps,
discussed previously, and varies between 10 and 40 feet thick, depending on location. The
clayey material beneath this is comprised of interdistributary materials deposited in a paludal
(quiet water) environment, dominated by clay, but with frequent clay stringers. This unit
pinches out in vicinity of the London Avenue Canal and increases in thickness to about 15
feet beneath the 17th Street Canal, three miles west. Occasional discontinuous lenses of pure
clay are often encountered which formed through flocculation of the clay platelets when the
swamp was inundated by salt water during severe hurricanes.
39
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
The area east of the Inner Harbor Navigation Canal (IHNC) is quite different (Figure
3.14), in that these deposits are dominated by finegrained lacustrine deposits deposited in
proto Lake Pontchartrain, and the Pine Island Sands are missing. These lacustrine materials
extend eastward and are characterized by clays and silty clays with intermittent silt lenses and
organics.
The lacustrine facies is underlain by the distinctive Pine Island Beach Sand, described
previously. These relict beach sands thicken towards the east, closer to its depositional
source. They reach a maximum thickness of about 30 ft. It thins westward towards Jefferson
Parish, where it is only about 10 feet thick beneath the 17th Street Canal, as shown in Figure
3.14. The Pine Island sands are easily identified by the presence of mica in the quartz sand,
and were likely transported from the mouth of the Pearl River by longshore drift (Saucier,
1963). Broken shells are common throughout the entire layer.
A bay sound deposit consisting of fine lacustrine clays begins just east of the Inner
Harbor Navigation Canal; it begins near the 40 foot depth, has about a 10 foot thickness and
continues to the west across the city, thickening along the way (Figure 3.14). It reaches its
greatest thickness of about 35 feet just east of the 17th Street Canal. It is interesting to note
that this area has experienced the greatest recorded settlement in the city, which may be
attributable to dewatering of the units above this compressible lacustrine clay, increasing the
effective stress acting on these materials (areas to the east are underlain by much more sand,
which is less compressible).
The Holocene age deposits reach their greatest thickness just east of the 17th Street
Drainage Canal where they are 80 feet thick (Figure 3.5). Undifferentiated Pleistocene
deposits lie below these younger deposits.
For the most part, this area sits below sea level with the exception of the areas along
old channels and natural levees. The MetairieGentilly Ridge lies above the adjacent portions
of the city because it was an old distributary channel of the Mississippi River (Figure 3.1upper). The same is true for the French Quarter and Downtown New Orleans, which are built
on the natural sand levee of the Mississippi River.
Geology from the Inner Harbor Navigation Canal to the east becomes exceedingly
complex. Although the surficial 10 feet consist of materials from an old cypress swamp, this
is an area dominated by the Mississippi River and its distributaries, especially the old St.
Bernard delta (See Figure 3.1lower). Distributaries are common throughout the area and
consist of sandy channels flanked by natural levees. 1015 feet of interdistributary materials,
mainly fine organic materials, are present between distributaries. Relic beaches varying in
thickness from 10 to 15 feet are present below the interdistributary deposits. These beaches
rest atop a 510 foot thick layer of nearshore deposits which are then followed by a thick
sequence of prodelta clays leading out into the Gulf of Mexico.
3.1.8
Faulting and Seismic Conditions
Subsidence of the Gulf Geosyncline has led to numerous “growth” faults in South
Louisiana. One group, the Baton Rouge Fault Zone (shown in Figure 3.7), is currently active
3 10
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
and passes in an eastwest direction along the north shore of Lake Pontchartrain. Localized
faulting is also common near salt domes. There has been no known faulting in the New
Orleans area which has been active in Holocene times. The area is seismically quiescent. The
earthquake acceleration with a 10% chance of being exceeded once in 250 years is about
0.04g.
3.2
Geologic Conditions at 17th Street Canal Breach
3.2.1 Introduction
The 17th St. Canal levee (floodwall) breach is one of New Orleans’ more interesting
levee failures. It is one of several levees that did not experience overtopping. Instead, it
translated laterally approximately 50 feet atop weak foundation materials consisting of
organicrich marsh and swamp deposits. Trees, fences, and other features on or near the levee
moved horizontally but experienced very little rotation, indicating the failure was almost
purely translational in nature.
3.2.2 Interpretation of Geology from Auger Borings
A series of continuously sampled borings was conducted and logged using 3inch
Shelby tubes in the vicinity of the 17th St. Outlet Canal levee failure on 212006 (east side)
and 272006 (west side) to characterize the geology of the materials serving as a foundation
for the levee embankments and floodwalls. Drilling on the east bank took place just behind
(east) of an intact portion of the levee embankment that had translated nearly 50 feet while
drilling on the west side took place directly across the canal from the middle of the eastern
breach. This drilling uncovered a wide range of materials below the embankments and
provided insights into the failure.
Drilling on the east side of the levee was started at approximately 23 feet above sea
level. A thin layer of crushed rock fill placed by contractors working for the U.S. Army
Corps of Engineers to provide a working surface at the break site was augered through before
reaching the native materials. Upon drilling at the east side of the levee, organic matter was
encountered almost immediately and a fetid swamp gas odor was noted. This organic matter
consisted of lowdensity peat, humus, and wood fragments intermixed with fine sand, silt, and
clay, possibly due to wind shear and wave action from prehistoric hurricanes. This area
appears to have been near the distal margins of a historic slough, as shown in Figure 3.16. At
46 feet, highly permeable marsh deposits were encountered and drilling fluid began flowing
from a CPT hole several feet away, indicative of almost instantaneous conductivity at this
depth. The CPT was sealed with bentonite before proceeding to prevent further fluid loss.
The bottom of this sample was recovered as a solid 3inch core of orangered cypress wood
indicating that this boring had passed through a trunk of stump of a former, but geologically
young, tree.
A suspected slide plane was discovered at a depth between 8.3 and 11 feet below the
ground surface depending on the location of the borings, indicative of an undulating slip
surface. Gray plastic clays appeared to have been mixed with dark organics by shearing and
3 11
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
this zone was extremely mushy and almost soupy in texture. The water content was very
high, on the order of 278%.
Organic rich deposits continue to a depth of about 20 feet below the surface while
showing an increasing clay and silt content. Most clays are highly plastic with a high water
content although there are lenses of lower plasticity clay, silt, and some sand. The variability
of grain sizes and other materials is likely due to materials churned up by prehistoric storms.
The clays are usually gray in color but vary and are olive, brown, dark gray, and black
depending on the type and amount of organic content. Some organic matter towards the base
of this deposit was likely roots that grew down through the preexisting clays and silts or tree
debris and that were mixed by prior hurricanes. Some woody debris came up relatively free
of clays and closely resembled cypress mulch sold commercially for landscaping purposes.
Full recoveries of material in this zone were rarely achieved in this organic rich zone. It
appears that the lowdensity nature (less than water) of these soils caused them to compress
due to sampling disturbances.
Most material below 21 feet was gray plastic clay varying from soft to firm and nearly
pure lacustrine in origin. This clay included many silt lenses which tended to be stiffer and
had some organics at 26 feet. It is likely that the silt and organics were washed into an
otherwise quiet prehistoric Lake Pontchartrain by storms.
Sand and broken shells showed up at 30 feet in depth and continued to increase in
quantity and size until 35.5 feet when the material became dirty sand with very little cohesion.
This hole was terminated at 36 feet. These sands appear to be the Pine Island Beach Trend
deposits, described in Section 3.1.3.
The geologic conditions beneath the 17th Street Canal breach are shown in Figures
3.17 thru 3.20. Figure 3.17 shows the relative positions of the cross sections presented in
Figures 3.18 and 3.19. Figure 3.18 is a geologic section through the 17th Street Canal breach,
extending into the canal. It was constructed using Brunton Compass and tape techniques
commonly employed in engineering geology (Compton, 1962). In this section the landside of
the eastern levee embankment translated laterally about 48 feet. The levee had two
identifiable fill horizons, separated by a thin layer of shells, likely used to pave the old levee
crest or the road next to the levee prior to 1915 (similar to the conditions depicted in Figure
4.18). A distinctive basal rupture surface was encountered in al the exploratory borings, as
depicted in Figures 3.18 and 3.19. This rupture surface was characterized by the abrupt
truncation of organic materials, including cypress branches up to two inches in diameter
(shown in the inset of Figure 3.18). The rupture surface was between ¾ and 1 inch thick, and
generally exhibited a very high water content (measured as 279% in samples recovered and
tested). This material had a liquid consistency with zero appreciable shear strength. It could
only be sampled within more competent materials in the Shelby Tubes. A brecciated zone
three to four inches thick was observed in samples immediately above the rupture surface.
This contained chunks of clay with contrasting color to the matrix materials, and up to several
inches across, along with severed organic materials.
The geologic cross section portrayed in Figure 3.19 was taken on the north side of the
same lot, using the same Brunton Compass and tape technique. It was located between 80 and
100 feet north of the previous section described above, as shown in Figure 3.17. In this
3 12
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
location the landside of the levee embankment translated about 52 feet laterally, to the east.
These offsets were based on tape measurement made from the chain link rightofway fences
along the levee crest. No less than four distinct thrust planes were identified in the field,
suggesting a planar, translational failure mode, as sketched in the cross section. As with the
previous section, the old swamp deposits are noticeably compressed beneath the levee
embankment, likely due to fill surcharge and the fact that the drainage canals have never been
drained over their lifetime (in this case, since 1858 or thereabouts, described in Section 4.6).
This local differential settlement causes the contact between the swamp deposits and the
underlying lacustrine clays to dip northerly, towards the sheetpile tips supporting the concrete
Iwalls constructed in 199394. There was ample physical evidence that extremely high pore
pressures likely developed during failure and translation of the levee block, in the form of
extruded bivalve shells littering the ground surface at the second toe thrust, as shown in
Figure 3.20 and indicated on the cross section (Figure 3.19).
Planar translational failures are typical of situations where shear translation occurs
along discrete and semicontinuous low strength horizons (Cruden and Varnes, 1996).
Additional evidence of translation is the relatively intact and undilated nature of the landside
of the failed levee embankment, upon which the old chain link rightoffence was preserved,
as well as a substantial portion of the access road which ran along the levee crest, next to the
concrete Iwall. Wherever we observed the displaced concrete Iwall in this area it was
solidly attached to the Hoesch 12 steel sheetpiles, each segment of which was about 23 inches
wide (as measured along the wall alignment) and 11 inches deep, with an open Zpattern. The
thickness of the sheets were about 7/16ths of an inch. The observed sheetpiles interlocks were
all attached to one another. The entire wall system was quite stiff and fell backward (towards
the canal) after translating approximately the same distance as the landslide of the levee
embankment. The sheetpiles and attached Iwalls formed a stiff rigid element. The sheetpiles
were 23 ft6 inches long and were embedded approximately 2 to 3 feet into the footings of
concrete Iwalls.
The geology of the opposite (west) bank was relatively similar except that the organics
persist in large quantities, to a depth of 36 feet. The marsh deposits appeared deeper here and
root tracks filled with soft secondary interstitial clay persisted to a depth of 39 feet. Sand and
shells were first encountered at 40 feet and cohesionless sand was found at 41 feet. This hole
was terminated at 42 feet.
3.2.3 Interpretation of Data from CPT Soundings
Six distinctive geologic formations are identified studying the Cone Penetrometer Test
(CPT) soundings which were done in the vicinity of 17th Street Canal: Fill, swamp/marsh
deposits, Intermixing deposits, lacustrine deposits, Pine Island beach sand deposits and Bay
Sound deposits. The description and coverage of these geologic formations from CPT
soundings are explained in the following paragraphs. These unit assignments are shown
graphically in Figure 3.21.
FILL: Fill is not present in all CPT soundings. It is characterized by stiff silty clay to
sandy clay and sandy silt with some silt lenses. It is differentiated from the swamp deposits by
having little or no organic matter in its content. Along the breached area, the fill appears to be
missing in the CPT soundings. Fill thickness is around 10 ft (down to 8 ft below sea level) on
3 13
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
the west bank of the 17th street Canal. Just north of the breached area (east bank), the
thickness of the fill ranges from 14 ft to 16 ft (down to 10 ft). Fill materials for the drainage
canals appear to have been placed in three sequences: 1) during the original excavation of the
various canals, between 18331878; 2) after the 1915 Grand Isle Hurricane; and 3) after the
October 1947 hurricane (the history of the drainage canals is described in Chapter 4, Section
4.6).
SWAMP/MARSH DEPOSITS: Marsh deposits consists of soft clays, organic clays
usually associated with organic material (wood and roots). The organic materials are readily
identifiable by observing the big jumps in the friction ratios of the CPT’s. The thickness of
swamp/marsh deposits is around 9.5 ft on the west bank of the canal and 4 to 6 ft on the east
bank of the canal. The depth at which swamp/marsh deposits encountered on banks ranges
from approximately 8.5' (on the west side) to 10' (on the east side), using
the NAVDD882004.65 datum.
INTERMIXING ZONE: This zone consists of mixture of soft clays, silt lenses with
little or no organic material. The thickness of intermixing zone ranges from 3 ft to 8.5 ft on
the east bank of the canal. No intermixing zone is interpreted on the west bank of the canal.
However the contact between marsh and intermixing zone is highly irregular and should be
correlated with borehole data.
LACUSTRINE DEPOSITS: Lacustrine deposits consist of clays to organic clays
with thin silt and fine sand lenses. No organic matter is found in these deposits. The thickness
of lacustrine deposits is around 1719 ft on the west bank of the canal and 1522 ft on the east
bank of the canal. The depth at which lacustrine deposits encountered ranges from 17 (on the
west side) to 1423 (on the east side).
PINE ISLAND BEACH TREND SANDS: Beach sand is identified by its sand and
silty sand content. It is easily recognized in the CPTs by a large jump in the tip resistance and
a drop in the pore pressure. The depth at which beach sand encountered ranges from 37 (on
the west side) to 36 ft (on the east side) and it has fairly uniform 6 ft of thickness.
BAY SOUND: This deposit contains stiff organic clays and stiff clays. It is easily
recognized in the CPTs by a large drop in the tip resistance and an increase in the pore
pressure. Bay sound deposits are only encountered on the east side of the canal and only top
of bay sound deposits encountered in this area –not bottom. The depth at which these deposits
encountered is around 42 ft (which appears to be uniform in this area).
3.3
Geologic Conditions at London Avenue Canal (North) Breach
3.3.1 Introduction
The London Ave. Outlet Canal Levee system catastrophically failed on its western
bank just south of Robert E. Lee Blvd. during Hurricane Katrina between 9 and 10 AM on
August 29, 2005. The hurricane induced a storm surge from the Gulf of Mexico that moved
into Lake Pontchartrain and subsequently backed up into the canal. The levee failed at one
location by translating laterally atop poor foundation materials, not by overtopping. The
3 14
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
break formed on the west bank levee just south of Robert E. Lee Blvd. The toe of this break
appears to have thrust over the surrounding landscape 68 feet in places.
The east bank levee directly opposite this break translated by about two feet, but did
not breach catastrophically. Here again, the movement was due to lateral translation,
reflecting instability within the foundation soils. An imminent failure was likely but
hydrostatic pressure was relieved by the break opposite this bank and a break on the east bank
further south near Mirabeau Ave. Floodwall panels here have been displaced, tilted, and
distressed.
Cohesionless beach sands from the micaceous Pine Island beach strand comprise the
majority of the deposits beneath the London Ave. Canal Levee. These sands were quickly
eroded and deposited in great quantities in the neighborhoods surrounding the breaks. Much
of the sand was also likely in the bottom of the canal prior to the breaks.
3.3.2 Geology Beneath the Levees
A series of continuously sampled borings was conducted and logged using 3inch
Shelby tubes where cohesive soil was present. Cohesionless sands were sampled using the
material recovered during the Standard Penetration Tests (SPT). CPTs were conducted
alongside many of the other borings.
The first two feet of material appeared to be topsoil heavily influenced by modern
vegetative growth. The material was a dark brown silty clay with many roots and organics
and a relatively low water content.
The next 0.65 feet contained highly plastic and waterrich organic clay and contained
what appeared to be the slide plane at 2.65 feet in depth. Although the slip surface was likely
deeper under the levee, it was thrusting to the surface at this point. There was a return to the
dark brown organic silty clay at this point, which continued to 3.1 feet where there was a
strong contact. A gray clayey sand remained in the last 0.5 feet of the tube.
From 46 feet appeared to be a deposit of shallow marsh materials transitioning to
beach sands from Lake Pontchartrain. The first part of the tube contained gray organic rich
clays and silts with a fetid odor and transitioned to a relatively clayey gray sand. Cohesion
dropped beyond 6 feet in depth and sampling was no longer possible using a Shelby tube.
Sampling continued using an SPT split spoon sampler down to 44 feet where clays were again
encountered. The entire layer of sand appeared to be beach sand. Shells were included
throughout the layer and most sand was mica rich, likely brought in by long shore drift from
the Pearl River. Shells were included throughout the layer and most sand was mica rich,
likely transported by longshore drift from the Pearl River. This is the “relic beach” of the
Pine Island Beach Trend described in Section 3.1.3.
The clay recovered from 4446 feet was, silty, bluegray in color, and very plastic.
Sand and shell fragments were mixed in with this clay, possibly due to wave action and
mixing due to storms. Additional boring logs show a lacustrine bay sound material at this
depth. No sampling was conducted by our team below this depth. All recovered sediment
was Holocene in age.
3 15
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Boring logs from Design Manual 19A (U.S. Army Corps of Engineers, 1984) show
similar results. In addition, the transitional layer of clayey sand is shown beneath the breach
but not below adjacent unfailed sections of the levee. The marsh deposits and transitional
zone extend up to 10 feet deeper beneath the breached levee (west) and distressed levee (east)
than below the unbroken portions of the levee. Marsh deposits begin near the surface and
transition to sand at around 1015 feet in depth. The sand continues to around 45 feet where
lacustrine bay sound material is found. This continues down to Pleistocene materials at 6575
feet.
3.4
Geologic Conditions at London Avenue (South) Canal Breach
3.4.1 Introduction
The London Ave. Outlet Canal levee system catastrophically failed on its eastern bank
just north of Mirabeau Ave. during Hurricane Katrina between 7 and 8 AM on August 29,
2005. This failure appears to have been induced by concentrated zone of underseepage,
because the failure was relatively deep, and did not extend over a long zone of the canal. Nor
was there any physical evidence of overtopping. The seepage appears to have been driven by
high water level in the canal, caused by the storm surge coming up the canal from its mouth
along Lake Pontchartrain.
Post failure reconnaissance revealed that micaceous sands from the Pine Island Beach
Strand were eroded from this breach and, possibly, from within the canal where they were
deposited throughout the surrounding neighborhood.
3.4.2 Geology Beneath the Levees
The section of levee incorporated in the London South breach is founded upon
geology similar to the northern London Ave. Canal failure. The levee was constructed upon
approximately ten feet of organicrich cypress swamp deposits. Borings by the Corps of
Engineers indicate that the swamp deposits extended three to five feet deeper below the
failure area than the areas immediately adjacent to the breach (north or south of it). Unlike
the London Avenue northern breach, where there is a transition of clayey sand between the
marsh deposits and the underlying Pine Island Trend sands, there is a more definite transition
at this location. These differences in foundation conditions are indicated on the boring logs
within Design Manual 19A (U.S. Army Corps of Engineers, 1984).
3.5
Geologic Conditions along the Inner Harbor Navigation Canal
3.5.1 Introduction
Levees surrounding the Inner Harbor Navigation Canal (IHNC) were overtopped and
breached catastrophically during Hurricane Katrina. Some of New Orleans’ worst devastation
occurred at two large breaches on the east side of the IHNC in the Lower Ninth Ward. These
breaches washed houses from their foundations, leaving many blocks of the neighborhood as
little more than piles of used lumber, destroyed automobiles, and other debris.
3 16
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
3.5.2 Geology
The geology beneath the IHNC levees is far more complex and variable than that of
the foundation materials at the London Ave. and 17th St. Canals. The foundation materials
here tend to be fluvially dominated by past distributaries of the Mississippi River with the
exception of the area near Lake Pontchartrain. Conditions near the lake more resemble those
under the London Ave. Canal but with a slightly thicker marsh deposit. The buried beach
deposit is present below the marsh and eventually transitions into prodelta clays.
As with most modern fluvial systems, the geology of this Holocene deposit is complex
and varies widely in both vertical and horizontal extent. The area was once covered by the
marshes and swamps once common to the area. Organic fat clays are dominant and contain
peat and other organic materials. Some wood is present but not in the quantities found at the
17th St. Canal site, indicating that marshes were more pronounced at this location. These
deposits vary in thickness between 1020 feet, depending on the location.
Interdistributary materials consisting largely of fat clays dominate much of the IHNC
geology below the marsh deposits. This layer, which also contains zones/lenses of lean clays
and silt, is approximately 3035 feet thick.
A complex estuarine deposit exists below the interdistributary layer and is comprised
of a complex mix of clays, silts, sands, and broken shell material. This deposit is about 30
feet thick and is underlain by Pleistocene deposits (undifferentiated, but commonly a stiff
clay). Cross sections from The New Orleans District’s Design Manual 02 Supplement 8 (U.S.
Army Corps of Engineers, 1968, 1969, 1971) do not always do a good job of differentiating
this material, but much of the material appears to be sand mixed with clays and silts. These
deposits lie at sufficient depth as to preclude their having any significant impact on levee
stability.
Abandoned distributaries cut across the IHNC in some locations. Materials in the old
channels are highly variable. Although basal units usually consist of sands, upper units are
heterogeneous layers of silts, clays, sandy silts, and silty sands. Natural levee deposits are
commonly found around these old channels.
3.6
Paleontology and Age Dating
3.6.1 Introduction
Micropaleontology was used in conjunction with carbon 14 dating to determine both
the age and depositional environment of the sediments below levee failure sites in New
Orleans, LA. Foraminifera, singlecelled protists that secrete a mineralized test or shell, were
identified as these organisms grow in brackish or marine settings but not freshwater. Their
presence in sediments indicate that they were deposited insitu or were transported from
brackish Lake Pontchartrain or marine environments by Hurricanes. Palynology, the
identification and study of organicwalled microfossils, commonly pollens and spores, was
conducted to aid in the recreation of paleoenvironments beneath the levees. Macrofossils of
the phylum Mollusca, including classes Gastropoda and Bivalvia are common in sands of the
Pine Island Trend (Rowett, 1958). Most recovered samples contained heavily damaged shells
or fragments.
3 17
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
3.6.2 Palynology
Although varying sediment types including clays, peats, and sands were studied,
similar palynomorphs were found throughout the samples. These samples came from
different depths and locations throughout New Orleans. The commonalities between the
sediments may be due to transportation of the palynomorphs by wind and water or the mixing
of materials by hurricanes. Pollens of the family Taxodiaceae, genus Cupressacites (cypress)
are common. Species of cypress are common in perennially wet areas such as swamps.
Cypress is common throughout the swamps of the Gulf Coast Region. Cypress wood,
including trunks, roots, and stumps, was unearthed by scour during the levee failures and
subsequent construction to temporarily patch the levees. Samples recovered in 3” Shelby
tubes commonly included cypress fragments resembling commercially available landscaping
mulch and cores of intact wood. Cypress trees are freshwater and die if exposed to salt water
for a prolonged amount of time.
Dinocysts/Dinoflagellates were also discovered among the samples taken for
palynology. Dinoflagellates are singlecelled algae belonging to the Kingdom Protista. They
live almost exclusively in marine and brackish water environments, with very few freshwater
species. The discovery of these organisms was not surprising, given the close proximity to
brackish Lake Pontchartrain (essentially a bay). On the other hand, several exclusively
marine species that live in the open ocean were recovered. These species were transported a
far distance inland, indicating transport by a catastrophic event, possibly a hurricane storm
surge or tsunami.
3.6.3
Foraminifera
Foraminifera were identified in the Pine Island Trend, a micaceous quartz beach sand
that was deposited in the Holocene Gulf of Mexico by the Pearl River of Mississippi. This
sand was subsequently formed into a large sand spit by long shore drift, separating Lake
Pontchartrain from the rest of the Gulf of Mexico (Saucier, 1994). Lake Pontchartrain is a
brackish body of water with only a small connection to the Gulf. Agglutinated, planispiral,
and uniserial foraminifera were discovered where the sand grades into the silts and clays
deposited in the low energy environments of Lake Pontchartrain. Although foraminifera are
abundant at these locations, their diversity is low. This is indicative of a stressed environment
and is not surprising, given the brackish nature of Lake Pontchartrain.
3.6.4
Carbon 14 age dating
We are awaiting the results of six C14 age dating by the NSFfunded age dating
laboratory at the University of New Mexico in Albuquerque, NM. These are samples of the
cypress wood and fibrous peats recovered at the 17th Street Canal failure area.
3.7
Mechanisms of Ground Settlement and Land Loss in Greater New Orleans
3.7.1 Settlement Measurements
URS Consultants (2006) in Baton Rouge recently completed a study for FEMA of the
relative ground settlement in New Orleans since 1895, using the Brown (1895) map, which
3 18
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
has 1 foot contours and extends north to the Lake Pontchartrain shoreline. This comparison
was made by creating Digital Elevation Models (DEMs) of the 1895 map (Figure 3.22)
relative to Mean Gulf Level against the 1999/2002 DEM extracted from LiDAR data and
New Orleans network of benchmarks. The resulting product was a map noting relative
settlement (in feet) between 1895 and 1999, shown in Figure 3.23. This study suggests that
the entire city has settled between 2 and 10 feet. During this same interim, sea level has risen
approximately 12 inches. The area with the greatest settlement (> 8 feet) was north of I610
in the Lakeview area and north of Mirabeau Ave. in the Gentilly area, exclusive of the 1931
fill along Lake Pontchartrain (which extends a half mile into the Lake).
3.7.2 Tectonic Subsidence
Tectonic subsidence is caused by sediment compaction at great depths (Figure 3.24).
Salt and muds flow towards the continental shelf. Pressure ridges and fold belts develop;
which are akin to sitting on a peanut butter and jelly sandwich and watching material ooze out
and shift. The Continental Slope and Shelf is blanketed by large subaqueous landslides.
3.7.3 Lystric Growth Faults
As compacting materials move seaward, the ground surface drops. If sediment is not
added at the ground surface, the seaward side of these features gradually subsides below sea
level. The delta’s lystric growth faults have been grouped into bands thought to be more or
less related to one another. The relatively recent emergence of the Baton Rouge Fault Zone
along the northern shore of Lake Pontchartrain, thence towards Baton Rouge, is the most
striking example, and one of the furthest inland (Figures 3.25 and 3.26).
3.7.4 Compaction of Surficial Organic Swamp and Marsh Deposits
The interdistributary sediment package covering the old back swamps around New
Orleans is highly compressible and the neighborhoods built on these materials exhibit obvious
signs of differential settlement. This is particularly true of the West End, Lakeview, City
Park, Fillmore, St. Anthony, Dillard, Milneburg, Pontchartrain Park, Desire, and Gentilly
neighborhoods flanking Lake Pontchartrain. Most of this settlement is ascribable to
oxidationinduced settlement of underlying peaty soils, caused by local drawdown of the
ground water table, as sketched in Figure 3.27. The amount of postdevelopment settlement is
moreorles proportional to the thickness of the peaty surface layer, shown in Figure 3.12. It
varies in thickness from a few feet to as much as 20 feet, depending on location (WPALA,
1937; Kolb and Saucier, 1982).
The mechanisms promoting surficial settlement in lower New Orleans are thought to
be: 1) drainage of the near surface soils, through simple nearsurface dewatering and the
storm water collection system; and 2) biochemical oxidation of organic materials above the
[lowered] water table. Simple drainage of the surficial peaty soils can induced consolidation
of up to 75% of their original thickness (Kolb and Saucier, 1982), which in of itself, could
account for up to 12 ft of settlement, if the local water table was lowered >15 feet. But,
biochemical oxidation continues afterwards, with greater severity during extended periods of
drought, as occurred in the late 1990searly 2000s around New Orleans. Oxidation continues
until only the mineral constituents of the soil are left remaining.
3 19
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Dense urban development also leads to increased subsidence because the absorptive
capacity of the peaty soils is decreased by the mass implementation of impervious surfaces,
such as streets, parking lots, sidewalks, roofs, driveways, etc. Increasing the area of
impervious surfaces decreases overall seasonal infiltration and increases the peak runoff
through hardened impervious surfaces. As a consequence, the Sewerage & Water Board of
New Orleans had to continually increase the capacity of their drainage collection, conveyance
and discharge system during the post1945 period. These examples are from the Lakeview
area adjacent to the 17th St. Canal failure, where the ground appears to have settled 10 to 16
inches since 1956.
The Lakeview and Gentilly neighborhoods were intensely developed in the post World
War II era, mostly between 194670 (although infilling of newer structures continued up
through 2005, as older structures were torn down). Most residential structures built in lower
New Orleans after the mid1950s are concrete slabs founded on wood pilings 6 to 8 inches in
diameter, driven about 30 feet deep (Waters, 1984). From inspection, it appears that the
ground beneath the foundations has settled 10 to 40 inches over the past 50 +/ years since
these homes were constructed. This development was accompanied by a lowering of the
ground water table to accommodate normal living conditions and combat mildew and mold in
the crawl spaces beneath the homes (Figure 3.28 upper). Since the historic groundwater
table was at or within a few inches of the ground surface in this area, the lowering of the
water table by 2 to 10 feet in this area hastened nearsurface settlement through oxidation of
the organic rich peat soils underlying the area.
As the peats oxidize, the ground settles, creating a depressed area beneath pile
supported homes (Figure 3.28upper). Groundwater pumping, drainage, and structural and
earthen surcharges all contribute to the observed settlement. Historic measurements of
ground settlement in the Kenner area of Jefferson Parish are shown in Figure 3.29.
During the 130 to 170 years since the drainage canals were constructed upon what
became the Lakeview and Gentilly areas, these channels have never been drained for any
significant period of time, because they were open to Lake Pontchartrain. As a consequence,
the peaty soils immediately beneath these canals (17th Street, Orleans, and London Avenue)
and Bayou St. John have not experienced significant nearsurface settlements like those
fomented by oxidation of peaty soils in the adjoining neighborhoods, although they have
experienced gross ground settlement due to the other causes described in Section 3.7.
This history of nearcontinual ground settlement necessitated raising of the old
drainage canal embankments on three occasions in the 20th Century, following hurricane
induced flooding from storm surges off Lake Pontchartrain, in: 1915, 1947, and 1965. Earth
fill was placed upon the levee embankments in 1915 and 1947. After flooding associated
with Hurricane Betsy in 1965 steel sheetpiles were used in selective zones to increase the
freeboard for Category 3 storm surge (a figure that shifts each decade, as new information and
models are developed). In the 1990s sheetpile–supported concrete Iwalls were constructed
along the crests of the drainage canals and on either side of the IHNC.
3.7.5 Structural Surcharging
An interesting aspect of the recent URS (2006) study for FEMA is the marked
increase in settlement noted in the Central Business District, where tall structures are founded
3 20
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
on deep piles. This area settled 5 inches in 100 years, but much less further away from the
city’s tallest and heaviest structures. The sandy natural levees along the Mississippi River
even settled 2 inches; likely due to surcharging by the Corps’ Mississippi River & Tributaries
Project (MR&T) sequences of levee enlargements, between 192860.
3.7.6 Extraction of Oil, Gas, and Water
Since the 1960s groundwater withdrawal has been recognized as contributing to
subsidence of the Gulf Coast area, especially adjacent to deep withdrawal points for industrial
consumption (Kazmann and Heath, 1968). More recently, R.A. Morton of the USGS has
blamed oil and gas extraction for the subsidence of the Mississippi Delta. Morton has
constructed convincing correlations between petroleum withdrawal and settlement rates on
the southern fringes of the delta, near the mouth of the Mississippi River (Morton, Buster, and
Krohn, 2002). But, other factors are likely involved as well, as petroleum withdrawal alone
cannot account for marked settlement well inland of Lake Pontchartrain, where little
withdrawal has occurred. Figure 3.30 presents Saucier’s (1994) map of the Mississippi Delta,
which summarizes the structural geologic framework of the area. This shows salt basins, salt
domes, and active growth faults that pervade the delta region. Solutioning of salt diapirs and
seaward migration of low density contrast materials likely exacerbate settlement, but more
slowly that fluid/gas withdrawal.
3.7.7 Coastal Land Loss
The U.S. Geological Survey’s National Wetlands Research Center (USGSNWRC)
has about 100 years of land loss information. Since 1973, satellites have allowed monitoring
of sediment expulsion from the delta and the nefarious shoreline, which is continuously
sinking. The USGSNWRC has been monitoring coastal land loss over the past 50 years
using 1956 and 1978 imagery published by Cahoon and Groat (1990) and LANDSAT
Thematic Mapper satellite imagery from 1993 and 2000 (Barras et al., 2003).
Coastal lands loss is a high visibility problem along the Gulf Coast, especially in the
Mississippi Delta.
•
•
•
•
•
USGS and NGS state that the approximate rate of subsidence is between 1/3” to ½”
per year; or about 4.2 ft/100 yrs
Sea level rise is running about 1 ft/100 yrs (Burkett, Zilkowsi, and Hart, 2003)
15% of New Orleans is already more than 10 ft below sea level (URS, 2006)
The average current rate of coastal land loss is between 25 and 118 square miles per
year (the record of 118 mi2 being a result of Hurricanes Katrina and Rita in 2005)
The 2050 Reclamation Plan would restore 25 to 30 mi2 over the next 40 to 50 yrs at a
cost of $14 billion
The USGS National Wetlands Research Center has determined that Hurricane Katrina
created as much new standing water area in the Mississippi Delta (below sea level) as
occurred naturally over the previous 50 years! This was due to increased traction shear,
which tore out large tracts of peat bogs, to depths of several feet (USGSNWRC, 2006).
3 21
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
3.7.8
Negative Impact of Ground Settlement on Storm Surge
As large tracts of land along coastal Louisiana sink below sea level, less protection is
afforded inland areas from the destructive impacts of storm surges caused by hurricanes. The
absolute level of storm surges on the Louisiana Coast is also likely exacerbated by the loss of
coastal vegetation, such as cypress swamps, which mollify wave energy through mechanical
obstruction and tortuous flow path (increased boundary shear) as high water sweeps onto the
land. The diminution of storm surge height would depend on the speed and duration of the
storm as it makes landfall, and the density and height of the cypress swamps and the
vegetation they support.
Many figures have been cited in the nontechnical literature in regards to this
“protective impact;” the most common being that every 41/2 miles of mature cypress swamp
absorbs one foot of storm surge coming from the Gulf (Hallowell, 2005). Although the
concept of storm surge mollification through turbulent boundary shear at the ground surface is
conceptually possible, we were unable to find any measurements that quantified this effect
through credible scientific study of historic storm events (NRC, 2006). Observations made
during and after Hurricane Katrina, may, however, help to fill this data void.
3.7.9 Conclusions about Ground Settlement
Multiple physical factors have combined to cause marked historic settlement of the
New Orleans area. These include:
1)
2)
3)
4)
5)
6)
7)
The average silt load of the Mississippi River (550 million tons [mt] per
year prior to 1950; now 220 mt/yr) causes continuous crustal loading of the
Mississippi River Delta, causing isostasydriven settlement, which has been
recognized since 1937 (Meade and Parker, 1985; Russell, 1940, 1967).
Tectonic compaction caused by sediment compaction at great depths, with
associated pressure ridges and fold belts.
Subsidence along the seaward side of lystric growth faults perturbing the
Mississippi Delta.
Drainage of nearsurface soils causing an increase in effective stress and
resulting primary consolidation
Oxidation of nearsurface peaty soils due to lowering of the groundwater
table in developed areas, or drainage of historic marshes and swamp lands.
This component is often exacerbated by New Orleans residents who
routinely fill in portions of their yards adjacent to protruding foundations
(Figure 3.28), driveways and sidewalks, creating additional loads on the
compressible materials lying beneath them.
Consolidation of soft compressible soils (with high water contents), due to
surcharging by earth filling and other manmade improvements.
Structural surcharging. Settlements measured in vicinity of downtown high
rise structures suggests that a portion of the observed settlement may also
emanate from deeper horizons, caused by loads transferred to those
horizons along friction piles and caissons for heavy structures.
3 22
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
8)
9)
3.8
Fluid extraction of oil, gas, and water from the subsurface. Extraction of
fluids and natural gas is a pressure depletion that increases effective
stresses acting on underlying sediments, hastening consolidation.
Solutioning of salt diapirs (salt domes) and seaward migration of low
density contrast materials (salt and mud), as well as large subaqueous slope
movements on the continental slope and shelf. When large volumes of
material move laterally, adjoining areas drop to compensate for the
volumetric strain.
References
American Institute of Professional Geologists. (1993). A Citizen’s Guide to Geologic
Hazards. Arvada, Colorado, 134 p.
Barras, J., Beville, S., Britsch, D., Hartley, S., Hawes, S., Johnston, J., Kemp, P., Kinler, Q.,
Martucci, A., Porthouse, J., Reed, D., Roy, K., Sapkota, S., and Suyhada, J. (2003).
Historical and Projected Coastal Louisiana Land Changes: 1978-2050, USGS Open
File Report 03334, U.S. Geological Survey National Wetlands Research Center,
Baton Rouge LA.
Brown, L.W. (1895). Contour Map of New Orleans (Plate 11), Report of the Advisory Board
on Drainage of the City of New Orleans, Louisiana, New Orleans Engineering
Committee.
Burkett, V.R., Zilkowski, D.B., and Hart, D.A. (2003). “SeaLevel Rise and Subsidence:
Implications for Flooding in New Orleans, Louisiana. ” USGS Subsidence Interest
Group Conference, Proceedings of the Technical Meeting, K. R. Prince and D.L.
Galloway (Eds.), Galveston TX, 6370.
Cahoon, D.R., and Groat, C.G. (1990). A Study of Marshy Management Practice in Coastal
Louisiana, Vol II, Technical Description, OCS study MMS 900076, Minerals
Management Service, New Orleans LA.
Coastal Environments, Inc. (2001). Southeast Louisiana Fault Study, Consultant’s Report for
U.S. Army Corps of Engineers, New Orleans District, Contract No. DACW2900C0034 (February).
Coleman, J.M. (1988). “Dynamic Changes and Processes in the Mississippi River Delta.”
Geological Society of America Bulletin, 100(7), 9991015.
Coleman, J.M. and Galiano, S. M. (1964). “Cyclic Sedimentation in the Mississippi River
Deltaic Plain.” Transactions of the Gulf Coast Association of Geological Societies, 14,
6780.
Coleman, J.M. and Robert, H.H. (1991). “The Mississippi River Depositional System: A
Model for the Gulf Coast Tertiary.” An Introduction to Central Gulf Coast Geology,
D. Goldthwaite (Ed.), New Orleans Geological Society, New Orleans LA, 99121.
Compton, R.R. (1962). Manual of Field Geology. John Wiley & Sons, New York.
Cruden, D.M., and Varnes, D.J. (1996). Landslide Types and Processes Special Report 247.
In A.K. Turner and R.L. Schuster, eds. Landslides Investigation and Mitigation.
Transportation Research Board. National Academies Press.
3 23
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Draut, A.E., Kineke, G.C., Velasco, D.W., Allison, M.A., and Prime, R.J. (2005). “Influence
of the Atchafalaya River in Recent Evolution of the Chenier Plain inner Continental
Shelf, Northern Gulf of Mexico.” Continental Shelf Research 25, 91112.
Dunbar, J.D., Blaes, M.R., Duitt, S.E., May, J.R., and Stroud, K.W. (1994). Geological
Investigation of the Mississippi River Deltaic Plain, Technical Report GL8415, U.S.
Army Engineer Waterways Experiment, Vicksburg MS.
Gould, H.R., and Morgan, J.P. (1962). “Coastal Louisiana Swamps and Marshes.” Geology of
the Gulf Coasts and Central Texas and Guidebook of Excursions, E. Rainwater, and
R. Zingula (Eds.), Houston Geological Society, Houston TX.
Gould, H.R. (1970). “The Mississippi Delta Complex” Deltaic Sedimentation: Modern and
Ancient, SEPM Special Publication 15, J.P. Morgan (Ed.), Society of Economic
Paleontologists and Mineralogists, Tulsa Oklahoma, 330.
Hallowell, C. (2005). Holding Back the Sea: the Struggle on the Gulf Coast to Save America,
Harper Perennial, New York.
Interagency Performance Evaluation Task Force (IPET). (2006). Performance Evaluation
Status and Interim Results.” Report 2 of a Series, Performance Evaluation of the New
Orleans and Southeast Louisiana Hurricane Protection System., U.S. Army Corps of
Engineers, Washington DC.
Kazmann, R.G., and Heath, M.M. (1968). “Land Subsidence Related GroundWater Offtake
in the New Orleans Area.” Transactions of the Gulf Coast Association of Geological
Societies, 18, 108113.
Kolb, C.R. (1958). Geology of the Mississippi River Deltaic Plain, Southeastern Louisiana,
Technical Report 3483, Vol. 1, U.S. Army Engineer Waterways Experiment Station,
Vicksburg MS.
Kolb, C.R. and Shockley, W.G. (1959). “Engineering Geology of the Mississippi Valley.”
ASCE Transactions, 124, 633656.
Kolb, C.R., Smith, F.L., and Silva, R.C. (1975). Pleistocene Sediments of the New OrleansLake Pontchartrain Area, Technical Report No. S756, U.S. Army Engineer
Waterways Experiment Station, Vicksburg MS.
Kolb, C.R., and Van Lopik, J.R. (1958). Geology of the Mississippi River Deltaic Plain,
Southern Louisiana, Technical Report No. 3483, U.S. Army Engineer Waterways
Experiment Station, Vicksburg MS.
Meade, R.H., and Parker, R.S. (1985). “Sediment in Rivers of the United States.” National
Water Summary 1984 – Water Quality Issues, U.S. Geological Survey, 4960.
Moore, N. R. (1972) Improvement of the Lower Mississippi River and Tributaries 1931-1972,
Department of the Army Corps of Engineers, Mississippi River Commission,
Vicksburg MS.
National Research Council. (2006). Drawing Louisiana’s New Map: Addressing Land Loss in
Coastal Louisiana, National Academies Press, Washington DC.
Russell, R.J. (1940). “Quaternary History of Louisiana.” Geological Society of America
Bulletin, 51, 11991234.
3 24
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Rowett, C.L. (1957). “A Quaternary Molluscan Assemblage from Orleans Parish, Louisiana.”
Transactions of the Gulf Coast Association of Geological Sciences, 7, 153164.
Russell, R. J. (1967). River Plains and Sea Coasts, Univ. of California Press, Los Angeles,
CA.
Saucier, R.T. (1963). Recent Geomorphic History of the Pontchartrain Basin. Louisiana,
Coastal Studies Series No 9, Louisiana State University Press, Baton Rouge LA.
Saucier, R.T. (1994). Geomorphology and Quaternary Geologic History of the Lower
Mississippi Valley. U.S. Army Corps of Engineers Waterways Experiment Station,
Vicksburg MS.
Terzaghi, K. (1956). “Varieties of Submarine Slope Failures.” Proceedings of the Eighth
Texas Conference on Soil Mechanics and Foundation Engineering, 141.
URS Consultants. (2006). A Century of Subsidence: Change in New Orleans DEMs Relative
to MGL 1895 to 1999/2002. GIS Maps Prepared for The Federal Emergency
Management Agency, URS Consultants, Baton Rouge LA.
U.S. Army Corps of Engineers. (1968) Lake Pontchartrain, LA and Vicinity Lake
Pontchartrain Barrier Plan, Design Memorandum No 2, General Supplement No. 8,
Inner Harbor Navigation Canal Remaining Levees, New Orleans District, New
Orleans LA.
U.S. Army Corps of Engineers. (1969) Lake Pontchartrain, LA and Vicinity Lake
Pontchartrain Barrier Plan, Design Memorandum No 2, General Supplement No. 8,
West Levee Vicinity France Road and Florida Avenue, New Orleans District, New
Orleans LA.
U.S. Army Corps of Engineers. (1971) Lake Pontchartrain, LA and Vicinity Lake
Pontchartrain Barrier Plan, Design Memorandum No 2, General Supplement No. 8,
Modification of Protective Alignment and Pertinent Design Information I.H.N.C.
Remaining Levees West Levee Vicinity France Road and Florida Avenue
Containerization Complex, New Orleans District, New Orleans LA.
U.S. Army Corps of Engineers. (1984). General Design London Avenue Outfall Canal,
Design Memorandum No. 19A, New Orleans District, New Orleans LA.
U.S. Army Corps of Engineers. (1990). General Design 17th St. Outfall Canal (Metairie
Relief), Design Memorandum No. 20, New Orleans District, New Orleans, LA.
U.S. Geological Survey (USGS)National Wetland Research Center(NWRC). (2006). Water
Area Changes in Southeastern Louisiana after Hurricanes Katrina and Rita Detected
with Landsat Thematic Mapper Satellite Imagery, Map ID: USGSNWRC 2006110049.
Waters, R. K. (1984). Handbook for Pile Supported Foundations, G.G. Printing Co., Kenner,
Louisiana LA.
Works Projects Administration of Louisiana. (1937). Some Data in Regard to Foundations in
New Orleans and Vicinity, New Orleans LA. 243 p.
3 25
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 3.1 (upper): Areal distribution of abandoned channels and distributaries of the Mississippi
River (from Kolb, 1958). The Metairie Ridge distributary channel (highlighted in red) lies
between two different depositional provinces in the center of New Orleans (shown in Figure 3.6).
Figure 3.1 (lower): Major depositional lobes identified in lower Mississippi Delta around
New Orleans, taken from Saucier (1994).
3 26
Independent Levee
Investigation Team
3 27
Figure 3.2: Northsouth geologic cross section through the central Gulf of Mexico Coastal Plain, along the Mississippi River
Embayment (from Moore, 1972). Note the axis of the Gulf Coast Geosyncline beneath Houma, LA, southwest of New Orleans. In
this area the Quaternary age deposits reach a thickness of 3600 ft.
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 3.3: Transverse cross section in a west to east line, across the Mississippi River Delta a
few miles south of New Orleans, cutting across the southern shore of Lake Borgne (modified
from Saucier, 1994). New Orleans is located on a relatively thin deltaic plain towards the
eastern side of the delta’s depositional center, which underlies the Atchafalaya Basin, west of
New Orleans.
Figure 3.4: Pleistocene geologic map of the New Orleans area, taken from Kolb and Saucier
(1982), modified from Kolb and Saucier (1982). The yellow stippled bands are the principal
distributory channels of the lower Mississippi during the late Pleistocene, while the present
channel is shown in light blue. The Pine Island Beach Trend is shown in the ochre dotted
pattern. Depth contours on the upper Pleistocene age horizons are also shown.
3 28
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 3.5: Contours of the entrenched surface of the Wisconsin glacial age deposits
underlying New Orleans, taken from Saucier (1994). Note the well developed channel
leading southward, towards what used to be the oceanic shoreline. This channel reaches a
maximum depth of 150 feet below sea level.
3 29
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 3.6: Areal distribution and depth to top of formation isopleths for the Pine Island
Beach Trend beneath lower New Orleans, modified from Saucier (1994).
3 30
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 3.7: Block diagram of the geology underlying New Orleans (modified from Kolb and Saucier,
1982). The principal feature dividing New Orleans is the Metairie distributary channel, shown here,
which extends to a depth of 50 feet below MGL and separates geologic regimes on either side. Note
the underlying faults, especially that bounding the northern shore of Lake Pontchartrain.
Figure 3.8: Block diagram illustrating relationships between subaerial and subaqueous deltaic
environments in relation to a single distributary lobe (taken from Coleman and Roberts, 1991). The
Lakeview and Gentilly neighborhoods of New Orleans are underlain by interdistributary sediments,
overlain by peaty soils lain down by fresh marshes and cypress swamps.
3 31
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 3.9: Sedimentary sequence caused by overlapping cycles of deltaic deposition, along a trend
normal to that portrayed in the previous figure (modified from Coleman and Gagliano, 1964). As long
as the distributary channel receives sediment, the river mouth progrades seaward. Lower New Orleans
lies on a deltaic plain with marsh and swamp deposits underlying the Lakeview and Gentilly
neighborhoods, and delta front deposits closer to MetairieGentilly Ridge, the nearest distributary
channel.
Figure 3.10: Portion of the 1849 flood map showing the mapped demarcation between
brackish and fresh water marshes along Lake Pontchartrain (taken from WPALA, 1937).
This delineation is shown on many of the historic maps, dating back to 1749.
3 32
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 3.11: 1816 flood map of New Orleans showing areal distribution of cypress swamps
north of the old French Quarter (from the Historic New Orleans Collection). These extended
most of the distance to the Lake Pontchartrain shore.
3 33
Independent Levee
Investigation Team
3 34
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Figure 3.12: Distribution and apparent thickness of surficial peat deposits in vicinity of New Orleans, taken from Kolb and
Saucier (1982) and Gould and Morgan (1962). Turn on side, landscape. Add the three drainage canals (blue dashed lines)
IHNC, MRGO, IGWW.
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 3.13: Geologic map of the greater New Orleans area, modified from Kolb and Saucier
(1982). The sandy materials shown in yellow are natural levees, green areas denote old
cypress swamps and brown areas are historic marshlands. The stippled zone indicates the
urbanized portions of New Orleans.
3 35
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 3.14: Geologic cross section along south shore of Lake Pontchartrain in the Lakeside,
Gentilly, and Ninth Ward neighborhoods, where the 17th Street, London Avenue, and IHNC
levees failed during Hurricane Katrina on Aug 29, 2005. Notice the apparent settlement that
has occurred since the city survey of 1895 (blue line), and the correlation between settlement
and nonbeach sediment thickness. This eastwest section was taken from Dunbar et al.
(1994).
Figure 3.15: Wood and other organic debris was commonly sampled in exploratory borings
carried out after Hurricane Katrina throughout the city. This core contains wood from the old
cypress marsh that was recovered near the 17th Street (Metairie Relief) Canal breach. Organic
materials are decaying throughout the city wherever the water table has been lowered, causing
the land surface to subside (photo by C. M. Watkins).
3 36
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 3.16: Overlay of 1872 map by Valery Sulakowski on the WPALA (1937) map,
showing the 1872 shoreline and sloughs (in blue) along Lake Pontchartrain. Although
subdivided, only a limited number of structures had been built in this area prior to 1946. The
position of the 2005 breach along the east side of the 17th Street Canal is indicated by the red
arrow.
3 37
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 3.17: Aerial photo of the 17th Street Canal breach site before the failure of August 29,
2005. The yellow lines (at middle right) indicate the positions of the geologic sections
presented in Figures 3.18 and 3.19, while cross sections AC’ are shown in Figure 3.21.
3 38
Independent Levee
Investigation Team
3 39
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Figure 3.18: Westtoeast geologic cross section through the 17th Street Canal failure approximately 60 feet north of the northern curb
of Spencer Avenue, close to the yellow school bus. A detailed sketch of the basal rupture surface is sketched above right. The slip
surface was about one inch thick with an extremely high moisture content (watery ooze). A zone of brecciation 3 to 4 inches thick
was above this. Numerous pieces of cypress wood, up to 2 inches diameter, were sheared off along the basal rupture surface.
Independent Levee
Investigation Team
3 40
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Figure 3.19: Westtoeast geologic cross section through the 17th Street Canal failure approximately 140 feet north of the
northern curb of Spencer Avenue, just north of the first surviving home next to the canal. Large quantities of bivalve shells were
extruded by high water pressure along the regressing toe thrusts (shown in Figure 3.20). Note the slight back rotation of the
distal thrust block.
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 3.20: Bivalve shells ejected by high pore pressures emanating from toe thrusts on
landside of failed levee at the 17 Street Canal (detail view at upper left). These came from a
distinctive horizon at a depth of 2 to 5 feet below the prefailure grade (photo by C.M.
Watkins).
3 41
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 3.21 upper: Stratigraphic interpretations between CPT soundings along western
embankment of the 17th Canal (in Jefferson Parish), opposite the breach on the east side. The
marshswamp deposits are dipping slightly towards Lake Pontchartrain, while the lacustrine
clays appear to be flat lying.
3 42
Independent Levee
Investigation Team
3 43
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Figure 3.21 lower: Stratigraphic interpretations and crosscanal correlations in vicinity of the 17th Street Canal breach on August 29,
2005. The swamp much appeared to be thinning northerly, as does the underlying Pine Island Beach Trend. The lacustrine clays
appear to thicken northward, as shown. The approximate positions of the flood walls (light blue) and canal bottom (dashed green) are
indicated, based on information provided by the Corps of Engineers (IPET, 2006).
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 3.22: Topographic map with one foot contours prepared under the direction of New
Orleans City Engineer L.W. Brown in 1895. This map was prepared using the Cairo Datum,
which is 21.26 feet above Mean Gulf Level.
3 44
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 3.23: Map showing relative elevation change between 1895 and 1999/2002, taken from
URS (2006). The approximate net subsidence was between 2 and 10+ feet, depending on
location. The brown colored zones along Lake Pontchartrain and the Mississippi River are
areas where substantive fill was placed during the same interim.
3 45
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 3.24: Block diagram illustrating various types of subaqueous sediment instabilities in the
Mississippi River Delta, taken from Coleman (1988).
Figure 3.25: Geologic cross section through the Gulf Coast Salt Dome Basin, taken from Adams
(1997). This shows the retrogressive character of young lystric normal faults cutting coastal
Louisiana, from north to south. The faults foot in a basementsaltdecollement surface of middle
Cretaceous age (> 100 Ma).
3 46
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 3.26: Structural geologic framework of southeastern Louisiana, taken from Coastal
Environments (2001). This plot illustrates the enechelon belts of growth faults forming more or less
parallel to the depressed coastline. The Baton Rouge Fault Zone (shown in orange) is graphic fault
scarp feature that has emerged over the past 50 years, north and west of Lake Pontchartrain.
Figure 3.27: Settlement of surficial peaty soils is usually triggered by lowering of the local
groundwater table, either for agriculture or urban development. Lowering the water table
increases the effective stress on underlying sediments and hastens rapid oxidation of organic
materials, causing settlement of these surficial soils (taken from AIPG, 1993).
3 47
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 3.28: Upper photo shows gross nearsurface settlement of homes in the Lakeview
neighborhood, close to the 17th Street Canal breach. Most of the homes were constructed
from 195675 and are founded on wood piles about 30 feet deep. The lower photo shows
protrusion of a bricklined manhole on Spencer Avenue, suggestive of at least 12 inches of
near surface settlement during the same interim (photos by J. D. Rogers).
3 48
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 3.29 – Record of historic settlement in the town of Kenner, which is characterized by
6.5 to 8 feet of surficial peaty soils (taken from Kolb and Saucier, 1982). The earlier episodes
of settlement were triggered by groundwater withdrawal (for industrial and municipal usage),
while the later episode was caused by drainage associated with urban development. This area
was covered by dense cypress swamps prior to development.
3 49
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 3.30: Structural geologic framework of the lower Mississippi River Delta, taken from
Saucier (1994). Growth faults (solid black lines) perturb the coastal deltaic plain, as do salt
domes (shown as dots). The nearest salt domes to New Orleans are 9 to 15 miles southwest of
New Orleans. This study did not uncover evidence of growth faults materially affecting any of
the levee failures from Hurricane Katrina, although such possibility exists.
3 50
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
CHAPTER FOUR: HISTORY OF THE NEW ORLEANS
FLOOD PROTECTION SYSTEM
4.1 Origins of Lower New Orleans
New Orleans is a deep water port established in 1718 about 50 miles up the main stem
of the Mississippi River, on the eastern flank of the Mississippi River Delta. New Orleans was
established by the French in 171718 to guard the natural portage between the Mississippi
River and Bayou St. John, leading to Lake Pontchartrain. The 1749 map of New Orleans by
Francois Saucier noted the existence of fresh water versus brackish water swamps along the
southern shore of Lake Pontchartrain.
The original settlement was laid out as 14 city blocks by 172123, with drainage
ditches around each block. The original town was surrounded by a defensive bastion in the
classic French style. The first levee along the left bank of the Mississippi River was allegedly
erected in 1718, but this has never been confirmed (it is not indicated on the 1723 map
reproduced in Lemmon, Magill and Wiese, 2003). New Orleans’ early history was typified
by natural catastrophes. More than 100,000 residents succumbed to yellow fever between
1718 and 1878. Most of the city burned to the ground in 1788, and again, in 1794, within
sight of the largest river in North America. The settlement was also prone to periodic flooding
by the Mississippi River (between April and August), and flooding and wind damage from
hurricanes between June and October. Added to this was abysmally poor drainage, created by
unfavorable topography, lying just a few feet above sea level on the deltaic plain of the
Mississippi River, which is settling at a rate of between 2 and 10 feet (ft) per century.
The tendency for flooding during late spring and summer runoff came to characterize
the settlement. The natural swamps north of the original city were referred to as “back
swamps” in the oldest maps, and “cypress swamps” on maps made after 1816. During the
steamboat era (post 1810), New Orleans emerged as the major transshipment center for river
borne to seaborn commerce, viceversa, and as a major port of immigration. By 1875 it was
the 9th largest American port, shipping 7,000 tons annually. In 1880, after completion of the
Mississippi River jetties (in 1879), New Orleans experienced a 65fold increase in seaborne
commerce, shipping 450,000 tons, jumping it to the second largest port in America (New
York then being the largest). New Orleans would retain its #2 position until well after the
Second World War, when Los AngelesLong Beach emerged as the largest port, largely on
the strength of its container traffic from the Far East. New Orleans remains the nation’s
busiest port for bulk goods, such as wheat, rice, corn, soy, and cement.
New Orleans has always been a high maintenance city for drainage. The city’s
residential district did not stray much beyond the old Mississippi River levee mound until
after 1895, when serious attempts to bolster the Lake Pontchartrain “back levee” and establish
a meaningful system of drainage were undertaken by the city. Most of the lowland cypress
swampland between MidTown and Lake Pontchartrain was subdivided between 19001914,
after the City established and funded a Drainage Advisory Board to prepare ambitious plans
for keeping New Orleans dry all the way to Lake Pontchartrain’s shoreline. This real estate
41
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
bonanza increased the City’s urban acreage by 700% and their assessed property values by
80% during the same interim (Campanella, 2002). Most of these lots were developed after the
First World War (191718). Another 1,800 acres was reclaimed from the south shore of Lake
Pontchartrain in 192831, between the mouth of the 17th Street Canal on the west and the
Inner Harbor Navigation Canal (IHNC) on the east. The entire area was subsequently built
out following the Second World War, between 194570.
4.2 Mississippi River Floods
The Mississippi River drains 41% of the continental United States, with a watershed
area of around 1,245,000 square miles (mi2), according to the U.S. Army Corps of Engineers.
This makes it the third largest watershed of any river in the world. Although its official length
is 2,552 miles (if measured from Itasca State Park in Minnesota), when combined with the
Missouri River (2,540 miles long), it is the longest river in North America, with a combined
length of 3,895 miles. Prior to 1950, the sediment load (suspended and dissolved) transported
by the Mississippi River averaged between 550 and 750 million tons per annum (Meade and
Parker, 1985). Since 1950, the average annual suspended discharge of the river has decreased
to 220 million tons/yr (Meade and Parker, 1985), because of the construction of dams and
maintenance of the navigation channel (which includes dredging). The Mississippi River now
ranks as the 6th largest silt load in the world.
The Mississippi’s flood plain upstream of Baton Rouge is an alluvial valley that, prior
to 1928, was periodically subject to inundation by flooding. Vast tracts of the flood plain
were periodically inundated. 26,000 square miles of land (mi2) was inundated during the
1927 flood; 20,312 mi2 in the 1973 flood, and 15,600 mi2 in the 1993 flood (which focused on
the lower Missouri watershed). 75% of the sediment deposited on the North American
continent is overbank flood plain silt, which spills onto the flood plain when floods spill over
natural or manmade levees. At its widest point in the Yazoo Basin, the Mississippi flood
plain is more than 80 miles wide.
4.2.1 Mississippi River is the High Ground
The river is the high ground in the Mississippi Embayment (Figure 4.1). A vexing
problem with a high silt load river is that it tends to build up its own bed, which prevents
drainage of the adjoining flood plains. Sediment is deposited on the adjoining lowlands when
the river spills up out of its channel during flood stage. Sediments are hydraulically sorted
during this process, becoming increasingly finegrained and soft with increasing distance
from the river channel, as sketched in Figure 4.2. Millions of acres of flood plain swamps and
marshlands in the Mississippi Embayment downstream of Gape Girardeau, MO were
reclaimed by mechanically excavated drainage ditches, beginning around 1910, when large
railmounted dragline excavators became available. This machinery was also employed for
levee construction on the MR&T Project (after 1928) as well as drainage work for agricultural
reclamation.
4.2.2 Flooding from the Mississippi River
A great number of floods have occurred in the lower Mississippi Valley during
historic time, including: 1718, 1735, 1770, 1782, 1785, 1971, 1796, 1799, 1809, 1811, 1813,
42
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
1815, 1816, 1823, 1824, 1828, 1844, 1849, 1850, 1851, 1858, 1859, 1882, 1892, 1893, 1903,
1907, 1908, 1912, 1913, 1916, 1920, 1922, 1923, 1927, 1929, 1932, 1936, 1937, 1945, 1950,
1957, 1958, 1973, 1974, 1975, 1979, 1983, 1984, 1993, and 1997.
The most damaging to New Orleans were those in: 1816, 1826, 1833, 1849, 1857,
1867, 1871, 1874, 1882, 1884, 1890, 1892, 1893, 1897, 1903, 1912, 1913, 1922, 1927, 1937,
1947, 1965, 1973, 1979, 1993, and 2005. But, the last flood of any consequence to affect the
City of New Orleans emanating from the Mississippi River was in 1859!
New Orleans was founded in 1718. In April 1719 the town’s founder Jean Baptiste le
Moyne, Sieur de Bienville, reported that water from the Mississippi River was regularly
inundating the new settlement with half a foot of water. He suggested constructing levees and
drainage canals, and soon required such drainage work of all the landowners. In 173435 the
Mississippi River remained high from December to June, breaking levees and inundating the
settlement.
Flood protection from the Mississippi River was originally afforded by heightening of
the river’s natural bank overflow levees (Hewson, 1870), like those shown in Figure 4.3.
Crevasses, or crevassesplays, (Figure 3.8) are radiating tensile cracks that form in the bank of
a river, natural levee, manmade levee, or drainage canal. Crevassesplays are often triggered
by underseepage along preferential flow conduits, such as old sandfilled channels or the
radiating distributaries of previous channel breaks. For these reasons, crevassesplays often
occur at the same locations repeatedly.
On May 5, 1816 the Mississippi levee protecting New Orleans gave way at the
McCarty Plantation, in presentday Carrollton, and within a few day water filled the back
portion of the city, extending from St. Charles Avenue to Canal and Decatur Streets, flooding
the French Quarter. The water was only drained after a new drainage trench was excavated
through Metairie Ridge and channels connecting to Bayou St. John.
On May 4, 1849 the Mississippi River broke the levee at the Suavé Plantation at River
Ridge, 15 miles upstream of New Orleans. Within four days this water reached the New
Basin Canal, and within 17 days was flooding the French Quarter in New Orleans proper,
flooding the area down slope (north of) of Bienville and Dauphine Streets. The 1849 flood
waters rose at an average rate of one foot every 36 hours, which allowed residents ample time
to evacuate. Uptown residents thought about severing the levee along the New Basin Canal to
prevent water levels building up on their side, but those living on the opposite side of the
canal threatened to prevent such measures using armed force. Shortly thereafter the New
Basin upper levee collapsed, diverting flood waters to Bayou St. John and thence, into Lake
Pontchartrain. A nine foot deep lake developed in what is now the City’s Broadmoor area,
flooding 220 city blocks and necessitating the evacuation of 12,000 residents.
The 1849 crevasse at Suavé Plantation was eventually plugged by driving a line of
timber piles and piling up thousands of sand bags against these on the landside of the pile
wall. This work was of unprecedented proportions until that time and took six weeks to
complete before the river’s waters were once again confined to their natural channel.
Drainage trenches were then excavated through Metairie [distributary] Ridge to channel
ponded water out to Lake Pontchartrain. By midJune 1849 the water was finally receding and
43
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
residents began reentering their flooded homes, spreading lime to combat mold, mildew, and
impurities.
Between 1849 and 1882, four major crevassesplays occurred at Bonnet Carré, on the
eastern bank of the Mississippi River, about 33 river miles upstream of New Orleans. The
Bonnet Carré crevasses left a large fanshaped imprint on the landscape. In fact, during the
flood of 1849, a 7,000footwide crevasse developed at Bonnet Carré which diverted flow
from the Mississippi into Lake Pontchartrain for more than six months. This breach had to be
filled so sufficient discharge could flow down the main channel to allow ocean going vessels
to reach New Orleans.
The 1849 floods were the last time that the eastern bank of the Mississippi River was
breached affecting New Orleans proper. In 1858 high water lapped over the east bank levee,
but this was followed a few days later by a break on the west bank of the river (at Bell
Plantation), which drew down the high water threatening New Orleans. The Bell Plantation
crevasse remained open for six months. In 1859 the rear portion of New Orleans again
flooded, between Carrollton and Esplanade Avenues, flooding onethird of the City between
January and March.
The City of New Orleans and the Mississippi River became important battlegrounds
during the American Civil War between 186165. Early in the conflict a principal goal of the
Union forces west of the Appalachian Mountains was to sever the Confederacy along the
Mississippi River. Union forces had a distinct advantage insofar as they retained most of their
naval power, allowing them to blockade Confederate ports. General Ulysses Grant achieved
considerable notoriety for his early campaigns up the Cumberland and Tennessee Rivers, and
later, in the successful siege of Vicksburg, which gave northern forces control of the
Mississippi, isolating 40,000 Confederate troops west of the river, where they played no
further significant role in the conflict. Grant recognized the pivotal military role of the great
river, because it was his Army’s vital supply line. Grant turned to his engineers on numerous
occasions and ordered the construction of cutoffs (Figure 4.4), some of which were
successful, while others, such as that a short distance downstream of Vicksburg, were not.
The success or failure of the manmade cutoffs depended on a number of factors, such
as time of year, severity of the spring flood, and ability to meter flows into the cutoffs trying
to control the erosion caused by dropping the water over oversteepened gradients. These
experiences were drawn upon soon after the war (Hewson, 1870) to create an inland empire
through drainage of low lying swamps and construction of thousands of miles of privately
constructed levees to keep the river from flooding reclaimed tracts.
During the post Civil War boom that witnessed significant reclamation of floodprone
tracts in the Mississippi flood plain, a pattern of protection emerged as the established cities
like New Orleans battled the Mississippi: that being of adjacent breaks, upstream at Bonnet
Carré and downstream, in Plaquemines Parish, often providing “safety valves” that reduced
high water in the river along the New Orleans waterfront. The western bank would breach
again in 1893, at the Ames Plantation in Marrero. Breaks in adjoining areas gradually gave
rise to rumors about levees being purposefully undermined to save the more valuable property
within the city, which reached epic proportions during the record flood of 1927, when the
44
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
levees adjoining Plaquemines Parish were dynamited no less than seven times, by City
officials worried that their own protective works would crumble and give way (Barry, 1997).
Army Engineer A.A. Humphreys and civilian engineer Charles Ellet were funded by
Congress in separate contracts to make a scientific examination of the Mississippi River in
1850. Ellet completed his work in 1851, but Humphreys did not complete his report until
1861, after suffering a nervous breakdown (Barry, 1997). Humphreys exerted significant
control of the Mississippi River as Chief of the Army Corps of Engineers between 18661879.
He was the father of the Corps’ flawed “levees only policy” of flood control, which remained
in effect till the 1927 flood, which triggered the creation of the Jadwin Plan, embodied in the
Federal Flood Control Act of 1928 (Morgan, 1971; Shallat, 1994). The “levees only policy”
maintained that the Mississippi River could be constrained within its natural low flow channel
by extending its natural levees upward, assuming the channel would downcut its bed
vertically during high flows, thus remaining in an artificially confined channel. This logic
was hopelessly flawed in that it ignored the river’s serpentine curvature, which causes it to
loop on itself in a seemingly endless series of “meander belts” across the floodplain. Because
of this curvature, the channel is seldom symmetrical (as portrayed in Figure 4.2), but
generally exhibits marked asymmetry, like that shown in Figure 4.5.
In 1871 the Mississippi River once again spilled its eastern bank at Bonnet Carré, 33
miles upstream of New Orleans. The massive break diverted much of the river’s flow into
Lake Pontchartrain, raising its level. A strong north wind pushed lake water up into the
Metairie and Gentilly ridges, filling the thenexisting system of drainage canals. A levee on
the Hagan Avenue (now the Jefferson Davis Parkway) drainage canal gave way, flooding the
back side of New Orleans, including the Charity Hospital, a town landmark.
The 1927 flood was the largest ever recorded on the lower Mississippi Valley (Figure
4.6). The deluge was preceded by a record 18 inches of rain falling on New Orleans in a 48
hour period in late March 1927, which was followed by six months of flooding. The levees
that were supposed to protect the valley broke in 246 places, inundating 27,000 square miles
of bottom land; displacing 700,000 people, killing 1,000 more (246 in the New Orleans area),
and damaging or destroying 137,000 structures.
There was an enormous public outcry for the government to do something more
substantive about flood control. Fearing the worst, the political leadership of New Orleans
sought relief by dynamiting the Mississippi levee in Plaquemines Parish, downstream of New
Orleans. By the time promises were made regarding damage compensation and the necessary
permission was granted, the flood had crested and begun to subside. No less than seven
sequences of dynamiting ensued, all promoted by fear. The initial dynamiting of the
Caernarvon levee below New Orleans with 30 tons of dynamite devastated much of St.
Bernard and Plaquemines Parishes, and their residents were never remunerated in any
meaningful way for their damages. The saddest aspect of the dynamiting was that it was
unnecessary, as several levees gave way upstream of New Orleans, one the very afternoon of
the dynamiting, and the river level at New Orleans never regained its maximum crest during
the remainder of that record year (Barry, 1997).
45
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
4.3
The Mississippi River and Tributaries Project 1931-1972
The Corps of Engineers’ Mississippi River & Tributaries (MR&T) Project was
authorized by Congress in the Flood Control Act of 1928, which emanated from the Great
Flood of 1927 on the lower Mississippi River. At the time of its introduction it was referred
to as The “Jadwin Plan,” because Major General Edgar Jadwin was the Army’s Chief of
Engineers at the time it was issued, on December 1, 1927 (Jadwin, 1928). It was incorporated
into the Federal Flood Control Act of May 15, 1928, which authorized $325 million to the
Mississippi River Commission (created in 1879) controlled by the Corps of Engineers to
provide for flood protection along the Mississippi River between Cape Girardeau, MO and
HeadofPasses, LA. In essence the Mississippi River Commission adopted the Mississippi
River & Tributaries Project, and the commission’s responsibilities, annual budget,
expenditures and importance increased by an order of magnitude, where it remains moreorless today. Actual construction did not begin until 1931, when the authorized funds were
finally appropriated by Congress.
The original flood control plan selected a project flood of 2,360,000 cubic feet per
second (cfs) at the mouth of the Arkansas River and 3,030,000 cfs at the mouth of the Red
River. These figures were about 11% greater than the record 1927 flood at the junction of the
Mississippi and Arkansas Rivers and 29% greater than 1927 flood at the junction of the
Mississippi and Red Rivers, 60 miles downstream of Natchez, MS.
The Jadwin Plan proposed four major elements to control the flow of the Mississippi
River. These were: 1) levees to contain flood flows wherever practicable, or necessary to
avoid razing large sections of existing cities and transportation infrastructure; 2) bypass
floodways to accept excess flows of the river, passing these into relatively undeveloped
agricultural basins or lakes; 3) channel improvements intended to stabilize river banks, to
enhance slope stability and commercial navigation; and 4) improvements to tributary basins,
wherever possible. This category included dams for flood storage reservoirs, pumping plants,
and auxiliary channels.
The main stem levees (Figure 4.7) were intended to protect the Mississippi alluvial
valley against flooding by confining the river to its low flow channel. The main stem, or so
called “federal levees,” extend 1,607 miles along the Mississippi River, with another 600
miles along the banks of the lower Arkansas, Red, and Atchafalaya Rivers.
A vexing problem with maintaining 1,552 miles of flood control levees in the lower
Mississippi Valley has been the complex and everchanging foundations upon which they are
founded (Figure 4.7). In addition, channel curvature promotes undercutting of the outboard
banks of bends, often depositing these materials in semilinear stretches of channel a short
distance downstream, because of lower gradients. This sediment reduces freeboard and raises
flow levels, often beyond design assumptions. Crevasses are often sandfilled distributary
channels that form preferred seepage paths beneath the flood plain during high flow. These
high permeability corridors lie beneath earthen levees like ticking time bombs, waiting to
explode (areas indicated by red arrows on Figure 4.8).
The 1928 Jadwin Plan also sought to emplace storage facilities wherever practicable in
the four principal watersheds bordering the lower Mississippi Valley: the St. Francis Basin in
46
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
southeastern Missouri and northeastern Arkansas; the Yazoo Basin in northwestern
Mississippi; the Tensas Basin in northeastern Louisiana; and the Atchafalaya Basin in
southern Louisiana. Five flood control reservoirs were constructed in these basins as part of
the MR&T Project: Wappapello Dam and Reservoir in the St. Francis Basin; and four dams in
the Yazoo Basin: Arkabutla, Sardis, Enid, and Grenada.
Bypass floodways were constructed by the Corps of Engineers. These included: 1) the
Birds PointNew Madrid Floodway between Cairo, IL and New Madrid, MO (which depends
on a fuse plug levee in lieu of a spillway; only used once, in 1937); 2) The Old River or Red
River Landing Diversion structure, intended to divert half the project flood (1,500,000 cfs)
from the main channel into the Atchafalaya River through the Morganza and West
Atchafalaya floodways; 3) The Bonnet Carré bypass and floodway, a concrete spillway
capable of diverting 250,000 cfs into Lake Pontchartrain during periods of high flow, about
30 miles upstream of New Orleans. The locations of MR&T structures in close proximity to
New Orleans are shown on Figure 4.9.
This was followed by numerous channel improvements and stabilization measures
which have been implemented as needed along the entire course of the navigable river
channel, to enhance river bank stability and commercial navigation. The Corps typically
employs channel cutoffs to shorten the river channel and increase hydraulic grades, which
reduces flood heights. They employ armored revetments to retard channel migration and
meandering. Countless dikes have been employed to direct the river’s flow, beneath the
channel surface. Annual dredging is required to maintain navigable channels, as sediment is
deposited by seasonal high flows. These activities have combined to reduce the annual
sediment yield of the river by 60% (Kesel, 2003).
The Bonnet Carré bypass and Old River Control Structure (into the Atchafalaya
Basin) are major elements of the MR&T Project that protect New Orleans from a Mississippi
River flood by reducing the volume of flow that passes the city. The Bonnet Carré spillway
was the first structural element of the MR&T Project to be constructed, in 1931, and initially
used during the 1937 flood. It is opened up whenever the river level exceeds 19.0 to 19.6 ft in
New Orleans and can draft off 250,000 cfs into Lake Pontchartrain.
The Old River Control Structure was not authorized by Congress until 1954. It was
intended to draft off 600,000 cfs of the Mississippi’s flow during an extreme flood event and
prevent capture of the Mississippi River by the Atchafalaya River, which would have
occurred naturally by 1975 (because the flow distance of the Atchafalaya to the Gulf of
Mexico is only onethird the distance taken by the present channel of the Mississippi; see
Fisk, 1952 and McPhee, 1989). The Corps constructed the Old River Control Structure and
lock from 196163. The project was intended to divert 30% of the Mississippi River Project
Flood into the Atchafalaya Basin. The Old River Control Structure has only been used once,
during the Flood of 1973, when it nearly failed catastrophically (MRC, 1975; Noble, 1976).
In the wake of this failure, the capacity was doubled with construction of an auxiliary
structure, completed by the Corps of Engineers in 1986, doubling the bypass capacity at Old
River into the AtchafalayaMorganza Basin to 1,220,000 cfs.
The average height of the MR&T levees above the natural levees in the Gulf Region is
about 16 ft (Kolb and Saucier, 1982). The crest of the flood protection levee along the eastern
47
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
bank of the Mississippi River is 24.5 feet MGL at Carrolton in New Orleans, as shown in
Figure 4.10. The Lake Pontchartrain protection levee varies between 13.5 and 18.0 feet MSL,
as shown in Figure 2.6, 2.7, and 4.10. All of the neighborhoods north of Metairie Ridge lie
below sea level. The worst flooding scenario for New Orleans would be a breach of the
Mississippi River levee because of its elevated position, which would engender rapid erosion
and high spill velocities, which could overwhelm the City’s lowest neighborhoods before
residents could effect an escape.
From its inception, the 1928 Flood Control Act has been modified every few years by
additional authorizations from Congress, usually based on modifications requested by the
Corps of Engineers. These included expenditures for establishment of an emergency fund for
maintenance and rescue work (1930) and acquisition of lands for floodways, etc. These early
changes resulted in the Flood Control (Overton Act) Act of 1936, which established a national
flood control policy to be administered by the Corps of Engineers, beyond the lower
Mississippi Valley. Even with these sweeping changes, more acts followed in quick
succession throughout the late 1930s and 1940s (for instance, a 1937 act authorized $52
million for strengthening of levees following the disastrous 1937 flood in the Ohio and
Mississippi Valleys). This pattern of amended flood control acts and authorized expenditures
continued throughout the 1940s, 50s, 60s, and 70s, usually following flood years.
Today, 3,714 miles of flood control levees have been authorized for construction
under the Mississippi River & Tributaries Project. 3,410 miles of levees have been completed
and 2,786 miles are in place to grade and section. On the main stem of the Mississippi River,
1,602 miles of levees have been completed. Work on the main stem levees of the Mississippi
River is approximately 89 percent complete and work on tributary levees is approximately 75
percent complete.
4.3.1 Dimensions of Navigation Channels Maintained by the Corps of Engineers on the
Lower Mississippi River
Over the next 60 years Congress added new river borne transport projects, extending
up the Mississippi drainage and elsewhere, creating an intricate system of barge commerce
that demands constant maintenance, clearing, patching, and dredging. In addition to ensuring
flood protection, the Corps of Engineers was also charged with maintaining yearround
navigation for the Port of New Orleans, which was the nation’s second largest port facility
when MR&T project work commenced in 1931.
After the mouths of the Mississippi River had been opened and maintained in a navigable
state (the first jetty was completed in 1879), navigation interests lobbied Congress to establish
and maintain “feeder” channels to the Mississippi River and deepen the main stem channel to
accommodate more modern vessels, with deeper draft. In 1945 Congress authorized the
development of a navigation channel for oceangoing traffic in the lower reaches of the
Mississippi River. Over the past 60 years this system has been expanded greatly through a
series of Congressional acts, until today it consists of 12,350 miles of navigable inland
waterways. The depths and widths of the Mississippi River channel between Baton Rouge
and the Gulf of Mexico have been established as:
48
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
•
•
•
•
•
•
•
•
•
4.4
Baton Rouge to New Orleans 40 by 500 feet
Port of New Orleans 35 by 1,500 feet, with portion 40 by 500 feet
New Orleans to Head of Passes 40 by 1,000 feet
In Southwest Pass 40 by 800 feet
In Southwest Pass Bar Channel 40 by 600 feet
In South Pass 30 by 450 feet
In South Pass Bar Channel 30 by 600 feet
Mississippi RiverGulf Outlet 36 by 500 feet
Mississippi RiverGulf Outlet Bar Channel 38 by 600 feet
Flooding of the New Orleans Area by Hurricanes
Hurricanes strike the Louisiana Coast with a mean frequency of two every three years
(Kolb and Saucier, 1982). Since 1759, 172 hurricanes have struck southern Louisiana
(Shallat, 2000). Of these, 38 have caused flooding in New Orleans, usually via Lake
Pontchartrain. Some of the more notable events have included: 1812, 1831, 1860, 1893,
1915, 1940, 1947, 1965, 1969, and 2005.
In 1722 a hurricane destroyed most of embryonic New Orleans and raised the river by
8 feet. Had the river not been running low prior to the storm, the river might have overtopped
its banks by as much as 15 feet. In 1778, 1779, 1780 and 1794 hurricanes struck the New
Orleans area destroying many buildings and sinking ships. The worst storm of the early years
was “The Great Louisiana Hurricane” of August 9, 1812. It rolled over the barrier islands and
drowned Plaquemines and St. Bernard Parishes and the area around Barataria Bay under 15
feet of water. The parade ground at Fort St. Phillip was inundated by 8 feet of water and the
shoreline along Lake Pontchartrain was similarly inundated, though this was far enough
below the French Quarter to spare any flooding of the City.
The back side of New Orleans was afforded some natural protection by the Metairie,
Gentilly, and Esplanade Ridges, which are recent distributary channels of the Mississippi
River. These “ridges” were originally about 4 feet higher than the surrounding marshland, but
much of the former cypress swamps and marshes (comprised of compressible peaty soils)
have settled as much as 10 feet over the past 110 years , while the ridges, being underlain by
sand, have only settled 1 to 2 feet. The ridges performed as quasi flood protection levees
from storm surges emanating from Lake Pontchartrain during hurricanes. But the ridge also
prevented drainage from moving between the old French Quarter and Lake Pontchartrain.
The Carondelet, or Old Basin, canal was excavated between Basin Street and Bayou St. John,
which formed the one low point between the elevated Metairie and Gentilly Ridge channels.
The Old Basin Canal drained the French Quarter and allowed smaller craft to transit through
the ridge to Lake Pontchartrain.
In June 1821 easterly winds surged off Lake Pontchartrain and pushed up Bayou St.
John, flooding fishing villages and spilling into North Rampart Street until the winds abated
and allowed the water to drain back into the lake. It was an ominous portent of things to
come.
49
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
On August 16, 1831 “The Great Barbados Hurricane” careened across the Caribbean,
striking the Louisiana coast west of New Orleans. The area south of town was again
inundated by storm surge, while a three foot surge entered the city from Lake Pontchartrain.
The Mississippi levee at St. Louis Street gave way, flooding the French Quarter. Heavy rains
accompanying this storm added to the flooding and boats were the only means of moving
about for several days.
Southeastern Louisiana suffered through three hurricanes during the summer and fall
of 1860. On August 8th a fast moving hurricane swept 20 feet of water into Plaquemines
Parish. The third hurricane struck on October 2nd making landfall west of New Orleans. It
inundated Plaquemines, St. Bernard, and Barataria, causing a significant storm surge in Lake
Pontchartrain which destroyed 20 lakeside settlements, washing out a portion of the New
Orleans and Jackson Great Northern Railroad. Surge from this storm overtopped the banks
along the Old and New Basin drainage canals and a levee along Bayou St. John gave way,
allowing the onrushing water to flood a broad area extending across the back side of New
Orleans.
Between 1860 and 1871 the city avoided serious flooding problems caused by
hurricanes. In 1871 three hurricanes caused localized flooding, which proved difficult to
drain. Flooding emanating from storm surges on Lake Pontchartrain during these storms
overtopped the Hagen Avenue drainage canal between Bayou St. John and New [Basin]
Canal, spilling flood waters into the MidCity area. City Engineer W. H. Bell warned the city
officials about the potential dangers posed by the drainage canals leading to Lake
Pontchartrain, because the MidCity area lay slightly below sea level (as seen on the 1895
Brown map in Figure 3.22).
The record hurricane of October 2, 1893 passed south of New Orleans and generated
winds of 100 mph and a storm surge of 13 feet, which drowned more than 2,000 people in
Jefferson Parish, completely destroying the settlements on the barrier island of Cheniere
Caminada. This represented the greatest loss of life ascribable to any natural disaster in the
United States up until that time. Seven years later, in August 1900, a hurricane passed
directly over Galveston, TX, demolishing that city and killing between 6,000 and 8,000
people, which remains the deadliest natural disaster in American history. Prior to impacting
Galveston, that hurricane tracked westerly parallel to the Gulf Coast about 150 miles south of
New Orleans. Its flood surges were noted along the Gulf Coast, including Lake
Pontchartrain’s south shore (Cline, 1926).
Prior to Katrina’s landfall in 2005, the most damaging hurricane to impact New
Orleans was the Grand Isle Hurricane of September 29, 1915, a Category 4 event which
produced winds as great as 140 miles per hour (mph) at Grand Isle. It slowed as it made
landfall and eventually passed over Audubon Park, seriously damaging structures across New
Orleans. Electrical power was knocked out, preventing the City’s new pumps from
functioning. The wave crest height on Lake Pontchartrain rose to 13 ft, easily overtopping 6foot high shoreline levee, destroying the lakefront villages of Bucktown (at end of 17th Street
Canal), West End, Spanish Fort, and Lakeview (these lakeside settlements were swallowed up
by the infilling of the Lake Ponchartain shoreline in 192831). The drainage canals were also
overtopped, flooding the city behind Claiborne, leaving MidCity and Canal Street under
4 10
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
several feet of water. This storm overwhelmed the City’s defenses so quickly that 275 people
were killed, mostly in the Lake Pontchartrain shoreline zone.
On September 19, 1947 an unnamed hurricane made landfall near the Chandeleur
Islands, producing wind gusts between 90 and 125 mph, with 1 minute maximum of 110 mph.
A storm surge of 9.8 ft reached Shell Beach on Lake Borgne. The runways at Moisant
Airport were covered by 2 ft of water while Jefferson Parish was flooded to depths of 3+ ft.
Sewage from an overwhelmed S&WB treatment plant stagnated in some of the drainage
canals, producing sulfuric acid fumes that caused staining of leadbased paint on some of the
homes in the Lakeview area, leaving them with unsightly black blotches. 51 people drowned
and New Orleans suffered more then $100 million in damages. City officials were unable to
clear floodwaters through the drainage canals in the Lakeview, Gentilly, and Metairie
neighborhoods for nearly two weeks. This was the first significant hurricane to strike New
Orleans which generated a large body of reliable storm surge data, which was subsequently
used in design of flood protection works by the Corps of Engineers (Figure 4.11). The New
Orleans TimesPicyaune prepared a map that showed reported depths and locations of
flooding in the 1947 hurricane.
After the 1947 storm, hurricane protection levees were heightened along the south
shore of Lake Pontchartrain and extended westward, across Jefferson Parish (constructed in
1949). In addition, the embankments along the old drainage canals were raised by earthfill to
protect the Orleans and Jefferson Parishes from future storm surges off Lake Pontchartrain.
The precise height of these additions depended on position and historic settlement up till that
time. The entire Lakeview area north of what is now Interstate 610 (excluding the area filled
by the Lakefront Improvement Project) was already more than 2 ft below sea level by the late
1930s (WPALA, 1937).
Hurricane Betsy was a fast moving storm that made landfall at Grand Isle, LA on
September 910, 1965. Wind meters at Grand Isle recorded gusts of up to 160 mph and a 15.7
ft storm surge that overwhelmed the entire island. Winds gusts up to 125 mph were recorded
in New Orleans along with a storm surge of 9.8 ft, which overwhelmed both sides of the Inner
Harbor Navigation Canal (IHNC), flooding the Ninth Ward, Gentilly, Lake Forest, and St.
Bernard Parish areas (Figure 4.12), as well as all of Plaquemines Parish, causing the worst
flooding since 1947, and revealing inadequacies in the levee protection system surrounding
the city. 81 people were killed by the storm (58 in Louisiana), which was the first natural
catastrophe in America to exceed $1 billion in damages (USACE, 1965). Damage in
southeast Louisiana totaled $1.4 billion, with $90 million of that being to New Orleans.
In October 1965 Congress approved a $2.2 billion public works bill that included $250
million for Louisiana projects and $85 million down payment for a system of levees and
barriers around New Orleans (Figure 4.13). This work included raising the Lake
Pontchartrain levee to a height of 12 ft above Mean Gulf Level (MGL) in response to the
flooding caused by Betsy. The Orleans Levee Board also let contracts to pound steel
sheetpile walls along the crests of their drainage canal levees to increase their effective height,
so storm surges on Lake Pontchartrain would not overtop the drainage canals (which had
occurred in 1915, 1947, and 1965, but without catastrophic loss of the canal levees). The
uncased sheetpiles were intended to be a temporary measure, awaiting a permanent solution
4 11
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
that envisioned placement of concrete flood walls using the sheetpiles as their foundations,
funded by the Federal government. These shortterm improvements spared the city from
similar flooding in 1969 when Hurricane Camille struck the area.
Prior to Katrina, the only other Category 5 hurricane to make landfall on the United
States was Hurricane Camille in August 1969 (the atmospheric pressure on landfall was
second only to the Labor Day Hurricane of 1935). Camille made landfall on August 17th, its
eye crossing the Mississippi Coast at Pass Christian, about 52 miles east northeast of New
Orleans. Wind velocities in the eye of the storm reached 190 mph, while gusts on land
exceeded 200 mph, causing most wind meters to fail (the highest recorded gust was 175 mph).
Camille annihilated the coastal communities between Henderson Point and Biloxi, and caused
extensive flooding of 3,900 mi2 of coastal lowland between lower Plaquemines Parish and
Perdido Pass, AL. The peak storm surge measured 25 feet above MGL near Pass Christian,
MS (a record), 15 ft in Boothville, LA, 9 ft in The Rigolets, and 6 ft in Mandeville, LA. The
death toll from Camille was 258 people, with 135 of these being from the Mississippi coast (9
were killed in Louisiana). 73,000 families either lost homes or experienced severe damage
and the official damage toll was $1.4 billion, with damages in Louisiana totaling $350
million. A particularly vexing aspect of Camille was that it occurred just four years after
Hurricane Betsy, which had been touted as something between a 1in200 to 1in300 year
recurrence frequency event (USACE, 1965).
On September 28, 1998 Hurricane Georges wrecked havoc across the Caribbean,
pummeling Haiti, the Dominican Republic, Puerto Rico and other islands. Georges appeared
to be headed straight for New Orleans, but suddenly turned east, making landfall near Biloxi,
MS on September 28th (about 68 miles east northeast of New Orleans). Georges produced
sustained winds of over 100 mph at landfall, generating a storm surge of 8.9 ft at Point à la
Hache, LA. Maximum storm surge along the Gulf Coast was 11 ft, in Pascagoula, MS.
Hurricane Georges severely eroded the Chandeleur Islands in outer St. Bernard Parish.
Despite forewarnings and evacuation orders 460 people were killed, all outside of Louisiana.
Dozens of camps not protected by levees were destroyed along the south shore of Lake
Pontchartrain. Hurricane Georges provided the last preKatrina test of the vulnerability of
New Orleans levee protection system to hurricanes, and efforts resumed to improve the levee
system along the canals that connect the city with Lake Pontchartrain.
4.5
Flooding of New Orleans Caused by Intense Rain Storms
As mentioned previously, the New Orleans area receives an average of about 52
cumulative inches of rainfall each year. In the winter of 1881 severe rainstorms caused
flooding of the downtown area, up to 3 feet deep. Rain storms of severe intensity also caused
significant flooding of New Orleans in 1927, 1978 and 1995.
The 1927 storm dumped 14 inches on Good Friday, overwhelming the Sewerage &
Water Board’s vaunted system of Wood pumps, at least temporarily. Uptown streets were
flooded, with the Broadmoor and MidCity areas inundated by 6 feet of water and 2 feet in the
old French Quarter. This storm occurred simultaneously with the onset of the record high
flows along the lower Mississippi River, which lasted almost six months.
4 12
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
On May 3, 1978 a line of rain squalls approaching New Orleans from the west became
stalled over the city when it intersected a stationary front sitting over Lake Pontchartrain. The
resulting storm dropped 10 inches of rain during the morning, with a peak sustained intensity
of two inches per hour rain. The runoff exceeded the aggregate capacity of the city’s pumps
operated by the S&WB, causing extensive flooding of low lying areas that lasted about 24
hours.
A series of intense rain storms struck Louisiana, Mississippi, and Alabama in two
consecutive sequences in March and April of 1980. The first storm occurred from March 26
to April 2nd, striking southeastern Louisiana and portions of Mississippi. The second storm
sequence rolled through the same area from April 11 to April 13, affecting much of
Mississippi, but especially intense in the area bounded by Baton Rouge and New Orleans to
Mobile, Alabama. The 2hour rainfall in Mobile on April 13 had a recurrence interval of 100
years. As a result of this rainfall, Mobile experienced the worst flash floods in the city's
history. In New Orleans flood waters being pumped into the London Avenue Canal
overtopped the eastern side of the Canal just south of Robert E. Lee Boulevard, where steel
sheetpiles providing additional flood freeboard had recently been removed. This was the
same portion of the northern London Avenue Canal which subsequently experienced incipient
failure during Hurricane Katrina in 2005, and it moved two feet laterally (the area shown in
Figure 4.24 upper).
On the evening of May 89, 1995 a cold front approaching New Orleans from the west
stalled after moving east of Baton Rouge. A nearly continuous chain of thunder storms befell
the New Orleans area, dropping 4 to 12 inches of rain across New Orleans. The storm’s
intensity overwhelmed the S&WB’s maximum pump capacity (47,000 cfs) and almost the
entire city experienced severe flooding, including the Interstate highways. More severe
storms struck the coast the following evening, but the rainfall was not as severe over New
Orleans proper, though the two day totals reached a record 24.5 inches in Abita Springs, LA.
The 1995 storm sequence had a duration of 40 hours and damaged 44,500 homes and
businesses, causing $3.1 billion in damages. This was the costliest single nontropical
weather related event to ever affect the United States.
4.6
New Orleans Drainage Canals
The drainage canals of New Orleans are a unique feature of the bowlshaped city that
are much older than most people realize. The city’s first drainage canal was the Old
Carondelet Canal originally excavated in 1794, by order of Spanish Governor Baron de
Carondelet. It was dug by convicts and slaves and it was later enlarged to accommodate
shallow draft navigation (row boats and keel boats) between the City and Lake Pontchartrain.
Its name was later changed to the Basin Canal because it terminated at Basin Street, in the
French Quarter. Its name was later changed to the Old Basin Canal. It was infilled in the
1920s, when it became Lafitte Avenue and railroad tracks were placed down the street’s
centerline. Figure 4.14 shows the systems of drainage ditches and canals established by
1829, leading to Bayou St. John.
4 13
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
The New Basin Canal was excavated by Irish immigrants in the early 1830s in the
American Sector, but an outbreak of yellow fever killed 10,000 workers. The New Orleans
City Railroad paralleled this canal in post Civil War era. The New Basin Canal was the first to
cut through Metairie Ridge. The severing of Metairie Ridge was a double edged sword, as
flood waters came up the Old Basin Canal and inundated the downtown area in 1871. The
portion south of Metairie Ridge was filled in the 1930s; and the remainder in the 1950s, with
the Pontchartrain Expressway replacing the old canal.
The six piece Topographic Map of New Orleans and Vicinity prepared by Charles F.
Zimpel in 183334 suggest that portions of the Orleans Canal had been excavated and were
proposed to be extended by that date to convey water from Bayou Metairie to Lake
Pontchartrain (Lemmon, Magill, and Wiese, 2003). The Turnpike Road ran along the west
side of this canal. In 1835 the New Orleans Drainage Company was given a 20year charter
by the city to drain the cypress swamps between the riverbank and Lake Pontchartrain. The
company consulted State Engineer George T. Dunbar and evolved a scheme to drain the area
using underground canals beneath prominent uptown streets which would collect water and
convey it down the natural slope to the Clairborne Canal and then to the newly completed
Orleans Canal (then called the Girod Canal) into Lake Pontchartrain. This ambitious scheme
was derailed by the financial panic of 1837, though a system of ditches were completed which
conveyed runoff from the French Quarter to the upper Orleans Canal, from which it had to be
transferred to Bayou St. John using steampowered pumps.
A review of historic maps (Figures 4.15 thru 4.17) suggests that the Upper Line
Protection Levee or 17th St. Canal along the OrleansJefferson Parish boundary was
excavated between 1854 and 1858 (shown as completed). The 17th Street Canal is not
indicated on the 1853 Pontchartrain Harbor and Breakwater Map, although the Jefferson and
Lake Pontchartrain Railroad is shown along the OrleansJefferson Parish boundary. The 1858
map shows the 17th Street canal just east of the railroad tracks and the new village of
Bucktown, along the shore of Lake Pontchartrain adjacent to the mouth of the 17th St. Canal.
The 1878 Hardee map (Figure 4.17) calls the 17th St. Canal the “Upper Line Protection Levee
and Canal.” 17th Street was renamed Palmetto Avenue in 1894. The early rail lines serving
the docks on Lake Pontchartrain remained in operation for many years after the Civil War
(Figure 4.16).
Disastrous outbreaks of yellow fever in the 1850s spurred new ideas to drain the
cypress swamps. Between 185759 City Surveyor Louis H. Pilié developed a drainage plan
using open drainage canals with four steampowered paddle wheel stations to lift collected
runoff into bricklined channels throughout lower New Orleans, which was poorly drained
because the MetairieGentilly Ridge presented a natural barrier between the downtown slope
and Lake Pontchartrain (Figure 4.19). In 1858 the Louisiana Legislature divided the city into
four “draining districts,” providing a commission for each district and a method of assessment
for the operation and maintenance of drainage facilities. These names of these were the New
Orleans First and Second, Jefferson City, and Lafayette Draining Districts (Beauregard,
1859). In 1859 the legislature mandated issuance of 30year bonds totaling $350,000 for each
of the four districts. This allowed a program of local taxation to fund the pumps and maintain
the four lift stations, which were called “draining machines.”
4 14
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
These steampowered pumping machines were located at: the Dublin machine at the
head of the New Canal (old 17th St.) at Dublin and 14th Streets; the Melpomene machine at the
head of the Old Melpomene Canal (at Melpomene and Claiborne); the Bienville machine at
the head of Bayou St. John (at Hagan and Bienville); and the London machine (just north of
Gentilly and London Avenues). These facilities became a city trademark for many years
thereafter. Shortly before the outbreak of the American Civil War in 1861, the legislature
passed another bill that allowed any of the draining districts to make special assessments to
make necessary repairs, based on the recommendations of their respective boards.
Figure 4.16 is a portion of the Map of New Orleans area completed under direction of
Brigadier General Nathaniel P. Banks of the Union Army in February 1863, during the
American Civil War. This map shows the position of the Jefferson and Lake Pontchartrain
Railroad along the 17th St. Canal alignment, but not the canal itself. It also shows the New
Basin Canal (a short distance east), the upper Orleans Canal, feeder canals emptying into
Bayou St. John, and the Pontchartrain Railroad (near today’s IHNC), which operated between
18311932, its northern terminus being named Port Pontchartrain.
The upper end of the London Avenue Canal appears to have been constructed in the
1860s, north of Bayou Gentilly. One of the aforementioned steampowered draining
machines was located near the intersection of London and Pleasure Street, which lifted water
from the upper London Canal into the cypress swamp near what is now Dillard University,
north of Gentilly Ridge. Based on a comparison of the 1873 Valery Sulakowski map and the
1878 Thomas Hardee maps, the lower London Avenue Canal appears to have been extended
out to Lake Pontchartrain sometime between 187378.
In 1878 City Engineer and Surveyor Thomas S. Hardee compiled the most accurate
map of the City to that date, after a yellow fever epidemic that year which killed 4% of New
Orleans’ population (which brought to City’s accumulated death toll to Yellow Fever in
excess of 100,000 people). The map sought to delineate improvements for the city’s drainage
system to enhance sanitation. It would take another two decades before a substantive
drainage plan eventually evolved.
The New Orleans drainage dilemma can be appreciated from a review of the earliest
cross section drawn through the city, reproduced in Figure 4.19. The Mississippi River’s
natural levees form the highest ground in New Orleans. The natural levee slopes northerly
towards Lake Pontchartrain. This slope is interrupted by the MetairieGentilly Ridge, a
geologicallyrecent distributary channel, lying between 3 and 6 feet above the adjacent swamp
land.
The protection levee along Lake Pontchartrain (Figure 4.19) was erected after the
1893 hurricane, which generated a storm surge of up to 13 feet (described in Section 4.4).
This protective structure was known as the “shoreline levee” and was 6 feet above the normal
surface of Lake Pontchartrain. The creation of this structure was a doubleedged sword: it
served to keep rising water from Lake Pontchartrain out of the city, but also prevented gravity
drainage from the city into the Lake, except through drainage canals, into which runoff must
be pumped to gain sufficient elevation to flow by gravity into the Lake. Discharge could not
be conveyed to Lake Pontchartrain during hurricaneinduced storm surges. The gravity of
this problem was not fully appreciated until the 1915 Grand Isle Hurricane.
4 15
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
4.7
City Adopts Aggressive Drainage System
The failure of the Hagan Avenue Canal levee in 1871 signaled the beginning of a
political crisis, hastened by hurricaneinduced surges on Lake Pontchartrain. The City sought
to consider a better solution than it had heretofore employed in providing for reliable drainage
to Lake Pontchartrain, and vice versa. New Orleans City Surveyor W.H. Bell warned of the
potential dangers posed by the big outfall drainage canals. He told city officials to place
pumping stations on the lakeshore, otherwise “heavy storms would result in water backup
within the canals, culminating in overflow into the city.” This prophetic warning was ignored
with catastrophic results during Hurricane Katrina.
A new attempt to construct an integrated drainage system was undertaken by the
Mississippi and Mexican Gulf Ship Canal Company, which excavated many miles of canals
in New Orleans between 187178, before going out of business. By 1878 the City assumed
responsibility for maintenance of a 36mile long system of drainage canals feeding into Lake
Pontchartrain. The city’s old network of steampowered paddlewheel lift stations could only
handle 1.5 inches of rainfall in 24 hours, which represented slightly more than a nominal 1year recurrence frequency storm. This meant that the city began suffering flooding problems
with increasing frequency because of insufficient runoff collection, conveyance, and
pumping/discharge capacity.
The drainage problem was greatly exacerbated by a growing sewage treatment crisis.
The City’s population grew from about 8,000 in 1800 to nearly 300,000 residents by 1900.
The need for space enticed development into the low lying cypress swamps, which were being
reclaimed by construction of shallow drainage ditches feeding into the newly completed
system of drainage canals. In the 1880s houses began to appear on the old marsh and swamp
areas below Broad Street. No one regulated the inflow to the drainage canals and there was
an abject lack of a modern sewerage collection, conveyance, treatment, or outfall system.
Residents on the high ground near the Mississippi River could install pipes that conveyed
their effluent to the Mississippi River, but this was not a practical option for people living
below Broad Street, which lay below the river level.
The drainage crisis grew throughout the 1880s. In 1890 the Orleans Levee Board
offered $2500 for the best drainage plan for the troubled city, but no suitable plans were
submitted because of the paucity of reliable topographic data. In the wake of this
disappointing result, newspaper editorials and civic leaders recognized the city could not
continue growing without a substantive effort to handle drainage and sewage. After several
more unsuccessful attempts to encourage someone credible to come forward with a plan, in
February 1893 the City Council created a Drainage Advisory Board (DAB) and provided
$700,000 to gather the necessary topographic and hydrologic data, study the situation, and
make recommendations on how the problems might be solved. The DAB sought to gather
together the City’s best and brightest engineers from public, private, and academic ranks.
Chief among this work was the preparation of an accurate topographic map of the city,
prepared under the direction of City Engineer L. W. Brown (shown in Figure 3.22).
The first DAB’s findings were presented to the city in January 1895 (Advisory Board,
1895; Kelman, 1998). The Drainage Board recommended that the city create a modern
system of drainage collection, conveyance, and discharge, which included street gutters, drop
4 16
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
inlets, buried storm drains beneath city streets, with gravity flow to the principal drainage
canals leading to Lake Pontchartrain. At that juncture, the conveyance problems became
unprecedented, insofar that the city would need to install a series of pump stations to convey
collected runoff into Lakes Borgne and Pontchartrain. The projected cost of such a system
would be enormous.
The following year (1896) the Louisiana legislature authorized the creation of the
Drainage Commission of New Orleans, which began preparing a comprehensive drainage
plan for the city, and, a corollary plan to fund such work. In 1897 the Drainage Commission
began issuing contracts for new pumping stations, an electric power generation station, and
the construction of additional feeder canals into the existing network of drainage canals.
In June 1899 voters passed a municipal bond referendum in a special election, which
allowed a property tax of two mils per dollar to fund municipal waterworks, sewerage and
drainage. With this revenue mechanism in place, the Sewerage & Water Board (S&WB) of
New Orleans was shortly thereafter established (in 1899) by the State Legislature to furnish,
construct, operate, and maintain a water treatment and distribution system and sanitary
sewerage system. In 1900 the Drainage Commission began realigning and shifting the
existing system of drainage canals, filling in a number of the crosscutting canals and feeder
canals which contained much stagnant water, which was encouraging the proliferation of
mosquitoes and summertime yellow fever epidemics. In 1903 the S&WB was merged with
the Drainage Commission to consolidate operations under one agency for more efficient
operations. The drainage infrastructure at this time is shown in Figure 4.20.
The combined organization retained the name Sewerage & Water Board (S&WB),
which it retains today. S&WB then set about the Herculean tasks at hand, which more or less
continued at a feverish pace until the early 1930s, when the economic downturn caused by the
Great Depression curtailed revenue. By 1905 the S&WB had completed 40 miles of drainage
canals (in addition to the 36 they inherited), constructed six new electrically powered
pumping stations and had a pumping capacity of 5,000 cfs, which represented about 44% of
the original plan. At this time the S&WB provided drainage for 34.4 mi2 of city area, all on
the eastern side of the Mississippi River.
As the S&WB tackled the tough drainage problems plaguing lower New Orleans,
rapid development of these low lying areas ensued, with the real estate values increasing
dramatically, with many of the city’s residents engaged in speculation, purchasing lots and
then selling them as prices inflated. Because of this, many of the lots in lower New Orleans
were developed in different eras instead of all at once, leading to the heterogeneity of
architectural styles and ages that have made New Orleans neighborhoods famous. An
unforeseen downside of the rapid pace of development was the increase in runoff which
accompanied the emplacement of impervious surfaces, such as streets, roofs, sidewalks, and
the like, which increased drainage problems, necessitating enlargement of pump capacity each
decade.
By 1910 the S&WB system was rapidly being overwhelmed and something needed to
be done to increase capacity. A. Baldwin Wood was a young Sewer & Water Board
mechanical engineer who joined the Sewer & Water Board as assistant manager of drainage
upon his graduation from Tulane University in 1899. Wood was a retiring and shy
4 17
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
personality who took on the various challenges facing the S&WB with unparalleled
enthusiasm and imagination. Within a few years (at age 27 in 1906) Wood filed his first
patent, for a 6ft diameter centrifugal water pump that was the largest of its kind in the world.
After this he invented an ingenious flapgate that prevented backflow when the pumps were
not in use.
In 1913 Wood made his greatest contribution to the continued growth of New Orleans
when he introduced his novel design for the lowlift “Wood Screw Drainage Pump,” a 12foot
diameter screw pump that employed an enormous impellor powered by a 25 cycle per second
(or Hertz, abbreviated as Hz) Alternating Current (AC) electrical motor. The motive power
was highly efficient, using 20 feet diameter Allis Chalmers dynamos that spin up to 87 rpm.
The lowlift screw pumps employ a siphon action to maximize hydraulic efficiency. This was
followed in 1915 by Wood’s patented Trash Pump, capable of pumping record volumes of
water as well as flotsam and trash without risk of shutting down the pumps (Junger, 1992).
This latter feature was of particular value in maintaining pumping during storm events, which
brought large volumes of organic debris into the drainage canals. In 1915 the City let a
$159,000 contract for thirteen patented Wood screw pumps, installing 11 of them in three
pump stations by the end of the year, when the Grand Isle Hurricane struck the city, causing
widespread flooding of the old back swamps, which already lay at sea level. By that time
(1915) there were 70 miles of drainage canals in place.
By 1926 the New Orleans S&WB was serving an area of 47 mi2 with a 560 mile long
network of drainage canals and storm drains with a total pumping capacity of 13,000 cfs.
This impressive infrastructure had been constructed over a period of 47 years at a cost of
$27.5 million (18791926). Up to this time (1926) most of the S&WB’s revenue had been
generated by the special twomill tax on all property and half of the surplus from the 1% debt
tax. As the city grew and the S&WB’s jurisdictional area increased to other areas adjacent to
the city, the tax structure saw a number of amendments. Today the S&WB is funded by a
number of sources, including three, six, and ninemil property taxes.
The integrated drainage network allowed the water table of the old cypress swamps to
be dropped so that subterranean cellars and burials became possible, and deaths from malaria
and typhoid dropped 10fold between 18991925. The City’s last bout with summertime
yellow fever was in 1905 (Campanella, 2002). During this same interim (191526), the port
authority saw enormous growth with the development of a massive Army Supply Depot along
the riverfront during the First World War (191718) and the longanticipated completion of
the Inner Harbor Navigation Canal (IHNC) between the river and Lake Pontchartrain in mid1923.
In the mid1920s Wood increased the capacity of his patented screw pump to 14 feet
diameter, using the same powerful siphon action to lift water. This increased the capacity of
each pump unit by almost 40%. His improved capacity screw pumps were eventually
marketed across the world; in China, Egypt, India, and Holland. Wood retired from the
S&WB in 1945 and died in May 1956.
4 18
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
4.7.1
Pre-Katrina Conditions and Maintenance by the S&WB
Today the S&WB is responsible for draining 95.3 mi2 of New Orleans and
neighboring Jefferson Parish, which receive an average annual rainfall of 52 inches per year.
The general layout of the drainage system is presented in Figure 4.22. The preKatrina
system was intended to handle an average annual discharge of 12.9 billion cubic feet of water
that had to be collected and pumped into Lake Pontchartrain, Lake Borgne, and the
Mississippi River. The City’s 22 main pump stations and 10 underpass pump stations still use
about 50 of A.B. Wood’s old pumps, and their system can lift an aggregate total of 47,000 cfs
of water under peak operating conditions (the State Department of Transportation maintains
the pumps for the General DeGaulle underpass at the Mississippi River Bridge ramps and on
the East Bank at the Pontchartrain Expressway at the Southern Railway tracks and Metairie
Cemeteries). A typical pump station (Pump Station No. 6) can lift 9,600 cfs using its old
Wood pumps. New Orleans also employs vertical pumps with impellors to lift water from
subterranean (below street) storm drains to the drainage canals, which outfall in Lake
Pontchartrain. The S&WB maintains 90 miles of covered drainage canals, 82 miles of open
channel canals, and several thousand of miles of storm sewer lines feeding into their system.
The S&WB maintains that their agency installed two sets of piezometers along the
canal in the early 1980s, but that these revealed little correlation between transient flow levels
in the canals and the adjacent piezometers. They took this result to mean that the canal
floored in materials of relatively low permeability. In 1988 the S&WB received a permit
from the Corps of Engineers to deepen and widen the 17th Street Canal, based on the
“positive” indicators garnered from the piezometers that had been installed a few years
previous. The Corps warned that dredging might weaken the stability of the canal, but a
system of monitoring pore water (groundwater) pressures adjacent to the canal was not
undertaken and the canal was substantially enlarged using a trackmounted excavator.
Although the S&WB system is highly efficient from an energy expenditure
perspective, the 25 Hz AC electrical power requires the board to produce its own electricity,
in lieu of purchasing 60 Hz AC off the national electrical power grid. As a consequence,
approximately 60% of the S&WB’s electrical power has to be generated locally, at their own
20 MW generator stations (Snow, 1992). Unfortunately, all of these generating stations are
located below mean Gulf level and subject to shutdown by flooding.
4.7.2 Damage to S&WB Facilities and Capabilities Caused by Hurricanes Katrina and
Rita
During Hurricane Katrina the following pump stations were incapacitated and closed
due to flooding: Pump Station #1 (2501 S. Broad Street), #3 (2252 N. Broad St.), #4 (5700
Warrington Dr.), #6 (345 Orpheum), #7 (5741 Orleans Ave.), #10 (9600 Haynes Blvd.), #14
(12200 Haynes Blvd.), #15 (Intercoastal Waterway), #16 (7200 Wales St.), and #19 (4500
Florida Ave.). These pump stations were gradually brought back online and were all at least
partially operational within six months. 100% pumping capacity had not been restored to the
S&WB system by the time of this writing (May 1, 2006). Drainage for Jefferson Parish, west
of the city, remained online in wake of Hurricanes Katrina and Rita. This failure of the
4 19
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
S&WB drainage system was without historic precedent, and pointed to fundamental flaws in
the drainage system, with respect to operational redundancy.
During Hurricanes Katrina and Rita the Eastbank Sewer Treatment Plant was also
closed (and has not reopened as of June 1, 2006). City residents were immediately advised to
boil water before using it by the city’s Department of Health and Hospitals immediately
following flooding of the city. This restriction was lifted for the neighborhoods west of the
IHNC on October 6, 2005 and for the New Orleans East, Southshore and Ventian Isles areas
on December 8, 2005. Water quality had not been restored to The Lower Ninth Ward in Zip
Code 70117 by the time of this writing (June 1, 2006).
4.7.3 Reclamation of the Mid-City Lowlands (early 1900s)
The MidCity area occupies a natural basin that formed between the levees of the
Mississippi River and Metairie Ridge. The City’s original network of pieshaped property
boundaries and streets converged on this area from their points of origin perpendicular to the
broad crescentshaped bend of the Mississippi River upstream of the French Quarter, from
which the city derives its motto “the Crescent City.” The area was a closed depression
(Figure 4.18), which had to fill up with water to drain into Bayou St. John, thence three miles
into Lake Pontchartrain. A series of feeder canals were excavated to convey drainage into
Bayou St. John and the New Basin Canal after the Civil War. But stagnant water occupied
these feeder ditches, promoting the existence of mosquitoes and yellow fever outbreaks,
which were recognized to favor poorly drained areas decades before the scientific connection
between the two was established (beginning around 1905).
In the early 1900s it was decided to begin filling the lowest areas of the MidCity area
to provide better drainage and accommodate growth into this area, which had been subject to
frequent flooding. Sand from Metairie Ridge and from dredging of nearby canals was used to
provide the fill material and the feeder canals in this area were filled in and replaced with
buried storm drain pipes beneath the streets (discussed in Section 4.7).
4.7.4 1915 Flood Triggers Heightening of Drainage Canal Levees
On September 29, 1915 The Grand Isle Hurricane lifted the water level in Lake
Pontchartrain to 13 feet above mean gulf level. The Lake Pontchartrain shoreline levee and
many of the drainage canals were overtopped and much of the lower city flooded, killing 275
people. The City’s new pump system was overwhelmed when the power generating stations
for the new Wood screw pumps were flooded. After the 1915 flood, Sewerage and Water
Board General Superintendent George Earl ordered the levees along the drainage canals to be
raised approximately three feet, while the Pontchartrain shoreline levee was also raised. It is
not known if this work was carried out by the S&WB or the Orleans Levee District.
4.7.5 The Lakefront Improvement Project (1926-34)
The southern shore of Lake Pontchartrain supported a number of small commercial
wharves and fishing camps during the late 19th Century, including Milneburg, Spanish Fort,
and West End. Shanties and structures along the shore were founded on wood pilings. The
4 20
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
old Lake Pontchartrain shoreline levee had been constructed along the south shore to protect
New Orleans from flood surges off the lake around 1893. This levee was overtopped by the
storm surge on Lake Pontchartrain during the Grand Isle Hurricane in 1915 (described in
Section 4.4). This levee was difficult to maintain because the shoreline was actively receding
southward, towards New Orleans (Figure 3.16). In 1921 the Orleans Levee Board were
granted increased powers by the state legislature to reinforce the Pontchartrain shoreline. In
1924 the board’s chief engineer, Colonel Marcel Garsaud, embarked on developing an
ambitious plan to construct a permanent seawall along Pontchartrain’s south shore and
reclaim several square miles of land by filling the gap between the new seawall and the
eroding shoreline.
In 1926 the levee board began construction of a temporary wooden bulkhead wall
constructed onehalf mile north of the existing shoreline, within Lake Pontchartrain. This
temporary structure extended two feet above mean gulf level (MGL). The nearshore area
between this bulkhead wall was initially backfilled to an elevation of +2 feet above MGL,
creating 1,800 acres of “made ground.” The fill material was sand taken from the floor of
Lake Ponchartrain, placed using hydraulic dredges. The wooden bulkhead was then raised
another two feet and hydraulic fill placed behind it to a level of +4 ft. This process was
repeated yet again, creating a fill platform 4 to 6 feet above MGL and up to 10 ft higher than
the old cypress swamps that subsequently became the Lakeview and Gentilly neighborhoods
(even higher than the MetairieGentilly Ridge). The reclamation plan envisioned the
construction of a permanent stepped concrete seawall along the new shoreline, replacing the
wooden bulkhead wall, and construction of this permanent barrier began in 1930.
To offset the hefty price tag of $27 million for this work, the levee board secured
special legislation (in 1928) creating the Lakefront Improvement Project, which allowed them
sweeping powers to reclaim land along the Pontchartrain shoreline. In 193132 another
sizable fill was placed along Lake Pontchartrain behind another concrete seawall to create an
additional 300acre fill for a municipal airport. This was christened Shusan (now Lakefront)
Airport, which has a 6,900 ft runway, used as a flight training facility during World War II.
When the lakefront improvement project was completed in 1934, a public debate
erupted as to how best utilize the reclaimed land. A battle soon developed between private
development, public access to the shoreline, and those forces promoting its adoption as open
space parkland. A compromise plan was eventually adopted which allowed public access for
recreation along with residential and public facility development (University of New
Orleans). The new acreage was sold to developers to help the levee board pay off the
construction bonds, and the Lakeshore, Lake Vista, Lake Terrace, and Lake Oaks
neighborhoods were developed between 19391960.
After the Second World War the Lakeview, City Park, Fillmore, Gentilly, and
Pontchartrain Park areas behind the lakefront emerged as desirable bedroom communities
with yacht harbors, parks, and pleasant summer breezes. This area experienced unprecedented
growth, between 194575, adding about 100,000 residents to the City.
4 21
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
4.7.6 Second Generation of Heightening Drainage Canal Levee Embankments (1947)
The hurricane of September 1947 caused storm surges of up to 10 ft above MGL along
the shores of Lakes Borgne and 5.5 ft along the south shore of Lake Pontchartrain which
overwhelmed levees in the Inner Harbor Navigation Canal (IHNC) and the old drainage
canals, within a mile of their respective mouths. After several of these drainage canal levees
were overtopped in 1947, the state’s congressional delegation asked the federal government to
assist in protecting the city (culminating in the Lake Pontchartrain and Vicinity Hurricane
Protection Plan passed by Congress in 1955). The Orleans Levee Board spent $800,000 to
raise its levees, including both sides of their drainage canals (with the exception of 17th Street,
the west side of which is owned by the Jefferson Levee Board). Sheet piles were also
reportedly used in by the port authority in the inner harbor area. We have not been able to
determine how much additional freeboard was added by filling and/or sheet pile extensions in
194748.
4.7.7 Federal Involvement with the City Drainage Canals (1955 – present)
Federal involvement in the city’s drainage canals began in 1955 with approval of the
Lake Pontchartrain and Vicinity Hurricane Protection Project by Congress. The Corps studied
the problems posed by the drainage canals, which had settled as much as 10 feet since their
initial construction in the mid19th Century. This settlement had necessitated two generations
of heightening following hurricaneinduced overtopping in 1915 and 1947. Each of these
upgrades likely added something close to three additional feet of embankment height to keep
water trained within the drainage canals and provide sufficient freeboard to prevent storm
surges emanating from Lake Pontchartrain from overtopping the canal levees. The maximum
design capacity of the three principal drainage canals (17th Street, Orleans, and London
Avenue) was about 10,000 cfs, but this figure was being reduced by settlement and
sedimentation problems.
The Corps had several nonfederal partners in the venture: the Orleans and Jefferson
Parish Levee Boards, and the Sewerage & Water Board of New Orleans. The levee districts
maintained the canals and the S&WB maintained the pump stations and controlled the
discharge in the drainage canals. If the S&WB pumped at maximum capacity, the increased
flow could accelerate erosion of the unlined canals, which floor in extremely soft soils. If
they didn’t pump much water, then the canals could fill up with sediment, and thereby
experience diminished carrying capacity. By the time the Corps got involved, a dense
network of single family residences abutted the drainage canals along their entire courses (the
canals are 21/2 to 31/2 miles long). The encroachment of these homes adjacent to the canal
embankments circumvented any possibility of using conventional methods to heighten the
levees, which is usually accomplished by adding compacted earth on the landside of the
levees (Figure 4.23, which would require the condemnation and removal of hundreds of
residences, which would be costly and timeconsuming (not to mention unprecedented).
In 1960 the Corps of Engineers New Orleans District office issued its initial report
detailing their plan for remedying the ongoing problems with the slowly sinking drainage
canals. The Corps plan opted to solve the drainage canal freeboard problem by installing tidal
gates and pumps at the drainage canal outfalls along Lake Ponchartrain. This obviated the
need for condemning all the homes built along the canal levees. The Corps soon found itself
4 22
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
embroiled in a clash of cultures and goals with the levee districts, the S&WB, and the local
citizenry, who flatly opposed the Corps’ proposal. The S&WB and local residents feared that
the tidal gates would malfunction, inhibiting outflow of pumped storm water, which would, in
turn, allegedly cause flooding.
The following year (1961) the Corps of Engineers unveiled a more grandiose plan to
provide hurricane flood protection for New Orleans by constructing large flow barriers at the
passes (The Rigolets) leading into Lake Pontchartrain, to prevent storm surges from reaching
the lake. This scheme was expensive, and never garnered sufficient political support to gain
appropriations (it was also proposed in the era before environmental assessments were
required).
The issue of how to address improvement of the drainage canals dragged on for
another 17 years. Between 196077 what few lots remained in lower New Orleans were
rapidly built out, and most of the post1970 development in New Orleans focused on the areas
east of the IHNC, in Jefferson Parish (west of New Orleans), and across the Mississippi River
(Algiers, etc). In 1977 the U.S. Circuit Court of Appeals ruled against the Corps of Engineers
plans for tidal gates at the mouths of the drainage canals because the Corps failed to examine
the impacts of alternative schemes. From this juncture, the Corps focus shifted to heightening
the drainage canal levees using concrete walls (Figure 4.24lower), which was what the
opposing groups desired. These walls were to be designed to withstand a Category 3 storm
surge with 12 ft tides and 130 mph winds.
Construction began in 1993, but the wrong benchmark datums were selected for the
contract drawings, so some of these walls were constructed almost two feet lower than
intended (IPET, 2006). Although the concrete flood walls were completed by 1999, concrete
skirt walls on several of the bridges crossing the drainage canals had not yet been completed
when Hurricane Katrina struck on August 29, 2005. So, the drainage canal system was not
fully “tight,” but it was generally believed that it could survive a Category 3 storm surge by
surviving 6 to 8 hours of overtopping. The design storm surge values used by the Corps of
Engineers are reproduced in Figure 4.25.
Records for the drainage canals in New Orleans indicate that between 19322005,
water levels in these canals exceeded a flow stage of greater than +4 ft MSL on at least 29
occasions; +5 feet was exceeded 13 times (including during Hurricanes Betsy in 1965 and
Camille in 1969; +6 ft was exceeded only three times (including during Hurricanes Juan in
1985 and Isadore in 2002); and exceeded +7 feet for the first and only time on August 29,
2005 , during Hurricane Katrina.
4.7.8
Hurricane Katrina strikes New Orleans – August 2005
A complex network of levees protected the City of New Orleans from flooding (Figure
4.26). New flood walls were constructed in the 1990s on the crowns of drainage canals and
the Inner Harbor Navigation Canal to accommodate functionality during high storm surges.
The walls in the lower Lakeview and Gentilly Districts topped out at +14 ft above MGL.
This system of flood walls quickly failed on the morning of August 29, 2005, when
water levels rose more than 7 feet above MSL, higher than ever previously recorded in the
4 23
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
drainage canals since 1932 (cited in previous section). Prior to Hurricane Katrina, the
drainage canals feeding into Lake Pontchartrain never exceeded a flow height of between 6
and 7 feet above MGL. Many of the recording tidal gages failed during Hurricane Katrina.
The incomplete record of the gage located closest to the 17th Street Canal failure is
reproduced in Figure 4.27. This record shows several interesting trends. The first is the
increase in diurnal high tide level each day after August 22nd. The second is a dramatic
departure from the normal tidal cycle beginning the day before Hurricane Katrina made
landfall, around 5 PM on August 28th. The third interesting aspect is the sharp increase in
surge level on the morning of August 29th, which is much steeper than the assumed design
storm surge for Lake Pontchartrain shown on the lowest curve in Figure 4.25.
4.8
Commercial Navigation Corridors
4.8.1 Inner Harbor Navigation Canal/Industrial Canal
Ever since the founding of the city by the French in 1718, the concept of a navigation
channel between the Mississippi River and Lake Pontchartrain had been proposed, which
would allow intercoastal commerce to connect with river and seaborne commerce traveling up
and down the Mississippi River. The Port Authority of New Orleans was established in 1896
as an agency of the State of Louisiana. The port engineers recognized that the problem with
establishing a water borne link was the fluctuating flow of the river, which raised and lowered
20 feet, depending on flood stage. The river was also 10 to 26 feet higher than the normal
level of Lake Pontchartrain, so some impressive locks would be needed to control the flow
between the river and the lake.
The idea never progressed too far until construction of the Panama Canal between
190614, which heralded advances in excavation and grading technology that allowed
widespread programs of public works, drainage, and flood control in the succeeding half
century. In July 1914 New Orleans received authorization from the state legislature to locate
and construct a deep water canal between the Mississippi River and Lake Pontchartrain,
which was supposed to boost the capacity of the port by as much as 100%. America’s entry
into the First World War triggered the rapid expansion of ship building facilities and
construction of an enormous Army Supply Depot along the river front. While the war was
still raging, a committee was formed early in 1918 to examine the feasibility of a connecting
canal, using the most modern technology. Their initial report was released in May 1918 and it
surprised everyone by envisioning a much larger project than most supposed, with the
creation of ship building facilities within a protected, fixedlevel harbor, increasing the
available wharf space by almost 60%. The canal would be 5.3 miles long and up to 1,600 ft
wide, located just downstream of the Army’s new riverfront Supply Center (about 2 miles
downriver and parallel to Elysian Fields Avenue). A key aspect of its location was the 1911
donation of a pieshaped tract of land owned by the Ursuline Nuns which covered about half
of the proposed route, contiguous with the Mississippi River.
The Port Authority’s Dock Board retained the services of the George W. Goethals
Company as consulting engineers, borrowing upon General Goethals renown as chief
engineer of the Panama Canal project a few years earlier. The local firm of J. F. Coleman
Engineering Co. performed most of the actual detailed design work, as well as assisting the
4 24
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Port Authority in construction management. Construction commenced on June 6, 1918. The
superior elevation of the Mississippi River dictated that excavation would necessarily proceed
from the lake side towards the river, and the massive locks, the project’s kingpin structure,
would be placed at the river end of the canal.
Excavation work initiated with the construction of parallel dikes on either side of the
proposed canal, from which hydraulic fill could be loosed through sluice pipes. Hydraulic
excavation was used wherever possible to excavate the channel, when the materials were
easily loosed (e.g low cohesion materials, such as gravel, sand, organic ooze and swamp
muck). When more resistant clay was encountered large front tower cableway dragline
excavators or conventional dragline excavators (Figure 4.29) were employed to scoop out the
clay and drag it up onto the dikes, which were gradually built up to become permanent
protective levees. The draglines employed 3.5 cubic yard buckets and could handle about 150
cubic yards per hour. From the onset, contractors battled problems with slope stability, as the
soft oozy soils constantly slid back into the excavation (Campanella, 2002). Buried cypress
stumps slowed progress by jamming suction dredges and stalling dragline buckets.
During construction the Port Authority decided to increase the size of the channel to a
minimum depth of 30 feet at low water, with a minimum bottom width of 150 feet and a
minimum channel width of 300 feet, roughly double the original design. Abreast of the new
wharves the bottom width was increased to 300 feet, with a minimum canal width of 500 feet
near piers and slips, and 600 feet adjacent to quays (Dabney, 1921). The canal excavation
was completed in just 15 months, in September 1919. Everyone’s attention then turned to the
lock structure, located 2,000 ft from the Mississippi River, at the south end of the canal. The
normal flow level of the river was 10 ft above that of Lake Pontchartrain, so cofferdams had
to be constructed on either end of the locks to allow safe access and dewatering of the
exposed foundations. The lock is 640 ft long and 74 ft wide. The footing excavations were
50 feet deep, where timber piles were pounded into the underlying sands. The lock structure
was finally completed on January 29, 1923, and dedication ceremonies for the entire Inner
Harbor Navigation Canal (IHNC) were convened on May 5th, 1923. The residents of New
Orleans often refer to the IHNC as the “Industrial Canal.”
Almost immediately upon completion, the Port Authority set about developing piers,
docks, and quays to increase cargo handling. Their first large structure was the Galvez Street
Wharf, which was 250 ft wide and 2,400 ft long, costing $1.8 million (1923 dollars),
completed in 1924. It was constructed of reinforced concrete and fitted with tracks for a local
Beltline railroad. The Port Authority also made available adjacent lands for use by industries,
but it took many years until the envisioned development occurred. The IHNC benefited from
the completion of the Intracoastal Waterway in the mid 1930s, as a cargo handing and
provisioning stop. This was an unforeseen benefit, serving smaller vessels, which provided
an economical means of transport prior to the establishment of the Interstate Highway
network in the 1960s.
The massive Florida Avenue Wharf was added during World War II while the Gentilly
Road section of the canal witnessed the sprawling expansion of shipbuilding facilities
operated by Andrew Jackson Higgins, who pioneered the development of wooden PT boats
and landing craft crucial to the war effort. Much of the area flanking the west side of the
IHNC was built out during World War II (Figure 4.29). The eastern side was developed
4 25
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
much later, after the Korean War (195053) and completion of the MRGO channel in 1964
(Figure 4.28). The immense France Road and Jordan Road Container Terminals (Berths 5
and 6) near the head of the MRGO channel were completed in the 1980s and 90s. The
narrow width of the 1923 lock (74 feet) has restricted the passage of commerce, in particular,
river barges, which often wait up to 36 hours to pass through.
4.8.2 Flooding problems around the IHNC
During the 1947 hurricane (Figure 4.11) a back protection levee adjacent to the IHNC
was overtopped at Tennessee Street, spilling 10 feet of water into the East Side of New
Orleans. Fortunately, the levee did not collapse, the area was undeveloped, and the flooding
was quickly cleaned up. There was also quite a bit of flooding in the Metairie and Jefferson
Parish areas, also attributable to temporary overtopping. There was a flood inundation map
published in the New Orleans Times-Picayune.
Both sides of the IHNC experienced breaks and overtopping during Hurricane Betsy
in September 1965. 6,560 homes and 40 businesses were flooded in water up to 7 ft deep on
the west side of the IHNC. The east side of the IHNC also failed, flooding the west end of St.
Bernard’s Parish. A map of the flood inundation of New Orleans caused by Hurricane Betsy
in September 1965 is shown in Figure 4.12. The Corps’ report on Hurricane Betsy (USACE,
1965) states that both internal levee failures and overtopping occurred along the Inner Harbor
Navigation Canal, on both the west and east sides. No details about the mechanisms of failure
were described, however.
The IHNC was heightened using steel sheetpiles and concrete Iwalls in the 1980s and
90s. On August 29, 2005 during Hurricane Katrina both sides of the IHNC were overtopped
by the storm surge converging on the IHNC from Lakes Borgne and Pontchartrain.
Sustained overtopping flow undermined the landside toe of the Iwalls, in places gouging
down as much as 5+ feet below the crest of the earthen levee. In addition, there was ample
physical evidence of underseepage at both the eastern IHNC breaches, in the form of linear
sand boils.
4.8.3 Gulf Intercoastal Waterway (GIWW)
The Intracoastal Waterway (GIWW) was originally conceived in 1808, but was not
authorized by Congress until 1919. The GIWW was excavated by dredge in the late 1930s to
a channel size measuring 9 ft deep by 100 feet wide, and completed between New Orleans
and Corpus Christi, Texas by mid1942. This was enlarged to 12 feet deep by 125 ft wide
channel and officially completed in June 1949. The GIWW forms a protected shipping lane
between Port Isabel, Texas (the Mexican border) and Apalachee Bay, Florida. The first 15%
of the Mississippi RiverGulf Outlet Channel follows the GIWW, which then diverges
northeastward, about five miles east of the Inner Harbor. The GIWW then runs east, towards
The Rigolets and onto the Mississippi coast.
4.8.4 Mississippi River Gulf Outlet
When the IHNC was completed in 1923 the Port Authority announced that it intended
to lobby the federal government to construct a Mississippi River Gulf Outlet (MRGO)
4 26
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
channel connecting to the IHNC, to increase shipping capacity (Dabney, 1921). The idea
didn’t surface appreciably until 1943, during the Second World War, when thousands of
amphibious assault craft and shallow draft vessels were being fabricated along the nation’s
inland waterways. The Corps of Engineers felt that a tidewater canal serving New Orleans
and the nation’s interior waterways would be able to compete with the Panama Canal for east
west shipping, crucial to the war effort (most industrial goods were manufactured in the
eastern United States, which was being shipped to the Pacific via the Panama Canal).
Competing priorities placed the project in limbo until the late 1940s, when it was resurrected.
In the early 1950s the project was repeatedly voted down in Congress, because of competition
with the St. Lawrence Seaway project between Canada and the U.S (approved in 1954).
After passage of the competing seaway, the Mississippi River Gulf Outlet (MRGO)
project was authorized by Congress in March 1956. Kolb and Van Lopik (1958) of the Corps
of Engineers prepared a geology report on the MRGO alignment in 195758. This study
showed that the upper 2 to 5 feet was mainly fibrous peat, although highly organic marsh
deposits extend to depths of between 5 and 16 feet. These highly compressible materials are
underlain by interdistributary and intratidal complex silts and clays over much of the proposed
alignment (Figure 4.31). They graded these materials as soft marsh (500 to 900% water
content), firm marsh (100 to 500% water content), and swamp substrate (highly organic peat
with 600 to 800% water content). They noted that the soft marsh and swamp substrate
materials would be unable to provide competent foundations for the protective levees
bordering the channel, and these same materials would be unsuitable for use in such
embankments.
During the first phase of dredging in 195859, 20 million cubic yards (mcy) of
material was excavated between the IHNC and Paris Road (now I510), essentially widening
the GIWW. In 195960 contractors excavated a “pilot channel” between the GIWW and
Breton Sound, excavating and placing 27 mcy of material. In the third and fourth phases
completed between 196065, 225 mcy were excavated between Paris Road and Breton Sound.
Dredge spoils were placed in a strip of land 4000 ft wide along a corridor paralleling the
southwest side of the MRGO channel in St. Bernard Parish. The dredge soils from the initial
excavations (195859) were placed on the land which now underlies the Jourdan Road
Container Terminal, near the intersection of the MRGO and IHNC.
The MRGO channel was excavated as 500feet minimum width channel with a
minimum (low tide) depth of 36 feet (excavated to 38 feet; accepted at 36 ft). The route of
the MRGO channel crosses 45 miles of delta marshland in Orleans and St. Bernard Parishes,
with another 30 miles of open (dredged) channel across Breton Sound. This offshore section
is slightly larger. Its 75 mile path is 37 miles shorter than that of the deep water navigation
channel connecting New Orleans to the Gulf of Mexico via Southwest Pass. The project was
finalized in 1968.
The flanking levees have experienced significant settlement since the project’s
completion, due to consolidation of prodelta clays underlying the flanking levee
embankments, as well as plastic sagging due to low strength and creep properties of
underlying organic material. The amount of settlement varies between 1.5 and 8 feet,
depending on location. Many estimates have been offered regarding the tectonic rate of
subsidence of the Mississippi Delta; from 0.4 ft/century (Saucier, 1963) to as much as 1.3
4 27
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
ft/century (Watson, 1982). The Corps of Engineers authorized two sequences of levee
heightening to keep pace with ongoing settlement, but the third was delayed by funding
problems and had not been emplaced when Hurricane Katrina struck in 2005.
Since its completion, the seaway has eroded to a width of 2000 ft in places (Coastal
Environments, 1984), due in large part to ship wakes in the relatively confined channel. In
addition, siltation necessitates ongoing dredging, which cost the Corps of Engineers about $16
million per year. Salt water intrusion along the channel has impacted adjacent marshes,
although significant quantities of salt water have not been conveyed inland during hurricanes,
because the channel’s width is relatively insignificant when compared to adjoining bodies of
water, such as Breton Sound and Lake Borgne.
During Hurricane Katrina the levees fronting the MRGO channel were overtopped by
the nearrecord storm surge that came from the east off of Lake Borgne. The overtopping
caused by the severe storm surge quickly eroded the MRGO frontage levees in those reaches
where the levees were comprised of materials with little of no cohesion and high organic
content. In long stretches the entire levee was washed away down to its original marsh
foundations without a trace (Figure 4.32).
4.9
Influence of Elevation Datums on New Orleans Flood Protection System
4.9.1
Introduction
Persistent subsidence of the Gulf Coast/Mississippi River Delta region has led to a
complex relationship between the various geodetic datums used during historic surveys of the
area. The underconsolidated and organic rich sediments of the Mississippi Delta are
continually subsiding due to their compressible nature, the biochemical oxidation of the
entrained organics, and all the other factors described in Section 3.7. Tectonic activity along
active normal faults is also contributing to subsidence of nearly the entire Gulf Coast region.
Rates of subsidence are highly variable throughout the region, resulting in a complex
relationship between different geodetic datums at benchmarks in the New Orleans area.
Subsidence combined with a slow rise in sea level (about 1 ft per century) has caused much of
the Gulf Coast Region surrounding New Orleans to drop ten or more feet relative to sea level
in historic times, both of which have made the city more vulnerable to tropical storms.
It is important to accurately determine elevations in relation to sealevel in order to
design and construct flood protection systems in areas vulnerable to tropical storms.
Unfortunately, outdated terrestrial datums were referenced when constructing many of the
floodwalls protecting New Orleans. Variations of the NGVD29 datum were used, which is
based on terrestrial reference points, not sea level. The use of the outdated datums also
neglected subsidence and sea level rise, resulting in a lesser protection height than intended in
the floodwall designs. The subsidence of the region has made the correlation of datums a
complex task. No single conversion factor may be used when converting between two
datums.
4 28
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
4.9.2
17th St. Outfall Canal
Between 1952 and 2005, there has been a 2.345 foot decrease in the elevation of the
benchmark ALCO at the mouth of the 17th St. Outfall Canal due to subsidence and adjustment
of datums. In 1952, the benchmark elevation was 8.235’ while it had decreased to 5.89’ by
2005 (post Katrina) according to the NGVD29 (1952) and LMSL (19831992) datums,
respectively.
When the concrete Iwalls were placed atop the 17th St. Outfall Canal Levees during
the 1990’s, their tops were to extend to an elevation of 14.0 feet according to the NGVD
datum. Contract reports do not specify which NGVD epoch was to be used in design and
construction. It is possible that NGVD29 (09 Apr 1965) was used. In addition, NGVD is a
terrestrial datum and is not directly referenced to sea level as is LMSL. The top of the 17th St.
Outfall Canal Floodwall is presently between 1.3 and 1.9 below the design level of 14.0 feet
according to LMSL (19831992). This is likely due to the use of an outdated datum (1.6 feet
of difference) and settlement of the levee embankments and floodwalls (0.3 feet).
4.9.3
London Ave. Outfall Canal
The floodwalls bordering the London Ave. Outfall Canal were also designed and built
during the 1990’s. According to contract documents, the NGVD29 (09 Apr 1965) datum was
used. The use of an outdated, terrestrial datum in conjunction with settlement has resulted in
the floodwall heights being 1.61.8 feet below their intended heights of 14.4 feet (LMSL
(19831992)).
4.9.4 Orleans Outfall Canal
The NGVD29 (01 Sep 1982) datum was referenced during the design and construction
of the Orleans Outfall Canal floodwalls in the 1990’s. Presently, the floodwalls surrounding
this canal are up to 0.8 feet lower than called for than the 14.014.9 foot elevation called for in
the designs (according to LMSL (19831992)).
4.9.5
Inner Harbor Navigation Canal – East Levee
Floodwalls were placed atop the Inner Harbor Navigation Canal’s East Levee in 1970. The
walls were to extend to 15.0 feet (MSL) according to the 1969 contract documents. MSL was
tied to an earlier terrestrial datum and the exact correlation to modern adjustments has yet to
be determined. Floodwalls presently reach heights between 12.3 and 13.2 feet according to
the LMSL (19832001) datum.
4.9.6 Inability to Apply Universal Corrections for Elevation Datums
Although subsidence has played a role in the differences between designed and actual
floodwall heights, most of the variance appears to have been caused by datum abnormalities.
It is standard engineering practice to use an NGVD datum to determine sea level. The use of
NGVD is not cause for concern in portions of the country away from coastlines but becomes
troublesome in areas at or just above sea level.
4 29
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Due to the highly variable rates of subsidence throughout the region, a common
conversion factor cannot be used to adjust between datums, even over a short distance. The
complex relationships between the various geodetic datums in the New Orleans Region are
not discussed in great detail in this report. A more thorough discussion of this subject is
presented in Chapter III of IPET’s Second Interim Report (IPET, April 2006).
4.10
Names of New Orleans Neighborhoods
Figure 4.34 presents the official neighborhood names recognized by the City of New
Orleans. Local residents also use local ward and district numbers, and parish names to
describe an area. A common example would be the Lakeview and Gentilly areas, which are
used in a general sense to describe the former Cypress swamplands that now are among the
City’s lowest lying areas. The “Lakeview district” more or less encompasses Lakewood West
End, Lakewood, Lakeview, Navarre, and City Park neighborhoods. The “Gentilly district”
more or less includes the Fillmore, St. Anthony, Dillard, Milneburg, Gentilly Terrace,
Pontchartrain Park and Gentilly Woods neighborhoods.
4.11
References
Advisory Board [New Orleans]. (1895). Report of the Drainage of the City of New Orleans by
the Advisory Board, Appointed by Ordinance No. 8327, Adopted by the City Council
[of New Orleans], November 24, 1893.
Barry, J. M. (1997). Rising Tide: The Great Mississippi Flood of 1927 and How It Changed
America, Simon & Schuster, New York.
Beauregard, P.G.T. (1859). Report on Proposed System of Drainage for First Draining
District, New Orleans. Chief Engineer’s Office, U.S. Corps of Engineers, New
Orleans LA.
Campanella, R. (2002). Time and Place in New Orleans: Past Geographies in the Present
Day, Pelican Publishing.
Chatry, F.M. (1961). “Flood Distribution Problems below Old River.” Transactions of the
American Society of Civil Engineers, 126(1), 106119.
Cline, I. M. (1926). Tropical Cyclones, Comprising an exhaustive study of features observed
and recorded in sixteen tropical cyclones which have moved in on gulf and south
Atlantic coasts during the twenty-five years, 1900-1924 inclusive, Macmillian Co.,
New York.
Coastal Environments, Inc. (1984). The Mississippi River Gulf Outlet: A Study of Bank
Stabilization. Consultant’s report to St. Bernard Parish Police Jury, U.S. Department
of Commerce National Oceanic and Atmospheric Administration, and State of
Louisiana Department of Natural Resources. Baton Rouge. 127 p.
Dabney, T.E. (1921). The Industrial Canal and Inner Harbor of New Orleans: History,
Description, and Economic Aspects of Giant Facility Created to Encourage Industrial
4 30
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Expansion and Develop Commerce. Board of Commissioners of the Port of New
Orleans.
Denny, M. (2002). Surveying Little Egypt. Point of Beginning. Vol. 27:8 (April 30, 2002).
Elliott, D.O. (1932). The Improvement of the Lower Mississippi River for Flood Control and
Navigation. Mississippi River Commission, St. Louis.
Fisk, H. N. (1952). Geological Investigation of the Atchafalaya Basin and the Problem of
Mississippi River Diversion, Vol. 1, U.S. Corps of Engineers Waterways Experiment
Station, Vicksburg LA.
Hewson, W. (1870). Principles and Practice of Embanking Lands from RiverFloods, as
applied to”Levees” of the Mississippi, 2nd Ed., D. Van Nostrand, New York.
Interagency Performance Evaluation Task Force (IPET). (2006). Performance Evaluation
Status and Interim Results, Report 2 of a Series, Performance Evaluation of the New
Orleans and Southeast Louisiana Hurricane Protection System. Final Draft (subject to
revision),10 March 2006. 322 p.
Jadwin, E. (1928). “The Plan for Flood Control of the Mississippi River in Its Alluvial
Valley.” Great Inland Water-Way Projects in the United States, R.A. Yound (Ed.),
Philadelphia, American Academy of Political and Social Science, CXXXV (January),
3544.
Junger, S. (1992). “The Pumps of New Orleans.” American Heritage Invention & Technology
, 8(2), 4248.
Kelman, A. (1998). A River and Its City: Critical Episodes in the Environmental History of
New Orleans, Ph.D. dissertation. Brown University, Province RI.
Kesel, R.H. (2003) “Human modification of the sediment regime of the lower Mississippi
River flood plain.” Geomorphology, 56, 325334.
Kolb, C. R. (1976). “Geologic Control of Sand Boils Along Mississippi River Levees.”
Geomorphology and Engineering, D.R. Coates (Ed.), Dowden, Hutchinson & Ross,
Halsted Press, 99113.
Kolb, C.R., and Saucier, R.T. (1982). “Engineering geology of New Orleans.” Geological
Society of America, Reviews in Engineering Geology, 5, 7593.
Kolb, C.R., and Van Lopik, J.R. (1958). Geological Investigations of the Mississippi RiverGulf Outlet Channel, Miscellaneous Paper No. 3259, U.S. Army Engineers
Waterways Experiment Station, Vicksburg LA.
Lemmon, A.E., Magill, J.T., and Wiese, J.R., Eds., Hebert, J.R., Cons. Ed. (2003). Charting
Louisiana: Five hundred years of maps. The Historic New Orleans Collection, New
Orleans.
Mansur, C. I., and Kaufman, R. I. (1956). “Underseepage, Mississippi River Levees, St. Louis
District.” Journal of the Soil Mechanics and Foundations Division, 82(1), 385406.
McPhee, J. A. (1989). The Control of Nature. Farrar, Straus, Giroux, New York.
Meade, R.H., and Parker, R.S. (1985). “Sediment in Rivers of the United States.” National
Water Summary 1984 – Water Quality Issues, U.S. Geological Survey, 4960.
4 31
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Mississippi River Commission (MRC), and Lower Mississippi Valley Division. (1975). The
Flood of ’73, Corps of Engineers, U.S. Army.
Moat, L.S, (1896). Frank Leslie’s Famous Leaders and Battle Scenes of the Civil War, Mrs.
F. Leslie, New York.
Moore, N. R. (1972) Improvement of the Lower Mississippi River and Tributaries 1931-1972,
Department of the Army, Corps of Engineers, Mississippi River Commission,
Vicksburg MS.
Morgan, A. E. (1971). Dams and Other Disasters: a Century of the Army Corps of Engineers
in Civil Works, Porter Sargent Publishers.
Noble, C. C. (1976). “The Mississippi River Flood of 1973.” Geomorphology and
Engineering, D.R. Coates (Ed.), Dowden, Hutchinson & Ross, Halsted Press.
Press, F., and Siever, R. (1997). Understanding Earth, 2nd Ed., W.H. Freeman & Co.
Saucier, R.T. (1963). Recent Geomorphic History of the Pontchartrain Basin. Louisiana,
Coastal Studies Series No 9, Louisiana State University Press, Baton Rouge LA.
Saucier, R.T. (1994). Geomorphology and Quaternary Geologic History of the Lower
Mississippi Valley, U.S. Army Corps of Engineers Waterways Experiment Station,
Vicksburg, MS.
Shallat, T. (1994). Structures in the Stream: Water, Science, and the Rise of the U.S. Army
Corps of Engineers, University of Texas Press.
Shallat, T. (2000). “In the Wake of Hurricane Betsy.” Transforming New Orleans and Its
Environs: Centuries of Change, C.E. Colten (Ed.), University of Pittsburgh Press,
Pittsburgh, 121137.
Snow, R.F. (1992). “Low and Dry.” American Heritage Invention & Technology, 8(2), 45.
Watson, C.C. (1982). An assessment of the lower Mississippi River below Natchez,
Mississippi. Ph.D. dissertation, Department of Civil Engineering, Colorado State
University, Fort Collins, Colorado. 162 p.
Williams, F. E. (1928). “The Geography of the Mississippi Valley.” Great Inland Water-Way
Projects in the United States, Annals of American Academy of Political and Social
Science, CXXXV (January), 714.
U.S. Army Corps of Engineers. (1965). Hurricane Betsy September 8-11, 1965, U.S. Army
Engineer Office, New Orleans LA
U.S. Army Corps of Engineers. (1984). Lake Pontchartrain, Louisiana, and Vicinity
Hurricane Protection Project. Reevaluation Study. New Orleans District, New
Orleans LA.
U.S. Army Corps of Engineers. (1987). Design Memorandum No. 17, General Design,
Jefferson Parish Lakefront Levee, Vol. 1, New Orleans District, New Orleans LA.
Works Projects Administration of Louisiana. (1937). Some Data in Regard to Foundations in
New Orleans and Vicinity, New Orleans LA. 243 p.
4 32
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 4.1: Typical cross section through the sandy bank levees of the Mississippi River,
illustrating how the river’s main channel lies above the surrounding flood plain, which were
poorly drained swamp lands prior to reclamation in the post Civil War era (from Williams,
1928).
Figure 4.2: Same typical cross section, showing the hydraulic sorting of sediments moving
away from the Mississippi River channel. The levee backslope zone lies between the elevated
levees and the poorly drained swamps. In New Orleans, the Carrollton, Uptown, French
Quarter, and Central Business Districts are situated on the natural levee and its backslope,
while the MidCity area was built on a levee flank depression between the Mississippi and
Metairie levees. The Lakeview, Gentilly, and Ninth Ward areas occupy the old cypress
swamps.
4 33
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 4.3: Natural levees exist along most perennial channels subject to periodic overbank
flooding emanating from a prominent low flow channel, as sketched above. Manmade levees
originated by piling up additional earthen fill on top of these natural levees (from Press and
Siever, 1997).
Figure 4.4: Union forces under General Grant cutting the levee near the state line of Louisiana
and Arkansas, 20 miles above Lake Providence (from Moat and Leslie, 1896). In describing
this activity, Moat and Leslie (1896) noted: “The soil is very tough, and will not wash away.
The levees consequently have to be blown up with gunpowder. The soil is then loosened with
spades.” Levees constructed of cohesive clay were found to be the most resilient, but those
constructed of other materials, such as overbank silt, peat, or organic ooze were easily eroded.
4 34
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 4.5: Asymmetric channel cross section typical of the Mississippi River, showing
slumping of the oversteepened banks on the outside of its turns and the relative position of the
river’s thalweg, the line connecting the lowest points along the bed of the river. River
mileage is measured along the thalweg, not along the river centerline, because this line more
accurately describes the actual flow path (from Fisk, 1952).
Figure 4.6: Map showing the lands inundated in Louisiana during the height of the great
Mississippi River Flood of 1927 (from the Historic New Orleans Collection). Concerns over
long term safety from flooding caused many businesses and financial institutions to depart
New Orleans to seemingly safer havens, such as Houston, TX (Barry, 1997).
4 35
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 4.7 – Cross section through a typical Corps of Engineers levee in an alluvial valley
(from Mansur and Kaufman, 1956). Analyses of levee stability depend in large measure on
various assumptions made about seepage conditions beneath and adjacent to such structures.
For instance, the coarse sand and gravel shown here may be 1000x more permeable than the
overlying medium sand.
Figure 4.8: A major problem with manmade levees constructed during the MR&T Project is
that they are necessarily constructed upon highly heterogeneous foundations, as portrayed
here (taken from Kolb, 1976). The sharp contrast between highly organic channel fills
(stippled zones) and natural levee sands and gravelly point bars promotes dangerous
concentrations of seepage and differential settlement.
4 36
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 4.9: Map showing principal elements of the Mississippi River & Tributaries Project
flood control for the lower Mississippi River Delta region (from Chatry, 1961). Note the
much shorter flow channel to the Gulf of Mexico along the Atchafalaya River as opposed to
the Mississippi River. The Mississippi River would have switched to this channel by 1975 if
the Old River Control Structure had not been constructed in 196163.
Figure 4.10 – This cross section illustrates how much of New Orleans lies below mean gulf
level, requiring every drop of rain water to be pumped out. The height of the Mississippi
River levee is +24.5 ft, MGL while the Lake Pontchartrain levee crests at +13.5 ft, MGL
(from Kolb and Saucier, 1982).
4 37
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 4.11: Stage hydrographs on Lakes Borgne and Pontchartrain from the September 1947
hurricane (from USACE DM17, 1987). The 10 foot surge on Lake Borgne was the highest
recorded value up to that time, though shortlived. A 13 foot surge was reported along lake
Pontchartrain during the 1915 Grand Isle Hurricane, but this was before storm surge recorders
were emplaced along the shorelines.
4 38
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 4.12: Portion of the flood inundation map from Hurricane Betsy in 1965, showing the
areas on either side of the Inner Harbor Navigation Channel which were affected by
overtopping, from storm surges on Lakes Borgne and Pontchartrain (from USACE, 1965).
4 39
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 4.13: South Lake Pontchartrain flood protection measures authorized by Congress in
the wake of Hurricane Betsy in 1965. These included heightening of the protective levees
along the IHNC and the Lake Pontchartrain shoreline to the OrleansJefferson Parish
boundary, and around Chalmette in St. Bernard’s Parish. This system was subsequently
enlarged to include the Pontchartrain levee all the way to the Bonne Carré Spillway and along
the principal drainage canals in New Orleans and Jefferson Parishes.
4 40
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 4.14: Plan of the City of New Orleans prepared by Francis Ogden in 1829. Note the
linear drainage canals feeding into Bayou St. John, thence into Lake Pontchartrain (from the
Historic New Orleans Collection).
4 41
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 4.15 Map of Sauvés Crevasse and the portions of New Orleans inundated by the
flooding of 1849, the last significant flood to affect the city emanating from the Mississippi
River. This 1849 map shows the extensive cypress swamps lying between the uptown and
French Quarter areas and Lake Pontchartrain. The Carondelet and New Orleans Canals are
clearly shown, but curiously omits the New Basin Canal (built in the 1830s). The map clearly
shows the projected path of the 17th Street Canal between Orleans and Jefferson Parishes,
suggesting it was being proposed (it appears to have been completed in 185758). The
Labarre Canal in Jefferson Parish (near today’s Bonnabel Canal) was likely never built (taken
from WPA, 1937).
4 42
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 4.16: By 1863 there were a series of eastwest feeder canals serving Bayou St. John
from the west side and a series of north northeasterly trending drainage canals in St. Bernard
Parish (from The Historic New Orleans Collection).
4 43
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 4.17: All 36 miles of drainage canals in the Lakeview and Gentilly areas are shown in
this portion the 1878 Hardee Map (courtesy of The Historic New Orleans Collection). The
canals are, from left: 17th Street, New Basin (infilled), Orleans, Bayou St. John, and London
Avenue, and the Lower Line Protection Levee.
4 44
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 4.18 – Photo taken in 1890 looking north along the “shell road” than ran along the
west side of the New Basin Canal, seen at extreme right. Note the modest height of the
original embankment, no more than 5 feet above the adjacent cypress swamp at left. The
original embankments were heightened after hurricaneinduced overtopping in 1915 and 1947
(image from the University of New Orleans Special Collections, New Orleans Views).
Figure 4.19: Cross section through New Orleans prepared by City Engineer L. W. Brown in
1895 (from the Historic New Orleans Collection). This shows the elevated position of the
Mississippi River and the MetairieGentilly Ridge distributary channel, which lies 3 to 6 feet
above the surrounding area. The green lines denote high and low levels in the river and Lake
Pontchartrain. Elevations are in the old Cairo Datum (21.26 ft above MGL).
4 45
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 4.20: Principal elements of drainage system infrastructure as it existed in 1903
(taken from Campanella, 2002). The 17th Street and London Avenue Canals had already
been in operation for several decades.
Figure 4.21: S&WB engineer A. Baldwin Wood standing next to one of his 14foot
diameter screw pumps in 1929 with several of the board’s secretaries sitting inside the
housing for scale (courtesy of the Sewerage & Water Board of New Orleans).
4 46
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 4.22: Principal elements of the preKatrina drainage system infrastructure as it existed
in 1992 (taken from Campanella, 2002). The aggregate pump capacity could have cleared the
city of flood waters in less than three days if the levees had simply been overtopped without
failing.
Figure 4.23: Evolution of the Corps of Engineers’ standard levee section, 1882 to 1972 (from
Moore, 1972). Earth embankments levees are generally heightened sequentially by
compacting additional soil on the land side of the embankments (each sequence of
heightening shown as different colors). Levees adjacent to drainage canals or perennial
channels are not raised on the river side of the embankment because excess moisture would
prevent meaningful compaction of the fill. Existing homes abutted the landside of the
drainage canal levees in New Orleans by the time the Corps of Engineers began analyzing
them in the 1960s.
4 47
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 4.24 (upper): View looking up the east side of the London Avenue Canal near
Robert E. Le Boulevard crossing showing the encroachment of homes against the slope
of the levee. This situation was common across New Orleans (photo by C. M. Watkins).
Figure 4.24 (lower): Concrete flood wall along the west side of the 17th Street Canal in
Jefferson Parish, where a street runs along the toe of the embankment. This scene is
typical of the concrete Iwalls constructed on steel sheetpiles driven into the crest of the
drainage canal embankments in New Orleans in the 1990s to provide additional flood
freeboard from hurricaneinduced storm surges (photo by J. D. Rogers).
4 48
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 4.25: Assumed Category 3 storm surge curves for the Gulf of Mexico shoreline, Lake
Borgne, and Lake Pontchartrain used by the Army Corps of Engineers for planning and
design purposes prior to Hurricane Katrina in 2005. Note the short duration of extreme
surges, about 12 hours duration above 5 ft MGL for Lake Pontchartrain (taken from USACE
DM17, 1987).
4 49
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 4.26: Schematic layout of levees and flood walls protecting the New Orleans area at
the time Hurricane Katrina struck on August 29, 2005 (from image by the New York Times).
Red arrows denote locations of levee failures.
4 50
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 4.27: Incomplete record of the Lake Pontchartrain tidal stage gage at West End, near
the mouth of the 17th Street Canal during the early stages of Hurricane Katrina (from U.S.
Geological Survey). This record shows the progressive building of tidal stages, days before
the storm made landfall. Significant “ramping” of the storm surge began on August 28th, with
the sharpest increase on the morning of August 29th, when the hurricane made landfall. The
gage failed when the lake level reached 5.3 ft, before the peak surge was recorded.
Figure 4.28: Mobile dragline constructing the MorrisonPicayuneville Levee about 25 miles
south of New Orleans in June 1931 (from Elliot, 1932). Tower draglines could excavate
materials up to a quarter mile away, dragging it back up onto the new levee.
4 51
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 4.29: Aerial oblique view of the Inner Harbor Navigation Canal between 196064,
after the entry to the Mississippi RiverGulf Outlet Channel had been enlarged (upper right),
connecting to the inner harbor area (photo from the Army Corps of Engineers).
4 52
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 4.30: Seepage crevasse splay exposed on the water side of the east levee of the IHNC
breach, likely from backdrainage of the floodinundated area of the Lower Ninth Ward, after
Hurricanes Katrina and Rita. This same section of the IHNC levee failed in 1965 during
Hurricane Betsy. Seepage crevasse splays tend to occur where high permeability materials
daylight, or come into proximity of the existing ground surface. They are easily recognized
by anomalous seepage and the birdfoot pattern of the splays, which are often filled with soils
displaced by seepage and hydraulic piping. (photo from U.S. Army Corps of Engineers).
4 53
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 4.31: Portion of a map of the upper MRGO channel adjacent to Lake Borgne from the
report by Coastal Environments, Inc. (1984). This shows the major soil subdivisions they
identified: soft marsh, firm marsh, and swamp substrate. Much of this material was
unsuitable for using in the adjoining levee embankments.
Figure 4.32: Area where the southwest bank of the MRGO channel levee within two
miles southeast of Bayou Dupree was completely swept away by overtopping from
Lake Borgne (photo by L. F. Harder).
4 54
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Datum
Ellet Dat um of 1850
Delt a Survey Dat um of 1858
Old Mem phis Dat um of 1858
Old Cairo Dat um of 1871
New Mem phis Dat um of 1880
Mean Gulf Level Dat um ( prelim inary) 1882
Mean Gulf Level Dat um of 1899
New Cairo Dat um of 1910
Mean Low Gulf Level Dat um of 1911
Conve r sion t o M e a n Se a Le ve l 1 9 2 9
unknown
0.86
- 8.13
- 21.26
- 6.63
0.318
0
- 20.434
- 0.78
Figure 4.33: Table relating correction factors used when comparing various historic datums in
the New Orleans area (Denny, 2002). Blanket corrections can no longer be made to adjust
elevations to NAVD882004.65, which is the most oft cited datum currently used in New
Orleans. The reason for these disparities is the gross differential settlement between reference
benchmarks, which can be significant (order of magnitude difference).
Figure 4.34: Official neighborhood names recognized by the City of New Orleans (taken from
Campanella, 2002). The Ninth Ward used to extend across the IHNC, but that portion east of
the IHNC has been renamed the “Lower Ninth Ward.”
4 55
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
CHAPTER FIVE: THE LOWER MISSISSIPPI REGION
AND PLAQUEMINES PARISH
5.1 Overview
Plaquemines Parish is the area where the last portion of the Mississippi River flows
out into the Gulf of Mexico (see Figures 2.6 and 5.1). Extending southeast from New
Orleans, Plaquemines Parish straddles both sides of the lower reaches of the Mississippi River
for about 70 miles out to the river’s mouth in the Gulf. This protected strip, with “river”
levees fronting the Mississippi River and a second, parallel set of “storm” levees facing away
from the river forming a protected corridor less than a mile wide, serves to protect a number
of small communities as well as utilities and pipelines. This protected corridor also provides
protected access for workers and supplies servicing the large offshore oil fields out in the Gulf
of Mexico.
It is an area that is sparsely populated, with a population of only about 27,000 people
in the entire parish just prior to Hurricane Katrina’s arrival (see Plaquemines Parish
Government Website: http://www.plaqueminesparish.com). Most of these people live in
small, unincorporated towns and villages along the river. Not only are these communities
subject to potential flooding from the Mississippi River, but they are also vulnerable to
flooding from hurricane surges because the parish extends so far out into the Gulf from the
mainland.
For flood protection from the Mississippi River, large federal project levees were
constructed along both sides of the river with design crest elevations of approximately +25
feet (MSL). For many of the communities lying closely alongside the Mississippi River
levees, “hurricane” or back levees were also constructed behind them to protect them from
hurricane surges coming from the Gulf. These hurricane levees were constructed with lesser
crest heights than the river levees, and typically had crest heights on the order of +17 to +18
feet (MSL). Thus, many of the homes in these areas are sandwiched between two sets of
levees: one along the river and the other behind the towns.
The Independent Levee Investigation Team was not able to devote significant time to
detailed investigations and analyses of the numerous individual levee failures that occurred
along this protected corridor. Accordingly, this chapter will present only a brief overview of
the performance of the flood defenses in this parish during Hurricanes Katrina and Rita.
As described previously in Chapter 2, Plaquemines Parish was the first developed area
to be severely affected by the large onshore storm surge as Hurricane Katrina approached the
southern coast in the early morning of August 29, 2005.
Hurricane Katrina devastated many of the Plaquemines Parish communities.
Hurricane Katrina was reported to have induced storm surges on the order of up to 20 feet in
this region, as shown in Figure 5.2. In addition, large storm waves atop this surge rose to
greater heights. This storm surge, and the waves that accompanied it, overtopped and
damaged many portions of the “storm” levees. Both the United States Army Corps of
51
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Engineers (see Figures 2.6 and 5.1) and the Plaquemines Parish Government website report
numerous breaches of the storm levees and widespread deep flooding and destruction.
Figures 5.3 through 5.12 show examples of the types of damage and flooding that
resulted from the overtopping and breaching of the protective hurricane levees.
Figure 5.3 shows an aerial view of the inundation of the hamlet of Myrtle Grove, on
the west side of the Mississippi River, as it appeared on September 25, 2005, one day after the
second Hurricane (Rita) again inundated this section.
Figure 5.4 shows an aerial photograph of a levee breach of the hurricane (back) levee
on the western side of the Mississippi River near the community of Sunrise. The breach
occurred at a “transition” between an earthen levee section with a sheetpilesupported
concrete Iwall, and a plain structural floodwall section. Failures at transitions between
different adjoining sections were relatively common throughout the affected area during
Hurricane Katrina.
Figure 5.5 shows an aerial photograph of a breach of the hurricane (back) levee at
another “transition” near the Hayes Pump Station. This time the failure occurred at a
sheetpile transition between an earthen embankment and a structural floodwall section, and
sheetpile to earthen embankment connection appears to have been the weak link.
Figure 5.6 shows a pair of large shrimp boats on Highway 23, near the foot of the
Empire High Rise Bridge. As illustrated by this photo, overtopping was quite severe, and
large objects were floated up onto, and sometimes over, the levees.
5.2 Point a la Hache
Point a la Hache is the parish seat for Plaquemines Parish and is located along the east
side of the Mississippi River. Storm surges from the east largely overwhelmed the back
levee, breached it in several places, and inflicted deep flooding and widespread destruction in
this town. Figure 5.7 presents an aerial photograph of one such breach taken on September
25, 2005 (from Plaquemines Parish Government Website). Shown in this photograph is a
temporary road constructed across the interim breach repair to facilitate access and repairs.
Figure 5.8 shows this same levee breach a few weeks later during the installation of a
sheetpile cutoff that was undoubtedly intended to be part of an interim, and perhaps
permanent repair. The team members viewing the installation believed that the sheetpile wall
was a good concept to affect a positive cutoff of seepage through the deeply scoured breach
and loose debris. However, during the installation, team members noted that the contractor
was having difficulty advancing the southern portion of the sheetpiles very far into the ground
using the equipment in use at the time of the team’s visit. It is hoped that the pilings ended up
being driven to their needed depths.
Residences in Pointe a la Hache were commonly inundated to depths of 12 to 18 feet
(see Figure 5.9). Inundation flooding was so great that water flowed across the community
from the east towards the Mississippi River, and even overtopped the Mississippi River levee
52
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
(at least with significant wave splashover) by several feet. Based on debris found on tractor
equipment left on the levee crown along the Mississippi River, overflows or splashover of up
to 4 feet were estimated. For most of the areas visited by our team, relatively little significant
damage was observed on the Mississippi River levees, possibly because the river sides of the
levees viewed by the team were paved with concrete slope protection (see previous Figure
2.17). Damage to the “storm” levees was significant at many locations, however,
Like many New Orleans residences, the small wooden homes in Pointe a la Hache
were commonly founded on cinderblock piers. As a result of the deep flooding and the flow
towards the Mississippi River, homes in Pointe a la Hache were commonly picked up and
floated away from their foundations. Many ended up being deposited on or across the
Mississippi River Levee as a result of storm surges flowing from the overtopped “storm”
levees towards the “river” levees alongside the Mississippi River (see Figures 5.10 through
5.12).
5.3 Erosion Studies
Although overtopping caused numerous breaches in the “storm” levees facing away
from the Mississippi River, less erosion was observed along most of the Federal “river”
levees. This may have been due in part to the fact that the riverside levee embankments
slope faces were paved with concrete slope face protection (as shown previously in Figure
2.17, which clearly shows this riverside slope face protection.) It may also have been due in
part to the fact that the backsides of these river levees, which had no formal slope face
protection, were at least partially protected from the full energy of the storm surge and the
wind driven waves by the obstacles presented by the “hurricane” levees, and by other
obstructions including buildings and trees, etc.
Nonetheless, it is a noteworthy performance on the part of these levee embankments,
and it merits further study. It is hoped that with further testing trends will emerge showing
that soil type and character, as well as placement and compaction conditions, can be used as a
relatively reliable basis for prediction of the level of vulnerability of levee embankment soils
to erosion and scour. Issues associated with erosion are discussed in more detail in Chapter 9.
5.4 Summary
Plaquemines Parish is the most obviously exposed populated and flood protected area
in the region. It juts out into the Gulf of Mexico much like a boxer’s chin, almost daring a
knockout blow.
Because Plaquemines Parish is so obviously exposed, the evacuation of the Parish was
unusually comprehensive prior to Katrina’s arrival. That was a good thing, as most of the
lower reaches of the Parish were catastrophically flooded. Massive damage was done to
homes and businesses in the many small and generally unincorporated townships, and there
was at least one major rupture in an oil transmission line. The best information available to
this investigation team at this time is that approximately 60 lives were lost in Plaquemines
Parish during hurricane Katrina.
53
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
The merits of expending Federal dollars to attempt to defend the full Parish, or even
large portions of it, in the face of ongoing regional subsidence, sea level rise, and increasing
projected hurricane intensity due to rising Gulf water temperatures, warrant further study.
Recent requests for up to $3 billion in Federal funds to repair and upgrade the levees for a
narrow strip of land into which less than 15,000 to 20,000 people are currently expected to
return would represent an expenditure of approximately $150,000 to $200,000 per capita. In
the mean time, large amounts of Federal funds are currently being expended to repair the
damaged levees in this Parish.
5.5 References
Interagency Performance Evaluation Task Force, (2006), “Performance Evaluation, Status
and Interim Results, Report 2 of a Series, Performance Evaluation of the New Orleans
and Southeast Louisiana Hurricane Protection System,” March 10, 2006.
Plaquemines Parish Website, (2006), http://www.plaqueminesparish.com
54
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Source: Modified after USACE
Figure 5.1: Map showing the levee protected areas along the lower reaches of the
Mississippi River (in the Plaquemines Parish Area).
55
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Source: IPET (2006)
Figure 5.2: Aggregated maximum storm surge elevations (maximum among all times).
56
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Source: http://www.plaqueminesparish.com/
Figure 5.3: Aerial photograph of inundated portion of Myrtle Grove along western
side of the Mississippi River. [September 25, 2005]
Sunrise: 29 21.62N 89 33.67W
Source: http://www.plaqueminesparish.com/
Figure 5.4: Aerial photograph of levee breach of storm (back) levee along western side of the
Mississippi River near the community of Sunrise. [September 25, 2005]
57
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Source: http://www.plaqueminesparish.com/
Figure 5.5: Aerial photograph of levee breach of storm (back) levee at leveetowall
transition near Hayes Pump Station. [September 25, 2005]
Source: http://www.plaqueminesparish.com/
Figure 5.6: Aerial view of two large shrimp boats deposited on Highway 23 at the foot of the
Empire High Rise Bridge.
58
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
East Pointe a La Hache: 29 35.68N 89
Source: http://www.plaqueminesparish.com/
Figure 5.7: Aerial photograph of levee breach of storm (back) levee East of Pointe
a la Hache. [September 25, 2005]
Photograph by Les Harder
Figure 5.8: Photograph of Sheetpile Cutoff Being Placed into Levee Breach of Storm
(Back) Levee East of Pointe a la Hache. [October 12, 2005]
59
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph by Les Harder
Figure 5.9: Photograph of flood elevation on trees landward of hurricane levee East of
Pointe a la Hache – illustrating that flood waters remained to large depths
for extended periods. [October 12, 2005]
Photograph by Les Harder
Figure 5.10: Photograph of Pointe a la Hache home deposited on Mississippi River levee
crown after storm surges overtopped the storm levee from the East (left)
towards the River – which is to the right in this photograph.
5 10
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph by Les Harder
Figure 5.11: Photograph of Pointe a la Hache homes deposited on Mississippi River Levee
after storm surges overtopped the levee from the East (left) towards the River
(right). [October 12, 2005]
Photograph by Les Harder
Figure 5.12: Photograph of Pointe a la Hache home site where a wood home was floated off
of its cinderblock piers. [October 12, 2005]
5 11
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
CHAPTER SIX: THE ST. BERNARD AND LOWER NINTH
WARD PROTECTED AREA
6.1 Introduction
As described previously in Chapter 2, St. Bernard Parish and the Lower Ninth Ward
are protected by a single continuous “ring” of levees that, together, constitute one of the three
main protected basins flooded by hurricane Katrina.
Figures 2.11 and 6.1 show the locations of the principal breaches and distressed
sections of the levee and floodwall system protecting this basin. Figure 6.2 shows the
inundation of this basin four days after the hurricane, on September 2, 2005. At the time
shown in this figure, the floodwaters have been partially drained out from the flooded basin,
and they are shown at elevation + 3 feet (MSL) [or +5 feet, NAVD 88.]
Cloud cover obstructed the taking of a good image of the flooding at its peak, but this
basin flooded very rapidly in the first hours of the main storm surge. The levees were
massively breached and catastrophically eroded on the northeastern flank; fronting the MRGO
channel and Lake Borgne. In addition, two large breaches occurred at the west end of this
protected basin, fronting the IHNC. The result was that this basin flooded extremely rapidly,
before the storm surge had subsided, and the resulting surgepushed floodwaters rose to an
elevation of approximately + 12 feet above mean sea level in this basin. As a result, even
homes and businesses located on ground well above sea level were inundated. Of course,
sites on lower ground were inundated to greater depths.
After the hurricane passed, and the storm surge had subsided, a number of “notches”
were deliberately excavated through several of the levees to facilitate drainage of ponded
floodwaters by simple gravity flow (as indicated by the yellow stars in Figure 6.1).
6.2 The Northeast Frontage Levees
As shown in Figures 2.9 through 2.11, the initial storm surge swelled the waters of
“Lake” Borgne (which is actually a bay, as it is connected directly to the Gulf of Mexico.) As
the eye of the hurricane then continued to the north, the counterclockwise swirl of the winds
pushed the elevated waters of Lake Borgne to the west, against the levees along the northeast
frontage of the St. Bernard protected basin. The result was catastrophic erosion of the levees
along much of this frontage, and the throughpassage of the floodwaters.
Figures 6.3 and 6.4 show two sections of the levees along this frontage after this event.
These are aerial views taken from significant elevation, and they each show many hundreds of
feet of levee section that have been catastrophically eroded. In Figure 6.3, the depression in
the foundation soils induced by the settlement of the nowvanished levee, and the erosion
produced by the turbulent flow across the original levee footprint, is the only sign of the
former presence of a levee. In Figure 6.4, a sheetpile curtain had been driven along the
centerline of the levee crest, to raise a section that had settled as an interim measure until the
61
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
final stage of fill placement could reraise this embankment section to the final design grade.
The levee embankment has eroded completely from both sides of these sheetpiles, and the
large diameter pipe in this figure was resting on the crest and slopes of the now vanished
levee and so serves as a visual template to show the size and shape (the outline) of the levee
section that is now gone.
Figure 6.5 shows another view of massive erosion along a long stretch of levees along
this “MRGO frontage” section, this time a bit farther to the south (nearer to the second
navigational lock structure at bayou Dupres.) Here the massive erosion is not as complete,
and portions of the levee embankment remain. In this photo, the eroded detritus can be
clearly seen to be strewn back behind the partially eroded levees, and the sandy (and shell
sand) nature of some of this eroded material is evident.
Figure 6.6 shows a ground level view of the sheetpiles from Figure 6.4. In this photo
it can be clearly seen that the sheetpiles, which had originally been driven to constant grade,
have settled differentially under the pounding of the storm surge and storm driven waves.
This would suggest that the cyclic wave loading may have caused pore pressure increases in
the fine, sandy foundation soils into which the sheetpiles were embedded, and that this (full or
partial) liquefaction reduced the bearing strength and stiffness of these foundation soils and
led to the observed differential sheetpile settlements as the sheetpiles were only lightly self
loaded with regard to vertical bearing and settlements.
LIDAR surveys were performed by the USACE to document the elevation of the levee
crest along the full 11mile long northeast (MRGO) frontage both before and after Katrina.
An example is shown in Figure 6.7, where the magenta line indicates the crest elevation prior
to Katrina, and the darker blue line indicates the crest elevation afterwards. The photo at the
top of this figure is a vertical (plan view) photographic image along the same section. The
two LIDAR surveys serve to show the amount of erosioninduced crest loss along this section,
and this can be correlated with the same locations in the photo at the top. Note the light
material streamed back behind the levees (on the “protected side”) in the corresponding
photo; representing eroded material from the levees strewn back into the inboard side
swamps.
Figure 6.7 includes the large (gated) reinforced concrete navigation control structure at
Bayou Bienvenue. A large barge was deposited on the crest of the levee immediately to the
north of this lock structure, and this can be clearly seen in Figure 6.7. Figure 6.8 shows a
second view, of the massive breach eroded at the contact between the lock structure and the
adjacent levee embankment.
Figure 6.9(a) is an oblique aerial of the smaller Bayou Dupres concrete navigation
structure situated farther to the south along this same MRGO levee frontage, showing a
similar massive eroded breach at the juncture between the northwest end of the concrete
structure and the adjoining earthen levee section. Figure 6.9(b) shows a second view, taken
from the eroded breach and looking to the inboard (protected) side along the north flank of
Bayou Dupres, showing the eroded detritus strewn inland from this breach. In this figure, it
can be clearly seen that large fractions of the eroded material consisted of shell sand fill. The
use of lightweight shell sand fill had been called for at this interface section in order to
62
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
minimize differential settlements between the embankment section and the adjacent concrete
lock structure. By minimizing these differential settlements, the formation of a small
settlementinduced gap between the levee and the lock structure would be prevented. As a
result of using the dangerously erodeable lightweight shell sand fill, however, a massive
eroded breach occurred instead.
The crest heights of the levees along much of this MRGO frontage section were
several feet below design grade at the time of Katrina’s arrival. This levee frontage was being
constructed in stages, to allow time for settlement of the evolving levees and for dissipation of
pore pressures (which results in progressive strength and stiffness gain in both the levee fill
and in the underlying foundation soils, so that the softer foundation soils can safely support
the increasing levee section height and weight of the next stage.) The USACE had reportedly
long requested appropriation of the funds necessary to place the final stage of fill and bring
this critical 11mile long section up to full design grade. That funding did not arrive in time.
The levees along this frontage were unusually vulnerable to erosion as they were
“sand core” levees, constructed largely using material available from the adjacent MRGO
channel excavation. Given the nature of the local soils at this location, much of that
excavated material consisted of sands and lightweight shell sands. These materials have a low
intrinsic resistance to erosion (see Chapters 9 and 10), and this led to a hazardous condition.
It is possible that the final fill stage, if it had arrived in time, might have provided a covering
veneer of compacted clay fill (with a higher resistance to erosion), but such a covering was
not in place. In addition, given the ferocity of the surge and storm waves that struck long
sections along this alignment; it is not clear that a relatively thin veneer of compacted clay
would have been sufficient to help very much.
As shown in the map of Figure 6.1, this levee frontage is one of only two locations
where the levees protecting the three main protected basins of New Orleans are exposed
directly to storm waves crossing a large body of Gulf waters (Lake Borgne) without the
protection of significant swamp grounds on their outboard sides. The swamp grounds (and
cypress trees) serve to damp the energy of the storm waves, reducing their height and
velocity, and thus their erosive potential. It was unfortunate that this section that was so
exposed to severe (unprotected) storm waves was also not yet up to full design grade, and that
large portions were comprised of highly erodeable sand and lightweight shell sand fill.
It should be noted that the only other section of levee protecting one of the three main
basins of New Orleans that was also exposed to open water storm waves (without significant
outboard side swamp and cypress protection) is the “sister” section to the north; at the
southeast corner of the New Orleans East protected basin (facing south, fronting Lake
Borgne.) As discussed in Chapter 7, that “sister” section was also constructed using dredge
spoils from the excavation of an adjacent shipping channel (the GIWW channel in that case),
and was also comprised largely of highly erodeable sands and shell sands. That section, too,
eroded catastrophically and represented the largest source of the floodwaters that
catastrophically flooded the New Orleans East protected basin.
The exact nature of the erosion and breaching that occurred along this frontage section
has not yet been fully agreed upon by the various investigation teams. It is the view of our
63
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
investigation that sections of this levee frontage appear to have eroded and begun to be
breached prior to the storm surge reaching its full height (of approximately +16 to +19 feet,
MSL) by as early as about 5:30 to 6:00 a.m.
Figure 6.10 shows the calculated hydrograph developed by IPET at this location,
showing storm surge rise vs. time at this location as estimated by IPET (IPET, Second Interim
Report, April 2006.) “Storm surge” is the mean water level between storm waves and
troughs, so the additional height of waves, plus “runup” as waves arrive at the levees must be
added to determine when and to what extent the waters overtopped the levees. This is further
complicated by the significant variations in crest elevation along this not yet completed levee
frontage. The analytical prediction of Figure 6.10 matches well with the similar numerical
hydrodynamic modeling performed by Team Louisiana (Kemp and Mashriqui, 2006), and
both models are fairly well calibrated against regional observations of water elevations at
numerous locations. The two investigation teams (IPET and Team Louisiana) differ
significantly, however, in their calculated wave heights and frequencies along this MRGO
frontage. IPET have calculated longer period storm waves typical of more “open ocean”
conditions, and Team Louisiana have calculated shorter period waves constrained by lack of
depth within the Lake Borgne embayment.
Figure 6.11 shows a schematic illustration of two different sets of erosion mechanisms
for the levees along this frontage. Figure 6.11(a) shows simple “sheet flow” overtopping.
This is a common mode of concern for many river levees, and also for many earth dams. In
this mode, as the water flows over the top and then flows like a sheet down the rearside slope
of the levee embankment, the velocity of flow down the rear slope face accelerates and the
shear stresses (erosive forces) induced by the flow increase with this increased velocity.
Accordingly, erosion is initially most pronounced low on the back slope face (where the flow
velocities become highest), and the embankment is eroded from the back side until the crest is
breached (whereupon rapid flow through the crest rapidly enlarges the original breach.) This
is the mechanism that is the customary principal design focus for the flood control levees in
this region; excepting the large rivers such as the Mississippi River where scour produced by
longitudinal flow of the river current itself is also a major concern.
Figure 6.11(b) illustrates two additional potential sets of erosion modes likely to have
been active along sections of the MRGO frontage levees. One is the attacking of the outboard
side (water side) face of the levee by storm waves. These high energy waves can scallop and
erode the outboard face. They can also rush up the face toward the crest, and can erode
“notches” in the crest from the front side. Subsequent waves can then pass through these
notches, especially as the storm surge continues to rise, and the flow can widen the notches
and also erode the back face levee slope (as discussed above as “sheet flow overtopping
erosion”.) This exploitation and widening of crest notches is called crenellation, after the
crenellation (notched shape) that often tops castle walls.
Figure 6.11(b) also illustrates seepage flow passing through the embankment section,
and then eroding soil as it exits through the lower portion of the back side slope face. This
“through flow” can cause significant erosion if the embankment soils are pervious, as was the
case along significant portions of the MRGO frontage levees. As this type of erosion occurs
primarily in the lower back slope face region that is also most prone to erosion by sheetflow
64
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
overtopping, it can be difficult to separate field evidence of these two types of erosion as to
cause.
The highly erodeable (and pervious) sands and shell sands that comprised significant
sections of the levees along this frontage were vulnerable to all three types of erosion, and
would have been expected to have been damaged by waves and by through flow from the
rising storm surge well before the storm surge actually overtopped some sections. Evidence
of front face scalloping erosion, and “notching” at the crest and front crest lip of levees along
this MRGO frontage section are presented in Figures 6.12 and 6.13.
Methods and procedures for calculation of rates of likely erosion due to the various
erosive mechanisms likely to have been operating along the critical MRGO levee frontage are
not wellestablished, and there is little agreement within the profession as to how the
erodeability of the various materials present (fill types, and fill placement and compaction
states.) A number of members of the ILIT team made their own estimates of likely rates of
erosion, based on their perceptions of the likely fractional content of various fill types, and the
types of erodeability data presented and discussed in Chapters 9 and 10, and in Appendix I.
These estimates also required judgmental assessment of through flow potential, wave runup
magnitudes and velocities, numbers of wave cycles at different times (and thus different storm
surge stage levels), etc.
The resulting estimates varied considerably, but all agreed that there was a high
likelihood that initial breaching would have initiated well before the storm surge approached
within several feet of the low points along the crests along this critical levee frontage. This
appears to correlate well with the observation that massive amounts of storm surge flows
filled and then pushed across the open swamplands behind the MRGO frontage levees, and
then crossed over the secondary (Forty Arpent) levee and filled the populous zones to the
south to elevations as high as +12 feet above mean sea level.
This is further supported by the observed behavior of the “sister” levee frontage
section at the southeast edge of the New Orleans East protected basin. This section, which
was also comprised in part of highly erodeable fill materials dredged from the adjacent
shipping channel excavation (in that case the GIWW channel), and which also fronted Lake
Borgne directly, without significant outboard side swamps or cypress to dam and suppress
wave energies, was clearly breached and admitted large volumes of floodwaters well before
the storm surge approached the levee crests. Timing along this “sister” section, and crest
heights and storm surge heights, are better documented (see Chapter 7, Section 7.3.2) than
along the MRGO frontage section, as the resultant New Orleans East flooding was
definitively noted and captured on videotape by workers at the nearby Entergy power plant.
As shown in Figure 6.1, it was intended that the levees along this outer frontage would
bear the brunt of the storm surge. Any overtopping flow, or even flow through localized
breaches, would then have available a wide swath of undeveloped swamp land into which it
could flow and pond. At the back side of this swampland a lower secondary levee (the Forty
Arpent Levee) was then situated to protect the populous areas to the south. Unfortunately, the
unexpectedly rapid and catastrophic erosion of this outer frontage levee allowed the storm
65
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
surge to flow virtually unimpeded across the open swampland before the storm surge had
begun to subside significantly.
The Forty Arpent levee was only a “secondary” levee, with crest heights on the order
of Elev. + 7.5 to + 10 feet (MSL), and it was not intended to have to face the full brunt of a
largely undiminished rising storm surge. As a result, the storm surge passed easily over this
secondary levee, and pushed rapidly into the populated areas of St. Bernard Parish, as
described previously in Chapter 2. As is arrived rapidly, and prior to significant abatement of
the storm surge, the floodwaters ponded to an unexpectedly high elevation of approximately
+12 feet above mean sea level. Homes and businesses on “high ground” (at elevations several
feet and more above sea level) were thus unexpectedly flooded, and the depth of flooding in
lowerlying areas was especially severe. The massive inrushing floodwaters also had large
lateral force, and pushed homes aside from their foundations (as shown previously in Figure
2.19), tossed cars like toys (see Figure 6.15), deposited large fishing boats in residential
neighborhoods (Figure 6.16), and left large branches of trees on the roofs of numerous homes
(e.g.: Figure 6.17).
Interestingly, the smaller (secondary) Forty Arpent levee was severely overtopped
along much of its length, but it suffered relatively little erosional damage as a result. This
appears to be because it was constructed of significantly better materials than the outer
(MRGO frontage) levees; the Forty Arpent levee appears to have been constructed primarily
of clay, with good intrinsic resistance to erosion. Figure 6.14 shows a section of the Forty
Arpent levee that was apparently significantly overtopped, but which suffered only slight
“cosmetic” erosional damage as a result.
The use of highly erodeable sand and shell sand fill was unfortunate along the exposed
MRGO frontage levee section, and the consequences were severe. Damage to the populated
areas of St. Bernard Parish was catastrophic, and the floodwaters from this populous area next
began to make their way westwards towards what was now the already doomed Lower Ninth
Ward.
6.3 The Two large Breaches on the East Bank of the IHNC at the Lower Ninth Ward
As the storm surge from Lake Borgne pushed westward along the eastwest trending
channel of the GIWW/MRGO that separates the St. Bernard and New Orleans East protected
basins, it raised the water levels in the IHNC and produced two massive breaches on the east
bank of the IHNC (at the western edge of the Lower Ninth Ward). These two breaches
occurred at approximately 7:30 to 7:45 a.m., at an IHNC water level of approximately Elev. +
14 to +14.5 feet (MSL), as shown in Figure 6.18 (which shows a hydrograph of measured
water levels vs. time in the IHNC channel.)
6.3.1 The IHNC East Bank (South) Breach at the Lower Ninth Ward
The larger of these two breaches was the south breach, and this is shown in Figure
6.19 (which is a repeat of Figure 2.13). This was a very long breach, nearly 900 feet in
length, and the inrushing waters entered the adjacent community with great force. As shown
66
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
in Figure 6.17, homes for several blocks were ripped from their foundations and scattered,
usually in splinters, eastward across the inboard neighborhood.
Figure 6.19 also shows the sheetpile curtain that had supported the floodwall at the
crest of the earthen levee at this section. It is interesting to note that the sheetpiles (which
were coldrolled steel sections) remained interlocked throughout the cataclysmic failure and
the ensuing hydrodynamic loading of the massive inrushing floodwaters. The concrete
floodwall is largely absent from the tops of these sheetpiles, as the sheetpiles have been
stretched out (like an accordion), flattening their bent flanges in order to accommodate the
extension imposed on them by the inrushing flow.
Figure 6.19 also shows a large steel barge that passed inward through this section, and
came to rest near the southern end of the breach. This raised the question as to which came
first; the barge or the breach?
Figure 6.20 shows the large barge, in its final resting position (prior to being cut apart
with torches to remove it) atop a small yellow bus. This was not the initial resting location of
this barge immediately after hurricane Katrina, however. Initially, after Katrina, the barge
had come to rest a bit farther to the east. It was then refloated several weeks later when the
temporary breach repair failed during the second hurricane surge produced by hurricane Rita
on September 24, 2005 (see Chapter 11), and came to rest at its current position at that time.
The small yellow school bus also arrived between hurricanes Katrina and Rita, having been
appropriated and used for interim transport and then abandoned in its location as shown.
There is a single large dent low on the side of the barge just around the left side of the
bow (not quite visible in Figure 6.20), and a pronounced scrape on the bottom of the barge at
that same location. Most of the concrete floodwall was failed in extension and flexure, with
its reinforcing steel (rebar) fairly extended. There was one single section of wall which
clearly evinced a major impact, however, and that was at the extreme southern end of the
breach. Figure 6.21 shows a closeup view of the floodwall at this location. The rebar is
compressed and bent, and the concrete crushed at this location. It was the consensus view of
our investigation team that the barge had scraped along the wall and then impacted the end of
the wall at this location.
As this was the extreme southern end of the very long breach; this impact was not the
cause of the breach and failure. Instead, the barge was apparently traveling southwards along
the IHNC (driven by the prevailing storm winds at that time) and was drawn into the breach
by the inflowing waters. The barge did not enter cleanly into the breach, but struck at the
south end before passing in.
That does not mean that the barge might not have struck the floodwall twice (or more
times) before finally impacting the southern end of the breach, but our investigation’s view is
that there are other modes of failure that would have been expected to fail this section without
any need for help from the barge, so that the likelihood is that the barge slipped its moorings
and was eventually drawn in through a breach that was already well developed.
67
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 6.22 shows the trench that was eroded by water that passed over the top of the
concrete floodwall at the south end of the large breach. (The barge can be seen at the right in
this photo.) Overtopping and scour occurred at both ends of this breach feature, and the
resulting scoured trenches reached depths of up to 5.5 feet in sections that did not
subsequently fail.
It is, of course, not possible to determine whether deeper
scouring/trenching might have occurred at the actual breach inception location, as the
embankment and foundation soils at the center of the breach were deeply scoured out by the
massive flows in through the breach. One of the potential failure modes evaluated by our
(ILIT) studies was the possibility that this scour had sufficiently laterally unbraced the
concrete floodwall (and its supporting sheetpile curtain) that the lateral force of the elevated
canal water was able to displace it laterals and foment a resulting breach.
Figure 6.23 shows our ILIT reinterpretation of the original boring data along this
section of the east bank of the IHNC, with the locations of the two large breaches indicated.
The boring data was far too sparse along this section for the importance of the design (the
inboard population and properties being protected) and for the complexity of the local
geology. In addition, widely spaced borings along the approximate levee centerline do not
provide an adequate basis for development of appropriate crosssections for analysis and
design. An effort was made to perform pairs of borings (one roughly at the crest and another
at the inboard toe) at selected locations so that crosssections could at least be attempted, but
this was still an inadequately sparse investigation. The foundation investigation for the design
of these levees and floodwalls was inadequate for a project of this scope and importance, and
the minor savings in drilling, sampling and testing are now dwarfed by the massive costs of
the failures that resulted; both property damages and loss of life.
Figure 6.24 shows the crosssection used for our analyses of this south breach. The
two preKatrina (“initial design”) borings, Borings B4 and B4T, were supplemented by
three additional CPT probes performed by the IPET investigation (IHBR6.05C, 5.05C and
16.05C), and two additional borings and a CPTU probe that were performed by our ILIT
investigation (Borings IHNCSBOR1 and CON1, and CPTU IHNCSCPT1). The cross
section of Figure 6.24 shows the tragic failure to extend the sheetpile curtain to sufficient
depth as to cut off underseepage flow through the laterally pervious “marsh” deposits at this
site.
The upper embankment fill is a moderately compacted imported clay, which is
underlain by an older “fat clay” (CH) fill apparently comprised of locally available lacustrine
clays. The upper foundation soils are then dominated by thick deposits of high plasticity
clays (CH), punctuated by two layers of marsh deposits, and there is a relatively thin but
continuous stratum of low plasticity silt (ML) underlying the lower marsh unit.
Subsequent to the completion of the levee embankment and floodwall, additional
sandy fill was placed on the outboard (water) side of the levee to raise the ground surface
slightly above mean canal water level. Some buildings and facilities had been constructed on
this made ground, but these had been removed prior to hurricane Katrina.
Figure 6.25 shows plots of data regarding strength properties vs. depth for the soils
from the silt layer down (from Elevations 19 to 50 feet, MSL) beneath (a) the levee crest,
68
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
and (b) at the inboard toe of the levee (under far lesser embankment overburden). The
detailed procedures and relationships used to process the CPTU data, and then to overlay the
additional UUTX data to develop these plots, are presented and discussed in detail in Chapter
8, and this will not be repeated here. The lower unit of lacustrine clay clearly shows two
overconsolidation “crusts” as a result of surface desiccation during early “stands” in the
accretion of these deposits, and they are more normally consolidated at greater depth. These
clays, in the end, do not appear to have participated in the failure that occurred.
Similarly, the relatively thin silt stratum (ML) also shows evidence of
overconsolidation, and this gives it sufficient strength that it too is uninvolved the failure.
Figure 6.26 shows similar plots regarding strength properties of the far more critical
upper foundation soil strata between elevations of approximately +0 to 20 feet (MSL). These
deposits, consisting of interlayered marsh and clay units, are the critical soils at this site.
As described in detail in Chapter 8, a number of different approaches were taken to the
processing of the available field and laboratory test data in order to evaluate and characterize
these soils. Based on the CPTU measurements within the marsh deposits (both at this site,
and at the 17th Street canal breach site) values of Bq were developed, and then based on the
relationships of Lunne et al. (1994) and Karlsrud et al. (1996), a value of Nkt = 15 was
selected for transposing the CPTU tip resistance values to the estimates of undrained shear
strength that are plotted in Figure 6.26. The resulting values were then converted to values of
Su/P as shown in the far right figure of Figure 6.26, and these appear to infer three
desiccationinduced overconsolidation profiles corresponding to surface exposure at three
times during the evolution of these deposits. The relationship of Mayne and Mitchell (1988)
was then used, again as described in Chapter 8, to crosscheck the resulting relationship
between Su/P vs. OCR as a function of Plasticity Index (PI, %) for these deposits using the
available UUTX laboratory test data. These were found to be consistent. Finally, the limited
available in situ vane shear test data, and the UUTX laboratory test data, was coplotted with
the CPTUbased strengths, and these too were judged to be consistent (with allowances for
sample disturbance and vane insertion disturbance in these soils of variable fibrous organic
content).
Similar processing resulted in selection of a value of Nkt = 15 for processing of the
CPTU data for the silty clay (CH/CL) stratum lying between the two “marsh” deposits. This
differs from the value of Nkt = 12 that was used to process the CPTU data for the deeper layer
of gray lacustrine clay of high plasticity, and it reflects the lower plasticity of this upper clay
unit. Once again, the limited available in situ vane shear test data and UUTX laboratory test
data were then coplotted with the strengths as interpreted by the CPTU, and were found to be
consistent (as shown).
Figure 6.27 shows the geometry and principal input parameters used to model and
analyze this section using the finite element analysis program PLAXIS (2004). The “soft
soil” constitutive model within PLAXIS was used to model all of the uppermost soil strata, so
that both undrained and partially drained conditions could be studied within an effective stress
framework. Shear strengths from Figures 6.25 and 6.26 were reduced by 15% in the marsh
strata, and by 20% in the clay strata, to account for differences between the field (in situ) test
69
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
conditions and the laboratory test conditions, and the direct simple shear (DSS) conditions
expected to dominate the critical field performance behavior in these analyses.
Initial analyses were performed to model the incremental construction of the levee
embankments in order to establish the initial stress conditions for the subsequent analyses of
the overall section performance and stability during hurricane Katrina’s storm surge loading.
Figure 6.28 shows the deformed mesh at the end of staged construction and consolidation
under the levee embankment loads. Overconsolidation stress profiles beneath the crest, and
beneath the inboard levee toe, well matched those from the available field data, and the
consolidation properties were iterated slightly until the final (postconsolidation) settled
profile matched well with the observed field configuration.
Analyses were then performed in which the water level within the canal was
progressively raised. Transmission of pore pressures beneath the wall (and beneath the
sheetpiles) was very rapid, and nearly “steady state” pore pressure conditions developed very
rapidly beneath the inboard side of the levee after each increase in water levels as the lateral
transmissivity of the marsh deposits was high, and the system was initially well saturated.
The rate of water level rise (and subsequent decline) in the canal was based on the hydrograph
of Figure 6.18.
Figure 6.30 shows conditions calculated just as the canal water level reached the top of
the concrete floodwall. Plotted in this figure (as color contours) are levels of relative shear
strain (shear strain developed, divided by shear strain to failure) within the levee embankment
and foundation soils. As shown clearly in this figure, two distinct failure mechanisms are
beginning to develop. The lower one is a shear surface concentrated at the interface between
the base of the upper gray clay (CH/CL) layer and the underlying layer of marsh deposits, and
the upper failure surface attempting to develop is concentrated at the interface between the top
of the upper marsh stratum and the lower levee embankment fill section. Both of these
mechanisms represent the results of underseepageinduced increases in pore pressures being
“trapped” at the bases of less pervious overlying strata. These pore pressure increases are
decreasing the strength and stiffness of the soils at these two critical interfaces.
At the water stage shown in Figure 6.30, a gap has begun to form at the outboard side
of the floodwall and its supporting sheetpile curtain. When effective tensile stress was
calculated between the floodwall/sheetpile wall and the adjacent soils, the analysis was
temporarily stopped, the tension was eliminated by changing the mesh details to insert a small
gap (and to insert hydrostatic water pressures within the gap), and the analysis was resumed.
This was done iteratively, as water levels continued to rise, so that the progressive
development of a waterfilled gap between the floodwall/sheetpile curtain and the outboard
section of the levee embankment could be modeled. At this section, within reasonable
parameter variations modeled, gap formation generally initiated at canal water levels on the
order of Elev. +11.5 to +13 feet (MSL), and the gap then tended to progress fairly rapidly to
the base of the sheetpiles (within the next 1 to 2 feet of water level rise in the canal).
Figure 6.31 shows calculated conditions for a canal water level at Elev. +14 feet
(MSL). At this stage, water is now overtopping the floodwall, the gap at the outboard side of
the sheetpile wall is developed to full depth, and stability failure is occurring on the
6 10
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
uppermost of the two potential failure surfaces. This upper failure is serving to “protect”
against further development of the lower failure surface (which can also be seen in this
figure.) If the upper failure surface is strengthened a bit, to prevent the upper failure, then the
lower failure becomes critical.
Figure 6.32 shows the postulated path to failure based on the finite element (PLAXIS)
analyses performed. In this figure, the Factor of Safety at any given surge height was
assessed by stopping the analysis at each stage of water level rise, and evaluating Factor of
Safety by means of progressive c – Ø reduction. Two sets of conditions were analyzed;
conditions in which a “gap” was allowed to form on the outboard side of the
sheetpile/floodwall (and the gap was allowed to fill with water as it opened), and a second set
of analyses without allowing the opening of this gap. The light blue diamonds in Figure 6.32
represent conditions without gapping, and the yellow circles represent conditions with
progressive opening of a waterfilled gap.
As shown in this figure, the gap begins to open as the storm surge rises near to the top
of the floodwall (at a surge elevation of about +11 to +12 feet, MSL), and the increasing
lateral push of the rising surge waters finally destabilizes the system at a surge elevation of
approximately +12 to +13 feet, MSL. This appears to agree closely with the observed field
timing and surge levels at failure.
These analyses also include the “excavation” of a trench at the levee crest at the rear
side of the floodwall representing the results of overtopping erosion at the north and south
ends of the breach. The depth of this eroded trench was taken as rapidly increasing from none
to 5 feet in depth as overtopping began to pass over the top of the floodwall. Additional
analyses were performed for eroded trench depths of up to 7.5 feet, but this did not
significantly affect the overall results; simple erosion of a scoured trench behind the
floodwall, even as deep as 7.5 feet, was not sufficient as to cause the observed failure and
breaching of this levee/floodwall section. The scoured trench behind the floodwall did
contribute a bit to the enhancement of lateral displacement (and resultant waterfilled
gapping) on the outboard side, but it does not appear to have been the principal factor at this
failure and breach site.
Additional analyses were performed to further evaluate both the seepage flow vs. time,
and the overall stability of this levee and floodwall section. Seepage analyses, as well as
conventional Limit Equilibrium analyses (by several methods, but these agreed closely and
results presented herein are for Spencer’s Method) were performed using the program
package GEOSLOPE/W.
Figures 6.33 and 6.34 show the crosssections and meshes used for conventional limit
equilibrium and coupled seepage analyses of this same breach section. As shown in Figure
6.35, the rapid lateral flow through the main marsh stratum distorts the flownet, carrying
pressures and equipotential contours along as it passes beneath the embankment. Figure 6.36
shows a closeup view of calculated pore pressure contours for a storm surge elevation of +14
feet (MSL). Over a considerable area at and inboard of the levee toe the net pore pressure
uplift forces are slightly greater then the weight of the relatively light soils present,
representing conditions prone to potential “uplift” or “blowout” at this critical location.
6 11
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 6.37 shows a closeup view of hydraulic gradients at this same canal surge
stage. As expected, the exit gradients calculated at the toe are slightly unstable with regard to
initiation of seepage erosion and piping for the relatively lightweight soils present.
A key question in these analyses is the rate at which rises in outboard side canal water
levels manifest themselves in the form of increased pore pressures beneath the inboard side of
the levee embankment. That, in turn, is largely a function of the lateral permeability modeled
within the marsh strata, and assumptions regarding degree of initial saturation.
It was our investigation team’s observation that lateral permeability was very high
within at least some of the substrata of these variable marsh deposits, both at the two east
bank IHNC breach sites at the edge of the Lower Ninth Ward, as well as at sites along the
drainage canals at the north end of the main (downtown) New Orleans protected basin.
Hydraulic response at nearby boreholes was very rapid, and evidence of the occurrence of
high water pressures and underseepage was noted at several locations. Investigators from the
IPET team were surprised by difficulties in dewatering a very shallow excavation to recover
large block samples of peaty “marsh” deposits at the 17th Street canal breach site for
subsequent centrifuge testing. In addition, persistent reports of underseepage and ponding of
waters along this IHNC frontage at the west edge of the Lower Ninth Ward, and contractor’s
significant problems with dewatering of excavations along this same frontage, all bespoke of
high lateral permeability within these strata.
The values of lateral permeability used in these analyses were based on experience
with similar geologic units from other regions, our own field observations, and the
accumulated reports indicating high lateral permeability. A bestestimated coefficient of
lateral permeability of kh ᄃ102 cm/sec was modeled for the most open of the marsh substrata,
and parametric sensitivity analyses were performed for values of kh that were five times
higher, and values that were an order of magnitude (factor of 10) lower.
Figures 6.38 and 6.39 show results of these sensitivity analyses. Transient flow
analyses were performed in which canal water levels were raised progressively, beginning
with assumed fully equilibrated (“steady state”) conditions with a canal water elevation of
about +5 feet (MSL) at ~11:00 p.m. on the night of August 28th (after many hours of relatively
slow surge rise to that level), then rising progressively to elevation +9 feet (MSL) by about
3:30 a.m. on the morning of August 29th, and then rising a bit more rapidly to elevation +14.4
feet (MSL) by about 8:30 a.m. (It should be noted that the failure and breach occurred at
about 7:45 a.m., but that these transient flow analyses were carried forward to at least 9:00
a.m. to more fully examine progressive flow and pore pressure development.)
Figure 6.38 shows calculated pore pressures vs. time at location 1, at the top of the
lower marsh stratum, directly below arrow “D” near the inboard toe of Figures 6.35 through
6.37. The horizontal light blue line at the top of this figure represents the “steady state”
conditions that would eventually develop for a canal water level rise to Elevation +14.4 feet
(MSL) if infinite time were allowed for full equilibration and development of steady state
flow. The lower diamonds represent calculated transient pore pressures at Location 1 for the
bestestimated lateral permeability of the marsh deposits, and for the upper and lower bound
permeabilities. As shown in this figure, the variation in permeability does not exert a major
6 12
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
influence on the pore pressures, given the relatively slow rate of canal water level rise, and
pore pressures within the main marsh deposit at the base of the inboard levee toe are on the
order of about 85% to 92% of full “steady state” pressures at the apparent time of failure (at
about 7:45 a.m.)
Figure 6.39 shows similar transient flow analyses to calculate pore pressure
development at various depths beneath the location of arrow “D” in Figures 6.35 through
6.37, using the bestestimated permeabilities. Again, pore pressure development is fairly
rapid, and lags only moderately behind outboard side canal water level rise.
Figure 6.40 shows the calculated gradients at the top of the lower marsh stratum (the
blue line) at about 7:45 a.m., based on bestestimated permeabilites, and the exit gradients at
this time as well (the red line.) The toe exit gradients are marginally unstable, given the
lightweight materials present, and represent conditions likely to give rise to the inception of
piping erosion.
Figure 6.41 shows the progressive development of pore pressures at the top of the
lower marsh stratum vs. time. As shown, there is a considerable area over which the
hydraulic uplift forces progressively grow to become somewhat larger than the total
overburden stresses; representing a condition that could lead to uplift and “blowout” at this
location.
Finally, Figures 6.42 and 6.43 shows analyses of limit equilibrium (Spencer’s Method)
for failure surfaces passing (a) along the interface at the top of the upper marsh stratum, and
(b) along the interface at the top of the (lower) main marsh stratum, for a canal water
elevation of +14 feet (MSL). Both sections are marginally unstable at this condition with
regard to lateral translation of the inboard portion of the levee embankment, pushed sideways
by the outboard side canal water pressures (including a waterfilled gap at the outboard side of
the sheetpiles), and in both cases the foundation soil strengths have been critically reduced by
underseepageinduced pore pressure increases.
Figures 6.42 and 6.43 likely overestimate the overall lateral translational stability at
this stage of canal water level rise, as it is likely that piping erosion would have at least
initiated at the inboard toe region by this stage, and the calculated hydraulic uplift pressures in
the inboard toe region are high enough that “buckling” of the passive toe block helping to
restrain the lateral translations of Figures 6.42 and 6.43 might further reduce the overall
stability.
As shown by these analyses, as the canal water level rises above about + 13 to +14
feet (MSL) this section becomes analytically unstable by a number of potential mechanisms,
all of them associated with underseepage flow passing beneath the sheetpile curtain. These
potential mechanisms are:
1. Seepage erosion and piping due to excessive exit gradients at the inboard toe.
2. Hydraulic uplift or “blowout” at the inboard toe.
6 13
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
3. Translational stability failure, as a result of reduction in strength of the foundation
soils at the inboard side due to underseepageinduced pore pressure increases.
Based on the length of the breach feature (approximately 900 feet), it is most likely
that the mechanism that won the race to failure at this site was translational instability due to
underseepageinduced pore pressure increases, and resulting strength reduction within the
inboard side foundation soils. It is certainly possible, however, that all three mechanisms
contributed at least in part. Figure 6.44 shows the postulated most likely path to failure, based
on both the finite element and the coupled transient flow/limit equilibrium analyses. The
postulated failure path proceeds up the “ungapped” limit equilibrium path at the right of
Figure 6.44 until a gap on the outboard side of the sheetpiles begins to open and fill with
water at a canal water elevation of about +12 to +13 feet (MSL). The failure mechanism then
transitions to the “waterfilled gap” limit equilibrium case (the leftmost line in Figure 6.44),
and as the canal water level continues to rise the overall section becomes unstable (in
underseepageinduced lateral translational foundation instability) at a canal water level of
approximately +14 feet (MSL).
This contradicts the initial conclusions of the Draft Final Report by the IPET
investigation (IPET; June 1, 2006), and also the initial hypotheses of the ASCE and NSF
sponsored field investigation teams; all of which favored the hypothesis that the failure and
breach at this site had resulted from overtopping flow over the floodwall which eroded a
trench along the back side of the wall (as shown previously in Figure 6.22), resulting in
laterally unbracing the wall so that it was then pushed over by the surge water pressures on its
outboard side.
Our investigation’s view is that, while overtopping and trenching were in fact
occurring, it was underseepageinduced instability that actually developed the more critical
mechanism that led to failure at this site.
The depth of overtoppinginduced trench erosion at the north and south shoulders of
the breach never reached depths greater than 4.5 to 5 feet. It might be inferred that a low spot
along the crest of the floodwall occurred at the breach location, and that somewhat deeper
erosional trenching resulted, but our finite element analyses show that even excavation of a
trench as deep as 7 to 8 feet by overtopping erosion does not sufficiently unbrace the wall as
to foment a lateral wall failure at surge heights overtopping the wall by as much as 1.5 feet.
Instead, the contribution of overtopping and erosion of a trench at the inboard toe of the
floodwall was more likely to have, at best, slightly accelerated the timing of this failure by
adding to the propensity of the floodwall to deflect laterally slightly and thus develop a “gap”
into which water could flow and then apply additional lateral pressure against the sheetpile
curtain to promote the lateral translational stability of the inboard side of the levee
embankment.
Our finite element analyses, performed with eroded “trenches” of various depths (from
none, to as much as 8 feet) suggest that the trench erosion likely helped to exacerbate the
initiation of “gapping” at the outboard side of the floodwall at a slightly lower canal water
elevation than would have occurred without this erosion, but that it was not the critical
contributor to this failure.
6 14
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
This is an important issue with regard to repair and reconstruction. The USACE has
expended considerable effort and resources to replace “Iwalls” with “Twalls”, and to install
concrete splash pads behind additional Iwall sections, in order to prevent failures due to the
mechanism of overtopping, erosion of a trench at the rear side of the Iwalls, and failure due
to the resulting unbracing of the wall sections. Although also useful, this will not also deal
effectively with the underseepage issues that appear to have been the actual cause of failure at
this site; and there appear to be unreasonably short sheetpile curtains (insufficient as to
effectively cut off underseepage flows) at other locations throughout the New Orleans
regional flood defense system. This is a potentially pervasive problem throughout the system,
and it should be evaluated systemwide, and remedied as necessary.
The IPET Final Draft Report notes that possible modes of failure initially considered
at this site included “sliding instability and piping and erosion from underseepage.” The
report then goes on to say
Piping erosion from underseepage is unlikely because the I-walls were founded in a
clay levee fill, a marsh layer made up of organics, clay and silt, and a clay layer.
Because of the thickness, the low permeabilities of these materials, and the relatively
short duration of the storm, this failure mode was considered not likely and was
eliminated as a possible mode of failure.
This greatly underestimates the permeability, and especially the laterally permeability
of the marsh deposits. It also continues the very dangerous assumption that underseepage was
not a serious problem for “short duration” storm surge loading that plagued the original
design of many sections of the New Orleans regional flood defense system, and led to use of
sheetpile curtains that were far too short to effectively (and safely) cut off underseepage
flows. At least four major failures (and breaches) that caused large portions of the overall
flooding damage and loss of life during hurricane Katrina appear to have been principally due
to lack of appreciation of underseepage, and resulting inadequate (short) sheetpile cutoffs.
These are the major breach at the west bank near the north end of the London Avenue
drainage canal (see Section 8.3.9), the major breach at the east bank of the London Avenue
drainage canal farther to the south (see Section 8.3.8), and the two breaches on the east bank
of the IHNC at the west end of the Lower Ninth Ward discussed in this current section and in
Section 6.3.2. Exoneration, a priori, of underseepage dangers should be discontinued
immediately, and underseepage analyses should be required for the full regional flood
protection system.
Demonstration that underseepage occurred at this site can be based on arguments of
analogous conditions and levee performance at this site, and at the London Avenue drainage
canal breach sites, as well as at the site immediately to the north (as described in the next
section.) It can also be based on the observed difficulties encountered by McElwee
Construction in dewatering an excavation near the breach site immediately to the north (due
to massive underseepage flow through the marsh deposits that were not adequately cut off at
that site either.)
In addition, as noted in the IPET Draft Final Report in discussion of the two massive
breaches at the west end of the Lower Ninth Ward:
6 15
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Although it is clear that the walls were overtopped, and that their stability was
compromised by the erosion that occurred, it is also clear that one of the east side
breaches occurred before the wall was overtopped. Eyewitness reports indicate that
the water level in the 9th ward near Florida Avenue was rising as early as 5:00 AM,
when the water level in the IHNC was still below the top of the floodwall. Stability
analyses indicate that foundation instability would occur before overtopping at the
north breach on the east side of the IHNC. This breach location is thus the likely
source of the early flooding in the 9th Ward. Stability analyses indicate that the other
three breach locations would not have failed before they were overtopped.
Unfortunately, even IPET’s own analyses do not suggest a high likelihood of failure of
the north breach section at the canal water levels present as early as 5:00 a.m. (approximately
Elev. + 9 feet, MSL), so this would not appear to be the explanation for the observed water in
the neighborhood. Instead, it is proposed that the observed water rise on the inboard
(protected) side near Florida Avenue was more likely the result of large underseepage flows
through the highly pervious “marsh” deposits along this frontage.
Finally, clear and uncompromising evidence of the high lateral permeability of these
deposits at this site is presented in Figure 6.45, which shows a welldeveloped classic
crevasse splay that resulted from reverse underseepage through these same highly pervious
marsh deposits as the ponded floodwaters drained out from the Lower Ninth Ward after the
hurricane passed.
The New Orleans District of the USACE must stop “assuming” that shortterm storm
surges do not pose a significant risk associated with underseepage, and should instead begin
assuming that such underseepage is a potential risk and that it must be addressed either: (1)
with testing and analyses, (2) by means of sheetpile curtains extended deeply enough to
effectively cut off potentially dangerous underseepage, or (3) by means of wider and heavier
levee embankments (including inboard side stability berms) and the use of filtered drains at
the inboard toe of the levee to “vent” and thus draw down the potentially dangerous pore
pressures in that vicinity.
6.3.2
The IHNC East Bank (North) Breach at the Lower Ninth Ward
Figure 6.46 shows an aerial view of the partially repaired breach that occurred just to
the north of the breach discussed in the preceding Section 6.3.1. This second breach feature
was a much shorter feature, with a length of only approximately 250 feet.
This narrower, deep failure had similar initial geometry and stratigraphy to that of the
far longer section immediately to its south, as shown by the crosssection in Figure 6.47. At
this section, there is only the one main marsh layer, but most of the other soil conditions are
very similar to those at the adjacent breach section to the south.
Figures 6.47 and 6.48 show the crosssection and finite element mesh used for limit
equilibrium and coupled seepage analyses of this section.
6 16
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Soils data at this site were sparse, and consisted of a single preKatrina boring (located
nearby but just offsite, and with higher embankment overburden loads than at the breach
site), two ILIT borings and one ILIT CPTU probe, and several additional IPET CPT probes.
We were never able to fully determine the locations of the IPET CPT’s in plan view, but it
was assumed that they were located in adequate proximity as to be “representative”, and their
elevations were known with good precision so that these data could be used to at least crosscheck the other data available. Crosschecking the limited data (mainly from two CPTU
probes) with the data from the breach site immediately to the south showed strong
compatibility; accordingly similar properties (and OCR profiles, etc.) were modeled for
similar soil units at this section.
Figure 6.49 shows the calculated flownet equipotential lines for a canal water surge
elevation of +14 feet (MSL). Once again, as with the larger breach just to the south, the flow
travels through the continuous marsh layer that was left frustratingly open to flow by the
shallow sheetpiles that were an inadequate cutoff at this site.
Figure 6.50 shows pore pressure contours for the same conditions as Figure 6.49.
Once again the hydraulic uplift pressures represent potential instability with regard to lifting
or “blowout” of the thin surficial strata of impervious and relatively lightweight soils
overlying the marsh stratum at and near the inboard toe.
Similarly, as shown in Figure 6.51, seepage exit gradients at and near the inboard toe
are massively unsafe with regard to the initiation of seepage erosion and piping in these
relatively lightweight soils.
And finally, as with the adjacent breach section to the south, the section is also
marginally unstable with regard to limit equilibrium (Spencer’s method), as shown in Figure
6.52, as a result of underseepageinduced pore pressures and resultant loss of strength. The
most critical failure surface this time passes through (and largely within) the main marsh
layer, though a secondary failure surface concentrated near the interface between this marsh
layer and the overlying clay layer has a nearly similar (unstable) factor of safety.
Figures 6.53 and 6.54 show calculated transient pore pressures (6.53) at the top of the
marsh stratum beneath the inboard toe of the levee, and (6.54) at various depths beneath the
inboard levee toe. As for the breach section immediately to the south, the upper and lower
bound lateral permeability estimates are also shown in Figure 6.53; and again a large fraction
of the overall rise in canal water levels has resulted in corollary water pressure increases at the
inboard toe region by about 7:00 to 8:00 a.m.
Figure 6.55 shows the progressive increase in pore pressure at the top of the marsh
stratum vs. time, and the pore pressures are high enough to pose a very high risk of hydraulic
uplift (or “blowout”) at the inboard toe region.
Figure 6.56 shows a potential path to failure by means of lateral translational
foundation instability; reaching a condition of marginal lateral instability (with full
6 17
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
development of a waterfilled gap at the outboard side of the sheetpile curtain) at a canal
water elevation of approximately +13 to +14 feet (MSL).
Here again, as with the larger breach section immediately to the south, this breach
section is analytically unstable by a number of potential mechanisms, all of them associated
with underseepage flow passing beneath the sheetpile curtain. These potential mechanisms
are:
1. Seepage erosion and piping due to excessive exit gradients at the inboard toe.
2. Hydraulic uplift or “blowout” at the inboard toe.
3. Translational stability failure, as a result of reduction in strength of the foundation
soils at the inboard side due to underseepageinduced pore pressure increases.
As with the larger breach to the south, it is our investigation’s position that despite the
fact that overtopping (and resultant erosion at the inboard toe of the floodwall) was also
occurring, this failure was the result of one or more of the underseepageinduced mechanisms
above. (Two or more of these may have acted in concert.)
This site has a welldocumented history of underseepage problems; McElwee
Construction had great difficulty dewatering an excavation at this site during earlier
construction, and residents of the neighborhood had also previously reported problems with
seepage at the inboard toe.
Based on the geometry of the postfailure configuration (see Figure 6.46), this narrow,
deep failure appears to have most likely caused by either by seepage erosion and piping, or by
a combination of hydraulic uplift (“blowout”) followed by piping. The calculated high exit
gradients, and the hydraulic uplift pressures at the inboard toe region, would strongly support
this.
The IPET interim draft report also concluded that foundation instability was the cause
of the failure and breach at this site. The failure mechanism favored in those analyses,
however, was based on a semirotational failure dominated by undrained shear failure through
the soft clays underlying the marsh stratum, as shown in Figure 6.58. IPET concluded that
this failure occurred at a relatively early stage, at a canal water level of only Elevation +9 feet
(MSL), and that this early failure accounted for observations of ponding of water along this
general levee frontage well in advance of the failure of the larger breach section to the south.
Figure 6.59 shows the IPET interpretation of shear strength data for this section, and
the red lines are the IPET shear strength profiles for stability analyses (IPET: June 1, 2006.)
In this figure, the values of undrained shear strength based on the CPT tip resistance data are
based on a CPT tip factor of Nk = 15. This appears to be an overly conservative value of Nk
within the lower clay stratum, as the CPTbased shear strengths within this stratum are
significantly lower then the trend based on the unconsolidatedundrained triaxial tests on
“undisturbed” samples obtained with a 5inch diameter thinwalled fixedpiston sampler. In
6 18
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
any case, the shear strength profile used by the IPET analyses within this layer is well to the
left (lower than) the vast majority of the data available.
Our own studies determined (based on Bq values from the CPTU) indicated that values
of Nkt = 12 were more appropriate for this lower soft clay (CH) layer, and the resulting reinterpretation of this CPT data based on Nkt = 12 is shown (with a dark blue trace) superposed over the previous Figure 6.59 in Figure 6.60. Similarly, the dark blue lines in Figure
6.60 show our (ILIT) interpretation of the layering at this location, and the light blue dashed
lines show our interpretations of shear strength vs. depth at this section. The IPET shear
strength interpretation, in addition to being low, was also based on the assumption that this
lower clay stratum was normally consolidated over its full depth. As shown previously in
Figures 6.25 and 6.26, our own interpretations showed several clear desiccationinduced
overconsolidation profiles near the middle and top of this clay layer, and additional moderate
overconsolidation near the base (likely to the base being welldrained and thus partially
overconsolidated due to secondary compression), and these are reflected in our ILIT shear
strength profile. In this figure, the CPTUbased shear strengths (based on Nkt = 12) can be
seen to be in better agreement with the other shear strength data, and the overall shear strength
vs. depth profile is more consistent with this data.
Repetition of the limit equilibrium analysis (Spencer’s Method) of the failure surface
shown in Figure 6.58, but using our own (ILIT) interpretation of undrained shear strengths
within the critical lower clay layer (Figure 6.60) results in a calculated factor of safety, even
conservatively assuming the presence of a waterfilled gap on the outboard side of the
sheetpile curtain, of FS = 1.89. This underestimates the actual overall Factor of Safety,
which should actually be on the order of 10% to 15% higher based on threedimensional
considerations for this narrow, deep failure. It is therefore not likely that a deep, semi
rotational failure occurred, early on and at a relatively low canal water level, at this site.
The need of the IPET analyses to provide an early failure at this north breach site in
order to explain the significant observed ponding of waters along this frontage prior to the
occurrence of the large breach farther to the south, and to do so without consideration of
underseepage as a potential source of this water, resulted in an apparently unrealistic analysis
and an indefensible failure mechanism.
If the IPET team had been made aware of the pervasive history of underseepage
problems along this frontage, they would surely have considered and analyzed underseepagerelated failure modes for the two large breaches along this section of the east bank of the
IHNC. This information was apparently not available to the IPET analysis team, however,
reflecting insufficient communication between groups and teams across the overly modular,
subteamorganized IPET studies. In addition, the pervasive failure of the New Orleans
District of the USACE to adequately consider and analyze underseepage during preKatrina
design of considerable portions of the regional flood protection system was continued in the
postevent IPET studies of these two failed sections.
The New Orleans regional flood protection systems need to be thoroughly reassessed,
and reanalyzed as necessary, with regard to potential additional underseepagerelated
6 19
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
vulnerabilities. And then these must be mitigated in order to develop a safe and reliable
overall system..
6.3.3
Summary
The two large breaches on the east bank of the IHNC (at the west end of the Lower
Ninth Ward) both occurred at sites where overtopping occurred. Despite the occurrence of
overtopping, and resultant erosion of trenches at the inboard sides of the concrete floodwalls,
this overtopping does not appear to have been the cause of the two failures. Instead, these two
failures appear to have resulted from underseepageinduced instability; either due to erosion
and piping at the inboard toe, “blowout”, or translational instability due to strength reduction
in the inboard side foundation soils due to underseepageinduced pore pressure increases.
This represents a potentially critical difference from the findings to date from the
Corps’ IPET study; as the remedy for overtopping, trench erosion, and unbracing at the top of
the floodwalls is very different from the remedy for underseepageinduced instability
problems. The USACE has invested large resources to replace “Iwalls” with “Twalls”, and
to install concrete splash pads behind additional “Iwall” sections. This is laudible, but it will
not also effectively mitigate underseepagerelated problems.
Remedies for the underseepage related problems revealed by these analyses would
include either extension of the sheetpile curtains to greater depths in order to more effectively
“cut off” underseepage, or widening of the levee embankments to the inboard side and
installation of filtered drains at the inboard toes in order to safely draw down the
underseepageinduced high pore pressures in that area.
Analyses of the IHNC failure sections, and of sections of the three drainage canals in
the main (downtown) New Orleans protected basin (see Chapter 8), have shown that
unreasonably short sheetpile curtains of too limited penetration as to effectively cut off
underseepage are likely to be endemic throughout many parts of the New Orleans regional
flood protection system. Indeed, the USACE at a number of breach repair sites is replacing
sheetpiles with (preKatrina) lengths of 18 to 24 feet with far longer (deeper penetrating)
sheetpiles with lengths of 60 feet and greater as part of the repair operations; an unusually
frank admission that significantly deeper sheetpiles were warranted at those sections.
There is now a need to review, and to reanalyze as necessary, essentially the entire
regional flood protection system with regard to potential vulnerability associated with
underseepage (and inadequately deep sheetpiles), and to remedy these problems at sites where
necessary in order to ensure overall safety of the system.
6.4
Summary and Findings
The catastrophic flooding of the St. Bernard and Lower Ninth ward protected basin
was primarily due to: (1) catastrophic erosion of the MRGO frontage levees, and (2) a pair of
large failures (and breaches) on the east bank of the IHNC at the west end of the Lower Ninth
Ward.
6 20
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
The catastrophic erosion of large portions of the nearly 11mile long MRGO frontage
levees was the result in large part of the use of unsuitable sand and shell sand fills (with low
resistance to erosion) for major portions of these embankments. Large portions of these fill
materials came from spoils dredged from the excavation of the adjacent MRGO channel, and
the shortterm savings achieved by the use of these soils now pale in comparison to the
massive damages and loss of life that resulted. Because these levees eroded so rapidly, and so
massively, the storm surge was able to push largely undiminished across a wide area of
undeveloped swampland behind the main frontage levees, cross a lower secondary levee (the
Forty Arpent levee) that has not been intended to have to face an undiminished rising storm
surge, and then charged into the populated zones of St. Bernard Parish with catastrophic
consequences.
Because it passed so quickly and so completely through the frontage levees, the surge
filled the St. Bernard basin to an elevation of +12 feet above sea level; inundating homes and
businesses located well above sea level that had expected to be safe, and inundating lower
lying properties to great depths.
The use of intrinsically highly erodeable fills, especially clean sands, and the even
more dangerous lightweight shell sands, should be reconsidered. The use of such materials as
levee embankment fill, especially without taking appropriate measures to mitigate the
erosional hazards associated with these (e.g.: sheetpile cutoff, erosion protection and armoring
of exposed slope faces and crests, etc.) is inadvisable when constructing levees intended to
protect large populations at risk.
The two large breaches at the east bank of the IHNC (at the west end of the Lower
Ninth Ward) both appear to have resulted not from overtopping, but rather from underseepage
beneath the inadequately deep sheetpile curtains at these two sections. Overall, four of the
eight most significant failures (breaches) that occurred during hurricane Katrina (the eight
breaches that caused the greatest damages and loss of life) appear to have been due to
inadequate attention to underseepage during initial design, and resulting sheetpile curtains that
were far too short as to suitably cutoff or minimize these underseepage flows (see also
Sections 8.3.8 and 8.3.9). This appears to be a widespread problem throughout the New
Orleans regional flood protection system; the entire system should be reevaluated with
respect to this potential hazard, and mitigation implemented as necessary.
6.5 References
GEOSLOPE/W (2004) “Complete Set of Manuals”, John Krahn (Edit.), Calgary, Alberta,
Canada.
IPET – Interagency Performance Evaluation Task Force (2006) Performance Evaluation
Status and Interim Results, Report 2 of a Series, Performance Evaluation of the New
Orleans and Southeast Lousiana Hurricane Protection System, March 10, US Army
Corps of Engineers.
6 21
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Karlsrud,K., Lunne, T and Brattlien (1996) “Improved CPTU interpretations based on block
samples”, Nordic Geotechnical Conference, 12 Reykjavik 1996. Proc, Vol.1, pp 195201.
Kemp, P., (2006), Personal Communication
Lunne, T., Lacasse, S., and Rad, N.S., (1994) “General report: SPT, CPT, PMT, and recent
developments in insitu testing”, Proceedings, 12th International Conference on soil
mechanics and foundation engineering, Vol.4, Rio de Janeiro, pp. 23392403
Mashriqui, H., (2006), Personal Communication
Mayne, P,W and Mitchell, J.K (1988) “Profiling of Overconsolidation Ratio in Clays by Field
Vane”, Canadian Geotechnical Journal, Vol 25, No 1, February, pp 150157.
PLAXIS Finite Element Code for Soil and Rock Analyses (2004). “Complete Set of
Manuals”, V8.2 , Brinkgreve yVermeer (Edit.), Balkema, Rotterdam, Brookfield.
U.S. Army Corps of Engineers. (1968) Lake Pontchartrain, LA and Vicinity Lake
Pontchartrain Barrier Plan, Design Memorandum No 2, General Supplement No. 8,
Inner Harbor Navigation Canal Remaining Levees, New Orleans District, New
Orleans LA.
U.S. Army Corps of Engineers. (1969) Lake Pontchartrain, LA and Vicinity Lake
Pontchartrain Barrier Plan, Design Memorandum No 2, General Supplement No. 8,
West Levee Vicinity France Road and Florida Avenue, New Orleans District, New
Orleans LA.
U.S. Army Corps of Engineers. (1971) Lake Pontchartrain, LA and Vicinity Lake
Pontchartrain Barrier Plan, Design Memorandum No 2, General Supplement No. 8,
Modification of Protective Alignment and Pertinent Design Information I.H.N.C.
Remaining Levees West Levee Vicinity France Road and Florida Avenue
Containerization Complex, New Orleans District, New Orleans LA.
6 22
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Figure 6.1: Map showing locations of levee breaches and distressed levee sections.
Independent Levee
Investigation Team
623
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Source: LSU Hurricane Center
Figure 6.2: Depth of flooding of St. Bernard Parish and the Lower Ninth Ward on Sept.
2nd (4 days after Hurricane Katrina).
624
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 6.3: Catastrophically eroded levee section along the northeast frontage of the St.
Bernard Parish protected basis, fronting the MRGO channel.
Figure 6.4: Catastrophic erosion of levee embankment leaving central sheetpile curtain;
also along the northeast frontage of the St. Bernard Parish protected basin
fronting the MRGO channel.
625
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 6.5: Extensive erosion of levees along the MRGO frontage at the northeast edge of the
St. Bernard Parish protected area.
Photo by: Dr. Juan Pestana
Figure 6.6: View of sheetpiles left behind at catastrophically eroded section of the MRGO
frontage levee.
626
Independent Levee
Investigation Team
g
627
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Figure 6.7: LIDAR survey data showing crest elevations before and after hurricane Katrina along a section of the MRGO
frontage levees at bayou Bienvenue, and corresponding postevent plan view aerial photo along the same section.
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Eroded Breach
Figure 6.8: Large eroded breach at the contact between the south end of the massive concrete
navigational lock structure at Bayou Bienvenue and the adjacent levee section.
628
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photo courtesy of the U.S. Army Corps of Engineers
Figure 6.9(a): Aerial view of the large erosional breach at the contact between the north
end of the concrete navigational lock structure and the adjoining levee
embankment at Bayou Dupres.
Figure 6.9(b): View looking to the inboard side from the breach shown in Figure 6.5 above;
showing eroded shell sand detritus deposited from the breach.
629
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 6.10: Approximate hydrograph of storm surge elevation (feet, MSL) vs. time at the
west end of Lake Borgne.
[IPET Interim Report, April, 2006]
Back-Slope
Critical Erosion
(a) Sheet flow overtopping erosion of the lower back slope face
Notching
Scalloping
Through-Flow
Erosion
(b) Wave erosion of the front face, and through-flow erosion of the lower back face
Figure 6.11: Schematic illustration of two different sets of modes of levee erosion.
630
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 6.12: Photo of outboard side waveinduced erosion on the MRGO levee frontage at
the northeast edge of the St. Bernard/Ninth Ward protected area.
Figure 6.13: Photo of outboard side levee erosion and crest “notching”, as well as crenellation,
due to storm wave erosion and overtopping along the MRGO frontage levees
at the northeast edge of the St. Bernard/Ninth Ward protected area.
631
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 6.14: View of the secondary (Forty Arpent) levee across the middle of the St. Bernard
protected basin.
Figure 6.15: View of tree limbs and detritus on roofs of homes in St. Bernard Parish.
632
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 6.16: Car tossed and flipped in St. Bernard Parish.
Figure 6.17: Boat deposited in neighborhood in St. Bernard Parish.
633
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Source: IPET Interim Report No. 2; April, 2006
Figure 6.18: Hydrographs showing measured (and photographed) water levels at gage stations
along the Inner Harbor Navigation Channel (IHNC).
634
Figure 6.19: Oblique view of the (south) levee break at the Inner Harbor Navigation Canal into the lower Ninth Ward.
Independent Levee
Investigation Team
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
635
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph by: Rune Storesund
Figure 6.20: Closeup view of the large barge that entered through the south breach at the
east bank of the IHNC at the west end of the Lower Ninth Ward.
Photograph by: Rune Storesund
Figure 6.21: Closeup view of crushed (impacted) concrete floodwall at the south end of the
south breach at the east bank of the IHNC at the west end of the Lower Ninth Ward.
636
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph by: Rune Storesund
Figure 6.22: Eroded trench at the rear (inboard) side of the floodwall at the south end of the
south breach at the east bank of the IHNC at the west end of the Lower Ninth
Ward.
637
NO 7 (1985)
Sta 44+00
EL +8.4’
Figure 6.23: Reinterpreted longitudinal subsurface soil profile, showing location of breach
section on the east bank of the Lower ninth Ward Levee.
NO 6 (1985)
Sta 42+00
EL +8.6’
BREACH ~200’
Independent Lev
Independent
Levee
ee
Investigat
Invest
igation
ion Team
New Orleans Levee Systemss
Hurricane Ka
Katrin
trina
a
July 31, 2006
638
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
IHNC- S – CPT- 1
Figure 6.24: Crosssection showing location of borings for Lower Ninth Ward, East Bank, South Breach.
639
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Shear Strength distribution for Grey CH
at 9th Ward, IHNC, South Breach, Levee Toe
Grey CH
at 9th Ward, IHNC, South Breach, Levee Toe
Su (psf)
0
500
1000
1500
Su/P'v
Su/
Nk = 12
2000
2500
3000
0
-15
0.1
0.2
0.3
0.4
0.5
0.6
Grey CH
at 9th Ward, IHNC, South Breach, Levee Toe
Nk = 12
Nk = 12
OCR
0.7
0.8
0.9
1
1
-15
1.5
2
2.5
3
3.5
-15
(su/'v) NC ~ 0.33
-20
-20
-20
SILT
SILT
SILT
IHNC-S-CPT-1 (ILIT)
IHNC-S-CPT-2 (ILIT)
IHNC-S-CPT-3 (ILIT)
4
5
-25
-25
-30
-30
-35
-40
-30
CH
-35
IHNC-S-CPT-1 (ILIT)
IHNC-S-CPT-2 (ILIT)
-40
Elevation (ft)
Elevation (ft)
CH
Elevation (ft)
-25
CH
-35
IHNC-S-CPT-1 (ILIT)
IHNC-S-CPT-2 (ILIT)
-40
CL with sand and silt lenses
CL with sand and silt lenses
CL with sand and silt lenses
-45
-45
-45
-50
-50
-50
Figure 6.25: Plots of (a) OCR vs. Depth and (b) Su vs. Depth for the soft gray marsh clay (CH) at the inboard toe and further to the
landside (not under levee embankment overburden pressure) – Lower Ninth Ward, IHNC East Bank, South Breach site.
640
4
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Shear Strength distribution for Marsh
at 9th Ward, IHNC, South Breach, Levee Toe
Marsh at 9th Ward, IHNC, South Breach, Levee Toe
Marsh at 9th Ward, IHNC, South Breach, Levee Toe
Nk = 15
Nk = 15
Su (psf)
0
500
1000
Nk = 15
Su/'
Su/Pvo
Su/P
Su/'
vo
1500
2000
2500
0
3000
5
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
2.2
2.4
2.6
2.8
0
3
IHNC-S-CPT-1 (ILIT)
IHNC-S-CPT-2 (ILIT)
IHNC-S-CPT-3 (ILIT)
UU-TX
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
2.2
2.4
IHNC-S-CPT-1 (ILIT)
IHNC-S-CPT-2 (ILIT)
IHNC-CPT-3 (ILIT)
Lab Vane
0
0
0
Marsh
Marsh
Marsh
-5
-10
CH/CL
-10
Elevation (ft)
-5
Elevation (ft)
Elevation (ft)
-5
CH/CL
CH/CL
-10
Marsh
-15
2.6
5
5
-15
-15
Marsh
Marsh
-20
-20
-20
-25
-25
-25
Figure 6.26: Plots of (a) OCR vs. Depth and (b) Su vs. Depth for the shallow marsh and clay deposits at the inboard toe and further to
the landside (not under levee embankment overburden pressure) – Lower Ninth Ward, IHNC East Bank, South Breach site.
641
2.8
3
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 6.27: Geometry and input parameters for FEM (PLAXIS) stability analyses for Lower Ninth Ward, IHNC East Bank, South Breach.
642
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 6.28: Deformed mesh after modeling staged construction for the levee and allowing for consolidation for Lower Ninth Ward,
IHNC East Bank, South Breach.
643
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 6.29: Deformed mesh for storm surge elevation of +13.5ft (MSL) for Lower Ninth
Ward, East Bank, South Breach.
644
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Storm Surge = +12.5 feet (MSL)
Figure 6.30: Normalized shear strain contours (shear strain divided by strain to failure) for a storm surge at Elev. + 12.5 feet (MSL) at
the Lower Ninth Ward, IHNC East Bank, South breach site; gapping at outboard toe of floodwall is developed fully
through the embankment fill material.
645
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Storm Surge = +14 feet (MSL)
Figure 6.31: Normalized shear strain contours (shear strain divided by strain to failure) for a storm surge at Elev. + 14 feet (MSL) at
the Lower Ninth Ward, IHNC East Bank, South breach site; gapping at outboard toe of floodwall is fully developed.
646
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Elevation, WL (ft)
Independent Levee
Investigation Team
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
0.7
0.8
0.9
1.0
1.1
1.2
1.3
1.4
1.5
FS
Figure 6.32: Calculated Factors of Safety for two modes based on PLAXIS analyses of the
Lower Ninth Ward South breach site (east bank IHNC) for various canal water
elevations; showing the bestestimated path to failure.
647
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
MA
MAT
TERIAL
Fill
PAR
PARAMETER
j#(pc
(pcff)
i#
c (psf)
Kh (ft/hr
t/hr))
Kh (cm/s)
m/s)
Kv
Kv//Kh
T#-#
T#-
105
105
0
90
900
0
1.17E17E-0
04
9.91
91E-07
E-07
1
0.35
0.35
Up
Upper
perC
CH
95
0
80
800
0
2.00E00E-0
04
1.69
69E-06
E-06
0.
0.333
333
0.35
0.35
Up
Upper
perM
Marsh
85
28
0
1.10E+
10E+0
00
9.31
31E-03
E-03
0.25
0.25
0.5
OCG
GreyC
CH
NCGrey
GreyC
CH
95
0
50
500
0
2.00E00E-0
04
1.69
69E-06
E-06
0.
0.333
333
0.35
0.35
95
0
Su/p: 0.28
0.28
2.00E00E-0
04
1.69
69E-06
E-06
0.
0.333
333
0.35
0.35
Lowe
LowerM
Mar
ars
sh
85
28
0
1.10E+
10E+0
00
9.31
31E-03
E-03
0.25
0.25
0.5
Silt
110
110
0
60
600
0
1.17E17E-0
04
9.91
91E-07
E-07
0.
0.333
333
0.41
0.41
Lean
LeanC
Clay
Sands
Sands
100
100
0
60
600
0
2.00E00E-0
04
1.69
69E-06
E-06
0.
0.333
333
0.38
0.38
120
120
30
0
1.00E+
00E+0
00
8.5E-0
8.5E-03
3
0.
0.5
5
0.42
0.42
10
1
Gaps
100
100
* Fredl
edlun
und
d et al,G
Green andCor
Corey,Va
Van Gen
Genu
uch
chten
ten
Canal Side
Land Side
e 6.33: Geotechnical cross-section for analysis of the IHNC east bank, south breach.
6-48
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 6.34: Finite difference mesh for seepage analyses for IHNC east bank, south breach.
6-49
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 6.35: Flow net for the south breach on IHNC; storm surge at 14.4ft (MSL). Head contours at 1-foot intervals of head.
Potential Hydraulic Uplift
Figure 6.36: Pressure contours for the south breach on IHNC. Storm surge at 14.4ft (MSL).
6-50
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Io: 1.0
Figure 6.3
Io: 0.8
7: Hydraulic gradients for the south breach on IHNC east bank; storm surge at 14.4ft (MSL). Maximum exit gradient at the
levee toe is io ᄃ 0.8 to 1.0.
6-51
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Pressure vs Time at top of Lower Marsh at Location D
Steady State developed at 14.4ft (MSL)
7:45 a.m.
1200
1100
92% of
Steady
State
Pressure (psf)
1000
900
800
700
600
4:30
Steady State
developed at 10ft
(MSL) (Initial
Condition)
Transient flow calculation at 14.4ft (MSL) and 4
hours. (89 to 97% of Steady State)
5:30
6:30
7:30
08/29/05 Time, CDT
Figure 6.38: Transient flow pore pressure generation for the south breach on IHNC east bank.
Pore Pressure vs Time and Depth in Location D
Best Estimate for Transient Flow
generated Pore Pressure
1600
1400
Pressure (psf)
1200
1000
Steady State
generated Pore
Pressure
800
600
400
200
0
4:30
-2.8ft
-8.0ft
-11.0ft
-11.0ft
5:30
6:30
-19.0ft
-19.0ft
-5.0ft
7:30
8:30
08/29/05 Time, CDT
Figure 6.39: Pore pressure generation at different times and depths at the inboard toe of the
south breach on IHNC east bank.
6-52
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
1 .0 0
0.92
0 .9 0
0 .8 0
0 .7 0
Top of Lower Marsh (1)
0 .6 0
0 .5 0
0 .4 0
0 .3 0
0 .2 0
0 .1 0
0 .0 0
Exit Gradient (2)
(2)
(1)
gradients at the south breach on IHNC east bank.
Figure 6.40: Hydraulic
6-53
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
1500
1400
Steady State at 14ft
Pore Pressure (psf)
1300
1200
SS at 14ft
5:30a.m . CDT
6:30a.m . CDT
7:30a.m . CDT
8:30a.m . CDT
1100
1000
900
800
700
Transient Flow at 14ft
600
500
50
100
150
200
250
300
Zone with Total
overburden stress
less than 1000psf
Second Marsh Layer
Figure 6.41: Pore pressure versus horizontal distance and time at the top of the second marsh layer; south breach on IHNC east bank.
6-54
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
FS = 0.99
Figure 6.42: Critical failure surface for the south breach on IHNCeast bank; storm surge at
14ft (MSL), failure through upper marsh layer with gap at front of sheetpiles
fully developed.
FS = 1.03
Figure 6.43: Deeper failure surface for the south breach on IHNCeast bank; storm surge at
14ft (MSL), gap at front of sheetpiles fully developed.
6-55
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
15
14
Storm Surge (ft-MSL)
(ft-MSL)
13
12
11
10
Lower Marsh, gap
generated
Lower Marsh, No Gap
generated
9
Upper Marsh, Gap
generated
8
0.6
0.8
1
1.2
1.4
1.6
1.8
2
FS
Figure 6.44: Calculated Factors of Safety for three modes based on SLOPE/W analyses of the
Lower Ninth Ward South breach site (east bank IHNC) for various canal water
elevations; showing the best-estimated path to failure.
6-56
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 6.45: Aerial view of the south breach at the east bank of the IHNC (at the west end of
the Ninth Ward), showing the crevasse splay generated by reverse drainage
flow.
[Photograph by U.S. Army Corps of Engineers]
6-57
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 6.46: Aerial view of the partially repaired north breach on the east bank of the IHNC
at the west end of the Lower Ninth Ward.
6-58
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
MA
MAT
TERIAL
j#(pc
(pcff)
i#
PA
PAR
RAMETER
c (ps
(psff)
Kh (ft
(ft//hr)
Kh (c
(cm/
m/s
s)
Kv/K
Kv/Kh
T#-#
T#-
Fill
Fill
105
0
900
1.17E-04
17E-04
9.91E-0
9.91E-07
7
1
0.35
0.35
CH
95
0
800
2.00E-04
00E-04
1.69E-0
1.69E-06
6
0.333
0.35
0.35
OC G
GrreyCCH
95
0
500
2.00E-04
00E-04
1.69E-0
1.69E-06
6
0.333
0.35
0.35
NCGrey
GreyCH
CH
Mar
Mars
sh
95
0
Su/p: 0.28
0.28
2.00E-04
00E-04
1.69E-0
1.69E-06
6
0.333
0.35
0.35
85
28
0
1.10E
10E+00
9.31E-0
9.31E-03
3
0.25
0.25
0.5
Lean
Lean Clay
100
0
600
2.00E-04
00E-04
1.69E-0
1.69E-06
6
0.333
0.38
0.38
Sandss
Sand
120
30
0
1.00E
00E+00
8.5
8.5E
E-03
-03
0.5
0.42
0.42
10
1
Gap
Gapss
100
* Fr
Fredlu
edlund et al, Gr
GreenaandCor
Corey
ey,,Va
Van Genuc
Genuch
hten
e 6.47: Geotechnical cross-section for analysis of the IHNC east bank, north breach.
6-59
Figur
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Finite difference mesh for seepage analyses for IHNC east bank, north breach.
Figure 6.48:
6-60
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 6.49: Flow net generation for the north breach on IHNC east bank. Storm surge at +14.4ft (MSL). Head contours at 1 foot
intervals of hydraulic head.
Potential Hydraulic Uplift
Figure 6.50: Pressure contours for the north breach on IHNC east bank. Storm surge at +14.4ft (MSL). Pore pressure contours
2
.
at intervals of 62.4 lb/ft
6-61
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Io: 1.0
Io: 1.5
Io: 2.5
Figure 6.51: Hydraulic gradients for the north breach on IHNC. Storm surge at +14.4ft (MSL). Maximum exit gradient on the
o ᄃ 1.0, and io ᄃ 1.5 to 2.5 at the lower toe.
upper levee toe is i
6-62
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
FS=0.98
Figure 6.52: Critical potential stability failure surface for the north breach on IHNC. Storm
surge at +14 ft (MSL).
6-63
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Pressure vs Time at top of Lower Marsh at Location D
1050
Steady State developed at 14.4ft (MSL)
1000
950
Steady State
developed at 10ft
(MSL) (Initial
Condition)
Pressure (psf)
900
850
90% of
Steady
State
800
750
700
Transient flow calculation at 14.4ft (MSL) and 4
hours. (87 to 93% of Steady State)
650
600
4:30
5:30
6:30
7:30
8:30
8/29/05 Time, CDT
Figure 6.53: Transient flow pore pressure generation for the north breach on IHNC east bank.
Pore Pressure vs Time and Depth in Location D
1800
Best Estimate for Transient Flow
generated Pore Pressure
Steady State generated
Pore Pressure
1600
1400
Pressure (psf)
1200
1000
800
600
400
200
0
4:30
-5.0ft
-8.0ft
-14.0ft
5:30
6:30
-21.0ft
-11.0ft
7:30
8/29/05 Time, CDT
Figure 6.54: Pore pressure generation at different times and depths at the inboard toe of the
north breach on IHNC east bank.
6-64
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
3.50
Exit Gradient (2)
3.00
2.50
Top of Lower Marsh (1)
2.00
1.50
1.00
0.50
0.00
(2)
(1)
Figure 6.55: Hydraulic gradients on the south breach on IHNC east bank.
6-65
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
1600
Steady State at 14.4ft
1400
1200
1000
Zone with Total
overburden.
stress less than
800psf
800
600
400
Steady State at 14ft
8:30a.m. CDT
7:30a.m. CDTr
6:30a.m. CDT
5:30a.m. CDTr
Transient Flow at 14ft
200
0
150.00
200.00
250.00
300.00
350.00
Figure 6.56: Pore pressure versus horizontal distance and time on the north breach on IHNC east bank.
6-66
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
15
With Gap
No Gap
14
13
Storm Surge (ft) MSL
12
11
10
9
8
7
6
5
0
0.5
1
1.5
2
2.5
3
FS
Figure 6.57: Factor of Safety vs. water elevation (ft, MSL) for the north breach, east bank of
the IHNC at the west end of the Lower Ninth Ward.
6-67
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 6.58: IPET shear strength profile; IHNC east bank/Lower Ninth Ward (North) breach.
6-68
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 6.59: Critical limit equilibrium stability failure mode from IPET Draft Final Report;
canal water elevation at +9 feet (MSL). Factor of Safety: 1.03.
[IPET; June 1, 2006]
6-69
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
IHNC-N-CPT-1 (ILIT)
Nk = 12 (CH), Nk = 15 (Marsh)
Sand
Clay (CH)
Marsh
ILIT best strength estimate
Figure 6.60: Re-interpretation of shear strength, and the ILIT shear strength profile.
6-70
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
CHAPTER SEVEN: THE NEW ORLEANS EAST
PROTECTED AREA
7.1
Introduction
Figure 7.1 shows the New Orleans East (NEO) protected area, a contiguously ringed
area that includes some of the lowest ground in the metropolitan region. This is a repeat of
Figure 2.4, and the blue stars again represent levee breaches, and the red stars locations of
significant levee distress. Multiple levee breaches and significant overtopping produced
complete flooding of this protected area, and the resulting damage was extensive.
The New Orleans East protected area had a preKatrina population of approximately
96,000 people residing in over 30,000 households. Most of these residences were located in
the western portion of the polder (protected area) between Lake Pontchartrain and Chef
Menteur Highway (Highway I10). The residential neighborhoods are suburban in character,
with many of the homes dating to the 1960s and 1970s. Ironically, a number of these homes
were built in response to the devastation inflicted by Hurricane Betsy in 1965, which had also
left much of New Orleans East submerged by floodwater. This protected area also includes
an industrial corridor located along its southern fringe, adjacent to the Gulf Intracoastal
Waterway (GIWW) which runs adjacent to its southern edge. The eastern limits of the
protected area are largely comprised of wetlands that border Lake Pontchartrain/Lake Borgne
water systems and/or the swamplands between them.
The New Orleans East protected area extends over approximately 70 square miles and
is bounded by Lake Pontchartrain to the north, the GIWW shipping channel to the south, and
the Inner Harbor Navigation Channel (IHNC) to the west. Lake Borgne abuts the south
facing levees at the southeast corner of this protected area.
Figure 7.3 shows the depths of flooding on September 2, four days after hurricane
Katrina, at a time when the water levels were at equilibrium with the still slightly swollen
waters of Lake Pontchartrain (Elev. ~ +1 foot, MSL), and this map of flooding depths thus
serves well to illustrate the distribution of ground elevations across this protected area.
Elevations typically range from approximately +10 feet to 8 feet (MSL), with the higher
elevation reaches located south of the Chef Menteur Highway. This Highway follows along a
ridge of “high ground” known as the Bayou Sauvage ridge which is the result of an earlier
river depositional channel (see Chapter 3), and this slight ridge serves to nearly separate the
large northern section of the protected area from a smaller basin to the south. This separation
was incomplete, however, as floodwaters managed to cross this ridge at a number of
locations.
The New Orleans East protected area encompasses some of the lowest elevation lands
in the greater New Orleans populated region, and the results of the full flooding of this
protected basin were thus catastrophic, especially with regard to damage to homes and
properties. As shown in Figure 2.12, loss of life was moderate (on the order of 120 persons,
to date), however, largely because of the relatively effective preevacuation of this exposed
7 1
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
outlying area, and the relatively moderate rate at which the waters eventually filled the low
lying populous areas at the western end of this protected area. Because the area flooded and
filled progressively over the course of the day on August 29, the storm surge subsided as it
filled and the eventual filling extended only to approximately +2 feet (MSL) in the populous
western end of the protected area; accordingly portions of the “high ground” along the
southwest edge of the protected area remained above water (as shown in Figure 7.4.) The
open, unpopulated eastern portion of the protected area initially filled to somewhat higher
elevations, however, as it was relatively rapidly filled by the massive beaching and erosion of
the New Orleans East back levees fronting the GIWW channel and Lake Borgne.
7.2
New Orleans East Hurricane Protection System
Figure 7.2 shows the results of a posthurricane assessment of the condition of the
primary levee system surrounding the protected area (IPET; March 10, 2006.) This protection
system, which includes earthen levees, Iwall, Twall, and sheet pile sections, was designed
by the USACE as part of the Lake Pontchartrain and Vicinity Hurricane Protection Project.
The NEO protected area also includes a secondary or "local" levee that separates the
developed portions of the region from the wetlands to the east (Figure 7.1). The primary
purpose of the secondary levee is interior drainage control rather than hurricane protection,
and it was of lesser height than the main frontage levees (elevations typically on the order of
+5 to +6 feet, MSL as opposed to elevations of +14 to +18 feet for the main perimeter
frontage levees.)
The New Orleans East hurricane protection system is divided for planning and
management purposes into individual segments, or "reaches," which are defined by physical
characteristics, elevation, and/or potential consequences. For consistency, the names assigned
to the individual reaches by the USACE will be used in this chapter. Figure 7.5 illustrates
these section designations, and also indicates the locations of other points that will be
discussed in this chapter.
The eastern edge of the protected area is defended by the New Orleans East Levee, an
approximately 8.5 mile long earthen levee segment consisting largely earthen levees with 3 to
4 horizontal: 1 vertical side slopes, fronted on the outboard side by cypress swamps and
wetlands. The southern boundary of the protected area (along the north bank of the eastwest
trending shared GIWW/MRGO channel) is defended by the New Orleans East Back Levee (to
the east) and the adjacent Citrus Back Levee (to the west). These two reaches, which together
measure approximately 18 miles in length, are largely comprised of earthen levee sections
interspersed with sections comprised of concrete floodwalls atop lower height earthen levee
sections and/or sheet pile wall segments. The IHNC East Levee is an approximately 3mile
reach primarily comprised of concrete floodwalls atop earthen levees. As its name implies,
the portion of the levee system separates the western edge of the protected area from the
adjacent IHNC. Continuing clockwise are the New Orleans Lakefront and Citrus Lakefront
Levees, which include both earthen levees and composite concrete floodwall/earthen levee
sections. Finally, the eastern 12.5 miles of the northern Lake Pontchartrain frontage is the
New Orleans East Lakefront levee, and earthen levee with geometry similar to that of the
adjoining New Orleans East Back Levee (just around the corner, along the eastern edge of the
protected area.)
7 2
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
7.3
Performance of the New Orleans East Hurricane Protection System
in Hurricane Katrina
7.3.1
Overview
Figures 7.1 and 7.2 show the locations of damage to the levee system surrounding the
New Orleans East (NOE) protected area. The most significant damage to the system occurred
to East Back Levee that fronts the GIWW and Lake Borgne. Here the storm surge completely
destroyed (and massively eroded) large expanses of earthen levee in the southeastern corner
of the NOE protected area. Additional smaller, but nevertheless significant breaches also
occurred along other portions of these NOE back levee reaches. As the storm surge next
passed west two significant levee breaches occurred, both due to overtopping, along the north
bank of the eastwest trending channel of the GIWW/MRGO. Damage (mostly in the form of
scour) also occurred along the IHNC East Levee and portions of the New Orleans Lakefront
Levee located near the Lakefront Airport as the storm surge raised the water levels within the
IHNC. Finally, the reverse (counterclockwise) swirl of the storm winds raised the levels
along the south shore of Lake Pontchartrain. Portions of the levee system fronting Lake
Pontchartrain, such as the New Orleans Lakefront, Citrus Lakefront, and New Orleans East
Lakefront Levees, generally performed well in the hurricane, as did most of the New Orleans
East Levee located to the east.
7.3.2 Chronology of Events in the New Orleans East Protected Area
It is believed that water first entered the NOE protected area between about 5:00 a.m.
to 5:45 a.m. on August 29 as a large section of earthen levee in the southeastern corner of the
protected area catastrophically eroded and breached, as a result of wave action and possible
seepage associated with the rising storm surge from Lake Borgne. The levee system at this
location was so severely damaged that it ultimately did little, if anything, to impede the storm
surge that later peaked at this location. Water entering the NOE protected area through this
breach then crossed the adjacent wetlands before being channeled, initially, by the Bayou
Sauvage ridge (high ground underlying Highway 90) to the west. Video footage (and
eyewitnesses) recorded at the Entergy Power Utility Plant near the Michoud Canal show this
inflowing water appearing to arrive from the east at approximately 6:15 a.m. Storm surge
simulations by the IPET team (IPET Report 2, March 10, 2006) indicate relatively low water
levels in the adjacent GIWW at the 6:00 a.m. hour, indicating that the water first arriving at
the Entergy plant did not result from simple overtopping of the levees closely adjacent to this
plant.
The storm surge then passed westward along the eastwest trending GIWW/MRGO
shared channel and produced levee damage and several smaller breaches on the north side of
the channel. These breaches added to the water already flowing into the area through the
major breaches in the southeast corner. The surge then continued westward reaching the
GIWW's “T” intersection with the IHNC channel. The surge passed to the north (and south)
along the IHNC, and damaged a number of sections along the IHNC frontage.
7 3
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
As the hurricane then passed northward to the east of New Orleans, the
counterclockwise direction of the storm winds also produced a storm surge southward
towards the shore of Lake Pontchartrain. The lake level rose, but largely stayed below the
crests of most of the lakefront levees. The lake rose approximately to the tops of the lakefront
levees at a number of locations, especially along the shoreline of New Orleans East, and there
was modest overtopping (storm surge + wave splashover) and some resulting erosion on the
crests and inboard faces of some lakefront levee sections along the Lake frontage. However,
there were no breaches in this area. Overtopping occurred over a section of floodwall near the
west end of the New Orleans East protected area lakefront, where the floodwall was lower
than the adjacent earthen levee sections. This, too, added to the flow into the New Orleans
East protected area, which was now beginning to fill with water even as the original storm
surges subsided. As shown in Figure 7.4, water depths ultimately approached 10 feet in area.
Sadly, some of the deepest waters were in the NOE protected area's principal residential
neighborhoods.
7.3.3 Damage to Levee System Frontages
The following sections summarize damage to the individual frontages of the levee
system (Figures 7.1 and 7.2).
For consistency, locations are referred to using the
designations assigned by the USACE Task Force Guardian levee system rebuilding team.
These names associated with each of the main levee sections are shown in Figure 7.5.
7.3.3.1 GIWW/Lake Borgne Frontage; the New Orleans East Back Levee
As shown in Figure 7.5, the New Orleans East back levee extends from the southeast
corner of the NOE protected area west along the GIWW waterway, and it fronts both the
GIWW channel and Lake Borgne as well. As noted earlier, the most severe damage to the
NOE Levee System occurred along an approximately 5,300 foot long section of the New
Orleans East Back levee, which is situated in the southeast corner of the protected area
(Figures 7.1 and 7.2). The protection system at this location consists of earthen levee sloped
at 4 horizontal: 1 vertical with a 10foot wide crown.
This damage to this segment of the levee system was similar to that which occurred
along the Mississippi River Gulf Outlet (MRGO) levees in St. Bernard Parish: entire sections
were completely eroded leaving virtually no trace of the original earthen levee (Figures 7.1
and 7.2). Figure 7.6 shows typical erosion along the eastern end of this levee frontage; the
levee embankment is entirely removed by erosion along much of this reach.
This NOE back levee frontage is a “sister” section to the MRGO levee frontage along
the northeast edge of the St. Bernard/Lower Ninth Ward protected area that also suffered
similarly catastrophic erosion along miles of its length (see Chapter 6, Section 6.2.) These
two levee frontages share a number of unfortunate, deadly characteristics. Both sections were
constructed in large part using materials from the excavation of the adjacent shipping
channels (the MRGO and the GIWW, respectively), and as a result both were comprised
largely of unacceptably highly erodeable soils; including large quantities of sands and
lightweight shell sands. (Figure 7.3 shows the official material designations for the
constructed perimeter levees surrounding the NOE protected area. All are nominally
7 4
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
compacted fills, except for the “hydraulic fill” section along the NOE back levee.) Both levee
frontages directly fronted the swollen waters of “Lake” Borgne (which is actually a bay, being
directly connected to the open Gulf of Mexico), and so both sections experienced storm waves
driven by winds that passed across large open distances; waves that gathered significant
energy. Both sections had little or no effective protection on the outboard side from swamps
or cypress groves, or other vegetation, etc., that could reduce the intensity of these waves.
And both sections appear to have failed catastrophically, and eroded massively, producing
massive breaches along thousands of feet through which passed a majority of the floodwaters
that so catastrophically devastated the St. Bernard/Lower Ninth Ward and the NOE protected
areas.
As described previously in Chapter 6, it is the conclusion of out ILIT investigation that
the MRGO frontage levees likely failed, and suffered significant breaching, well before they
experienced significant overtopping. The discussion of potential erosion mechanisms
presented in Section 6.2 is applicable again here, and is worth revisiting on the part of the
reader.
Whereas our investigation concluded that the MRGO frontage levees were apparently
compromised before they were significantly overtopped, with the “sister” levees along the
NOE back levee frontage it can be conclusively demonstrated that massive failures occurred
prior to overtopping.
Figure 7.7 shows hydrographs of calculated (modeled, backcalculated) water levels
vs. time during and after hurricane Katrina’s passage, as calculated by IPET, for locations at
and near the NOE back levee frontage. Similar calculations by Team Louisiana give similar
results. The storm surge at the western end of Lake Borgne rose fairly slowly to Elev. +4 feet
(MSL), then as the eye of the storm approached more closely it rose rapidly and peaked at
about Elev. +16 to +18 at about 8:30 a.m.(CDT; local New Orleans time.) After peaking, the
storm surge dropped rapidly at this location. [Many of the hydrographs in this report, and
others, are based on GMT (Greenwich Mean Time), and so must be converted to CDT (local
time). Similarly, the hydrographs of Figures 7.7 and 7.9 are based on the NGVD datum, and
actual MSL elevations are approximately 1.7 feet lower. Some adjustment to elevations as
shown are being inferred herein, as the calculated elevations of Figures 7.7 and 7.9 may be a
bit low (on the order of about a foot or so) based on field observations and similar calculations
by Team Louisiana.]
Figure 7.8 shows calculated maximum storm surge (and also storm surge + wave)
elevations, again based on IPET analyses, and also levee crest heights along this frontage.
This figure shows that peak surge + waves might have overtopped this frontage at several
locations at the eastern end, and at the far west end as the GIWW and MRGO “funnel” necks
down to become the joint, eastwest trending shared GIWW/MRGO channel.
There is well established evidence, however, that significant breaching had already
occurred between about 5:00 a.m. to 6:00 a.m. Eyewitnesses, and a hand held video, clearly
show that significant floodwaters approached from the east and arrived at the Entergy power
plant located along the north side of the GIWW/MRGO waterway at 6:15 a.m., and that the
depth of water increased rapidly over the next few minutes (indicating a large source.)
7 5
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 7.9 (top) shows the location of this power plant. There are only three possible
breaches/sites that could have been the source of these welltimed floodwaters; (1)
overtopping, and two breaches, along the Citris levees (along the GIWW/MRGO channel, to
the west, (2) local overtopping adjacent to the power plant itself, and (3) the massive breaches
at the southeast corner of NOE, along the NOE back levees fronting Lake Borgne. Given the
crest heights, and water elevations vs. time, it can be established that the overtopping required
for options (1) and (2) above did not begin until well after 7:00 a.m., so the only likely source
of these floodwaters appears to be the massively eroded sections of the NOE back levee
frontage.
Floodwaters from these breaches would have been channeled by the Bayou Sauvage
ridge (high ground underlying Highway 90), and would have come west around the top of the
Michoud Canal to the Entergy power plant fairly rapidly. Allowing for the distances
involved, there must have been significant breaching and inflow by at least 6:00 a.m., and
likely earlier. Water levels along this frontage would only have been on the order of Elev. +8
to +10 feet (MSL) by 6:00 a.m., and would not have passed over (even with wave runup) the
levees along this frontage (with crest elevations of +15.5 to +19 feet, MSL.) Accordingly, it
appears that significant levee failures, and breaching, occurred prior to significant
overtopping.
Like the MRGO frontage levees discussed in Section 6.2, this catastrophic failure was
due primarily to the use of inappropriate, highly erodeable levee embankment fill materials,
including sands and lightweight shellsands. As discussed in Section 6.2, the actual
mechanisms of erosion that led to this failure are likely to have included wave scour on the
outboard sides, wave runup and resulting notching and crenellation of the levee crests,
exploitation of this by splashover overtopping, and throughflow erosion (which would have,
initially, been most pronounced low on the back or protected side of the levees.) These
mechanisms, working alone or in combination, appear to have compromised the earthen
levees well before the storm surge peaked, and therefore, well before the levees were
overtopped in the conventional sense of the word.
Damage to the NOE back levee reach also occurred further west, between the interior
secondary levee and the Michoud Canal. A sheetpile levee “transition” section located near
Pump Station 15 deflected and tilted inward (i.e., toward the protected side, see Figure 7.10),
as the result of overtoppinginduced erosion at the base of the backside of the sheetpile wall.
Sheet piling was used at these locations to transition between concrete floodwall and full
height earthen levee sections. The tops of the damage sheet pile wall had preKatrina
elevations that were less then the immediately adjacent concrete floodwall sections, and hence
scour at this location was worsened by preferential overtopping during the peak of the storm
surge. Further to the west near the Air Products Corporation site, a similar sheet pile
transition section overturned and collapsed in response to scour and the associated loss of
passive resistance on the protected side (Figures 7.12 and 7.13). Once again, the top of the
damaged section was at a lower elevation then adjacent levee segments resulting in highly
concentrated overflow (and resulting scour, that laterally unbraced the sheetpile wall) at this
location. Note that there is little or no evidence of overtopping erosion adjacent to the failed
sheetpile transition section. This is one of numerous cases wherein the adjacent long reaches
7 6
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
of fullheight earthen levee and concrete floodwalltopped levee both performed well, but
where inadequate attention was paid to effecting a safe “transition” between these two major
project elements; a tragic failure of attention to detail, and an adverse product of the
piecemeal process by which these massive and complex levee systems are constructed in
individual segments and stages.
7.3.3.2 The Michoud Area and the Citrus Back Levee
The Michoud area levee systems site extends along the GIWW from Michoud Slip to
(and around) the Michoud Canal. The site is located below and immediately west of the
Interstate 510/Highway 47 bridge near the Entergy New Orleans Corporation's power plant.
Scour was noted at the base of the rear side of the concrete floodwalls surrounding both
Michoud Slip and Michoud Canal; however, breaching did not occur at this location and
overall system performance was good (Figure. 7.11). In addition to the video of early
morning flooding here highlighted earlier, mounted security cameras later captured dramatic
images of levee overtopping during the peak of the storm surge (see Figures 7.17 and 7.18.)
West of the Michoud sector, the remainder of the levee reaches along the north bank
of the GIWW/MRGO channel constitute the main Citrus back levee section. As the risen
waters of Lake Borgne were pushed west along the shared GIWW/MRGO channel,
overtopping occurred along considerable lengths of the Citrus back levee frontage. Many
earthen embankment sections sustained this overtopping with little or no damage, while
adjacent sections suffered variable amounts of overtoppinginduced erosion on their back
(inboard side) slopes, but without full breaching.
A major failure did occur along this frontage, at the Citrus back levee floodwall. This
site is located in the industrial corridor south of Chef Menteur Highway along the GIWW.
Because its protection system consists of a relatively short floodwall segment situated
between longer stretches of fullheight earthen levee, the site provides a unique opportunity to
compare the performance of different types of levees subjected to identical storm surge
loadings. The levee system at the site principally consists of an approximately 3000 foot long
Iwall with a short (~ 80 feet) Twall section, and a 50foot long Tsection with a steel gate.
The adjacent earthen levee sections are sloped at 4 horizontal: 1 vertical and include a 10 foot
wide crown. The Iwall tilted and deflected significantly in response to the rising storm surge.
Deflection along the 3000foot length of the concrete Iwall section from severe (i.e., almost
completely tilted over, Figure 7.15) to moderate (i.e., lateral movement of several feet, with
limited tilting, Figure 7.14). Deflections were generally greater near the eastern and middle
segments of the floodwall.
Scour trenches developed along the full length of the floodwall on the protected side,
as overtopping cascaded over the tops of the floodwalls. In many instances, these trenches
were located several feet from the base of the wall (indicating progressive tilting of the
floodwalls, and thus the waters falling farther to the inboard side) and some had widths of 7
feet or more. A massive scour hole was found behind to the most tilted segment of the Iwall
system. Localized scour was also noted at the western edge of the Iwall where it connects to
the earthen levee, representing yet another example of an inadequate “transition” detail
connecting two disparate sections. These scourinduced trenched reduced the lateral support
7 7
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
for the sheetpiles and the concrete floodwall they supported, and the lateral forces of the
outboard side storm surge pushed the laterally unbraced floodwalls sideways. Figure 7.14
shows the eroded trench at the inboard side of a floodwall section that experienced only
limited movement; note the heave of soils immediately at the toe of the sheetpiles/floodwall.
Figure 7.15 shows a view of the outboard side of a floodwall section that was nearly
completely overturned. In this figure, the “gap” between the sheetpiles and the nondisplaced
outboard side levee embankment toe can be clearly seen. As discussed in numerous other
sections of this report, the formation of this waterfilled gap served to increase the lateral
forces acting against the outboard side of the sheetpile/floodwall.
Postevent topographic maps of the area show a localized low area close to the large
scour hole. The tilting of the wall effectively reduced its top elevation, which is likely to have
attracted additional overtopping at this location, causing localized erosion that ultimately
developed into the large scour hole. This may have, in turn, further exacerbated tilting of the
floodwall due to loss of passive soil resistance. It is worth noting that damage to the levee
system at this location was almost entirely limited to the relatively short floodwall segment.
The adjacent earthen levee segments performed well despite having been subjected to an
identical storm surge loading.
As noted above, the floodwall protection system included two isolated segments
which were Twall segments, both of which performed well (i.e., little if any permanent
deflection) despite the scour that occurred along their bases. This suggests that the increases
lateral and rotational stability and stiffness provided by the battered structural piles supporting
these Twall sections were very useful at this location.
The earthen levee sections east and west of the floodwalls also performed well (i.e., no
breaching or significant distress), though at some sections, particularly to the east of the
floodwalls, isolated scour holes developed along the levee slopes on the protected side. One
of the worst of these is shown in Figure 7.16. The soil exposed in these scours indicated the
levees were comprised of largely cohesive materials, and this likely explains their favorable
performance with regard to successfully resisting erosion and full breaching (failure) during
sustained overtopping.
Figure 7.17 shows a still image from a security videotape showing significant
overtopping of the earthen levee adjacent to the Entergy power plant, immediately east of the
highway bridge to the St. Bernard parish. Figure 7.18 shows the same site after the hurricane
had passed. The overtopping had produces moderate damage, but again no beaching of the
levee crest and no failure at this location. Erosionrelated performance was generally more
favorable than these two examples along the earthen levees that comprised most of the Citrus
levee frontage, and many sections showed no indication of overtopping erosion whatsoever.
7.3.3.3 The IHNC Frontage (IHNC East Levee)
The levee system located along the IHNC is primarily comprised of conventional
floodwalltopped levee sections interspersed with a number of gate and transitions structures.
Overtopping occurred along almost all of this levee frontage. Overall performance was good
along most of this frontage, with only one major breach at the extreme north end of this reach.
7 8
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
There were also, however, numerous partially evolved erosional problems at “transitions”
along this frontage, and some of these might have been more serious if the inboard side had
not already been filling with water from breaches at other locations.
Figure 7.19 shows a typical example of overtoppinginduced scour behing a concrete
floodwall along this frontage. This was common along this frontage, but no full failures
resulted. It is not possible to know with certainty to what extent this type of erosional damage
was limited by the fact that waters were likely already accumulating at the inboard sides of
these floodwalls due to overtopping and breaches at other locations.
Figures 7.20 through 7.22 show several examples of the 8 locations along this frontage
where erosion occurred, but did not develop fully to the point of “failure”, at transitions
between adjoining flood system elements. Transitions between full height earthen levees and
adjacent, composite levee/floodwall sections, and transitions between levees and concrete
gate structures (with rolling steel floodgates), were routinely problematic in this regard, and it
was common to find partially developed erosion problems at both ends of most gate structures
along this frontage. Inadequate attention to transition details, especially to lateral embedment
of transitions, and differences in top elevations of adjoining elements, were common. Also
disconcerting were sites where the eroded materials appeared to be comprised, at least in part,
of lightweight shellsands; materials notorious for lack of erosion resistance that have no
place in these levees protecting large populations.
At all locations, these “transition” erosional features were partially developed, and so
no full failures developed. This initially puzzled our field teams, until we learned that
floodwaters had been already rising on the inboard (protected) side of levees and floodwalls
while the overtopping was occurring; effectively reducing the gradient across these erosional
features and minimizing the progression of the erosion. These are features that warrant
significant additional attention during reconstruction, as these features might otherwise prove
far more dangerous in future events if the inboard side is not already flooding.
At the north end of the IHNC frontage, at the corner where it joins the Lakefront
levees, a full breach did occur. This was a complex “transition” section where three utilities
consisting of (1) a major highway (the I10), (2) an adjacent active railroad line, and (3) a
surface roadway between these two, all cross the federal perimeter levees. This transition is
rendered even more complex by the fact that it is the “corner” of the NOE protected area.
Figure 7.23 shows this location in plan view. Significant overtopping occurred along
a nearly milelong section of the Lakefront levee that had an unexpectedly low floodwall crst
height, and this flow passed through the gravel ballast of the railroad embankment (a local
low spot, as it was pervious) and eroded the adjacent earthen perimeter levee. This flow also
eroded the transition between a concrete floodwall and the adjoining earthen levee section
beneath the elevated highway, as shown in Figure 7.24.
7.3.3.4 The New Orleans Lakefront and Citrus Lakefront Levee Frontages
The lakefront levee systems include both earthen levees and composite
levee/floodwall sections. With one exception, these performed well. This exception was a
7 9
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
nearly milelong section of floodwall at the west end, behind the Old Lakefront Airport. This
section had a unexpectedly low floodwall crest elevation, and it experienced significant
localized overtopping, and resultant scour at the inboard side toe of the concrete floodwall, as
shown in Figure 7.26. This overtoppinginduced scour did not produce a failure, however, so
the overtopping flow simply added to the misery of an area that was already flooding as a
result of numerous failures that had already occurred to the south.
Only modest damage, primarily in the form of scour, occurred along the remainder of
the Lake Pontchartrain frontage. The levee system along this reach was comprised of both
floodwall and conventional earthen sections. Storm surge simulations indicate that the lake
levels were close to but not greater then the top of the levees, and therefore the scour most
likely resulting from wave splash over rather than sustained sheetflow overtopping. Figure
7.27 shows one of the few locations where minor repairs had to be made for erosion. Figure
7.28 shows a second location where limited overtopping produced minor erosional damage.
Overall, the performance of levees along the Lakefront, east of the Old Lakefront Airport, was
very good.
7.3.3.5 The New Orleans East Levee Frontage
Similar performance was also noted along the eastern levee frontage, which is
buffered from the nearby lake systems by a large stretch of wetlands to the east. Figure 7.29
shows a postevent view of a typical levee segment along this frontage. No damage at all was
noted along most of this frontage, and only limited erosion at a few locations. This was
despite evidence suggesting that overtopping had occurred along at least some portions of this
frontage. This favorable performance was likely due to: (1) the use of compacted, clayey fill
for the levee embankments (materials with a high resistance to erosion), and (2) the presence
of significant widths of swamps and cypress and other vegetation on the outboard sides of the
levee (which served to buffer the wave action.)
The only notable damage that occurred in this area was scour in a floodwallearthen
levee transition section that was part of a railroad gate structure. This produced a minor
“breach”, but given the massive flows that were admitted through the catastrophically eroded
lengths of the New Orleans East back levee immediately to the south, this was a relatively
unimportant feature in this event. It does, however, provide yet another example of problems
with handling of “transitions”, and the site should be reasessed and mitigated as it might
represent a more serious potential vulnerable point in future events if the inboard side lands
are not already rapidly filling with floodwaters.
7.4
Summary of Findings for the New Orleans East Protected Area
The key findings of this chapter may be summarized as follows:
ク
The catastrophic breaching of the New Orleans East Bask Levee System in the
southeast corner of the polder was responsible for much of the flooding of the New
Orleans East protected area. While there is limited data as to the exact time that the
breach developed, the available evidence strongly suggests this occurred well in
advance of the peak of the storm surge. This implies that the levee at this location
7 10
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
failed not in response to simple overtopping, but rather as a result of wave action
and/or throughseepage erosion, and this levee frontage appears to have been
significantly compromised, related to the rising water levels in the GIWW. The use of
fill materials known to be highly erodeable, from the excavation of the adjacent
GIWW shipping channel, resulted in shortterm cost savings that are, in hindsight,
difficult to justify against the massive damages and the loss of life engendered by the
catastrophic erosion and failure of these levees.
ク
ク
ク
7.5
With the notable exception of the levee system in the southeast corner, the
conventional fullheight earthen levees that protect most of the New Orleans East
protected area performed quite well. This is despite, in some cases, significant
overtopping that occurred during the peak of the storm surge.
The performance of concrete floodwalls was uneven. In some cases these systems
performed well even when overtopped (e.g. along the IHNC frontage). In other
situations (e.g. collapsed Citrus Back Levee Floodwall) the performance was
unsatisfactory.
Levee transition sections and gate structures were routinely problematic. Common
problems, often because of the differences in elevation between adjacent sections,
which resulted in concentrated or preferential overtopping. In many instances,
damage also occurred at these locations because of the contrast in erosion resistance
between adjoining sections (e.g. flood wallearthen levee transitions).
References
Interagency Performance Evaluation Task Force, (2006), “Performance Evaluation, Status
and Interim Results, Report 2 of a Series, Performance Evaluation of the New Orleans
and Southeast Louisiana Hurricane Protection System,” March 10, 2006.
IPET, (2006), “Performance Evaluation of the New Orleans and Southwest Louisiana
Hurricane Protection System, Draft Final Report of the Interagency Performance
Evaluation Task Force, Volume VI – The Performance – Interior Drainage and Pumping,”
available online: https://ipet.wes.army.mil/, date accessed: June 1, 2006.
7 11
Independent Levee
Investigation Team
7 12
Figure 7.1: Map showing principal features of the main flood protection rings or “protected areas” in the New Orleans area.
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Source: Modified after USACE, 2005
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 7.2: Damage locations in the NOE protected area (base map from USACE.) Color
indicates severity of damage, with red being the worst. [IPET; March 10, 2006]
Figure 7.3: Construction materials and methods, New Orleans East. [IPET; June 1, 2006]
7 13
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Source: LSU Hurricane Center, 2006
Figure 7.4: Depth of flooding of New Orleans East on September 2nd (4 days after Hurricane Katrina)
7 14
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
New
Orleans
East
Levee
Lakefront
Levees
Michoud
Canal
IHNC
Levees
New Orleans East
Back Levee
Citrus Back
Levee
Figure 7.5: Principal sections of the New Orleans East perimeter defense levees; including
the Lakefront Levees, the New Orleans East Levee, the New Orleans East Back
Levee, the Michoud Canal, the Citrus Back Levee, and the IHNC Levees.
7 15
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photo courtesy of USACE
Figure 7.6: Some of the most severe damage to the New Orleans regional levee system
occurred along this section of the New Orleans East Back levee, which is situated in
the southeast corner of the protected area, facing south toward Lake Borgne.
7 16
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 7.7: Approximate hydrograph of storm surge elevation (feet, MSL) vs. time at the
west end of Lake Borgne.
[IPET Interim Report; April, 2006]
Figure 7.8: PreKatrina crest elevations, and various estimates of storm surge + wave height;
New Orleans East back levee facing Lake Borgne
[IPET; June 1, 2006]
7 17
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 7.9: Timing of observed flooding at Entergy Powerplant and storm surge at New
Orleans East back levee breach (southeast corner fronting Lake Borgne.)
[Times shown are UTC or Greenwich Mean Time. Elevations shown are in feet,
NGVD29.]
7 18
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph by J. Wartman
Figure 7.10: Deflected and tilted sheet pile sections near Pump Station 15.
Photograph by J. Wartman
Figure 7.11: Scour at the base of floodwalls near the Michoud Canal.
7 19
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
ef
Photo by Dr. Les Harder
Figure 7.12: Failed sheetpile transition at the Air Products Corporation site; NOE back levee.
Photo courtesy of USACE
Figure 7.13: Second view of failed sheetpile transition.
7 20
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph by J. Wartman
Figure 7.14: Significant lateral deflection of the Citrus Back Levee floodwall, seen from the
inboard (protected) side. Note the heave adjacent to the displaced sheetpiles and
wall.
Photograph by J. Wartman
Figure 7.15: Deflection and tilting of another section of the Citrus Back Levee Floodwall,
this time viewed from the outboard side. Note the gap between the outboard levee
toe section and the sheetpile curtain.
7 21
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph by J. Wartman
Figure 7.16: Scour varied greatly along the Citrus Back Levee. It was significant on the back
(inboard side) slope of the levee at this location; nearly breaching the levee crest.
7 22
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Sill photo from security video at Entergy Powerplant
Figure 7.17: Still image from security videotape taken at Entergy power plant showing
overtopping adjacent to the I510/Hwy 47 Bridge on the NOE Back Levee.
Photograph by Rune Storesund
Figure 7.18: PostKatrina photo of the same levee section shown above in Figure 7.17.
7 23
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph by J. Wartman
Figure 7.19: Minor scour along the base of the IHNC floodwall. Note the boat pushed
against the outboard (flood) side of the wall.
Photograph by Rune Storesund
Figure 7.20: One of numerous examples of partially exploited erosive vulnerability at a
“transition” section along the IHNC levee frontage; in this case a transition from a
gated concrete floodwall to a full height earthen levee section.
7 24
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph by Francisco SilvaTulla
Figure 7.21: Another example of partially exploited erosive vulnerability at a “transition”
section along the IHNC levee frontage; in this case a transition from a roadway
floodgate to a full height earthen levee section.
Photograph by Rune Storesund
Figure 7.22: Erosion at the east bank IHNC CSX Rail Crossing.
7 25
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Area of Detail
Lakeview
airport
Location of
Figure 7.24
Active
railroad line
Direction of storm
surge overtopping
flow through
“transition area”
Scour of earthen
levee shown in
Figure 7.25
Figure 7.23: Stormsurge induced overtopping traveled through the granular gravel ballast for the
railroad line and eroded the railroad line embankment, which served as a transition
levee between the concrete floodwall and the earthen levee shown in Figure 7.25.
[Base image from Google Earth, 2006]
7 26
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph by Rune Storesund
Figure 7.24: IHNC levee near the Lakefront Airport adjacent to the railroad section from Figure
7.23, showing erosional failure and scour at transition to concrete floodwall
protecting highway support.
Scour of
earthen levee
Location of
Figure 7.24
Direction of Storm Surge
Overtopping Flow in Fig. 7.23
Photograph by Rune Storesund
Figure 7.25: Significant erosion was observed on the levee adjacent to (and behind) the
floodwall shown in Figure 7.24. The storm surge overtopped the floodwall and
railroad ballast and failed the earthen levee behind the railroad.
7 27
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph by J. Wartman
Figure 7.26: Scour near the base of a floodwall near the Lakefront Airport.
7 28
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph by Rune Storesund
Figure 7.27: Lakefront levee near the Jahncke Pump Station outfall structure, where minor
overtopping erosion occurred. These levees performed well and only minor,
surficial damage was observed.
Photograph by Rune Storesund
Figure 7.28: Observed scour at the Jahncke Pump Station outfall structure, Lakefront. Scour
was limited to areas of soilstructure interfaces, and no full breach occurred.
7 29
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph by Rune Storesund
Figure 7.29: Condition of levees east of HWY 11 (location 3 on Figure 10.6) in October 2005.
These levees performed exceptionally well and were not eroded during Hurricanes
Katrina or Rita.
7 30
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
CHAPTER EIGHT: THE ORLEANS EAST BANK (DOWNTOWN) AND
CANAL DISTRICT PROTECTED AREA
8.1 Overview
The most populous of the four major protected areas that suffered significant flooding
during Hurricane Katrina was the Orleans East Bank (downtown) protected area. As shown
in Figures 2.4, 8.1 and 8.2, the Orleans East Bank (downtown) section is one contiguously
protected section. This protected unit contains the downtown district, the French Quarter, the
Garden District, and the “Canal” District. The northern edge of this protected area is fronted
by Lake Pontchartrain on the north, and the Mississippi River passes along its southern edge.
The Inner Harbor Navigation Canal (also locally known as “the Industrial Canal”) passes
along the east flank of this protected section, separating the Orleans East Bank protected
section from New Orleans East (to the northeast) and from the Lower Ninth Ward and St.
Bernard Parish (directly to the east.) Three large drainage canals extend into the Orleans East
Bank protected section from Lake Pontchartrain to the north, for the purpose of conveying
water pumped north into the lake by large pump stations within the city. These canals, from
west to east, are the 17th Street Canal, the Orleans Canal, and the London Avenue Canal.
Figure 8.2 shows how this single, contiguously protected unit can be subdivided into
several localized subbasins separated by a series of ridges, levees and canals. The base map
of Figure 8.2 is the flooding map of Figure 2.17 (repeated here) which shows the flooding on
September 2, four days after the passage of Katrina. The elevation of the top of the
floodwaters in this figure is Elev. +3 feet (NAVD 88). This is approximately the peak
flooding, and the depths of flooding shown at this point in time reflect the underlying basin
topography and thus serve well to illustrate how the overall protected zone can be
approximately subdivided into four separate zones or subbasins.
The original city of New Orleans had been founded on the high ground adjacent to the
Mississippi River (along the southern edge of this protected area.) The river “climbs” within
its own channel, periodically depositing overbank sediment deposits which form “natural
levees” to constrain its path, until it rises above the surrounding countryside. Then,
periodically, the river breaks through its own “natural levees” and takes a new path to the
Gulf. The riverbank deposits thus represent the highest ground locally, and it was here that
the city began.
As shown in Figure 8.2, this high ground adjacent to the river now comprises much of
the expensive Garden District, and much of downtown New Orleans and the historic French
Quarter as well. Due to their elevation (typically Elev. +2 feet above Mean Sea Level and
higher) these areas remained largely unflooded. Most of the remainder of this large and
densely populated protected zone lies at lower elevations, however, and so most of the rest of
this zone was flooded.
As also shown in Figure 8.2, a ridge of high ground known as Metairie Ridge
separates the lowlying northern (Canal District) portion of this protected area from the
81
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
southern half. Metairie Ridge is the result of a previous river “stand”, and resulting river
deposits. The Metairie Ridge did not quite successfully separate the northern and southern
halves of this protected section during Katrina; flow passed over the ridge at a number of
locations carrying floodwaters from the catastrophic northern drainage canal breaches into
much of the rest of the protected area to the south. This flow over (across) the Metairie Ridge
was noted by eyewitnesses (Van Heerden, 2006), and is also confirmed by calculations of
flows through the various breaches.
As described previously in Chapter 2, the initial breaches in this protected area
occurred along the eastern flank (on the west bank of the IHNC). Several breaches occurred
along this frontage. These breaches allowed significant amounts of water to flow into the
adjacent neighborhoods, but these breaches were noncatastrophic; they breaches did not
scour to a depth below mean sea level, so that as the storm surge subsequently subsided the
flow inwards through these breaches was eventually halted as the IHNC water levels fell
below the (mean sea level plus) “lips” of these breaches. Our current estimates, based on
simplistic calculations of flow and surge heights vs. time, suggest that approximately 10% to
20% of the eventual flow into the overall Orleans East Bank (Downtown) protected area came
through these breaches. Similar calculations, in a bit more detail, by Team Louisiana suggest
that approximately 12 to 15% of the overall floodwaters eventually filling the Orleans East
Bank protected area came through the breaches on this east bank of the IHNC (Mashriqui,
2006).
The vast majority of the flow into the Orleans East Bank came through the three
subsequent, catastrophic breaches in the drainage canals at the northern edge of the Orleans
East Bank protected area. As shown in Figures 8.1 and 8.2, one catastrophic breach occurred
on the 17th Street drainage canal and two catastrophic breaches occurred on the London
Avenue drainage canal. These all eroded (scoured) to depths well below mean sea level, and
so continued to admit flow into the city from Lake Pontchartrain well after the initial storm
surge had subsided. The drainage canal located between these two (the Orleans Canal) did
not suffer any breaches, but the southern end of this canal was unfinished and a “gap” (low
area) in the floodwall at the southern end of this canal allowed water to flow freely into New
Orleans for a number of hours during the peak of the storm surge.
It was, however, mainly the flow through the three catastrophic breaches in the 17th
Street and London Avenue drainage canals that accounted for approximately 85% of the
flooding that slowly filled this Orleans East Bank protected area during and after the storm.
Flow from the canals overfilled the northern basin and eventually also flowed over the
Metairie Ridge and into the other zones shown as flooded in Figures 8.1 and 8.2. This
flooding continued to progress after the initial storm surge had subsided, and flooding in the
southern portions of this protected zone continued to worsen overnight and into the three days
that followed, finally equilibrating with the slightly inflated water levels in Lake Pontchartrain
on Thursday (September 1.)
As discussed in Chapter 2, this flooding had catastrophic consequences, accounting for
approximately half of the total loss of life in this event, and a similar share of the economic
damages as well. The performance of the flood protection system in this Orleans East Bank
82
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
(downtown) basin is thus of great importance, and was studied in some detail by this
investigation.
8.2 Performance of the Flood Protection System Along the
West Bank of the Inner Harbor Navigation Channel (IHNC)
8.2.1 An Early Breach at About 4:45 a.m.
As described previously in Chapter 2, the first levee breach and failure in the
metropolitan New Orleans area appears to have occurred along one of the banks of the IHNC.
Figure 8.3 shows water elevations at three gage stations as well as at a manual water
elevation station in the IHNC as the hurricane storm surge initially began to raise the water
levels throughout the IHNC region on the morning of August 29th. As the storm began to
approach the coast, water levels within the IHNC began to rise. By about 4:30 a.m. the water
level within the IHNC had risen to approximately +9 to +9.5 feet (MSL). Then, at
approximately 4:45 a.m., two of the gauges near the Highway I10 bridge registered a sudden
change in the otherwise relatively constant rate of rise in water levels. The U.S. Geological
Survey (USGS) gage at this location shows a precipitous drop in water levels at
approximately 4:45 a.m. The Orleans Levee District gage was “sampled” less frequently, but
it also shows a reduction in rate of local water level rise between about 4:45 and about 5:00
a.m.
These gage readings appear to indicate that a levee breach occurred at about 4:45 a.m.
near the I10 Bridge across the IHNC channel, resulting in a local and temporary drawdown
of the otherwise rising water levels in the IHNC.
A number of levee breaches occurred during Hurricane Katrina along the northsouth
channel of the IHNC, so there is no shortage of candidate sites for this breach.
Many of the partial breaches and distressed levee sections of New Orleans East fronting
the IHNC (on the east bank of the IHNC) were relatively minor features, with minor flow
potential, and could not have accounted for the significant changes observed in the gage
readings shown in Figure 8.3. In addition, a number of these features showed evidence of
erosion and scour specifically due to overtopping, indicating that water elevations
significantly greater than +9 feet (MSL) eventually occurred at their locations.
Similarly, the timing(s) of the occurrences of the two large breaches on the east side of
the IHNC at the edge of the Ninth Ward are wellestablished by eye witnesses as well as by
“stopped clock” data, and these two major breaches appear to have occurred considerably
later at about 7:45 a.m.
A significant breach occurred on the west side, behind the main Port of New Orleans,
due to overtopping and erosion of soil support for an Iwall (see Section 8.2.3.1.). The
elevation of the Iwall, and the observed overtopping erosion, indicate that this failure
occurred later in the morning as well when the storm surge had risen high enough to pass
water over the top of this floodwall (see Section 8.2.3.1.).
83
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
That leaves only three candidate breach sites that might have caused the drop in water
level rise shown at about 5:00 a.m. in Figure 8.3.
One of these is the breach on the east side of the IHNC at the CSX railroad crossing and
roadway crossing over the levee, as described in Chapter 7.
A second candidate site is the pair of breaches that occurred closely adjacent to each
other at the south end of the main Port of New Orleans, as described in Section 8.2.3.3. These
were large breaches, and might well have had sufficient flow as to account for the drop in
water level rise shown in Figure 8.3. In addition, these sections were constructed of highly
erodeable lightweight shellsand fill, and might well have eroded early due to throughpassage
of seepage flows through the “earthen” levee embankment as the storm surge rose (but prior
to full overtopping of the levee embankment at this location.) This is discussed further in
Section 8.2.3.3.
A third candidate breach site is the west bank of the IHNC at the CSX railroad
crossing, as described in Section 8.2.2. At this location, a steel “storm gate” on rollers had
been damaged by a train accident several months prior to Hurricane Katrina, and was away
for repair. In lieu of this missing gate, a sandbag levee crest section had been constructed in
the opening left by the missing floodgate. The sandbags washed out at some point during
Katrina, and this may have been the early breach reflected by the gage readings shown in
Figure 8.3. At this same site, flow along the juncture between the railroad embankment and
the adjacent embankment fill supporting an asphalt paved roadway passing over the earthen
Federal levee resulted in erosion and scour that produced a second breach feature at
essentially this same site, as is also described in Section 8.2.2.
In the end, based on the information currently available to this investigation team, any
of these three candidate breach sites might have been responsible for the for the observed
gage level drops shown in Figure 8.3.
8.2.2
The CSX Railroad Breach
As shown in Figure 8.2, the CSX railroad crosses the IHNC channel immediately to
the south of the I10 Highway bridge. On both the east and west banks of the IHNC, the
railroad passes through the levee system by means of a gate through a structural concrete
floodwall. Steel gates are used to close these openings during storms.
Figure 8.4 shows the concrete structural floodwall on the west side of the IHNC, at the
east edge of the Orleans East Bank (Downtown) protected area. Note that there is no steel
gate shown in this photograph. The steel gate at this location had been damaged by a train
accident several months prior to Katrina’s arrival, and it was away for repair at the time of the
hurricane.
In lieu of this missing steel gate, a temporary sandbag “levee” was erected across the
opening. At some point during the storm this sandbag “levee” section either was pushed over
by the rising storm surge or was overtopped and washed away.
84
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
In addition, erosion occurred at the juncture between the railroad embankment fill and
the fill supporting an adjacent roadway passing over the earthen federal levee at this location,
as shown in Figure 8.5. This roadway passed over the levee crest to provide access to port
facilities on the outboard (water) side of the Federal levee system. This is shown in Figure
8.5, which is a view from the inboard side of this breach showing the erosion of the roadway
fill. The elevated I10 highway bridge is at the left of this photo, and the CSX railroad is just
to the left (north) of the roadway. The roadway fill at this location was comprised largely of
highly erodeable lightweight “shell sand” fill; a material not suitable for levee fill in a levee
protecting a large population (especially without sheetpile cutoff or similar features to prevent
erosion.) The flow appears to have passed initially through the pervious gravel ballast
supporting the train rails (which is the “low point” at this complicated location), and then
undermined the less competent fill beneath the roadway. The resulting flow through the
eroded breach then passed to the inboard (protected) side and made its way into the adjacent
neighborhood.
The erosion and scour at this conjoined pair of breach locations did not erode the base
(lips) of these breach features to a level below mean sea level. Accordingly, although flow
passed through this pair of features for a number of hours, the flow eventually ceased as the
storm surge (water level rise in the IHNC) eventually subsided.
The failure at this site is an excellent example of a failure produced by multiple
adjoining jurisdictions, and a lack of overall coordination of the various system elements
constructed and operated by each. The Federal levee system was “penetrated” here by both
the railway and the Port roadway, and the interactions of the pervious railway ballast and the
highly erodeable roadway fill combined to fail the overall flood protection system at this
location. Lack of coordination, and lack of authoritative oversight, of these disparate
organizations and their disparate system components was a critical problem here.
It should be further noted that this same site had also failed catastrophically in 1965
during hurricane Betsy, so that the refailure of this same location represents a daunting case
of lack of progress and learning over the intervening 40 years. As discussed in Chapter 7, the
east bank CSX rail crossing, which also failed during hurricane Katrina, was also a “repeat”
failure (as it, too, had failed during hurricane Betsy in 1965.) The continued failure to
recognize and suitably address the hazards associated with these complex “penetrations”,
despite their demonstrated history of previous failure, is difficult to understand.
In addition, it is interesting to note that the steel gate was allowed to be removed for
repair, rather than requiring it to be fixed in place until a suitable replacement gate (or at least
interim replacement gate) could be fabricated and be brought in, so that trains could continue
to operate. This created an obvious potential hazard to the safety of the very large community
inboard of this rail crossing; placing the safety of many at increased risk. In hindsight; that
was a decision that is difficult to justify.
8.2.3 Breaches and Distressed Sections at the Port of New Orleans
Three breaches occurred to the south of the CSX railroad breach on the west side of
the IHNC at the main Port of New Orleans. Several additional levee and floodwall sections
85
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
were “distressed” or damaged, but did not fully breach along this same section. These breach
and distress sites along this reach are jointly indicated as the suite of “Industrial Canal
Overwash Sites” in Figure 8.2.
As the storm surge raised the water levels in Lake Borgne, and then pushed the
elevated waters (and flow) westward through the “funnel” at the east end of the east/west
trending GIWW/MRGO channel between New Orleans East and St. Bernard Parish, large
flows and a major rise in water elevations pushed westward along the GIWW/MRGO channel
to this channel’s “T” intersection with the IHNC channel, and raised the water levels within
the IHNC channel.
This resulted in rising waters rushing directly at the west bank of the IHNC, coupled
with overall raising of the water levels throughout the IHNC region. This produced distress,
and several breaches, on the west side of the IHNC in the general vicinity of the main Port of
New Orleans. The subsections that follow will describe each of these in turn.
8.2.3.1 Breach at Rail Yard Behind the Port of New Orleans
The northernmost of these features was a breach in a combined earthen levee and
concrete Iwall section, as shown in Figures 8.6 and 8.7. This breach occurred behind the
main Port of New Orleans, just to the south of the juncture between the eastwest trending
GIWW/MRGO channel and the IHNC, so the water pressures and overtopping from the
lateral flow from the eastwest trending GIWW/MRGO channel were particularly severe at
this location, as indicated in hydrodynamic modeling by Team Louisiana (Mashriqui, 2006.)
At the time of our field team’s arrival in late September, this site was already under
repair. The field team arrived at this site on the morning of September 30, 2005, and at that
time the trench that had been scoured behind the wall on the north end of the breach had been
“filled” with clayey backfill and additional backfill had been placed behind the wall to form
an additional buttressing berm, as shown in Figure 8.6. A temporary access road had also
been placed through the breach, as is also shown in this photo.
Figure 8.7 shows conditions on the north side of this breach, at the same point in time
(on September 30.) The interim repair efforts had not yet reached the north side of the breach,
and the mechanisms that contributed to this failure were still clearly evident here. As shown
in Figure 8.7, significant overtopping had passed over the concrete Iwall and then cascaded
down the backside, resulting in erosion of a “trench” at the base of the backside of the Iwall.
It should be noted that water falling over an 8 foot high Iwall strikes the ground at a velocity
on the order of about 20 to 25 feet per second; sufficient to cause rapid erosion at the point of
impact.
The initial height of the compacted embankment fill on the backside of the Iwall prior
this erosion can be clearly seen in Figure 8.7 by the soil markings on the Iwall at the left of
the photograph. The depth of this erosion (scour) from the elevation of the top of the pre
event Iwall/soil crest contact to the base of the eroded trench was 4.5 feet at the location of
the photographer taking the photo of Figure 8.7, and it deepened progressively towards the
actual breach location approximately 25 feet to the North. Just before reaching the actual
86
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
displaced Iwall section shown in Figure 8.7 this depth of erosion was approximately 5.5 feet,
so that the depth of erosion at the location of the actual Iwall failure was likely on the order
of 5.5 to 6.5 feet.
The depth of the sheetpiles was unusually shallow at this location, as shown in Figure
8.8. The Iwall “stickup” had not been large at this location, and it was not felt that very long
sheetpiles were needed to support this Iwall by means of cantilever action given its relatively
short unsupported length (stickup). There was no sign of lateral embankment movement at
this site, and the sheetpiles and Iwall showed no signs of flexure on their vertical axis (along
their length from top to bottom.) The Iwall failed by rigid body “toppling” laterally towards
the inboard (protected) side in a “rigid, posthole” toppling mode as it became progressively
unbraced by the erosion of the supporting soil at the inboard toe. Eventually, it became
unable to support the water pressures on the outboard (canal) side due to the storm surge and
hydrodynamic forces, and the Iwall toppled far enough to permit catastrophic erosion at the
main breach section.
As shown in the crosssection of Figure 8.8, the sheetpiles at this section were only 14
feet in length, and were tipped at a base elevation of approximately 6 feet (MSL). As the
overtopping water cascaded over the top of the concrete Iwall, the resulting trench eroded to
a depth of approximately 6 feet below the original wall/soil crest contact, and the critical
section achieved approximately the geometry shown in Figure 8.8. Soil properties are not
well established at this location, as site specific investigation was not possible within the
budget and time constraints of this independent investigation. Accordingly, soil stratigraphy
and soil properties used in our analyses are inferred from the original design data available
from the USACE.
Based on the field observation that no major embankment foundation failure was
observed, the most significant properties for analysis of this section were the sheetpile
sections (which were PZ22) and the properties of the engineered embankment fill (which was
a moderately compacted silty clay of medium plasticity). Shear strength of the embankment
fill was modified to determine the strength (and stiffness) at which the observed failure would
be expected to occur, and it was found that the Iwall section would be marginally unstable
with a fill strength of approximately 600 to 1000 lb/ft2. This appeared to be well
representative of the strength of the observed fill, and the failure mode (shown in Figure 8.9)
matched well with the field observations. Figure 8.9 shows the results of Finite Element
Analyses (FEA) performed using the program PLAXIS; in this case for embankment shear
strength of approximately 800 lb/ft2, and with a trench to a depth of 6 feet behind the
floodwall. This Figure shows calculated displacements, and the rigid toppling mode of wall
failure can be clearly seen. Lateral failure of the Iwall results in large part from shearing at
the transition between the base of the embankment fill and the underlying foundation soils.
It appears that this failure could have been prevented, simply and at little incremental
cost, by installation of concrete “splash pads” or other erosion protection at the base of the
inboard side of the Iwall to prevent the observed erosion. Similarly, this failure would have
been prevented if the floodwall had been a “Twall” section, as illustrated in Figure 8.10(b),
rather than the less expensive “Iwall” section, as illustrated in Figure 8.10(a). The Iwall
sections are supported laterally only by the cantilever action of their supporting sheetpile
87
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
walls, and this cantilever action is adversely affected by this erosion. Twalls, on the other
hand, have lateral concrete stems at their bases and these are supported both laterally and
rotationally by battered piles (providing a much higher level of rotational resistance.)
Instead a significant breach occurred, and floodwaters passed through the adjacent
railroad yard and into the adjacent neighborhoods for a number of hours. This breach was
located well inboard from the actual edge of the IHNC channel, however, and the breach did
not erode its front lip to a depth below mean sea level. Hence, this flow eventually ceased as
the storm surge subsequently subsided later in the morning of August 29th.
8.2.3.2 Erosional Distress at Floodgate Structure Behind the Port of New Orleans
Just a few hundred yards to the south of the breach described in the previous section,
significant erosional distress occurred at a concrete Iwall and floodgate structure behind the
Port of New Orleans. As shown in Figure 8.11, this concrete wall and steel gate structure
provided access from the rail yard (on the protected side) to Port facilities (on the water side)
which can be closed off by means of a rolling steel floodgate.
Significant erosional distress occurred at each end of this floodgate structure as it
“transitioned” to join the earthen levee and floodwall at each end. An example, at the north
end of this structure, is shown in Figure 8.12. In this figure, the canal is on the left and the
“protected” side is on the right. The trenchlike feature at the outboard side toe of the
floodwall is the “gap” left when the wall displaced to the right as overtopping eroded a trench
on the right side of this wall, laterally unbracing the very short sheetpiles and wall. New fill
has been placed on the right side (as an interim repair), so this erosion is no longer visible.
The concrete wall of the gate structure has not displaced, as it is supported on a Twall basis,
and the rotational stiffness of the battered piles has been sufficient to prevent wall rotation.
The erosion at the juncture between the concrete gate structure and the adjacent
concrete Iwall was locally exacerbated by the disparity in top elevations between these two
walls; which acted to locally concentrate the overtopping flow. Erosional distress of this sort,
at the “transitions” between differing elements of the flood protection system, was a recurring
theme in the damages caused by Hurricane Katrina.
8.2.3.3 Two Adjacent Erosional Embankment Breaches at the North End of
the Port of New Orleans
Additional erosional “distress” and two large erosional breaches occurred slightly
farther to the south, at the southern end of the main Port of New Orleans.
Figures 8.13 and 8.14 show two views of a large breach through an earthen levee at
the contact (“transition”) between the levee and a concrete floodwall section. As shown in
Figure 8.14, the embankment fill material is lightweight shellsand, a material known to be
unusually highly erodeable. This type of shellsand material performed notably poorly at a
number of locations during Hurricane Katrina, and is a material not suitable for construction
of critical levees protecting large populations. At this location, no provisions (e.g. a sheetpile
cutoff, etc.) had been made to prevent catastrophic erosion of this shellsand fill due to either
88
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
overtopping or througherosion (erosion due to throughflow prior to full overtopping.) In the
absence of an eyewitness, it was not possible to discern from evidence at this site whether the
embankment was actually overtopped, or whether flow through this highly erodeable fill
caused progressive erosional failure prior to full overtopping.
Figure 8.15 shows a second large erosional breach, less than 50 yards from the breach
shown in Figures 8.13 and 8.14. This embankment section was also comprised largely of
highly erodeable shellsand fill. A large scour hole can be seen to the right, immediately
inboard of this large breach. The massive flows have rippled the asphalt tarmac, and detritus
from the eroded shell sand fill is scattered over a large area. As shown in this photo, some of
this shell sand detritus has been scooped back into the breach as part of the initial repair.
Although these two adjacent erosional breaches were both of good size, they were
both located some distance inboard from the IHNC channel, and neither eroded a pathway all
the way back to the IHNC channel that was continuously below sea level. As a result,
although both breaches admitted significant volumes of water into the adjacent
neighborhoods, flows through these two breaches eventually ceased as the storm surge
subsequently subsided.
8.2.4 Summary and Findings
The breaches along the west bank of the IHNC were each “non catastrophic” as none
of them eroded or scoured to such depth that their lip dropped below mean sea level.
Accordingly, although they admitted significant volumes of floodwaters into the greater
Orleans East Bank (downtown) protected area, these flows eventually ceased as the storm
surge subsided. Together, these features appear to have contributed approximately 10% to
20% of the overall volume of floodwaters that eventually flowed into the Orleans East Bank
(downtown) protected area.
Although they were each “noncatastrophic”, these features each had the potential to
cause significant localized flooding and damage. They were also each the result of
engineering lapses and/or lapses in oversight during design and construction; none of the
failures in this area should have occurred at the storm surge and wind/wave loadings produced
at these locations by Hurricane Katrina had proper design and construction features been
included.
The removal of the steel floodgate at the CSX Railroad crossing, and the inadequate
sandbagging of the resulting “gap” in the overall regional flood protection system should not
have been permitted. The steel gate should have been immediately replaced with a suitable
and serviceable temporary replacement until the original gate could be repaired and returned.
Instead it was missing for approximately three months of the hurricane season. In view of the
events during Katrina, it is difficult to justify the decision to remove the gate and thus
maintain the “operability” of the railroad line when it placed the “operability” of the flood
protection system, and the safety of the community, at risk.
Similarly, the confluence of the CSX railroad embankment and the adjacent roadway
both passing over/through the Federal levee system immediately to the south of the I10
89
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
bridge represents one of many “transitions” between disparate flood protection system
elements that performed poorly as an apparent result of lack of appropriate oversight and/or
poor design with regard to how abutting elements of the system joined at their edges. In
addition, highly erodeable shell sand fill was used at this roadway location without suitable
cutoff by means of sheetpiles, etc., representing a hazardous condition that should have been
caught and remedied prior to Katrina’s arrival.
The erosional “distress” that occurred at the junctures of structural Iwall sections and
the structural Twall gate structure at the rail yard behind the Port of New Orleans represent
additional examples of inadequate attention to details at “transitions” between adjacent
sections of differing type and geometry.
The two large erosional breaches at the south end of the Port of New Orleans were
clearly the result of use of inappropriate fill materials (highly erodeable lightweight shellsand
fill) in earthen embankment sections with no suitable provisions to reduce the obvious risk of
catastrophic erosion and breaching. This, too, should have been spotted and remedied prior to
Katrina’s arrival. It is not clear whether these two breach sections were overtopped by the
rising storm surge, or whether the embankment sections eroded as a result of “through flow”
prior to full overtopping as the waters rose within the IHNC.
Finally, the Iwall section breach behind the Port of New Orleans was largely the
result of overtopping and subsequent erosion at the base of the inboard toe of the concrete I
wall. This failure could have been prevented, at relatively little incremental cost, by
installation of concrete splash pads or other erosion protection at the inboard toe of this
floodwall. In addition, there was ample right of way available to construct a somewhat wider
(and heavier) levee embankment on the inboard side of this Iwall. The incremental cost of
doing so would have relatively small, and that too would likely have prevented this failure.
8.3 The Canal District Failures
8.3.1 Introduction
As the eye of the hurricane began to pass to the northeast of New Orleans, the
counterclockwise swirl of the storm winds caused a surge in water levels along the southern
end of Lake Pontchartrain. The storm surge along the Pontchartrain lakefront (which peaked
at about 9:00 to 9:30 a.m. at an elevation of about +10 feet, MSL) did not produce water
levels sufficiently high as to overtop the crests of the concrete floodwalls atop the earthen
levees lining the three drainage canals that extend from just north of downtown to Lake
Pontchartrain; the 17th Street Canal, the Orleans Canal, and the London Avenue Canal. Three
major breaches occurred along these canals, however, and these produced catastrophic
flooding of large areas within the Orleans East Bank protected area (as shown in Figure 8.2.)
The first major breach along the drainage canals occurred near the south end of the
London Avenue canal, between about 7:00 to 8:00 a.m. The second breach occurred near the
north end of the London Avenue canal, and the best current estimates of the timing of this
breach are between about 7:30 to 8:30 a.m. The third major breach occurred near the north
8 10
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
end of the 17th Street canal. The main breach here occurred between about 9:00 to 9:15 a.m.,
but this may have been preceded by earlier visually observable distress at this same location.
All three of these breaches rapidly scoured to depths well below mean sea level, so they
continued to transmit water into the main Orleans East Bank (downtown) protected area for
three days after the initial peak storm surge subsided. More detailed discussions and analyses
of these catastrophic drainage canal breaches are presented in the sections that follow.
The resulting flooding of the main Orleans East Bank (Downtown) protected area was
catastrophic, and resulted in approximately half of the 1,293 deaths attributed (to date) to the
flooding of New Orleans by this event. Contributions to this flooding came from the
overtopping and breaches along the IHNC channel at the east side of this protected area, as
described previously in Section 8.2, but the majority of the flooding (approximately 80% to
90% of it) came from the three catastrophic failures along the drainage canals at the northern
portion of this protected area.
In addition, one of the drainage canals (the Orleans Canal) had not yet been fully
“sealed” at its southern end, so that floodwaters flowed freely into New Orleans during the
peak of the storm surge through this unfinished drainage canal. A section of levee and
floodwall approximately 200 feet in length had been omitted at the southern end of this
drainage canal, so that despite the expense of constructing nearly 5 miles of levees and
floodwalls lining the rest of this canal, as the floodwaters rose along the southern edge of lake
Pontchartrain, the floodwaters did not rise fully within the Orleans Canal; instead they simply
flowed freely into downtown New Orleans.
By about 9:30 a.m. all of the levee failures had occurred, and the main Orleans East
Bank (downtown) protected area was slowly filling with water. As the northern end filled
from the three catastrophic breaches along the drainage canals, water eventually began to pass
over low spots in the Metairie Ridge and flowed into the southern zones within this protected
area as well.
The sections that follow present more detailed examinations of the performance of the
flood protection system in the “Canal District”.
8.3.2 The Lining of the Drainage Canals
There were a number of lapses, errors and poor decisions that led to the catastrophic
breaches along the drainage canals and thus the flooding of the main section of metropolitan
New Orleans. Several of these began right at the start, in the aftermath of Hurricane Betsy (of
1965) and the flooding caused in New Orleans by that event.
The decision was made, in the wake of Hurricane Betsy, to raise the level of flood
protection throughout the region. The three drainage canals (the 17th Street, Orleans and
London Avenue canals) were problematic in this regard, however, due to limited rightofway
adjacent to the existing embankments lining these canals.
As described in Chapter 3, the USACE argued (correctly as it turned out) that the low
rise levees lining the canals were not adequately stable as to sustain a significant raising, and
8 11
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
that the preferred solution would be to place storm gates at the north ends of the three canals
which could be closed in the event of a Hurricane to prevent storm surge rise within the
canals.
This proposal was bitterly contested by the local Water and Sewerage Board, who
were concerned that the gates would be under the control of the local Levee Board, and that
they might therefore be impeded in their efforts to operate the massive pumps to “unwater”
the city from heavy rainfall (which is also a source of frequent, though noncatastrophic,
flooding problems in New Orleans.)
The USACE was, in the end, not allowed to install the floodgates, which would have
been the technically superior solution, largely as a result of the internecine distrust between
the local Levee Board and the local Water and Sewerage Board. In response, the USACE
attempted to “exempt” the three canals from the otherwise contiguous levee system around
the main metropolitan Orleans East Bank (downtown) protected area.
As discussed in Chapter 3, lobbying by State and local interests next resulted in a
Senate rider (inserted clause) on a bill that unexempted the three canals and specifically
required the USACE to raise the level of flood protection along these three canals. This was
the first of a number of causative errors that would prove catastrophic here. The canals would
remain open to hurricaneinduced storm surges at the south end of Lake Pontchartrain;
essentially “allowing the enemy (storm surges) right into the backyard” of metropolitan New
Orleans.
A second problem now arose. The existing levees were low, and they were relatively
narrow as well. Homes had been constructed throughout the area, and the private property at
the inboard (protected side) toes of these existing levees left inadequate space for construction
of wider levees. Accordingly, a decision was made to raise the level of flood protection by
adding reinforced concrete floodwalls to the crests of the existing earthen embankments.
This, in effect, represented a decision to work within the narrow space available rather
than purchasing additional property to allow construction of wider, and more stable, levee
sections. That was a second issue that contributed significantly to the catastrophic failures
that occurred along the drainage canals.
It also resulted in difficulties with regard to both maintenance and inspection, as
private homes at the toes of the levees often had property lines interfering with inspection of
conditions at the inboard (protected side) toes. In some locations, private property (mainly
people’s back yards) extended up the inboard slope faces of the levee embankments, and trees
grown on these faces and at the inboard toes of the levees represented an obvious hazard both
with regard to seepage erosion and also with regard to the possibility that trees would blow
over (in water softened ground) during hurricanes. This would leave large voids (the sizes of
their root balls) at a very dangerous location (right at the inboard toes of the levees) at a time
when storm surges in the canals were simultaneously rendering seepage erosion, and inboard
slope stability, very tenuous. During Hurricane Katrina a number of large trees did indeed
topple, leaving dangerous voids at the toes and on the inboard slope faces of the levees along
these canals.
8 12
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
In addition, along some sections private homeowners excavated and constructed in
ground swimming pools in close proximity to the inboard toes of these levees, effectively
partially undermining them and rendering them less stable. This, too, should have been
prevented.
The abutting private properties also led to inspection difficulties, as inspection of
conditions immediately inboard of the levee toes is of great importance and private property
rights largely prevented inspectors from walking these critical areas. Reports of seepage and
wetness at some locations were made to the local Water and Sewerage Board (who were
responsible for “unwatering”, and were thus the group to whom such reports were made), but
this investigation team has not been able to determine whether these were then passed along
to the local Levee Board or to the USACE, to whom they might have represented
unanticipated seepage problems warranting further investigation. Certainly the USACE has
stated that they were unaware of such reports.
Lack of appropriate control of conditions at the inboard levees toes, and lack of
suitable access for inspection and maintenance at the inboard toes, represented additional
inadvisable sources of increased hazard.
8.3.3 The E99 Sheetpile Wall Test Section:
In order to effect the raising of the flood protection levels within the narrow rightofway available, the decision was made to erect concrete floodwalls at the crests of the existing
earthen levee embankments. To facilitate the analysis and design of these challenging
sections (on narrowly confined rightsofway, and on very difficult foundation soil
conditions) the New Orleans District of the USACE made an admirable decision to construct
a test section and perform a fullscale test of this type of design.
Very similar (difficult, swampy, riverine delta) soil conditions exist nearby in the
Atchafalaya river basin (approximately 80 miles to the west), and a site in this area was
selected. A sheetpile “Iwall” and contiguous sheetpile curtain, was constructed on the
inboard side stability berm of a federal levee in the Atchafalaya basin in a configuration that
was very similar to the eventual installation of similar sheetpilesupported concrete floodwalls
at the crests of the lowrise levees along the drainage canals in New Orleans (Foott and Ladd,
1977). The swampy foundations soils at this test site were remarkably similar to those at the
north end of the 17th Street canal in New Orleans. A sheetpile cofferdam was constructed
adjacent to the fullscale test section, and was filled with water to load the test section’s Iwall
and its supporting sheetpile curtain.
Two important lessons were learned from this test, and from subsequent analyses (e.g.:
Jackson, 1988; Foott and Ladd, 1977; Oner, Dawkins and Mosher, 1997; Oner, Dawkins,
Mosher and Hallal, 1997). One was that a gap opened between the sheetpile curtain and the
outboard side earthen embankment during loading (by raising of the outboard side water
level), and then water penetrated into this gap. This effectively cut the supporting
embankment in half, and the water pressures applied against the lower sheetpile sections
helped to push the inboard half of the embankment, as well as the Iwall and its supporting
sheetpile curtain, towards the inboard (protected) side. This was a failure mechanism that had
8 13
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
not traditionally been considered in the local design of floodwall systems in the New Orleans
District. The other lesson was the shape of the failure surface, which was more curved than
the deeply plunging threewedge “planar” failure surfaces considered in the “method of
planes” used for analysis of these types of sections in the New Orleans District of the
USACE.
Unfortunately, despite publication of these important findings in both internal USACE
reports as well as in electronic professional journals (e.g.: Oner, Dawkins and Mosher, 1997;
Oner, Dawkins, Mosher and Hallal, 1997), and despite the fact that these studies had been
undertaken to facilitate the design of the challenging floodwalls along the drainage canals and
the IHNC, neither of these lessons were then incorporated in the subsequent design of the
floodwalls along the 17th Street, Orleans and London Avenue drainage canals, nor along the
IHNC.
8.3.4 Field Tests for Assessment of Underseepage Risk at the Canals
The USACE also commissioned a pair of local permeability tests at two selected
sections along the drainage canals to assess the rate at which changes in water levels within
the canals were transmitted through the soils beneath the embankments. The intent here was
to assess whether or not it would be necessary to drive the sheetpile curtains deep enough to
“cut off” such underseepage flows for the transient loading conditions represented by a storm
surge that would raise and then lower the canal water levels within a matter of hours.
The two sections selected were instrumented with piezometers at a series of stations
orthogonal to the canals so that the water levels (phreatic surface) could be observed. The
canal sections were then excavated to increase the crosssection available for pumping flows.
It was assumed that this excavation and deepening would remove the sediment that “sealed”
the canals, and would result in an increase in the observed phreatic surface. If little rapid rise
was observed, then that would indicate that the increased hydraulic pressures of a transient
storm surge would not propagate rapidly under the levees.
There were two critical flaws to this reasoning. One was the assumption that two such
tests could suitably characterize the highly variable soil conditions along many miles of the
three drainage canals (and also the IHNC). The other was that this testing program failed to
note the alternate possibility that the canals were not well “sealed” at all; in which case simply
excavating the canals to greater depth would result in no net change in the observed phreatic
surface in the piezometers installed inboard at the test sections (the canal water levels would
be unchanged by the excavation of the canal bases, and if “steady state” seepage conditions
were already established based on full connectivity between the canals and the inboard toe
areas then no net change in phreatic surfaces would be observed.)
When the canals were excavated, no significant change in inboard water levels was
noted, and it was concluded that underseepage would not pose a significant risk for a short
lived (transient) storm surge. That would prove to be a very serious error, and would result in
sheetpiles throughout the system (the three drainage canals and the IHNC as well) routinely
being far too short to adequately cut off underseepage. Several major failures along the
8 14
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
drainage canals and the IHNC would result from underseepage during Hurricane Katrina, and
the short sheetpiles continue to pose a risk to the remaining sections today.
8.3.5 Water Levels Within the Canals During Hurricane Katrina
Figure 8.16 shows the calculated peak storm surge heights along the southern shore of
Lake Pontchartrain based on the most recently available IPET analyses (IPET Report No. 2:
April, 2006). These are in close agreement with similar analyses by Team Louisiana along
the canal frontage (Kemp and Mashriqui, 2006). As shown in this figure, the storm surge was
estimated to be a bit higher at the west end of the “Canal District” than at the east. The water
elevations shown in Figure 8.16 are based on the NGVD 29 datum, and must be reduced by
about 1 foot to be compatible with the approximate local Mean Sea Level datum used in this
report. With this adjustment, the projected peak water levels at the northern ends of the
drainage canals are on the order of +10 to +11 feet (MSL) based on these hydrodynamic
analyses.
Figure 8.17 shows locations at which relatively reliable high water marks near the
mouth of the 17th Street Canal [IPET Report No. 2, 2006]. These high water locations were
selected so as to be affected as little as possible by wave action, so that the water levels
recorded would be the mean surge height (without wave action.) Based on these data, the
IPET study concluded that the maximum storm surge rise at the mouth of the 17th Street Canal
was on the order of +11 feet (NAVD 882004.66 datum, which is approximately MSL).
Figure 8.18 shows a hydrograph of estimated water elevations vs. time within the 17th
Street Canal, based on the hydrodynamic calculations performed by IPET and on observations
of water levels at nearby sites [IPET Report No. 2; April, 2006]. This hydrograph peaks at an
assumed height of approximately +11 feet, and it peaks fairly sharply between about 9:00 to
10:00 a.m.
Based on the watermark data, our own field observations, and observations and data
provided by Team Louisiana (Kemp, Mashriqui and Van Heerden, 2006), our team feel that
these are realistic estimates of the surge heights near the mouths of the three key drainage
canals (the 17th Street, Orleans, and London Avenue canals), but that they likely slightly
overestimate the water levels. Our team has assumed a peak surge height of approximately
+10 to +10.5 feet (MSL) at the mouth of the 17th Street Canal, and slightly lesser heights of
on the order of +9.5 to +10 feet (MSL) at the mouths of the Orleans and London Avenue
canals.
Accordingly, the hydrograph of Figure 8.18, but with a slight reduction of peak surge
height (to approximately +10 to +10.5 feet, MSL in the 17th Street Canal, and +9.5 to +10
feet, MSL in the Orleans and London Avenue Canals) will be used for these current studies.
8 15
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
8.3.6 The Orleans Canal
As described previously in Chapter 3, the U.S. Army Corps of Engineers had lobbied
and fought for many years to install floodgates to close off the three drainage canals (the 17th
Street, Orleans, and London Avenue canals) during hurricanes so that storm surges would not
push their way up into these canals. That would have been a superior technical solution, but it
was not allowed as there was internecine fighting between the local Levee Board (who are in
charge of “protection”; including levees, walls and floodgates) and the local Water and
Sewerage Board (who are in charge of “unwatering” by means of pumping for both rainfall
and other flooding.) The Water and Sewerage Board were concerned that the floodgates
would not be under their control, and so their ability to pump out rainwater from rainstorms
(also a cause of flooding in New Orleans) might be obstructed.
As a result of the two disparate local Boards being unable to resolve their differences
in the interest of the greater Public good (and safety), the sides of all three drainage canals
were instead lined with floodwalls topping the earthen levees along both sides. This, in
effect, opened many additional miles of narrow levees and floodwalls atop difficult (and often
marshy) foundation soil conditions to storm surges; greatly increasing vulnerability by
“allowing the enemy right into the backyard” of this protected area.
An extreme example of the dangers resulting from the poor interaction between the
local Water and Sewerage Board and the local Levee Board occurred at the south end of the
Orleans Canal.
At the south end of this canal, the main pumping plant crosses the end of the canal as a
“T”. Levees and floodwalls provide storm surge protection to an elevation of approximately
+13 feet (MSL) along essentially the full length of both sides of the canal, except at the
southern end.
The pumping plant is a brick masonry building that was constructed in 1903, and it
houses several of the large capacity Woods pumps of that same era. When the water level
within the canal rises three to four feet above normal, the operators report that water seeps
through the wall of the building that fronts the canal. It is clear that raising the water level
significantly higher against the brick face of this old structure would induce water pressures
that could collapse this wall.
The obvious solution would have been either: (1) for the Levee Board to construct a
floodwall across the south end of the canal, joining to the levees and floodwalls lining the east
and west banks of the canal, thus sealing the end of the canal and simultaneously protecting
the ancient structure, or (2) for the Water and Sewerage Board to construct a stronger wall, to
achieve the same two purposes.
Neither happened.
The Levee Board did not construct the wall to protect the property of the Water and
Sewerage Board (and the safety of the Public by closing the base of the canal), and the Water
and Sewerage Board did not assist the Levee Board by closing off the end of the canal (and
8 16
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
protecting their own building at the same time.) Instead, an opening of approximately 200
feet in length was left “open” on the east side at the south end of the otherwise continuous
levee and floodwall system lining the rest of this canal. A concrete “spillway” section
occupies this gap, to prevent erosion from further exacerbating the flows emanating from this
hole in an otherwise continuous flood protection system.
Figure 8.19 is a view of the south end of the Orleans canal, showing the brick masonry
pumping house, the levees and floodwalls on both sides of the canal, and the “gap” at the
south end of the east bank (on the left side of this photo.) Figure 8.20 shows this “gap” from
the outboard side, with elevations of key features indicated. Figure 8.21 shows an oblique
view from rotation of threedimensional LIDAR survey measurements (see Appendix A) of
this same section. All dimensions, and elevations, are captured by this LIDAR dataset to an
accuracy of approximately ±0.1 feet (or less). The “spillway” section across the open gap has
a crest elevation of approximately +6.8 feet (MSL), with a marginally lower “low spot”
slightly to the north of the concrete “spillway” section at Elev. +6.5 feet (MSL). The “gap”
thus represents a long opening in the otherwise contiguous levees and floodwalls along many
miles of both sides of this canal, and with a top elevation of approximately 6 feet below the
top of the adjacent floodwalls topping the levees (permitting overflow at approximately Elev.
+6.5 to +6.8 feet, MSL.)
As a result, while the storm surge along the southern shore of Lake Pontchartrain was
raising the water levels within the full lengths of the adjacent 17th Street and London Avenue
drainage canals, the rising storm surge (after reaching an elevation of approximately +6.5 feet,
MSL) simply caused floodwaters to flow freely into the heart of New Orleans through this
“gap” in the flood protection system.
The opening left at the south end of the Orleans canal resulted in lower water levels
toward the south of the canal, but did little to alleviate the storm surge rise at the north end.
The lack of failures along the north end of the canal must therefore have been the result of
more favorable embankment and floodwall geometries and/or foundation soil properties than
occurred along failed sections of the nearby London Avenue and 17th Street drainage canals.
On both sides of this canal there was considerably more right of way available, and the
earthen levee embankments along the Orleans canal are considerably wider than those along
either the 17th Street or London Avenue canals. Figures 8.22 and 8.23 show views of these
levee and floodwall sections along the east and west sides of the Orleans canal. The
embankment widths shown in these photos are in strong contrast to the narrower
embankments (and crowding from adjacent homes and yards) along the London Avenue and
17th Street canals, as shown for example in Figures 8.109, and 8.24 and 8.30, respectively.
An additional factor working in favor of the stability of the Orleans Canal levees and
floodwalls was the fact that the relationship between effective soil overburden stress and
resulting soil shear strength in the soft clayey and organic marsh soils near the north end of
the Orleans Canal embankments and floodwalls had been better treated during initial analysis
and design than it was for 17th Street Canal embankments and floodwalls for similar soils.
8 17
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
8.3.7 The 17th Street Canal
8.3.7.1 The Breach on the East Bank
(a) Introduction
One of the most catastrophic failures during Hurricane Katrina was a breach near the
north end of the 17th Street Canal, on the east side, just to the south of the Hammond Highway
bridge. The location of this breach is shown in Figures 8.1 and 8.2.
Figure 8.24 (which is a repeat of Figure 2.14) shows the use of military helicopters to
place oversized bags of gravel into this breach. This photo shows a number of important
features at this breach site. In this photo, it can be clearly seen that the inboard side of the
levee embankment (on the “protected” side of the floodwall) has translated laterally to the
east (to the right in this photo, which is taken looking north.) The translated embankment
section is relatively intact along the northern twothirds of this breach, and appears to have
swung much like a door about the northern end. Severe scour and damage to structures on the
inboard (“protected”) side at the south end of this feature support this mode; the major rush of
inflow was concentrated near the southern end of this breach.
A number of borings and Cone Penetration Test (CPT) probes were performed at this
site by the IPET investigation, by Team Louisiana, and by the ILIT investigation team. In
addition, several borings had been performed earlier, as part of the initial design studies for
the raising of the floodwalls at this location. Figure 8.25 is an approximate plan view of this
site, showing the locations of the borings and CPT performed by our (ILIT) investigation.
This plan view also shows the locations of a number of important features that help to shed
light on the causes and mechanism of this failure. Figure 8.25(a) shows the approximate
locations of borings and CPT probes performed at this site by the IPET investigation.
Figure 8.26 shows two views of a crosssection through the heart of this breach along
Section AA′ from Figure 8.25. Figure 8.26(a) shows this crosssection before the failure, and
Figure 8.26(b) shows this same section after the failure. Nearby cultural features (including
buildings, fences, and floodwall sections) as well as boring logs and CPT probes are projected
to this cross section for graphical clarity.
As shown in Figures 8.25 and 8.26, the intact levee segment near the center of the
breach moved laterally approximately 49 feet. To the inboard side (“protected” side) of the
displaced levee embankment sections, three sets of exiting toe overthrust features were
mapped, as also shown in these figures.
As shown in Figure 8.26(b), the breach was the result of a translational failure of the
inboard section of the embankment, pushed laterally by the water pressures exerted by the
storm surge in the canal acting on the outboard face of the floodwall and sheetpile curtain.
Figure 8.27 illustrates the sequence of movements associated with this failure, again for the
crosssection through Section A A′. As discussed in the text sections that follow, the rising
waters in the canal pushed laterally against the floodwall and eventually (progressively)
opened a gap between the floodwall and the outboard section of the levee embankment, as
8 18
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
illustrated in Figure 8.27(b). Water then entered this gap, and increased the lateral push
against the sheetpile curtain and floodwall. A shear failure then occurred in the foundation
soils beneath the embankment, and the embankment section along with the sheetpile
curtain/floodwall slid inboard, pushed laterally by the storm surge as illustrated in Figures
8.27(c) and (d).
Figure 8.28 is an oblique aerial view of this breach section, showing tops of the Iwall
sections that “pushed” the inboard section of the earthen embankment (driven by water
pressures on their outboard sides), and then toppled backwards towards the canal as the
translating levee embankment section finally came to rest and as water pressures equilibrated
when the neighborhood filled with water and the storm surge eventually subsided. It also
shows two sections of floodwall at the northern end of the failure (the near end in this photo)
toppled forward (toward the “protected” side) by the inrushing floodwaters at the north end of
this breach.
Figure 8.29 shows the tops of the Iwall sections at the very southern end of the
breach, which were also left “toppled forward” (towards the inboard, or “protected” side) by
the inrushing floodwaters passing through the breach opening.
Figure 8.30 shows a collapsed metal shed, with a corrugated roof, that was pushed
against the side of the home at 6914 Belaire Drive by “plowing” at the toe of the laterally
translating earthen embankment section, as is also shown in Figures 8.26 and 8.27.
Figure 8.31 shows a foundation slab at the toe of the failed section, immediately to the
south of the home at 6914 Belaire Drive. The final exiting toe thrust feature rises just at the
near end of this slab, which was partially laterally displaced despite being supported by piles,
as shown previously in Figures 8.26 and 8.27. Scour caused by the floodwaters also left an
erosional depression beneath and behind this slab, resulting in the “pond” shown in the
background of Figure 8.31. Also clearly visible in this photo are blocks of peat that were
scoured from the foundation strata by the inrushing floodwaters.
Figure 8.32 shows a piece of one of the exiting toe thrusts (Toe Thrust #1, from Figure
8.26(b)) at a location between the slab of Figure 8.31 and the home and collapsed shed of
Figure 8.30. Figures 8.33 and 8.34 show two views of the other two toe thrust features which
occur farther to the inboard (protected) side of the failure (Toe Thrusts #2 and #3, from Figure
8.26(b)).
The general failure mode involved water pushing on the canal side of the floodwall,
resulting in the opening of a gap between the sheetpile curtain/floodwall and the outboard side
of the earthen embankment. Water then flowed into this gap, and the resulting water
pressures pushed the inboard half of the earthen embankment (and the sheetpile
curtain/floodwall) sideways. This “cutting the embankment in half, opening a gap, filling it
with water, and then pushing the inboard half of the embankment (along with the sheetpile
curtain/floodwall)” mode of failure had not been considered or analyzed during the original
design of the floodwalls along the drainage canals. It was, however, not an unexpected mode
of failure as it had been clearly evinced in the E99 fullscale test section experiment near
8 19
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Morgan City in the nearby Atchafalaya basin in 1977 (as described previously in Section
8.3.3.)
In the second IPET interim report (IPET; April 1, 2006) this mode was selected as the
likely mode of failure based on stability analyses and centrifuge model testing performed as
part of the IPET studies. Our own investigation team had favored this failure mode from the
time of the initial postevent field observations in September and October of 2005. It was
apparent that this mode had been in operation at this site based on the field observations made
at that time. In addition, the same mode had also been in operation, and was “frozen” in place
as a partially developed or incipient failure, on the east bank near the north end of the London
Avenue drainage canal (see Section 8.3.8), and the field evidence also clearly indicated that
this same “half embankment with a waterfilled crack pushing laterally” had been the mode of
failure at the large breach on the west bank near the north end of the London Avenue Canal
(see Section 8.3.8). Also, we had read the E99 fullscale test section reports, and were aware
of the likelihood of this mechanism.
The deeper question is: What was the underlying mechanism that produced the
observed failure within the foundation soils beneath the embankment?
Here the findings of our investigation differ significantly from those of the second
IPET interim report, and those of the IPET Draft Final Report of June 1, 2006 as well. The
IPET report’s finding was that the failure was the result of a largely rotational failure,
shearing mainly through the soft gray clays occurring beneath the organic, marshy layers that
support the base of the embankment. Our own studies found that there were two mechanisms
that were each capable of producing the failure and breach, and that the margins of safety
associated with each of these did not differ by large amounts. The actual failure that occurred
followed the weakest and least stable of these two mechanisms, and was a largely
translational failure along a relative thin but laterally continuous stratum of weak and highly
sensitive organic clayey silt silty clay embedded within the “marsh” layer as shown in the
crosssections of Figures 8.26 and 8.27.
An examination of the various soil units, the various potential failure modes, and
analyses and explanation of the findings as to the nature of the actual failure mechanism,
follow.
(b) Geotechnical Analyses of the Failure
As shown previously in Figures 4.14 through 4.17, the north end of the 17th Street
Canal is situated atop largely paludal marsh clays and organic marsh deposits, in an area long
riven with erosional drainage features associated with the Lake Pontchartrain basin.
Figures 8.35(a) and (b) present two additional crosssections showing conditions prior
to the failure along Sections BB and CC in Figure 8.25. As shown, the foundation soil
conditions differed somewhat, but were largely similar along the width of the breach (failure)
section.
As shown in Figures 8.26, 8.27 and 8.35, the levee embankment was comprised of two
distinct soil fill zones. The upper embankment was a moderately compacted imported brown
8 20
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
clay fill, placed in the early 1970’s. This fill had raised the preexisting levee, which was
comprised largely of locally available gray clay fill from the local swamp deposits.
Placement of the original layer of gray clay fill dated back to the previous century, and these
earlier “historic” fills had consisted of simply piling up locally available gray paludal marsh
clays without compaction.
These two embankment fill zones were underlain by a layer of “marsh” deposits. This
was actually a relatively complex and layered zone, consisting of strata of peaty organics
interbedded with soft, sensitive organic clayey silts and plastic clays with very high water
contents, and of varying organic and fibrous organic contents. Cypress tree root systems were
common in this mixed “marsh” layer, apparently representing two distinct “stands” or levels
of cypress marsh as shown in Figure 8.26(a), and these root systems often interfered with
drilling and sampling.
The “marsh” layer was underlain by a transitional layer of progressively less organic
soils, with fewer fibrous and peaty inclusions and an increasing fraction of soft, plastic gray
clays. Beneath this transitional “intermixing” zone, the foundation consisted of soft, weak
gray paludal marsh clays (CH) of high natural water content (natural water contents of wo ᄃ
85 to 95 %.) These marsh clays were both weak and “sensitive”. Sensitivities (the ratio of
peak undrained shear strength vs. residual undrained shear strength) were typically on the
order of 2 to 6.
These soft gray clays were underlain by fine sands. These sands are adequately strong
and competent relative to the softer (and weaker) overlying soil units that they were not
involved in the failure. Similarly, although these sands were relatively pervious, this was not
a significant issue at this site as they occurred at sufficient depth that they were effectively
“capped” by the relatively thick low permeability layer of soft gray clays.
The plan views of Figures 8.25 and 8.25(a) show the locations of borings and CPT
probes performed for the original design studies, as part of the IPET investigation, and as part
of our own studies. The predesign and IPET borings generally used 5inch diameter thin
walled fixedpiston samplers to obtain samples. Most shear strength data reported from both
efforts that are currently available to our investigation team are the result of unconsolidatedundrained triaxial tests (UUTX) performed on these samples, although some samples were
tested in unconfined compression (qunc). A limited number of insitu vane shear test results
(VST) were also reported for some sites.
Our own field investigations involved primarily the use of 3inch diameter thin
walled, fixedpiston Shelby tube samples, and laboratory UUTX tests were performed on
many of these samples. The Shelby tubes were “modified” prior to use to eliminate the “rollin” at the cutting end that produces overcutting and then allows lateral expansion of the
sample during sample entry into the tubes. It has been shown (e.g. Lunne and Lacasse, 1994)
that the use of this type of constant tube diameter, sharpedged, thinwalled fixed piston
sampling with good technique can greatly reduce the disturbance otherwise associated with
sampling of the soft, sensitive clayey soils of principal concern at this site.
8 21
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Some of the borings were sampled continuously and the samples were extruded onsite
to examine the stratigraphy and geology in detail. Some of the borings were not sampled at
all; instead insitu vane shear tests were performed at selected depths within these boreholes.
Some of the samples were retrieved and brought to the laboratory for testing. Finally, some
of the samples were subjected to rather unusual laboratory vane shear testing, and this will be
discussed in detail a bit later in this section.
The boring logs for all borings performed as part of these current studies are presented
in Appendix B. Laboratory test data, including laboratory vane shear strength test data, are
presented in Appendix D. Insitu vane shear strength test data for tests performed within the
borings is summarized in Appendix D, and on the boring logs in Appendix B.
In addition to the borings, insitu and laboratory vane shear and laboratory testing,
both the IPET investigation and our own team performed a number of piezocone Cone
Penetration Test (CPTU) probes. Logs of the CPTU probes performed as part of our studies
are presented in Appendix C.
Figure 8.36 shows a summary of the shear strength test data available to our
investigation team for the embankment fill at and near the 17th Street drainage canal breach
site. Shear strength of the embankment fill is not of significant direct importance for the
conventional overall stability analyses that will follow, as the embankment fill “went for the
ride” and was carried along on shear surfaces that sheared through lower, weaker foundation
soil units. The strength data from Figure 8.36 was of some importance, however, in selection
of properties to model the nonlinear stiffness of these embankment soils in finite element
modeling of this levee and floodwall section. The heavy line shown in Figure 8.34 is the
shear strength modeled through the embankment fill along the embankment centerline
(directly beneath the levee crest) in these studies. The strength lines representing CPT data in
Figure 8.36 are based on interpretation of the CPT data using a cone tip factor of Nk = 12 in
the upper, brown clay fill and Nk = 12 in the lower, gray clay fill as well.
Figure 8.37 shows an example of the available CPTU data beneath the central portion
of the levee embankment, for “Marsh”, “Intermixing Zone” and “Gray Clay” strata shown in
the crosssections of Figures 8.27 and 8.35. Figure 8.38 shows a similar example of CPTU
data, but this time for locations outboard of the toe of the levee embankment. As expected,
shear strengths are notably lower here, due to lesser effective vertical stresses resulting from
lesser overburden loads.
As shown in Figures 8.37 through 8.40, there are distinct differences between the
“gray clay” strata and the “marsh” strata, and these will therefore be treated separately.
Beginning with the deeper unit, the gray clays, it must be observed that our
interpretation differs somewhat from that presented in the second IPET interim report. The
IPET report assumed that these clays were normally consolidated as they had been protected
from desiccation by the overlying swamp deposits. Our interpretation differs, as we found
three separate “stands” in the evolution of this layer of soft gray clays and the overlying
marsh deposits, with three corollary desiccationinduced overconsolidation profiles associated
with these.
8 22
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
The shear strength data based on UUTX tests from the initial design studies, as well as
the IPET studies, showed considerable scatter and this was considered likely to reflect the
issues associated with sampling disturbance for these soft, sensitive soils. The CPTU data
from both the IPET and our own (ILIT) studies, on the other hand, appeared far more
consistent within this stratum (as shown in Figures 8.37 and 8.38.) Figure 8.39 shows a
typical plot of the pore pressure parameter Bq from a CPTU performed through the crest of
the levee (Bq = Δu/(qt σvo)), highlighting the value of Bq where the clay appears to be
normally consolidated. Figure 8.40 then shows these values of transposed onto the
relationships of Lunne et al. (1985) and Karlsrud et al. (1996) to determine appropriate values
of the cone tip factor Nkt for conversion of CPT tip resistance to undrained shear strength. As
shown in this Figure, the value determined for this stratum was approximately Nkt = 12.
B
Figure 8.41 then shows the values of [Su/P]OC/[ Su/P]NC vs. OCR determined for
minerologically similar Mississippi River clays of similar depositional history in Atchafalaya
as determined by Foott & Ladd (1977). The SHANSEP exponent for these similar clays was
found to be λ = 0.75, a relatively normal value for clays of this plasticity and character.
Using a value of Nkt = 12, and λ = 0.75, the CPTU data within the soft gray clay
foundation stratum was then processed to develop plots of Su/P vs. depth, and OCR vs. depth,
for CPT beneath the full height of the levee (Figure 8.42) and inboard of the levee toe where
effective overburden stresses were significantly lower (Figure 8.43).
As shown in Figures 8.42 and 8.43, the results show a pleasingly consistent pattern.
The clay inboard of the levee toe clearly evinces three “stands” of the marsh development,
with three OCR profiles associated with surficial desiccation. The clays beneath the levee
embankment loads show just the residual tips of these same three OCR “crusts”, as the clays
have been further loaded by the placement of the overlying embankment fill and so are more
nearly normally consolidated over most of the stratum. Near the base of this stratum, the
clays inboard of the levee toe show a minor degree of overconsolidation associated with
secondary compression (as verified by subsequent consolidation analyses using the program
PLAXIS which successfully modeled the evolution of this site and accurately reproduced this
basal OCR profile).
In establishing the plots shown in Figures 8.42 and 8.43, the value of (Su/σv)NC = 0.31
was found to best fit the data. This is a fairly normal value for clays of this plasticity, and it
was exactly the same value found by Foott & Ladd (1977) for the minerologically similar
clays at Atchafalaya.
The green lines in Figure 8.44 shows the resulting profiles of Su vs depth within the
soft gray clay foundation stratum (a) beneath the crest of the levee, and (b) inboard of the
levee toe, based on Su/P = 0.31 and λ = 0.75. Also plotted on this figure are the CPTU tip
resistance data converted to Su based on Nkt = 12, and the results of UUTX tests on
“undisturbed” ILIT samples, lab vane tests (LVT) on ILIT samples and in situ field vane
shear strength tests (FVT). The overall “fit” to all the data is generally very good.
Figure 8.45 then repeats Figure 8.44, but adds the rest of the available IPET and pre
Katrina strength data (including UUTX, FVST and CPT data.) For the “toe” region some
8 23
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
adjustment of this data is necessary in viewing this figure, as some of the IPET data is located
such that some portion of the embankment overburden stresses slightly increase the shear
strengths for some of the “toe” data; as a result these data (including the CPT) tend to drift to
the right (to the stronger side) a bit, especially at depth. Overall, these additional data also
well support the relationships developed.
Figure 8.46 then shows the selected value of (Su/P)NC = 0.31 for UUTX, field vane and
lab vane tests plotted vs. data for other clays (Ladd, 2003). It also shows the value of (Su/P)NC
for direct simple shear (DSS) tests on the minerologically similar Atchafalaya clays by Foott
and Ladd (1973). Both sets of data fit well with the overall background relationship implied
for other clays. This suggests that an appropriate scaling factor for the Su values for
conversion from “triaxial” conditions to the DSS stress path conditions that will better
represent the stability and deformation analyses for this embankment and floodwall system is
approximately 0.80 to 0.84, as shown in Figure 8.46. A value of Su,dss = 0.82 x Su,tx was used
for this soft gray clay in these studies.
As an additional check, the value of (Su/P)NC = 0.31 (for triaxial and in situ vane shear)
determined for these clays was also checked against other clays (Figure 8.47.)
Figure 8.48 shows similar treatment of the derivation of the CPT cone factor Nkt based
on Bq, this time for the “marsh” deposits overlying the soft gray foundation clays. Based on a
value of Bq = 0.25 to 0.40, a value of Nkt = 16 was determined and used to process the CPT
data for this unit.
A second approach was also used to also develop profiles of Su/P vs depth and OCR
vs. depth for these marsh deposits, as shown in Figure 8.49. The relationship of Mayne and
Mitchell (1978) was used, in conjunction with the available UUTX, LVST and FVST data to
iteratively develop relationships for Su/p vs. depth and OCR vs. depth, as a function of
Plasticity Index (PI, %) over the range PI ᄃ 55% to 140%, which encompasses the range
observed in this complex soil unit. The resulting relationships confirm the classic desiccated
OCR crust profile shown previously in Figure 8.41 for this “marsh” deposit.
Figure 8.50 then shows the resulting interpretation, based on all available data, of
strength vs. depth within this complex marsh unit for conditions (a) beneath the overburden of
the central embankment, and (b) inboard of the toe of the levee. The green lines in this figure
represent the final interpreted soil shear strength profiles at these two indicative locations. As
with the soft clays, these strengths were, finally, further slightly reduced by multiplying them
by a factor of 0.82 to develop the DSStype strengths needed for the stability analyses
performed in these studies.
The red zone near the center of the “marsh” deposits shown in Figure 8.50 is a thin
layer of soft, highly sensitive organic silty clay that varies slightly in depth across the profile
(and so is thinner than it appears in this figure.) This was the material in which the main
lateral translational shear failure occurred at this site.
Figure 8.52 shows a sample of this thin layer at one of three boreholes within the slide
region (near to the large, relatively intact displaced levee block) that captured a sheared
8 24
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
sample of this layer. The material is completely remolded and sheared to a fully residual
condition with negligible remaining strength, and unidirectional extension and tearing of
organic fibers across the sheared zone clearly indicate the shear failure within this sample.
This layer is typically only one to several inches in thickness, but was found to be
laterally continuous across essentially the full site (as well as at the distressed section on the
opposite, west side of the canal.) It is exceedingly difficult to spot, and to sample, because it
is closely overlain (and even partially mixed with) a layer of leaves and twigs and bark that is
typically also one to several inches in thickness, as illustrated on the auger stem in Figure
8.51. The very dark, shiny material also coating the auger stem in this photo is the sensitive
organic silty clay and indicates that we have just drilled through the layer in question (and so
now have to move our hole laterally a few feet and redrill to attempt to sample it.)
This layer of sensitive organic silty clay is the result of a previous major storm that
churned up organics and sediments, mixed them with the locally prevalent clays, and also
greatly (temporarily) increased the salinity of the water so that the ensuing deposit is
unusually heavily flocculated. The result is a material of low strength and extremely high
sensitivity (sensitivities of between about 10 and 20+.)
The same storm was accompanied by winds that knocked down leaves and twigs and
bark (and other organic detritus), accounting for the closely overlying layer of organic
impediments that “mask” this thin layer.
Figure 8.53 shows a plan view of the site, highlighting with red the 10 locations at
which this layer was positively identified. It was not always possible to positively identify
this thin layer in CPT, as the strength of this layer is not much less than that of the closely
overlying and underlying soils; it is the combination of low strength and high sensitivity that
made this thin layer so dangerous. “Thin layer” effects also made spotting this layer in CPT
(based on tip resistance) difficult. The best initial “marker” or signature of the presence of
this layer was found to be a positive spike in friction ratio; as the sleeve continued to drag
through the overlying and underlying deposits but the tip resistance dipped a bit.
Figure 8.54 shows a photo of an “undisturbed” sample of this sensitive organic silty
clay. The local clays have a gray, peanut butterlike appearance and consistency. They are
not highly shiny, but rather semiglossy, and their stiffness and texture are not unlike peanut
butter. The sensitive organic clay, on the other hand, is dark and has a very shiny and
translucent appearance; much like “jelly”, as shown in Figure 8.54. In Figure 8.54, hints of
the organic detritus that closely overlies and masks access to this thin layer can also be seen.
Two approaches were taken to attempt to characterize the strength (and stress
deformation) behavior of this material. At any location, the precise depth of this layer was
first determined by drilling to encounter it. One approach was then to move the drill rig
laterally several feet and to redrill to within approximately one foot of this layer. A 3foot
long Shelby tube, 3inches in diameter (and modified to eliminate the turnin that produces
overcutting at the mouth) was then used, with a fixed piston system, to drive the tube
approximately two feet past the target layer so that more competent underlying soils would
“plug” the bottom of the tube and permit careful withdrawal of a sample. Otherwise, the
8 25
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
samples remoulded upon attempted withdrawal and slopped out of the base of the tube
making sample recovery nearly impossible.
The samples thus obtained were then taken to the lab at the University of California at
Berkeley, where they were subjected to an unusual process, as illustrated in Figure 8.55. The
tubes were cut off in 2inch increments, and a small spoon was used to carefully dig ahead
into the remaining tube. When the telltale organic detritus was encountered digging stopped
and the organic material was handplucked form the tube to daylight the underlying sensitive
layer. A lab vane shear test was then performed.
The second method used to evaluate the strength of this material also began by pre
locating the precise depth of this layer, usually by sacrificially “oversampling” it (to plug the
base of the tube to foment retrieval) and then extruding the sample to determine the precise
location of the layer. A second, adjacent hole was then carefully hand augered, and an in situ
vane shear test was performed using a shallowbladed vane. Insertion disturbance, and
obstruction by unremoved organic detritus (mixed in the top of the layer) sometimes defeated
this effort, often making multiple attempts necessary. Unacceptable insertion disturbance was
apparent when the characteristically brittle peak to residual transition was absent and the
material exhibited only residual strength.
Figure 8.56 shows typical stressdisplacement plots for tests on the thin layer of highly
sensitive organic silty clay, and on the local deposits of sensitive gray clay. As shown in
Figure 8.56(b), which shows normalized behavior in the form of shear strength divided by
maximum shear strength on the vertical axis, the sensitive organic clay was more highly
brittle, failed at lower displacement, and exhibited even more pronounced sensitivity and
rapid postpeak strength degradation. It was the combination of low strength, and this very
brittle sensitivity, that caused this material to “capture” the failure surface at this site.
Finite element analyses were performed for this levee and floodwall section using the
program PLAXIS. Figure 8.57 shows the principal parameters and the mesh used for these
analyses. The gray foundation clays (CH) and the “marsh layer” were modeled using the
“soft soil” effective stress model within PLAXIS, and the soil parameters used were fitted to
the values of Su/p vs. OCR as described previously to match the evaluated strengths of these
units and their distribution.
It was necessary to establish the stress state at the end of incremental construction and
consolidation of the embankment and foundation. Initial overconsolidation profiles due to
desiccation and secondary compression were input, and embankment construction was
modeled in two stages (the “historic” fill, and the more recent engineered top fill), and both
the OCR vs. depth and the settlement pattern (the bowl shaped pattern at the base of the oldest
fill) were well matched to the observed field conditions. Figure 8.58 shows the settlements
calculated at the end of initial construction and consolidation.
The front lip of the embankment was then “excavated” and the floodwall installed (as
with the actual field case), and displacements were rezeroed to prepare for the remaining
analyses to follow.
8 26
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Water levels within the canal were incrementally raised, and within a range of
stiffnesses considered reasonable it was found that initiation of “gapping” between the
outboard toe of the floodwall and the outboard embankment section typically initiated at a
surge elevation of between about 7.5 to 8 feet, as illustrated in Figure 8.59. This Figure
shows normalized shear strain contours, with the red color indicating shear strains equal to or
greater than the shear strain to “peak” shear strength (and thus localized failure.) As shown in
this figure, with a water elevation of +8 feet (MSL) gapping has opened partially down the
front face of the sheetpile curtain (on the outboard, or water side), and the thin, sensitive
organic silty clay layer has already sheared to failure along a short segment inboard of the
crest of the levee.
If one looks very carefully at Figure 8.59, a second “lighter” area can be seen beneath
this shear zone, representing the beginning of shear deformations along a more “rotational”
shear surface passing through the deeper soft gray foundation clays (CH). A dashed line has
been added to indicate this surface. This deeper, and more rotational failure surface has a
calculated factor of Safety only slightly higher than that calculated for the upper sensitive
organic silty clay layer, and this deeper surface represents the failure mechanism favored by
the IPET studies reported to date.
As the analysis began to calculate the progressive development of tensile effective
stresses between the front of the sheetpiles and the soil, the mesh was revised to model the
development of a “gap” between these, and the intrusion of water into the gap as well. Once
this “gapping” began, it then developed rapidly. Figure 8.60 shows the situation with an
additional foot of storm surge rise to Elev. + 9 feet (MSL) based on our best estimates of the
soil parameters. As shown in this figure, the gap has now extended nearly to the base of the
sheetpiles. Further extension of the gap is temporarily held up by the malleability of the
marsh soils, but further gapping does not provide significant additional lateral water pressures
against the front of the sheetpile curtain because the lateral permeability of the “marsh”
deposits is relatively high. At this stage, the shear failure along the thin layer of sensitive
organic silty clay is well developed, and embankment movements are now significant. This
figure also shows quite clearly the deeper, more rotational failure surface that represents the
second least stable mechanism at this site (the mechanism favored to date by the IPET
studies.)
Figure 8.61 shows calculated displacements for a surge height of 8.5 feet, with
displacements exaggerated times two for clarity. Initially, the floodwall tilts slightly forward
as it compresses the soils a bit. As sliding then develops, the floodwall base begins to move
along with the displacing embankment and the whole moving mass (inboard embankment
section, floodwall and sheetpile curtain) displace laterally together, as shown previously in
Figure 8.27.
Figure 8.62 shows the Factors of Safety calculated (by cØ reduction) using PLAXIS
for a variety of water levels in the canal. Three cases are presented: (1) failure dominated by
the thin layer of sensitive organic silty clay, but without gapping between the sheetpile curtain
and the outboard side soils, (2) a more rotational failure through the deeper soft gray
foundation clays, again without gapping, and (3) failure dominated largely by the upper
8 27
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
sensitive organic silty clay layer, but this time with a waterfilled gap on the outboard side of
the sheetpile curtain.
As shown in this figure, the Factors of Safety for the upper lateral shear failure, and
the deeper more rotational failure, are not very different. The heavy red line shows the best
estimated path to failure at this site. Based on these analyses, it appears that gapping would
have developed at a surge height of between about 7.5 to 9 feet (MSL), and the intrusion of
water into this gap would have increased the lateral forces and rapidly driven the section to
instability.
Figure 8.63 shows the crosssection and principal soil properties used to perform more
classical limit equilibrium analyses (using Spencer’s Method, crosschecked against
Morgenstern’s and Janbu’s Methods) using the program SLOPE/W.
Figure 8.65 shows the most critical failure mode for the “no gapping” case with a
surge height to Elevation +6 feet (MSL). The PLAXIS analyses had shown very little
likelihood of gapping at this water elevation, and this probably represents the best estimate of
Factor of safety for this surge height. As shown, the calculated Factor of Safety is FS = 1.51
for this case, and as shown in Table 8.1, the associated probability of failure for this surge
height is approximately Pf = 0.01. These calculated low probabilities of gapping and of
failure are reassuring, as the water in the canal had previously reached an elevation of
approximately +6 to +6.5 feet (MSL) during previous storm surges, and no gapping or failure
had occurred in those events.
Figures 8.66 and 8.67 show the most critical failure surfaces for a surge to Elev. +9.5
feet (MSL) for (a) a shallow translational failure dominated by sensitive organic silty clay
layer, and (b) a deeper, more rotational failure through the soft gray foundation clays. In both
analyses, a waterfilled gap was modeled at the outboard side of the sheetpile curtain. This
water elevation is approximately the maximum elevation achieved (maximum surge at this
location is estimated by our team to be approximately Elev. +9.5 to +10 feet, MSL). The
calculated Factors of safety are again similar for both modes, and the shallow lateral
translation along the sensitive organic silty clay again provides the lower Factor of Safety.
Figure 8.68 shows calculated Factors of Safety for various water elevations (Spencer’s
Method) for the four cases of principal interest: (a, b) lateral translation along the sensitive
organic silty clay layer, with and without a waterfilled gap, and (c, d) deeper and more
rotational failure, again with and without a waterfilled gap. The solid red line again shows
the bestestimated path to failure at this site, this time based on the suite of limit equilibrium
analyses.
It is challenging to make an estimate of the probability of failure at any given canal
water level, as there are numerous uncertainties involved, and some of these are cross
correlated. The principal uncertainties are those associated with shear strengths of the
foundation soils, and also with the “representative” shear strength that can be mobilized at any
given moment by the very sharply strainsoftening soils (especially the highly sensitive, thin
organic silty clay layer.) Additional significant uncertainties are those associated with the
likelihood (and severity) of opening of the waterfilled gap at the outboard side of the
8 28
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
floodwall and its supporting sheetpile curtain, and the unit weights of some of the soils.
These were not all accurately reflected in these probabilistic estimates, and as a result the
overall uncertainty is expected to have been somewhat underestimated.
One important set of variables are the shear strengths of the various soil units
controlling each of the potential instability modes. Each of the conventional limit equilibrium
analyses performed was performed using probabilistic variation of these shear strengths. The
coefficient of variability in soil shear strengths was taken as lognormally distributed, and was
estimated as approximately COV ᄃ 30% for the soft gray clays, and COV ᄃ 40% for the thin,
sensitive organic silty clay layer. The resulting distributions of probable factor of safety are
shown (approximately) graphically in Figure 8.68(a) for both the “ungapped” case and the
case of a waterfilled gap at the outboard side of the sheetpile curtain. These are only
approximate, as they do not precisely fit themselves to any single wellknown distribution.
The next critical uncertainty is the probability of cracking. The probability of
cracking cannot be calculated or evaluated in any closedform manner, and so requires a
judgmental estimate based on the preceding finite element analyses (and supported in part by
the observed field behavior). It should be noted that the stiffnesses used in the PLAXIS
analyses to estimate inception of cracking are a bit time dependent, so that a slower rising and
falling storm surge would be a bit more deleterious here. The inception of cracking was not
taken as the point at which cracking “occurred”; instead cracking was taken to be significant
when the crack propagated more than halfway towards the base of the sheetpile curtain.
Figure 8.68(b) shows the judgmentally derived estimates of probability of significant crack
formation as a function of rising canal water elevation. The upper and lower bounds shown
were inferred to represent approximately ± 3ε values.
Monte Carlo simulation was used to estimate the distribution of the factor of safety
considering the analysis with and without a gap and the probability of a gap forming. The
formulation is as follows:
P(FS) @ P(FS NG NG)P(NG) . P(FSG G)P(G)
This equation reads; the distribution of the factor of safety is equal to the conditional
distribution of the factor of safety for the levee with no gap multiplied by the probability of
there being no gap, plus the conditional distribution of the factor of safety for the levee with
gap multiplied by the probability of there being a gap. The conditional distribution of the
factor of safety with or without a gap is based on the mean and standard deviation from
stability calculations. For the probability of gapping, which can also be considered a
transition function from no gap to gap conditions, a mean probability function and upper and
lower bounds were estimated from Figure 8.68(b). The above equation is for any single canal
water elevation.
All the distributions were treated as Gaussian based on observation of the data. The gap
and no gap conditions were considered statistically independent scenarios. A Monte Carlo
simulation was run for 10,000 samples. Typical simulation results are shown in Figure
8.68(c) for a single depth increment. The first plot is a histogram of the simulation results of
the factor of safety for no gap conditions [FSNG|NG], the second is for gap conditions
8 29
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
[FSG|G], the third is the probability distribution of a gap (no gap) occurring [P(G) and
P(NG)=(1P(G))], and the fourth is the total distribution of the factor of safety [P(FS)].
Figure 8.68(a) showed the distributions of factor of safety for the gap and nogap cases
(separately) as a function of rising canal water levels. Figure 8.68(d) repeats this figure as a
background, but adds the now calculated distributions of conjugate overall factor of safety as
a function of rising canal water levels, showing how the conjugate distribution “transitions”
from the ungapped to the waterfilled gap case. Based on these approximate simulations, the
resulting probabilities of failure at any given canal water elevation are then as shown in Table
8.1.
As shown in Table 8.1, the probability of failure was found to be very low for surge
heights of less than about Elev. + 7 feet (MSL), and they rise rather quickly as the surge
elevation passes above about + 8.5 feet (MSL). Failure at the estimated actual maximum
surge elevation of approximately + 9.5 to + 10 feet (MSL) is calculated to have had a
likelihood, on the order of Pf ᄃ 0.8 to 0.9. Failure at the originally intended “design” surge
height of Elev. + 12.5 feet (MSL) was essentially certain.
Finally, Figure 8.69 shows a comparison between the observed failure mode and the
rotational mode determined by IPET. There have been a number of IPET representations (to
date) of their failure mode, each varying slightly as to actual depth and dimensions, but all
were semirotational failures through the soft gray clay stratum underlying the marsh deposits.
The rotational surface shown in Figure 8.69 is a somewhat “average” representation of these
various failure surfaces, from both the second interim report (IPET, April 1, 2006) and the
more recent Draft Final Report (IPET, June 1, 2006.) The rotational IPET failure is
superimposed, as carefully as possible, onto our own investigation’s more detailed cross
section. The two modes are not wholly dissimilar, and both lead to low factors of Safety.
More detailed examination of the IPET mode, however, shows it to be problematic
with regard to agreement with key field evidence. The rotational IPET mode would have left
the chain link fence at the edge of the crest road (on the displaced intact levee block) rotated
backwards, but as shown clearly in Figure 8.26 (and Figure 8.69(a)) this crest fence was
essentially perfectly vertical at the end of the displacements. Massive rotation would have
been necessary to produce the observed very large lateral displacement of the upper “intact
crest” section of the levee (lateral displacement of up to 50 feet), and this would not have
been feasible with the IPET failure mode. Also, the IPET rotational mode would have
significantly backrotated the floodwall; but the floodwall instead traveled the full lateral
distance (50 feet) in contact with the displacing levee embankment section, and then toppled
backwards as the water pressures began to equilibrate (as illustrated in the top of Figure 8.69,
and in Figures 8.28 and 8.69(a). The IPET mode also fails to explain the large lateral extent
of the mapped toe exit features, and the multiple toe thrust features, as shown in Figures 8.26,
8.27, and 8.32 through 8.34.
Most importantly, the shear failure along the thin, highly sensitive organic silty clay
stratum was confirmed at several locations based on remoulding and also unidirectional
extension and tearing of organic fibers; conclusive evidence as to the occurrence of massive
unidirectional shear failure along this stratum within the marsh sequence.
8 30
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
The failure of the IPET investigation to discover the critical thin stratum of sensitive
organic silty clay that was the principal culprit in the failure at this site represents an
important lesson both for current geotechnical practice at large, and also for subsequent
design studies for the levees and floodwalls of the new Orleans regional flood protection
system. The IPET studies drove numerous geotechnical borings and CPT probes right
through this stratum (see Figure 8.25(a)) but did not discover it. That was, in large part,
because the crews performing the field borings and CPT were “separated” from those who
had performed the initial IPET postevent forensic field studies, and both subteams were
separated from the expert “engineering geology” team also working on the IPET studies. The
analysis subteam was a fourth, separate group. There were geological experts on the IPET
team who could certainly have pointed out the possibility, and even likelihood of such a layer
if they had been asked. Instead the four subteams performed their tasks largely separately,
without adequate interchange of knowledge and findings.
Our own investigation team took a wholly different approach. We began by carefully
assessing the visually observable surface forensic evidence at this site in the wake of the failure,
and by backtracking through the original (preKatrina) field and lab data for this site. We also
studied the challenging geology of the region (including seminal publications by the USACE’s
geological experts.) Based on all of this, our site team (which included senior investigative team
members right out on the drill rigs) went in search of an unusual layer, of high sensitivity, and
considerable lateral extent, that would have been capable of producing a lateral translational
stability failure with toe thrust features extending to unusually great distances inboard of the
original levee toe. The suspected depth of this stratum, inboard of the levee toe, was fairly
shallow; probably between several feet to as much as 10 feet at most. We encountered and
identified the critical layer with our first boring, and then sampled and tracked it across the site in
a total of 11 borings and CPT’s.
Two important lessons here are: (1) the importance of fully integrating all phases of field
investigation, laboratory testing, analysis and design (and the team members performing these),
and (2) the importance of suitably involving expert engineering geologists in all phases of site
investigation and site characterization, as well as the other project phases. These two things are,
unfortunately, not always done in contemporary geotechnical practice, and they are also not the
norm for many contemporary Corps design studies which tend to be relatively segmented (as were
the IPET studies in this case.)
Overall, it can be concluded that there were two potentially critical failure modes at
this site, but that the lateral translational failure along the sensitive organic silty clay layer
within the “marsh” deposits was the weaker of the two, and that this was the mode of failure
that actually occurred at this site.
(c) Initial Section Design Studies
The obvious next question to address is then how the original design studies failed to
note this. The answer is a bit complex as a number of poor judgements and errors contributed
to the misperception of the original “design” section as being adequately stable (and reliable)
8 31
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
for targeted design canal water elevations significantly higher than those that caused the
actual failure (the design canal water level was Elev. +12 feet, MSL). The original design
studies have been reviewed, and the following are significant errors and poor judgements
during initial design that contributed to this failure:
1.
Figure 8.70 shows the longitudinal cross section along the segment of the east
bank of the 17th Street Canal as developed for the original design studies. An early
error in the design process was the use of borings that were too widely spaced to
attempt to characterize challenging and complex foundation geology. The savings
achieved by not performing more borings now appear miniscule relative to the cost
of the catastrophe that has ensued. [It should also be noted, however, that even the
borings that were performed appear to have been sufficient as to correctly predict
the failure, if the resulting data had been suitably processed and then used in the
ensuing analyses.]
2.
The longitudinal section of Figure 8.70 was prepared by the USACE, and was
based on a number of assumptions; including the assumption the “marsh” deposits
were typically flatbottomed. The history of previous drainage channel erosion
across this area would lead to the expectation of likely nonlevel transitions even
for swamp bottoms, and Figure 8.71 shows our own team’s reinterpretation of the
original (sparse) longitudinal data to develop an alternative longitudinal subsurface
soil profile. This difference in interpretation might be considered the second
problem at this site during original design.
3.
The USACE then passed the design on to outsourced engineers, who developed
the strength data and interpretations for analysis of stability of the intended levee
and floodwall section. A major problem occurred here, as data from far too large a
lateral distance was eventually transposed to the design analysis crosssection. In
the vicinity of the actual failure, there are only 5 sample locations shown within
the critical “marsh deposits” (in the 4 borings shown intersecting this unit.)
4.
Two of the sample locations shown within the “marsh” deposit of Figure 8.69
were nonrecovered samples, and at approximately the same depth in nearly
adjacent borings. This is the location of the sensitive organic silty clay layer that
actually caused this failure and breach. Failure to note the importance of the non
recovery of testable samples, and in two nearly adjacent borings at essentially the
same elevation, should have represented a red flag and an effort should have been
made to further investigate this location.
5.
Figure 8.72 shows the stability calculations for the critical section nearest to the
actual breach and failure. The limit equilibrium method used for these was the
“Method of Planes”, a threewedge analysis with conservative side force
assumptions. This method continues to be preferred by the New Orleans District
of the USACE, but it is now a relatively archaic anachronism given the availability
of more accurate methods and the availability of the simple computer programs
necessary to run these. The method itself provides a slightly conservative answer
so long as the most critical failure surface can be closely represented by the steeply
plunging wedges at the front and back, and by the horizontal surface in between.
In the original design analyses, layers were assumed to be laterally horizontal, so
8 32
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
this analysis was a good fit for the crosssections analyzed. Unfortunately, the
actual stratigraphy was not horizontally layered (see for example any of the cross
sections analyzed in these current studies), so this method was poorly suited to the
finding of the failure mechanism that was actually most critical.
6.
And the assumption of laterally horizontal layering was itself a major problem too.
It was born of necessity, as no borings had been performed significantly off the
embankment centerline alignment to permit development of full lateral cross
sections. Again, the minimal savings on exploration and testing costs here pale
relative to the costs of the catastrophe that ensued. Stratigraphy is a vitally
important issue, especially given the low strengths of many of the foundation soils.
Looking at the crosssections at the 17th Street canal breach site as analyzed in this
current study, for example, one will note a subtle “bowl shaped” settlement pattern
at the base of the embankment fill, and a corresponding bowl shape to the critical
sensitive organic silty clay layer just beneath it. Without this “bowl shape”, the
original embankment would have been unstable during initial construction; it
would have slid sideways on the sensitive layer if that layer had been horizontal.
Instead the layer dipped in the center, so that the evolving embankment would
have had to slide up a small slope (up a hill) to fail during construction. Minor
changes in stratigraphy details can have a major impact on overall stability on
these soft, weak soils. Use of “assumed” horizontal layers therefore missed a
vitally important element of the problem.
7.
Figure 8.73 shows the now wellcirculated summary of strength data for stability
analyses at this section. The data are based on UU triaxial tests and on vane shear
tests. Scatter in the data is considerable, and is likely due in large part to sampling
disturbance issues for these sensitive soils. Most samples were obtained from
borings through the crests of the levees (the most accessible location) and so
represent strength information for locations under full embankment overburden
stresses. The solid lines in this figure show the strength interpretation used in the
actual design analyses. This line represents an unconservative assessment of the
data points presented, in both sides of the figure, even without allowance for the
additional effects of overburden stress reduction away from the levee centerline.
This interpretation is especially unconservative at elevations of between + 10 feet
to – 10 feet (Cairo datum) in the figure at the right, and between 10 feet to 30
feet (NGVD datum) in the figure at the left. These both represent the same 20 foot
range of critical elevations, which correspond approximately to Elev. 10 feet to 30 feet (MSL), and this is the region in which strengths are important in the
“Method of Planes” analysis performed for this location in the original design
studies. As shown in Figure 8.74, a majority of the available shear strength data is
lower than the shear strength actually used for the stability analyses in this critical
depth range; violating customary “Corps” procedures in this regard. (Corps
procedures generally require that approximately 1/3 of the data fall below the
strength used for analysis and design, and that 2/3 of the data be greater.) As
shown in Figure 8.70, the resulting calculated Factor of Safety was found to be FS
= 1.30….., barely enough to satisfy the design criteria which required a FS of at
least 1.3 for the case of “transient” storm surge loading. It is very difficult to
8 33
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
justify the apparently unconservative strengths selected in this critical elevation
range based on the data presented.
8.
Figure 8.74 is a repeat of Figure 8.73, but with additional red and blue lines added
to illustrate another major error made in determination of shear strengths for
stability analyses. Shear strengths of soils are very strongly a function of effective
overburden stress, so the samples obtained from beneath the overburden of the
embankments would consistently overestimate the strengths under the levee toes,
and in the “free field” out beyond the levee toes. This fundamental principle of
soil mechanics was wellknown in local practice in the New Orleans region at the
time that these analyses were performed. However, it was ignored in the original
design studies at this section, and the result was a massive additional increase in
the unconservative error in the overall stability analyses. The blue lines on Figure
8.74 represent our own team’s assessment (as described in preceding sections) of
the shear strength vs. depth beneath the crest of the levee, and the red lines
represent our assessment of the shear strength vs. depth inboard of the levee toe.
The contrast is very significant, and the unconservatism involved in the misuse of
strengths from “beneath the full levee overburden” to model conditions beneath
and inboard of the levee toe is readily apparent.
9.
Despite having adroitly invested significant funds and effort in the E99 test
section (near Atchafalya; see Section 8.3.3) to perform a very welldesigned full
scale field test on appropriate foundation soil conditions, the results of this field
test of a model floodwall/sheetpile curtain in a levee embankment founded on
weak marshy soils were not subsequently used (as had been intended.) The failure
mechanism disclosed by this field test was the opening of a gap at the outboard
side of the sheetpile curtain, the filling of this gap with water, and thus the
resulting exertion of increased lateral water pressures against the sheetpile curtain.
This mechanism, which proved to be the actual field failure mechanism at this site,
was not among the suite of cases/mechanisms analyzed in the original design
studies.
10.
And the use of a design Factor of Safety of only 1.3 was also a major problem. As
discussed in detail in Chapters 11 and 12, this was far too low a value for a system
protecting a large urban population. This value has a history of development that
is traced in Chapter 12 back to use for design of levees protecting agricultural
lands in the first half of the last century, and failure to update this in the face of
both the passage of time and the increased level of potential consequences
associated with flood protection of a major urban area was a significant lapse that
left little room for the other errors and poor judgements cited above.
Calculations using the data available at the time of the initial design, and using
analysis methods widely available and in common use at that time (though not necessarily
within the new Orleans District of the USACE), clearly indicate that this section would be
expected to be unstable at canal water levels less than those for which the design was intended
(water level of less than Elev. +12 feet, MSL). The more sophisticated analyses employed in
these current (ILIT) studies give more precise answers, but this level of sophistication was not
necessary to demonstrate overall deficiency of the original design.
8 34
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
8.3.7.2 Distressed Section on the West Bank
There is a “distressed” levee and floodwall section on the west bank of the 17th Street
Canal, across from the large breach discussed above. This “distress” was visually minor, but
this section was studied both as a check of the ramifications of “minor” visually observable
distress, and also because it provided an opportunity to see if the same analysis methods that
correctly predicted the failure on the east bank could also accurately predict the observed
performance of a second section that it was hoped would be somewhat similar.
Figure 8.75 shows measurement of observed lateral wall offset at the point of
maximum offset. Wall tilt is less than 0.75 inches, and the maximum lateral offset is
approximately 3.5 inches.
As shown previously in Figure 8.25, only a few borings and CPT were performed at
this distressed section on the west bank of the canal, so data is sparse. Figure 8.77 presents
the interpreted crosssection used for analysis at this site. The same basic sequence of strata
observed on the east bank are again present, but the details of the stratigraphy differ a bit.
Passing quickly through intermediate details (as were presented in detail in Section
8.3.6), the same procedures were used to process and interpret the limited available data, and
this was supplemented by the knowledge gained from across the canal. Figures 8.78 and 8.79
show an example determination of the value of Nkt = 12 for the soft gray foundation clay
(CH), and this matches with this same deposit on the east bank. Using the same methods, and
the same SHANSEP exponent λ = 0.75, Figure 8.80 shows the iterative processing of the
CPTU data to develop profiles of Su/p vs depth, and OCR vs depth for this clay unit. These
too match well with the east bank deposit data.
Figure 8.81 then presents our SHANSEPbased profiles of strength vs. depth (a)
beneath the crest, and (b) at the toe, along with the available strength and CPT data. The fit
with the available data is excellent.
Figure 8.82 shows the use of the correlation proposed by Mayne and Mitchell to
develop profiles of Su/P vs depth and OCR vs. depth within the “marsh” deposits overlying
the soft gray clays. This matches well with the CPTUbased interpreted OCR profile within
this stratum, and with the data from the east bank as well.
Figure 8.83 presents the resulting overall profiles of strength vs. depth within the
marsh deposits (a) beneath the crest, and (b) at the toe, along with all available data (including
CPT tip resistances interpreted using Nkt = 16. The thin layer of sensitive organic silty clay
was encountered in one boring, again at the approximate midpoint in the “marsh deposits,
and again closely overlain by leaves and twigs. This sample is shown in Figure 8.76.
Strengths for this thin layer were based on Su/P values from the east bank deposit. This thin
layer was not critical at this west bank site, as the sheetpiles penetrated well below this
sensitive layer and so forced a deeper, more rotational failure through the soft gray clays to be
the most critical mode.
835
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Once again, all shear strengths determined represented triaxial or vane shear strengths,
and these were reduced slightly (multiplied by a factor of 0.84) to develop shear strengths
suitable for the direct simple shear (DSS) dominated shear surfaces to be evaluated.
Figures 8.84 and 8.85 show the most critical failure surfaces (without gapping) for a
storm surge level of +9 feet (MSL) for failure (a) to the top of the soft gray clay, and (b)
within the lower marsh deposits. These both give low Factors of Safety, but the failure
through the lower marsh strata is the more critical case.
Figures 8.84 and 8.85 show these same two potential failure modes, again for a storm
surge elevation of +9 feet (MSL), but this time with an assumed waterfilled gap at the
outboard face of the sheetpile curtain. Once again the lower marsh units present the more
critical mechanism.
Figure 8.88 shows calculated Factors of Safety vs. canal water elevation for the failure
through the lower marsh stratum, both with and without gapping. The heavy red line in this
figure shows the best estimate of the likely critical failure path, based on these limit
equilibrium analyses. It is judged that gapping is most likely to be initiated at surge
elevations of approximately +10 to +11 feet (MSL) as the Factor of Safety (without gapping)
drops below about 1.25 to 1.35. Gapping was relatively unlikely during Katrina (max surge
level ~ +10 feet, MSL), and indeed no gap could be seen.
Based on these analyses, probabilities of failure were again estimated using the same
procedure as described previously in Section 8.3.7.1. Figure 8.88(a) shows distributions of
factor of safety as a function of rising canal water elevations for the “waterfilled gap” and the
“ungapped” cases, and Figure 8.88(b) shows the resulting estimated distributions of the
overall conjugate factor of safety for this west bank section.
Table 8.2 then presents the resulting estimated probabilities of failure vs. canal water
elevation. The probability of failure at the actual peak Katrina water elevation of
approximately + 10 feet (MSL) was low, but it was not negligible. Moreover, it would have
increased rapidly with even minor additional increase in canal water level. The probability of
failure becomes very high at the “design” water level of +12.5 feet (MSL).
It should also be noted that the marsh soils have likely been sheared (and thus
softened) a bit, and that the overall strength of this section was therefore likely somewhat
degraded by the loading it received during Katrina. Accordingly, it may not perform quite as
well in subsequent loading in the future.
This levee and floodwall section protects the large population of the still undamaged
Jefferson Parish. If the canal floodgate currently being installed, and future control of
pumping, cannot guarantee that canal water levels will never exceed about Elev. +5 to 6 feet
(MSL), then this section should be remediated.
836
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
8.3.8
The Breach Near the South End of the London Avenue Canal
A major breach occurred on the east bank, near the south end of the London Avenue
Canal, as shown in Figures 8.1 and 8.2. Figure 8.89 shows an oblique aerial view of this
breach under repair. The breach was approximately 80 feet in length, and it scoured to
significant depth. Sands eroded and transported by the inrushing floodwaters blanketed the
neighborhood inboard of the breach to considerable depth over a surprisingly wide area, as
shown for example in Figure 8.90.
Figures 8.91 and 8.92 show the floodwall sections at the south and north ends of the
breach, respectively. In these photos it can be seen that these wall sections have not displaced
(translated) laterally towards the inboard (“protected”) side; instead they have simply
“dropped” into the hole eroded by the scour of the breach flow.
Clearance for the footprint of the levee and floodwall was very limited, and the
neighboring homes and their back yards encroached closely on the levee. Levee maintenance
was very poor along this section, and numerous large trees had been allowed to grow along
the inboard toe. Many of these were actually rooted part way up the inboard slope face of the
levee embankment itself, as shown in Figure 8.93 which is a view looking north from the
breach location. These trees at the inboard toe represented an unacceptable risk as they can
be blown over by storm winds, creating sudden voids that represent favorable paths for
concentration of seepage flows and erosion in the critical toe area. Also, when they die the
rotting root system can leave voids that can pose a significant hazard with regard to seepage
and erosion in the critical inboard toe area.
Several large trees did topple at this site during Katrina, but in the absence of
eyewitnesses it is not possible to be certain if they toppled before the breach, or as a result of
erosion and scour after the breach opened. Figure 8.94 shows toppled trees at this site. Two
large trees from the levee toe area within the breach footprint toppled during this event.
This breach was much shorter in length than the large breaches at the 17th Street
Canal, the north end of the London Avenue Canal, and the southern breach on the IHNC at
the west side of the Ninth Ward (each of which were hundreds of feet in length.) Instead, like
the northern breach at the IHNC at the west end of the Ninth Ward, this was a narrow and
deep breach; suggesting that underseepage rather than foundation instability may have been
the key issue here.
As discussed previously in Chapter 4, the geology of the London Avenue canal differs
significantly from that of the north end of the 17th Street Canal. The buried sand “ridge” runs
laterally across the canal region, as shown in Figure 4.10 in Chapter 4, and relatively thick
sand strata occur at shallow depths in the London Avenue Canal (and the south Orleans
Canal) region. On the south side of this buried sand ridge, the sands tend to be dense as a
result of wave action and energy from the Gulf side. On the lee side (the north side), the
sands, especially at shallow depth, were protected and tend to be looser.
Figure 8.95 shows the locations of borings and CPT probes performed by the ILIT
investigation at this site. Figure 8.96 shows a crosssection through the breach, based on our
837
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
own (ILIT) data as well as IPET data and data available prior to Katrina. The embankment
has a modern (engineered fill) crown consisting of lightly compacted clay and silty clay,
underlain by older fill of more variable composition. The embankment section rests atop
variable “marsh’ deposits consisting primarily of variably interbedded clays and organics.
This “marsh” stratum is relatively thin, with a thickness of only 3 to 4 feet at the inboard toe,
and it is underlain by about 2 to 3 feet of soft gray clay (CH).
This thin surficial marsh and clay “crust” is underlain by deep deposits of medium
dense and then dense sands. In addition to the sheetpile curtain supporting the current
concrete floodwall, there is an older sheetpile curtain on the outboard side that used to support
a previous small floodwall at this location.
Strengths of the marsh deposits and the thin layer of underlying clay were determined
based on the available data, and the resulting strength characterizations are summarized in the
table within Figure 8.97, along with the estimated friction angles for the underlying sand
units. Stability analyses showed high factors of safety with regard to “landslide type
instability failure”, even for steady state seepage conditions at the maximum storm surge
height of approximately Elev. +9 feet (MSL). Figure 8.106 shows the most critical potential
slide surface for these worst case steady state seepage conditions. It was concluded that this
breach was unlikely to have resulted from conventional foundation stability failure.
Numerous analyses of seepage were performed, varying the horizontal and vertical
permeabilities of the various soil units and strata (in both the horizontal and vertical
directions) over ranges considered reasonable for these soils. For all reasonable ranges of
conditions, it was found the soils in the inboard toe area were vulnerable to erosion and
potential piping at storm surge levels of less than Elev. +9 feet (MSL).
An example is shown in Figure 8.98, which shows the flownet and flow velocity
vectors for a surge to Elev. +9 feet (MSL). Ranges of values of in situ permeability were
modeled for the sandy strata (in transient flow analyses), and it was concluded for reasonable
ranges of lateral permeabilitites that nearly full equilibration of pore pressures (greater than 90
to 95% equilibration) at the inboard side levee toe region would occur within 30 minutes or
less of outboard side canal water level rises. Given the rate at which the outboard side canal
waters rose (see Figure 8.18), steady state seepage analyses were considered to provide an
accurately (to slightly conservative) basis for assessment of underseepage pore pressures. The
analysis shown in Figure 8.98 thus represents steady state flow conditions.
Figure 8.99 is a closeup from this figure showing localized conditions in the vicinity
of the levee and floodwall. The sheetpiles are nowhere near deep enough to be effective in
reducing massive underseepage flows through the pervious sands, and exit gradients near the
inboard toe are unsafe with regard to erosion and the initiation of potential piping.
Figure 8.100 shows pore pressure contours from this same flow analysis. Hydraulic
uplift forces at and just inboard of the toe exceed the weight of soil overburden, suggesting
the possibility that hydraulic uplift ruptured the less pervious thin clay and marsh crust
causing a “blowout” failure in this toe area.
838
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.101 shows hydraulic gradients for this same flow analysis. The exit gradients
at the inboard toe are on the order of io ᄃ 0.5, representing a factor of safety with respect to
erosion of approximately
FS =
γb / ( io • γw )
where γb is the buoyant unit weight of soil, γw is the unit weight of water, and io is the exit
gradient. For the lightweight marsh soils, with light buoyant unit weights, the calculated
factor of safety is on the order of FS ᄃ 0.8 to 1.05 for the conditions shown in Figure 8.101.
Any “bunching” or localized constriction of the flownet near the exiting face would further
exacerbate the tendency to initiate erosion and the beginning of piping. Given the high
variability of the thin surficial marsh deposits that “cap” this site, erosion and piping are
highly under these conditions.
Figures 8.101 through 8.105 illustrate how such erosion can rapidly escalate as the
flownet converges on even a slight void (Figure 8.102) to rapidly increase the localized exit
gradient and accelerate the erosion process (as occurs progressively as the erosion enlarges
the hole at the inboard toe in Figures 8.103 through 8.105.) This is actually a three
dimensional process, so the rate of acceleration of this erosion and “piping” process is
actually more severe than can be properly illustrated in these twodimensional figures.
Figure 8.107 is a schematic illustration of this process. As the flownet increasingly
converges, and erosion continues to accelerate, and the erosion literally tries to “tunnel” back
under the levee embankment. This produces slumping and periodic collapses into the opening
void, and the process continues to accelerate until the crest is finally breached, at which point
the inrushing flows rapidly further scour the breach.
An additional possibility is that this type of erosion process may have been
exacerbated by the toppling of a tree near the levee toe, as illustrated schematically in Figure
8.108. Flow towards the toe (and the trees rootball zone) weakens the ground and thus
weakens the tree’s resistance to pullout failure under storm wind loading. Many trees toppled
in this manner during the hurricane. If the tree near the toe topples, it created a large void
toward which the exiting flownet would rapidly converge, initiating or greatly accelerating the
type of erosion and piping process described above.
Figure 8.109 shows another view of this breach section, this time from the waterside
and in late September of 2005. In this photo it can be clearly seen the breach is a very narrow
feature, deeper at the north end (to the left in this view). On the inboard side our field team
felt that the evidence suggested that the breach initiated either as a seepage erosion “blowout”
or similar near the north end of the feature. There was a large tree that was uprooted at that
location, but it could not be determined whether the tree fell before or after (as a result of) this
failure and breach.
In the end, this breach scoured to significant depth and was then rapidly buried by the
emergency embankment repair section, so there is no conclusive evidence left with which to
determine which of the above described possible mechanisms (in detail) caused the actual
839
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
failure. It is apparent, however, that this failure was the result of underseepage and erosion of
some form. The lack of sufficient sheetpile depth as to adequately reduce underseepage flows
and toe exit gradients was an engineering lapse, and so was allowing the rampant growth of
large trees in the inboard toe area.
The original design analyses for this section were performed by an outsourced
engineering consultant, and were reviewed by the USACE (USACE; DM19A.) In these
analyses, the canalside phreatic level was taken at the full design level (Elev. +12 feet,
MSL), and the phreatic level at the inboard side levee toe was taken at Elev. 5 feet (MSL).
Based on our investigation’s transient flow analyses, for reasonable ranges of in situ lateral
permeability, for the full (design) canal water elevation of +12 feet (MSL), the phreatic level
at the inboard side levee toe due to underseepage would have actually been on the order of +2
to +5 feet (MSL). This represents a large increase in underseepageinduced uplift pore
pressures and exit gradients, and is the principal difference between the preKatrina “design”
analyses and our investigation’s postKatrina forensic analyses at this section.
8.3.9
The Breach and Distressed Sections Near the North End of the London Avenue Canal
An additional major breach occurred on the west bank near the north end of the
London Avenue Canal, as shown in Figures 8.1 and 8.2. This too was a catastrophic breach
as it rapidly scoured below mean sea level and so was one of the three large drainage canal
breaches that continued to push water into downtown New Orleans for three days after
Hurricane Katrina’s passage.
Figure 8.110 shows an aerial view of the breach on the west bank. There was also a
“distressed” section on the opposite side (on the east bank) that represents an incipient failure
in progress; this failure was arrested in a partially developed state by the failure of the west
bank section (which drew down the water level and thus saved the east bank.)
Figure 8.111 shows a view looking south along the canal, with the emergency repair
embankment section on the west bank on the right, and the incipient failure section on the left
side. If one looks closely, the floodwall on the left (east) side can be seen to be leaning away
from the canal in this photo.
This was one of the most challenging sites for our investigation. Foundation soil
conditions, and embankment and floodwall geometries, were similar on both sides of the
canal. One side failed catastrophically, and the other appears to have begun to fail but to have
been saved by the failure on the opposite bank. It was a challenge to develop a model that
would predict the failure of the west bank before the east bank failure was able to fully
develop. There are also a variety of data and evidence suggestive of a number of potential
failure and distress modes evident at both sites (both sides of the canal), and sorting through
these posed a significant challenge as well.
Figure 8.112 shows a view of the main breach on the west bank, taken from the south
end of the breach on the outboard (water) side. In this photo it can be clearly seen that the
waterside toe section of the earthen embankment is still in place, and that the
840
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
floodwall/sheetpile curtain and the inboard side of the earthen levee have been separated from
it and pushed to the inboard side.
Figure 8.113 shows conditions at the inboard toe of the failed embankment section on
the west shoreline. The small clubhouse shown had originally been at the same elevation as
the nearly adjacent house, but was lifted nearly 7 feet vertically by the displacements during
the failure. Some initial field investigators suggested that this was evidence of rotational
movement, but our investigation found that this clubhouse (and the ground upon which it
stood) was raised vertically by heave due to “plowing” as the main levee embankment
displaced laterally (without rotation.) The confined uplift region, and its “humped” nature,
are clearly evident beneath the small clubhouse in this photo.
Figure 8.114 shows a view of the inboard toe of the “distressed” (displaced)
embankment and floodwall section on the east shoreline, taken on the outboard (water) side.
As shown in this photo, the concrete floodwall leaned away from the canal, and a gap with a
maximum width of 2.5 feet (and a common width of 1.5 to 6 feet) opened between the
outboard side of the earthen levee embankment and the concrete floodwall (and its supporting
sheetpile curtain.)
Figure 8.115 shows the other side of this same floodwall section. As shown in this
photo, the displaced floodwall leaned to the inboard with a readily discernable tilt of up to 8º.
The next photo, Figure 8.116, shows conditions along the inboard base of the floodwall (at the
feet of the photographer who took the photo of Figure 8.115.) A series of apparent
“sinkholes” occurred along the inboard side contact between the concrete floodwall and the
crest of the earthen levee at this location.
Figure 8.117 shows conditions at the inboard toe immediately below the sinkholes of
Figure 8.116. A prominent sand boil feature, with sandy ejecta, occurred at this location.
Less apparent, but important, was the hummocky wrinkling of the nearly level ground inboard
of the toe of the levee, and the slight overthrust feature adjacent to the sand boil. This
overthrust feature was apparently missed by many field investigators, but our team noted it
and went back and excavated it during our subsequent field boring, sampling and CPT
program and found that it was indeed the toe thrust of the beginning of a translational
instability feature.
Figure 8.118(a) shows a crosssection through the west side breach prior to Katrina,
and Figure 8.118(b) shows this same section after the failure. The failure on the west side
was a translational failure of the embankment, sliding along the interface between the
foundation sands and the overlying less pervious layer of silty clay (CL/ML).
Figure 8.119(a) shows a crosssection through the east side “distressed” section prior
to Katrina, and Figure 8.119(b) shows this same section after the hurricane. The displacement
and tilting of the floodwall was the result of the initiation of slippage, once again at the
interface between the foundation sands and the overlying less pervious layer of silty clay.
Unlike the west bank, this slippage progressed only enough to produce displacements of
approximately 1.5 to 2.5 feet, whereupon these movements were arrested as the failure and
breaching on the opposite bank rapidly drew down the canal water level and reduced the
lateral push against the sheetpile curtain and floodwall.
841
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.120 shows a plan view of both sides of the canal, indicating the locations of
the borings and CPT performed as part of this investigation.
Figure 8.121 shows the longitudinal subsurface soil profile developed along this
section of levee on the west bank during the original design studies, and Figure 8.122 shows
the reinterpretation of this section by this study team based on the original boring data.
Figures 8.123 and 8.124 show the same pairing of profiles for the east bank side.
Processing of the available geotechnical data was performed using essentially the
same methods and procedures as were described in detail in the preceeding sections, and
much of the detail will be omitted here in the interest of brevity.
Figures 8.125 and 8.126 show the best estimated profiles of strength vs. depth and
Su/P vs depth on the west bank (breach) side for profiles (a) beneath the full levee
embankment overburden, and (b) inboard of the levee toe.
Figure 8.127 shows estimated friction angles across the transition from the base of the
silty clay stratum (CL/ML) into the underlying clayey sands and sands. Friction angles were
estimated from the CPT data using two correlations, and they were also estimated based on
the SPT data available from the borings. Also shown are the results of two direct shear tests
performed on “undisturbed” samples as part of these studies.
Figure 8.109(a) shows an “undisturbed” sample from the transition across the silty
clay into the underlying sands. As shown in this figure, this transition was semigradational
rather than abrupt. The base of the silty clay layer is underlain by fine clayey sands with
variable fines content. Near the contact the fines content is high enough that the clayey fines
dominate the shear strength behavior. The fines content rapidly decreases over the next 6
inches or so, and eventually the fines content of the remainder of the layer remains relatively
stable at between 5% to 10%. The green line in this figure represents our best estimate of the
approximate operative effective friction angle through this zone.
It was not possible to discern with certainty the elevation to which pore pressures
arising from underseepage passing beneath the sheetpile curtain through the more open,
pervious sands at depth due to the transient rising storm surge penetrated (vertically) upwards
into this transition zone. Accordingly, various combinations of partial pore pressure
development may be postulated at different elevations across this transition, and these may be
paired with various effective friction angles to evaluate the shear strength within this narrow,
and critical zone.
Several combinations were postulated and analyzed in these studies. Higher (more
completely penetrating) pore pressures more nearly approaching steady state flow are clearly
appropriate at the base of this transition zone, and these would be paired with friction angles
on the order of Ǿ ᄃ 30 to 32º. A few inches higher in the transition zone the effective friction
angle would be somewhat lower, but this would be offset by reduced penetration of pore
pressures, resulting in largely similar estimates of resultant frictional shear strength. In the
end, an effective friction angle of 31º was selected, and this was coupled with assumed rapid
842
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
development of steady state pore pressures as the storm surge rose. (For reasonable ranges of
in situ permeability of the deeper, more open and pervious sands and with reasonable ranges
of specific storage for these initially saturated deposits; pore pressure development at the
inboard side toe region within the pervious deeper sands was approximately 65 to 90%
developed within two hours of outboard side (canal) water level increases.)
Figures 8.128 through 8.130 show the same sequence of figures, this time for
conditions on the east bank (distressed) side of the canal. Once again the transition between
the silty clay and the underlying clayey sand is the critical region. As with the west bank, an
effective friction angle of 31º was selected for analysis, and this was coupled with assumed
rapid development of full steady state underseepage as the storm surge rose within the canal.
Figure 8.131 shows the analysis crosssection and principal soil properties modeled
for analysis of the west bank breach site. Analyses were performed using both finite element
analysis methods (again using the program PLAXIS) and limit equilibrium methods
(Spencer’s Method).
Figure 8.132 shows normalized shear strain contours for the west bank (breach)
section at a storm surge level of Elev. +9 feet (MSL). Gapping initiated at the outboard side
of the floodwall and its supporting sheetpile curtain initiated in this analysis at a canal surge
elevation of between +7 to +8 feet (MSL), and was fully developed by a surge elevation of +
9 feet, as shown in this figure.
Figure 8.133 shows normalized shear strain contours for the east bank (distressed)
section, this time for a slightly higher surge to elevation +10 feet (MSL). This the upper
bound estimate of the surge elevations achieved during Katrina. Gapping developed in this
east bank section at a surge elevation of between +7 to +8 feet, and was fully developed by a
surge elevation of +9.5 feet (MSL).
These conditions produce a predicted failure of the west (breach) side at a surge
elevation of approximately +9.5 feet (MSL) in these PLAXIS analyses, and the east side
displaces a bit (with associated lateral displacement and tipping of the floodwall) but remains
barely stable to a surge elevation of +10 feet (MSL).
Figures 8.134 and 8.135 show a simultaneous analysis of both sides of the canal, and
the predicted (“best estimated” properties and flow) conditions for a storm surge to elevation
+9 feet (MSL). Figure 8.134 shows normalized shear strain contours, and Figure 8.135 shows
the associated predicted deformations and displacements. The west side has failed
catastrophically, and the east side section is “distressed” (with lateral displacements of
approximately 2 to 3 feet and some tilting of the floodwall. This closely matches the field
observations.
Figure 8.136 shows the associated PLAXISbased prediction of the critical path to
failure for each side of the canal. Once gapping occurs, the extra “push” of the water in the
gap is sufficient to destabilize the west bank at a surge height of approximately +9 to +9.5 feet
(MSL), but the east bank section remains barely stable until a surge height of +10 feet (the
upper bound of the estimated surge height that actually occurred).
843
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figures 8.137 through 8.159 repeat these same analyses, this time using classic
seepage analyses to predict pore pressures and gradients resulting from the underseepage
flows as the storm surge rises, and limit equilibrium analyses (Spencer’s Method) coupled
with these predicted pore pressure and gradient conditions to evaluate overall stability for
both sides of the embankment. Once again, rapid development of essentially full steady state
underseepage was assumed, and an effective friction angle of 31º was modeled at the interface
between the silty clay and the underlying clayey sand.
Figures 8.142 and 8.146 show the most critical failure surfaces on the west bank
(breach) side for a surge elevation of +9 feet (MSL), with and without gapping respectively.
Figures 8.153 and 8.157 show the same two cases for the east bank (distressed) side, again for
a surge height of +9 feet (MSL).
Figure 8.159 summarizes the results of these limit equilibrium analyses for both sides
of the canal, and the heavy red lines show the estimated most critical paths to failure. Once
again the west bank side fails at a surge height of slightly less than +9 feet, but the east bank
(distressed) side remains barely stable at this surge elevation. The blue horizontal dashed line
in this figure represents our investigation team’s best estimated surge elevation in the canal at
the time of the breach and failure of the west bank section.
These analyses show that the observed behaviors were not the result of underseepage
and resultant piping erosion. The behaviors on both sides of the canal were, instead, the result
of lateral translational instability (and incipient instability), with the critical potential failure
mode on both banks being lateral translational sliding on the interface between the silty clay
and the underlying clayey sands. This sliding was made possible by the high porewater
pressures in the foundation soils at and near the base of the inboardside levee toe due to
underseepage.
This exactly fits with the observed field data. The “sinkholes” at the crest of the
embankment on the east side were the result of tilting of the slightly displaced floodwall, and
the resulting opening of a gap between the floodwall and the embankment into which
embankment soils could fall. This correlates with the observation the “sinkhole features”
were all narrow, and were all parallel and adjacent to the floodwall (see Figures 8.114 and
8.117.)
8.3.10 Summary and Findings
A large number of critical errors and poor judgements jointly contributed to the
catastrophic failures that occurred along the drainage canals. There were conceptual errors in
the layout and fundamental design of the levees and floodwalls, there were policy and funding
issues that greatly reduced the level of safety of the overall system, and there were
engineering errors in the analysis and design of individual sections.
No one organization, agency or group of individuals had a monopoly on their
contribution to this disaster. Federal government (including the Congress), the Corps of
844
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Engineers, local government and local oversight agencies (including the local Levee Board
and the local Water and Sewerage Board), and outsourced engineering firms all contributed.
The resulting system failed catastrophically, and at multiple locations. And it failed at
significantly less than the intended levels of “design” (storm surge) loading. Moreover, it is
clear that additional sections were saved from failure only by the catastrophic failures of
nearby breaches, which drew down the water levels and so reduced the loading on additional
potentially unstable levee and floodwall sections.
The results of these failures were catastrophic. The vast majority (approximately
80%) of the eventual floodwaters that flowed into the main Orleans East Bank (downtown)
protected area came through the breaches in the drainage canals. These flows overfilled the
subbasin north of the Metairie Ridge, and then crossed this ridge and flowed into the
southern areas as well where they greatly exacerbated flooding that had already occurred as a
result of overtopping and failures of levees and floodwalls along the west side of the IHNC.
In the absence of the drainage canal failures there would still have been localized flooding and
damage near the IHNC, but this would have been minor relative to the eventual damages that
resulted when the canal breaches filled a majority of the overall basin.
The localized flooding near the IHNC would have posed relatively little threat of loss
of life; the damages would have been (relatively) limited and the floodwaters could have been
pumped out in a matter of days. Instead, roughly half of the 1,293 fatalities (to date)
attributed to flooding of the New Orleans region occurred in the Orleans East Bank
(downtown) protected basin, and a roughly similar fraction of the devastating regional
economic damages as well.
The following is a listing of critical errors and poor judgements and decisions that
contributed significantly to the poor performance of the drainage canal levees and floodwalls
during Hurricane Katrina:
1. The decision not to install floodgates at the north ends of the three drainage canals to
prevent uncontrolled water level rise due to storm surge within the canals was largely
the result of poor interaction between the local Levee Board and the local Water and
Sewerage Board, and their inability to resolve their differences in the interests of the
greater Public good (and safety). Lawsuits by environmentalists aginst this system
also worked against the floodgates. As a result, the canals remained open to storm
surges; essentially inviting the enemy (storm surge) into a poorly protected section of
the interior of the protected ring around metropolitan New Orleans.
2. The decision not to purchase additional land (right of way) to permit widening of the
levees required that the system be extended vertically without allowing provision of
additional levee width and mass with which to resist the increased floodwater forces
associated with the increased height. Shortterm savings here resulted in tens of
billions of dollars in losses.
3. Similarly, the failure to garner access and control of property at the inboard (protected
side) toe of the levees prevented full and proper inspection of this critical area. It also
845
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
led to unacceptable risk associated with growth of trees on the inboard side levee
slopes and toes, and the literal undermining of levee toes by excavation of inground
swimming pools in this critical inboard toe area.
4. The designers failed to take advantage of critical lessons from an expensive and well
directed research program that involved construction of a fullscale model levee and
floodwall on nearly identical foundation soils in the nearby Atchafalya basin. This
model was loaded to failure, and the failure mode observed involved opening of a gap
on the outboard side of the floodwall, water entering into the gap, and subsequent
pressures on the floodwall and sheetpiles pushing the inboard side section of the
earthen embankment sideways (the “cut the cake in half and slide it” failure mode).
This failure mode was neglected in the subsequent design of the levees and floodwalls
lining the canals, and at least two of the catastrophic failures (breaches), and two
additional “incipient” failures were the result of this failure mechanism.
5. The designers also failed to take account of the influence of stress history and
effective overburden stresses on the strengths of the foundation soils beneath a number
of the embankments. Furthermore, they deviated from USACE policy by using
average shear strengths (not strengths slightly lesser than the average data), and by
“averaging” strengths across lateral distances that were too large. These errors and
shortcomings in the determination and selection of soil shear strength parameters
played a critical role in the catastrophic failure of the east bank near the north end of
the 17th Street canal.
6. Optimistic assumptions, and misinterpretation of two field tests, led to the assumption
that system permeability was low enough that underseepage would not be a critical
issue during “transient” (shortterm) rises in canal water levels during hurricane
induced storm surges. This was a critical error, and it resulted in inadequate sheetpile
lengths throughout the drainage canals (especially the London Avenue Canal), and
along the IHNC. These sheetpile curtains routinely extend to insufficient depths as to
adequately “cut off” underseepage flows, and the resulting underseepage flows were
principal contributors to the catastrophic failures observed at both of the major breach
sections on the London Avenue canal. These inadequate cutoffs continue to be a
potentially critical issue at other sections that did not (yet) breach during hurricane
Katrina, and they appear to have been a principal factor in the two massive breaches
on the east bank of the IHNC (at the edge of the Lower Ninth Ward; see Chapter 6) as
well.
7. Insufficient site investigation was performed for the design of these critical systems
protecting a major metropolitan population. Given the difficult and complex
foundation soil conditions, additional borings and testing would have represented a
very modest incremental expenditure, and would have greatly improved the
information available as a basis for analysis and design of these important sections.
8. Errors and poor judgements were made in engineering analysis and design of these
sections. Soil properties were extrapolated laterally over inappropriately large
distances, and without adjustment for the resulting uncertainties. Archaic analysis
846
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
techniques were employed (the Method of Planes), and projectspecific research (a full
scale test embankment and floodwall in the nearby Atchafalya basin) was ignored,
resulting in failure to analyze the failure mode (“cut the cake in half and slide it”) that
proved critical for at least two of the catastrophic drainage canal breaches, and likely
also for the two massive breaches at the east side of the IHNC (adjacent to the Ninth
Ward.) This mode was also evident at an additional “incipient” failure section on the
London Avenue canal that was saved from failure, by the failure of the even weaker
section on the opposite shoreline (which immediately drew down the local water
levels.) The stability of the entire canal system should be considered potentially
suspect until it can be properly reevaluated with regard to this potentially critical
mechanism.
9. Design review was inadequate. Errors and questionable judgements that would have
been expected to be caught and challenged by a properly convened independent
external review panel went unchallenged. On one occasion when reviewers from the
USACE Division level in Vicksburg did catch and challenge such issues, they were
rebuffed by the local District Chief who declared those issues to be a matter of
“judgement”.
10. The Factor of Safety (FS) used for design of these vital levees and floodwalls was set
at only FS = 1.3 for the case of “transient” storm surge loading. As discussed in detail
in Section 8.3.7.1 this is inappropriately low for systems critical for the safety of large
populations, and for the difficult and challenging foundation soils conditions of the
region. This issue is discussed in significantly more detail in Chapters 11 and 12.
11. Congressional funding (appropriations) were problematic. Funding was irregular and
somewhat unpredictable, representing a difficult basis for design and construction of a
system intended to be contiguous (seamless) and to protect a large metropolitan
population. Strategic decisions, and conceptual design, were often driven by a need
for frugality. In addition, when appropriations did arrive, some elements of the system
had to be further streamlined for economy. The relatively minor savings achieved
now pale in comparison to the many tens of billions of dollars in losses that ensued.
12. Pace of funding was also problematic. At the time of Katrina’s arrival, the flood
protection system in the canal district was still incomplete…. fully 51 years after the
flooding from Hurricane Betsy that inspired the inception of construction of the
improved flood protection system. Three of the bridges across the drainage canals still
had not yet had their side walls raised, so three “holes” remained in an otherwise
contiguous system. These “holes” at the bridges were not critical during Hurricane
Katrina only because: (a) the storm surge was less than the full design load case, and
(b) catastrophic nearby breaches (failures) occurred. (An additional “hole” in the
system, at the south end of the Orleans Canal, was yet another result of dysfunctional
interactions between the local levee board and the local water and sewerage board (as
discussed previously in item #1 above.)
Many of the issues above, from conceptual design issues through engineering analysis
details and even selection of appropriate Factors of Safety would have been expected to be
847
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
challenged by a properly convened independent review panel. Unfortunately, in the current
system with myriad local interests and no strong local entity able to convene appropriate
levels of unbiased expert review capability, this critical element was absent during the design
and construction of these important flood protection system elements.
In addition, there was a lack of centralized authority, and of clear areas of
responsibility. Involvement of a significant “local” institutional presence of significant
stature and resources was lacking. The local levee board lacked the resources and funding to
mount serious review of the Federal plans and designs, and the mandate to challenge
problems that should have been apparent at early stages.
In the end, the performance of the flood protection system along the three drainage
canals was unacceptable, and resulted in catastrophic loss of life and property throughout a
major metropolitan region.
8.4 References
Burk & Associates, Inc (1986) Geotechnical Investigation of London Avenue Outfall Canal,
OLB Project No 20490269, Vol I and II, prepared for the Board of Levee
Commissioners of the Orleans Levee District, New Orleans, Louisiana.
Foott, R and Ladd, C.C (1977) “Behavior of Atchafalaya levees during construction”,
Geotechnique,27, No2, pp 137160.
GEOSLOPE/W (2004) “Complete Set of Manuals”, John Krahn (Edit.), Calgary, Alberta,
Canada.
IPET – Interagency Performance Evaluation Task Force (2006) Performance Evaluation
Status and Interim Results, Report 1 of a Series, Performance Evaluation of the New
Orleans and Southeast Lousiana Hurricane Protection System, January 10, US Army
Corps of Engineers.
IPET – Interagency Performance Evaluation Task Force (2006) Performance Evaluation
Status and Interim Results, Report 2 of a Series, Performance Evaluation of the New
Orleans and Southeast Lousiana Hurricane Protection System, March 10, US Army
Corps of Engineers.
Jackson, R. B., (1988), “E99 Sheet Pile Wall Field Load Test Report”, Technical report No.1,
US Army Engineer Division, Lower Mississippi Valley, Vicksburg, MS
Jamiolkoski, M, Ladd, C.C., Germaine, J.T., and Lancellotta, R. (1985) “New developments
in field and laboratory testing of soils.” Proc. 11th International Conf. on Soil
Mechanics and Foundation Eng., San Francisco, 1, 57154.
Karlsrud, K., Lunne, T and Brattlien (1996) “Improved CPTU interpretations based on block
samples”, Nordic Geotechnical Conference, 12 Reykjavik 1996. Proc, Vol.1, pp 195201.
848
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Kaufman, R and Weaver, F (1967) “Stability of Atchafalaya Levees”, Journal of the Soil
Mechanics and Foundations Division, ASCE, SM 4, pp 157 176.
Kemp, P., and Mashriqui, H. (2006). Personal Communication
Lacasse, S., Ladd, C.C., and Baligh, M.M. (1978) “Evaluation of field vane, Dutch cone
penetrometer and piezometer probe testing devices.” Res. Report R7826, Dept. of
Civil Eng., MIT, 375 pp.
Ladd, C.C and DeGroot, D. 2003. “Recommended Practice for Soft Ground Site
Characterization”, Arthur Casagrande Lecture, 12th Panamerican Conference on Soil
Mechanics and Geotechnical Engineering, June, Boston.
Lunne, T., Lacasse, S., and Rad, N.S., (1989) “General report: SPT, CPT, PMT, and recent
developments in insitu testing”, Proceedings, 12th International Conference on
soilmechanics and foundation engineering, Vol.4, Rio de Janeiro, pp. 23392403
Mashriqui, H., (2006), Personal Communication
Mayne, P,W and Mitchell, J.K (1988) “Profiling of Overconsolidation Ratio in Clays by Field
Vane”, Canadian Geotechnical Journal, Vol 25, No 1, February, pp 150157.
Oner, M., Dawkins, W.P., and Hallal, I., (1997) “Soil Structure Interaction Effects in
Floodwalls”, Electronic Journal of Geotechnical Engineering
Oner, M., Dawkins, W.P., and Mosher, R., (1997) “Shearing Method for Soil Structure
Interaction Analysis in Floodwalls”, Electronic Journal of Geotechnical Engineering
Robertson, R.K, and Campanella, R.G. (1983) “Interpretation of Cone Penetration Tests. Part
I: Sand”, Canadian Geotechnical Journal, Vol 20, No 4, Nov, pp 718733.
Seed, R.B. (2005) “CE 270B Class Notes”, University of California, Berkeley.
PLAXIS Finite Element Code for Soil and Rock Analyses (2004). “Complete Set of
Manuals”, V8.2 , Brinkgreve yVermeer (Edit.), Balkema, Rotterdam, Brookfield.
USCE (1989) London Avenue Outfall Canal, Design Memorandum No 19A General Design,
Vol 1 and Vol 2, Louisiana, Department of the Army, New Orleans District, Corps of
Engineers, New Orleans, Louisiana.
USCE (1990) Orleans Parish (Jefferson Parish) 17th St. Outfall Canal, Metairie Relief, Design
Memorandum No 20 General Design, Vol 1 and Vol 2, Louisiana, Department of the
Army, New Orleans District, Corps of Engineers, New Orleans, Louisiana.
Van Heerden, I., (2006), Personal Communication
849
Independent Levee
Investigation Team
850
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Figure 8.1: Map showing principal features of the main flood protection rings or “protected areas” in the New Orleans area.
[Modified after USACE, 2005]
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
: Levee Breach
: Distressed Levee Section
The Canal District
CSX Railroad
Breach
Metarie Ridge
Sites
Orleans Canal
The French Quarter
Mississippi River
Garden District
Downtown New Orleans
Figure 8.2: Plan View of the Orleans East Bank (Metro) protected area showing approximate
depth of flooding on September 2, 2005.
[Modified after Mashriqui, 2006]
851
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Source: IPET Interim Report No. 2; April, 2006
Figure 8.3: Hydrographs showing measured (and photographed) water levels at gage stations
along the Inner Harbor Navigation Channel (IHNC).
Photograph by Rune Storesund
Figure 8.4: Concrete storm gate at the west bank of the CSX Railroad crossing of the IHNC,
after Hurricane Katrina, showing the steel floodgate missing.
852
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph by Rune Storesund
Figure 8.5: View from the inboard side of the CSX Railroad breach site from Figure 8.4
showing scour of the roadway fill adjacent to the railway embankment fill.
Photograph by Rune Storesund
Figure 8.6: The Iwall section breach behind the Port of New Orleans, showing floodwall
failure.
853
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph by Rune Storesund
Figure 8.7: Second view of the Iwall failure at the Port of New Orleans showing the trench
scoured by overtopping flows.
854
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.8: Deformed mesh of Iwall failure at the IHNC behind the Port of New Orleans.
Relative
Shear Strain
Figure 8.9: Finite element analysis of Iwall failure at the IHNC behind the Port of New
Orleans, showing shear strain contours at point of wall failure.
855
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Source: USACE
Figure 8.10(a): Typical reinforced concrete Iwall atop a sheetpile curtain.
Source: USACE
Figure 8.10(b): Typical reinforced concrete Twall atop a sheetpile curtain.
856
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph by Rune Storesund
Figure 8.11: Concrete floodwall and steel floodgate structure at railroad yard behind the
Port of New Orleans.
Photograph by Rune Storesund
Figure 8.12: Erosion, gapping and offset of adjacent wall sections at the east end “transition” of
the concrete floodwall and gate structure from Figure 8.11.
857
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph by Rune Storesund
Figure 8.13: Erosional breach behind the southern end of the Port of New Orleans at contact
between embankment section and structural concrete monolith.
Photograph by Rune Storesund
Figure 8.14: Closeup view of the highly erodeable shell sand fill at the breach shown above
in Figure 8.12.
858
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph by Rune Storesund
Figure 8.15: View of second large erosional breach behind the southern end of the Port of
New Orleans.
859
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Source: IPET Report No. 2, April, 2006
Figure 8.16: Approximate contours of maximum storm surge heights along the southern end
of Lake Pontchartrain. [Note: Elevations in Feet, NGVD29]
860
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Source:
IPET Report No. 2; April, 2006
Figure 8.17: High water marks near the mouth of the 17th Street Canal.
Source: IPET Report No. 2, April, 2006
Figure 8.18: Hydrograph proposed by IPET for the 17th Street drainage canal.
861
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
West Bank Floodwall
Pump House Building
Spillway Section Across the “Gap”
End of East Bank Floodwall
Photograph by Rune Storesund
Figure 8.19: View of the pump house, levees and floodwalls, and the “gap” at the south end
of the Orleans canal.
Elev. ᄃ + 13.2 feet (MSL)
Elev. ᄃ + 6.5 feet (MSL)
Elev. ᄃ + 6.8 feet
Elev. ᄃ + 12.4 feet (MSL)
Photograph by Rune Storesund
Figure 8.20: Side view of the “gap” at the south end of the Orleans canal.
862
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Pump House
South end
of spillway
Spillway
North end
of spillway
Figure 8.21: Oblique view of LIDAR survey of the southern end of the Orleans Canal,
showing the “gap” on the east side where the levee and floodwall that
should have sealed the end are omitted, and a “spillway” replaces them.
863
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph by Rune Storesund
Figure 8.22: View of the earthen levee embankments and floodwalls along the east side of the
Orleans Canal (view looking to the North.)
Photograph by Rune Storesund
Figure 8.23: View of the earthen levee embankments and floodwalls along the west side of
the Orleans Canal (view looking to the North.)
864
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.24: Oblique aerial view of the breach at the 17th Street Canal.
865
Figure 8.25: Plan view for the 17th Street Canal showing approximate ILIT boring and CPT locations
Independent Levee
Investigation Team
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
866
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Figure 25(a): Plan view of 17th Street Canal showing locations of IPET borings and CPT
Independent Levee
Investigation Team
867
Independent Levee
Investigation Team
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Figure 8.26: Cross-section through the 17th Street Canal breach showing conditions (a) before the hurricane, and (b) after the breach and failure.
8-68
Independent Levee
Investigation Team
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
869
Figure 8.27: Crosssections through the 17th Street Canal breach showing conditions as: (a) prestorm surge, (b) a gap opens between
the outboard side of the earthen embankment and the sheetpile curtain (and Iwall), (c) the elevated water levels increase within this
gap pushing the Iwall and the inboard side of the embankment laterally away from the canal, and (d) final configuration at the end
of displacement after Iwall topples backwards towards the canal.
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.28: View of the 17th Street Canal breach site showing the tops of the Iwall sections
that “pushed” (and followed) the displaced levee embankment section, and
then toppled backwards towards the canal.
Figure 8.29: View of the 17th Street canal breach site showing the Iwall sections at the south
end of the breach toppled forward (to the inboard side) by the inrushing flood
waters. Note the severe scour inboard of the failure at this location.
870
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
6914 Bellaire Dr.
Collapsed Shed
Figure 8.30: Side view of the collapsed shed pushed into the house at 6914 Belaire Drive as
illustrated in the crosssection of Figure 8.27.
Peat blocks
Photograph by Joseph Wartman
Figure 8.31: View of the foundation slab shown in the crosssection from Figure 8.27.
871
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.32: View of the exit of the upper failure surface (shear surface No.1 from Figure
8.27) at the inboard toe of the laterally translated embankment (toe scarp No.
1 in Figures 8.26 and 8.27).
[IPET Interim Report No. 2, April 1, 2006]
872
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.33: View of toe scarps No’s. 2 and 3 at the inboard edge of the lateral translational
failure at the 17th Street drainage canal.
Figure 8.34: Additional view of toe scarps No’s. 2 and 3 at the inboard edge of the lateral
translational failure at the 17th Street drainage canal.
873
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.35: Two additional crosssections through the breach section on the east bank of the
17th Street canal showing: (a) CrossSection BB′ and (b) CrossSection CC′.
874
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Leveee
Investigat
Investigation
ion Team
17th Str. Canal, East Bank, Levee Crest
Su (psf)
0
500
1000
1500
2000
2500
3000
4
Model
0
Engineering Fill (CL)
-4
-8
Non-engineering Fill (CH)
Elevation (ft)
-12
-16
-20
-24
-28
UUTX (Kaufman et al, 1967)
-32
UU-TX
Lab Vane (peak)
-36
17-CPT-3A, Nk=12
17-CPT-4A, Nk=12
-40
Figure 8.36: Summary
Summary of shear strength data within the 17th Street drainage canal levee
em
embankm
bankmeent fill at and near the
the east
east bank breach section.
875
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
qc (tsf)
0
5
fs (tsf)
15
10
20
0
5
0
0.5
1
Δu (tsf)
Rf (%)
1.5
2
0
5
10
15
20
0
5
5
5
0
0
0
-5
-5
-5
-10
-10
-10
-15
-15
-15
-20
-20
-20
-25
-25
-25
0.5
1
Fill
-5
-15
-20
Marsh
Deposits
Intermixing Zone
Grey Clay
(CH)
-30
Elevation (ft)
Elevation (ft)
-25
-30
-30
Elevation (ft)
-10
Elevation (ft)
Grey Clay
(CL)
-30
-35
-35
-35
-35
-40
-40
-40
-40
Uo
-45
-45
-45
-45
-50
-50
-50
-50
-55
-55
-55
-55
-60
-60
-60
-60
-65
-65
-65
-65
Example of CPT test data for locations beneath the levee embankment at and near the 17th Street canal breach section.
Figure 8.37:
876
1.5
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
qc (tsf)
0
5
10
fs (tsf)
15
20
0
0.5
1
Δu (tsf)
Rf (%)
1.5
2
0
5
10
15
20
0
5
5
5
5
0
0
0
0
-5
-5
-5
-5
-10
-10
-10
-15
-15
-15
-15
-20
-20
-20
-20
-25
-25
-25
-10
Grey Clay
(CL)
Marsh Deposits
0.5
1
-30
-30
-30
Elevation (ft)
Grey Clay
(CH)
Elevation (ft)
Elevation (ft)
-25
Elevation (ft)
Intermixing Zone
-30
-35
-35
-35
-35
-40
-40
-40
-40
Uo
-45
-45
-45
-45
-50
-50
-50
-50
-55
-55
-55
-55
-60
-60
-60
-60
-65
-65
-65
-65
Example of CPT test data for locations inboard of the toe of the levee embankment at and near the 17th Street Canal
Figure 8.38:
breach section.
877
1.5
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Group 3 & 4 (Crest, outside breach, with CPT data)
Bq = Gu/(qT-vvo)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0
1
n-5
CPT-4A
Elevation (ft)
-10
-15
-20
Bq = 0.48 to 0.68
-25
-30
0.1
-35
1
CPT-4A
OCR
-40
Figure 8.39: Pore Pressure parameter Bq vs. OCR and vs. depth within the soft gray clay
(CH) foundation layer under full embankment overburden load.
Figure 8.40: CPT cone factor Nkt based on Bq (after Lunne et al., 1985 and Karlsrud et al.,
1996) for the soft gray clay (CH) under full embankment overburden load.
878
10
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
10
(S u/v 'v)/(S u/v 'v)NC
Data DSS test, Grey
Clay (CH) @
from Foott & Ladd,
1977
(Su/v'v)/(Su/v'v)NC=OCR 0.75
1
1
OCR
10
Figure 8.41: Relationship between Su/P vs. OCR for mineralogically similar gray
marsh clays (CH) at Atchafalia.
[Foott & Ladd, 1977]
879
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Su/v 'vo
OCR
1
2
3
4
0
5
0
0
-5
-5
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
(su / v'v)NC ~ 0.30
-10
-10
-15
-15
-20
-20
-25
-25
-30
-30
CPT-4A (qc)
CPT- 3A (qc)
-35
-35
CPT 4 (u2)
CPT-4A
CPT-3A
-40
-40
Figure 8.42: Plots of (a) OCR vs. Depth and (b) Su vs. Depth for the soft gray marsh clay
(CH) beneath the full embankment overburden pressure.
880
Transition & Grey Clay (CH)
Elevation (ft)
(Nk = 12)
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Su/v 'vo
OCR
2
3
4
5
0
0
-5
-5
-10
-10
-15
-15
-20
-20
-25
-25
-30
-30
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
(Nk = 12)
(su /v'v)NC ~ 0.31
CPT-2
-35
-35
CPT-2
-40
PLAXIS (Soft Soil Model
prediction)
-40
Figure 8.43: Plots of (a) OCR vs. Depth and (b) Su vs. Depth for the soft gray marsh clay
(CH) at the inboard toe and further to the landside (not under levee
embankment overburden pressure) – 17th Street Canal breach site.
881
1
Transition & Grey Clay (CH)
Elevation (ft)
1
0
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Su (psf)
Su (psf)
0
200
400
600
800
0
1000
UU-TX
Lab Vane (peak)
400
600
800
1000
17-CPT-3A, Nk=12
17-CPT-4A, Nk=12
UU-TX
Field Vane (peak)
Lab Vane (peak)
Field Vane (peak)_Disturbed Area
17-CPT-2, Nk=12
-20
-20
-25
Best estimate
-30
Elevation (ft)
-15
-25
Best estimate
Transition & Grey Clay (CH)
-15
Transition & Grey Clay (CH)
Elevation (ft)
200
-10
-10
-30
-35
-35
-40
-40
(a) Beneath the crest of the levee
(b)
Beneath the inboard toe
Profiles of shear strength vs. depth within the soft gray foundation clay at the 17th Street Canal breach
site (a) beneath the crest of the levee, and (b) inboard of the toe of the levee.
Figure 8.44:
882
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
a) beneath the crest
b) at or near inboard toe of levee
Figure 8.45: Profiles of shear strength vs. depth within the soft gray foundation clay at the 17th Street Canal breach site (a) beneath the
crest of the levee, and (b) at and near the inboard toe of the levee, including the ILIT, IPET and preKatrina data.
883
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Estimated(UU,FVT & LVT)
0.30~0.31
0.80~0.84
0.24~0.26
Data DSS (from Foott
& Ladd, 1977)
Figure 8.46: Undrained shear strength for UUtriaxial loading vs. undrained shear
strength for DSS loading.
[Base figure from Ladd, 2003]
17th Street
Canal
Figure 8.47: Su/P for in situ field vane testing vs. Plasticity Index (PI) for normally
consolidated clays (OCR = 1).
[Base Figure from Ladd, 2003]
884
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
from CPT 4-4A
@ -10.5 ~ -14.70 ft
Bq ~ 0.28 - 0.35
Figure 8.48: CPT cone factor Nkt based on Bq (after Lunne et al., 1985 and Karlsrud et al.,
1996) for the marsh deposits under full embankment overburden load;
17th Street Canal breach site.
B
885
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Su/v 'v
STRENGTH ON TOE OF LEVEE
MARSH DEPOSIT
0
0.2
0.4
0.6
0.8
1
1.2
STRENGTH ON TOE OF LEVEE
MARSH DEPOSIT
1.4
1.6
1.8
1.0
2
-7
3.5
4.0
-7
60
Elevation (ft)
Elevation (ft)
3.0
-5
-5
100
140
-9
-11
> PI
-13
-11
-13
-15
-17
-17
-19
PI ~ 60
100
140
-9
-15
Consolidation data
(Modjeski, 1971)
-19
ILIT - Field Vane Marsh
INLIT
IPET - UUTX Masrh
IPET
ILIT - UC Marsh
ILIT -UTX Marsh
INLIT
ILIT - LV Marsh
INLIT
IPET-UUTX CH
ILIT - UC CH
IPET
ILIT - FV CH
INLIT
ILIT - UTX CH
INLIT
ILIT - LV CH
INLIT
-21
PI ~ 62%
PI ~ 103%
PI ~ 141%
Consol. data (Foott & Ladd, 1977)
-23
-25
Figure 8.49:
2.5
-3
-3
-23
2.0
-1
-1
-21
1.5
OCR
-25
Undrained shear strength vs. depth, and OCR vs. depth, within the marsh deposits beneath the inboard toe based on
th
Street Canal breach site.
Mayne and Mitchell (1988) – 17
886
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Su (psf)
Su (psf)
0
200
400
600
800
0
1000
200
400
600
800
1000
0
0
UUTX (Kaufman et al, 1967)
ILIT -UTX Marsh
ILIT - LV Marsh
ILIT - Lab Vane Marsh
IPET UC
17-CPT-3A, ILIT
17-CPT-4A, ILIT
-5
-5
17-18.05C, IPET
ILIT - Field Vane
ILIT -UU TX
ILIT - Lab Vane
IPET UU-TX
IPET UC
17-CPT-2, ILIT
17-11C, IPET
17-12C, IPET
17-2.05C, IPET
17-3.05C, IPET
Elevation (ft)
-10
Sensitive Zone
Marsh
-15
-10
Sensitive Zone
-15
-20
-20
-25
-25
(a)
Beneath the crest of the levee
Marsh
Elevation (ft)
17-4a.05C, IPET
(b) At the inboard toe
Shear strength vs. depth within the marsh deposits at the 17th Street Canal breach site (a) beneath the crest
of the levee, and (b) at the inboard toe.
Figure 8.50:
887
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.51: Shiny dark brown to black sensitive organic clay on auger stem and (inset)
closeup view of leaves and twigs; 17th Street Canal breach site.
Figure 8.52: Shear zone within disturbed area, middle of 17th Street Canal breach, adjacent to
displaced block at Elev. 9 feet (MSL).
888
Figure 8.53: Plan view of the 17th Street Canal breach site showing locations at which the shear zone and/or the
sensitive layer” within the marsh deposits was positively identified, sampled, or tested.
Independent Levee
Investigation Team
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
889
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
17th Str. Canal, West Bank
Figure 8.54: Photo of an “undisturbed” sample of the sensitive organic silty clay.
890
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.55: Laboratory vane shear testing of the thin layer of sensitive organic silty clay.
891
New Orleans Levee Systems
Hurricane Ka
Katrin
trina
a
July 31, 2006
Independent Lev
Independent
Levee
ee
Investigat
Invest
igation
ion Team
Lab Vane Test Results
for 17th Str. Canal, East Bank
1.00
0.90
0.80
CH/OH, OCR=1
0.70
CH, OCR= 2 to 3
0.60
τ/Su
Sensitive Layer, OCR=1
0.50
Sensitive Layer, OCR=1.5
0.40
0.30
0.20
0.10
0.00
0
1
2
3
4
5
6
7
8
9
10
11
12
9
10
11
12
Displacement (in)
Lab Vane Test Results
for 17th Str. Canal, East Bank
800
700
Shear Strength, τ (psf)
600
Sensitive Layer, OCR=1.5
500
CH/OH, OCR=1 to 3
400
Sensitive Layer, OCR=1
300
200
100
0
0
1
2
3
4
5
6
7
8
Displacement (in)
Figure 8.56: Stressdisplacem
Stressdisplacement
ent behavior
behavior for soft gray clay (CH and CH/OH) and
“sensitive”
e” layer of organic clay within marsh
“sensitiv
marsh deposit.
892
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Geotechnical parameters for the Finite Element Analysis Model
ID
Soil Model
Name
Type
CF
Upper CH
SC
Mohr Coulomb
ML-silt (Compacted Fill)
CH-Upper Fat Clay
SC (Sand)
Undrained
Undrained
Drained
ID
Soil Model
Name
Type
Soft Soil
Marsh 2
Marsh 1
Intermixing zone
Grey Clay
Grey Clay
Grey Clay
Undrained
Undrained
Undrained
Undrained
Undrained
Undrained
M2
M1
IZ
CH1
CH2-a
CH2-b
g_unsat
[lb/ft^3]
105
90
110
g_sat
[lb/ft^3]
115
95
110
k_x
[ft/day]
0.0028
0.00028
0.28
k_y
[ft/day]
0.00028
0.000028
0.28
nu
[-]
0.35
0.35
0.30
E_ref
[lb/ft^2]
234000
180000
1000000
c_ref
[lb/ft^2]
900
600
0.01
phi
[ °]
0.001
0.001
38
g_unsat
[lb/ft^3]
80
80
85
90
90
90
g_sat
[lb/ft^3]
80
80
85
95
95
95
k_x
[ft/day]
28.3
28.3
28.32
0.00028
0.00028
0.00028
k_y
[ft/day]
2.8
2.8
2.8
0.000028
0.000028
0.000028
lambda*
[-]
0.21
0.21
0.10
0.17
0.17
0.17
kappa*
[-]
0.033
0.033
0.02
0.03
0.03
0.03
n_ur
[-]
0.15
0.15
0.15
0.15
0.15
0.15
K0nc
[-]
0.60
0.60
0.61
0.63
0.63
0.63
M
[-]
1.90
1.90
1.27
1.24
1.24
1.24
Figure 8.57: Parameters (and model) used in PLAXIS model for the 17th Street Canal breach site.
886
c_ref
[lb/ft^2]
0
0
0
0
0
0
phi @ pref
[ °]
36
36
23
22
22
22
OCR
2.25
1.1
3
2.2
1.25
1.5
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.58: Deformed mesh at the end of initial construction of embankment and consolidation; 17th Street Canal breach section.
887
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Relative Shear Strain
( j/jmax)
Figure 8.59:
Normalized shear strain contours (shear strain divided by strain to failure) for a storm surge at Elev. + 8 feet (MSL)
At the 17th Street Canal breach site; initiation of gapping at outboard toe of floodwall.
888
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Relative Shear Strain
( j/jmax)
Normalized shear strain contours (shear strain divided by strain to failure) for a storm surge at Elev. + 8 feet (MSL)
at the 17th Street Canal breach site; gapping at outboard toe of floodwall is now developed to full depth.
Figure 8.60:
889
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
.
Displaced mesh for storm surge height at Elev. + 8.5 feet (MSL) at the 17th Street Canal breach site; displacements are
Figure 8.61:
increased by a factor of 3 in this figure.
890
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
06
Storm Surge, ft (MSL)
Independent Lev
Independent
Levee
ee
Investigat
Invest
igation
ion Team
12
11
10
9
8
7
6
5
4
3
2
1
0
Marsh Drained
Marsh Undrained
Marsh Drained, sensitive
layer
sensitive layer, full gap
0.5
0.6
0.7
0.8
0.9
1.0
1.1
1.2
1.3
FS
Figure 8.62: Calculated Factors
Factors of Safety for three modes
modes based on PLAXIS analyses of
th
1
t he1 7 Street Canal bbrreach section for various canal water elevations;
elevations;
showing the
the bestestim
bestestimated
ated path
path to ffailure.
ailure.
891
1.4
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
ID
Soil Model
Type
j#
c
phi
[lb/ft^3]
[lb/ft^2]
[ °]
0
0
0
0
0
0
0
0
Upper fill
Lower fill
Marsh, beneath crest
Marsh, free field
Sensitive layer, beneath crest
Sensitive layer, beneath toe
Sensitive layer, free field
Overconsolidated Grey CH
Normally consolidated Grey CH
Mohr-Coulomb
Mohr-Coulomb
Mohr-Coulomb
Mohr-Coulomb
Mohr-Coulomb
Mohr-Coulomb
Mohr-Coulomb
Mohr-Coulomb
SHANSEP
Undrained
Undrained
Undrained
Undrained
Undrained
Undrained
Undrained
Undrained
Undrained
110
85
80
80
80
80
80
90
90
800
550
375
200
240
180
70
400
Sand
Mohr-Coulomb
Drained
110
0
Su/p'
0.26
33
Geotechnical Parameters for the Limit Equilibrium Analyses for 17th Street Canal levee, East Bank.
Figure 8.63: Crosssection and parameters used for conventional stability analyses of the breach section on the east side of the
17th Street Canal.
892
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
th
Figure 8.64: Crosssection at 17 Street Canal breach site for static slope stability analyses.
FS
FS==1.51
1.51
Figure 8.65:
Stability analysis (worst case) for storm surge to Elev. + 6 feet (MSL) at the 17th Street Canal breach section.
893
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
FS = 0.83
Figure 8.66: Stability analysis for actual observed failure mechanism at the 17th Street Canal breach section, with storm surge
at Elev. + 9.5 feet (MSL) and with fully developed crack at the outboard side of the sheetpile/floodwall.
FS = 1.11
Figure 8.67:
Stability analysis for shear failure through the deeper soft gray clays (CH) at the 17th Street Canal breach
section, with storm surge at Elev. + 9.5 feet (MSL) and with fully developed crack at the outboard side of the
sheetpile/floodwall.
894
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
12
Sensitive Layer, Full Gap
Sensitive Layer, No Gap
Weak Clay, Gap
Weak Clay, No Gap
Storm Surge
+9.5ft (MSL)
Storm Surge Elevation, ft (MSL)
10
8
6
4
2
0
0.0
0.5
1.0
1.5
2.0
2.5
FS
Figure 8.68: Factor of safety for four modes of failure at the 17th Street Canal breach site as
a function of canal water elevation based on static Limit Equilibrium slope
stability analyses: again showing the bestestimated path to failure.
895
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Gap mean
Gap +stdev
Gap -stdev
No Gap mean
No Gap +stdev
No Gap -stdev
FS=1.0
9.5
9.0
Storm Surge (Ft), MSL
8.5
8.0
7.5
7.0
6.5
6.0
5.5
5.0
4.5
0.5
0.6
0.7
0.8
0.9
1.0
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
1.9
Factor of Safety
Figure 8.68(a): Distributions of Factor of Safety for the Ungapped and WaterFilled Gap
Cases as a function of increasing canal water level; 17th Street Drainage
Canal, east bank breach section.
11
Storm Surge (Ft), MSL
10
9
8
7
Best estimate
Upper
Lower
6
5
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
1.00
Probability of Gap Formation
Figure 8.68(b): Estinated probabilities of formation of a significant waterfilled gap as a
function of increasing canal water level; 17th Street Drainage Canal, east
bank breach section.
896
2.0
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.68(c): Example calculation (Monte Carlo simulation) for a canal water elevation
of +5 feet (MSL).
9.5
Gap mean
No Gap mean
Transition
Transition +stdev
9.0
Storm Surge (Ft), MSL
8.5
Transition -stdev
FS=1.0
8.0
7.5
7.0
6.5
6.0
5.5
5.0
4.5
0.5
0.6
0.7
0.8
0.9
1.0
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
1.9
Factor of Safety
Figure 8.68(d): Distribution of Factor of Safety vs. canal water level with progressive
transition from ungapped to waterfilled gap conditions; 17th Street Canal,
east bank breach section.
897
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Storm Surge
Elevation (ft), MSL
9
8
7
6
5
Probability of
Failure
88.1%
25.6%
7.9%
0.01%
0.00%
Table 8.1: Probability of failure for 17th Street Canal, East Bank
898
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.69: Comparison between deep rotational failure through the soft gray clay (CH) and translational failure along the sensitive
clay layer within the marsh deposits as actually observed.
8 - 99
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Initial longitudinal subsurface profile used for initial design at the 17th Street Canal breach site.
Figure 8.70:
8 100
[USACE, DM20, Vol. 1, 1990]
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Soil Profile
Figure 8.71: Reinterpreted longitudinal subsurface soil profile, showing location of breach section on the east side of the 17th Street
Canal, and with nontested samples highlighted.
8 101
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Use of the “Method of Planes” in original design stability analysis calculations; 17th Street Canal.
Figure 8.72:
8 102
[USACE, DM20, Vol.1, 1990]
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.73: Profile of shear strength vs. depth used in original stability analyses for design at the 17th Street Canal breach site.
[USACE, DM20, Vol.1 & 2, 1990]
8 103
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Failure Plane
Failure Plane
Figure 8.74: Profile of shear strength vs. depth used in original stability analyses for design at the 17th Street Canal breach site, and this
investigation team’s best estimated profiles of undrained shear strength vs. elevation (a) beneath levee crest [blue line],
and (b) beneath levee toe [red line].
[USACE, DM20, Vol.1 & 2, 1990]
8 104
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.75: View of floodwall displacement on the west bank of the 17th Street Canal.
Figure 8.76: View of sample of “sensitive” clay layer within the marsh soils at the west
side of the 17th Street Canal.
8 105
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.77:
Analysis crosssection for the site on the west bank of the 17th Street Canal.
8 105
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Bq from CPT
at 17th Str. Canal We st Bank, (le v e e cre st)
Bq from CPT
at 17th Str. Canal We st Bank, (le v e e cre st)
10
1
17-cpt-10 GE = 4.31
5
17-cpt-10 GE = 4.31
0
-5
Bq
-10
Elevation (ft)
-15
-20
-25
Bq = 0.35 to
0.60
-30
0.1
1
10
-35
OCR
-40
Figure 8.78: Bq for the soft gray clay at the site on the
West side of the 17th Street Canal.
-45
B
-50
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Bq
from CPT 10 & 11
@ -25 ~ -34ft
B q ~ 0.35 - 0.6
Figure 8.79: Nkt and NΔu for the soft gray clay at the site on the west side of the 17th Street
Canal based on CPTU.
8 106
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
(17th west, crest, distressed section)
(17th we st, cre st, distre sse d section)
10
10
17-cpt-10 GE = 4.31
17-cpt-10 GE = 4.31
5
5
0
-5
-5
-10
-10
-15
-15
Elevation (ft)
0
-20
-25
-20
Grey Clay, CH
Elevation (ft)
Nk = 12
-25
(Su/σv')NC = 0.31
-30
-30
-35
-35
-40
-40
-45
-45
-50
-50
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1
Su/P
1.5
2
2.5
3
3.5
4
4.5
5
OCR
Su/P vs. depth and OCR vs. depth for soft gray clay (CH) beneath the crest of the embankment for the site on
th
Street Canal.
Figure 8.80:
the west side of the 17
8 107
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Su strengths from CPT, Grey CH
at 17th Str. Canal West Bank, (levee toe)
IPET and EUSTIS data
-15
-20
-20
-25
-25
Elevation (ft)
-15
-30
Grey Clay, CH
Elevation (ft)
Su strengths from CPT, Grey CH
at 17th Str. Canal West Bank, (levee crest)
Nk = 12 ILIT data
-30
-35
-35
17-cpt-10 GE = 4.31
UC-TX, IPET
17-CPT-11 GE = 4.31
UU-TX, IPET
UU-TX, ILIT
UU-TX, EUSTIS
Lab Vanes, ILIT
-40
-40
0
500
1000
1500
0
2000
Su (psf)
1000
1500
2000
Su (psf)
Beneath the levee crest
(a)
500
(b) Beneath the inboard of the levee toe
Shear strength vs. depth within the soft gray clay (CH) for the site on the west side of the 17th Street Canal
(a) beneath the crest of the levee, and (b) beneath the inboard side toe of the levee embankment.
Figure 8.81:
8 108
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Ref: Mayne, P.W and Mitchell,
J.K (1988)
Su/P
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
2.2
2.4
-1
-3
23
-5
19
20
-7
100
81
20
2.4
2.2
1.2
0.6
0.34
0.27
0.22
0.2
0.2
0.2
PI (%)
100
180
2.2
0.9
0.6
0.4
0.4
0.4
0.4
3
1.4
0.8
0.6
0.55
0.55
5.2
12.5
11.5
6.3
3.1
1.8
1.4
1.1
1.0
1.0
1.0
2.4
1.8
5.3
2.2
1.4
1.0
1.0
1.0
1.0
5.5
2.5
1.5
1.1
1.0
1.0
STRENGTH ON LEVEE CREST MARSH DEPOSIT
108
108
OCR
180
-11
1.0
209
209
2.0
3.0
4.0
5.0
6.0
7.0
-4
-13
-15
-6
> PI
Elevation (ft)
Elevation (ft)
-9
elev (ft)
-5
-6
-7
-8
-9
-10
-11
-12
-13
-14
OCR= dVST (Su/v'v)
dVST = 22 PI-0.48
Su/v'v
STRENGTH ON LEVEE CREST MARSH
-17
-19
-8
-10
-12
-21
-14
PI ~ 20%
-23
UU-TX ILIT
Lab Vane ILIT
-16
PI ~ 100%
PI ~ 180%
-25
-18
Figure 8.82: Su/P and OCR estimation from PI and vane shear tests for the marsh deposits at the
th
Street Canal.
site on the west side of the 17
8 109
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Su stre ngths from CPT, M arsh
at 17th Str. Canal We st Bank, (le v e e toe )
IPET and EUSTIS data
0
-5
-5
-10
-10
Elevation (ft)
0
Sensitive layer
-15
Sensitive layer
Marsh
Elevation (ft)
Su strengths from CPT, Marsh
at 17th Str. Canal West Bank, (levee crest)
Nk = 16 ILIT data
-15
-20
-20
17-cpt-10 GE = 4.31
17-CPT-11 GE = 4.31
UU-TX, ILIT
UU-TX, EUSTIS
Lab Vanes, ILIT
-25
-25
0
500
1000
1500
0
2000
500
1000
1500
2000
Su (psf)
Su (psf)
(a) Beneath the crest of the levee
Beneath the inboard toe of the levee
(b)
Figure 8.83: Shear strength vs. depth within the marsh deposits at the site on the west side of the 17th Street Canal (a) beneath
the crest of the levee, and (b) beneath the inboard toe of the embankment.
8 110
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
FS = 2.09
Figure 8.84: Stability analysis of the west side of the 17th Street Canal with water at Elev.
+ 9 feet (MSL) for case of rotational failure through the soft gray clay (CH)
with no gap developed.
FS = 1.47
Figure 8.85: Stability analysis of the west side of the 17th Street Canal with water at Elev.
+ 9 feet (MSL) for case of failure through the base of the “marsh”layer
(with no gap developed).
8 111
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
FS = 1.47
Figure 8.86: Stability analysis of the west side of the 17th Street Canal with water at Elev.
+ 9 feet (MSL) for case of rotational failure through the soft gray clay (CH)
(with waterfilled gap).
FS = 1.16
Figure 8.87: Stability analysis of the west side of the 17th Street Canal with water at Elev.
+ 9 feet (MSL) for case of failure along the base of the “marsh” layer
(with waterfilled gap).
8 112
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
12
11
Storm Surge Elevation, ft (MSL)
10
9
8
7
6
5
Gap
No gap
4
0.50 0.60 0.70 0.80 0.90 1.00 1.10 1.20 1.30 1.40 1.50 1.60 1.70 1.80 1.90 2.00
FS
Figure 8.88: Factor of safety vs. water level on the west bank of the 17th Street Canal.
8 113
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Gap mean
Gap +stdev
Gap -stdev
No Gap mean
No Gap +stdev
No Gap -stdev
FS=1.0
11.5
11.0
10.5
Storm Surge (Ft), MSL
10.0
9.5
9.0
8.5
8.0
7.5
7.0
6.5
6.0
5.5
5.0
4.5
0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 2.1 2.2 2.3 2.4 2.5
Factor of Safety
Figure 8.88(a): Distributions of factors of safety for the “Waterfilled Gap” and the
“Ungapped” cases as a function of rising canal water elevations; 17th Street
Drainage Canal; west bank distressed section.
11.5
Gap mean
No Gap mean
Transition
Transition +stdev
11.0
10.5
Storm Surge (Ft), MSL
10.0
Transition -stdev
FS=1.0
9.5
9.0
8.5
8.0
7.5
7.0
6.5
6.0
5.5
5.0
4.5
0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 2.1 2.2 2.3 2.4 2.5
Factor of Safety
Figure 8.88(b): Distributions of conjugate overall factors of safety for the 17th Street
Drainage Canal; west bank distressed section.
8 114
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Storm Surge
Elevation (ft), MSL
11
10
9
8
7
6
5
Probability of
Failure
59.2%
38.7%
4.9%
0.10%
0.013%
0.007%
0.0%
Table 8.2: Probability of failure for 17th Street Canal, West Bank
8 115
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.89: View of failure and breach on the east bank near the south end of the London
Avenue Canal.
Figure 8.90: View of sands piled in neighborhood inboard of the south London Avenue Canal
breach.
8116
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figures 8.91 and 8.92: Views of floodwall panels “dropping” into the eroded void at the
north and south ends of the south London Avenue canal breach.
8117
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.93: View of trees at the inboard levee toe immediately to the north of the end of the
south breach in the London Avenue Canal.
Figure 8.94: View of toppled trees at the London Avenue Canal south breach site.
8118
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.95: Plan view of the London Avenue Canal (South) breach site.
8119
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.96: Crosssection through the breach near the south end of the London Avenue Canal.
8120
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
ID
Upper fill
SM/SC w/org anics, crest
SM/SC w/org anics, free field
Lean organic silty clay, crest
Lean organic silty clay, free field
Medium Dense Sand
Dense Sand
Soil M odel
Mohr-Coulomb
Mohr-Coulomb
Mohr-Coulomb
Mohr-Coulomb
Mohr-Coulomb
Mohr-Coulomb
Mohr-Coulomb
Type
Undrained
Undrained
Undrained
Undrained
Undrained
Drained
Drained
[lb/ft^3]
100
90
90
100
100
100
110
c
[l b/ft^2]
800
250
200
500
300
0
0
phi
[ °]
0
0
0
0
0
33
35
k
[ ft/hr ]
1.20E-03
0.011
0.011
3.60E-03
3.60E-03
1.100
0.600
Figure 8.97: Geotechnical crosssection for analysis of the London Avenue south breach.
8121
Figure 8.98: Flow vectors and head contours for seepage analysis of the south London Avenue Canal breach for a surge at
Elev. + 9 feet (MSL).
Independent Levee
Investigation Team
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
8122
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.99: Closeup of flownet for storm surge at + 9ft (MSL) in the London Avenue
Canal south breach. (Equipotential contours at intervals of one foot of head)
Figure 8.100: Pressure contours for Storm Surge at 9ft (MSL) in the London Avenue Canal
South breach. (Pressure contours every 250 lb/ft2)
8123
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
io= 0.5
io= 0.7
Figure 8.101: Hydraulic gradients for storm surge at +9ft (MSL) in the London Avenue
Canal south breach. (Maximum toe exit gradient io= 0.7)
io= 0.8
Figure 8.102: Hydraulic gradients for storm surge at +9ft (MSL) in the London Avenue Canal
south breach: initiation of erosion at levee toe. (Maximum gradient io= 0.8)
8124
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
io= 0.9
Figure 8.103: Hydraulic gradients for storm surge at +9ft (MSL) in the London Avenue Canal
south breach; development of erosion at levee toe. (Maximum gradient io= 0.9).
io= 1.0
Figure 8.104: Hydraulic gradients for storm surge at +9ft (MSL) in the London Avenue Canal
south breach; development of erosion at levee toe. (Maximum gradient io = 1.0.)
8125
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
io= 0.5
io= 0.7
Figure 8.101: Hydraulic gradients for storm surge at +9ft (MSL) in the London Avenue
Canal south breach. (Maximum toe exit gradient io= 0.7)
io= 0.8
Figure 8.102: Hydraulic gradients for storm surge at +9ft (MSL) in the London Avenue Canal
south breach: initiation of erosion at levee toe. (Maximum gradient io= 0.8)
8124
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.107: Schematic illustration of progressive erosion development at inboard toe.
8127
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.108: Schematic illustration of toppling of tree at inboard toe of levee.
8128
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.109: View of the breach near the south end of the London Avenue Canal from the
canal side in late September of 2005.
Figure 8.109(a): Sample across transition zone at London Avenue Canal, South Breach
Section
8129
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.110: Overhead view of the breach section on the west bank near the north end of the
London Avenue Canal.
Figure 8.111: Oblique aerial view of the breached section on the west bank of the London
Avenue Canal (North) and the “distressed” section on the east bank (on the
left in this photo, which is taken looking to the south.)
8130
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.112: View of the west bank breach near the north end of the London Avenue Canal from
the south end showing the outboard side embankment section still in place.
Small Clubhouse
Boil Ejecta
He small wood
Figure 8.113: View of the inboard of the displaced embankment on the west side of the
London Avenue canal showing heaving at the toe (and beneath the small
wooden clubhouse), and the boil ejecta in front of the heave feature.
8131
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.114: Waterfilled gap at outboard side of east bank distressed section.
Figure 8.115: Inboard leaning floodwalls on the other side of the floodwalls shown above in
Figure 8.112.
8132
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.116: Sinkholes along the contact at the inboard side base of the floodwall and the
levee crest.
Figure 8.117: Sand ejecta from toe boil, and edge of toe thrust feature; London Avenue canal
(north) east bank distressed section.
8133
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.118: PreKatrina and PostKatrina crosssection for London Avenue Canal North, west bank breach section.
8134
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.119: PreKatrina and PostKatrina crosssection for London Avenue Canal North,
east bank, distressed section.
8135
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.120: Approximate planview of London Avenue Canal, North, showing location of borings and CPT.
8136
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
BREACH
Figure 8.121: Initial longitudinal subsurface profile used for initial design at the London Avenue Canal North, west breach site.
[USACE, DM19A, Vol. 1, 1989].
8137
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Soil Profile
Figure 8.122: Reinterpreted longitudinal subsurface soil profile, showing location of breach section on the west bank of the London
Avenue Canal, North.
8138
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
DISTRESS
Figure 8.123: Initial longitudinal subsurface profile used for initial design at the London Avenue Canal North, east bank distressed site.
[USACE, DM19A, Vol. 1, 1989].
8139
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Soil Profile
Figure 8.124: Reinterpreted longitudinal subsurface soil profile, showing location of
distressed section on the east bank of the London Avenue Canal.
8140
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Shear strengths - London Ave. North (WB), Nk = 16
Levee Toe
Shear strengths - London Ave. North (WB)
Levee Crest
0
0
UU-TX, ILIT
Lab Vane, ILIT
UU-TX, ILIT
LACW-CPT-3, ILIT
LACW-CPT-4, ILIT
Lab Vane, ILIT
NLON-1.05C, IPET
NLON-2.05C, IPET
UU-TX IPET
NLON-12.05C, IPET
-2
-2
LACW-CPT-1, ILIT
Fill
Direct Shear Test
-4
-4
-6
-6
M a rsh
-8
Elevation (ft)
Elevation (ft)
M a rsh
-10
-8
CL/ML
-10
CL/ML
-12
-12
SC
SC
-14
-14
SP
-16
-16
SP
-18
-18
0
500
1000
1500
0
2000
Shear strength (psf)
500
1000
1500
2000
Shear strength (psf)
a) beneath crest of levee
b) at or near toe of levee
Figure 8.125: Bestestimate for shear strength (Su) from CPTdata for London Avenue Canal
west bank breach section (a) beneath the crest of the levee and (b) at or near
the toe of the levee.
8141
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Su/P - London Ave. North (WB)
Levee Crest
Su/P - London Ave. North (WB), Nk = 16
Levee Toe
0
0
-2
-2
Fill
-4
-4
-6
-6
-8
Elevation (ft)
Elevation (ft)
Marsh
Marsh
-10
-8
CL/ML
-10
CL/ML
-12
-12
SC
SC
-14
-14
LACW-CPT-3, ILIT
LACW-CPT-4, ILIT
-16
-16
NLON-1.05C, IPET
LACW-CPT-1, ILIT
NLON-2.05C, IPET
NLON-12.05C, IPET
-18
-18
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
0
5
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
Su/P
Su/P
( a) beneath crest of levee
(b) at or near toe of levee
Figure 8.126: Bestestimate for Su/P from CPTdata for London Avenue Canal west bank
breach section a) beneath the crest of the levee and b) at or near the toe of the
levee.
8142
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
10
LACW-CPT-2
LACW-CPT-3
5
equivalent friction angle from
CPT data (Olsen and Farr, 1986)
LACW-CPT-4
φ from SPT
0
Direct Shear Test
elevation (ft)
-5
-10
SC
-15
-20
-25
SP
-30
-35
-40
0
5
10
15
20
25
30
35
40
45
50
55
60
φ (degrees)
Figure 8.127: Estimation of friction angle for cohesionless materials for London Avenue
Canal west bank breach section.
Notes: a) The continuous lines are from Robertson and Campanella (1983)
b) Friction angle from SPT is based on R.B. Seed’s table
c) The direct shear test was performed at the UC Berkeley geotechnical laboratory
8143
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Shear strengths - London Ave. North (EB) Levee Midslope, Nk = 16
Shear strengths - London Ave. North (EB)
Levee Toe, Nk = 16
0
0
LAC-CP-1, IPET
UU-TX, ILIT
Field Vane (peak), ILIT
LAC-CP-2, IPET
Lab Vane, ILIT
LAC-CP-3, IPET
-2
LAC-CPT-1, ILIT
-2
LAC-CPT-2, ILIT
LAC-CPT-4, ILIT
Fill
Direct Shear Test
-4
-4
-6
-6
M a rsh
-8
Elevation (ft)
Elevation (ft)
M a rsh
Best Estimate
-10
-12
-8
CL/ML
-10
Best Estimate
-12
SC
CL/ M L
-14
-14
SC
-16
-16
-18
-18
0
500
1000
1500
0
2000
Shear strength (psf)
500
1000
1500
2000
Shear strength (psf)
a) beneath crest of levee
b) at or near toe of levee
Figure 8.128: Bestestimate for shear strength (Su) from CPTdata for London Avenue Canal
distressed section (east bank) a) beneath the crest of the levee and b) at or near
the toe of the levee.
8144
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Su/P - London Ave. North (EB)
Levee Mid-slope
Su/P - London Ave. North (EB)
Levee Toe
0
0
LAC-CP-1, IPET
LAC-CPT-1
LAC-CPT-2
LAC-CP-2, IPET
LAC-CPT-4
LAC-CP-3, IPET
-2
-2
Fill
-4
-4
-6
-6
Marsh
-8
Elevation (ft)
Elevation (ft)
Marsh
-10
-8
-10
CL/ML
-12
-12
SC
(Su/P)NC = 0.3
CL/ML
-14
-14
SC
-16
-16
-18
-18
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
Su/P
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
Su/P
a) beneath crest of levee
b) at or near toe of levee
Figure 8.129: Bestestimate for Su/P from CPTdata for London Avenue Canal distressed
section (east bank) a) beneath the crest of the levee and b) at or near the toe of
the levee
8145
5
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
0
LAC-CPT-2
LAC-CPT-1
φ from SPT
-5
equivalent friction angle from
CPT data (Olsen and Farr, 1986)
-10
elevation (ft)
SC
-15
-20
SP
-25
-30
-35
-40
0
5
10
15
20
25
30
35
40
45
50
55
60
φ (degrees)
Figure 8.130: Estimation of friction angle for cohesionless materials for London Avenue
Canal distressed section (east bank).
Notes: a) The continuous lines are from Robertson and Campanella (1983)
b) Friction angle from SPT is based on R.B. Seed’s table
8146
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
London Avenue Canal, North, East Bank
Geotechnical parameters for the Finite Element Analysis Model
ID
fill
marsh
CL/ML
SC/SM
SP_1
SP_2
SP_3
CH
Soil Model
Mohr Coulomb
Description
Type
(Non-engineered Fill / Fill)
marsh
silty low-PI clay (traces organic)
clayey, silty sand
loose sand
medium dense sand
dense sand
grey clay (Bay Sound)
Undrained
Undrained
Undrained
Drained
Drained
Drained
Drained
Undrained
γunsat
[lb/ft3]
70
50
70
90
90
100
110
95
γsat
[lb/ft3]
90
80
90
100
100
110
115
105
kx
ky
nu
[ft/day]
2.80E-03
28.3
2.80E-03
2.33
2.33
2.33
2.33
2.80E-03
[ft/day]
2.80E-04
2.83
2.80E-04
1.15
2.33
2.33
2.33
2.80E-04
[-]
0.35
0.35
0.35
0.3
0.25
0.25
0.25
0.35
φ
Eref
[lb/ft2]
1.00E+06
2.80E+05
9.00E+05
5.00E+05
9.00E+05
1.00E+06
1.50E+06
1.00E+06
cref
[lb/ft2]
800
550
250
0.001
0.001
0.001
0.001
600
[ °]
0.001
0.001
0.001
30
33
36
40
0.001
Eref
[lb/ft2]
1.00E+06
2.80E+05
9.00E+05
5.00E+05
9.00E+05
1.00E+06
1.50E+06
1.00E+06
cref
[lb/ft2]
800
450
300
0.001
0.001
0.001
0.001
600
[ °]
0.001
0.001
0.001
29
33
36
40
0.001
London Avenue Canal, North, West Bank
Geotechnical parameters for the Finite Element Analysis Model
ID
fill
marsh
CL/ML
SC/SM
SP_1
SP_2
SP_3
CH
Soil Model
Mohr Coulomb
Description
(Non-engineered Fill / Fill)
marsh
silty low-PI clay (traces organic)
clayey, silty sand
loose sand
medium dense sand
dense sand
grey clay (Bay Sound)
Type
Undrained
Undrained
Undrained
Drained
Drained
Drained
Drained
Undrained
γunsat
[lb/ft3]
70
50
70
90
90
100
110
95
γsat
[lb/ft3]
90
80
90
100
100
110
115
105
West Bank
kx
ky
nu
[ft/day]
2.80E-03
28.3
2.80E-03
2.33
2.33
2.33
2.33
2.80E-03
[ft/day]
2.80E-04
2.83
2.80E-04
1.15
2.33
2.33
2.33
2.80E-04
[-]
0.35
0.35
0.35
0.3
0.25
0.25
0.25
0.35
φ
East Bank
Figure 8.131: Geometry and input parameters for FEM analyses for London Avenue Canal, North (east and west banks)
8147
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Relative
Shear Strain
(j/jmax)
Figure 8.132: Normalized shear strain contours (shear strain divided by strain to failure) for a
storm surge at Elev. + 9 feet (MSL) at the London Avenue Canal breach site
(west bank); gapping at outboard toe of floodwall is developed to full depth.
Relative
Shear Strain
(j/jmax)
Figure 8.133: Normalized shear strain contours (shear strain divided by strain to failure) for a
storm surge at Elev. + 10 feet (MSL) at the London Avenue Canal distressed
site (east bank); gapping at outboard toe of floodwall is developed.
8148
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Relative Shear
Strain
( j/jmax)
Figure 8.134: Normalized shear strain contours (shear strain divided by strain to failure) for a storm surge at Elev. + 9 feet (MSL) at the
8149
London Avenue Canal (east and west banks).
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.135: Deformed mesh for a storm surge elevation + 9 feet (MSL), London Avenue Canal (east and west banks). Displacements
8150
are exaggerated for clarity.
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
London Avenue Canal, North (East Bank)
London Avenue Canal, North (West Bank)
12
12
wNo
ithout
Gap crack
wNo
ithout
Gap crack
wFull
ith Gap
crack
Storm Surge
+9.5ft (MSL)
10
Storm Surge Elevation, ft (MSL)
10
Storm Surge Elevation, ft (MSL)
Gap
w Full
ith crack
8
6
4
8
6
4
2
2
0
0
0
0.5
1
1.5
2
2.5
3
0
3.5
FS
0.5
1
1.5
2
2.5
3
3.5
FS
Figure 8.136: Calculated Factors of Safety for two modes based on PLAXIS analyses of the London Avenue Canal breach and distressed
section for various canal water elevations; showing the bestestimated path to failure.
8151
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
,
ID
Upper fill
Marsh, crest
Marsh, free field
CL/ML, crest
CL/ML, free field
SC
Loose Sand
Dense Sand
Stiff Clay
Soil Model
Mohr-Coulomb
Mohr-Coulomb
Mohr-Coulomb
Mohr-Coulomb
Mohr-Coulomb
Mohr-Coulomb
Mohr-Coulomb
Mohr-Coulomb
Mohr-Coulomb
Type
Undrained
Undrained
Undrained
Undrained
Undrained
Drained
Drained
Drained
Undrained
j
[lb/ft^3]
100
80
80
85
85
100
100
105
90
c
[lb/ft^2]
800
475
200
500
300
0
0
0
600
phi
[ °]
0
0
0
0
0
30
33
36
0
k
[ ft/hr ]
1.17E-04
1.17
1.17
2.00E-04
2.00E-04
0.097
0.333
0.09
2.00E-04
Figure 8.137: Geometry and input parameters for Limit Equilibrium and Steady State seepage
analyses for London Avenue Canal North, West bank.
Figure 8.138: Finite Difference mesh for Steady State seepage Analyses for London Avenue
Canal North, West bank.
8152
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.139: Flow net generation without the gapping in the outboard toe of the floodwall,
London Avenue Canal North, West bank. Storm surge at 9ft (MSL).
Figure 8.140: Pore water pressure contours without the gapping in the outboard toe of the
floodwall, London Avenue Canal North, West bank. Storm surge at 9ft (MSL).
8153
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.141: Hydraulic gradient contours without the gapping in the outboard toe of the
floodwall, London Avenue Canal North, West bank. Storm surge at 9ft
(MSL). Exit gradient at the inboard toe is 0.20.
FS = 2.14
Figure 8.142: Critical failure surface without the gapping in the outboard toe of the floodwall,
London Avenue Canal North, West bank. Storm surge at 9ft (MSL).
8154
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.143: Flow net generation with the gapping in the outboard toe of the floodwall,
London Avenue Canal North, West bank. Storm surge at 9ft (MSL).
Figure 8.144: Pore water pressure contours with the gapping in the outboard toe of the
floodwall, London Avenue Canal North, West bank. Storm surge at 9ft (MSL).
8155
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.145: Hydraulic gradient contours with the gapping in the outboard toe of the
floodwall, London Avenue Canal North, West bank. Storm surge at 9ft (MSL).
Exit gradient at the inboard toe is 0.32.
FS = 1.06
Figure 8.146: Critical failure surface with the gapping in the outboard toe of the floodwall,
London Avenue Canal North, West bank. Storm surge at 9ft (MSL).
8156
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
12
Gap
No gap
11
Storm Surge Elevation, ft (MSL)
10
Storm Surge
Elevation
+9.5ft (MSL)
9
8
7
6
5
4
0.50
1.00
1.50
2.00
2.50
3.00
FS
Figure 8.147 : Calculated Factors of Safety for two models based on Limit Equilibrium
Analyses of the London Avenue Canal, North West bank.
8157
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
ID
Upper fill
Marsh, crest
Marsh, free field
CL/ML, crest
CL/ML, free field
SC
Loose Sand
Dense Sand
Stiff Clay
Soil Model
Mohr-Coulomb
Mohr-Coulomb
Mohr-Coulomb
Mohr-Coulomb
Mohr-Coulomb
Mohr-Coulomb
Mohr-Coulomb
Mohr-Coulomb
Mohr-Coulomb
Type
Undrained
Undrained
Undrained
Undrained
Undrained
Drained
Drained
Drained
Undrained
j
[lb/ft^3]
100
80
80
85
80
100
100
115
95
c
[lb/ft^2]
800
550
300
275
200
0
0
0
600
phi
[ °]
0
0
0
0
0
30
33
36
0
k
[ ft/hr ]
1.17E-04
1.17
1.17
2.00E-04
2.00E-04
0.097
0.333
0.09
2.00E-04
Figure 8.148: Geometry and input parameters for Limit Equilibrium and Steady State seepage
Analyses for London Avenue Canal North, East bank.
Figure 8.149: Finite Difference mesh for Steady State seepage Analyses for London Avenue
Canal North, East bank.
8158
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.150: Flow net generation without the gapping in the outboard toe of the floodwall,
London Avenue Canal North, East bank. Storm surge at 9ft (MSL).
Figure 8.151: Pore water pressure contours without the gapping in the outboard toe of the
floodwall, London Avenue Canal North, East bank. Storm surge at 9ft (MSL).
8159
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.152: Hydraulic gradient contours without the gapping in the outboard toe of the
floodwall, London Avenue Canal North, East bank. Storm surge at 9ft (MSL).
Exit gradient at the inboard toe is 0.20.
Figure 8.153: Critical failure surface without the gapping in the outboard toe of the floodwall,
London Avenue Canal North, East bank. Storm surge at 9ft (MSL).
8160
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.154: Flow net generation with the gapping in the outboard toe of the floodwall,
London Avenue Canal North, East bank. Storm surge at 9ft (MSL).
Figure 8.155: Pore water pressure contours with the gapping in the outboard toe of the
floodwall, London Avenue Canal North, East bank. Storm surge at 9ft (MSL).
8161
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 8.156: Hydraulic gradient contours with the gapping in the outboard toe of the
floodwall, London Avenue Canal North, East bank. Storm surge at 9ft (MSL).
Exit gradient at the inboard toe is 0.25.
Figure 8.157: Critical failure surface with the gapping in the outboard toe of the floodwall,
London Avenue Canal North, East bank. Storm surge at 9ft (MSL).
8162
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
12
Gap
No gap
11
Storm Surge Elevation, ft (MSL)
10
Storm Surge
Elevation
+9.5ft (MSL)
9
8
7
6
5
4
0.50
1.00
1.50
2.00
2.50
3.00
FS
Figure 8.158 : Calculated Factors of Safety for two models based on Limit Equilibrium
analyses of the London Avenue Canal, North East bank.
8163
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
12
12
Gap
No gap
11
11
10
10
Storm Surge Elevation, ft (MSL)
Storm Surge Elevation, ft (MSL)
Gap
9
8
7
Storm Surge
+9.5ft (MSL)
9
8
7
6
6
5
5
4
4
0.50
No gap
1.00
1.50
2.00
2.50
0.50
3.00
1.00
1.50
2.00
2.50
3.00
FS
FS
Figure 8.159: Calculated Factors of Safety for two modes based on Limit Equilibrium Analyses of the London Avenue Canal breach and
distressed section for various canal water elevations; showing the bestestimated path to failure.
8164
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
CHAPTER NINE: EROSION TESTS ON
N EW ORLEANS LEVEE SAMPLES
9.1
Erodibilty: A Definition
Erodibility is a term often used in scour and erosion studies. Erodibility may be
thought of as one number which characterizes the rate at which a soil is eroded by the flowing
water. With this concept erosion resistant soils would have a low erodibility index and erosion
sensitive soils would have a high erodibility index. This concept is not appropriate; indeed the
water velocity can vary drastically from say 0 m/s to 5 m/s or more and therefore the
erodibility is a not a single number but a relationship between the velocity applied and the
corresponding erosion rate experienced by the soils. While this is an improved definition of
erodibility, it still presents some problems because water velocity is a vector quantity which
varies everywhere in the flow and is theoretically zero at the soil water interface. It is much
preferable to quantify the action of the water on the soil by using the shear stress applied by
the water on the soil at the watersoil interface. Erodibility is therefore defined here as the
relationship between the erosion rate z& and the hydraulic shear stress applied w (Figure 9.1).
This relationship is called the erosion function z& (w). The erodibility of a soil or a rock is
represented by the erosion function of that soil or rock. This erosion function can be obtained
by using a laboratory device called the EFA (Erosion Function Apparatus) and described
later.
9.2
Erosion Process
Soils are eroded particle by particle in the case of coarsegrained soils (cohesionless
soils). In the case of finegrained soils (cohesive soils), erosion can take place particle by
particle but also block of particles by block of particles. The boundaries of these blocks are
formed naturally in the soil matrix by microfissures due to various phenomena including
compression and extension.
The resistance to erosion is influenced by the weight of the particles for coarse grained
soils and by a combination of weight and electromagnetic and electrostatic interparticle
forces for fine grained soils. Observations at the soil water interface on slow motion
videotapes indicates that the removal of particle or blocks of particles is by a combination of
rolling and plucking action of the water on the soil.
9.3
Velocity vs. Shear Stress
The scour process is highly dependent on the shear stress developed by the flowing
water at the soilwater interface. Indeed, at that interface the flow is tangential to the soil
surface regardless of the flow condition above it; very little water if any flows perpendicular
to the interface. The water velocity in the river is in the range of 0.1 to 3 m/s, whereas the bed
shear stress is in the range of 1 to 50 N/m2 (Figure 9.2) and increases with the square of the
water velocity. The magnitude of this shear stress is a very small fraction of the undrained
shear strength of clays used in foundation engineering (Figure 9.3).
91
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
It is interesting to note that such small shear stresses are able to scour rocks to a depth
of 1,600 m, as in the case for the Grand Canyon over the last 20 million years at an average
scour rate of 9.1x106 mm/hr. This leads one to think that even small shear stresses if applied
cyclically by the turbulent nature of the flow can overcome, after a sufficient number of
cycles, the crystalline bonds in a rock and the electromagnetic bonds in a clay. This also leads
one to think that there is no cyclic stress threshold, but that any stress is associated with a
number of cycles to failure. (Gravity bonds seem to be an exception to this postulate, because
it appears that gravity bonds cannot be weakened by cyclic loading.) This postulate
contradicts the critical shear stress concept discussed later.
The profile of the water velocity versus depth in the flow (Figure 9.4) indicates a
maximum velocity at the free surface and a zero velocity at the bottom of the flow. This zero
velocity boundary is due to the fact that the water does not flow below the flow bottom. While
the velocity is zero at the bottom, the shear stress is maximum because the shear stress is
proportional to the slope of the velocity profile versus depth. This is explained in Figure 9.4.
One can think of the water element in contact with the bottom as a simple shear test on water.
Since water is a Newtonian fluid, the shear stress that it develops is proportional to the rate at
which it is sheared. This governs the equation in Figure 9.4.
9.4
Erosion Threshold and Erosion Categories
The critical velocity is the velocity at which the soil starts to erode. The critical shear
stress is the shear stress at which the soil starts to erode (Figures 9.1 and 9.2). Below these
values there is no erosion, above these values the soil erodes at a certain rate. This threshold
of erosion is very useful in engineering but it is not obvious that such a clear threshold truly
exists physically. Indeed a sample of granite, for example, has a very high critical shear stress.
Yet common sense tells us that a pebble made of granite and left under a dripping faucet for
20 million years would develop a hole. In this case, the critical shear as we conceive it would
not have been reached yet the rock would have been eroded. The reason for the hole in the
pebble may be that there is no such thing as a cyclic threshold for materials and that cyclic
stresses even very small can destroy any material bonds; it is only a matter of the number of
cycles to break the bond. So one has to accept a practical definition of the critical shear stress.
The critical shear stress is defined here as the shear stress corresponding to a rate of erosion of
1 mm/hr in the Erosion Function Apparatus. Values of critical shear stresses are shown in
Figure 9.2.
If the critical shear stress is exceeded, it becomes important to know how fast the soil
is eroding at a given velocity. The relationship between the erosion rate and the velocity or the
interface shear stress is a function. In order to quantify this erosion function using a single
number, the following scheme is proposed. It consists in placing the erosion function on the
erosion chart of Figure 9.5 and deciding what erodibility category fits best for the soil
considered. This approach holds promise to use only one number to characterize a function.
Work is ongoing to tie a number of soil types into erodibility categories.
9.5
Erodibility of Coarse-Grained Soils
Clean sands and gravels erode particle by particle. This has been observed on slow
motion videotapes. Three mechanisms seem to be possible: sliding, rolling, and plucking.
92
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
A simple sliding mechanism (Figure 9.6) consists of assuming that the soil particle is a
sphere, that the resultant force exerted by the water on the soil particle is a shear force parallel
to the eroding surface, and that neighboring particles do not exert forces on the particle being
analyzed because they move at the same rate. Electromagnetic and electrostatic forces
between particles are neglected because the analysis is done for a sand or a gravel particle. As
the velocity increases, the shear stress w imposed by the water on the particle becomes large
enough to overcome the friction between two particles staked on top of each other, and sliding
takes place. The critical shear stress wc is the threshold shear stress at which erosion is
initiated. Referring to Figure 9.6, horizontal equilibrium leads to (White 1940):
wc Ae = W tan φ
(1)
where Ae = effective friction area of the water on the particle; W = submerged weight of the
particle; and φ = friction angle of the interface between two particles. If the particle is
considered to be a sphere, (1) can be rewritten as
wc d (sD502/4) = (us uw) g (sD503/6) tan φ
or
wc = 2 (us uw) g D50 (tan φ)/3d
(2)
(3)
where d = ratio of the effective friction area over the maximum cross section of the spherical
particle; D50 = mean diameter representative of the soil particle size distribution; us and uw =
mass density of the particles and of water, respectively; and g = acceleration due to gravity.
Eq. (3) shows that the critical shear stress is linearly related to the particle diameter. Briaud et
al. (1999b) showed experimentally for a sand and a gravel tested in the EFA that an
approximate relationship is:
wc (N/m2) = D50 (mm)
(4)
Using (3) and (4), and assuming reasonable values for us, uw, g, and φ, leads to a value of d
equal to about 6. This value is many times higher than would be expected and shows that the
sliding mechanism is not the eroding mechanism, or at least not the only one involved.
A simple rolling mechanism (Figure 9.7) consists of assuming that the soil particle is a
sphere, that the resultant force exerted by the water on the soil particle is a shear force parallel
to the eroding surface, that neighboring particles do not impede the process, and that rotation
takes place around the contact point with the underlying particle. Electromagnetic and
electrostatic forces between particles are neglected because the analysis is done for a sand or a
gravel particle. At incipient motion and referring to Figure 9.7, moment equilibrium around
the contact point O leads (White 1940) to:
wc Ae a = W b
(5)
or
wc (d sD502/4) (D50/2 + D50(cose)/2) = (us uw) g (sD503/6) (D50(sine)/2) (6)
or
(7)
wc = 2 ((us uw) g D50 sine)/(3d (1 + cose))
Eq. (7) confirms that wc is linearly proportional to D50. For reasonable values of us, uw, and g,
and for d = 1, using (4) and (7) leads to e values equal to about 10–120, which is indicative
of a loose arrangement; indeed, the sand and the gravel tested to obtain Eq. 4 were placed in a
very loose condition in the EFA. Therefore it appears that rolling is more reasonable a
mechanism than sliding. Equation 7 tends to indicate that, while wc is linearly proportional to
D50, the proportionality factor may depend on the relative density. The dominant value of the
93
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
angle e can be obtained from a contact angle distribution diagram such as the ones shown in
Figure 9.8.
A simple plucking mechanism consists of assuming that the particles are cubes with a
side a. The water pressure on top of the cube is ut and the water pressure at the bottom of the
cube is ub. If it is assumed that all particles are plucked up at the same time, the differential
pressure between the top and bottom necessary to initiate plucking of the particle or block of
particles is:
W = (ubut) a2
(8)
or
us g a = ubut
(9)
The differential pressure ubut is made up of the hydrostatic differential water pressure (ubut)o
and the differential water pressure created by the flow Δu
or
ubut = (ubut)o + Δu
(10)
For a particle with a = 1 mm, the hydrostatic differential water pressure (ubut)o is 10 N/m2 .
This hydrostatic differential water pressure reduces the weight of the particle to its buyant
weight. The additional differential water pressure necessary to pluck the particle away Δu is
15 N/m2. This value of Δu is equivalent to 1.5 mm of water and it is easy to conceive that
such a small differential pressure can be developed. It is created dynamically by the water
flow including the fluctuations and the turbulence in the water. These pressure fluctuations
are very difficult to measure (Einstein and ElSamni, 1949 and Apperley, 1968). These
pressure fluctuations can be calculated through advanced numerical simulations.
These simplistic analyses of the sliding, rolling, and plucking mechanisms help to
clarify the important factors affecting the incipient motion of coarse grained soils. However,
they are not reliable for prediction purposes, and today experiments are favored over
theoretical expressions to determine wc for example. Shields (1936) ran a series of flume
experiments with water flowing over flat beds of sands. He plotted the results of his
experiments in a dimensionless form on what is now known as the Shields diagram. This data
as well as other data on sand gathered at Texas A&M University are plotted in Figure 9.2 as
critical shear stress wc versus mean grain size D50. Eq. (4) is shown in Figure 9.2 and seems to
fit well for sands. Shields did not perform any experiments on silts and clays. The data
developed for silts and clays at Texas A&M University show that Eq.(4) is not applicable to
fine grained soils and that D50 is not a good predictor of wc for those types of soils.
There seems to be consensus in using the shear stress applied by the water to the soil
at the soil water interface as the major parameter causing erosion. It is likely that the hydraulic
normal stress or pressure created by the water at that interface also contributes to the process..
Nevertheless, the use of the shear stress only has remained common practice and the role of
the normal stress that generates bursts of uplift forces during turbulent flow has yet to be
included in common approaches to scour.
9.6
Erodibility of Fine-Grained Soils
In the case of silts and clays, other forces come into play besides the weight of the
particles; these are the electrostatic and Van der Waals forces. Figure 9.10 and 9.11 show
cartoons of the forces and pressures acting on the soil particle in the general case. The water
94
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
pressure uw surrounds the particle if the soil is saturated. The contact forces fci exist at the
contact point and have normal as well as shear components. The electrostatic and Van der
Walls forces fei are also shown on the figure. Figure 9.10 refers to the case where the water is
not moving. In this case the water pressure is smaller on the top of the particle than on the
bottom of the particle but the difference is not significant. This difference is equal to the
hydrostatic pressure difference due to the height of the particle and creates the buoyancy of
the soil particle. In Figure 9.11, the water is moving and the difference between the top and
bottom water pressure has increased. Note that the water pressure uw and therefore the uplift
force on the particle is a function of time t and fluctuates during the flow. The cartoon shows
a situation where the water pressure may be such that the particle weight is overcome.
The electrostatic forces are likely to be repulsive because clay particles are negatively
charged. Van der Waals forces are relatively weak electromagnetic forces that attract
molecules to each other (Mitchell 1993); although electrically neutral, the molecules form
dipoles that attract each other like magnets. The Van der Waals forces are the forces that keep
H2O molecules together in water. The magnitude of these Van der Waals forces can be
estimated by (after Black et al. 1960):
f(N/m2) = 1028 / d(m)4
(11)
where d(m) = distance in m between soil particles; and f = attraction force in N/m2. By
multiplying f by the particle surface area, one can obtain the interparticle force. Table 9.1
shows the value of these forces for a sand and a clay particle.
In both cases the soil particle was assumed to be spherical and the distance between
particles was taken equal to the particle diameter. While such an evaluation of the Van der
Waals force can only be considered as a crude estimate, the following observations regarding
the numbers in Table 9.1 are interesting. First, the ratio between the weight and size of the
sand particle and the clay particle are similar to the ratio between the weight and size of a
Boeing 747 and a postage stamp; therefore, if the critical shear stress is proportional to the
particle weight, the critical shear stress for clays should be practically zero. Second, the ratio
between the Van der Waals force and the weight of the sand particle indicates that the Van der
Waals force is truly negligible for sands. Third, the same ratio for the clay particle, while 1017
times larger than for sand, also indicates that the Van der Waals forces are negligible
compared with the weight of the clay particle. This would lead one to think that the critical
shear stress, wc , is essentially zero for clays. Note that the electrostatic forces have not been
calculated here but since they are predominantly repulsive they would decrease, if anything,
the attraction due to the Van der Waals forces. Other phenomena give cohesion to clays; they
include water meniscus forces, such as those developing when a clay dries, and diagenetic
bonds due to aging, such as those developing when a clay turns into rock under pressure over
geologic time. Because of the number and complexity of these bonds, it is very difficult to
predict wc for clays empirically on the basis of a few index properties. Several researchers
however have proposed empirical equations for wc in clays, such as Dunn (1959) and Lyle and
Smerdon (1965).
One problem associated with measuring τc is determining the initiation of scour. When
the particles are visible to the naked eye, it is simple to detect when the first particle is
scoured away. For clays this is not the case, and various investigators define the initiation of
95
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
scour through different means; these vary from ‘‘when the water becomes muddy’’ to
extrapolation of the scour rate versus shear stress curve back to zero scour rate. Table 9.2
shows a variety of measured τc values. The lack of precise definition for the initiation of scour
may be in part responsible for the wide range of values.
Beyond the critical shear stress, a certain scour rate z˙ (mm/hr) is established. This
scour rate is rapid in sand, slow in clay, and extremely slow in rock. The example of the
Grand Canyon rock cited earlier leads to a value of z˙ equal to 9.1 x 106 mm/ hr, whereas fine
sands erode at rates of 104 mm/hr as measured in the EFA. Clays scour at intermediate rates
with common values in the range of 1 to 1,000 mm/hr. The high scour rate in sand exists
because once gravity is overcome, no other force slows the scour process down. The very low
scour rate in rock exists probably because it takes a large number of shearstress cycles
imposed by the turbulent nature of the flow to overcome the very strong crystalline bonds
binding the rock together. Note that rock scour can also occur at larger rates if the rock is
fractured and the water flow provides very high velocities as in the case of the downstream
end of high dam spillways. The low scour rate in clays is probably associated with the fact
that it takes a large number of shear stress cycles to overcome the electromagnetic bonds
created by the Van der Waals forces between clay particles. Even though these bonds are
relatively weak, as discussed previously, they are sufficient to slow the scour process
significantly. The scour rate z˙ versus shear stress τc curve (Figure 9.1) is used to quantify the
scour rate of a soil as a function of the flow. Several researchers have measured the rate of
erosion in cohesive soils; most have proposed a straight line variation (Ariathurai and
Arulanandan, 1978), while some have found S shape curves (Christensen, 1965). Some of the
rates quoted in the literature are given in Table 9.3.
Some of the factors influencing the erodibility of fine grained soils are listed in
Table 9.4. Although there are sometimes conflicting findings, the influence of various factors
on cohesive soil erodibility is shown in Table 9.4 when possible.
The critical shear stress of coarse grained soils is tied to the size of the particles and
usually ranges from 0.1 N/m2 to 5 N/m2. The rate of erosion of coarse grained soils above the
critical shear stress increases rapidly and can reach tens of thousands of millimeters per hour.
The most erodible soils are fine sands and silts with mean grain sizes in the 0.1 mm range
(Figure 9.2). The critical shear stress of fine grained soils is not tied to the particle size but
rather to a number of factors as listed in Table 9.4. The critical shear stress of fine grained
soils however varies within the same range as coarse grained soils (0.1 N/m2 to 5 N/m2) for
the most common cases. One major difference between coarse grained and fine grained soils
is the rate of erosion beyond the critical shear stress. In fine grained soils (often called
cohesive soils), this rate increases slowly and is measured in millimeters per hour. This slow
rate makes it advantageous to consider that erosion problems are time dependent and to find
ways to accumulate the effect of the complete velocity history rather than to consider a design
flood alone.
9.7
Erodibility and Correlation to Soil Properties
There is a critical shear stress wc below which no erosion occurs and above which
erosion starts. This concept while practically convenient may not be theoretically simple.
96
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Indeed, as seen on Figure 9.1, there is no obvious value for the critical shear stress. The
critical shear stress is arbitrarily defined as the shear stress which corresponds to an erosion
rate of 1 mm/hr. The critical shear stress is associated with the critical velocity vc. One can
also define the initial slope Si = (d z& /dw)i at the origin of the erosion function. Both wc and Si
are parameters which help describe the erosion function and therefore the erodibility of a
material.
In coarse grained soils (sands and gravels), the critical shear stress has been
empirically related to the mean grain size D50 (Briaud et al., 2001a).
wc (N/m2) = D50 (mm)
(12)
For such soils, the erosion rate beyond the critical shear stress is very rapid and one flood is
long enough to reach the maximum scour depth. Therefore there is a need to be able to predict
the critical shear stress to know if there will be scour or no scour but there is little need to
define the erosion function beyond that point because the erosion rate is not sufficiently slow
to warrant a time dependent analysis.
In fine grained soils (silts, clays) and rocks, equation 12 is not applicable (Figure 9.2)
and the erosion rate is sufficiently slow that a time rate analysis is warranted. Therefore it is
necessary to obtain the complete erosion function. An attempt was made to correlate those
parameters, wc and Si, to common soil properties in hope that simple equations could be
developed for everyday use. The process consisted of measuring the erosion function on one
hand and common soil properties on the other (water content, unit weight, plasticity index,
percent passing sieve no. 200, undrained shear strength). This lead to a database of 91 EFA
tests (Table 9.5) which was used to perform regression analyses and obtain correlation
equations (Figure 9.12 to 9.15). All attempts failed to reach a reasonable R2 value.
The fact that no relationship could be found between the critical shear stress or the
initial slope of the erosion function on one hand and common soil properties on the other
seems to be at odds with the accepted idea that different cohesive soils erode at different rates.
Indeed if different clays erode at different rates then the erosion function and therefore its
parameters should be functions of the soils properties. The likely explanation is that there is a
relationship between erodibility and soils properties but that this relationship is quite
complicated, involves advanced soil properties, and has not been found. Instead, it was found
much easier to develop an apparatus which could measure the erosion function on any sample
of cohesive soil from a site. This apparatus was called the Erosion Function Apparatus or
EFA.
9.8
The EFA: Erosion Function Apparatus
The EFA (Briaud et al. 1999, Briaud et al., 2001a) was conceived by Dr. Briaud in
1991, designed in 1992, and built in 1993 (Figure 9.16). The sample of soil, finegrained or
not, is taken in the field by pushing an ASTM standard Shelby tube with a 76.2 mm outside
diameter(ASTMD1587). One end of the Shelby tube full of soil is placed through a circular
opening in the bottom of a rectangular cross section pipe. A snug fit and an Oring establish a
leak proof connection. The cross section of the rectangular pipe is 101.6 mm by 50.8 mm. The
pipe is 1.22 m long and has a flow straightener at one end. The water is driven through the
pipe by a pump. A valve regulates the flow and a flow meter is used to measure the flow rate.
97
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
The range of mean flow velocities is 0.1 m/s to 6 m/s. The end of the Shelby tube is held flush
with the bottom of the rectangular pipe. A piston at the bottom end of the sampling tube
pushes the soil until it protrudes 1 mm into the rectangular pipe at the other end. This 1 mm
protrusion of soil is eroded by the water flowing over it.
9.8.1 EFA test procedure
The procedure for the EFA test consists of
1. Place the sample in the EFA, fill the pipe with water, and wait one hour.
2. Set the velocity to 0.3 m/s.
3. Push the soil 1 mm into the flow.
4. Record how much time it takes for the 1 mm soil to erode (visual inspection)
5. When the 1 mm of soil is eroded or after 30 minutes of flow whichever comes
first, increase the velocity to 0.6 m/s and bring the soil back to a 1 mm protrusion.
6. Repeat step 4.
7. Then repeat steps 5 and 6 for velocities equal to 1.0 m/s, 1.5 m/s, 2 m/s, 3 m/s, 4.5
m/s, and 6 m/s. The choice of velocity can be adjusted as needed.
9.8.2 EFA test data reduction
The test result consists of the erosion rate dz/dt versus shear stress w curve (Figure 9.1,
and 16). For each flow velocity v, the erosion rate dz/dt (mm/hr) is simply obtained by
dividing the length of sample eroded by the time required to do so.
dz/dt = h/t
(13)
Where h is the length of soil sample eroded in a time t. The length h is 1 mm and the
time t is the time required for the sample to be eroded flush with the bottom of the pipe (visual
inspection through a Plexiglas window). After several attempts at measuring the shear stress w
in the apparatus it was found that the best way to obtain w was by using the Moody Chart
(Moody, 1944) for pipe flows.
w = f u v2/8
(14)
Where w is the shear stress on the wall of the pipe, f is the friction factor obtained from
Moody Chart (Figure 9.17), u is the mass density of water (1000 kg/m3), and v is the mean
flow velocity in the pipe. The friction factor f is a function of the pipe Reynolds number Re
and the pipe roughness ε/D. The Reynolds number is Re = vD/q where D is the pipe diameter
and q is the kinematic viscocity of water (106 m2/s at 200C). Since the pipe in the EFA has a
rectangular cross section, D is taken as the hydraulic diameter D = 4A/P (Munson et al., 1990)
where A is the cross sectional flow area, P is the wetted perimeter, and the factor 4 is used to
ensure that the hydraulic diameter is equal to the diameter for a circular pipe. For a
rectangular cross section pipe:
D = 2ab/(a + b)
(15)
Where a and b are the dimensions of the sides of the rectangle. The relative roughness
ε/D is the ratio of the average height of the roughness elements on the pipe surface over the
pipe diameter D. The average height of the roughness elements ε is taken equal to 0.5D50
where D50 is the mean grain size for the soil. The factor 0.5 is used because it is assumed that
the top half of the particle protrudes into the flow while the bottom half is buried into the soil
98
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
mass. During the test, it is possible for the soil surface to become rougher than just 0.5 D50;
this occurs when the soil erodes block by block rather than particle by particle. In this case the
value used for ε is estimated by the operator on the basis of inspection through the test
window. Typical EFA test results are shown on Figure 9.1 for sand and then clay.
9.9
Some Existing Knowledge on Levee Erosion
9.9.1 Current Considerations in Design
The US Army Corps of Engineers’ design manual (USACE, 2000) outlines the steps
followed in the design and construction of levees (Table 9.6). The procedure does not include
an evaluation of the erodibility of the soils used for the levees. A more indepth discussion of
design requirements is presented in Chapter 10.
9.9.2 Failure Mechanism
Flowing water exerts a tractive shear stress along the soilwater interface. The erosion
process begins when this tractive shear stress exceeds the resistive force of the backslope soil
(AlQaser, 1991). Hanson et al. (2003) describe four stages of erosion during the overtopping
of cohesive embankments (Figure 9.18):
Minor headcut movement up to the downstream embankment crest; surface
erosion occurs.
Stage II: Headcut progresses from the downstream embankment crest to the
upstream embankment crest.
Stage III: The crest lowers and breach formation begins as the headcut continues to
migrate upstream of the embankment crest.
Stage IV: Erosion of the breach opening has progressed to near the base of the
upstream toe of the embankment; driven by erosion of the sidewalls and
development of an overhang, resulting in episodic mass failures and breach
widening (Hunt et al., 2004).
Stage I:
Erosion typically occurs adjacent to some change or interruption in the flow pattern
(Ralston, 1987). The turbulence associated with the flow disturbance breaks down the
protective boundary laminar flow layer. This leads to the occurrence of full hydraulic stress
intensity as well as rapid stress reversals, greatly increasing the erosion rate.
Gradually varied flow also leads to nonuniform erosion along the backslope
producing overfalls. The overfall will advance progressively headward as long as the
remaining embankment material can support the dam crest and upstream slope (Figure 9.19).
The base of the overfall will deepen and widen.
As the eroding vertical overfall face advances headward, the overflow crest elevation
will lower, cutting into the adverse grade of the upstream slope. This erosion pattern will
continue and progress until the flow pattern changes into a free surface flow (Figure 9.20,
AlQaser, Ruff, 1993). Headward advance of the overfalls is due to a combination of the
following:
99
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
1) Insufficient soil strength to stand vertically due to the height of the face, stress
relief cracking, and induced hydrostatic pressure in the cracks
2) Loss of foundation support for the vertical face due to the waterfall plunging effect
and its associated lateral and vertical scour. As the vertical overfall gets higher,
impact energy of the water fall increases, the rate of erosion increases and the
scour hole becomes larger. The supporting foundation of the overfall face and
sidewalls is thus removed.
The erosion pattern of embankments using noncohesive soil is affected by the
existence and location of a cohesive soil zone. For purely noncohesive embankments, the
erosion occurs on a uniform, but gradually flattening gradient. This erosion pattern can be
modeled using the theory of tractive stress. The breach development is consistent with the
principle of minimum rate of energy dissipation for streams (Coleman et al., 2002). Breaches,
like streams, tend to alter their geometry in order to produce a minimum rate of energy
dissipation. When the embankment includes a zone of cohesive soil, the overfall development
will be retarded. If the zone is symmetrical, erosion will behave similar to that of a cohesive
soil embankment. If the zone is an upstream sloping section, the overflow crest will degrade.
This is due to undermining of the downstream noncohesive zone. Portions of the
overhanging cohesive zone will subsequently break off as the allowable bending moment is
exceeded.
9.9.3 Numerical modeling
Erosion computer models are used to describe and quantify the complexities
associated with an embankment breach. OVERFALL, a computer program developed by
AlQaser (1991), predicts the heights and numbers of overfalls along the backslope of an
overtopped embankment. Key features of breaches can also be reproduced with SIMBA, or
SIMplified Breach Analysis (Temple et al., 2005). This model has been verified against
embankment breach tests. Presently, SIMBA is only capable of addressing homogeneous
embankment conditions. Future work will allow for applications to nonhomogeneous field
conditions, though.
Breach and discharge characteristics can be modeled and predicted with BREACH
(Fread, 1988). BREACH allows for predictions of the size, shape, and time of formation of
an earthen dam breach. A breach outflow hydrograph is also provided from the analysis. The
extent of the enlargement, the peak outflow, and the time to peak flow are determined by the
internal friction angle and the cohesive strength of the embankment soil. The BREACH
model was verified by comparing the results of the model and several overtopping failure
tests. These tests were conducted in different countries at varying scales with different
homogeneous materials and construction practices. A summary of dam break numerical
models that can be used for gradual failure is shown in Table 9.7.
9.9.4 Laboratory Tests
Nairn (1986) conducted twodimensional flume tests to study cohesive shore erosion.
Tests were conducted on artificial clay, composed of a bentonitesilt mixture, with and
without an overlying veneer of sand. Surprisingly, the flume tests with sand did not lead to
failure as the sand acted as an armor over the clay. Tests without sand, however, produced
9 10
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
responses close to those observed in the field. Table 9.8 displays the erosion rate results for
the flume tests conducted by Nairn. Dodge (1988) also conducted laboratory flume tests to
study erosion of a clayey sand (Figure 9.21). The results of the tests were not verified with
field observations; they serve to provide a qualitative assessment of erosion.
AlQaser (1991) performed two laboratory tests to study progressive failure of an
overtopped embankment (Figure 9.22). Both tests had the same design, but differed in the
percent of sand in the soil. The first sample consisted of 80% clay and 20% sand. The other
had 50% clay and 50% sand. Results show that the presence of more clay in the soil mixture
leads to a greater vertical overfall height. The soil with more sand in the mixture, however,
resulted in more horizontal overfall regression. It is concluded, therefore, that the physical
and the geometrical properties of the embankment affect the number and heights of the
developed overfalls.
9.9.5 Field Tests
Hanson, Cook, and Hahn (2001) describe preliminary evaluation of the headcut
migration rates during overtopping and breaching tests on largescale models. The headcut
advance threshold was evaluated based on an energy dissipation term:
E @ qj w H
(16)
Where q = unit discharge, jw = unit weight of water, and H = Headcut height. The
headcut migration rates for each test section were evaluated and compared to measured soil
properties, such as erodibility and soil strength. The results show that as soil strength
decreases, the headcut migration rate increases (Figure 9.23).
The breaching of noncohesive homogeneous embankments under constantreservoir
levels was studied using flume tests (Figure 9.24) by Coleman et al. (2002). This experiment
simulated the failure of an embankment restricting a very large upstream reservoir. A small V
breach was initiated and grew as erosion took place. A wide range of uniform noncohesive
soils were tested. The quantitative findings of these tests have not been verified by the results
from large scale embankments.
It was found from the flume tests conducted by Coleman et al. (2002) that erosion
progresses from primarily vertical to lateral in nature. This occurs as the breach channel
invert approaches the foundation level. The channel invert slope will flatten as it rotates
about a fixed pivot point, XP, on the embankment (Figure 9.25).
The location of this pivot point is a function of the embankment sediment size. In plan
view, the breach channel develops into an hourglass (or Venturi) shape (Figure 9.26). The
curvature of the channel increases with time until the embankment foundation impedes the
vertical erosion of the breach. This leads to an increase in the rounding of the approach and
exit channels.
After the preliminary studies of 2001, Hanson, Cook, et al. (2003) performed a second
study of the headcut migration and erosion widening rates during overtopping. They used
largescale models and three soils including two nonplastic (SM) silty sand materials and a
9 11
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
(CL) lean clay. The width of the breach during testing was evaluated using photographic
measurements of the model embankment (Figure 9.27). Details of the testing indicated that
headcut erosion was an important erosion process in the failure of cohesive embankments. It
can influence the breach initiation time, breach formation time, breach width, peak discharge,
and the overall outflow hydrograph.
The headcut migration rates (Figure 9.28), as well as the erosion widening rates
(Figure 9.29), show a direct correlation to the compaction water content. The rate of breach
widening was found by taking the linear regression of the breach width measurement from
left bank to right bank versus time (Figure 9.30). The observed breach widths during testing
were equal to two to five times the dam height. Figure 9.31 indicates that the head cutting
rate for stages II and III of the erosion process is larger than the widening rate at the
beginning but becomes approximately equal to it towards the end of the breaching process.
9.9.6 Factors Influencing Resistance to Overtopping
For a given soil, Hanson et al. (2003) show that erodibility correlates well with
compaction water content, energy, density, and texture. By contrast, Cao et al. (2002) using a
large data base found no relationship between common soil properties and the erodibility of
cohesive soils. Dodge (1988, Figure 9.32) gave some trends of erodibility for cohesive soils
using the plasticity chart. The FHWA (Chen, Cotton, 1988) also presents a plot of
permissible shear stresses for cohesive soils based on the plasticity index (Figure 9.33).
Choliaras et al. (2003) concludes that the main measure of erosivity of overland flow
is shear stress flow. He states that the increase of erosion rate is linear with shear stress of
flow. He adds that for low values of surface shear stress, the erodibility of a soil decreases
with increasing soil strength while for high values of surface shear stress, the erodibility of the
soil increases with increasing surface strength.
According to Fread (1988), the growth of a breach is dependent on the soil properties
of the dam. Unit weight, friction angle, and cohesive strength are shown to influence the size,
shape, and time of formation of a breach. The amount of grass cover on the dam is also a
factor in breach formation.
The results of the research performed by AlQaser (1991) point to poor compaction as
a source of breaching. According to model tests conducted by Dodge (1988), the volume of
scour produced during flow can be decreased by increasing the compaction of the soil.
Similarly, for clay soils, an increase in density reduces erodibility (Choliaras et al.,
2003). For silty and sandy soils, the density or compaction of the soil does not significantly
influence erodibility.
1953 Levee (Dike) failures in Netherlands
The Netherlands is a country of 8.5 Million people and 26% of them live below mean sea
level protected by levees (Gerritsen, 2006). The following is a summary of an excellent article
by Gerritsen in GeoStrata (2006) which describes the 1953 disaster and the steps taken since
9 12
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
then by the Netherlands. Prior to 1953 the dikes were at a height equal to the maximum
recorded water level plus 0.5 m. The height of some of the levees had been increased by
constructing concrete walls along the levee crest. During World War II, the levees were used
as a defense system and many holes were dug to that effect. After the war, the damage done to
the levees was not adequately repaired.
On January 31, 1953, a North Sea storm combined with hide tide and raised the water level to
unprecedented height and 150 levee breaches occurred. During that storm, 1836 people died,
100,000 people evacuated, tens of thousands of livestock perished, and 136,500 hectares were
inundated. The levee breaches were attributed to sustained wave overtopping. The land side of
the levees was typically at a steeper slope (1v to 1.5h or 1v to 2h) than the sea side (1v to 3h
or more). The failure process initiated from the land side and progressed backward towards
the sea side. One sign of imminent failure was a longitudinal crack forming along the crest of
the levee which was quickly filled by the rushing water.
On February 18, 1953, a committee was formed called the Delta Committee with the task of
ensuring that such a disaster would not happen again. The committee chose to solve the
problem not by increasing the height of the levees but rather by recommending the Delta Plan.
This plan consisted of closing the shoreline completely through a series of permanent barriers
to be built over a 20 year period. In 1975, due to political pressures from the fishing industry,
the barriers were changed from complete damming to moveable storm surge barriers to be
closed only in the event where a North Sea storm would coincide with a high tide.
The Netherlands now requires that the flood protection systems satisfy the following
ク
ク
ク
Be able to resist a storm surge with a probability of occurrence of 1/10,000 for the
Province of Holland;
Be able to resist a storm surge with a probability of occurrence of 1/4000 for less
populated coastal areas; and
Be reviewed and evaluated every 5 years with associated recommendations to be
constructed in the following 5 years.
9.9.7 Influence of Grass Cover on Surface Erosion
Grass makes a difference in the resistance to surface erosion (Figure 9.34). The
physical vegetative coverage on slopes provides increased resistance through underground
soil reinforcement and surface protection (Li and Eddleman, 2002). Root systems aid slope
stabilization through soilroot interaction. The mechanics of rootreinforcement are similar to
the basic mechanics of engineering reinforcedearth systems (Coppin and Richards, 1990).
The vegetation root growth reinforces the upper soil layers increasing the soil shear strength
by over 33 % (Bhandari et al., 1998). Many researchers have developed theoretical models of
rootreinforced soils, including Gray and Leiser (1982), Greenway (1987), Coppin and
Richards (1990), Styczen and Morgan (1995), and Wu (1995). In general, the vegetative
methods for surface erosion control include two types: herbaceous and woody. They all have
the following four mechanisms in controlling surface erosion (Gray 1974; Greenway, 1987;
Coppin and Richards, 1990):
9 13
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
1.
2.
3.
4.
Restraint: The root system binds the soil particles. The foliage residues restrain
soil particle detachment via shallow, dense root systems, consequently
reducing sediment transport.
Retardation: The foliage and stems increase the surface roughness and slow
surface runoff.
Interception: The foliage and plant residues absorb the rainfall energy by
intercepting the raindrops to reduce raindrop impacts.
Transpiration: Absorption of soil moisture by plants delays the initiation of
saturation and increases shear strength by reducing porepressures.
The level of vegetation for protecting the soil depends on the combined effects of
roots, stems and foliage (Coppin and Richards, 1990). Woody vegetation installed on slopes
and streambanks provides resistance to shallow massmovement by counterbalancing local
instabilities. In order to achieve optimum stabilization, vegetation must establish quickly and
solidly. For biotechnical stabilization techniques that only use vegetative materials, the
stabilization is vulnerable at the early stage but becomes stronger as the vegetation is
established (Li and Eddleman, 2002). For techniques that combine plant and inert materials
such as dead wood, rocks or geosynthetics, inert materials support major loads at the early
stage. As the vegetation matures, root systems will bind soils, inert materials and vegetation
altogether on the slope or streambank, and increase the safety factor of structural protection
(Biedenharn et al., 1997).
From the engineering perspective, vegetation’s use on slopes or streambanks may not
be always ideal. Trees planted on certain parts of levees may have roots undermining the
levee stability (USACE, 1999). Greenway (1987), and Coppin and Richards (1990) have
analyzed vegetation’s engineering functions and determined that its effects are both adverse
and beneficial, depending on the circumstances. Therefore, selecting appropriate plant type
becomes very critical in such conditions. This can be done by testing at large scale facilities
such as the one at Texas A&M University which grows grass and tests it on slopes of various
geometries.
Johnston (2003) prepared the chart of Figure 9.35 which gives the allowable shear
stress at the interface between the soil and the water flowing on a slope. Different covers are
represented including bare soil, grass covered, geosynthetic matting, hard armor. The depth D
is the depth of water flowing over the slope S. Note that overall the range of slope covered is
fairly shallow.
9.10 Soil and Water Samples Used for Erosion Tests
A total of 11 locations were identified for studying the erosion resistance of the levee
soils. Emphasis was placed on levees which were very likely overtopped. These locations are
labeled S1 through S15 for Site 1 through Site 15 on Figure 9.36. The samples were taken by
pushing a Shelby tube when possible or using a shovel to retrieve soil samples into a plastic
bag. For example at Site S1, the drilling rig was driven on top of the levee, stopped at the
location of Site 1, a first Shelby tube was pushed with the drilling rig from 0 to 2 ft depth and
then a second Shelby tube was pushed from 2 to 4 ft depth in the same hole. These two
Shelby tubes belonged to boring B1. The drilling rig advanced a few feet and a second
9 14
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
location B2 at Site S1 was chosen; then two more Shelby tubes were collected in the same
way as for B1. This process at Site S1 generated 4 Shelby tube samples designated
ク
ク
ク
ク
S1B1(02ft)
S1B1(24ft)
S1B2(02ft)
S1B2(24ft)
Four such Shelby tubes were collected from sites S1, S2, S3, S7, S8, and S12. In a
number of cases, Shelby tube samples could not be obtained because access for the drilling rig
was not possible (e.g.: access by light boat for the MRGO levee) or pushing a Shelby tube did
not yield any sample (clean sands). In these cases, grab samples were collected by using a
shovel and filling a plastic bag. The number of bags collected varied from 1 to 4. Plastic bag
samples were collected from sites S4, S5, S6, S11, and S15. The total number of sites sampled
for erosion testing was therefore 11. These 11 sites generated a total of 23 samples. One of the
samples, S8B1(24ft), exhibited two distinct layers during the EFA tests and therefore lead
to two EFA curves. All in all 24 EFA curves were obtained from these 23 samples: 14
performed on Shelby tube samples and 10 on bag samples. The reconstitution of the bag
samples in the EFA is discussed later.
Water salinity has an effect on erosion. The salinity of the water was determined by
using the soil samples collected at the sites. Samples S11 and S15 were selected because one
was on the Lake Pontchartrain side and the other on the Lake Borgne side. The procedure to
obtain theconsisted of:
1. Dry the soil (about 70 g) in an oven for 12 hr
2. Weigh a quantity of soil, e.g. 10 g and place it in a PE bottle
3. Add deionized (DI) water in the ratio of 2 ml water for one sample and 5 ml water for
another sample to each gram of soil
4. Soil: DI water = 10 g: 20 ml or 10g: 50 ml
5. Shake the bottle to thoroughly mix the soil and water
6. Allow the soil to settle for 12 hr
7. Use a pH meter (Orion model 420 A) to measure the pH and a calibrated conductivity
meter (Corning model 441) to measure the conductivity of the water.
8. Perform a calibration of the conductivity meter by using known concentrations of salt.
9. Use the conductivity to salinity calibration curve to obtain the salinity of the water
created in steps 1 to 7.
Then it becomes necessary to correct the salinity of this water because the amount of
water added to the soil for the salinity determination test does not correspond to the amount of
water available in the soil pores in its natural state (in the levee). This is done by calculating
the amount of water available in the pores of the samples in its natural state. This requires the
use of the void ratio and the degree of saturation of the samples calculated using simple phase
diagram relationships. The results obtained are shown in Table 9.9.
9 15
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
9.11 Erosion Function Apparatus (EFA) Test Results
9.11.1 Sample Preparation
No special sample preparation was necessary for the samples which were in Shelby
tubes. The Shelby tube was simply inserted in the hole on the bottom side of the rectangular
cross section pipe of the FEA (described previously).
For bag samples obtained by using a shovel to collect the soil, there was a need to
reconstruct the sample. These samples were prepared by recompacting the soil in the Shelby
tube (Figure 9.37). The same process as the one used to prepare a sample for a Proctor
compaction test was used. Since it was not known what the compaction level was in the field,
two extreme levels of compaction energy were used to recompact the samples. The goal was
to bracket the erosion response of the intact soil.
For the high compaction effort (100% of Modified Proctor compaction effort), the
sample was compacted in an 18inch long Shelby tube as follows:
1) The total sample height was 6 inches. The sample was compacted in eight layers.
2) To form each layer, the soil was poured into the Shelby tube from a height of 1
inch above the top of the tube.
3) The soil was compacted using a 10 lb hammer (Modified Proctor hammer) with a
drop height of 1.5 feet. Each layer was compacted by 8 hammer blows, i.e. 8
blows/layer.
4) This process was repeated until a 6 inch sample was obtained.
5) The corresponding compaction energy was equal to the Standard Modified Proctor
Compaction energy.
For the low compaction effort (1.63% of Modified Proctor compaction effort), the sample was
compacted in an 18inch long Shelby tube as follows.
1. The total sample height was 6 inches. The sample was compacted in eight layers.
2. To form each layer, the soil was poured into the Shelby tube from a height of 1
inch above the top of the tube.
3. The soil was compacted using a 10 lb hammer (Modified Proctor hammer) with a
drop height of 1 inch. Each layer was compacted by 3 hammer blows, i.e. 3
blows/layer.
4. This process was repeated until a 6 inch sample was obtained.
5. The corresponding compaction energy was 1.63% of the Standard Modified
Proctor Compaction energy.
9.11.2 Sample EFA Test Results
The procedure described earlier was strictly followed for the EFA tests. The results
were prepared in the form of a word file report and an accompanying excel spread sheet
detailing the data reduction and associated calculations. The main result of an EFA test is a
couple of plots: one is the plot of the erosion rate versus mean velocity in the EFA pipe, the
other is the plot of the erosion rate versus shear stress at the interface between the soil and the
9 16
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
water. These two plots are collected in Appendix A for all 24 EFA tests. Figs. 9.38 and 9.39
show two examples of results for a very erodible soil and a very erosion resistant soil.
9.11.3 Summary Erosion Chart
In an effort to give a global rendition of the EFA results, an erosion chart was created.
The blank erosion chart has been presented earlier and is reproduced here for convenience
(Figure 9.40). This chart allows one to present the erosion curves in a way which categorizes
the soils according to one erosion category. Category 1 is very erodible and refers to soils
such as clean fine sands. Category 5 is very erosion resistant and refers to soils such as some
of the highly compacted and well graded clays.
Figure 9.41 shows the erosion chart populated with the EFA results for all 24 EFA
tests. The legend contains the sample/test designation which starts with the site number
(Figure 9.36), followed by the boring number, the depth, and letter symbols including SW,
TW, LC, HC, and LT. SW stands for Sea Water and means that the water used in the EFA test
was salt water at a salinity of approximately 35000 ppm. TW stands for Tap Water and means
that the water used in the EFA test was Tap Water at a salinity of approximately 500 ppm. LC
stands for Low Compaction, refers to bag samples only, and means that the sample was
prepared using 1.6% of Modified Proctor compaction effort. HC stands for High Compaction,
refers to bag samples only, and means that the sample was prepared using 100% of Modified
Proctor compaction effort. LT stands for Light Tamping and refers to the preparation of some
bag samples used in some early tests; it is very similar to the LC preparation.
One of the first observations coming from the summary erosion chart on Figure 9.41 is
that the erodibility of the soils obtained from the New Orleans levees varies widely all the
way from very high erodibility (Category 1) to very low erodibility (Category 5). This
explains in part why some of the overtopped levees failed while other overtopped levees did
not. This finding points to the need to evaluate the remaining levees for erodible soils (weak
links).
9.11.4 Influence of Compaction on Erodibility
Several of the bag samples were tested at two extreme compaction efforts: 100%
Modified Proctor and 1.6% Modified Proctor. Because the low and high compaction samples
originated from the same bag of collected soil, it is reasonable to assume that the samples are
very similar. The EFA tests results aimed at identifying the influence of the compaction effort
are isolated in Figure 9.42. Sample S4 shows a major influence of the compaction effort on
the erodibility. Indeed, the low compaction sample is at the border between Category 1 and
Category 2 (very high to high erodibility) while the high compaction sample is at the border
between Category 4 and Category 5 (very low to low erodibility). However, Samples S15 and
S11 do not show much difference between the high compaction and the low compaction.
The index properties of the samples tested are presented in a following section.
Sample S4 is a high plasticity silt. It has 90.47 % fines, a plasticity index of 30, and a USCS
classification of MH. Sample S11 is a clean uniform sand It has 0.1 % fines, and a USCS
classification of SP. Sample S15 is a silty sand. It has 29.89 % fines, and a USCS
classification of SM. These three data points tend to indicate that compaction has a more
9 17
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
significant influence on erodibility for some soils (higher fines content) than for others (lower
fine content).
9.11.5 Influence of water salinity on erodibility
Salinity can have an influence on the erodibility of a soil. Several of the samples were
tested by using water at two extreme salt concentrations: 35000 ppm to simulate sea water and
500 ppm to simulate water with a very low salt concentration. Because the samples used to
check the influence of the water salinity originated from different Shelby tubes at two
different depths (02 ft and 24 ft), it is possible that the samples may have had different
erodibility to start with. This may have clouded the influence of the water salinity.
The EFA tests results for the tests aimed at identifying the influence of the water
salinity are isolated in Figure 9.43. Conclusions are difficult to draw because the samples may
not be from the same soil. One sample (S8B1) actually was made of two separate layers
which had two different erosion functions and lead to two EFA curves for the same Shelby
tube.
Nevertheless, the following observations can be made. Samples S12 show that an
increase in water salinity increases the resistance to erosion, samples S2 and S8 show no
influence, while samples S1 and S7 show a reverse influence of the water salinity. The index
properties of the samples tested are presented in a following section. All samples exhibit a
high clay content.
9.12 Index Properties of the Samples Tested in the EFA
A set of index property tests were performed on the samples used in the EFA. Some of
the tests were performed by Soil Testing Engineering in Baton Rouge, the remainder of the
tests were performed at Texas A&M University. Table 9.10 shows a summary of the results as
well as the classifications according the Unified Soil Classification System. As can be seen
there are no gravels, and mostly sands, silts, and clays.
9.13 Levee Overtopping and Erosion Failure Guideline Chart
In an effort to correlate the results of the EFA erosion tests with the behavior of the
levees during overtopping flow, Figure 9.44 was prepared. It seems reasonably sure that the
levees at sites S4, S5, S6, and S15 were overtopped and failed. At the same time it seems
reasonably sure that the levees at sites S2, and S3 were overtopped and resisted remarkably
well. The dark circles on Figure 9.44 correspond to samples taken from levees that were
overtopped and failed by erosion while the open circles correspond to samples taken from
levees that were overtopped and held up well during that overtopping.
Figure 9.44 shows a definite correlation between the EFA tests results and the
behavior of the levees during overtopping. Figure 9.45 was generated from Figure 9.44 as a
levee guideline for erosion resistance during overtopping. It is suggested that such EFA
erosion tests should be used in the future to predict levee behavior and ensure erosion
resistance to overtopping. In addition, this type of testing can be performed on an increased
variety of soils, and with varied compaction conditions, to develop generalized relationships
9 18
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
between soil types and soil characteristics, placement and compaction conditions, and
resistance to erosion.
9.14 Summary
The results of this pilot study show that we are well able to correlate soil type, soil
characteristics, and placement and compaction conditions with embankment performance
with regard to erodibility. Moreover, we are also able to perform erodibility tests of specific
soils, and for specific placement and compaction conditions, and the results of these tests
appear to correlate well with observed field performance with regard to erodibility during
levee overtopping in the New Orleans region during hurricane Katrina.
Accordingly, it is clearly possible to identify and avoid the use of materials that can be
expected to perform poorly with regard to erosion resistance, and it is feasible to design
engineered embankments with a high intrinsic resistance to erosion. That would have been
very useful in the New Orleans regional flood protection system, and it appears that avoiding
the use of highly erodeable levee embankment fills, and using instead embankment fills
engineered to provide improved erosion resistance, would likely have significantly reduced
both damages and loss of life in this event.
There is more to be done in further developing and refining these testing procedures
and developing generalized correlations between material characteristics and placement
conditions vs. erodibility, and also with regard to the development of corollary analysis
methods and procedures for making engineering assessments regarding likely rates of erosion,
etc., as a function of overtopping intensities, geometries and durations. Nonetheless, it
appears that we are able to include resistance to erosion as a deliberately engineered feature of
levee embankments. As this adds a potentially important source of additional system
resilience, this should be considered in the future for flood protection systems defending large
populations at risk.
9.15 References
AlQaser, G., and Ruff, J. F., (1993), “Progressive failure of an overtopped embankment,”
Proceedings of the 1993 National Conference on Hydraulic Engineering, Hydraulic
Division of ASCE, pp. 19571962.
AlQaser, G.N., (1991), Progressive Failure of an Overtopped Embankment, Ph.D.
dissertation, Colorado State University, Fort Collins, CO.
Apperly, L.W., 1968, “Effect of turbulence on sediment entrainment”, PhD dissertation,
University of Auckland, Auckland, New Zealand.
Ariathurai, R., and Arulanandan, K. (1978). ‘‘Erosion rates of cohesive soils.’’ J. Hydr. Div.,
ASCE, 104(2), 279–283.
Arulanandan, K. (1975). ‘‘Fundamental aspects of erosion in cohesive soils.’’ J. Hydr. Div.,
ASCE, 101(5), 635–639.
9 19
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Arulanandan, K., Loganathan, P., and Krone, R. B. (1975). ‘‘Pore and eroding fluid influence
on surface erosion of soil.’’ J. Geotech. Engrg. Div., ASCE, 101(1), 51–66.
ASTM D1587, American Society for Testing and Materials, Philadelphia, USA.
Bhandari, G., Sarkar, S. S., and Rao, G. V. (1998). Erosion control with geosynthetics, Geo
horizon: State of art in geosynthetic technology, A.A. Balkema, Rotterdam,
Netherlands.
Biedenharn, D. S., Elliott, C. M., and Watson, C. C. (1997). The WES Stream Investigation
and Streambank Stabilization Handbook. US Army Engineer Waterways Experiment
Station, Vicksburg, MS.
Briaud J.L., Ting F., Chen H.C., Cao Y., Han S.W., Kwak K., “Erosion Function Apparatus
for Scour Rate Predictions”. Journal of Geotechnical and Geoenvironmental
Engineering, ASCE, Vol.127, No. 2, 2001a, pp.105113.
Briaud J.L., Ting F., Chen H.C., Gudavalli R., Kwak K., Philogene B., Han S.W., Perugu
S., Wei G.S., Nurtjahyo P., Cao Y.W., Li Y., “SRICOS: Prediction of Scour Rate at
Bridge Piers, TTI Report no. 29371 to the Texas DOT, 1999, Texas A&M University,
College Station, Texas, USA.
Briaud, J. L., Chen, H. C. Kwak K., Han SW., Ting F., “Multiflood and Multilayer Method
for Scour Rate Prediction at Bridge Piers”, Journal of Geotechnical and
Geoenvironmental Engineering, ASCE, Vol.127, No. 2, 2001b, pp.105113.
Briaud, J.L., Ting, F., Chen, H.C., Gudavalli, S.R., Perugu, S., and Wei, G., “SRICOS:
Prediction of Scour Rate in Cohesive Soils at Bridge Piers”, ASCE Journal of
Geotechnical Engineering, Vol.125, 1999, pp. 237246.
Cao Y., Wang J. ,Briaud J.L., Chen H.C., Li Y. Nurtjahyo P., 2002, “EFA Tests and the
influence of Various Factors on the Erodibility of Cohesive Soils”, Proceedings of the
First International Conference on Scour of Foundations, Texas A&M University, Dpt.
of Civil Engineering, College Station.
Cao, Y., “The Influence of Certain Factors on the Erosion Functions of Cohesive Soil”,
Master Thesis, 2001, Texas A&M University, College Station, Texas, USA
Chen, Y.H., Cotton, G.K., (1988), “Design of Roadside Channels with Flexible Linings,”
Federal Highway Administration, Hydraulic Engineering Circular No. 15.
Choliaras, I.G., Tantos, V.A., Ntalos, G.A., Metaza, X.A., (2003), “The influence of soil
conditions on the resistance of cohesive soils against erosion by overland flow,”
Journal of International Research Publication, Issue 3.
Christensen, B. A. (1965). ‘‘Discussion of ‘Erosion and deposition of cohesive soils,’ by E.
Partheniades.’’ J. Hydr. Div., ASCE, 91(5), 301– 308.
9 20
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Coleman S.E., Andrews D.P., Webby M.G., (2002), “Overtopping Breaching of Non
Cohesive Homogeneous Embankments”, Journal of Hydraulic Engineering, Vol. 138,
No. 9, ASCE, Reston, Virginia, USA.
Coppin, N. J., and Richards, I. G. (1990). Use of Vegetation in Civil Engineering.
Butterworths, London.
Dodge, R.A., (1988), Overtopping Flow on Low Embankment Dams – Summary Report of
Model Tests, RECERC883, U.S. Bureau of Reclamation, Denver, Colorado.
Dunn, I. S. (1959). ‘‘Tractive resistance to cohesive channels.’’ J. Soil Mech. and Found.
Div., ASCE, 85(3), 1–24.
Einstein, H.A., ElSamni, E.S., 1949, “Hydrodynamic forces on a rough wall”, Review of
modern physics, 21, 520524.
Enger, P. F., Smerdon, E. T., and Masch, F. D. (1968). ‘‘Erosion of cohesive soils.’’ J. Hydr.
Div., ASCE, 94(4), 1017–1049.
Fread, D.L., (1988), BREACH: An Erosion Model for Earthen Dam Failures, National
Weather Service, National Oceanic and Atmospheric Administration, Silver Spring,
Maryland.
Gerritsen, H. (2006), “The 1953 Dike Failures in the Netherlands, GeoStrata, MarchApril
2006, p1821, GeoInstitute of the American Society of Civil Engineers, Reston,
Virginia, USA.
Gray, D. H. (1974). “Reinforcement and stabilization of soil by vegetation.” J. Geotech. Engr.
Div., ASCE, 100(GT6), 695699.
Gray, D. H., and Leiser, A. T. (1982). Biotechnical Slope Protection and Erosion Control.
Van Nostrand Reinhold, New York.
Gray, D. H., and Sotir, R. B. (1996). Biotechnical and Soil Bioengineering Slope
Stabilization: a Practical Guide for Erosion Control. John Wiley & Sons, Inc., New
York.
Greenway, D. R. (1987). Vegetation and slope stability. In: M.G. Anderson, K.S. Richards
(Eds.). Slope Stability: Geotechnical Engineering and Geomorphology, John Wiley &
Sons, New York, pp. 187230.
Hanson, G.J., Cook, K.R., (2004), “Apparatus, test procedures, and analytical methods to
measure soil erodibility in situ,” Applied Engineering in Agriculture, American
Society of Agricultural Engineers, Vol. 20, No. 4, pp. 455462.
Hanson, G.J., Cook, K.R., Hahn IV, W., (2001), “Evaluating headcut migration rates of
earthen embankment breach tests,” Written for Presentation at the 2001 ASAE Annual
International Meeting, Paper No. 01012080, Sacramento, California.
9 21
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Hanson, G.J., Cook, K.R., Hahn IV, W., Britton, S.L., (2003), “Evaluating erosion widening
and headcut migration rates for embankment overtopping tests”, Written for
Presentation at the 2003 ASAE Annual International Meeting, Paper No. 032067, Las
Vegas, Nevada.
Hanson, G.J., Cook, K.R., Hahn IV, W., Britton, S.L., (2003), “Observed Erosion Processes
During Embankment Overtopping Tests,” Written for presentation at the 2003 ASAE
Annual International Meeting, Paper No. 032066, Las Vegas, Nevada.
Hanson, G.J., Cook, K.R., Hunt, S., (2005), “Physical modeling of overtopping erosion and
breach formation of cohesive embankment,” Transactions of the ASABE, Vol. 48, No.
5, pp. 17831794.
Hanson, G.J., Simon, A., (2001), “Erodibility of cohesive streambeds in the loess area of the
Midwestern USA”, Hydrological Processes, Vol. 15, pp. 23–38.
Hanson, G.J., Temple, D.M., Morris, M., Hassan, M., Cook, K., (2005), “Simplified Breach
Analysis Model For Homogeneous Embankments: Part II, Parameter Inputs And
Variable Scale Model Comparisons,” Proceedings Of 2005 U.S. Society On Dams
Annual Meeting And Conference, Salt Lake City, Utah, pp. 163174.
Hunt, S.L., Hanson, G.J., Cook, K.R., Kadavy, K.C., (2004), “Breach Widening Observations
from Earthen Embankment Tests,” American Society of Agricultural
Engineers/Canadian Society for Engineering Annual International Meeting, Ottawa,
Canada, Paper No. 042080.
Johnston A. (2003), “Range of Allowable Shear Stresses by Depth and Slope”, personnal
communication.
Kelly, E. K., and Gularte, R. C. (1981). ‘‘Erosion resistance of cohesive soils.’’ J. Hydr. Div.,
ASCE, 107(10), 1211–1224.
Li, M.H., and Eddleman, K. E. (2002). “Biotechnical engineering as an alternative to
traditional engineering methods: a biotechnical streambank stabilization design
approach.” Landscape and Urban Planning 60(4), 225242.
Lyle, W. M., and Smerdon, E. T. (1965). ‘‘Relation of compaction and other soil properties to
erosion and resistance of soils.’’ Trans., ASAE, American Society of Agricultural
Engineers, St. Joseph, Mich., 8(3).
Mitchell, J.K., 1993, “Fundamentals of Soil Behavior”, 2nd Ed., John Wiley and Sons, New
York.
Moody L.F., "Friction Factors for Pipe Flow", Transaction of the American Society of Civil
Engineers, 1944, Vol. 66, Reston, Virginia, USA.
Munson, B. R., Young, D. F., and Okiishi, T. H. (1990). Fundamentals of fluid mechanics.
Wiley, New York.
9 22
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Nairn, R.B., (1986), “Physical Modeling of Wave Erosion on Cohesive Profiles,”
Proceedings, Symposium on Cohesive Shores, Burlington, Ontario.
Ralston, D.C., (1987), “Mechanics of embankment erosion during overflow,” Proceedings of
the 1987 National Conference on Hydraulic Engineering, Hydraulics Division of
ASCE, pp. 733738.
Richardson, E. V., and Davis, S. R. (1995). ‘‘Evaluating scour at bridges.’’ Rep. No. FHWA
IP90017 (HEC 18), Federal Highway Administration, Washington, D.C.
Shaikh, A., Ruff, J. F., and Abt, S. R. (1988). ‘‘Erosion rate of compacted Namontmorillonite
soils.’’ J. Geotech. Engrg., ASCE, 114(3), 296– 305.
Shields, A. (1936). ‘‘Anwendung der AehnlichkeitsMechanik und der Turbulenzforschung
auf die Geschiebebewugung.’’ Preussische Versuchsanstalt fu¨r Wasserbau and
Schiffbau, Berlin (in German).
Shields, A., “Application of similarity principles, and turbulence research to bedload
movement,” 1936, California Institute of Technology, Passadena (translated from
German).
Smerdon, E. T., and Beasley, R. P. (1959). ‘‘Tractive force theory applied to stability of open
channels in cohesive soils.’’ Res. Bull. No. 715, Agricultural Experiment Station,
University of Missouri, Columbia.
Styczen, M. E., and Morgan, R. P. C. (1995). Engineering properties of vegetation. In: R.P.C.
Morgan, R.J. Rickson (Eds.). Slope Stabilization and Erosion Control: A
Bioengineering Approach. E & FN Spon, London, pp. 558.
Temple, D.M., Hanson, G.J., Neilsen, M.L., Cook, K.R., (2005), “Simplified Breach Analysis
Model For Homogeneous Embankments: Part I, Background And Model
Components,” Proceedings Of The 2005 U.S. Society On Dams Annual Meeting And
Conference, Salt Lake City, Utah, pp. 151161.
USACE (1999). Guidelines for Landscape Planting and Vegetation Management at
Floodways, Levees, and Embankment Dams. Engineering Manual 11102301, U.S.
Army Corps of Engineers, Washington, D.C.
USACE, (2000), “Engineering and Design – Design and Construction of Levees,” Publication
No. EM 111021913.
Vanoni, V. A., “Sedimentation Engineering”, ASCEManuals and Reports on Engineering
Practice no. 54, 1975, prepared by the ASCE task committee.
White, C.M., 1940, “The Equilibrium of grains on the bed of a stream”, Proc. Royal Soc. of
London, 174(958), 322338.
Wu, T.H. (1995). Slope stabilization. In: R.P.C. Morgan, R.J. Rickson (Eds.). Slope
Stabilization and Erosion Control: A Bioengineering Approach. E & FN Spon,
London, pp. 221264.
9 23
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Scour Rate vs Shear Stress
100.00
100.00
10.00
10.00
Scour Rate (mm/hr)
Scour Rate (mm/hr)
Scour Rate vs Velocity
1.00
Porcelain Clay
PI=16%
Su=23.3 Kpa
0.10
Porcelain Clay
PI=16%
Su=23.3 Kpa
0.10
0.01
0.01
0.1
10000.00
10.0
Scour1.0
Rate vs Velocity
Velocity (m/s)
0.1
100.0
1.0
Scour Rate
vs Shear10.0
Stress
100.0
2
Shear Stress (N/m )
10000.00
1000.00
1000.00
Scour Rate (mm/hr)
Scour Rate (mm/hr)
1.00
100.00
10.00
1.00
Sand
D 50 =0.3 mm
0.10
100.00
10.00
1.00
Sand
D50 =0.3 mm
0.10
0.01
0.01
0.1
1.0
10.0
0.1
100.0
1.0
10.0
Shear Stress (N/m2 )
Velocity (m/s)
Figure 9.1: Erodibility function for a clay and for a sand.
Figure 9.2: Critical shear stress versus mean soil grain size.
9 24
100.0
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 9.3: Magnitude of shear stresses involved in various fields of engineering.
Figure 9.4: Velocity and shear stress within the flow depth.
9 25
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
100000
Very High
Erodibility
I
10000
Erosion
Rate
(mm/hr)
High
Erodibility
II
1000
Medium
Erodibility
III
Low
Erodibility
IV
100
Very Low
Erodibility
V
10
1
0.1
1.0
10.0
Velocity (m/s)
Figure 9.5: Erosion Categories.
Figure 9.6: Particle diagram for a simple sliding mechanism.
9 26
100.0
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 9.7: Particle diagram for a simple rolling mechanism.
Figure 9.8: Contact angle distributions in coarse grained soils.
9 27
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 9.9: Particle diagram for a simple plucking mechanism.
Figure 9.10: Forces and pressure on particle: no flow condition
9 28
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 9.11: Forces and pressure on particle: flow condition.
9 29
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Table 9.1: Gravity and Van der Waals Forces for Sand and Clay Particles
Diameter d (m)
Weight W (N)
Van der Waals attraction FVDW (N)
FVDW/W
Sand particle
Clay particle
2 x 103
1.1 x 103
7.85 x 1023
7.1 x 1020
1 x 106
1.36 x 1013
3.14 x 1016
2.3 x 103
Table 9.2: Measured Critical Shear Stress in Clays
Range of τc (N/m2)
Authors
Dunn (1959)
Enger et al. (1968)
Hydrotechnical Construction, Moscow (1936)
Lyle and Smerdon (1965)
Smerdon and Beasley (1959)
Arulanandan et al. (1975)
Arulanandan (1975)
Kelly and Gularte (1981)
2–25
15–100
1–20
0.35–2.25
0.75–5
0.1–4
0.2–2.7
0.02–0.4
Table 9.3: Measured Erosion Rates in Clay
Authors
Results
Inferred scour rate
(mm/hr)*
Richardson, Davis (1995)
Maximum scour depth
reached in days
14 g/cm2/min
0.30.8 N/m2/min
0.0050.09 g/cm2/min
10100
0.00570.01 g/cm2/s
100180
Arulanandan et al. (1975)
Shaikh et al. (1988)
Ariathurai, Arulanandan
(1978)
Kelly, Gularte (1981)
3001200
924
1.527
* Erosion rate dz/dt = (weight loss rate per unit area dw/a dt)/(unit weight j)
9 30
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Table 9.4: Factors Influencing the Erodibility of Fine Grained Soils
When this parameter increases
Erodibility
Soil water content
Soil unit weight
Soil plasticity Index
Soil undrained shear strength
Soil void ratio
Soil swell
Soil mean grain size
Soil percent passing sieve #200
Soil clay minerals
Soil dispersion ratio
Soil cation exchange capacity
Soil sodium absorption ratio
Soil pH
Soil temperature
Water temperature
Water chemical composition
decreases
decreases
increases
increases
increases
decreases
increases
increases
increases
increases
Table 9.5: Database of EFA tests
Woodrow Wilson Bridge (Washington)
South Carolina Bridge
National Geotechnical Experimentation Site (Texas)
Arizona Bridge (NTSB)
Indonesia samples
Porcelain clay (manmade)
Bedias Creek Bridge (Texas)
Sims Bayou (Texas)
Brazos River Bridge (Texas)
Navasota River Bridge (Texas)
San Marcos River Bridge (Texas)
San Jacinto River Bridge (Texas)
Trinity River Bridge (Texas)
9 31
Tests 1 to 12
Tests 13 to 16
Tests 17 to 26
Test 27
Tests 28 to 33
Tests 34 to 72
Tests 73 to 77
Tests 78 to 80
Test 81
Tests 82 and 83
Tests 84 to 86
Tests 87 to 89
Tests 90 and 91
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Si vs. w
CSS vs. w
10000.00
35.00
30.00
1000.00
R2 = 0.0928
25.00
100.00
20.00
15.00
10.00
R2 = 0.0245
10.00
1.00
5.00
0.00
0.00
0.10
20.00
40.00
60.00
80.00
100.00
0.00
20.00
40.00
W ( %)
60.00
80.00
100.00
w( %)
Figure 9.12: Erosion properties as a function of water content.
Si vs.Su
CSS vs. Su
10000.00
45.00
40.00
1000.00
35.00
30.00
CSS(Pa)
100.00
25.00
2
R = 0.1093
20.00
10.00
15.00
10.00
1.00
5.00
0.10
0.00
0.00
20.00
40.00
60.00
80.00
100.00
120.00
0.00
140.00
Su(kPa)
20.00
40.00
60.00
80.00
100.00
S u( k P a )
Figure 9.13: Erosion properties as a function of undrained shear strength.
9 32
120.00
140.00
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Si vs.PI
CSS vs .PI
45.00
200.00
40.00
180.00
160.00
35.00
140.00
30.00
120.00
25.00
100.00
20.00
80.00
15.00
60.00
10.00
40.00
R2 = 0.056
R2 = 0.0011
5.00
20.00
0.00
0.00
0.00
20.00
40.00
60.00
80.00
0.00
100.00
20.00
40.00
60.00
80.00
100.00
P I ( kP a)
P I ( %)
Figure 9.14: Erosion properties as a function of plasticity index.
CSS vs.#200
Si vs.#200
25.00
10000.00
20.00
1000.00
15.00
100.00
10.00
10.00
R^2=0.2077
R2 = 0.1306
5.00
1.00
0.10
0.00
0.00
20.00
40.00
60.00
80.00
100.00
0.00
120.00
20.00
40.00
60.00
80.00
# 2 0 0 ( %)
# 2 0 0 ( %)
Figure 9.15: Erosion properties as a function of percent passing sieve #200.
9 33
100.00
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Source: Briaud et al (2001)
Figure 9.16: EFA (Erosion Function Apparatus).
Source: Munson et al (1990)
Figure 9.17: Moody Chart.
9 34
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Table 9.6: Procedures for Levee Design and Construction (USACE, 2000)
Major and Minimum Requirements
Step Procedure
1
Conduct geological study based on a thorough review of available data including analysis of aerial
photographs. Initiate preliminary subsurface explorations.
2
Analyze preliminary exploration data and from this analysis establish preliminary soil profiles, borrow
locations, and embankment sections.
3
Initiate final exploration to provide:
a. Additional information on soil profiles.
b. Undisturbed strengths of foundation materials.
c. More detailed information on borrow areas and other required excavations.
4
Using the information obtained in Step 3:
a. Determine both embankment and foundation soil parameters and refine preliminary sections where
needed, noting all possible problem areas.
b. Compute rough quantities of suitable material and refine borrow area locations.
5
Divide the entire levee into reaches of similar foundation conditions, embankment height, and fill material
and assign a typical trial section to each reach.
6
Analyze each trial section as needed for:
a. Underseepage and through seepage.
b. Slope stability.
c. Settlement.
d. Trafficability of the levee surface.
7
Design special treatment to preclude any problems as determined from Step 6. Determine surfacing
requirements for the levee based on its expected future use.
8
Based on the results of Step 7, establish final sections for each reach.
9
Compute final quantities needed; determine final borrow area locations.
10
Design embankment slope protection.
Source: Hanson et al. (2003)
Figure 9.18: Headcut Location as a Function of Time.
9 35
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Source: Ralston (1987)
Figure 9.19: Stages of Progressive Erosion.
Source: AlQaser, Ruff (1993)
Figure 9.20: Progressive Failure of an Overtopped Embankment.
9 36
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Table 9.7: Summary of Dam Break Computer Models (AlQaser, 1991)
Model
(Yr of Publ)
Cristofano
(1965)
Hydrodynamic
Approach
Sediment
Transport
Empirical
relation
Solution
Algorithm
Manual
iterative
Breach
Morphology
Constant
width
Harris &
Wagner
(1967)
BRDAM
Brown &
Rogers (1977,
1981)
DAMBRK –
Fread (1977)
Broadcrested
weir hydraulic
relation
Lou (1981)
Ponce &
Tsivoglou
(1981)
St. Venant
system of
equations
Numerical
Broadcrested
weir hydraulic
relation
Parabolic
shape
Linear pre
determined rate
of erosion
Rectangular,
triangular,
trapezoidal
Empirical
relation
Regime type
relation
between top
width and
flow
rate
Rectangular
changing to
trapezoidal
Preissmann’s 4
point finite
difference
MeyerPeter
and Mueller
bedload
formula
BREACH –
Fread (1984)
BEED –
Singh
1989)
Schoklitsch
bed
load formula
EinsteinBrown
bedload
formula
Numerical
iterative
linear
predetermined
rate of erosion
9 37
Other
Features
No tailwater
effects, no
sloughing
Grain size,
critical
discharge
value,
breach
dimensions and
slope
No tailwater
effects, no
sloughing
Failure duration
tim, terminal
size and shape
of breach
Coefficients of
the regime
relation, critical
shear stress
Critical shear
stress grain size
cohesion
friction angle
Friction angle,
dimensionless
shear stress 1/ψ
Rectangular or
trapezoidal
Froelich
(1990)
Characteristic
Parameters
Proportionality
constant, angle
of repose
Dam height
above the
breach volume
of water in the
reservoir
No
sloughing
Tailwater
effects and
sloughing
are
included
Neglects the
triggering
mechanism
of failure,
sloughing is
incorporated
Regression
relations to
predict
breach
parameters
and time of
failure
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Table 9.8: Results from Flume Tests (Nairn, 1986)
Run
Description
Duration
A
no bluff, composite
slope, sand veneer
6 hrs.
Erosion
Rate
(m3/m hr)
0.0066
B
bluff, composite
slope, toe sub
merged 3 cm, sand
veneer
6 hrs.
0.0046
C
bluff, composite
slope, toe submerged
5 cm, no sand
6 hrs.
0.0127
F
bluff, constant
1:20 slope, toe
at NWL, no veneer
3.5 hrs
6.5 hrs.
0.0109
0.0098
G
bluff, constant
1:20 slope, toe
at NWL, sand veneer
1 hr.
negligible,
profile
armoured with
sand
Source: Dodge (1988)
Figure 9.21: Laboratory Test Facility.
9 38
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Source: AlQaser (1991)
Figure 9.22: Testing Facility.
Source: Hanson et al. (2001)
Figure 9.23: Migration Rate vs. Unconfined Compression Tests.
9 39
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Source: Coleman et al. (2002)
Figure 9.24 : Experimental Setup.
Source: Coleman et al. (2002)
Figure 9.25: Longitudinal Profiles Along Breach Channel Centerline for
MediumSand Embankment.
9 40
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Source: Coleman et al. (2002)
Figure 9.26: Geometry Parameters for Breached Embankment.
Source: Hanson et al. (2003)
Figure 9.27: Photographic Measurements of Erosion Width.
9 41
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Source: Hanson et al. (2003)
Figure 9.28: Headcut Migration Rate vs. Compaction Water Content.
Source: Hanson et al. (2003)
Figure 9.29: Rate of Erosion Widening vs. Compaction Water Content.
9 42
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Source: Hunt et al. (2004)
Figure 9.30: Breach Width vs. Time.
Source: Hanson et al., (2003)
Figure 9.31: Headcut Migration Rate vs. Rate of Widening.
9 43
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Source: Dodge (1988)
Figure 9.32: Erosion Characteristics with respect to Plasticity.
Source: Chen & Cotton (1998)
Figure 9.33: Permissible Shear Stress for Cohesive Soils.
9 44
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 9.34: Difference in erosion resistance between grass cover and no grass cover.
9 45
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 9.35: Range of shear stresses allowable on slopes for different covers.
Source: Johnston (2003)
9 46
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 9.36: Location of samples.
Table 9.9: Salinity and pH of water associated with the samples
Sample S11
Sample S15
Typical sea water
Typical tap water
pH
Salinity (ppm)
8.61
8.09
7.9
7.0
3287
4199
30000 to 35000
500
9 47
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Soil
1 inch
3inch diameter
Shelby Tube
~ 17 inches
Piston
~ 1 inch
Sample Preparation
(Note: Dimensions Indicative Only)
Figure 9.37: Soil preparation by recompaction for bag samples.
9 48
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
EFA Test Results for Sample No. S4-(0-0.5ft)-LC-SW
Sample Type: Bulk Sample
Water Salinity: 36.1 PPT (Salt Water)
Compaction Effort: Low = 1.6% Modified Proctor Compaction
E ro sio n Rate(m m /h r)
Erosion Rate vs.Shear Stress
100000.0 S4-(0-0.5ft)-LC-SW
10000.0 Depth(ft):0-0.5
D50 = -mm
1000.0
100.0
10.0
1.0
0.1
0.1
1.0
10.0
Shear Stress (Pa)
E ro sio n Rate(m m /h r)
Erosion Rate vs.Velocity
100000.0 S4-(0-0.5ft)-LC-SW
10000.0 Depth(ft):0-0.5
D50 = -mm
1000.0
100.0
10.0
1.0
0.1
0.1
1.0
Velocity (m/s)
Figure 9.38: EFA test results for sample S4 (00.5 ft), low compaction, salt water.
9 49
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
EFA Test Results for Sample No. S3-B3-(0-1ft)-SW
Sample Type: Shelby Tube
Water Salinity: 36.4 PPT (Salt Water)
Compaction Effort: N/A
E ro sio n Rate(m m /h r)
Erosion Rate vs.Shear Stress
10.0
S3-B3-(0-1ft)
Depth(ft):0-1
Su = 94kPa
1.0
0.1
0.1
1.0
10.0
100.0
1000.0
Shear Stress (Pa)
E ro sio n Rate(m m /h r)
Erosion Rate vs.Velocity
10.0
S3-B3-(0-1ft)
Depth(ft):0-1
Su = 94kPa
1.0
0.1
0.1
1.0
Velocity (m/s)
Figure 9.39: EFA test results for sample S3B3 (01 ft), salt water.
9 50
10.0
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
100000
Very High
Erodibility
I
10000
Erosion
Rate
(mm/hr)
High
Erodibility
II
1000
Medium
Erodibility
III
Low
Erodibility
IV
100
Very Low
Erodibility
V
10
1
0.1
1.0
10.0
100.0
Velocity (m/s)
Figure 9.40: Erosion Chart.
100000
10000
Very High
Erodibility
I
1000
Erosion
Rate
(mm/hr)
High
Erodibility
II
Medium
Erodibility
III
Low
Erodibility
IV
100
10
Very Low
Erodibility
V
1
0.1
0.1
S1-B1-(0-2ft)-TW
S2-B1-(2-4ft)-SW
S3-B3-(0-1ft)-SW
S5-(0-0.5ft)-LT-SW
S7-B1-(2-4ft)-SW
S8-B1-(2-4ft)-L2-SW
S12-B1-(0-2ft)-TW
S15-CanalSide-(0-0.5ft)-HC-SW
1.0
Velocity (m/s)
10.0
S1-B1-(2-4ft)-SW
S3-B1-(2-4ft)-SW
S4-(0-0.5ft)-LC-SW
S6-(0-0.5ft)-LC-SW
S8-B1-(0-2ft)-TW
S11-(0-0.5ft)-LC-TW
S12-B1-(2-4ft)-SW
S15-Levee Crown-(0-0.5ft)-LT-SW
100.0
S2-B1-(0-2ft)-TW
S3-B2-(0-2ft)-SW
S4-(0-0.5ft)-HC-SW
S7-B1-(0-2ft)-TW
S8-B1-(2-4ft)-L1-SW
S11-(0-0.5ft)-HC-TW
S15-Canal Side-(0-0.5ft)-LC-SW
S15-Levee Crown-(0.5-1.0ft)-LT-SW
Figure 9.41: EFA test results for 24 levee samples.
9 51
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
100000
10000
Very High
Erodibility
I
1000
Erosion
Rate
(mm/hr)
High
Erodibility Medium
II
Erodibility
III
Low
Erodibility
IV
100
Very Low
Erodibility
V
10
1
0.1
0.1
1.0
10.0
Velocity (m/s)
100.0
S-4-(0-0.5)-LC-SW
S-4-(0-0.5ft)-HC-SW
S11-(0-0.5ft)-LC-TW
S11-(0-0.5ft)-HC-TW
S15-Canal Side-(0-0.5ft)-LC-SW
S15-CanalSide-(0-0.5ft)-HC-SW
Figure 9.42: Influence of compaction on erodibility.
100000
10000
Very High
Erodibility
I
1000
Erosion
Rate
(mm/hr)
High
Erodibility Medium
II
Erodibility
III
Low
Erodibility
IV
100
Very Low
Erodibility
V
10
1
0.1
0.1
1.0
10.0
Velocity (m/s)
S-1-B-1-(0-2ft)-TW
S-1-B-1-(2-4ft)-SW
S-2-B-1-(0-2ft)-TW
S-2-B-1-(2-4ft)-SW
S-7-B-1-(0-2ft)-TW
S-8-B-1-(0-2ft)-TW
S-7-B-1-(2-4ft)-SW
S-8-B-1-(2-4ft)-SW
S-12-B-1-(0-2ft)-TW
S-12-B-1-(2-4ft)-SW
100.0
Figure 9.43: Influence of water salinity on erodibility.
9 52
S1-B1-(0-2ft)-TW
Clay with hard clay grain mixture
CH
20.23
15.37 31.66
S1-B1-(2-4ft)-SW
S2-B1-(0-2ft)-TW
S2-B1-(2-4ft)-SW
S3-B1-(2-4ft)-SW
S3-B2-(0-2ft)-SW
S3-B3-(0-1ft)-SW
S4-(0-0.5ft)-LC-SW
S4-(0-0.5ft)-HC-SW
S5-(0-0.5ft)-LT-SW
S6-(0-0.5ft)-LC-SW
S7-B1-(0-2ft)-TW
S7-B1-(2-4ft)-SW
S8-B1-(0-2ft)-TW
S8-B1-(2-4ft)-L1-SW
S8-B1-(2-4ft)-L2-SW
S11-(0-0.5ft)-LC-TW
S11-(0-0.5ft)-HC-TW
Clay with rootlets
Clay with rootlets
Clay
Clay
Clay with some sand
Clay
Clat with some sand
Clay with some sand
Silt-Clay
Sand w/Some Clay
Clay
Clay with hard clay grain mixture
Clay with 1.5" thick grass on top of sample
Clay with 2 layers
Clay with 2 layers
Sand
Sand
CH
CL
CL
CL-CH
CH
CL-CH
19.10
19.74
20.26
17.60
20.20
17.16
13.87
17.69
SP
CH
CH
CH
CH
CH
SP
SP
S12-B1-(0-2ft)-TW
Clay with decomposed wood
S12-B1-(2-4ft)-SW
S15-CanalSide-(0-0.5ft)-LC-SW
S15-CanalSide-(0-0.5ft)-HC-SW
Clay
Sand w/Some Clay
Sand w/some clay
S15-LeveeCrown-(0-0.5ft)-LT-SW
Sand w/Some Clay
S15-LeveeCrown-(0.5-1ft)-LT-SW
Sand w/Some Clay
Tested at Texas A&M University Tests Performed by STE
%
Organic %
LL PL PI
fines
Content% fines
71 25 46
3.09
15.69
17.00
16.71
13.86
17.26
13.95
10.42
13.23
21.77 - 56 19 37
16.11 - 46 17 29
21.23 69.1 41 16 25
27.00 - 48 17 31
31.66 - 69 23 46
23.00 - 32 12 20
33.14
90.5 60 30 30
33.14
1.91
16.94
1.62
2.50
2.60
21.85
13.45
17.39
16.52
17.71
18.74
18.74
12.30
13.26
18.15
12.79
13.73
13.42
13.38
14.00
14.00
12.23
13.12
20.40 54.4 - - 5.21 8.9 - - 26.65 - 68 24 44
23.04 - - 32.34 - - 33.87 - - 33.87 - 54 21 33
1.02
- - 0.1
1.02
- - -
0.69
0.71
3.78
7.14
2.28
15.37
0.32
0.35
CH
14.77
10.19 44.94
-
67 27 40
16.91
MH-CH
17.56
13.85
12.64 38.94
12.21 13.43
-
58 32 26
- - -
5.28
19.63
17.31 13.43
13.29
11.94 11.29
13.57
12.46
CL
SM
8.93
-
29.9
8.16
LL
PL
PI
89.9
65
22
43
67.2
49
17
32
90.3
54
19
35
-
-
-
-
-
-
-
-
90.1
78
32
46
97.3
85
36
49
-
-
-
-
92
67
21
46
-
-
-
-
1.28
-
-
-
-
-
-
2.16
-
-
-
1.01
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Classification
Table 9.10: Results of the index property tests.
9 53
Soil Description
Independent Levee
Investigation Team
jt
jdry
w (%)
(kN/m3) (kN/m3)
Sample
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
100000
Very High
Erodibility
I
10000
High
Erodibility
II
Medium
Erodibility
III
1000
Erosion
Rate
(mm/hr)
Low
Erodibility
IV
100
Note:
- Solid circles =
levee breaches
- Empty circles =
no levee damage
10
Very Low
Erodibility
V
1
0.1
0.1
1.0
Velocity (m/s)
10.0
100.0
S2-B1-(0-2ft)-TW
S2-B1-(2-4ft)-SW
S3-B1-(2-4ft)-SW
S3-B2-(0-2ft)-SW
S3-B3-(0-1ft)-SW
S4-(0-0.5ft)-LC-SW
S5-(0-0.5ft)-LT-SW
S6-(0-0.5ft)-LC-SW
S15-Canal Side-(0-0.5ft)-LC-SW
S15-CanalSide-(0-0.5ft)-HC-SW
S15-Levee Crown-(0-0.5ft)-LT-SW
S15-Levee Crown-(0.5-1.0ft)-LT-SW
Figure 9.44: EFA test results and overtopping levee failure/no failure chart.
100000
Very High
Erodibility
I
10000
1000
Erosion
Rate
(mm/hr)
High
Erodibility
II
PRONE TO
FAILURE BY
OVERTOPPING
100
Medium
Erodibility
III
Low
Erodibility
IV
TRANSITION
ZONE
PRONE TO
Very Low
RESIST
Erodibility
OVERTOPPING
V
10
1
0.1
0.1
1.0
10.0
Velocity (m/s)
Figure 9.45: Proposed guidelines for levee overtopping.
9 54
100.0
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
CHAPTER TEN: ENGINEERING OVERVIEW;
EARTHEN LEVEES AND FLOODWALLS
10.1.
Overview
The vast majority of flood protection for the greater New Orleans area is strongly
dependent upon the presence and ability of earthen levees to separate large water bodies, such as
Lake Pontchartrain, Lake Borgne, the Mississippi River, and the Gulf of Mexico, and
appurtenant channels and canals, from inundating developed land areas and causing flooding of
homes and businesses. Earthen levee flood protection systems not having redundancy can be
viewed as series systems, where failure at one location, or failure of one component, can result in
catastrophic failure of the entire flood protection system and result in tragic loss of life, damage
to fundamental infrastructure (basic services such as water, sewage, and electricity), and
substantial devastation and economic impact to the immediate and surrounding regions. These
systems can be in place for a short duration (a few years) or for a very long duration (hundreds of
years). In order to ensure the desired level of flood protection system performance, identification
and mitigation of “weak links” in the system is crucial in order to maintain longterm system
integrity.
The earthen levees are supplemented and extended at many locations by means of more
“structural” components comprised of concrete and steel. Steel sheetpile curtains are routinely
used either to extend a “cut off” barrier to retard underseepage flow beneath levees, or to provide
support for reinforced concrete floodwalls at the crests of earthen levees. In some cases, the
sheetpile curtains are extended vertically above the earthen crests without concrete to simply
extend the crest elevation as an interim measure until a more permanent crest raising can be
implemented. The concrete floodwalls are used to achieve increased crest height without the
extra weight of additional earthen levee fill, and/or without the need to widen the earthen levee
embankment section to accommodate additional earthen levee fill in situations where the
available “footprint” is limited. Concrete walls are also employed to provide frames for gates
(usually steel gates) that can be opened to allow traffic to pass through (e.g.: automobiles, trains,
ships, etc.) and then closed when storms arrive.
Few studies have systematically analyzed actual longterm performance of earthen levees
and/or composite leveefloodwall systems to confirm effective design parameters, assumed
loading conditions, and actual performance after major flooding events. Additionally,
evaluations of component transitions (i.e. earthen levee to concrete structure transitions),
erodeability overtopping, wavescour, and effective inspection programs have not been well
documented and are critical components for high reliability flood protection systems.
Chapter 10 builds upon the technical lessons from the previous chapters, and establishes
additional findings as well as background to facilitate the presentation of “lessons learned” and
“recommendations” with regard to design and construction of these types of regional flood
protection systems that will then be presented and discussed in Chapter 11.
The main goals of this Chapter are to: (1) provide a brief overview of some of the
principal design procedures and standards employed in the development of preKatrina regional
10 1
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
flood protection system, (2) identify several critical “weak link” features not adequately
addressed in current design methods widely used in that region so that appropriate design
modifications can be implemented to improve levee performance, (3) establish an erodeability
testing methodology which can be used to assess existing earthen levees, (4) present some
limited comments regarding “unwatering” (pumping), and (5) present comments and
observations regarding emergency and interim levee and floodwall reconstruction efforts in the
wake of hurricane Katrina.
10.2.
Potential Levee Failure Mechanisms
There are numerous failure mechanisms that can result in the failure and breaching of
earthen levees and/or floodwalls, and the resultant catastrophic flooding of protected areas.
These failure mechanisms can occur as a single mode, or as a combination several different types
of failure modes acting in unison. Levees can fail as a result of damage to the levee itself, if the
foundation on which the levee is constructed fails, or as a result of failure of a floodwall for a
composite levee/floodwall section. An abbreviated overview of many of the potential failure
mechanisms of interest in the greater New Orleans area is presented here:
10.2.1. Structural Causes
This category includes potential failure mechanisms where the dominant issue is either
the strength and stability of the levee embankment and/or foundation soils, or the structural
capability of “structural” elements (e.g. sheetpile curtains, floodwalls, or gates) and/or their
interaction with the levees and foundations soils that support them. Such mechanisms include:
Slope Instability – If the levee embankment soils and/or the underlying foundation
materials that support the levee are weak, or become destabilized, a slope failure can develop and
result in catastrophic failure of the levee. Slope failures can be minor or they can be significant
enough to result in the catastrophic failure of the levee system. A number of catastrophic
failures occurred during hurricane Katrina due to this mechanism. Slope failures can be sub
divided into separable classes as
(a) Bearing capacity failure; Failure of the weak foundation soils to vertically support the
weight of the levee embankment. This is most common during construction (before
the foundation soils have time to consolidate and gain strength under the embankment
load.) A failure of this type just recently occurred on a section of levee under
reconstruction in Plaquemines Parish on May 29, 2006.
(b) Lateral translational stability failure; Failure by sliding laterally, usually as a result of
being “pushed” by elevated water pressures on the water (canal) side.
(c) Deeper, rotationaltype stability failure; This can also be caused by the “push” of
elevated water on the outboard (canal) side, but these types of failures can also occur
due to undercutting of the canal side of the levee by dredging operations, or by storm
surge scour or river flow.
Structural Failures of Walls, Sheetpiles or Gates – Simple structural failure as a result of
structural elements inability to safely bear the forces and loads exerted against them. There
appear to have been no structural failures of this type, except in cases where other types of
10 2
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
embankment or foundation failure, or overtopping erosion and resultant lateral unbracing of
floodwalls, occurred.
Structural Impacts – Structural impacts occur when physical objects collide with the
levee. This can occur during storm events when boats or barges become loose from their
moorings and are driven into the levee by wind or water forces, or simply from accidental boat
impacts due to operator error.
10.2.2. Causes due to Hydraulic Forces
This category includes failure mechanisms where the dominant parameters involve
groundwater flow and pore pressure. Among these are:
Underseepage/Instability – As shown in Figure 10.1 (red lines), if the underlying
foundation materials that support the levee are adequately permeable, water can quickly travel
through these porous materials as the water head differential between the outboard and inboard
sides of the levee increases. This underseepage raises the pore pressures within the soils at the
inboard side, and this in turn reduces the shear strengths of these soils. This can result in
catastrophic failure by means of resultant slope instability (as described in the previous section).
Bottom Heave or “Blowout” This is a variant of underseepage, but involves the
hydraulic pressure, rather than simple erosion. An increase in water pressure caused by a storm
surge can travel through a permeable zone in a levee’s foundation. If the water pressure exceeds
the total overburden pressure at the landside toe of the levee, then the (impervious) soil
overburden at that location can be displaced (heaved) by the water pressure, producing a large
void into which subsequent flow will rush (rapidly exacerbating localized erosion and failure).
This is often referred to as a “blowout” failure.
Erosion and Piping – As shown in Figure 10.1 (blue lines), erosion and piping occurs
when the localized hydraulic gradient becomes large enough to “pull” soil grains from their
location, and when there is no soil on that soil’s downflow side that can “filter” these soil grains
and thus hold them in place. Erosion and piping can be subdivided into two subclasses as:
(a) Exit seepage erosion and piping; This is one of the most common causes of levee
failures worldwide. If underseepage flow (and/or flow laterally through the levee embankment)
becomes sufficient as to raise the exit gradient, then there is little or no resistance to erosion of
soil particles at the point of water exit (either low on the inboard side levee slope face, or on the
ground surface at and just inboard of the levee toe). Once the seepage gradient is able to exert
enough “drag” on the soil grains to overcome the stability due to their selfweight, erosion
begins. As soon as erosion begins, the local flownet rapidly converges on the “hole” that begins
to develop, as the water moves towards a preferential shortcut in its effort to escape. That, in
turn, increases the local gradient and thus accelerates the erosion. The result is that erosion can
rapidly ‘eat back’ a tunnel (or “pipe”) beneath the levee [hence the name “piping”.] Soils from
above routinely slough and fall into this rapidly developing “pipe”, but they are usually
immediately washed out by the ever increasing flow until, finally, the embankment ruptures and
erodes through catastrophically. An illustration of this was provided previously in Figure 8.105.
10 3
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
(b) Internal seepage erosion; This can occur internally within either the levee, or within
the foundation soils. As water flows through these soils, smaller/finer soil particles can be
“washed” out resulting in the internal erosion of the levee or foundation soil. Enough internal
erosion can lead to the collapse and subsequent fullblown “washout” failure of the levee. For
levees constructed of layers with significantly different permeabilities, the layer with the highest
permeability becomes the main “conduit” by which the water flows through the levee. This
concentrated flow can lead to higher water velocities through the levee and more rapid
degradation. Appropriate control of soil gradation, and use of appropriate soil gradations within
adjacent embankment sections (which is called “filtering”), are the keys to prevention of internal
erosion.
The major design standards specifically address both internal erosion and piping as well
as exit seepage erosion and piping, and levees are required to be engineered against these failure
mechanisms. Mitigation strategies can include utilization of low permeability materials (such as
clay), provision of underseepage “cut offs” (e.g. sheetpile curtains extending below the levees),
internal drains or filters to safely “vent” pore pressures while filtering soil grains to hold them in
place, widenend embankments and use of inboardside stability berms to lengthen the flow path
(and thus reduce exit gradients at the inboard side toe), control of soil gradation to prevent
internal erosion, etc.
Effectively mitigating erosion and piping can be hampered by the presence of burrowing
animals that can carve intricate tunnel networks within the earthen levees. Effective detection
and corresponding correction of these animalinduced internal erosion channels is very
challenging, and many levee failures throughout the world are a result of this failure mechanism.
Exit erosion and piping can also be exacerbated by the presence of trees low on the inboard side
levee faces and/or at the inboard toe. Trees that die can leave root hole paths that can exacerbate
erosion and piping in this critical area. Trees that blow over during storm winds (and/or due to
weakening of their roots’ foundation soils due to wetting) can suddenly leave large voids that can
serve as initiation points for rapid advancement of localized exit erosion and piping. As a result,
it is common practice to prevent growth of trees (as a “maintenance” issue) in this critical
inboard toe area.
10.2.3. Causes Involving Surficial Erosion
These include the various forms of surface erosion which can occur due to surface water
flowing over (across) or against the exposed surface of the levee.
(a) Overtopping – As shown in Figure 10.2, overtopping occurs when the water level on the
outboard side of the levee exceeds the crest elevation of the levee. The inboard side of the levee
acts as a spillway for the overtopping water and damage is inflicted on the levee as a result of
water scour. Levees are not generally designed for overtopping and as a result, if overtopping
does occur, they are can be highly susceptible to catastrophic failure unless overtopping duration
and intensity are limited and erosionresistant materials are used to construct the levee (or unless
erosion protection, in the form of “armoring” is applied to the exposed crest and faces of the
levee.) As water passes over the top of the levee, its velocity increases as it runs down the back
side, and the erosive shear force of the water increases with this increase in velocity.
10 4
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Accordingly, overtopping erosion is usually most severe initially low on the back slope face.
Eventually, as the crest is “notched”, the resulting flow through the crest can also scour rapidly
and further exacerbate the erosive process.
(b) Sharp Overtopping and Jetting – This is more likely to occur on levees with floodwalls.
As shown in Figure 10.3, jetting occurs when the water level on the outboard side of the levee
exceeds the top of wall elevation for structural walls that are founded within the earthen levee.
Unlike overtopping of a conventional earthen levee, the floodwall acts as a weir and water falling
over the wall impacts the levee in a concentrated stream that is much more energy intensive than
conventional overtopping. For typical New Orleans floodwalls, the water impact velocities are
on the order of 6 to 8 m/s. Levees are not generally designed for overtopping and jetting and as a
result, if overtopping and jetting does occur, a deep scour trench can rapidly develop against the
landside face of the floodwall. This reduces the earth pressure providing lateral support for the
wall, making the wall highly susceptible to potential catastrophic lateral failure (when pushed by
the water pressures on its outboard side). This erosive trenching can be prevented by installing
“splashpads”, coarse riprap, or other energy dissipating devices at the inboard side toe of the
floodwall, and the use of highly erosionresistant materials to construct the crown section of the
levee is also advisable here.
(c) Lateral Surface Erosion – As shown in Figure 10.4, lateral surface erosion generally
occurs on the outboard side of the levee and is the result of water flowing past the levee face, or
against the face of the stream channel banks below. If the imposed shear stress from the water
abrading against the soil face is high enough, soil scour occurs and the integrity of the overall
levee is significantly reduced. Levees that are exposed to chronic water flow, such as river
levees, are generally designed and constructed with armoring or erosion protection to minimize
scourinduced surface erosion. In general, wellcompacted levees constructed of highplasticity
clays are much more resistant to surface erosion than uncompacted cohesionless soils (e.g.
“clean” sands) and silty sands. Surface protection such as riprap, concrete pads, soilcement
reinforcement, and select vegetation coverings are typical methods used to protect levee faces
from surface erosion.
(d) Wave Impacts – As shown in Figure 10.5, wave impacts can cause significant erosion to
levee faces. Waveinduced erosion consists of runup (sloshing up and down of water as a result
of staggered wave arrival) and “minijetting” when the crest of the waves breaks on the levee
face. Levees that are anticipated to be impacted by waves are generally designed with armoring
to prevent damage from wave impacts.
The aforementioned failure mechanisms are not intended to be an exhaustive list, but
rather to highlight common potential failure modes of interest for the levees in the New Orleans
region.
10.3.
Design Standards
Design standards are not just the primary means by which earthen levees are designed;
they are also the main metric by which proposed levee design and construction projects are
assessed and critiqued by reviewers. Incomplete, inaccurate, or inappropriate design standards
can lead to actual field performance which is less than desired. As part of this study, current
10 5
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
earthen levee design standards from the United States Army Corps of Engineers and the United
States Federal Emergency Management Administration were reviewed. A summary of the
design guidelines for the USACE and FEMA are presented in Sections 10.4 and 10.5,
respectively. This was not a fully exhaustive review (as that would have been beyond the scope
of our current study), but this issue was studied as it provides important context for
understanding some of the designs and decisions executed in the development of the preKatrina
New Orleans regional flood protection system.
10.3.1. United States Army Corps of Engineers Design Standards
The primary manual and summary of design standards for earthen levees for the United
States Army Corps of Engineers (USACE) is EM 111021913, “Engineering and Design –
Design and Construction of Levees.” This design manual covers the topics of: field
investigations, laboratory testing, (fill) borrow areas, seepage control, slope design and
settlement, levee construction, and special considerations (such as pipelines and other utility
crossings, access roads and ramps, levee enlargements, junctions with concrete closure
structures, and other special features such as landside ditch construction and levee vegetation
management.)
10.3.1.1 Primary Design Procedure
The design procedure and requirements for levee design are established by EM 11102
1913. The outlined design procedure provides guidance from the initial preliminary evaluation
through final design. These requirements (the principal “steps” in the design process) are
summarized in Table 10.1.
Further engineering analysis guidance for the design of levees is provided in the
following manuals:
テ
テ
テ
Slope Stability Analyses
Settlement Analyses
Levee/Structure Transitions
EM 111021902
EM 111021904
EM 111022502
The punch list of design steps identified in Table 10.1 provides an overview of design
parameters and principal steps for levee design. EM 111021913 prescribes required Factors of
Safety for slope stability of newly designed levees, existing levees, and other embankments and
dikes. These Factors of Safety vary from 1.0 for shortterm loading conditions to 1.4 for long
term (steady state conditions). Specific design criteria are not provided for settlement and
erosionsusceptibility.
In addition to these design parameters, material specifications and construction
procedures, critical elements in the actual lifespan performance of levees, have also been
defined by the USACE. These components are described in more detail in the following
sections.
10 6
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
10.3.1.1.
Material Selection
Acceptable soils for the construction of levees (borrow materials) are defined by EM
111021913 as “any soil is suitable for constructing levees, except very wet, finegrained soils
or highly organic soils.” Choosing a material type is generally a function of accessibility and
proximity to the project area. The design guidelines emphasize that studies should be performed
to ascertain the insitu moisture contents of the borrow materials. It is noted that “the cost of
drying borrow material to suitable water contents can be very high, in many cases exceeding the
cost of longer haul distances to obtain material that can be placed without drying.” Thus, any
materials may be used in the construction of levees so long as they are not overly wet fine
grained soils or highly organic soils. As will be discussed later in this chapter, our field
observations and laboratory testing clearly show that a highperformance levee must be
constructed from superior materials, and that utilization of more marginal materials (as allowed
by the design guidelines) can result in catastrophically poor performance.
10.3.1.2
Required Levee Soil Compaction
Three general types of engineered earthen levees are presented in the EM 111021913
design criteria. These are compacted, semicompacted, and uncompacted levees. The USACE
notes that, traditionally, compacted levees are usually used for areas of high property values
and/or high land use, high populations, and for steepsloped embankments with controlled
compaction during construction which are utilized on good foundation conditions. Areas of low
values, poor foundations, or high rainfall during the construction season generally warrant
specification of semicompacted or uncompacted levees.
According to the USACE design guidelines, compacted levees are required to be
constructed in areas where strong embankments of low compressibility are needed adjacent to
concrete structures or forming parts of highway systems. Compacted levees require specification
of appropriate ranges of fill material water content during compaction (with respect to standard
effort optimum water content), initial loose lift thickness of typically 6 to 9 inches, compaction
equipment type (e.g. sheepsfoot rollers, rubbertired rollers, etc.), and either the number of
compaction passes to attain a given percent compaction or standard maximum density or
specification of the minimum required density (relative compaction).
Semicompacted levees are recommended by the USACE to be constructed in areas where
there are no space limitations and thus steepsloped embankments are not required, where onsite
foundation soil conditions are relatively weak and unable to support steepsloped embankments,
where underseepage conditions require a wide base, and/or where the water content of borrow
materials or rainfall during construction does not allow for the proper compaction of levee fill
material.
Semicompacted levees require the specification of lift thickness (typically
approximately 12 inches) and are compacted by the movement of hauling and spreading
equipment, or by sheepsfoot or rubbertired roller compaction equipment.
The USACE recommends uncompacted levees only to be used for temporary/emergency
use. These levees are constructed by fill cast or dumped in place as thick layers with little or no
spreading or compaction. Hydraulic fill by dredge, often from channel excavations, is a common
fill borrow source for uncompacted levees. Hydraulic fills are known to be highly susceptible to
erosion upon overtopping and are not recommended to be used in the normal construction of
10 7
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
levees, except in locations where the levees are protecting agricultural areas whose failure would
not endanger human life or for zoned embankments that include impervious seepage barriers.
10.3.1.3 Embankment Geometry
Embankment geometry specified by the USACE design guidelines is controlled either by
material selection and compaction efforts during construction, or by the foundation soil
conditions. Maximum side slopes for levees are 1V on 2H. These steepsided levees are
required to be constructed from highgrade borrow materials that are compacted near optimum
moisture content and with appropriate compaction equipment. Levees with nonideal borrow
materials, such as sand levees, are required to have much shallower side slopes (on the order of
1V to 5H) to prevent damage from seepage and wave action. These design guidelines assume
that this geometry and associated levee material will then be adequately resistant to scour and
erosion, but as demonstrated by the numerous levee failures during Hurricane Katrina, this is not
reliably always the case.
Final top of levee elevations must also account for future settlements, as determined by
EM 111021904. In the past, the USACE specified a certain freeboard distance between the
final top of levee elevation and the design storm water level to account for hydraulic,
geotechnical, construction, operation, settlement and maintenance uncertainties. The updated
design procedures set forth in EM 111021913 are riskbased, and are assumed to directly
account for hydraulic uncertainties and establish a nominal level of protection.
10.3.1.4
Identified Potential Failure Modes for Design
The principal causes of potential levee failures, as identified by EM 111021913, consist
of the following mechanisms (but see also Section 10.2 of this chapter):
テ
テ
テ
テ
Overtopping;
Surface erosion;
Internal erosion (piping); and
Slides within the levee embankment or the foundation soils.
Considerable discussion is presented in the design manual to mitigate effects of internal
erosion/piping (see EM 111021913 Chapter 5 – Seepage Control). Guidance on overtopping,
surface erosion, and slides within the levee embankment or the foundation soils is not well
developed in this design manual. However, guidance is provided for the augmentation of soil
cement protection applied to exposed slopes, susceptible to erosion.
10.3.1.5
Erosion Susceptibility
Although not directly addressed or identified in EM 111021913, general guidelines for
erosion susceptibility of finegrained cohesive soils are presented in EM 111021100 [Coastal
Engineering Manual Part III], EM 111021100 [Coastal Engineering Manual Part VI], and
“Channel Rehabilitation: Processes, Design, and Implementation,” (1999). These manuals
provide insights on erosion and critical values of average overtopping discharges. They provide
valuable design information for levees situated in coastal areas, and these design concepts should
be applied to urban levees situated in coastal areas. Levee stretches such as the MRGO and the
10 8
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Orleans East southeastern levees abut Lake Borgne (an extension of the Gulf of Mexico) and are
susceptible to wave attack during strong storms.
For coastal grass covered seadikes and protected embankment seawalls, EM 11102
1100 indicates that no damage occurs for overtopping discharges of less than about 0.15 ft3/s per
foot. Significant damage is expected for overtopping flows greater than 0.35 ft3/s per foot. The
overtopping discharge flows were based on wave runup exceeding the crest of the embankment
or floodwall crest, and include estimated impact forces associated with the wave action
impacting the embankment. These values were based on field studies conducted both in the
United States and in the Netherlands. Discharge flow values are not based on sustained
overtopping discharge as a result of the mean storm water level rising above the crest of the
embankment or floodwall.
Maximum permissible velocities for flow within river and stream channels are
summarized in USACE (1999), and are based on field research from 1915 to about 1926.
Permissible velocities (for a canal type section with an average depth of 3 feet) are presented in
Table 10.2.
Equivalent shear stress in this Table was calculated using the following equation
(Munson et al, 1990):
ww = KuV2/2
[Equation 10.1]
In this correlation, the shear stress (ww) imposed on the surface exposed to the water flow
is a function of the surface roughness (K), the density of the fluid (u), and the velocity of the
fluid (V). Based on this table, permissible water velocities vary between 1.5 ft/s for highly
erosion susceptible materials to as much as 6 ft/s for highly erosion resistant materials.
Correspondingly, allowable shear stresses vary from a low of 2.4 lb/ft2 for highly erosion
susceptible materials to as much as 36 lb/ft2 for highly erosion resistant materials. Again,
erosion plus jetting can lead to impact velocities with erosive potential about 3 to 4 times the
maxima above. Although these design guidelines are available for use, they do not appear to
have been incorporated into the design of the MRGO frontage and New Orleans East coastal
levees fronting Lake Borgne.
10.3.2 United States Federal Emergency Management Agency (FEMA) Design Standards
Separate from the USACE levee design guidelines, design criteria for levee systems
required by the United States Federal Emergency Management Agency (FEMA) are presented in
the Title 44, Volume 1, Part 65 of the Code of Federal Regulations. These criteria establish the
minimum standards to which levees must adhere in order to satisfy the 100year level (referred
to as the base flood) of protection mandated by FEMA. The main design criteria for FEMA
approved levees are: freeboard, closures/transitions, embankment protection, embankment and
foundation stability, settlement, interior drainage, and other specialty design criteria deemed
appropriate by FEMA for unique situations.
10 9
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
10.3.2.1
Freeboard
Levees constructed adjacent to rivers are mandated to have a minimum freeboard of three
feet above the water surface level of the base flood. In areas where the levee is constructed
adjacent to structures, such as bridges, an additional one foot of freeboard is required extending
100 feet to either side of the structure. Levees constructed on the coast must have a minimum
freeboard of one foot above the height of the calculated one percent wave or the maximum wave
runup (whichever is greater) associated with the 100 year stillwater surge elevation. This
category best fits some portions of the New Orleans Hurricane Protection System. Exceptions
may be granted, based on sitespecific engineering studies, but a freeboard of less than two feet
is not deemed acceptable under any circumstance.
10.3.2.2
Closures
Closures refer to openings within the flood protection system. These closures can be for
through traffic (such as railroad traffic which is frequently grade controlled and can not easily be
diverted over levees), for pipeline crossings, or for maintenance purposes. FEMA requires all
closures to be structural parts of the overall flood protection system during operation, and that
they be designed in accordance with sound engineering practice.
10.3.2.3
Embankment Protection
Engineering analyses are required to be performed to demonstrate that no appreciable
erosion of the levee embankment will occur during the base flood due to currents or waves, and
that any anticipated erosion will not result in failure of the levee embankment or foundation
either directly or indirectly through seepage or subsequent instability. Specific factors to be
analyzed to determine the adequacy of embankment protection are: expected flow velocities
(especially in constricted areas), expected wind and wave action, ice loading, impact of debris,
slope protection techniques, duration of flooding at various stage and velocities, embankment
and foundation materials, levee alignment, bends, transitions, and levee side slopes. The FEMA
guidelines do not, however, provide guidance regarding acceptable performance
criteria/standards for the identified embankment protection factors to be evaluated.
10.3.2.4
Embankment and Foundation Stability
Stability analyses for levee embankments are required to be submitted that demonstrate
the adequacy of both shortterm and longterm slope stability of flood protection levees.
Stability analyses are required to include the expected seepage during the storm loading
conditions, and to demonstrate that seepage into, beneath or through the embankment will not
result in unacceptable stability performance. FEMA provides for the use of the USACE Case IV
(as defined by EM 111021913, “Design and Construction of Levees”) as an additionally
acceptable engineering analysis method. The required factors for evaluation include: depth of
flooding, duration of flooding, embankment geometry and length of seepage path at critical
locations, embankment and foundation materials, embankment compaction, penetrations, other
design factors affecting seepage (such as drainage layers), and other design factors affecting
embankment and foundation stability (such as interior berms). The FEMA guidelines do not,
however, provide quantitative guidance regarding acceptable performance criteria/standards for
the identified stability modes to be evaluated.
10 10
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
10.3.2.5
Settlement
Once levees have been constructed to the specified crest elevation, their ability to provide
the desired degree of flood protection against the base flood is often dependent to large extent on
settlements due to timedependent compression of the foundation materials beneath the levee. In
order to demonstrate the adequacy of the crest elevation over the intended service life, FEMA
requires that engineering analyses be submitted that assess the potential and magnitude of future
losses of freeboard as a result of levee settlement, and demonstration that freeboard will be
maintained within the minimum freeboard requirements for the duration of the intended levee
service period. Detailed analysis procedures, such as those specified in the USACE EM 11102
1904, “Soil Mechanics Design – Settlement Analyses,” are expected. The required factors for
evaluation include: embankment loads, compressibility of embankment soils, compressibility of
foundation soils, age of the levee system, and construction compaction methods. There are no
specific provisions for regional subsidence, tectonic subsidence, nor for potential water level rise
due to longterm climate change.
10.3.2.6
Interior Drainage
FEMA requires that the protected side of the flood protection system be capable of
draining onsite water. An analysis is required to be submitted that identifies the source(s) of
potential flooding, the extent of the flooded area, and, if the average depth of flooding is greater
than onefoot, the watersurface elevation(s) of the base flood. The analysis is required to be
based on the joint probability of interior and exterior flooding and the capacity of facilities (such
as drainage lines and pumps) for draining interior floodwater.
10.3.2.7
Other Design Criteria
In areas where levee systems have relatively high vulnerabilities, or other unique
situations, FEMA may require other design criteria and analyses be submitted for review and
approval. The rationale for the requirement of additional analyses will be provided by FEMA.
The review and subsequent evaluation standard of the analyses for the specified design criteria
are to be based on “sound engineering practice.”
10.3.2.8
Other FEMA Requirements
In order for the levee flood protection system to be recognized by FEMA as providing
protection for the base flood, additional requirements, beyond the established design procedures
and criteria, are required to be in place. Maintenance and operation plans are required to be
submitted that detail how the flood protection system will be maintained and operated during its
service period. In addition, FEMA has certification requirements which require that a registered
professional engineer certify the levee design, and that certified asbuilt plans of the completed
levee be submitted. Federal agencies with responsibility for levee design may also certify that
the levee has been adequately designed and constructed to provide the desired degree of
protection against the base flood.
10 11
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
10.4
Storm Surge and Wave Action During Hurricane Katrina
During Hurricane Katrina, the earthen levees were subjected to storm surges and wind
generated wave action. Accurately determining the magnitude of these forces is reliant on
numerical simulations and modeling with calibration from field data such as inplace
instrumentation that recorded data during Hurricane Katrina as well as posthurricane field
assessments, such as highwater marks. The most reliable storm surge and wave action
information collected and recorded during Hurricane Katrina was captured by instrumentation
installed at select locations within the greater New Orleans area. However, the number of
instrumentation locations was extremely limited, and as a result, little reliable storm surge and
wave action information is available from this source. Many of the instruments were damaged
during the storm and only partial records were collected.
Instruments that recorded useful data used to establish storm surge and wave action
information were located at the following locations (IPET 2006):
テ
テ
テ
テ
テ
テ
Lake Pontchartrain near 17th Street Canal (hydrograph & wave characteristics)
Pump station #6 on the 17th Street Canal (hydrograph)
Lake Pontchartrain at the Lakefront Airport (camera-based hydrograph)
Inner Harbor Navigation Channel at I10 (hydrograph)
Inner Harbor Navigation Channel at the Lock (hydrograph)
Gulf Intracoastal Waterway at I510 (hydrograph)
A detailed review and reconstruction of the storm surge and wave action during
Hurricane Katrina based on the data from the installed instruments, measured high water marks
and interviews was completed by IPET (2006). The storm surge and wave action information
presented by IPET was used in our performance evaluation of the levees. A discussion of the
maximum storm surges is presented in the following section along with an overview of our field
reconnaissance and levee condition survey and mapping.
10.5 Field Reconnaissance and Levee Condition Mapping with Regard to Levee Erosion
Field reconnaissance was a vital part to assessing and understanding the performance of
the earthen levee flood protection systems. Multiple field visits were performed by the team to
visually observe and evaluate the performance of the levee systems. The initial levee
assessments occurred between September 29 and October 15, 2005. The principal purposes of
these initial site investigations were to perform an initial survey of major damage areas, to
perform initial forensic studies at the major failure (breach) sites and at nearby, more successful
sites, and to note and record time sensitive data and observations before ongoing emergency
repair operations obscured vital stormrelated levee system performance information. The
results of our initial observations and findings are presented in Seed et al., (2005).
Subsequent to the initial field reconnaissance and forensic studies, a series of field survey
explorations have been performed to extend the initial condition surveys and to collect physical
samples for testing to ascertain susceptibility to erosion. The variable performance of the
earthen levee flood protection components during hurricane Katrina provide a unique learning
10 12
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
opportunity in that many of the levee system elements were overtopped, impacted by moving
objects and debris (such as steel barges and fishing boats), and/or attacked by windgenerated
waves. Some sections performed extremely well, while other sections performed poorly. As a
result, there is a valuable opportunity to draw empirical lessons regarding the interactions
between water and wave loadings, embankment and foundation soils and geometry, and
performance.
Figure 10.6 shows the extents of the formal visual reconnaissance (the dashed black line)
that was completed as part of our followon study on this issue. Due to access, schedule, and
funding limitations, the Independent Levee Investigation Team was not able to complete a full
and comprehensive survey of the entire greater New Orleans area for this element of our studies.
Locations of noteworthy performance have been identified in the numbered boxes on Figure
10.6, and these are discussed in further detail below (refer to Figure 2.6 in Chapter 2 for a
summary of design elevations for the flood protection system). Please note that these locations
are intended only to represent typical findings and are not intended to summarize the complete
performance of the overall flood protection system. In addition, the specified design flood
protection system component crest elevations may not be the actual crest elevations at the time
of hurricane Katrina’s arrival due to factors such as incomplete staged construction,
consolidation and settlement, regional subsidence, difficulties in correlating datums and
elevation benchmarks, etc.
It is important to emphasize that accurately determining elevations in the greater New
Orleans area is extremely complicated. Factors that exacerbate the problem include regional
subsidence, localized consolidation settlement, progressive settlement of benchmarks used to
establish regional datums, and the temporal variation in time of completion of individual project
sections. A tremendous effort has been undertaken by the IPET team to “equalize” all the
locations that are part of the New Orleans Flood Defense System and to merge the many
componentspecific elevations to one common projectspecific elevation datum.
Following is a summary and discussion of levee performance at select locations along the
perimeter of the New Orleans Flood Defense System sections traversed in this section of our
studies (as shown in Figure 10.6.) A detailed and more comprehensive assessment of the as
reconstructed levee system is recommended to be conducted upon completion of the ongoing
repairs and upgrades.
10.5.1 Location 1 – Lakefront Airport
Location 1 is situated near the intersection of Downman Road and Hayne Boulevard,
south of the old Lakefront Airport. At this location, an earthen levee connects to a railroad
bridge and vehicular underpass and an adjoining concrete floodgate structure, and the levee is
situated parallel to an active railroad line. This is a highly complex “penetration” (where several
access ways pass through or across the federal levee; including the rail line and the roadway),
and a complex set of “transitions” where disparate flood protection elements abut each other and
must perform well in combination. Unfortunately, this is one of numerous sites where these
elements were not adequately detailed and coordinated, and performance was unacceptably poor
and a breach occurred at this location.
10 13
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
High water marks, as reported by IPET (2006), at this location reached approximately
Elevation +12 feet (MSL). The design elevation of the levee system at this location was
Elevation +13.5 feet (NGVD29). Exact datum conversions in this area are not clearly
established and are still under review by the IPET team, but the design elevation has been
identified as Elevation +11.8 feet (MSL), resulting in some degree of overtopping at this
location.
Stormsurge induced overtopping traveled through the low spot at this complex
transition/penetration, which was the granular gravel ballast for the railroad line, and this flow
eroded the railroad line embankment, which served as a transition levee between the concrete
floodwall (design Elevation + 13.5 feet MSL = +11.8 feet NAVD 882004.65) and the earthen
levee (design Elevation +14.5 feet MSL = +12.8 feet NAVD882004.65) shown in Figure 10.7.
Figure 10.8 shows the location where overtopping occurred resulting in significant scour around
the floodwall and Figure 10.9 provides a view across the railroad line where the railroad line
embankment was eroded allowing for the terminus of the earthen levee to be scoured. Note that
at the time of our visit, the railroad embankment had already been repaired by railroad personnel.
Performance factors of the levee system that impacted the performance of the flood
protection components included the following: (1) unprotected highpermeability ballast (gravel)
which allowed high water levels to seep through the gravel ballast and erode the supporting
railroad embankment, (2) inadequate transition details between the flood protection components
which allowed for low points to be exploited, and (3) the presence of embankment and levee
materials that were not erosion resistant, resulting in scour as a result of overtopping. Without
redesigning this transition area, future performance at this location (under similar or more severe
storm surge conditions) is anticipated to be poor, and it will likely breach again.
10.5.2 Location 2 – Jahncke Pump Station Outfall
Location 2 is situated near the intersection of Hayne Boulevard and Jahncke Road, near
Lake Pontchartrain. At this location a concrete outfall structure protrudes through the flood
protection levee. High water marks, as reported by IPET (2006), at this location reached a
maximum Elevation of +12 feet (NAVD882004.65). The design elevation of the levee system
at this location was Elevation +14.5 feet (NGVD29). Exact datum conversions in this area are
not clearly established and are still under review by the IPET team, but the design elevation has
been identified as Elevation +12.8 feet (NAVD882004.65), resulting in a minor degree of
overtopping at this location. Our field reconnaissance verified that minor overtopping occurred
at this location, as can be seen in Figures 10.10 and 10.11.
Figure 10.10 provides an eastward looking view. Small patches can be seen on the levee
crest where minor erosion occurred. Figure 10.11 presents a view of scourrelated erosion
behind the concrete outfall structure transition.
Performance factors of the levee system that impacted the performance of the flood
protection components included the following: (1) placement of rip rap boulders along the Lake
Pontchartrain margin which aided in damping winddriven waves approaching the levee, (2) the
presence of the active railroad line which also aided in damping windwaves, and (3) utilization
10 14
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
of moderately erosionresistant embankment materials (moderately compacted clayey soils and
sandy clay soils.)
As shown in Figures 10.10 and 10.11, performance at this location was acceptable despite
the moderate overtopping; only limited erosion occurred at select locations across the earthen
levee crest and on the back face, and additional erosion occurred preferentially at several
soil/structure interfaces. None of this erosion was sufficient as to result in a breach of the flood
defenses at this location. The moderate erosion that did occur, however, suggests that more
severe overtopping would likely be more problematic and might well cause full breaching if this
section is not upgraded for erosion resistance.
10.5.3 Location 3 – Eastern Perimeter of New Orleans East
Location 3 is situated approximately 0.6 miles east of Highway 11 and approximately 1
mile north of Chef Menteur Highway (Hwy 90). In this vicinity, the flood protection system
consists primarily of earthen levees that are protected by both lowlying swamplands and trees
on both the outboard (water) side and the inboard (protected) side of the levee. High water
estimates, as determined through numerical modeling by IPET (2006), suggest that the water at
this location reached a maximum Elevation of approximately +16 feet (NAVD882004.65). The
design elevation of the levee system at this location was Elevation +14.5 feet (MSL). Exact
datum conversions in this area are not clearly established and are still under review by the IPET
team, but the design elevation has been identified as Elevation +12.4 feet (NAVD882004.65),
resulting in significant and relatively sustained overtopping at this location.
Figure 10.12 provides a southward looking view. The overall condition of the levees in
this area is excellent and no observable damage or erosion was encountered. As part of local
improvements after hurricane Katrina, an outfall access structure, near Hwy 90, was outfitted
with a rockgabion transition zone to minimize scour around the concrete access structure, as
seen in Figure 10.13. This is an excellent idea, and should serve to further improve the
performance of this detail in future events.
The excellent performance at this site was likely due to a number of factors, including:
(1) the presence of lowlying swamp areas which aided in damping windwaves approaching the
levee, (2) the presence trees and shrubs outboard and inboard of the levee which also aided in
damping windwaves, and (3) utilization of moderately to highly erosionresistant embankment
materials.
10.5.4 Location 4 – Southeast Corner of New Orleans East
Location 4 is situated at the southeast corner of the New Orleans East polder. In this
vicinity, the flood protection system consists primarily of earthen levees adjacent to the GIWW,
fronting Lake Borgne. A small stretch of lowlying swamp protects the outboard side of the
levees in this area, but it affords relatively little effective protection from winddriven waves
when “Lake” Borgne swells with storm surges from the Gulf.
High water marks, as determined by IPET (2006) using numerical simulations, suggest
that water levels at this location reached a maximum Elevation of approximately +16 feet
10 15
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
(NAVD882004.65). The design elevation of the levee system at this location was Elevation +19
feet (MSL). Exact datum conversions in this area are not clearly established and are still under
review by the IPET team. As discussed previously in Chapter 7, significant erosion and
breaching occurred at this location, and this length of levee frontage was the single largest point
of ingress for the floodwaters that eventually inundated the New Orleans East protected basin.
The IPET studies have ascribed this massive erosion principally to overtopping, but it is the view
of this investigation that considerable erosion also occurred due to wave action prior to full
overtopping, and that throughlevee seepage and underseepage may also have played a role at
some locations (see Chapter 7).
Figure 10.14 shows one of a number of eroded zones or “slots” of the original levee that
was breached and scoured as a result of stormsurge induced erosion during the hurricane. The
levee embankment at this section was comprised largely of material dredged from the excavation
of the adjacent GIWW channel, and large portions of this embankment appear to have had little
resistance to erosion.
Figure 10.15 shows completed levee rehabilitation work at the southeast corner. At the
time of our visit, construction activities had shifted approximately 1 mile west and consisted of
bellydump trucks placing borrow material which was being spread by bulldozers and track
walked. Dump trucks were also directed to travel over the newly placed levee, employing the
semicompaction technique defined in EM 111021913.
Performance factors of the levee system that impacted the unacceptable performance of
the earthen levees along this frontage included the following: (1) lack of slope protection (and
crest protection) to minimize erosion due to both stormdriven waves and overtopping flow, (2)
the site’s location adjacent to the “open” and relatively deep waters of Lake Borgne allowing for
significant winddriven waves to form and scour the flood side of the levee (and to notch the
crest), (3) the lack of useful protection from outboard side swamp and/or cypress groves, etc., to
reduce the energy and intensity of winddriven waves, and (4) the utilization of embankment fill
soils of variable erosion resistance (and permeability) so that both winddriven wave erosion and
throughflow erosion, as well as overtopping (wave splashover) erosion are all likely to have
been active at this location.
This levee frontage performed disastrously poorly, and these factors that contributed to
that unacceptable performance must be eliminated in the future.
10.5.5 Location 5 – Entergy Michoud Generating Plant
Location 5 is situated along the GIWW/MRGO shared channel, immediately beneath the
Hwy 47/Paris Road bridge. In this vicinity, the flood protection system consists primarily of
earthen levees. High water marks, as reported by IPET (2006), at this location reached a
maximum Elevation of +16.3 feet (NAVD882004.65). The design elevation of the levee system
at this location was Elevation +15 feet (MSL). Exact datum conversions in this area are not
clearly established and are still under review by the IPET team, but the design elevation has been
identified as Elevation +13.2 feet (NAVD882004.65), resulting in moderate to significant
overtopping at this location.. Our field reconnaissance verified that overtopping in fact occurred
at this location.
10 16
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 10.16 shows actual overtopping of the levee as captured by a security video
camera at the Entergy Michoud Generating Plant during Hurricane Katrina. Figure 10.17
presents a postHurricane Katrina view of the same section of levee shown in Figure 10.16.
Only minor damage occurred on the protected side, with a majority of the damage appearing to
have resulted from wave reflection from the adjacent bridge abutment. Excepting the section
where some minor erosion occurred on the inboardside slope of the levee, this levee frontage for
many hundreds of feet in each direction showed little evidence of damage from relatively
sustained, moderate to significant overtopping flows. The overall condition of the levees in this
area was good and no major damage was encountered.
The good performance of this embankment in the face of sustained moderate overtopping
was likely due to several factors including: (1) utilization of moderately to highly erosion
resistant embankment materials (clay), and (2) the small fetch of the GIWW/MRGO channel at
this location which limited the height of the windgenerated waves.
10.5.6 Location 6 – GIWW/MRGO Southern Shoreline Levee
Location 6 is situated along the ICWW/MRGO interchange beneath the Hwy 47/Paris
Road bridge, on the opposite side of the channel (on the south bank) from Location 5. In this
vicinity, the flood protection system consists primarily of earthen levees with a concrete
floodwall beneath the bridge that connects the eastern and western levee segments. High water
marks, as reported by IPET (2006), at this location reached a maximum Elevation of +16.3 feet
(NAVD882004.65). The design elevation of the levee system at this location was Elevation +14
feet (MSL). Exact datum conversions in this area are not clearly established and are still under
review by the IPET team, but the design elevation has been identified as Elevation +13.2 feet
(NAVD882004.65), resulting in significant and likely relatively sustained overtopping at this
location. Our field reconnaissance verified that overtopping occurred at this location.
The overall condition of the levees in this area was good and no major damage was
encountered. The concrete floodwall constructed beneath the Hwy 47/Paris Road bridge, due to
its top of wall elevation being lower than that of the neighboring earthen levees, acted as a weir
during the high water period and “sucked” in nearby steel barges, as shown in Figure 10.18.
Despite the collision impact of the barges with the concrete wall, the system performed well.
Some scourrelated damage was observed at the transition between the concrete flood wall and
the earthen levee. Figure 10.19 presents an eastward looking view of the levee, just west of the
washed up barges. East of the Hwy 47/Paris Road bridge, Figure 10.20 shows a gas processing
barge that collided with the earthen levee. The impact did not result in significant damage to the
levee.
The good performance of the levee system at this location was likely due in large part to
the utilization of moderately to highly erosionresistant embankment materials; these erosion
resistant (clayey) materials were also capable of absorbing impact loads from the barges,
allowing the barges to come to rest on the levee without breaching it. The small fetch of the
GIWW /MRGO canal at this location may also have limited the height of the windgenerated
waves, thereby minimizing waveinduced erosion of the levee materials.
10 17
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
During a subsequent visit to this site in March of 2006, we observed that a concrete apron
was being installed by the USACE around the transition between the concrete floodwall and
earthen levee. This transition detail is anticipated to minimize future erosion at this transition
area during a future severe event and reduce the risk of an erosioninduced full breach of the
levee.
10.5.7 Location 7 – Bayou Bienvenue Control Structure
As shown in Figure 10.6, Location 7 is situated along the northern end of the MRGO
levee frontage, near the northeast corner of the St. Bernard Parish protected basin. In this
vicinity, the flood protection system consists primarily of earthen levees that connect, with a
concrete control structure (with steel floodgates) across Bayou Bienvenue. High water marks, as
reported by IPET (2006), at this location reached a maximum Elevation of +18.4 feet (NAVD88
2004.65). The intended eventual design elevation of the levee system at this location was
Elevation +17.5 feet (MSL). Exact datum conversions in this area are not clearly established and
are still under review by the IPET team, but the design elevation has been identified as Elevation
+13.2 feet (NAVD882004.65). As discussed in Chapter 6 (see Section 6.2), this levee frontage
was incomplete at the time of hurricane Katrina’s arrival, as a final crest raising to offset
consolidationinduced settlements has not yet been implemented. As a result, major overtopping
occurred at this location.
Figure 10.21 shows an aerial photograph of the Bayou Bienvenue control structure. The
floodwall to the north of the control structure performed very well, withstanding an impact load
from a steel barge which became lodged atop the concrete flood wingwall attached to the
concrete control structure.
The southern side of the control structure did not perform well; massive erosion and
scour produced a major breach at the contact between the concrete control structure and the
adjacent earthen levee. The southern side of the control structure was built using mainly spoils
from the excavation of the adjacent MRGO channel that are more erosionsusceptible than the
clays on the northern side of the control structure. It is important to note that both sides of this
control structure were subjected to similar loading conditions and overtopping occurred on both
sides and as a result, this site offers a unique example of the importance of erosionresistant soil
materials. In addition, the southern portion of the control structure abuts the abandoned Bayou
Bienvenue channel, as shown in Figure 10.21. It is not conclusive whether the backfill materials
into the abandoned channel impacted the performance of the control structure, but further
investigation should be employed to determine the performance factors for this side of the
control structure.
Figure 10.22 shows a close up view of the flood control gate structure that acted as a weir
as the water overtopped the flood protection system. Significant scour and erosion was observed
around the structure. Upon a follow up visit in March of 2006, splashpads had been installed
behind the flood gate structure. In addition, the steel barge had been removed from the concrete
control structure and a vast sea of riprap protection installed around the control structure.
Figure 10.23 presents a picture of the installed splash pads and Figure 10.24 presents a view of
the control structure with the barge removed and placement of riprap. Figure 10.25 presents a
10 18
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
schematic of the mapped scour around this concrete structure observed in October of 2005.
Around the ends of the concrete wall, about 10 feet of soil have been eroded.
Factors of the levee system that impacted the performance of the flood protection
components included the following: utilization of moderately to highly erosionresistant
embankment materials on the northern end of the control structure, utilization of moderately to
highly erosionsusceptible embankment materials on the southern end of the control structure,
and possible effects of the old Bayou Bienvenue channel abandonment backfill materials on the
southern portion of the control structure. Future performance at this location under similar or
more severe conditions is anticipated to be good for the northern half of the control structure and
poor for the southern half of the control structure. Significant overtopping should be expected
for larger storm surge events.
10.5.8 Location 8 – Mississippi River Gulf Outlet
Location 8 is situated along the western edge of the MRGO, south of the Bayou
Bienvenue Control structure and north of the Bayou Dupre Control structure. In this vicinity, the
flood protection system consists primarily of earthen levees constructed from excavated
materials from the MRGO channel. High water marks, as reported by IPET (2006), at this
location reached a maximum Elevation of approximately +18 feet (NAVD882004.65). The
design elevation of the levee system at this location was Elevation +17.5 feet (MSL), however,
reports indicated that these levees were not fully completed and had crest elevations that were 3
to 4 feet lower than the specified design elevation. In addition, exact datum conversions in this
area are not clearly established and are still under review by the IPET team, but the design water
level has been identified as Elevation +12.7 feet (NAVD882004.65). During Hurricane Katrina,
moderate to major overtopping occurred at this location. Our field reconnaissance verified that
moderate to major overtopping occurred.
Figure 10.26 shows a United States Geological Survey topographic map of the MRGO
area. The identified “spoil area” corresponds to the zone of poor levee performance. Figure
10.27 shows aerial photography taken by NOAA in early September 2005 along the MRGO and
shows severe erosion/breaches in the levee and barges that floated over the top of the levee and
came to rest inside St. Bernard Parish after water elevations receded. Figure 10.28 shows close
up aerial photographs of the severely eroded levees.
A bank erosion study was performed by the USACE (1988) that identified the presence
of highly erosionsusceptible soils within the MRGO alignment. Merchant shipping traffic that
traversed the MRGO created wakeinduced waves and drawdown that were eroding the channel
banks, resulting in the widening of the MRGO from an intended 650 feet to an actual average
width of 1,500 feet, more than double the design width. Comments submitted from the Lower
Mississippi Valley District on the report made the following comment in response to selecting
the bank erosion mitigation alternative of decommissioning the MRGO:
The alternative to completely close the MRGO waterway should be
evaluated….This alternative will control all future channel maintenance problems
by controlling bank erosion, preventing the associated biological resource
problems, preventing saltwater intrusion, and lessening the recreational losses.
10 19
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
In addition to solving the aforementioned problems, it will also reduce the
possibility of catastrophic damage to urban areas by a hurricane surge coming up
this waterway and also greatly reduce the need to operate (and could possibly
eliminate) the control structures at Bayous Dupre and Bienvenue
Slope protection measures were recommended to aid in stabilizing these highly erosive
deposits against waveinduced erosion. At the time of Hurricane Katrina, slope protection
measures along the flood side of the MRGO levee were not in place. As identified in the above
comments, the Hurricane Katrina stormsurge massively eroded the levees and resulted in
catastrophic failure.
Performance factors of the levee system that impacted the performance of the flood
protection components included the following: utilization of highly erosive embankment
materials, lack of appropriate surface slope protection to minimize erosion of the flood side of
the levee during the stormsurge, and asconstructed crest elevations below design elevations
allowing for significantly higher water overtopping heights. Future performance, based on prior
performance, at this location under similar or more severe conditions is anticipated to be poor
unless improved materials and construction methods are used. We were unable to sample this
location and test the materials for erodibility, but we did perform a follow up visual
reconnaissance in April of 2005, and at that time we were very favorably impressed at the calibre
of the work, and of the materials, being placed and compacted. During that followon
reconnaissance, we did not detect the use of unsuitable fills, but we would like to return to
sample and test the soils used to construct the levee in this section of the MRGO.
10.5.9 Location 9 – Bayou Dupre Control Structure
Location 9 is situated approximately 6.5 miles southeast of the Bayou Bienvenue Control
structure, on the west side of the MRGO. In this vicinity, the flood protection system consists
primarily of earthen levees that connect via a concrete and steel flood access structure to a
concrete control structure across Bayou Dupre. High water marks, as reported by IPET (2006),
at this location reached a maximum Elevation of +17 to +22 feet (NAVD882004.65). The
design elevation of the levee system at this location was Elevation +17.5 feet (MSL). Exact
datum conversions in this area are not clearly established and are still under review by the IPET
team, but the design water elevation has been identified as Elevation +12.7 feet (NAVD88
2004.65), resulting in moderate to major overtopping at this location. Our field reconnaissance
verified that moderate to major overtopping occurred.
Figures 10.29 and 10.30 show aerial photographs of the Bayou Dupree control structure.
The area to the south of the control structure performed very well, while the northern side of the
control structure did not perform well, with significant erosion and scour as a result of the
overtopping. The northern portion of the control structure abuts the abandoned Bayou Dupre
channel, as shown in Figure 10.31. It is not conclusive whether the backfill materials into the
abandoned channel impacted the performance of the control structure, but further investigation
should be employed to determine the performance factors for this side of the control structure.
Figures 10.32 and 10.33 show aerial photographs of repair operations underway at
Bayou Dupre in January 2006. As can be seen in Figure 10.33, sand borrow material has been
10 20
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
imported to be used in the backfilling repair operations in the deep scour pools on the north side
of the control structure.
Performance factors that impacted the performance of the flood protection components at
this location included the following: utilization of highly erosive embankment materials, lack of
appropriate surface slope protection to minimize erosion of the flood side of the levee during the
stormsurge, asconstructed crest elevations below design elevations allowing for significantly
higher water overtopping heights, and possible effects of the old Bayou Dupre channel
abandonment backfill materials on the northern side of the concrete control structure.
10.5.10 Location 10 – St. Bernard Parish Interior Levee (Forty Arpent Levee)
Location 10 is situated north of the Corinne Canal and approximately ¾ miles east of
Hwy 47/Paris Road. In this vicinity, the flood protection system was designed to be a secondary
containment system for potential overtoppingrelated flooding behind the MRGO frontage levees
and to act as a barrier against rainwater that is discharged into the swamp area. This part of the
flood protection system consists primarily of earthen levees with a design Elevation of +8.0 to
9.0 feet (MSL). The actual elevation of this system during Hurricane Katrina was on the order of
5 to 8 feet (MSL) (note that IPET did not establish NAVD882004.65 elevations at this
location). High water marks were not reported by IPET at this location. Based on our field
reconnaissance, it was apparent that major overtopping occurred at this location.
Figure 10.34 shows an eastward looking view of the earthen levee. Although this levee
was significantly overtopped, it did not experience significant damage or erosion. Figure 10.35
shows a fishing boat that was washed over the levee shown previously in Figure 10.34 and came
to rest in a residential neighborhood. The clearly excellent erosion resistance of this secondary
levee, despite significant overtopping, was the result of use of cohesive, clayey soils as the levee
fill material. The ability of these soils to sustain significant overtopping without catastrophic
erosion is an important lesson.
Upon a subsequent visit to this location in March of 2006, we observed that the levee had
been improved and raised by several feet to a new Elevation of +10 feet (MSL). Figure 10.36
presents the same view as in Figure 10.34, but 5 months later. Based on our observations, it
appeared that cohesive soils and semicompaction construction methods were used for this
improvement. These soils appeared to be in a relatively erodeable condition as initially placed,
but it also appeared likely that significant wetting from rainfall would improve their resistance to
erosion from overtopping.
The performance factor that most significantly influenced the observed surprisingly good
performance of the flood protection levee embankment at this site was the utilization of
moderately to highly erosionresistant embankment materials. Future performance of the levee
embankment at this location under similar or more severe conditions is also anticipated to be
good, however, significant overtopping should be expected for larger storm surge events which
may cause the MRGO levees to breach, as this secondary levee has a crest elevation well below
that of the main MRGO frontage levees.
10 21
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
10.5.11 Summary of Observed Performance Factors with Regard to Erosion Due to Overtopping
Based on observations from our field reconnaissance and review of aerial photographs, it
is apparent that the performance and post Hurricane Katrina conditions of the earthen levee
systems varied significantly, from good performance in areas with major overtopping to poor
performance in areas with minor degrees of overtopping.
Table 10.3 presents a summary of the 10 locations evaluated as part of this study. Most
of the levees studied in this pilot study were overtopped as a result of the large storm surge that
rushed onshore. The magnitude of the storm surge and resulting overtopping did not, however,
singularly dominate the observed performance of the levees. Additional factors were the degree
of outboard side protection afforded by cypress swamps (which diminished wave energies), and
the intrinsic resistance of the levee embankment and foundation soils to erosion.
Table 10.3 lists these factors, as well as the observed performance of the sections studied.
As with the ILIT team’s overall observations and studies of failures and successful performances
throughout the New Orleans Flood Defense System, it is apparent that the use of highly
erodeable soils such as cohesionless (sandy) soils represents a potentially unacceptably
hazardous condition, and that the use of suitably compacted cohesive, clayey soils with relatively
high intrinsic resistance to erosion can provide a measure of ductility and resilience to an
otherwise brittle system (levees that can overtop for some number of hours, without
catastrophically eroding and breaching.)
10.6
Erosion Susceptibility Evaluation
To date, the field of scour and erosion has not been well characterized and there is a
paucity of welldefined field case studies that relate actual performance to design parameters.
Both the USACE and FEMA design guidelines do not specifically provide acceptability criteria
for erosion susceptibility due to the lack of comprehensive knowledge in this area. As a result of
the NOFDS being “loaded to failure,” it has provided an unfortunate opportunity to recognize
lessons learned and improve our body of knowledge for the performance of levee flood
protection devices.
It is important to note the magnitude of devastation caused to the regional flood
protection system as a result of erosion. The levee system along the MRGO, which is the main
protection mechanism for the 100,000+ citizens of St. Bernard Parish (and the Lower 9th Ward)
against storm surges from the Gulf of Mexico, was catastrophically degraded as a result of
erosion during Hurricane Katrina. Similarly catastrophic erosion occurred at the “sister” section
of levee at the southeast corner of the New Orleans East protected basin. Figure 10.37 shows a
comparison of two LIDAR surveys of the MRGO levee system at the Bayou Bienvenue control
structure, near the intersection of the GIWW at the northeast corner of St. Bernard Parish.
Effects of subsidence can be clearly seen in the elevation differences on the north side of the
control structure and the control structure itself between 2000 and immediately following
Hurricane Katrina in 2005. The levee on the north side of the Bayou Bienvenue control structure
was largely undamaged. The levee on the south side of the control structure was catastrophically
damaged. The magnitude of the erosion has been highlighted and white “splotches” of displaced
levee materials can be seen in the aerial photograph.
10 22
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
There are many factors that influence the erosion susceptibility of soils, and a more
comprehensive discussion of these was provided in Chapter 9. Fundamentally, soil erosion is
controlled by the resisting characteristic of the soil (including soil type and character, soil fabric
and structure, insitu density, etc.) and eroding forces (including the magnitude and duration of
the shear stress applied to the soil, impact or jetting pressures, etc.) from the contacting (eroding)
fluid.
A sampling and laboratory testing program was devised upon completion of our field
levee condition survey and mapping in October of 2005, to try to understand and characterize the
properties associated with the levees that performed well and the characteristics of the levees that
did not perform well during Hurricane Katrina. The intent of this study was to better understand
the nature of the levee materials that performed well during the extreme conditions in order to
provide recommendations on how to improve the sections of earthen levees that did not perform
well.
Insitu samples were collected and from select levee sites during January and February of
2006. The selected sampling sites included levees that performed very well during Hurricane
Katrina, levees that performed moderately well, and levees that did not perform satisfactorily.
Figures 10.38 and 10.39 identify the locations where samples were collected for laboratory
analyses.
Erosion susceptibilities of the soils were characterized using a state of the art erosion
index testing method, developed by Dr. JeanLouis Briaud at Texas A & M University, known as
an Erosion Function Apparatus (EFA). This test method required undisturbed samples to be
sampled from the field and be carefully transported back to Texas A&M University for analyses.
As described in Chapter 9, the EFA is a test that determines the shear stress and velocity
of flowing water required to erode soil from a cylindrical tube that is slowly advanced into a
rectangular pipe of flowing water. The more erosion resistant the soil, the faster the water (and
the higher the shear stress) is required to flow in the rectangular pipe in order to erode the soil
sample. A diagram of the EFA is presented in Figure 10.40. The measured shear stress at the
point at which the soil begins to erode is defined as the critical shear stress. Shear stress less
than the critical shear stress will not result in erosion, whereas applied shear stresses in excess of
the critical shear stress will result in erosion. Determination of the erodibility index is useful in
completing analyses for overtopping and surface erosion.
Upon completion of the test, the erodibility index of the soil was defined and the rate of
erosion as a function of applied shear stress (or velocity) established. This relationship can then
be compared with anticipated shear stresses the soil will experience in the field. If the estimated
field shear stresses are less than the shear stress required to erode the soil, no erosion is
anticipated to occur. If the field shear stresses exceed the laboratory determined critical shear
stress, the erodibility index provides a means by which to estimate the magnitude of the overall
erosion.
10 23
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
In addition to the erosion testing itself, additional engineering characteristics of the
earthen levee sections under study were also characterized. These characteristics included the
following:
テ
テ
テ
テ
テ
テ
Gradation, including passing the Number 200 sieve (ASTM D422)
Hydrometer (ASTM D422)
Atterberg Limits Determination (ASTM D4318)
Unconfined Compression (ASTM D2166)
Dry density and moisture content determination (ASTM D4937/2216)
Maximum dry density determination (ASTM D1557)
Table 10.4 presents a summary of the locations where samples were collected for
analyses, and these locations are shown in Figures 10.38 and 10.39.
The samples were collected by pushing an approximately 3inch diameter steel (Shelby)
tube into the ground to retrieve soil samples using a geotechnical testing drill rig. Sites 4, 5, and
6 were located along the MRGO section of levee that suffered severe overtopping and erosion.
At these sites the levee materials were collected in a soil sample bag and reconstituted back in
the laboratory due to the highly disturbed nature of the levee materials.
The erosion susceptibilities of all the samples collected are presented in Figure 10.41.
The test result designations are based on the Site Number, the boring number at the site, the
depth interval over which the sample was collected, and additional sample notes. Thus, a sample
marked as S1B1(02ft)TW indicates that the sample came from Boring 1 at Site No. 1 (Levee
east of HWY 11 and North of HWY 90) from a depth of 0 to 2 feet below the crest of the levee.
The results of the EFA test results are also presented in Table 10.4. The EFA test results
matched very well with the observed performance in the field. Areas where the levee
performance was observed to be good generally had low to very low erosion susceptibilities. In
areas where the levee performance was poor, the materials had a high to very high erosion
susceptibility.
The effects of material compaction were also evaluated. Previous work has been
performed in this area and a design guideline prepared by FHWA (1988). Figure 10.42 shows
that for a material of a given plasticity index, the permissible shear stress increases nearly tenfold when the material is properly compacted.
Figure 10.43 shows the dramatic impact proper compaction can have on the erodibility of
some types of soils. Materials sampled from the MRGO levee were tested at two compaction
levels: low compactive effort and high compaction effort. The corresponding results speak
volumes to the importance of compaction in earthen levees. The lowcompaction sample was
found to be very highly erodible, whereas the highcompaction sample exhibited low erodibility
characteristics. High compaction effort is based on 9095% relative compaction per the
Modified Proctor test (ASTM D1557).
Figure 10.44 provides a summary from the pioneering erodibility work performed by Dr.
Briaud et al. at Texas A & M University as part of these studies by (see Chapter 9). Soils that
fell within the very high to high erodibility categories are prone to failure by overtopping. Soils
10 24
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
that fell within the medium erodibility catergory fell in a transition zone, and soils that fell within
the low to very low erodibility categories were shown to be resistant to erosion induced failure as
a result of overtopping. These laboratory test results were wellvalidated by actual levee
performance during Hurricane Katrina.
10.7 Observed Failure Modes During Hurricane Katrina
Table 10.5 presents a summary of the types of failure mechanisms of earthen levees that
were observed in the greater New Orleans area flood defense system, including the ten locations
identified above. In addition, the required design evaluations per the USACE and FEMA
guidelines have also been summarized in this table.
Overtopping, jetting, internal erosion and piping, underseepage and piping,
underseepageinduced instability, and lateral and semirotational foundation instability failure
mechanisms were all observed in the greater New Orleans region during Hurricane Katrina.
There were multiple locations where trees that were rooted within the levee zone had fallen over
and may have contributed to the failure of the levee. Liquefaction may have been a partial
contributor to some of the failures along the MRGO levee system. The forces associated with
the breaking waves impacting the MRGO levee may have been sufficient to induce liquefaction
in the relatively weak foundation materials.
Some notable attributes were observed to be associated with levees that performed well
during the hurricane. These attributes included:
テ
テ
テ
テ
Utilization of erosionresistant soils for levee construction;
Gradual soil/structure transition zones (rock gabions around concrete structures);
Presence of lowlying swamp and vegetation on the outboard sides to dampen windwaves;
and
Presence of riprap protection.
Some notable attributes were observed to be associated with a number of levees that
performed poorly during the hurricane. These attributes included:
テ
テ
テ
テ
Utilization of low erosionresistant construction material;
Transitions between different flood protection component types;
Lack of surface slope protection for erosionsusceptible soil levees; and
Abandoned channel backfills underlying levees.
10.8 Establishment of Design Criteria and Acceptable Performance
Varying degrees of levee reliability and performance are required based on the assets that
the levees are protecting. Early in the last century, most U.S. levees protected agricultural farm
lands, where the consequences of levee breaching and subsequent flooding resulted primarily in
the loss of crops. With the growth of urban areas into lowlands adjacent to rivers and coastlines
10 25
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
over the past two to three hundred years (the first urban levee was constructed in New Orleans in
1718), and especially over the past century, levees have become increasingly important defense
mechanisms for industry, vital infrastructure elements such as drinking water, sewage, and
electricity transmission, and, most importantly, large populous regions (cities and towns) and
thus human life. Little guidance is provided by either the USACE or FEMA design criteria as to
what are acceptable design standards for highconsequence urbanized areas vs. lowconsequence
agricultural areas.
The current approach to establishing design standards utilized by the USACE is
conducting costbenefit analyses. This financial/risk evaluation procedure analyzes the cost of
achieving a certain level of protection and compares it with the recognized benefit associated
with that level of protection. Unfortunately, the cost/benefit model used by the USACE for levee
systems does not account for: (1) the loss of human life, (2) economic losses to cities, counties,
and states as a result of a nonoperational and nonfunctional revenue base (e.g.: businesses shut
down due to damage and lack of utilities, lack of a work force to operate businesses, lack of a tax
base due to the displaced residents, and lack of tourists, etc.), and (3) numerous other costs and
losses known collectively as “secondary” and “tertiary” costs/losses associated with economic
ripples that spread farther a field from the immediate locale of the disaster in question (as
opposed to the more easily quantifiable primary or “direct” losses associated with system failure
and consequent flooding.) These secondary and tertiary losses become increasingly important as
the scope of a disastrous failure increases; these uncounted losses are most pronounced for full
blown “catastrophes.”
This is unfortunate, as it results in systematic undervaluation of the likely benefits of
investing effort and resources to prevent disasters before they occur. In the case of hurricane
Katrina, estimates of “losses” due to the catastrophic flooding of approximately 85% of the
greater New Orleans region vary significantly at the time of this writing, but most independent
estimates are on the order of ~$100 to $200 billion, and some estimates range as high as $400
billion. A clear outlier is the recent estimate of damages proffered by the IPET study’s Draft
Final Report (IPET; June 1, 2006) which estimated these damages at approximately $25 billion.
This was largely a function of the procedures which the USACE is required to employ in making
potential loss projections; these systematically undercount expected losses, and they also
undercounted actual losses in the wake of Katrina. This systematic underevaluation of projected
(and real) losses provides a poor basis for subsequent decisionmakers (e.g. federal and local
government) to base decisions regarding appropriate allocation of resources to defend against
risk and threats.
10.8.1 USACE Risk Management Approach
In response to budget constraints, increased situations requiring costsharing, and general
public concern for the performance and reliability of completed projects, the USACE has
evaluated the use and methodology of riskbased analyses (especially as related to the
geotechnical components of these projects). A seminar was convened in 1983 by the USACE in
order to “incorporate more information into the safety assessment [of projects] than [traditional]
factor of safety methods.” A more recent evaluation was undertaken, and the results of this
effort are presented in Engineering Technical Letter (ETL) 11102556, published on May 28,
10 26
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
1999. This study recognized that there are inherent uncertainties associated with infrastructure
problems and that the total effect of risk and uncertainty on a project’s economic viability should
be examined in order for “conscious decisions” to be made reflecting “explicit tradeoffs
between risk and cost.”
According to ETL 11102556, major sources of uncertainty that require evaluation
include the following:
テ
テ
テ
テ
テ
テ
テ
Uncertainty in loadings;
Uncertainty in engineering analysis parameters;
Uncertainty in analytical models (model bias);
Uncertainty in performance;
Conversion of empiricallyderived performance modes;
Frequency and magnitude of physical changes or failure events; and
Conditions of unseen features.
ETL 11102556 identifies special situations uniquely applicable to geotechnical
problems that result in uncertainties with large magnitudes:
テ
テ
テ
テ
テ
Natural earthen materials generally exhibit high variability in composition and
engineering properties;
Engineering characteristics of soils can exhibit high variability due to
composition, deposition, sampling, and field & laboratory testing procedures;
Engineering analyses can be performed assuming total stress (excluding the
effects of groundwater) or effective stress (including the effects of groundwater).
As a result, groundwater uncertainties may either be included or excluded in the
analyses;
Consideration of spatial correlation of soil properties is required due to the
variability of deposition history; and
The spatial scale of the project (as much as tens of miles long for levees) requires
“sectioning” of the system into subcomponents.
These uncertainty factors can result in very large ranges and broad distributions for
parameter bounds. For example, a mean soil shear strength value, determined based on a
subsurface field sampling and laboratory testing, which is used to evaluate the stability of levee
slopes may naturally vary between ± 30% of the mean value to as much as ± 75% of the mean
value. These broad distributions significantly impact the reliability of the resulting calculated
answer.
The report summarized recommended target reliabilities for expected performance levels.
These target reliabilities are presented in Table 10.6.
The approximate median Factors of Safety associated with the established expected
performance levels were added to the target reliability indices presented in ETL 11102556
according to the following formula:
F.S.50 = e (evlnRS)
10 27
[Equation 10.2]
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
In this formulation, e#is the safety index, R is the capacity, and S is the loading/demand. A fairly
typical coefficient of variation of 30% is assumed in both the loading and capacities, and a
lognormal distribution for the system and the loads are assumed.
It is interesting to note that this approach had not yet been applied to the New Orleans
regional flood defense system. Instead, the system had been designed using more “traditional”
approaches, including the common use of a required Factor of Safety of FS ユ 1.3 for stability
analyses associated with the shortterm (“transient”) conditions produced by hurricaneinduced
storm surge, winds and waves.
If this is then compared with the projections of Table 10.6, the expected performance for
these design criteria would be anticipated to be “unsatisfactory”, with an approximately
estimated 7% probability of failure. It is, of course, a matter of judgement as to how many
individual segments and intersections comprise the overall New Orleans regional flood
protection system, and also how accurately the “typical” coefficients of variation for both
loadings and resistances characterize each of these, but the overall accuracy of the projection of
expected performance based on the simplified estimates of Table 10.6 could be argued to be
welljustified by the multiple and catastrophic failures of the flood protection system that
occurred during hurricane Katrina.
10.8.2 Other RiskBased Approaches
Alternative approaches to establishing design levels exist. John Christian (2004)
summarized studies that backcalculated the annual probability of failure based on failures of
actual engineered systems by Baecher et al. The lives lost or financial losses associated with the
failures were plotted against the backcalculated annual probabilities of failure. This plot
provides a mechanism by which to ascertain the “targeted” level of performance (the targeted
level of reliability) of a given engineered structure, based on historic practice in a number of
diverse fields. The resulting plot is presented in Figure 10.45.
Additional examples include both the Netherlands and Hong Kong, which have risk
based decision making tools (Figure 10.46) as part of their planning process to establish
acceptable levels of safety for engineered systems protecting significant populations based on the
expected number of fatalities, as well as expected financial “losses” (Christian 2004). Failures
that might impact large populations and may result in large numbers of fatalities (e.g.: greater
than 1,000) are required to have very low annual probabilities of failure.
Using the risk management planning relationship (Figure 10.46) developed by the Hong
Kong Government Planning Department as an example, a proposed engineered system that has
the potential to result in 1,000 fatalities would have an acceptable risk (based on an annual
frequency of occurrence) of 108, the range over which the principle “As Low As Reasonably
Prudent” (ALARP) reliability level is recommended varies from annualized Pfpa of 108 to 106,
and a risk with a Pfpa of less than 106 is considered unacceptable for any case.
10 28
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Table 10.7 presents a summary of calculated Annual Return Periods (in years), based on
the probability of occurrence limits recommended by the Hong Kong Government Planning
Department. This example highlights the sociological decision made in Hong Kong that high
consequence events should occur very infrequently, with an annual return period of 1 million
years! Although this may not be realistic due to natural and anthropogenic uncertainties, the
premise of varying acceptable risk as a function of consequences is rational and feasible.
The Dutch have welldeveloped riskbased approaches for targeting the reliability of
flood protection systems. Like the southern Louisiana region, the Dutch face two distinct types
of flooding risk; river floods and flooding from catastrophic North Sea storms and their
associated storm surge, waves and winds. The Dutch lost a large fraction of their nation to ocean
storm flooding in the mid 1950s, and determined to develop rationally riskbased flood defense
systems to prevent recurrence of similar catastrophic flooding in the future. The levels of
targeted flood defense levels, in terms of return periods, are interesting in contrast to the
approximately 200 to 300year recurrence level that was nominally targeted for New Orleans.
The Dutch use a recurrence level of flood loading on the order of 1,00010,000 years for river
floods of populous areas (major towns and cities), and 10,000 years for ocean storms. For less
populated (largely agricultural) areas, recurrence levels targeted for flood protection design are
on the order of 5001,000 years.
All of these targeted levels of protection greatly exceed the targeted levels for the New
Orleans regional flood protection systems.
Another good way to look at targeted levels of reliability for flood protection is to reexamine Figure 10.45. This figure is replotted as Figure 10.47, with a red crosshatched region
showing approximately the level of reliability associated with the New Orleans regional flood
protection system at the time of Katrina’s arrival. Based on the studies of the ILIT team, it is
inferred that the New Orleans flood defense system would have been expected to fail
catastrophically about once every 50 to 100 years. The consequences of the failures during
Katrina were on the order of 1,300 lives lost, and/or about $100 to $200 billion in economic
losses.
A red dashed line indicating the approximate levels of reliability targeted by current U.S.
practice with regard to dams is also shown in Figure 10.47. It is interesting to note that current
U.S. practice for dams would have called for approximately three orders of magnitude higher
level of reliability than that which was apparently provided by the New Orleans regional flood
protection system (a factor of approximately 1,000 times higher reliability or assured safety.)
That is largely because “dams” are generally assumed to be associated with large consequences
in the event of failure. Few U.S. dams, however, have likely consequences of failure larger than
those associated with the failure of the New Orleans regional flood defense systems during
hurricane Katrina. This would appear to suggest that the flood defense systems for major
metropolitan areas should be engineered to have targeted levels of reliability on a par with those
commonly targeted for U.S. dams.
10 29
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
10.9 Brief Comments on Drainage and Pumping
As discussed in Chapter 4, the New Orleans area is situated below sea level and
significantly below the river level in the Mississippi River. The proper functioning of the
drainage and pumping system is critical to ensure the population of New Orleans is not
inundated by groundwater, rainwater, or floodwater. Essentially every drop of rain that falls into
New Orleans has to be pumped out, because most of the developed metropolitan and suburban
regions are below mean sea level. In addition, underseepage constantly passes in beneath the
perimeter levee systems, and has to be pumped out as well.
Accordingly, despite the potential critical floodfighting contribution of the drainage and
pumping systems, “the pumping stations have not been considered to be part of the hurricane
protection system except in a few instances where the buildings are structural part of a levee or
floodwall” (IPET, 2006). Based on the observed poor performance of the pumping systems
during Hurricanes Katrina and Rita, this perception of the drainage and pumping system must be
changed in order to provide more reliable system performance in the future.
There are about 80 pumping stations (IPET, 2006) in the four study areas, and a majority
of these pumping stations are more than over 50 to 100 years old. The pump stations are powered
by an assortment of electrical supply systems, and most of the older pump stations are powered
by antiquated 25HZ power generation facilities constructed in the late 1910’s and early 1920’s.
As a result of Hurricane Katrina, approximately onethird of the total system pumping capacity
within the study areas was lost after the passing of the hurricane. Only 16% of the pumping
stations were fully operational during the hurricane.
An extensive review of the performance of the pumping and drainage systems in the
Jefferson, Orleans, Plaquemines, and St. Bernard parishes was beyond the scope of our ILIT
studies. Such an evaluation was performed by IPET, however, and the results are presented in
their draft final report (IPET; June 1, 2006). An overview map of the parishes and regions
studied by this element of the IPET studies is presented in Figure 10.48.
Figure 10.49 presents a detailed view of pump stations within Jefferson parish. The
maximum pumping capacity within Jefferson parish is 48,460 cubic feet per second (cfs) by a
total of 27 pumping stations that drain an area of 73,500 acres (IPET, 2006). Figure 10.50
presents a detailed map of pump stations within Orleans parish. The maximum pumping
capacity within Orleans parish is 48,900 cfs from a total of 23 pumping stations that drain an
area of 60,000 acres (IPET, 2006). Figure 10.51 presents a detailed map of pump stations within
Plaquemines parish. The maximum pumping capacity within Plaquemines parish is 12,065 cfs
from a total of 21 pump stations that drain an area of 55,000 acres. Figure 10.52 presents a
detailed map of pump stations within St. Bernard parish. The maximum pumping capacity
within St. Bernard parish is 7,000 cfs from a total of 8 pump stations that drain an area of 17,620
acres.
IPET identified four major failure modes for pump station malfunctions during Hurricane
Katrina. These consisted of (1) loss of operational staff as a result of evacuation orders at
pumping stations that required manual operation, (2) loss of potable water to lubricate and cool
pumps during operation as a result of municipal water distribution system malfunction, (3) loss
10 30
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
of electricity to power the pumping station, and (4) shutdown or disruption of the pump
facilities as a result of flooding. Pumping failure from evacuations, flooding, loss of electricity,
and loss of lubrication (potable water) accounted for 46%, 26%, 8%, and 4% loss of total
pumping capacity, respectively (Figure 10.53). These failures and breakdowns do not reflect
multiple failure modes, such as a pump station being evacuated, only later to be overwhelmed as
a result of breaches in the flood defense system (i.e. a break in the drainage canal wall such as at
17th Street Canal).
In addition to these four failure modes, there are three other significant pump and
drainage systembased failure modes that impacted the proper performance of the pump and
drainage system during Hurricane Katrina. These failure modes are (1) reverse flow (Figure
10.54) through the pumping stations due to inadequate pump discharge elevation clearance (and
a lack of internal reverseflow protection), (2) loss of drainage capabilities as a result of breach
of the drainage canal [such as within the 17th Street Drainage Canal, the Orleans Canal, and the
London Avenue Drainage Canal as shown in Figure 10.55], and (3) lack of sufficient temporary
storage capacity [such as within St. Bernard parish between the interior levee and the MRGO
exterior levee], where discharged water is ponded behind the MRGO levee until the gates of the
two control structures can be opened to allow the stored water to drain into Lake Borgne. It is
interesting to note that water discharged from the Lower 9th Ward must drain through
approximately 10.5 miles of bayou (Figure 10.55) until it is finally discharged from the protected
area through the control structure gates, which are not configured to allow discharge of water
during storm events, instead of having the pump station discharge directly into the much closer
IHNC about 650 feet from the pump station.
Mitigation of these systemic flaws and performance inhibitors will require significant
effort, and should prompt a full reevaluation of the pumping and unwatering system
configuration and details in order to develop an improved system that can function reliably both
during and after major storm events.
10.10 Conclusions
Hurricane Katrina resulted in the catastrophic flooding of the greater New Orleans area.
Although the magnitude of the storm surge that overwhelmed the levee flood defense system was
greater than the capacity of the system, the extent of the devastating damage could have been
greatly minimized if the system had been robustly designed. There were many miles of earthen
levees that were significantly overtopped, but did not breach catastrophically. These levees that
did not breach were only overtopped for a few hours’ duration and the quantity of water that did
flow over the levees could have been pumped out of the protected area utilizing the existing
drainage network and pump infrastructure (see Chapter 4). The levees that were not able to
withstand overtopping breached catastrophically, allowing the full magnitude of the storm surge
to overwhelm the protected area.
Design guidelines need to be updated to ensure the design and construction of robust
levee systems. All failure mechanisms must be acknowledged and included in the design
evaluation. All levees should be designed to withstand overtopping. Material selection and
compaction are critical components to ensure adequate performance and appropriate
10 31
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
specifications for material selection and compaction should be developed and should be
incorporated into the design guidelines.
The current design guidelines sponsored by both the USACE and FEMA assume that
overtopping does not occur and does not require safety in the event of overtopping of the levees.
A probabilistic approach should be utilized to determine the appropriate factor of safety
for the design of these levee systems. Accounting for uncertainties in demands on the system
(height of storm surges, wave impacts, etc.) as well as uncertainties in the capacity of the levee
system (erosion resistance, foundation stability, etc.) must be included in the safety evaluation of
the levee system.
The current design guidelines sponsored by both the USACE and FEMA are based on
deterministic factor of safety levels that do not account for a broad range of uncertainties nor do
they account for mechanisms to ensure an appropriate level of safety based on the consequences
of failure.
10.11 References
Christian, John T. (2004), “Geotechnical Engineering Reliability: How well do we know what
we are doing?” J. of Geotechnical and Geoenvironmental Engineering, Vol. 130, No. 10,
October 1, 2004.
Google Earth, (2005), earth.google.com, <accessed: March 21, 2006>
Hong Kong Government Planning Department. (1994). “Potentially hazardous installations.”
Hong Kong planning standards and guidelines, Chapter 11, Hong Kong.
Interagency Performance Evaluation Task Force, (2006), “Performance Evaluation, Status and
Interim Results, Report 2 of a Series, Performance Evaluation of the New Orleans and
Southeast Louisiana Hurricane Protection System,” March 10, 2006.
IPET, (2006), “Performance Evaluation of the New Orleans and Southwest Louisiana Hurricane
Protection System, Draft Final Report of the Interagency Performance Evaluation Task
Force, Volume VI – The Performance – Interior Drainage and Pumping,” available online:
https://ipet.wes.army.mil/, date accessed: June 1, 2006.
Munson, Bruce R., Donald F. Young, and Theodore H. Okiishi, (1990), “Fundamentals of Fluid
Dynamics – Third Edition,” John Wiley and Sons, Toronto, Canada, ISBN 0471170240.
Seed, Raymond B., et al., “Preliminary Report on the Performance of the New Orleans Levee
Systems in Hurricane Katrina on August 29, 2005,” Report No. UCB/CITRIS 05/01,
November 17, 2005.
Topozone.com, (2006), www.topozone.com, <accessed March 21, 2006>
United States Army Corps of Engineers, (2006), MRGO Levee Reconnaissance Handout,
“MRGO Center Profile Line,” April 8, 2006.
United States Army Corps of Engineers, (2002), EM 111021100 (Part III), “Coastal
Engineering Manual Part III,” April 30, 2002.
10 32
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
United States Army Corps of Engineers, (2002), EM 111021100 (Part VI), “Coastal
Engineering Manual Part VI,” February 28, 2005.
United States Army Corps of Engineers, (2003), EM 111021902, “Engineering and Design –
Slope Stability,” October 31, 2003.
United States Army Corps of Engineers, (1999), “RiskBased Analysis in Geotechnical
Engineering for Support of Planning Studies,” ETL 11102556, May 28, 1999.
United States Army Corps of Engineers, (1999), “Channel Rehabilitation: Processes, Design and
Implementation,” July 1999.
United States Army Corps of Engineers, (1990), EM 111021904, “Engineering and Design –
Settlement Analyses,” September 30, 1990.
United States Army Corps of Engineers, (2000), EM 111021913, “Engineering and Design –
Design and Construction of Levees,” April 30, 2000.
United States Army Corps of Engineers, (1989), EM 111022502, “Engineering and Design –
Retaining and Floodwalls,” September 29, 1989.
United States Army Corps of Engineers, (1988), “Mississippi RiverGulf Outlet, St. Bernard
Parish, La., Bank Erosion, Reconnaissance Report,” February 1988.
United States Army Corps of Engineers, (1983), “Proceedings, Seminar on Probabilistic
Methods in Geotechnical Engineering,” Miscellaneous Paper GL8326, September 1983.
United States Army Corps of Engineers, (1966), “Lake Pontchartrain, LA. and Vicinity,
Chalmette Area Plan, Design Memorandum No. 3, General Design,” November, 1966.
United States Code of Federal Regulations, 44CFR65.10, October 1, 2005.
Versteeg, M. (1987) “External Safety Policy in the Netherlands: An approach to risk
management,” J. Hazard. Mater., 17, 215221.
10 33
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Table 10.1 Summary of Major Levee Design Steps
Step
Procedure
1
Conduct geologic study based on a thorough review of available data including analysis of aerial
photographs. Initiate preliminary subsurface explorations.
2
Analyze preliminary exploration data and from this analysis establish preliminary soil profiles, borrow
locations, and embankment sections.
3
Initiate final exploration to provide:
a. Additional information on soil profiles
b. Undisturbed strengths on foundation materials
c. More detailed information on borrow areas and other required excavations
4
Using the information obtained in Step 3:
a. Determine both embankment and foundation soil parameters and refine preliminary sections where
needed, noting all possible problem areas.
b. Compute rough quantities of suitable material and refine borrow area locations.
5
Divide the entire levee into reaches of similar foundation conditions, embankment height, and fill
material and assign a typical trial section to each reach.
6
Analyze each trial section as needed for:
a. Underseepage and through seepage.
b. Slope stability.
c. Settlement.
d. Trafficability of the levee surface.
7
Design special treatment to preclude any problems as determined from Step 6. Determine surfacing
requirements for the levee based on its expected future use.
8
9
10
Based on the results of Step 7, establish final sections for each reach.
Compute final fill quantities needed; determine final borrow area locations.
Design embankment slope protection.
Table 10.2 Permissible canal velocities with average flow depth of 3 feet
Material
Fine sand (noncolloidal)
Sandy loam (noncolloidal)
Silt loam (noncolloidal)
Alluvial silt (noncolloidal)
Ordinary firm loam
Fine gravel
Stiff clay
Alluvial silt (colloidal)
Coarse gravel (noncolloidal)
Shales and hardpans
Clear water, no
detrius (ft/s)
Water transporting
colloidal silts (ft/s)
Equivalent Shear
Stress1 (lb/ft2)
1.5
1.75
2
2
2.5
2.5
3.75
3.75
4
6
2.5
2.5
3
3.5
3.5
5
5
5
6
6
2.4 6.3
3.1 6.3
4.0 9.0
4.0 12.3
6.3 12.3
6.3 25.0
14.1 25.0
14.1 25.0
16.0 36.0
36.0
1
Assuming a roughness constant equal to 1 and fluid consisting of seawater.
10 34
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Table 10.3: Performance summary for selected levee locations.
Design
Maximum
Water
Storm Surge
Location
Overtopping
Elevation
Elevation
(ft)1
(ft)2
[13.5]
1 Lakefront Airport
12
Minor
11.8
[14.5]
2 Jahncke Pump Station Outfall
12
Minor
12.8
Immediate PostHurricane
Levee Condition2
Poor
Adequate
3 Eastern Perimeter of New
Orleans East
[14.5]
12.4
~18?
Moderate to
Major
Very Good
4 Southeast Corner of New
Orleans East
[19.0]
13.0
~18?
Moderate to
Major
Poor
5 Entergy Michoud Generating
Plant
6 IWW/MRGO Southern Levee
7 Bayou Bienvenue Control
Structure
8 Mississippi River Gulf Outlet
9 Bayou Dupre Control
Structure
10 St. Bernard Parish Interior
Levee (Forty Arpent Levee)
[15.0]
13.2
[14.0]
13.2
[17.5]
13.2
[17.5]
12.7
(~103)
[17.5]
12.7
(~103)
[8]
~64 (~33)
16
16
18
Moderate to
Major
Moderate to
Major
Moderate to
Major
Good
Good
Good/Poor
1722
Major
Poor
1722
Major
Poor
Not
Established
Major
Very Good
1
Elevations converted from NGVD29 elevation (in brackets) to equivalent NAVD88(2004.65) elevation, from IPET (2006)
2
Based on NAVD88(2004.65) vertical datum. From IPET (2006)
3
Elevation at the time of Hurricane Katrina was below the design elevation
4
A conversion between the original design elevation from the NGVD29 to the new NAVD88(2004.65) elevation was not available from IPET
2
This is an assessment of conditions immediately after the hurricane, before significant repair and reconstruction.
10 35
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Table 10.4 Summary of sampling locations for laboratory testing
Site
Latitude Longitude
Description
(oW)
No.
(oN)
1
Levee east of Hwy 11 and North of Hwy 90 30.0895
89.8587
2
Entergy Powerplant
30.0065
89.9389
3
MRGO North Control Structure (North)
29.9996
89.9170
4
MRGO Levee (northern section)
Not Established
5
MRGO Levee (middle section)
Not Established
6
MRGO Levee (southern section)
Not Established
7
St. Bernard Parish South
29.8769
89.7818
8
St. Bernard Parish North
29.9558
89.9466
9
Lakefront Airport Transition Levee
30.03344
90.026
10
Hayne Blvd
30.05908 89.96697
11
12
13
14
15
16
17
Hayne Blvd and Paris Road (Beach)
Orleans East Southeast RR Transition
Orleans East Southeast Corner
Intracoastal Waterway North (New Levee)
Intracoastal Waterway North (Remaining
Levee)
Levee west of Entergy Plant
St. Bernard Parish (Middle)
Note: Geographical coordinates based on WGS84 datum.
10 36
30.07577
30.06156
30.04481
30.03542
89.94467
89.83352
89.83089
89.85399
Erosion
Performance
Good
Good
Good
Poor
Poor
Poor
Good
Good
Moderate
Good
Not
Applicable
Poor
Poor
Unknown
30.02707
30.00465
29.92541
89.87448
89.95062
89.8948
Poor
Good
Good
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
(Table 10.4, Continued) Summary of sampling locations for laboratory testing
Site
No.
Description
Performance
1
2
3
4
5
6
7
8
9
10
Levee east of Hwy 11 and North of Hwy 90
Entergy Powerplant
MRGO North Control Structure (North)
MRGO Levee (northern section)
MRGO Levee (middle section)
MRGO Levee (southern section)
St. Bernard Parish South
St. Bernard Parish North
Lakefront Airport Transition Levee
Hayne Blvd
Low (IV)
Low to Very Low (IVV)
Low to Very Low (IVV)
High (II)
High (II)
High (II)
Medium (III)
Medium (III)
Not Tested
Not Tested
11
Hayne Blvd and Paris Road (Beach)
12
13
14
Orleans East Southeast RR Transition
Orleans East Southeast Corner
Intracoastal Waterway North (New Levee)
Intracoastal Waterway North (Remaining
Levee)
Levee west of Entergy Plant
St. Bernard Parish (Middle)
Good
Good
Good
Poor
Poor
Poor
Good
Good
Moderate
Good
Not
Applicable
Poor
Poor
Unknown
Poor
Very High to High (III)
Good
Good
Not Tested
Not Tested
15
16
17
10 37
EFA Erosion Susceptibility
Determination
Not Applicable
High to Medium (IIIII)
Not Tested
Not Tested
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Table 10.5 Levee Failure Mechanisms
Failure Mechanism
USACE Guidelines
FEMA Guidelines
Overtopping
Jetting
Internal Erosion and Piping
Lateral Surface Erosion
Wave Impacts
Structural Impacts
Slope Failures
Sliding
Underseepage
Liquefaction
Bottom Heave/Blowout
Not allowed
Not allowed
Design criteria provided
Protection required
Protection required
Not addressed
Design criteria provided
Design criteria provided
Design criteria provided
Not directly addressed
Not directly addressed
Not allowed
Not allowed
Analyses required
Protection required
Protection required
Not addressed
Analyses required
Analyses required
Analyses required
Not directly addressed
Not directly addressed
Observed in
Greater New
Orleans Area
Yes
Yes
Yes
Possibly
Yes
Yes
Yes
Yes
Yes
Possibly
Yes
Table 10.6 Target Reliability Indices
Expected Performance
Level
High
Good
Above average
Below average
Poor
Unsatisfactory
Hazardous
Beta (e)
Probability of Unsatisfactory
Performance
Approximate Median
Factor of Safety1 (F.S.50)
5.0
4.0
3.0
2.5
2.0
1.5
1.0
0.0000003
0.00003
0.001
0.006
0.023
0.07
0.16
2.5
2.1
1.7
1.6
1.4
1.3
1.2
Table 10.7 Risk Levels for a System with the Potential for 1,000 Fatalities
Risk Level
Pfpa
Annual Return Period (yrs)
Acceptable
<108
>100,000,000
106 to 108
1,000,000 to 100,000,000
ALARP
Unacceptable
6
>10
< 1,000,000
10 38
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 10.1: Underseepage in highly permeable underlying foundation materials (red lines) can
result in the catastrophic failure of the levee in that once the foundation materials have been
eroded, the levee (which may be completely undamaged) has no underlying support and falls
into the resulting void and essentially washes away.
Internal erosion and piping (blue lines) occurs in levee materials that have high permeabilities
(such as sand and gravel) and allow for water to rapidly flow from high pressure areas to low
pressure areas. As the water flows through the levee, smaller/finer soil particles are “washed”
out of the levee resulting in the internal erosion of the levee. Enough internal erosion of the
levee can lead to the collapse and subsequent “washout” of the levee.
Figure 10.2: Overtopping occurs when the water level on the outboard side of the levee exceeds
the crest elevation of the levee. The inboard side of the levee acts as a spillway for the
overtopping water and damage is inflicted on the levee as a result of water scour.
10 39
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 10.3: Jetting occurs when the water level on the outboard side of the levee exceeds the
top of wall elevation for structural walls that are founded within the earthen levee. Unlike
overtopping of a conventional earthen levee, the floodwall acts as a weir and water impacts the
levee in a concentrated stream that is much more energy intensive than conventional
overtopping.
Figure 10.4: Surface erosion generally occurs on the outboard side of the levee and is the result
of water flowing past the levee face. If the imposed shear stress from the water abrading against
the soil levee face is high enough, soil scour occurs and the integrity of the overall levee is
significantly reduced.
Figure 10.5: Wave impacts can cause significant erosion to levee faces. Waveinduced erosion
consists of runup (sloshing up and down of water as a result of staggered wave arrival) and
“minijetting” when the crest of the waves breaks on the levee face.
10 40
Independent Levee
Investigation Team
10 41
Figure 10.6: Map showing the extents of the visual reconnaissance (dashed black line) of the earthen levee systems performed
between October 2005 and March 2006. Locations of notable performance are identified in the numbered boxes.
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Source: Modified after USACE, 2005
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Area of Detail
Lakeview
airport
Location of
Figure 10.8
Active
railroad line
Direction of storm
surge overtopping
flow through
“transition area”
Scour of earthen
levee shown in
Figure 10.9
Image from Google Earth, 2006
Figure 10.7: Stormsurge induced overtopping traveled through the granular gravel ballast for
the railroad line and eroded the railroad line embankment, which served as a transition levee
between the concrete floodwall and the earthen levee shown in Figure 10.8.
10 42
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph by Rune Storesund
Figure 10.8: Lakefront levee near the Lakefront Airport at location 1 (as indicated on Figure
10.7) where overtopping occurred and significant scour around the floodwall was observed.
Scour of
earthen levee
Location of
Figure 10.8
Direction of Storm Surge
Overtopping Flow
Photograph by Rune Storesund
Figure 10.9: Significant erosion was observed on the levee behind the floodwall shown in Figure
10.8. The storm surge overtopped the floodwall and railroad ballast and failed the earthen levee.
10 43
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph by Rune Storesund
Figure 10.10: Lakefront levee at location 2 (as indicated on Figure 10.6) where minor
overtopping occurred. These levees performed well and only minor, surficial damage was
observed.
Photograph by Rune Storesund
Figure 10.11: Observed scour at the Jahncke Pump Station outfall structure (location 2 as
indicated on Figure 10.6). Scour was limited to areas of soilstructure interfaces, and no full
breach occurred.
10 44
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph by Rune Storesund
Figure 10.12: Condition of levees east of HWY 11 (location 3 on Figure 10.6) in October 2005.
These levees performed exceptionally well and were not eroded during Hurricanes Katrina or
Rita.
Photograph by Rune Storesund
Figure 10.13: Levee rehabilitation work (near location 3 on Figure 10.6) after Hurricane Katrina
included reinforcement and protection of soilstructure interactions with rockgabion transition
zones.
10 45
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Zone of levee eroded during
Hurricane Katrina.
Photograph by Rune Storesund
Figure 10.14: Zones of earthen levees at the southeast corner of the New Orleans East polder
(location 4 on Figure 10.6) were washed out during Hurricane Katrina allowing water to rush in.
Photograph by Rune Storesund
Figure 10.15: Levee rehabilitation work (location 4 on Figure 10.6) post Hurricane Katrina. The
“semicompaction” construction approach was employed, utilizing earth moving equipment to
also compact the placed material as it transported borrow materials to the construction site.
10 46
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph courtesy the Entergy Corporation
Figure 10.16: Photographs (location 5 on Figure 10.6) captured by a security camera at the
Entergy Michoud Generating Plant beneath Route 47, on the IWW/MRGO show active
overtopping during Hurricane Katrina.
Photograph by Rune Storesund
Figure 10.17: Post Hurricane Katrina view of the earthen levee shown in Figure 10.16. Only
minor damage on the protected side was observed, with the majority of the damage a result of
wave reflection from the bridge abutment at the righthand side of the picture.
10 47
Independent Levee
Investigation Team
10 48
Figure 10.18: The presence of a concrete floodwall beneath the Hwy 47/Parish Road Bridge acted as a weir during the overtopping
stages of the storm and “sucked” in nearby barges. Scour can be observed behind the concrete floodwall, but there is only minor
scour damage visible at the earthen levee/concrete floodwall transition. Overall, this system performed well (also considering the
impact associated with the barges).
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Photograph courtesy ngs.woc.noaa.gov/Katrina/KATRINA0000.HTM
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph courtesy Lee Wooten
Figure 10.19: Woody debris and steel barges “washed up” on the southern ICWW/MRGO levee
just west of Route 47 (location 6 on Figure 10.6). This is a side view from the aerial photograph
presented in Figure 9.24.
Photograph courtesy Francisco SilvaTulla
Figure 10.20: A gas processing barge “washed up” on the southern IWW/MRGO levee just east
of Route 47 (location 6 on Figure 10.6). This levee was also overtopped and was not
significantly damaged by the barge impact.
10 49
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Flood control gate
shown in Figure 10.22
Photograph courtesy Les Harder
Figure 10.21: Aerial photograph of the Bayou Bienvenue Control Structure (location 7 on
Figure 10.6). The northwestern half of the control structure levee system performed extremely
well (withstanding significant impact loads from the steel barge), while the southeastern portion
suffered severe erosion.
Photograph by Rune Storesund
Figure 10.22: This flood control gate acted like a weir as water overtopped the structure during
the storm surge. Significant scour and erosion was observed around the flood control gate
structure.
10 50
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph by Rune Storesund
Figure10.23: By March of 2006, after Hurricane Katrina, splash pads had been installed at this
flood control gate to mitigate erosion impacts if any future overtopping of the flood protection
system occurs. This photograph was taken at the same location as Figure 10.22.
Photograph by Rune Storesund
Figure 10.24: The northern side of the Bayou Bienvenue control structure has beenrepaired, and
then heavily reinforced with new riprap transported to New Orleans from Kentucky. This
photograph was taken about 6 months after Hurricane Katrina and the barge, which can be seen
in Figure 10.21, has been removed.
10 51
Independent Levee
Investigation Team
10 52
Scour pattern behind around the floodgate structure at the Bayou Bienvenue Control Structure shown in Figure
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Figure 10.25:
10.22.
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Zone of excellent MRGO
levee performance
Zone of poor MRGO levee
performance due to high
erodability of construction
(fill) materials
Bayou Bienvenue
Control Structure
“transition”
Topographic map from Topozone.com
Figure 10.26: A U.S.G.S. topographic map showing the presence of “spoils” along the MRGO.
As seen in Figure 10.28, the levee materials to the southeast of the Bayou Bienvenue Control
Structure performed poorly.
10 53
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph courtesy ngs.woc.noaa.gov/Katrina/KATRINA0000.HTM
Figure 10.27: Aerial photograph taken by NOAA in early September 2005 along the MRGO
showing severe erosion/breaches of the earthen levee and transport and deposition of large
barges over the levee as a result of the storm surge (location 8 on Figure 10.6).
Photograph courtesy Les Harder
Figure 10.28: Close up aerial photograph of severely eroded MRGO levees at location 8 on
Figure 10.6.
10 54
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph courtesy Les Harder
Figure 10.29: Control structure at Bayou Dupre (location 9 on Figure 10.6) suffered extensive
scour and erosion during the storm surge and overtopping conditions associated with Hurricane
Katrina.
Photograph courtesy Les Harder
Figure 10.30: Another aerial photograph of the scour and erosion damage to the Bayou Dupre
control structure.
10 55
Independent Levee
Investigation Team
10 56
Scour Zones adjacent to the
control structures
Figure 10.31: Areas of severe scour and erosion damage to the two control structures appear to correspond to the alignment of the
abandoned original channel.
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Source: USACE (1966)
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph courtesy Robert Bea
Figure 10.32: Aerial photograph taken of the repair operations at Bayou Dupre in January 2006.
Photograph courtesy Robert Bea
Figure 10.33: Close up aerial photograph showing backfilling operations at Bayou Dupre.
10 57
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph courtesy Rune Storesund
Figure 10.34: Eastward looking view of the secondary earthen levee (constructed to
approximately Elevation +6.0 feet MSL) immediately to the north of the Corinne Canal, east of
Paris Road (location 10 on Figure 10.6). This levee was significantly overtopped and did not
experience significant damage.
Photograph courtesy Rune Storesund
Figure 10.35: A fishing boat was washed over the levee shown in Figure 10.34 and landed in
this Chalmette residential neighborhood within St. Bernard Parish.
10 58
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Photograph courtesy Rune Storesund
Figure 10.36: The same levee shown in Figure 10.34, where a fishing boat washed over the
levee during Hurricane Katrina, has now been raised from an elevation of approximately +6 feet
MSL to an elevation of approximately +10 feet MSL by March of 2006.
10 59
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Subsidence between
2000 and 2005.
Erosion as a result of
Hurricane Katrina.
Figure 10.37: This is a comparison of two LiDAR surveys of the MRGO levee system at the
Bayou Bienvenue control structure, near the intersection of the GIWW at the northeast corner of
St. Bernard Parish. Effects of subsidence can be clearly seen in the elevation differences on the
north side of the control structure and the control structure itself between 2000 and immediately
following Hurricane Katrina in 2005. The levee on the north side of the Bayou Bienvenue
control structure was largely undamaged. The levee on the south side of the control structure
was catastrophically damaged. The magnitude of the erosion has been highlighted and white
“splotches” of displaced levee materials can be seen in the aerial photograph (USACE 2006).
10 60
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Figure 10.38: Levee sample sites within the Orleans East Protected Area.
Independent Levee
Investigation Team
10 61
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Figure 10.39: Levee sample sites within the St. Bernard Parish Protected Area.
Independent Levee
Investigation Team
10 62
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 10.40: The EFA (Erosion Function Apparatus) as developed by Briaud et al. Soil
samples are advanced into a rectangular tube of flowing water, creating a shear stress on top of
the inserted soil sample. The velocity of the water is increased until the critical shear stress is
achieved, where the soil begins to actively erode.
10 63
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
100000
Very High
Erodibility
I
10000
1000
Erosion
Rate
(mm/hr)
High
Erodibility
II
Medium
Erodibility
III
Low
Erodibility
IV
100
10
Very Low
Erodibility
V
1
0.1
0.1
S1-B1-(0-2ft)-TW
S2-B1-(2-4ft)-SW
S3-B3-(0-1ft)-SW
S5-(0-0.5ft)-LT-SW
S7-B1-(2-4ft)-SW
S8-B1-(2-4ft)-L2-SW
S12-B1-(0-2ft)-TW
S15-CanalSide-(0-0.5ft)-HC-SW
1.0 Velocity (m/s) 10.0
S1-B1-(2-4ft)-SW
S3-B1-(2-4ft)-SW
S4-(0-0.5ft)-LC-SW
S6-(0-0.5ft)-LC-SW
S8-B1-(0-2ft)-TW
S11-(0-0.5ft)-LC-TW
S12-B1-(2-4ft)-SW
S15-Levee Crown-(0-0.5ft)-LT-SW
100.0
S2-B1-(0-2ft)-TW
S3-B2-(0-2ft)-SW
S4-(0-0.5ft)-HC-SW
S7-B1-(0-2ft)-TW
S8-B1-(2-4ft)-L1-SW
S11-(0-0.5ft)-HC-TW
S15-Canal Side-(0-0.5ft)-LC-SW
S15-Levee Crown-(0.5-1.0ft)-LT-SW
Figure 10.41: Summary of all EFA test results. The test result designations are based on the
Site Number, the boring number at the site, the depth interval over which the sample was
collected and sample notes. The EFA test results indicate that the materials used in levee
construction varied from very high to very low erodibility, which matched the observed
performance of these levees.
10 64
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Graphic after FHWA 1988
Figure 10.42: The effects of compaction are clearly evident in this figure from this FHWA
design guideline. For the same material (with a plasticity index of 20) a tenfold increase in
shear stress capacity can be achieved by properly compacting the material.
10 65
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
100000
10000
Very High
Erodibility
I
1000
Erosion
Rate
(mm/hr)
100
High
Erodibility Medium
II
Erodibility
III
Low Compaction
Low
Effort
Erodibility
IV
Very Low
Erodibility
V
10
1
High Compaction Effort
0.1
0.1
1.0
10.0
Velocity (m/s)
100.0
S-4-(0-0.5)-LC-SW
S-4-(0-0.5ft)-HC-SW
Figure 10.43: Here materials sampled from the MRGO were tested at two levels of compaction.
This figure shows the dramatic impact proper compaction has on the erodibility of soils.
Materials sampled from the MRGO levee were tested at two compaction levels: low compactive
effort and high compaction effort. The corresponding results speak volumes to the importance of
compaction in earthen levees. The lowcompaction sample was found to be very highly erodible,
whereas the highcompaction sample exhibited low erodibility characteristics.
10 66
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
100000
Very High
Erodibility
I
10000
1000
Erosion
Rate
(mm/hr)
High
Erodibility
II
PRONE TO
FAILURE BY
OVERTOPPING
100
10
Medium
Erodibility
III
Low
Erodibility
IV
TRANSITION
ZONE
PRONE TO
Very Low
RESIST
OVERTOPPING Erodibility
V
1
0.1
0.1
1.0
10.0
100.0
Velocity (m/s)
Figure 10.44: Resulting guideline table for evaluating erosion susceptibility of soils used for
levee construction developed by Dr. Briaud from Texas A & M University.
10 67
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 10.45: Plot of backcalculated annual probability of failure vs. lives lost and $ lost
(note that either the “lives lost” or “$ lost” axes are used, they are not
intended to be used in conjunction). From Christian, 2004.
10 68
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 10.46: FN diagrams adopted by the Hong Kong Planning Department (left) and FN
diagram as proposed for planning and design use in the Netherlands (right). From Christian,
2004.
10 69
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
New Orleans
U.S. Dams
Figure 10.47: Figure 10.45 repeated, with approximate level of reliability and consequences
for the New Orleans regional flood protection systems indicated, and with
current U.S. practice for dams highlighted (red dashed line.)
10 70
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 10.48: Overview of the drainage and pumping parishes studied as part of the IPET (2006)
performance evaluation. The primary study areas were Jefferson, Orleans,
Plaquemines, and St. Bernard parishes. [IPET; June 1, 2006]
Figure 10.49: Detailed map of pump stations within Jefferson parish.
[IPET; June 1, 2006]
10 71
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
.
Figure 10.50: Detailed map of pump stations within Orleans parish. [IPET; June 1,2006]
Figure 10.51: Detailed map of pump stations within Plaquemines parish. [IPET; June 1, 2006]
10 72
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 10.52: Detailed map of pump stations within St. Bernard parish. [IPET, June 1, 2006]
Source: IPET, 2006
Figure 10.53: Performance of the pumping system (and causes of pumping capacity loss)
during and after hurricane Katrina.
10 73
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Reverse Flow
Protected
Side
Water flows back through the
pumps when the floodside water levels
exceed the pump outfall.
Figure 10.54: Conditions where the water level on the outboard side of the pumping station is
higher than the pump infrastructure, resulting in reverse flow from the flood
side back into the “dry” side. [IPET; June 1, 2006]
Figure 10.55: Breakdown of the pumping and drainage system as a result of structural failures
within the drainage canals. Flood waters and discharge water from the
pumping station flow back into the protected area where structural failures
have occurred.
10 74
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 10.56(a): The current drainage configuration for the lower 9th Ward in St. Bernard
parish starts at the Florida Street pump stations and flows approximately 10.5
miles through Bayou Bienveue, then through a control structure to the
MRGO.
Figure 10.56(b): The alternative; pumping into the IHNC 650 feet to the west.
10 75
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
CHAPTER ELEVEN: SUMMARY OF ENGINEERING LESSONS
11.1
Introduction
This chapter presents a summary overview of the principal technical lessons and
findings from this investigation. The next three chapters that follow then carry forward a
study of the underlying organizational, institutional, political, economic, human factors and
decisionmaking issues that arise in conjunction with these “engineering” lessons and
findings.
11.2
Overarching Strategic Issues
11.2.1 Targeted Levels of Safety and Reliability
Figure 11.1 shows a “risk plot” with the vertical axis representing the annual
likelihood of failure, and the horizontal axis representing the expected cost of such failure
either in dollars (bottom axis) or in lives lost (top axis). This figure shows the ranges of risk,
or reliability, representing common practice for a number of areas of human endeavor.
Highlighted with a heavy red dashed line near the bottom is current U. S. practice in the field
of dam engineering.
Also shown on this plot is our investigation’s view of the level of reliability associated
with the New Orleans regional flood protection systems prior to hurricane Katrina. Our best
estimate, based on information currently available, and given the targeted design levels and
the flaws and vulnerabilities embedded in the system, is that the preKatrina system was likely
to fail catastrophically approximately every 30 to 75 years. The cost of the failure (in
hurricane Katrina) was on the order of $100 to $200 billion in losses, and approximately
1,500 lives were lost.
There is a stark contrast between the levels of reliability for which major U.S. dams
are engineered, and the level of reliability of the New Orleans levee systems. This is true, to
only slightly varying degree, for most levee and flood protection systems across the entire
nation.
“Dams” are engineered to very high levels of reliability because their potential failure
would threaten large numbers of lives, and large economic losses as well. Few dams protect
(or threaten) populations as large as the combined greater New Orleans and adjoining
Jefferson parish region, however, and simple logic would suggest that flood protection
systems defending large populations like this should be targeted at similar levels of safety or
reliability.
As indicated by the large arrow in Figure 11.1, the difference between the level of risk
(or reliability) of the New Orleans regional flood protection systems and conventional U.S.
dam practice is approximately three orders of magnitude; a factor of roughly 1,000 times safer
and more reliable. Reliability is a function of two subelements: (1) the targeted level of
11 1
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
loading(s) to be handled, and (2) the reliability with which that target is met. Achieving
significantly higher levels of reliability for complex regional flood protection systems like
that of the New Orleans region would require major improvements in both subelements.
The New Orleans regional flood protection systems were never specifically targeted at
any given level of reliability with regard to formal definition of storm levels for design (e.g.:
the Standard Project Hurricane was never quantified as a “100year storm” or a “300year
storm”, etc.), and this was a lapse, as it put the design of the regional flood protection system
out of step with current practice. Our assessment in plotting the preKatrina New Orleans
case in Figure 11.1 is that the system as it existed was likely to be failed by roughly the 30
year to the 75year (average recurrence interval) storm.
The actual design intent was for more, though how much more was never formally
defined.
In addition, the system did not perform as intended; multiple failures occurred at
levels of storm surge and wave loading that were less than or equal to what many of the failed
system elements had been intended to safely handle.
If the system is reengineered to safely (successfully) handle a 100year storm (storm
loading likely to be exceeded typically once every 100 years), then the large red area of
Figure 11.1 would move to the location shown by the light blue area in Figure 11.2. That
would not bring the level of safety and reliability anywhere near to current U.S. practice for
dams. Targeting a 1,000year level of flood protection, and achieving that level, would result
in the darker blue zone in Figure 11.2.
There are some significant challenges involved at the decisionmaking and policy
levels regarding appropriate levels of safety (e.g. storm levels) for which such systems should
be designed, and the degree to which resources should be committed to achieve this.
The other element of risk/reliability is the degree to which targeted design levels are
successfully achieved. Levels of success were not good in hurricane Katrina; numerous
failures occurred at storm surge and wave levels less than or equal to those for which the
flood protection system elements were intended to be designed. These failures occurred
because margins for error (e.g.: design Factors of Safety) were inappropriately small, because
decisions were made to reduce costs in exchange for increased levels of risk, and because of
errors and lapses in design, construction and maintenance. These types of “engineering”
issues are discussed in Section 11.3 that follows.
11.2.2 Funding and Appropriations
A second set of overarching considerations are those associated with the Byzantine
process by which large, complex, regionalscale flood protection systems are conceived,
approved, designed, funded, constructed, maintained and operated. No useful discussion of
engineering challenges can proceed without noting the tremendous additional difficulties that
arise due to these types of “nonengineering” issues.
11 2
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 11.3 presents a very simplified schematic illustration of the principal
“technical” steps involved in creating and operating a regional flood protection system (left
side of the figure, in blue), and also the corollary political, organizational, institutional,
engineering and construction units and entities (right side of the figure, in green) that must
interact to foment this process. This is discussed in detail in Chapters 12 through 14, so we
will simply note here that it is not realistic to assume that we can achieve the significant
improvement of the safety and reliability of the New Orleans regional flood protection system
that appears warranted simply by making adjustments to the “technical” side of this figure.
Simply revising design manuals and engineering procedures, etc., cannot possibly achieve the
significant improvement in system reliability that should be sought; significant improvements
on the righthand side of this figure will be needed as well.
These issues are addressed in Chapters 12 through 14, but several key issues warrant
special mention at this stage. The first of these is funding and appropriations; the allocation
of resources to the creation and operation of the regional flood protection system. This
allocation of resources is, properly, the domain of the decisionmaking bodies involved;
elected representatives (government) at both the federal and local levels. Unfortunately, these
elected officials often lose track of the ramifications of their decisions with regard to complex
technical systems created and operated over long periods of time.
It should not take 50 years to construct a critical system providing lifesafety for a
region with a population of nearly one million people (the greater New Orleans/Jefferson
parish region.) The regional flood protection system was incomplete at the time of Katrina’s
arrival; it was intended to be complete by the year 2015, fully 50 years after its inception in
response to the catastrophic flooding of New Orleans produced by hurricane Betsy in 1965.
We need to do better.
Apart from the obvious need to more rapidly and effectively provide protection for
large numbers of citizens, these types of extended construction periods (covering multiple
decades) wreak havoc with the actual engineering and construction of the intended systems
themselves. As noted in the IPET Draft Final Report (IPET; June 1, 2006), the New Orleans
regional flood protection system was largely a system in name only. Having been constructed
over fourplus decades, and in innumerable individual segments and sections, it was
optimistic to expect that the various interconnecting elements would function perfectly well
together. Stretching the construction over multiple decades posed major challenges with
regard to progressive loss of institutional memory and expertise, and it required excessive
segmentation of systems that needed to function literally seamlessly as contiguous defenses.
Another difficult issue was the nearly constant pressure to reduce costs. Decisions
that produced reductions in the costs of specific flood protection elements routinely resulted
in corollary increases in levels of risk; the increased likelihood that the system would not
perform well when eventually tested. As discussed in Section 11.3, this type of tradeoff
between shortterm cost reductions and increased risk now appears very hard to justify, as it
contributed significantly to many of the specific failures and breaches during hurricane
Katrina, and resulted in catastrophic losses that now dwarf the shortterm savings by two
orders of magnitude and more.
11 3
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
11.3
Principal Engineering Findings and Lessons
11.3.1 Introduction and Overview
Figure 11.4 shows an overview of the New Orleans area, indicating the locations of
the principal failures, breaches, and distressed sections of the New Orleans regional flood
protection system studied in this investigation. Plaquemines parish (along the lower reach of
the Mississippi River) is not included in this figure; instead it is indicated by the large arrow
at the bottom. The individual features, and groups of features, in Figure 11.4 are numbered
for purposes of discussion. Table 11.1 presents a summary of issues at each of these
locations, using the same numbering scheme as Figure 11.4.
The New Orleans regional flood protection system failed massively and
catastrophically during hurricane Katrina. Depending on how one counts individual breaches
(or groups breaches extending along long frontages that were massively eroded and scoured),
the number of failed sections was somewhere between three dozen to 50plus.
For an overview of performance, the system can be roughly subdivided into four
zones.
At the southern end, the flood protection systems in Plaquemines Parish were
massively overwhelmed by storm surges and waves significantly more severe than they had
been designed to handle.
At the east flank (fronting Lake Borgne), and in the central region (along the IHNC
and the GIWW/MRGO channels) the storm loadings were approximately equal to those for
which the system was intended to be designed. Design loading conditions were exceeded at
some locations, especially along the Lake Borgne frontages, but intended design levels were
only slightly exceed and the system might have been expected to perform better. Instead,
massive and catastrophic breaches occurred at multiple locations. These were principally the
result of one or more of the following: (1) insufficient crest heights (which led to, and
exacerbated, overtopping problems), (2) use of inappropriate materials at some locations
(materials with very poor resistance to erosion), and (3) other engineering lapses and
oversights.
At the north end, along the Lake Pontchartrain frontage, storm surge levels and waves
presented lesser levels of loading than the system elements were intended to safely handle.
System performance was good along most of the lake frontage itself, but three catastrophic
breaches occurred along the drainage canals at the north end of the main (downtown) New
Orleans protected basin, and these were the principal source of approximately 85% of the
floodwaters that catastrophically inundated most of that basin. These three failures, together,
accounted for nearly half of the overall loss of life in this event, and a similar fraction of the
overall economic losses and property damages. These three major failures on the drainage
canals were the result of engineering failures in design.
Multiple issues and challenges contributed to the numerous individual failures, and
these have been discussed in Chapters 3 through 10. The discussion that follows will select
highlights with regard to lessons that can be extracted from this event with an eye towards
effecting better system performance in the future.
11 4
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
11.3.2
Plaquemines Parish
Plaquemines parish is a narrow, highly exposed, and sparsely populated set of
corridors along the edges of the lower reaches of the Mississippi River, extending south from
New Orleans to the river’s outlet in the Gulf. This protected strip, with “river” levees
fronting the Mississippi River and a second, parallel set of “storm” levees facing away from
the river forming a protected corridor less than a mile wide, serves to protect a number of
small communities as well as utilities and pipelines. This protected corridor also provides
protected access for workers and supplies servicing the large offshore oil fields out in the Gulf
of Mexico.
The flood protection systems of lower Plaquemines parish were massively overtopped
and overwhelmed by storm surge and storm waves that significantly exceeded design levels,
and multiple breaches and failures resulted (see Chapter 5.)
Plaquemines parish is sparsely populated, with a preKatrina population of only about
27,000 people. Given increased public awareness of risk and exposure, less than half of these
are expected to return. There are few engineering lessons to be learned from the experience
of Plaquemines parish; when even wellconstructed levees and floodwalls are sufficiently
massively overwhelmed, failures will occur. The lesson, if anything, is one of humility in the
face of nature. If there is an engineering lesson here, it is:
1. Not all areas can be protected, and it is not economically reasonable to commit out of
scale resources to the protection of some areas. We must learn to choose our battles.
Federal policy across the nation over the past two decades has been moving
increasingly away from out of scale expenditures to protect, or to insure, small populations
living on marginal lands at high risk with respect to flooding. Plaquemines parish will
represent an interesting case in this regard.
11.3.3 The East Flank; New Orleans East and the St. Bernard/Lower Ninth Ward
Protected Areas
Major cities are different. A de facto decision has already been made to reconstruct
New Orleans, and to upgrade its regional flood protection systems. Accordingly, it is now
incumbent upon us to do all that we can to extract important engineering lessons from the
Katrina experience, and to see that these lessons are suitably applied to efforts to improve the
levels of safety and reliability of the regional flood protection systems in future events.
The main breaches that were the principal source of flooding for both the St.
Bernard/Lower Ninth Ward protected area and the New Orleans East protected area were the
levee frontages facing “Lake” Borgne (which is actually a bay, as it is connected directly to
the Gulf of Mexico.) These are sections #2 (and 2a) and #3 in Figure 11.4, and in Table 11.1.
These two sections shared a number of fatal characteristics. Both sections were
constructed largely using materials dredged from the excavations for the adjacent shipping
11 5
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
channels (the MRGO channel and the GIWW channel, respectively), and as a result both
levee frontages including large sections of levees comprised in large part of materials known
to perform very poorly with regard to erosion. These unacceptably erodeable materials
included sands and lightweight shellsands, and the massive and catastrophic erosion of these
materials caused the rapid failure of great lengths of levees along both the MRGO and GIWW
frontages. Another commonality was the lack of swamps or cypress groves on the outboard
sides of these levee frontages; features that would have served to dampen (reduce the energy
and intensity) of stormdriven waves attacking these frontages. Finally, these two frontages
were also, unfortunately, the two frontages that were most directly exposed to severe wind
driven waves across a large body of open water.
As a result, these two frontages failed catastrophically, and were massively eroded
along multiple miles of frontage, creating long breaches through which the hurricane storm
surge passed easily, and with devastating consequences for the communities of these two
protected areas.
Interestingly, adjacent levee sections along these same frontages, although also
overtopped, performed well; suffering relatively minor erosion and continuing to provide
protection as the storm surge subsided after the period of overtopping during the relatively
shortlived peak of the storm surge. These betterperforming sections were levees comprised
of compacted, clayey soils; soils known to have far higher intrinsic resistance to erosion (see
Chapters 9 and 10.)
We know a great deal about the soil types, and the placement and compaction
conditions, that lead to differing types of performance with regard to erosion. Moreover, we
are now increasingly able to perform specific tests of these materials, and to make reliable
engineering assessments of expected behavior with regard to erosion (see Chapters 9 and 10.)
Important lessons here include the following:
2. The use of materials excavated from the adjacent shipping channels resulted in some
initial cost-savings, but these minor cost-savings were multiple orders of magnitude less
than the subsequent damages that occurred when these levee sections failed. Short-term
cost savings in construction need to be balanced against the consequent increases in
risk (the consequent reduction in likely reliability) for the resulting built system.
3. Highly erodeable embankment (and foundation) materials represent an intrinsic hazard,
and their use should be avoided in flood protection systems defending significant
populations.
4. When the use of such materials cannot be avoided, then great care should be taken to
protect the sections by means of internal cut-offs, filters, and slope face protection
(armoring) on the front and back faces and on the crest as well. Even then, the use of
erosion-resistant soils is to be preferred if at all possible.
5. Levees (and composite levee/floodwall sections) can be designed to safely withstand some
degree of overtopping, and for some period of time. Hurricane storm surges, unlike river
floods, typically have their “peak” over only a limited number of hours. Given the
economic challenges of designing flood protection systems for very high levels of
11 6
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
(infrequently occurring) storm loading; the alternative of designing flood protection
systems to perform safely without admitting any water into the protected areas for an
fairly high levels of loading, but with sufficient resilience that they can be “overtopped
safely” (overtopped for a while during a peak storm surge, but not erode and fail
catastrophically) so that the system will continue to provide protection as the peak of an
unusually large storm surge passes and then subsides might also be considered. Some
water would enter the protected area(s), but the amount would be limited and it could be
pumped out afterwards with minimal risk to life, and manageable property damages.
At some locations (e.g. adjacent to large concrete navigation gate structures) the use of
the lightweight shellsands was deliberately specified in order to minimize differential
settlements and “gapping” at the contact between the concrete gate structures and the levee
embankment. Prevention of this differential settlement by use of these highly erodeable
materials led to massive eroded breaches at the south side of the bayou Bienvenue gate
structure, and at the north side of the bayou Bienville structure along the MRGO frontage; an
example of solving one problem while exacerbating another. Thus
6. The consequences of any engineering decisions need to be considered on a system basis;
there is a long history of engineering failures based on unintended consequences.
Another disturbing issue was the fact that many sections of the regional flood
protection system had levee crest elevations, and concrete floodwall elevations, that were
below intended design grade. This was largely a result of the 40plus years that the design
and construction of the system had been underway, and the fact that benchmarks and datums
for elevation control had progressively decreased in elevation as part of largescale overall
regional subsidence over that extended period of time. As a result, many sections of the
regional system had crest heights and floodwall heights as much as 1 to 2 feet below intended
design grade. [An excellent treatment of this issue with regard to datums and regional
subsidence is presented in the IPET Draft Final Report; IPET, June 1, 2006.] This “loss” of
levee crest and floodwall height exacerbated problems associated with overtopping.
In addition, the critical levees along the MRGO frontage at the northeast edge of the
St. Bernard Parish/Lower Ninth Ward protected area were well below grade along much of
their length. This was not an engineering lapse, nor a problem associated with subsiding
datums. These levees were being constructed in stages, to allow for settlements and
consolidation (to increase the strengths of the foundation soils prior to adding the next stage
of levee embankment fill.) At the time of Katrina’s arrival, the USACE had long been
requesting funds to place the final stage of fill along this frontage. Now it is too late.
It can be argued that this represents a tragic example of the intrinsic risk associated
with overlong project durations for these types of massive, regional scale projects. Also that
both White House and Congressional attention lapsed, and funds that could have been
provided to complete these important levees were instead deferred, as issues elsewhere drew
more urgent attention. Thus
7. If we resolve to create and operate important flood protection systems to defend large
populations, then we should commit sufficient funds and diligence to consummate this
11 7
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
construction within a reasonable time span. Otherwise: (1) we are leaving populations at
risk, as partially completed protection is no protection at all, (2) we invite problems that
will naturally arise as a result of over-segmentation of discrete project elements that must
perform perfectly well together as a contiguous system, and (3) overall system
coordination and integrity will inevitably suffer as a result of progressive loss of
institutional memory during the extended period of design and construction.
Finally, both of these levee frontages were “backed” by lower height secondary levees
that were rapidly overtopped by the massive flows passing through the long breaches in the
frontage levees. Indeed, the frontage levees failed so rapidly, and so early, that they did little
to blunt the storm surge. The secondary levees were never intended to have to deal with an
undiminished storm surge, and they were quickly overtopped along much of their lengths
(though they were comprised of better, clayey materials and suffered admirably little damage
at most locations from this massive overtopping.)
The lack of height of these internal “secondary” levees represented a wasted
opportunity to provide defense in depth.
8. As the regional system is now repaired and improved, further raising of these secondary
levees would provide a potentially valuable second line of defense for the populous
communities behind them. If the federal government will not fund this, then local
interests should consider doing this on their own.
These secondary levees are wellsituated, and are currently comprised of good,
erosionresistant materials. The USACE has recently helped to raise the secondary levee
across the middle of St. Bernard parish (the Forty Arpent levee) to Elev. +10 feet (MSL).
These secondary levees should be considered part of a system along with the frontage levees,
and with the ridges of high ground within their protected areas as well. The frontage levees
should be designed to safely resist a considerable level of storm surge. In the unlikely event
that an even greater storm surge exceeds this level, then they should be designed to overtop
without eroding catastrophically, and the secondary levees should be designed and sized to
entrap and water overtopping the frontage levees and so protect the populous communities.
That would require coordination of levee heights, and likely of oversight agencies as well.
This, in turn, represents an example of an element of engineering that was sadly
missing throughout much of the regional flood protection system; the asking of the vital
question: “What if?”
There was a persistent lack of ductility and resilience throughout the regional flood
protection system. Over and over, system elements and sections were designed to some
specified level of loading, but no thought was given as to what would happen if that level was
exceeded. Over and over, we studied sections that failed catastrophically, in a brittle manner,
when for little or no increase in cost the section designs could have been modified to safely
accommodate some minor exceedance of the design loading conditions, and where similarly
relatively minor modifications could have rendered even more severe exceedances of the
specified design loads at least far less devastating in terms of consequences for the protected
communities. To a large degree, this failure to ask the important “what if?” question is a
11 8
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
function of the rules and regulations that govern the creation of these large systems. This
needs to change.
9. Instead of working to pedantically prescribed “design levels” mandated by Congress, the
USACE should be allowed (and encouraged) to constantly ask the important “What if?”
question. When that leads to the awareness that minor additional effort and expense
would likely result in massive improvement in overall system performance and reliability,
then a feed-back mechanism should be established to allow advantage to be taken of this.
That would be an invaluable step forward, in many areas of federal operations.
11.3.4 The Central Region; the IHNC and the GIWW/MRGO Channel Frontages
The storm surge that swelled the waters of Lake Borgne was driven west along the
eastwest trending shared GIWW/MRGO channel to the “T” intersection with the IHNC,
raising water levels within these channels and resulting in overtopping at many locations
along both banks of both of these channels. Despite this overtopping, the performance of
many of the levees and floodwalls along these channels was excellent. Several major failures
occurred, but most of these were not caused by overtopping, but were instead the result of
other issues as described below.
Along the eastwest trending GIWW/MRGO channel, overtopping produced minor to
moderate erosional distress at a number of locations, but no full failures (breaches) of full
height earthen levee embankments occurred. The levees along both banks of this channel
appeared to be comprised primarily of compacted, clayey soils, and the good performance of
these materials in the face of moderate overtopping was encouraging.
Two breaches occurred along the north bank of this channel, at Sites #4a and #4b in
Figure 11.4 (and Table 11.1.) The first of these (Site #4b) was a breach at an inadequate
“transition” between a long reach of full height earthen levee as it joined (abutted) a midrise
earthen levee reach with a sheetpilesupported concrete floodwall at its crest. The transition
between these two adjacent project sections was a simple sheetpile wall, of lesser height than
either the earthen levee to the west, or the floodwall section to the east. As a result,
overtopping was most severe over the top of the short sheetpile wall transition section, and
this overtopping preferentially eroded a deeply scoured trench at the rear side of the sheetpile
wall. This, in turn, reduced the lateral support at the back side of the sheetpile wall section,
and the laterally unbraced sheetpile wall was then pushed sideways by the storm surge water
pressures on its front (outboard) side, and it failed.
This was one of many examples of inadequate detailing of “transitions” between
adjacent major project elements that resulted in poor performance, and breaches at a number
of locations. Thus
10. “Transitions” where adjacent project elements join together were routinely problematic
throughout much of the regional flood protection system. Successful design and
construction of two adjoining project sections counts for little if the connection between
11 9
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
them is not also successfully consummated. Significantly more attention needs to be paid
to these “transitions”.
The second breach along this frontage (Site #4a) was a failure of a concrete floodwall.
This floodwall was mainly a simple sheetpilesupported Iwall, but it had two short sections
of Twall with battered piles to provide increased rotational and lateral support. The Twall
sections performed well, but major lengths of the Iwall did not. The walls were overtopped,
and the water cascaded over the walls and eroded trenches at the back sides of the walls.
This, in turn, laterally unbraced the walls and some sections were laterally displaced while
other sections rotated, some to a nearly fully horizontal position.
Like the situation described at Site #4b above, this overtopping did not have to result
in erosion and unbracing of the floodwall. Installation of splash pads, or other erosion
protection, at the rear side of the walls to prevent erosion by the water passing over the tops
of the walls would have represented a relatively minor additional expense (estimated at less
than 5% of the overall project section cost). The USACE felt that splash pads were
disallowed; they were instructed by Congressional edict to design for a specified water level,
and to install splash pads would be to provide for a higher level than that which had been
authorized. In hindsight, everyone regrets this dilemma. To its credit, the USACE has
already undertaken (even prior to full authorization) to install splash pads behind Iwalls
and/or to replace Iwalls with Twalls (which have their own splash pads as a result of their
invertedT shape.) Thus
(9. Repeated): Instead of working to pedantically prescribed “design levels” mandated by
Congress, the USACE should be allowed (and encouraged) to constantly ask the
important “What if?” question. When that leads to the awareness that minor additional
effort and expense would likely result massive improvement in overall system
performance and reliability, then a feed-back mechanism should be established to allow
advantage to be taken of this.
11. Concrete floodwalls can be designed to be safely overtopped to some considerable
degree, and advantage can be taken of this in design and construction of an improved
overall regional flood protection system.
Further to the west, the rise in water levels within the IHNC produced overtopping at
numerous locations, and it produced a number of failures as well. The overtopping was not
directly related, however, to the most significant of these failures.
Two of the largest failures during hurricane Katrina were the pair of failures on the
east bank of the IHNC, at the west end of the Lower Ninth Ward (Sites #6a and 6b.) These
two major breaches were studied in detail, as presented in Chapter 6; Sections 6.3.1 and 6.3.2.
Although overtopping occurred along much of this IHNC frontage, and although this
overtopping produced erosion at the inboard base of the concrete floodwall (and Iwall
section) at the location of the south breach, both sections failed as a result of underseepage
rather than overtopping.
11 10
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
The large south breach, hundreds of feet in length, failed as a result of underseepage
induced pore pressures which weakened the foundation soils beneath the inboard toe of the
levee embankment, and resulted in translational stability failure of the embankment section
(pushed laterally by the risen waters in the canal.) The northern breach, which was a
narrower, deeper failure, was the result of underseepageinduced hydraulic uplift (“blowout”)
at the inboard toe and underseepageinduced toe erosion and piping.
These conclusions contradict the findings of the IPET Draft Final Report of June 1,
2006, which found the failure of the south section to be the result of overtopping, scour at the
inboard toe of the floodwall, and resultant lateral unbracing of the floodwall. The northern
failure was attributed to deeperseated semirotational failure of the foundation, primarily
through a layer of soft clays, and this failure was assumed to have occurred surprisingly early,
in order to explain observations of large amounts of water collecting on the inboard side
along this frontage. The IPET report also mentions that underseepageinduced failure
mechanisms were not studied, as the foundation soils were too impervious.
There was a long history of underseepagerelated problems along this frontage, and it
is likely that the IPET study would have pursued these if they had been so informed. Instead,
they hued to the history of the local New Orleans District of the USACE, and “absolved”
underseepage as a potential failure mechanism at these two important sites without bothering
to perform formal analyses to study the possibility.
The IPET analyses of their preferred failure modes make no technical sense, and defy
the available data regarding the strengths and stiffnesses of the soils involved (see Sections
6.3.1 and 6.3.2.) Moreover, there is a long history of problems associated with underseepage
along this frontage, including both citizen issues with ponded waters and contractor’s
difficulties with dewatering for construction. The most stunning demonstration of the high
lateral permeability of the “marsh” deposits at this location, however, is a visually spectacular
reverse crevasse splay (see Figure 6.45 ) produced beneath the temporary repair embankment
section by the relatively small reverse flow gradients as the protected had nearly fully
drained.
As shown in Figures 6.24 and 6.47, the relatively short sheetpiles at these two sections
failed to achieve adequate cutoff of underseepage flow through “marsh” strata that were
tantalizingly just below the bases of these sheetpile curtains. That was a repeated theme in
this event, as underseepageinduced failures, as a result of inadequate sheetpile depths (where
moderate extensions of the sheetpiles would likely have prevented failure) also occurred at
the two major, and devastating, breaches on the London Avenue drainage canal (Sites #10
and #11a), and an additional underseepageinduced “incipient” failure began to develop on
the east bank of the London Avenue canal (Site #11b) but halted as the section on the west
bank failed first, drawing down the water levels.
Inadequate sheetpile depths, as a result of overly optimistic assumptions regarding the
permeability of foundation soils, were thus found at five of the most important sites,
including four of the most damaging breaches that occurred during hurricane Katrina. It is
time for the New Orleans District to come to grips with the potential severity of the
underseepage problem. This is a major issue, not only because it represents a likely
11 11
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
remaining source of potential vulnerability throughout the remainder of the system, but also
because it may obviate current cost projections for further improvement of the regional
system. The types of steps needed to remedy underseepagerelated vulnerability are very
different from the types of actions already being undertaken to reduce overtopping
vulnerability.
As mentioned above, the USACE has already taken major steps to add concrete splash
pads and/or to replace concrete Iwalls with Twall sections. This was laudable, and a very
useful step with regard to addressing potential vulnerability associated with overtopping. It
does not, however, also address the potential vulnerabilities associated with underseepage.
At a recent press briefing a USACE representative at the IHNC east bank, at the west end of
the Ninth Ward, indicated the massive concrete splash pads newly installed and remarked that
“King Kong himself could not come over the top of that wall”. Unfortunately, Katrina did
not so much come over the top of that wall, as she passed beneath it.
Important lessons here include:
12. It is important not to exonerate any failure mechanism(s) a priori, not before thorough
analysis. This is true both in design, and in forensic investigations. In all cases, and
especially in design, all failure mechanisms must be considered potentially guilty until
either proven innocent, or until mitigated by appropriate design provisions.
13. Underseepage is one of the common modes of levee failure, and it appears to represent a
considerable potential source of vulnerability throughout much of the New Orleans
regional flood protection system. In addition to five sites studied in detail because
failures (or incipient failures) due to underseepage occurred there, numerous additional
sites were reviewed in a cursory manner and our investigation team was routinely struck
by the surprisingly shallow depths of the sheetpile curtains, and the manner in which
potential concerns regarding underseepage appeared to have been wished away during
design. The installation of massively longer (deeper) sheetpiles, often 60 feet in length
and more, replacing original sheetpiles less than thirty feet in length, as part of repairs at
numerous breach sites represents a de facto and unusually frank admission as to the
systemic inadequacy of pre-Katrina sheetpile penetrations. This is a potentially serious
source of continuing risk to the regional system, and it may also obviate current cost
projections for upgrading the regional system (and recent appropriations for this purpose
as well.)
14. There is an urgent need to perform a thorough, system-wide review of potential
underseepage-related vulnerability.
In addition to the two major breaches at the east bank of the IHNC (at the west end of
the Lower Ninth Ward), three additional breaches occurred on the west bank of the IHNC.
These three breaches are Sites #5a, b and c. These were the first breaches to admit
floodwaters into the main (downtown) protected basin. None of these three breaches
managed to erode or scour a path back to the IHNC with a base consistently below sea level,
however. Accordingly, although these breaches admitted floodwaters briefly while the water
level within the IHNC was elevated, all three breaches subsequently ceased inflow as the
storm subsequently subsided only a few hours later.
11 12
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Site #5a was the west bank of the CSX railroad crossing. The failure at this site is
particularly galling, as this site also failed during hurricane Betsy in 1965, so the repeat
failure represents a very disconcerting failure to learn.
The rail crossing is part of a complex “penetration” through the federal IHNC frontage
levees, as an adjacent roadway serving outboard side Port facilities crosses over the top of the
federal frontage levee adjacent to the railroad line. Our investigation team was unable to
learn just exactly who was overall in charge at this complex site with multiple overlapping
jurisdictions; and that is a problem.
The rail line passes through a concrete Twall structure with a rolling steel floodgate,
so that the floodgate can be closed during storms to complete the perimeter frontage
protection. Unfortunately, the steel gate had been damaged by a railroad accident several
months prior, and it had been taken away for repairs. Accordingly, a temporary “sandbag
levee” was erected across the missing floodgate opening; this washed away at some stage
during hurricane Katrina. It is not clear who was in authority here, but the decision to remove
the steel floodgate and allow trains to continue to operate, rather than affixing the damaged
gate in place until it could be replaced, placed the entire community of the main (downtown)
New Orleans protected basin (a population of approximately 250,000+) at risk. In hindsight,
this was a difficult decision to justify.
Fortunately, the missing gate was not a principal source of flooding for the main
(downtown) protected area. The main breach at this location was actually the result of
composite action of the railroad embankment and the adjacent roadway section. Water
appears to have passed first through the pervious gravel ballast at the top of the railroad
embankment (representing the local “low spot” with regard to stopping flow), and it then
eroded and undermined the adjacent roadway, resulting in a full breach. The levee
embankment underlying the roadway appeared to consist in part of highly erodeable sands
and shellsands, and the presence of such highly erodeable soils without cutoff or other
provisions to prevent catastrophic erosion was very illadvised.
Lessons here include:
15. Someone needs to be overall in charge at “penetrations” and “transitions” where
multiple groups and functions intersect, and where overlapping responsibilities result.
Whoever is in charge needs both to be made responsible for the overall situation, and
they need to be granted adequate authority as to successfully execute that responsibility.
16. The continued operation of trains cannot be allowed to be considered more important
than the safety of major urban populations.
(7, repeated.) Highly erodeable embankment (and foundation) materials represent an
intrinsic hazard, and their use should be avoided in flood protection systems defending
significant populations.
(8, repeated.) When the use of such materials cannot be avoided, then great care should be
taken to protect the sections by means of internal cut-offs, filters, and slope face
protection (armoring) on the front and back faces and on the crest as well. Even then, the
use of erosion-resistant soils is to be preferred if at all possible.
11 13
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
The other two breaches along this frontage occurred at the south end of the main Port
of New Orleans. Both breaches occurred in full height earthen levee embankment sections
that were comprised entirely of lightweight shellsand fill. This was shocking to our
investigators, and the lessons here are simple. Again
(7, repeated.) Highly erodeable embankment (and foundation) materials represent an
intrinsic hazard, and their use should be avoided in flood protection systems defending
significant populations.
(8, repeated.) When the use of such materials cannot be avoided, then great care should be
taken to protect the sections by means of internal cut-offs, filters, and slope face
protection (armoring) on the front and back faces and on the crest as well. Even then, the
use of erosion-resistant soils is to be preferred if at all possible.
An additional set of partially developed erosional features occurred at a number of
locations on the east bank of the IHNC at the west edge of the New Orleans East protected
area. These are grouped together as “Site #7” in Figure 11.4 (and Table 11.1.) Overtopping
occurred along much of this frontage, but the erosional distress systematically occurred at
“transitions” between adjacent, disparate system elements (e.g. at transitions between full
height earthen embankments and adjacent gated concrete floodwall segments, etc.) Here,
again, it was transitions rather than the main segments themselves that proved problematic.
Our field investigation team were initially puzzled that these multiple features all
appeared to be partially developed erosional features, on their way to failure but failing to
reach full failure. As our studies progressed, however, it became clear that the lands on the
inboard side were already filling with floodwaters as these features were developing, and this
reduced the gradients and the durations of flow. It is not possible to know whether any of
these features might have developed into full breaches if the inboard side lands had been more
successfully defended against flooding from breaches that occurred at other locations.
Lessons here thus include
(10, repeated.) “Transitions” where adjacent project elements join together were routinely
problematic throughout much of the regional flood protection system. Successful design
and construction of two adjoining project sections counts for little if the connection
between them is not also successfully consummated. Significantly more attention needs
to be paid to these “transitions”.
17. The multiple transitions along this frontage suffered erosional damage but did not fully
fail (and breach. But they may not have been fully tested as floodwaters were likely
already rising on the inboard side lands as a result of massive breaches at other
locations. These transitions should therefore be thoroughly re-evaluated as part of
ongoing flood protection system upgrades.
11.3.5 The Lake Pontchartrain Frontage, and the Drainage Canals
As the eye of the hurricane finally passed to the northeast of New Orleans, its counter
clockwise swirling winds drove a final storm surge south along the shoreline of Lake
Pontchartrain, along the north edge of the city.
11 14
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
This final storm surge produced some degree of overtopping at several locations along
the lake frontage levees of the New Orleans East protected area, and two failures occurred
(Sites #8a and b.) Site #8a was another complex “penetration” where multiple rights of way
passed, together, through the federal levee perimeter. These included an elevated State
highway, yet another railroad line, and a groundlevel roadway. These three elements
interacted poorly together; flow through the pervious railroad embankment ballast
undermined the connection between a concrete floodwall protecting a support for the elevated
highway and the adjoining surface roadway, and flow across these features also eroded an
adjacent section of the Federal perimeter earthen levee. Once again, it was not clear who, if
anyone, was overall in charge at this complex site. Thus
(15, repeated.) Someone needs to be overall in charge at “penetrations” and “transitions”
where multiple groups and functions intersect, and where overlapping responsibilities
result. Whoever is in charge needs both to be made responsible for the overall situation,
and they need to be granted adequate authority as to successfully execute that
responsibility.
The second site was a long section of floodwall whose crest was surprisingly low.
Overtopping occurred along this low section over approximately a mile of floodwall length,
despite a lack of persistent, sustained overtopping at any adjacent sections along this lakefront
frontage. Significant scour occurred at the rear base of this floodwall, and a minor breach
occurred at one location where floodwall panels shifted a bit as a result.
Farther to the west, the storm surge along the Pontchartrain Lake frontage levees at the
north end of the main (downtown) New Orleans protected area did not produce meaningful
overtopping along the lake frontage. This storm surge did, however, raise the water levels in
three drainage canals that emptied into the Lake…. and three major breaches occurred along
these drainage canals. These three breaches all rapidly scoured to well below sea level, and as
a result floodwaters continued to flow in through these breaches for three days (even after the
storm surge had subsided) eventually equilibrating with the still slightly elevated waters of
Lake Borgne on Thursday, September 1st. These floodwaters infilled much of the main
(downtown) New Orleans protected basin, resulting in roughly half of the overall deaths
during hurricane Katrina, and a similar fraction of the damages as well.
The three drainage canals should never have been exposed to storm surge rise. The
USACE had fought for years to install storm gates at the north ends of the three canals, but
had been defeated (outmaneuvered in Congress) by local interests as a result of dysfunctional
interactions and distrust between the local Levee Board (who were nominally responsible for
perimeter levee protection) and the local Water and Sewerage Board (who were responsible
for pumping and “unwatering” of New Orleans.) Every drop of rainwater that falls into New
Orleans has to be pumped out, as the city is largely below sea level. Rainfall, and constant
levee underseepage, are the principal concerns for “unwatering” on the part of the Water and
Sewerage Board in most years, while the Levee Board is concerned primarily with providing
protection during infrequent river floods and hurricanes. Accordingly these two organizations
have differing principal focuses.
11 15
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
That led to dysfunctional interaction between them, distrust, and eventually even
animosity. The Water and Sewerage Board, concerned that storm gates would be “perimeter
protection” under the control of the Levee Board (and thus might possibly not be opened
promptly when rainfall required pumping out through the drainage canals) fought successfully
to have Congress decline the USACE’s request for construction of storm gates at the heads of
the canals.
Unfortunately, the construction of the floodgates would have been the superior
technical solution. Instead, the canals remained “open” to Lake storm surges, and the three
canals thus represented daggers pointed at the heart of New Orleans.
The USACE then attempted to exempt the three drainage canals (the 17th Street canal,
the Orleans canal, and the London Avenue canal) from federal responsibility. Local interests
again outmaneuvered the Corps, and Congress specifically declared these to be a federal
responsibility; they required the USACE to raise the levels (elevations) of protection along the
sides of these three canals.
The USACE correctly pointed out that the “footprints” available for levees along the
sides of these canals (especially the 17th Street and London Avenue canals) were insufficient,
as homeowners’ properties abutting the canals encroached on the existing levees; in some
cases property lines extended up the levee slopes to the edges of the narrow crests. There was
insufficient room available to safely widen these levees in order to add to their heights.
This too was overruled, and the USACE was directed to raise the levels (elevations)
of protection along these canals, within the existing (inadequate) “footprints” available. The
results were catastrophic. Lessons here include
18. The USACE is the lead Federal agency with expertise regarding levees and flood control.
The USACE needs to be resolutely vocal and persistent in declining to undertake actions
that it considers to be unsafe. Congress needs to better heed due warning from the
Corps. The interactions between Congress and the Corps need to involve improved giveand-take; the Corps needs to be allowed to better assert strongly held professional
opinions.
19. Local interests, and special interests, cannot be permitted to “outmaneuver” legitimate
technical concerns with regard to Public Safety. The local Levee Board, and the local
Water and Sewerage Board, should have been required to resolve their personal
differences in the greater interest of Public safety.
20. The USACE felt that the path they were directed to follow was unsafe. They could have
simply refused to take that path. That might have required resignations on the part of the
leadership; those would have been honorable resignations.
Having been essentially ordered to raise the levees (and floodwalls) along the three
drainage canals, the USACE recognized that this posed significant technical challenges.
Accordingly, they next performed a very welldirected fullscale experiment in the nearby
Atchafalia River basin (on soil conditions very closely mirroring the challenging geology of
11 16
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
the drainage canals) in which a concrete floodwall (Iwall) was modeled using a plain
sheetpile wall. This test section (the E99 test section) was constructed on the berm of an
Atchafalia levee, with the berm height closely modeling the existing levee heights along the
three drainage canals. A sheetpile cofferdam was constructed, and filled with water (to model
storm surge loads against the sheetpile/floodwall).
This important largescale field experiment clearly showed that under storm surge rise,
a “gap” was likely to form between the sheetpile curtain supporting concrete Iwalls, and that
this gap would fill with water; significantly increasing the lateral pressures applied by water
pressures against the sheetpiles/floodwalls. This mechanism subsequently figured in all three
of the catastrophic drainage canal failures that occurred during hurricane Katrina, and in a
number of other failures at other sites during Katrina as well. The failure to include this
potential failure mode in the subsequent analysis and design of large elements of the regional
flood protection system proved disastrous.
Unfortunately, the important lessons from this expensive and welldirected fullscale
field test were never subsequently incorporated into the design of the floodwalls used to raise
the protection elevations along the three drainage canals. Thus
21. Our investigation uncovered a persistent failure to learn; to adapt to technical advances,
and even to heed the results of the USACE’s own research, on the part of the New
Orleans District. Outdated analysis methods, and strongly held views (which proved to
be in error) were key failings in the design and construction of the flood protection
system at a number of locations.
Another particularly important location was the south end of the Orleans drainage
canal (Site #9.) Although it was located between the 17th Street drainage canal and the
London Avenue drainage canal (catastrophic breaches occurred on both of these canals), no
breaches occurred on the Orleans canal. Instead, storm surge waters simply flowed freely into
the heart of New Orleans through an unfinished “gap” in the floodwalls lining this canal. A
section of concrete floodwall approximately 200 feet in length at the south end of this canal
was “omitted”; rendering the miles of floodwalls lining the remainder of the canal somewhat
superfluous.
The omission of the last several hundred feet was done to protect the ancient (1904)
brick building housing the several giant Woods pumps that pumped waters from the
neighborhood into the canal (and thus into Lake Pontchartrain.) This brick building forms a
“T” at the south end of the canal, closing the canal. When the canal water levels rise more
than about five of six feet (e.g.: during pumping) , water seeps actively through the walls of
the old brick building, and it is clear that significantly further rises in water levels would
threaten to buckle the wall. Thus, either: (1) the Levee Board (who were responsible for
“protection”) would have had to erect a barrier to protect the Water and Sewerage Board’s
pump house , or (2) the Water and Sewerage Board would have had to expend their own
resources to erect this protection themselves; helping out the Levee Board in the process by
closing the end of the canal. As a result of internecine battling between these two agencies,
and their inability to resolve their differences in the interest of the greater common good (and
Public safety), neither occurred. Instead a gap was left in the floodwall (to control maximum
11 17
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
possible canal water elevations (at Elev. +9 feet) and a “spillway” section was constructed
across this gap until the matter could be further resolved.
(19, repeated.) Local interests, and special interests, cannot be permitted to “outmaneuver”
legitimate technical concerns with regard to Public Safety. The local Levee Board, and
the local Water and Sewerage Board, should have been required to resolve their personal
differences in the greater interest of Public safety.
Eventually, however, floodwall systems were designed and constructed to raise the
crest elevations of the levees along these three canals. A number of engineering errors, poor
judgements, and poor decisions occurred during this process, and these too contributed to the
three catastrophic breaches that occurred (at Sites #10, 11a, and 12b.) In addition, two
“incipient” failures nearly occurred, but which were “saved” by nearby failures that rapidly
drew down the canal water levels. One of these “near failures”, on the west side of the 17th
Street canal (Site #12a) would have resulted in flooding of a considerable portion of heavily
populated Jefferson parish, and would have significantly increased the overall damages (and
likely loss of life as well) from hurricane Katrina.
The first mistake was the failure to secure adequate rightofway to widen the levees,
to provide adequate embankment mass and weight as to sustain the increased lateral water
forces that would be imposed by taller floodwalls atop the levee crests. This also meant lack
of access and control over some of the inboard side levee faces and the critical inboard toe
regions; rendering both inspections and necessary maintenance difficult.
These would both have disastrous consequences. Failure to purchase adequate right
of way contributed significantly to inadequate lateral stability at the massive breaches at the
17th Street canal (Site #12b) and at the west bank near the north end of the London Avenue
canal (Site #11a.) Lack of control, and lack of access for inspection of the critical inboard toe
areas led to rampant growth of trees at the toes (a known hazard), and even to excavations for
swimming pools near the inboard toes of the levees in this critical region. This uncontrolled
growth of trees appears to have contributed to the large failure on the east bank of the London
Avenue canal (Site #10.)
(2, repeated.) The failure to purchase adequate right-of-way in what had become (expensive)
developed neighborhoods resulted in initial project savings, but these cost-savings were
multiple orders of magnitude less than the subsequent damages that occurred when these
levee sections failed. Short-term cost savings in construction need to be balanced
against the consequent increases in risk (the consequent reduction in likely reliability)
for the resulting built system.
22. The inboard toe region is a critical area with regard to both inspections and
maintenance. Uncontrolled vegetation growth and other obstructions to maintenance
and inspections need to be precluded. Large trees can die and leave rotted root systems
that provide dangerous paths for seepage, and during hurricanes strong winds (and
ground wetting which reduces root anchorage) routinely lead to toppling of trees. Trees
on the inboard levee faces and the inboard toe regions can thus fall, leaving sudden voids
that can cause or exacerbate “blowouts’ and/or erosion and piping failures. Conditions
11 18
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
on the inboard side slopes and toes of considerable lengths of the levees lining both the
London Avenue and 17th Street canals represented clear potential hazards in these
regards.
23. In addition, it is customary policy for the USACE to require serviceable crest roads at
the tops of levees to provide access for inspection, maintenance and emergency repairs.
Given the failure to acquire adequate right-of-way, and resulting narrow crest widths,
this was waived.
The failure at the 17th Street canal was a lateral translational failure of the levee
embankment, with the principal shear surface constrained by a thin, weak, and highly
sensitive layer of organic clayey silt. Only one to several inches in thickness, this layer
resulted from a previous hurricane that passed through; churning up organic matter, mixing it
with the local silts and clays, and depositing a layer heavily flocculated clayey silt due to the
storminduced temporary increase in salinity of the local waters. This layer, which was only
one to several inches in thickness, was wellhidden by an overlying layer of sticks and twigs
and leaves, representing stormblown detritus from the causative hurricane. This overlying
layer obstructed conventional geotechnical sampling of this thin, sensitive layer, and also
clear detection of this layer by conventional CPT.
The presence of this critical stratum went undetected by the original design studies,
and by the postevent IPET forensic studies as well. The failure to detect this layer in both
studies, despite drilling numerous boreholes through it, and pushing multiple CPT through it
as well, was largely a result of employing “common practice” in which the field drilling (and
CPT) were performed by personnel without special experience or geological expertise. The
IPET team certainly had expert geological engineers, with experience with these types of
strata, who could have usefully advised this process, but they were sidelined with other tasks
(including writing up “geology” sections for the report.) There was segmentation (or
compartmentalization) of the work both in the original design studies and in the subsequent
forensic studies. Field personnel performing the drilling were insufficiently directed by
engineers who had performed the important initial postevent forensic inspections, expert
geologic input was insufficient, and the eventual analysts were not properly appraised of the
full pertinent details by the other subteams.
That contrasts sharply with the approach taken by our (ILIT) team at this site. Despite
considerable prior experience with these types of deposits, a thorough study was made of local
geological nuances prior to beginning drilling and sampling (and CPT). Expert senior team
members were present at the field boring and sampling, and the CPT, including specifically
toplevel expertise in geological engineering. Careful initial (immediate postevent) field
forensics had already led to the suspicion that the failure was a lateral translational failure,
controlled by a weak (and likely highly sensitive) layer occurring at a depth of approximately
3 to 8 feet beneath the inboard toe, producing laterally exiting toe features (including exiting
overthrusts) to unusually great distances beyond the levee toe. Having studied the local
geology, a highly sensitive organic clayey silt or silty clay layer (which might be very thin,
and likely screened by overlying wind blow organic detritus) was a leading potential suspect.
Our first boring discovered the failure stratum, and we then proceeded to follow it across the
site (including sampling it at locations within the failure zone, at the toe of the displaced intact
11 19
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
levee block, where the layer was clearly unidirectionally sheared and remoulded;
incontrovertibly the failure surface.)
The lessons here include:
24. Engineering geology is of vital importance. Always has been, and always will be. It must
be interwoven throughout all phases of geotechnical works; from pre-study, through site
investigation, and through analysis and design as well. Geologists are too often treated
as second class citizens, and some geotechnical firms no longer even “need” them.
Failure to avail ourselves of expert geological insight, and at all stages of a project, is to
needlessly imperil the effort.
25. Increasingly, the trend in “modern” practice is to segment geotechnical works;
separating field investigation (e.g. borings and sampling, CPT, etc.), laboratory testing,
analysis, and design. These elements need to be seamlessly interwoven, and iterative
cross-communication between the personnel performing these needs to be thorough.
Sadly, that is increasingly not the case; not only in government works, but in common
(private) civil practice as well. This segmentation can be more “cost effective”…. The
risk is that something will be missed.
In addition, review of the original design studies showed ten additional engineering
lapses and/or questionable judgements at this site, as enumerated in Chapter 8; Section
8.3.7.1(c).
These included extrapolation of data across excessive distances, failure to
recognize “red flags” such as failure to recover samples at the same elevation in nearby
borings (the elevation where the critical, sensitive, and verydifficulttosample organic
clayey silt stratum occurred), etc. Readers are directed to this section for a full listing.
Several key lessons include the following:
Basic principles of soil mechanics were neglected, as the influence of increased effective
stress beneath the centerline of the levee embankment was ignored, and soil shear strengths
beneath the levee toes (where effective overburden stresses were smaller) were
overestimated as a result. Shear strengths were extrapolated over lateral distances that were
too great, and sometimes over excessive vertical distances. Shear strength profiles used for
design calculations were not welljustified at certain, critical, elevation ranges by the data
available.
An archaic analysis method, the Method of Planes, was used for most stability calculations.
This method (involving three blocks or wedges, and a conservative side force assumption
between wedges) provides a demonstrated conservative answer for cases to which it can be
applied. It is inflexible with regard to geometry, however, and was unable to deal with non
level stratigraphy and curvilinear failure surfaces. More modern and flexible methods were
in common use at the time of these design studies, and should have been employed.
The formation of a waterfilled “gap” between the sheetpile curtain and the outboard portion
of the levee embankment was not considered among the potential failure modes; despite the
welldirected E99 test section fullscale experiment that had shown this mode to be of
concern.
11 20
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
The design Factor of Safety for overall lateral stability during “transient” storm surges was
only 1.30. That was far too low to allow an adequate margin for errors and uncertainties.
That design Factor of Safety had evolved from tradition, and dated back to the middle of the
last century, at which time it was selected for design of levees providing protection for
agricultural lands (not populous regions). The design standard had not been updated, nor
adapted for levees protecting large populations.
26. All of these problems would have been expected to be caught and challenged by a
competent panel of independent technical reviewers. Instead, reviews of the largely
locally “outsourced” engineering design were performed internally within the USACE.
The mobilization of suitable, independent expert review capability is one of the most
important steps that can be taken to enhance the likelihood of improved system
reliability and performance in future events.
27. There was a persistent failure to learn, and to adapt to technical advances, within the
local New Orleans District that affected performance of the regional flood protection
system at numerous sites. Difficulties in recognizing potentially important “new” issues
has continued in some cases since the hurricane; “That’s not how we do it” needs to
cease to be an issue…. On the heels of system-wide failure, changes are in order. The
USACE needs to ensure that the New Orleans District is adequately technically staffed
for the magnitude and technical difficulty of the challenges it faces with regard to the
engineering design and construction of critical flood protections systems in a region with
exceptionally challenging geology, and in the face of both local and federal governmental
assistance/interference as well. Suitable technical advances need to be studied, and
embraced if appropriate. Upgrading personnel, and education and training, will also be
important.
28. Design standards, especially with regard to targeted levels of system reliability, need to
be reconsidered (see Figures 11.1 and 11.2). This is already underway; the USACE is
performing a comprehensive re-assessment of design procedures and standards, and
treatment of flood protection systems on a risk-based systems basis is anticipated to be an
important element of this. That is a very promising development.
Two additional large failures (and breaches) occurred on the London Avenue drainage
canal. The failure on the east bank, near the south end (Site #10) was the result of
underseepageinduced erosion and piping and/or underseepageinduced hydraulic uplift at the
inboard toe (“blowout”), and it may have been exacerbated by a large tree at the inboard toe
of the levee that blew over during the storm (at approximately the location of the failure.)
The failure on the west bank, near the north end, was an underseepageinduced lateral
embankment stability failure; the embankment slid laterally, pushed by the increased canal
water pressures, and shearing occurred along foundation soils whose strengths had been
reduced by underseepageinduced pore pressure increases (and resultant reductions in
effective stress.)
The principal lessons to be learned from these two additional cases are repeats of
lessons cited previously, and will not be repeated here again.
11 21
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
11.4 References
Christian, John T. (2004), “Geotechnical Engineering Reliability: How well do we know
what we are doing?” J. of Geotechnical and Geoenvironmental Engineering, Vol. 130,
No. 10, October 1, 2004.
IPET, (2006), “Performance Evaluation of the New Orleans and Southwest Louisiana
Hurricane Protection System, Draft Final Report of the Interagency Performance
Evaluation Task Force, Volume VI – The Performance – Interior Drainage and Pumping,”
available online: https://ipet.wes.army.mil/, date accessed: June 1, 2006.
11 22
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
New Orleans
U.S. Dams
Figure 11.1: Risk plot showing the estimated preKatrina risk associated with the New
Orleans regional flood protection system, and customary risk levels for
current U.S. Practice with dams.
[Baseline Figure from Christian, 2004]
11 23
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
New Orleans
(Pre-Katrina)
100-Year
Protection
1,000-Year
Protection
U.S. Dams
Figure 11.2: Risk plot showing the estimated preKatrina risk associated with the New
Orleans regional flood protection system, and customary risk levels for
current U.S. Practice with dams. Also shown are projected New Orleans
risk levels for “successful” 100year and 1,000year storm and flood design.
[Baseline Figure from Christian, 2004]
11 24
ywvutsrpon
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Need t o Change How Regional Flood Prot ect ion
Syst em s are Creat ed and Maint ained
Concept ion
Congress
Plannin g
U.S.Arm y Corps
of Engineers
Analysis and
Design
Out sourced
Engineering
Const ruct ion
Out sourced
Const ruct ion
St at e & Local
Governm ent and
Oversight Ag encies
Operat ion and
Maint enance
A Reliable Syst em t hat Works ( Acce pt able Risk)
Figure 11.3: Engineering and organizational elements intrinsic to the creation, operation
and maintenance of major U.S. regional flood protection systems in the
Mississippi river basin.
11 25
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
1 1 a ,b
1 2 a ,b
8 a ,b
4 a ,b
7
3
10
9
2a
5 a ,b,c
6 a ,b
2b
1
1
( Pla qu e m in e s Pa r ish )
usrqnm
Figure 11.4: Summary of principal failures, breaches, and other locations of interest. (Blue stars mark breaches,
red stars mark locations of distress.)
[Base map provided by the USACE]
11 26
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Table 11.1: Summary of Principal Damage Features Studied
Site No.
or
Group
No.
General Location
Failure Mechanism and Cause
(This Study: ILIT)
Failure Mechanism
and Cause (IPET;
June 1, 2006)
Severity of
Consequences
1
Plaquemines Parish
Multiple failures resulting principally from massive
and sustained overtopping.
Same as ILIT
Catastrophic
2a
MRGO Frontage, St.
Bernard Parish
Massive erosion and scour along many miles of
levees due to waves, overtopping, throughflow,
and use of highly erodeable embankment materials.
Overtopping erosion
Catastrophic
2b
South MRGO Frontage,
St. Bernard Parish
Erosion and scour along isolated sections of levees
due to waves, overtopping, throughflow, and use
of highly erodeable embankment materials.
Overtopping erosion
Catastrophic
3
GIWW Frontage,
southeast corner of New
Orleans East
Massive erosion and scour along many miles of
levees due to waves, overtopping, throughflow,
and use of highly erodeable embankment materials.
Overtopping erosion
Catastrophic
4a
North Bank of GIWW;
the Citrus back levee
floodwall
Overtopping of concrete Iwall, resulting in scour
behind the wall, which was pushed sideways by
elevated water levels.
Same as ILIT
Moderate
4b
North Bank of GIWW;
sheetpile "transition"
wall
Overtopping of concrete Iwall and sheetpile
transitions resulting in scour behind the wall, which
was pushed sideways by elevated water levels.
Same as ILIT
Moderate
5a
CSX Rail Crossing,
west bank of IHNC
Poor coordination and poor interaction of multiple
elements at a complex "penetration." Also, use of
highly erodeable fill materials.
Rail gate absent
Moderate
5b
Earthen levee
embankment near south
end of Port
Use of highly erodeable fill materials.
Overtopping erosion
Moderate
5c
Second earthen levee
embankment near the
south end of the Port
Use of highly erodeable fill materials.
Overtopping erosion
Moderate
11 27
Comments
Storm surge and waves
generally exceeded design
levels.
Consequences would have been
more severe if other sections of
the protected basin had not
failed.
Consequences would have been
more severe if other sections of
the protected basin had not
failed.
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Table 11.1 (cont’d)
Site No.
or
Group
No.
General Location
6a
East bank of IHNC at
edge of the lower Ninth
Ward: South Breach
6b
Failure Mechanism
and Cause (IPET;
June 1, 2006)
Severity of
Consequences
Underseepageinduced lateral transitional failure.
Overtopping of Iwall
and scour
Catastrophic
East bank of IHNC at
edge of the lower Ninth
Ward: North Breach
Underseepageinduced erosion and piping and/or
underseepageinduced hydraulic uplift at inboard
toe.
Deep, semirotational
foundation failure
through soft clays
Catastrophic
7
Cluster of minor erosion
features at "transitions,"
east bank of IHNC
Erosion at inadequately detailed “transitions”
between adjoining, disparate flood protection
system segments.
---
Minor
8a
Erosional breach at
northwest corner of
New Orleans East
Erosional failure of another complex
“penetration” where multiple interacted poorly as
they crossed the perimeter levee system.
---
Moderate
8b
Overtopping and breach
at floodwall behind Old
Lakefront Airport
Same as ILIT
Moderate
9
“Missing” floodwall
section at south end of
the Orleans Canal
Floodwall section omitted due to poor interactions
between local oversight agencies.
---
Minor
10
South breach on the east
bank of the London
Avenue drainage canal
Underseepageinduced erosion and piping and/or
hydraulic “blowout” at the inboard toe.
Same as ILIT
Catastrophic
Failure Mechanism and Cause
(This Study: ILIT)
Overtopping of a surprisingly low floodwall,
resulting in erosion behind the floodwall.
11 28
Comments
Consequences might have been
more serious if other sections of
the protected basin had not
failed.
Consequences might have been
more serious if other sections of
the protected basin had not
failed.
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Table 11.1 (cont’d)
Site No.
or
Group
No.
General Location
11a
North breach on the west
bank of the London
Avenue drainage canal
11b
Failure Mechanism
and Cause (IPET;
June 1, 2006)
Severity of
Consequences
Underseepageinduced lateral stability failure.
Same as ILIT
Catastrophic
Incipient failure on the
east bank
Underseepageinduced lateral stability failure
was beginning, when the west bank failure
occurred and drew down the canal water level.
---
Negligible
12a
Breach on the east bank
of the 17th Street Canal
Lateral translational levee foundation failure on
a highly sensitive layer of organic clayey silt
embedded within “marsh” deposits.
12b
Incipient failure on the
west bank
Deeper, semirotational foundation failure
within soft clays.
Failure Mechanism and Cause
(This Study: ILIT)
11 29
Deeper, semi
rotational foundation
failure within soft
clays
---
Comments
This was nearly a fourth
catastrophic failure along the
drainage canals.
Catastrophic
Near miss
This would have flooded large
portions of the adjoining and
heavily populated Jefferson
parish.
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
CHAPTER TWELVE: ORGANIZED FOR FAILURE
We reflect on the 9/11 Commission's finding that the most important failure
was one of imagination. The Select Committee believes Katrina was primarily a
failure of initiative. But there is, of course, a nexus between the two. Both
imagination and initiative - in other words, leadership - require good
information. And a coordinated process for sharing it. And a willingness to use
information - however imperfect or incomplete - to fuel action.
Hundreds of miles of levees were constructed to defend metropolitan New
Orleans against storm events. These levees were not designed to protect New
Orleans from a category 4 or 5 monster hurricane, and all of the key players
knew this. The original specifications of the levees offered protection that was
limited to withstanding the forces of a moderate hurricane. Once constructed,
the levees were turned over to local control, leaving the USACE to make
detailed plans to drain New Orleans should it be flooded.
The Local sponsors - a patchwork quilt of levee and water and sewer boards were responsible only for their own piece of levee. It seems no federal, state, or
local entity watched over the integrity of the whole system, which might have
mitigated to some degree the effects of the hurricane. When Hurricane Katrina
came, some of the levees breached - as many had predicted they would - and
most of New Orleans flooded to create untold misery.
A Failure of Initiative
Final Report of the Select Bipartisan Committee
U.S. House of Representatives, 109th Congress (2006)
12 1
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
12.1
Introduction
This chapter summarizes results of studies performed by members of the Independent
Levee Investigation Team (ILIT) into the organizational and institutional factors associated with
failure of the Flood Defense System for the greater New Orleans area (NOFDS).
Over a period of eight months following failure of the NOFDS on 29 August 2005, the
ILIT examined more than 2,800 documents, conducted more than 220 interviews, and reviewed
more than 370 inputs from the general public. During the past 8 months, there have been a large
number of extensive investigations into the reasons for the failure of the NOFDS. The ILIT made
full use of results from these investigations. These results were combined with results from the
ILIT investigations to formulate the primary findings documented in this chapter; organized for
failure.
Chapter 13 outlines our thoughts on future organizational developments; organizing for
success. Chapter 14 summarizes background on engineering a longterm NOFDS and the
associated engineering guideline developments; engineering for success.
Appendix F presents a synopsis of the history of developments in the NOFDS between
1965 and 2005, summarizes background on understanding failures of engineered systems, and
provides key quotations and results from other studies of failure of the NOFDS. Results from
studies of the engineering and organizational aspects associated with future developments of a
NOFDS are summarized in Appendix G. Appendix H, by Dr. Edward Wenk, Jr., documents a
study of How Safe is Safe? - Coping with Mother Nature, Human Nature and Technology's
Unintended Consequences.
12.2
ク
ク
Purposes
The ILIT studies have two purposes:
to understand how and why the failure of the NOFDS developed, and
to understand alternatives to reduce the likelihoods and consequences of such future
catastrophes.
If we can adequately understand the mistakes of the past, then perhaps we have a chance
to avoid making them in the future.
The ILIT approach in this study was to include historical and organizational
institutional issues, political and budgetary considerations, decision making, utilization of
technology, and the evolving societal, governmental, and organizational priorities over the life of
the NOFDS. One cannot develop an adequate understanding of the failure of the NOFDS without
understanding both the engineering and organizational factors that were interwoven in
development of this failure.
12.3
Failure of the New Orleans Flood Defense System
Of particular importance in this diagnosis is the organizational institutional Technology
Delivery System (TDS) that was used to develop the NOFDS. This TDS is comprised of three
major components: (1) Government (federal, state, local), (2) Industry, and (3) the Public. All of
these components and elements are interconnected through a complex series of multiple
12 2
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
connections that represent information and communication transmission. Inputs to the system
include technical information, human and natural resources, capital, manufactured goods and
services and values and preferences. Outputs from these components are represented in the
NOFDS including its intended and unintended consequences.
The Government component is represented by agencies from all three branches
(executive, legislative, judicial) at federal, state, and local levels. There are important multiple
connections among federal, state, and local (parish, city) agencies. In the case of the NOFDS, the
primary agencies are the Corps of Engineers, the Louisiana Department of Transportation and
Development, and the parish levee boards and sewerage and water boards. All of these
government agencies are interconnected with a multitude of other federal, state, and local
agencies. These parts of the TDS were summarized by the Select Bipartisan Committee to
Investigate the Preparation for and Response to Hurricane Katrina (2006):
Several organizations are responsible for building, operating, and maintaining
the levees surrounding metropolitan New Orleans. USACE generally contracts to
design and build the levees. After construction USACE turns the levees over to a
local sponsor. USACE regulations state that once a local sponsor has accepted a
project, USACE may no longer expend federal funds on construction or
improvements. This prohibition does not include repair after a flood. Federally
authorized flood control projects, such as the Lake Ponchartrain project, are
eligible for 100 percent federal rehabilitation if damaged by a flood.
The local sponsor has a number of responsibilities. In accepting responsibilities
for operations, maintenance, repair, and rehabilitation, the local sponsor signs a
contract (called Cooperation Agreement) agreeing to meet specific standards of
performance. This agreement makes the local sponsor responsible for liability for
that levee. For most of the levees surrounding New Orleans, the Louisiana
Department of Transportation and Development was the sate entity that originally
sponsored the construction. After construction, the state turned over control to
local sponsors. These local sponsors accepted completed units of the project from
1977 to 1987, depending on when the specific units were completed. The local
sponsors are responsible for operation, maintenance, repair, and rehabilitation of
the levees when the construction of the project, or a project unit, is complete.
In development of the NOFDS, the Corps of Engineers had the primary responsibilities
for development of the concepts, design, and construction (Collins 2005; National Academy of
Engineering 2006). Once construction was completed, the operations and maintenance were then
turned over to the responsible state and parish agencies. At the federal level, the Corps of
Engineers had important interfaces with the executive branch (e.g., Department of Defense and
the White House), the legislative branch (Congress), and the judicial branch. Important interfaces
also developed with state, parish, and city government agencies, industry, and with the general
public. The Industry component is represented by commercial enterprises that are involved
throughout the lifecycle of the system including concept development, design, construction,
operation, and maintenance. The Public component is represented by national, state, and local
individuals and groups that are concerned with and influenced by the NOFDS.
12.4
Extrinsic Factors
Failure of the NOFDS is firmly rooted in Extrinsic factors associated with human and
organizational performance (Appendix F; Rasmussen 1997; Svedung and Rasmussen 2002; Bea
12 3
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
2006). Causes of the NOFDS failure spanned the full spectrum of organizational failures:
cultures, communications, lack of knowledge, use of existing technology, structure and
organization, management, leadership, monitoring and control, and mistakes. Mistakes involved
breakdowns in perceptions, interpretations, decisions, discrimination, diagnoses, judgments, and
actions. In several notable cases, doing things right and doing the right things apparently were
surrendered to getting the job done in an expedient way. These observations were summarized
by the Select Bipartisan Committee to Investigate the Preparation for and Response to Hurricane
Katrina (2006):
Both USACE and the local sponsors have ongoing responsibility to inspect the
levees. Annual inspections are done both independently by USACE and jointly
with the local sponsor. In addition, federal regulations require local sponsors to
ensure that flood control structures are operating as intended and to continuously
patrol the structure to ensure no conditions exist that might endanger it. Records
reflect that both USACE and the local sponsors kept up with their responsibilities
to inspect the levees. According to USACE, in June 2005, it conducted an
inspection of the levee system jointly with the state and local sponsors. In
addition, GAO reviewed USACE's inspection reports from 2001 to 2004 for all
completed project units of the Lake Ponchartrain project. These reports indicated
the levees were inspected each year and had received 'acceptable' ratings.
However, both the NSF-funded investigators and USACE officials cited instances
where brush and even trees were growing along the 17th Street and London
Avenue canals levees, which is not allowed under the established standards for
levee protection. Thus, although the records reflect that inspections were
conducted and the levees received acceptable ratings, the records appear to be
incomplete or inaccurate. In other words, they failed to reflect the tree growth,
and of course, neither USACE nor the local sponsor had taken corrective actions
to remove the trees.
Complex formal and informal organizations developed that involved a multiplicity of
federal, state, parish, city, commercial industrial, and public enterprises. These organizations
had vastly different means, methods, and resources that evolved in different ways at different
times. Executive, legislative, and judicial forms of government provided a primary framework
for interactions with commercial, industrial, public, and private enterprises. Malfunctions within
and between these organizational elements provided the primary element responsible for the
failure of the NOFDS (Government Accountability Office 2005, Members Scholars of the Center
for Progressive Reform 2005, Select Bipartisan Committee to Investigate the Preparation for and
Response to Hurricane Katrina 2006, Townsend 2006, Houck 2006). Ineffective leadership and
management were evident before and after failure of the NOFDS. Leonard and Howitt observed
(2006):
The leadership failures that contributed to the events we witnessed on the Gulf
Coast last August and September began long, long before Katrina came ashore. It
literally took centuries to make the mistakes that rolled together to make Katrina
such a vast natural and human-made calamity. First, for hundreds of years,
people have been constructing and placing large amounts of precious (human
lives) and expensive (infrastructure, homes, communities) value in New Orleans
and along the Gulf Coast in the known path of severe storms. Second, for
decades, we have been living with inadequately designed, built, or maintained
12 4
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
man-made protections (levees, building codes, pumps, and so on), and have
pursued policies and interventions that actively contributed to the destruction of
the natural buffers (salt marshes, dunes, and other natural barriers) against the
hazards created by placing value in harm's way. Third for years - at least since
9/11, but even before that - we have known that we had systems of preparation
and response that would prove inadequate against truly large scale disasters.
Fourth, in the days and hours before Katrina's landfall, we failed to mobilize as
effectively as we might have those systems that we did have in place. And fifth, the
days following the impact, we did not execute even the things that we were
prepared to do as quickly and smoothly as we should have. How do we not, in the
future, find ourselves again with those same regrets? Our work needs to begin
with a judicious and honest assessment of threats, followed by investments in
prevention and mitigation and by construction of response systems that will be
equal to a larger of class of disturbances than we have previously allowed
ourselves to contemplate.”
In development of the analysis of Extrinsic factors involved in failure of the NOFDS, it is
important to recognize that while the Corps of Engineers was primarily responsible for design
and construction of the NOFDS and the local state and parish organizations (e.g., Department of
Transportation and Development, Levee Boards, Sewerage and Water Boards) were primarily
responsible for operations and maintenance, these organizations were subjected to a wide variety
of influences and constraints provided by their executive, legislative, judicial and public
constituents. The responses of these multiple organizations to provide an adequate NOFDS was
clearly lacking. The Senate Committee on Homeland Security and Governmental Affairs report
supports this (Leonard and Howitt 2006).
For many years, the Corps of Engineers was severely criticized for delays and cost
increases in the Lake Pontchartrain and Vicinity Hurricane Protection Project (Government
Accountability Office 1972, 1982, 2005b, 2005c; Carter 2003, 2005a, Carter and Sheikh 2003,
Carter et al 2005). Many of these delays and cost increases were reflections of challenges posed
by local cost sharing and participation requirements. Local participation and funding
requirements introduced additional problems as did interactions with the general public.
Additional complexities were added by federal and state legislative, executive, and judicial
participation in the developments (lots of 'managers' with different goals, objectives, means, and
methods).
At the federal level a long and complex process is required to identify, define, select, and
develop projects and secure funding authorizations (Carter and Hughes 2005, Carter 2005d).
Historic problems exist with project backlogs, increases in funding requirements, reprogramming
actions to manage project funds, and even the fundamental basis for project selection; cost
benefit analyses (Government Accountability Office 1983, 2003, 2005). In short, the Corps must
operate in a world that is not of its own making. Outside pressures on the Corps have been
negative as well as positive in terms of their effects on performance.
This study indicates that the historic procedures utilized to develop the costbenefit
analyses employed by the Corps of Engineers were and are seriously flawed. These procedures
are apparently responsible for some of the seemingly illogical elements in the NOFDS. All costs
and all benefits are not incorporated into these analyses (General Accountability Office 2003,
Heinzerling and Ackerman 2002). These analyses fail to recognize many important
12 5
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
considerations, uncertainties and projected future developments (Government Accountability
Office 2003).
Because of the multitude of recognized deficiencies presently incorporated into
traditional Corps costbenefit analyses, flawed information is provided to policy makers to help
them make wise decisions regarding provision of financial resources to develop an adequate
NOFDS. Within the executive branch, the Office of Management and Budget has the
responsibility to ensure the quality of costbenefit analyses and resource recommendations, yet
the deficiencies were not effectively addressed.
This study indicates that many of the flaws that were introduced into the NOFDS came
from flawed decision making regarding provision of financial resources by many organizations
at many levels, times, and places. The exceedingly complex and flawed organizational system
and its decisions regarding provision of resources that evolved during development of the
NOFDS was a primary cause of the failure of the NOFDS. The Corps of Engineers does have
and use advanced methods to evaluate costs, benefits, and risks for flood damage reduction
studies and dam safety (National Research Council 1983; U.S. Army Corps of Engineers 1996;
Powers 2005). Application of these advanced methods was not, however, in evidence in the
background available on development of the NOFDS. The substantial body of technology
developed to assist risk management decision making (e.g., Fischhoff et al. 1981; Wenk 1989;
Shapira 1995; Molak 1997; Kammen and Hassenzahl 1999; Spouge 1999; Moteff 2004; Jordan
2005; see Appendix H) should be further developed, codified and applied by the Corps to assist
policy makers in making decisions regarding development and maintenance of levees and flood
protection systems.
For many years, the Corps of Engineers has been subjected to extreme pressures at the
federal and state levels to do more with less (Government Accountability Office 1997; Office
and Management and Budget 2006); to do their projects better, faster, and cheaper; and improve
project management (planning, organizing, leading, controlling). The organization's attempts to
respond to all of these frequently conflicting pressures has introduced organizational turbulence
and diversion of attention and resources that continues the present time. The Corps of Engineers
developed a plan to reengineer itself (U.S. Army Corps of Engineers, 2003b). However, it is
clearly struggling with all of its constraints to achieve key elements in this plan (Office of
Management and Budget 2006). Our study indicates that as in the case of NASA (Appendix F)
technical and engineering superiority and oversight was compromised in attempts to respond to
all of these constraints and pressures; especially those pressures for increased efficiency and
decreased costs. Adequate quality and reliability in the constructed works has suffered and will
continue to suffer until these challenges are successfully addressed.
Evidently the organizations responsible for the various parts of the lifecycle of the
NOFDS did not have effective process auditing procedures (Knoll 1986). They did not have
incentive systems that discouraged excessive and inordinate risk taking that could lead to less
than desirable quality and reliability. Quality standards did not meet or exceed the referent
standards required for a high quality and reliable NOFDS. These organizations did not correctly
assess the risks associated with given problems or situations; apparently, situational awareness
was frequently lost. These organizations lacked strong command and control systems as
evidenced with appropriate rules and procedures, effective selection and training of personnel,
decisions being made in the right ways at the right times by the right people, effective
redundancy (robustness) to create tolerance to organizational defects, and maintenance of
situational awareness for appropriate action. In general, these organizations performed as Low
12 6
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Reliability Organizations (Weick and Sutcliffe 2001). Effective leadership and management was
lacking (Townsend 2006; Collins 2005; Government Accountability Office 2006). Such
organizational malfunctions were summarized by Irons (2005):
The evidence indicates the U.S. Army Corps of Engineers knew about the threat of
breaches, as opposed to overtopping, since the early 1980s. Moreover, all
concerned agencies, including those at the local, state, and federal levels, knew
about the threat of overtopping and consequent flooding in even a Category 3
hurricane.
Basic flaws in the design of the levee protection system were first recognized over
two decades ago, before the wetlands were so diminished. An outside contractor,
Eustis engineering, was the first to express concerns about the levee vulnerability
to breaching in the early 1980s. In 1981, the New Orleans Sewerage & Water
Board developed a plan to improve street drainage by dredging the 17th Street
Canal. The Corps of Engineers issued permits to do the dredging in 1984 and
1992, though the Corps was not a partner in the Project. Eustis Engineering
contracted to do a design study for Modjeski and Masters, the consulting
engineers on the project, and performed soil investigations on a section of the 17th
Street Canal from south of the Veterans Memorial Boulevard bridges to just north
of those structures. They found that 'the planned improvements to deepen and
enlarge the canal may remove the seal that has apparently developed on the
bottom and side slopes, thereby allowing a buildup of such pressures in the sand
stratum.' Eustis' concerns about a 'blow-out', or breach, of the levee were strong
enough that the company recommended test dredging before the final design.
…The most puzzling point about the dredging project is that the Corps of
Engineers planned to follow the project by raising the floodwall from 10 feet to
14.5 feet. It is unclear whether the Corps paid attention to the contractor's
concerns since most of the documents related to the work remain unavailable to
the public. Although the Corps of Engineers was not a direct partner in the
dredging, it was aware of the work and knew it would have an impact on its later
project. Indeed, contractors working for the Corps on the later project raised
their own concerns about the soil and foundations of the levee.
Reports indicate that key sections of the levee system's soil and foundation,
particularly the floodwall on the 17th Street Canal where much of the serious
flooding occurred, posed serious problems for the contractors involved. Court
papers from 1998 show that Pittman Construction indicated to the Corps of
Engineers as early 1993 that the soil and foundation for the walls were 'not of
sufficient strength, rigidity and stability' to build on. The construction company
claimed that the Corps of Engineers did not provide it with complete soil data
when it developed a bid on the levee project.
…Engineers now say the difficulties Pittman Construction faced were early
warning signs that the Corps of Engineers ignored. The Corps of Engineers
officially disputed the points made by Pittman Construction regarding the soil
condition, though it now seems clear that the crucial breaches in New Orleans
occurred in levees where the floodwall foundations were not as deep as the canals
and that the Corps of Engineers was aware of the issue.… Would an organization
12 7
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
with processes in place to support ongoing learning, and surprise-avoidance, fail
to recognize the legitimacy of the contractor's point rather than argue about
purely budgetary issues related to the contract?
Principal knowledge related malfunctions centered on inappropriate use of existing
technology (“unknown knowables”) and inadequate measures to disclose unknowns throughout
the lifecycle of the NOFDS. Examples include the subsidence and settlements of critical flood
protection elements in the NOFDS including those of the floodwalls along the drainage canals
(which are now known to be about two feet below intended elevations), the Industrial Canal
floodwalls (about three feet below intended elevations), and the levee elevations along the
Mississippi River Gulf Outlet (MRGO) that front the south side of Lake Borgne (about two to
three feet below intended elevations) (Interagency Performance Evaluation Task Force, 2006a,
2006b). Concerns regarding settlements and subsidence were expressed early in the
development of the NOFDS, but apparently no effective action was taken to quantify the
regional subsidence and settlements and to make appropriate adjustments to the NOFDS. Even
though information was developed by the National Geodetic Survey that the reference
benchmarks being used as controls in construction of the NOFDS were in excess of one foot low,
the decision was made in August 1985 to use the benchmarks “current at the time of construction
of the first increment of the project” (1965) (Chatry 1985). The report of the Senate Committee
on Homeland Security and Governmental Affairs observed (2006):
In Designing, constructing and maintaining the hurricane-protection system the
Corps did not adequately address: (a) the effects of local and regional subsidence
of land upon which the protection system was built; and (b) then-current
information about the threat posed by storm surges and hurricanes in the region.
Another important example of knowledge development and utilization malfunctions is
that of the overtopping and breaching of the levees and flood control structures along the MR
GO. Of particular importance is the stretch of levee that defends St. Bernard Parish between the
bayou Bienvenue and bayou Dupre flood control structures. This stretch of levee was badly
damaged during hurricane Katrina as were sections where the levee joined the flood control
structures (Seed et al., 2005). The current work of the ILIT indicates that the sections adjacent to
the flood control structures breached where the construction had covered the original bayou
channels. The design of the junctions between the flood control structures and the earth levee
were not sufficient to withstand the surge and wave action developed during hurricane Katrina.
In a similar manner, the levees between bayous Dupre and Bienvenue were not able to withstand
the waves and surge that developed across lake Borgne; they were severely breached and
massively eroded. The ILIT indicates that the wave and surge velocities that preceded the arrival
of the peak surge likely were initiating breaching before these levees were overtopped so that
when the levees were overtopped, the breaches could be readily expanded and allow large
volumes of water to enter the protected areas in St. Bernard parish. In contrast, the performance
of the levee north of bayou Bienvenue to its intersection with the GIWW was markedly different.
Our studies indicate that this performance resulted from a combination of factors that included
superior soils used in construction of this stretch of levee (highly erosion resistant) and natural
protection (water velocity reduction) afforded by adjacent wetlands on the outboard side of the
levee.
The MRGO is a 76mile long navigation channel connecting the Gulf of Mexico to the
Port of New Orleans Inner Harbor Navigation Canal via the GIWW. The channel bisects the
marshes of lower St. Bernard Parish and the shallow waters of Chandeleur Sound. Construction
12 8
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
of this shipping channel/canal was authorized by Congress in 1956. Its construction was started
in 1958 and completed in 1965. Many people contend that the MRGO played a prominent role
developing the flooding of St. Bernard parish and East New Orleans during hurricane Betsy
(1965). Before, during, and after construction of the MRGO (48 years) many concerns were
expressed regarding the effects of this canal on the adjacent wetlands and on its potential
focusing effect on storm surge propagation into the IHNC. Originally conceived as a way to get
deep draft ships to the Port of New Orleans facilities in the IHNC, it failed to realize its
commercial justification because of changes in ships (deeper drafts) and because of the need for
almost continuous dredging to keep the channel open. The channel also allowed the highly saline
waters from the Gulf of Mexico to intrude into the adjacent fresh and brackish water wetlands
and marshes, destroying many in the process (estimated more than 20,000 acres of marsh have
been destroyed). In 1988, the St. Bernard Parish Council unanimously adopted a resolution to
close MRGO because it constituted a threat to public health and safety. In October 2004, the
Louisiana Legislature passed a resolution urging closure of the MRGO and immediate
implementation of remedial measures to address the risk posed by the MRGO.
Available information and soil sampling conducted during this study indicates that the
levees between bayous Bievenieu and Dupre originally were constructed from dredge spoil
deposited during the construction of the MRGO (A. Theis, personal communication, January
2006) (see Chapters 6 and 9). The USACE’s own design documentation states that the materials
used are potentially susceptible to erosion (USACE, DM___, 19__). Additional construction
was proposed to increase the height of the levee at the time of hurricane Katrina. While these
materials were highly susceptible to scour and erosion, the ILIT study has failed to discover
documentation of plans or proposals for armoring this levee prior to hurricane Katrina.
Given the recognized degradation of the protection afforded by wetlands (Hallowell
2001), recognition of the erodability of the levee soils, the lack of provision of protection for the
levee soils, recognized deficiencies in the design criteria, the continued challenges of keeping
these levees at their authorized elevations (significant subsidence and compression), and the
repeated expressions of concerns for the adequacy of these protective works, the performance of
this part of the NOFDS was a “predictable surprise.” The Member Scholars of the Center for
Progressive Reform (2005) arrived at similar conclusions.
Rejection and misuse of technology are evident in the history of the NOFDS. Interactive
risk assessment and management approaches (e.g., Quality Assurance and Quality Control)
(Knoll 1986; Loosemore 2000) to help detect, analyze, and correct knowledge related challenges
apparently failed for a wide variety of reasons including excessive authority gradients, low task
and situational awareness, excessive professional courtesy, culturalsocietal morays, excessive
beliefs, deficiencies in communications, and deficiencies in resource and task management. Irons
observed (2005):
The U.S. Army Corps of Engineers is historically an insular agency, known for doing
things its own way. It is not possible to say whether surprise-avoidance processes
are in place at the Corps of Engineers, until the public receives more access to
internal documents. The failure of Corps' staff to recognize and prioritize the
challenges of levee upgrades and receding wetlands to the city of New Orleans, and
surrounding areas, strongly suggests that surprise-conducive processes characterize
its organization. The Corps' organization has over the past few decades outsourced
more work, lost many engineers to private industry, and consequently suffered a
diminished capacity to attract top-notch engineers.
12 9
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
New Orleans had dodged the bullet many times, with the major force of hurricanes
skirting around the area. Nevertheless, most people with a reason to know about it
were aware that a Category 3 hurricane posed a severe threat to the New Orleans'
levee protection system, and a Category 5 hitting land as a Category 4, as with
Katrina, posed a catastrophic threat.
The occurrence of a hurricane like Katrina was not unexpected in New Orleans;
neither were the complications faced in the aftermath of the storm. Given this
understanding, and the neglect in preparing for a hurricane like Katrina, as well as
the ineffective response preparations, it seems reasonable to assert that Katrina as
well as its aftermath was a predictable surprise. The threats posed by the hurricane,
and the likely aftermath, were well known and unsurprising to most who thought
about the hurricane threat to New Orleans. Unfortunately, much of the local, state,
and federal leadership, especially the U.S. Army Corps of Engineers, appears to
have remained complacent about preparing the levees for a catastrophic hurricane.
All of these Extrinsic factors represent corporate failures in making decisions that
involved all components of the TDS, including the public. The right things were tradedoff for
the wrong things at the wrong times and in the wrong ways. The failure of the NOFDS has all of
the same ingredients found in previous catastrophic failures and accidents (Appendix F). It
involved many different people and organizations developing a wide variety of malfunctions
(e.g., decisions) over a long period of time (40 years). While a majority of these malfunctions
were embedded during the concept and design phases, early warnings that indicated ‘all was not
well’ as the NOFDS progressively developed were not detected, analyzed, and corrected. When
hurricane Katrina finally tested the flawed, defective, and deficient NOFDS, it failed
catastrophically producing the single most catastrophic failure of an engineered system in the
history of the United States.
12.5
Intrinsic Factors
Intrinsic factors representing natural variability and analytical modeling uncertainties also
played roles in the failure of the NOFDS (Vick 2002; Bea 2006). There are fundamental flaws in
the basic criteria and guidelines that were used to design the NOFDS. These flaws include
engineering elements that address:
ク
ク
Design demands for the elements of the NOFDS; including the Standard Project Hurricane
(SPH) conditions (surge heights in the NOFDS, frequency of occurrence, and lack of explicit
recognition of the likely effects of more intense hurricanes) (Select Bipartisan Committee to
Investigate the Preparation for and Response to Hurricane Katrina 2006, ASCE 2006a). Even
though studies after 1972 indicated the need for increases in the design flood protection
elevations due to greater surge and wave heights, these were not reflected in revised design
guidelines (Brouwer 2003; Carter 2005a, 2005b).
Design capacities for the elements of the NOFDS; including engineering guidelines used to
design and construct the levees and floodwalls (e.g., analyses of levee stability, levee
stability factors of safety, analyses of floodwall/sheetpile stability, deformation and stresses,
floodwall design factors of safety, provision for deformations in the floodwalls during surge
loading, provisions for robustness defect and damage tolerance and failsafe performance,
and provisions for subsidence) (Select Bipartisan Committee to Investigate the Preparation
12 10
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
ク
for and Response to Hurricane Katrina 2006; Vartabedian and Braun 2006; Irons 2005;
ASCE 2006b).
Configuration of the elements that comprise the NOFDS as an integrated flood defense
system (Seed et al. 2005, ASCE 2006a). Many failures of the NOFDS occurred at a variety of
types of interfaces in the physical elements such as interfaces between earth levees and
concrete and steel flood protection elements and between the flood control structures and
pump station structures (Carter 2005a). Flood discharge pumps were not sufficiently
protected from backflow and exacerbated flooding. Many vulnerabilities were found at
transitions and interfaces between flood protection elements and/or where other infrastructure
elements were involved (Seed et al. 2005). The NOFDS was not an integrated, coherent
system; rather “it is a jointed series of individual pieces conceived and constructed
piecemeal” (ASCE 2006a).
12.5.1 Standard Project Hurricane
The heart of most Corps of Engineers hurricane protection projects since the 1960s has
been the Standard project Hurricane, or SPH (Carter 2005a, Government Accountability Office
2005a, 2005b). The SPH was developed by the National Weather Service and the Corps of
Engineers at the request of Congress in the 1950s “to provide generalized hurricane
specifications that are consistent geographically and meteorologically for use in planning,
evaluating and establishing hurricane design criteria for hurricane protection works” (Grahan and
Nunn 1959). Attempts to describe the SPH in terms of Categories has lead to confusion because
the SPH preceded development of the SaffirSimpson hurricane scale (Categories). Depending
on what characteristic of a hurricane is referenced, the SPH for the NOFDS can vary from a
Category 2 to a Category 4 storm.
A primary goal of the SPH was to compare hurricane protection standards from region to
region (Perdikis 1967). This standardized approach led to disparities within a particular region.
The SPH model excluded storms that were deemed to be inordinately severe. For example, the
1979 revision of the SPH removed two particularly severe storms from the data base; hurricane
Camille of 1969 and the Labor Day Hurricane of 1935. Experience shows that excluding outlier
data is not appropriate in the context of dealing with extreme hazards. In addition, a higher
standard of protection was specified for facilities and areas “where high winds, waves and storm
surge could pose a threat to the public health and safety from a hurricaneinduced accident at a
nuclear power plant” (Schwert et al 1979). This shows that the SPH criteria includes an implicit
costbenefit assessment. This implicit assessment prevents policymakers (and the public they
represent) from determining whether an extreme event is worth guarding against by excluding
the possibility that such an event will or can occur. The following quotations indicate the
interpretations that developed through the history of development of the NOFDS regarding what
the system’s SPH represented.
ク “The Standard Project Hurricane wind field and parameters represent a
‘standard’ against which the degree of protection finally selected for a hurricane
protection project may be judged and compared with protection provided at
projects in other localities.” (Graham and Nunn 1959).
ク “The project is designed to protect against the Standard Project Hurricane
moving on the most critical track. Only a combination of hydrologic and
12 11
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
meteorological circumstances anomalous to the region could produce higher
stages. The probability of such a combination of occurring is, for all practical
purposes, nil.” (U.S. Army Corps of Engineers 1974).
ク “The SPH is a steady state hurricane having a severe combination of values of
meteorological parameters that will give high sustained wind speeds reasonably
characteristic of a specified coastal location. By reasonably characteristic is meant
that only a few hurricanes of record over a large region have had more extreme
values of the meteorological parameters.” (National Weather Service 1979).
ク “The SPH was expected to have a frequency of occurrence of once in about
200 years, and represented the most severe combination of meteorological
conditions considered reasonably characteristic for the region.” (Government
Accountability Office 2005).
As can be seen, over time the SPH went from being a general indicator of threat levels to
a guarantee of safety. The methods used to define the SPH were buried, along with their potential
flaws and questionable assumptions. Because it became the “gold standard” of flood system
performance, the SPH served to prevent uptodate reanalysis of the true risks of catastrophic
flooding of the NOFDS.
Recent work indicates that the probability that a hurricane will pass within 75 miles of
New Orleans in any given year is about 12.5 percent, or about once every eight years (URS
2005). The likelihood of a major hurricane (Category 3 and above) are about 3.2 percent per
year, or about once every 30 years. These projections do not account for the current period of
intensified hurricane activity (Klotzbach and Gray 2006). Thus, the history of development and
evolution of the SPH did not provide an adequate basis to understand the risks associated with
catastrophic flooding of the NOFDS. The report of the Senate Committee on Homeland Security
and Governmental Affairs observed (2006):
For several years, the Corps has inaccurately represented to state and local
officials and to the public the level of protection that the hurricane system
provided. The Corps claimed the system protected against a fast-moving Category
3 storm even though: (a) there was no adequate study or documentation to
support this claim; and (b) information known to or provided to the Corps
demonstrated that the claim was not accurate.
Some industries (e.g. offshore engineering) that must deal with hurricane related hazards
have developed specific design conditions (e.g., wave or surge height) and associated forces
based on specified annual return periods (e.g., 100 years) (Bea 1990, 1998, 2001). These design
conditions are chosen based on their potential effects (e.g., forces, water surface elevations) on
the structures to be designed. The design conditions and prescribed forces are supplemented with
safety factors (e.g., 2 to 4) that help assure that the resulting system can perform acceptably in
much more intense conditions (frequently identified as Ultimate Limit State conditions) (Bea
1990). For important industrial facilities, these Ultimate Limit State conditions have return
periods in the range of 1,000 to 10,000 or more years (Vick 2002; Baecher and Christian 2003,
U.S. Army Corps of Engineers 1999, Tekie and Ellingwood 2003). For example, typical modern
offshore structures in the Gulf of Mexico that are evacuated in advance of hurricanes are
designed to be able to resist forces from hurricanes that have return periods of more than 1,000 to
5,000 years (Bea 1996). Structures that cannot be evacuated are designed to be able to resist
forces from hurricanes and storms that have return periods in the range of 5,000 to 10,000 years
12 12
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
(Spouge 1999). In a similar vein, the Dutch currently provide protection against flooding of the
Netherlands for events that represent the worst storm that could be expected to affect the area
with return periods in the range of 1,000 to 10,000 years (Versteeg 1988, Netherlands Water
Partnership 2005).
The SPH evolved to represent the most severe storm the government should guard
against when designing hurricane protection projects. The SPH came to represent not only a
method for comparative assessment of storm risks between geographic areas, but also a design
standard that carried its own assurance of adequate reliability. For a variety of reasons, the
concept of storms much more intense than the SPH was not allowed to explicitly enter the
engineering process, even though the development of the SPH also involved a Probable
Maximum Hurricane (PMH) (National Weather Service 1979):
The PMH is a hypothetical steady state hurricane having a combination of values
of meteorological parameters that will give the highest sustained wind speed that
can probably occur at a specified coastal location. One of several possible uses
of the values of meteorological parameters is to compute maximum storm surge
at coastal points when the hurricane approaches along the most critical track
[authors’ emphasis].
Thus, it was clearly recognized that the SPH did not represent a maximum set of
conditions for design against hurricane conditions. It is also clear that general public was not
informed about the flooding risks that the selection of the SPH as a basis for design implied. In
many cases, even though very inexpensive defenses could have been provided for the potential
for hurricane surges exceeding those of the SPH (e.g., splashpads behind Iwalls and other
similar floodwalls sensitive to overtopping erosion), these defenses were not provided.
Another important element of the SPH was that it was not revised as knowledge
improved after the 1960s. Authorization constraints and engineering restraints were provided to
us as an explanation for this (bureaucratic engineering). Tremendous strides in the meteorology
and oceanography of hurricanes were made during the 1970’s, and these improvements in
technology continue to evolve to the present time (Simpson 2003, Interagency Performance
Evaluation Task Force 2006a, 2006b). However, the SPH remains essentially the same as it was
when it was initially defined in the 1950s and early 1960s. The natural variability in hurricane
conditions, and the ability of these conditions to exceed the design norms of the SPH, and how
these norms were translated in design resulted in many of the failures observed in the NOFDS in
the wake of hurricanes Katrina and Rita.
12.5.2 Failure Modes and Safety Factors
A primary obligation of an engineer is to anticipate failure modes in the element,
component, or system being engineered and then provide measures to prevent those failure
modes from developing or from developing catastrophic results (Petroski 1985, 1994; Harr 1987;
Wenk 1989, 1995, 1998; Appendix H). This obligation requires two primary things: (1)
anticipation of possible failure modes, and (2) provision of defenses in depth to prevent and/or
mitigate those failure modes.
The design demand for a particular component in the NOFDS, when combined with a
prescribed safety factor and associated analytical models and procedures, determines the
Ultimate Limit Strength of that element. When combined with an assessment of the intrinsic
uncertainties (natural, model, parametric, state), the ratio of the Ultimate Limit Strength to the
12 13
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
design demand (factor of safety) reflects the reliability (or probability of failure) associated with
that component (Bea 1990; Kulhawy 1996; Duncan 2000; Whitman 2000; Vick 2002; Christian
2004; Lacasse 2004).
For example, a factor of safety of 1.3 was specified by the Corps of Engineers as the
minimum acceptable safety factor for drainage canal levee lateral stability for the “transient”
loading conditions represented by hurricaneinduced storm surge and waves (U.S. Army Corps
of Engineers 1988, 1989, 1990, 2000). Our examination of the basis for this factor of safety
indicates that it was developed in the 1950s for levees used primarily to defend sparsely
populated agricultural areas (Wolff, 2005). This factor of safety is embodied in current Corps of
Engineers guidelines for the design of levees and assessment of slope stability (U.S. Army Corps
of Engineers 2000, 2003). In the case of the drainage canal levees and those along the IHNC, and
the floodwalls constructed on and in these levees, the design demand was determined by the total
lateral force represented by the canal water level determined on the basis of the SPH (U.S. Army
Corps of Engineers 1994).
In the 1990s the Corps of Engineers developed very advanced analytical methods to
assess the reliability of important flood control components such as levees and dams (U.S. Army
Corps of Engineers 1996, 1999, National Research Council 1983). These methods were validated
with field and laboratory test data and field performance data (U.S. Army Corps of Engineers
1999; Wolff 1999; Duncan 2000). Application of these methods entailed an assessment of the
inherent uncertainties for different failure modes and identification of target reliabilities for these
modes (Wolff, 1999). Analytical methods were developed for both elements and assemblies of
elements that represented flood control components and systems. The issues associated with
‘target’ or acceptable reliabilities were also addressed. These methods were used to define
reliabilitybased design and maintenance factors of safety. Application of these methods for
important flood protection facilities defending highly populated areas and the Corps of Engineers
levee stability analysis procedures indicated the need for factors of safety that substantially
exceeded those actually used for the levees and associated floodwalls that defended the NOFDS
drainage canals. A need for “Factors of safety” of 2 to 3 and greater were indicated for very
important facilities (annual target Safety Indices in the range of 3 to 4). Similar safety factors
were identified by other investigators for similar facilities (Bea 1990, Duncan 2000, Whitman
2000, Vick 2002, Christian 2004, Lacasse 2004). Apparently a technology lag (breakdown in
technology transfer) or rejection of technology (Sowers 1993) developed and persisted in the
design guidelines used for levees and floodwalls in the NOFDS. As a result, standard design
Factors of Safety were inadequate, and overall system reliability was compromised as a direct
result.
Following Hurricane Katrina a similar technology lag was identified as one of the causes
of the failure at the 17th Street canal. Both the ILIT (Seed et al. 2005) and the Corps of
Engineers Interagency Performance Evaluation Task Force analyses (Interagency Performance
Evaluation Task Force 2006b) of this failure concluded that a failure mode developed that was
not recognized by the designers. This finding lead to the official contention that this was a
“design failure.” The information developed by the ILIT clearly indicates that this failure was a
result, not a cause.
This failure mode involved lateral deflection of the concrete floodwall and the sheet piles
that supported that floodwall. This deflection resulted in separation between the stiff supporting
sheet piling and the soft soil of the levee on the outboard side (flood side) of the wall. Water was
then able to enter the gap and exert additional lateral forces against the lower portions of the
12 14
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
sheetpiles and thus on the remaining (inboard) ‘half’ of the levee (and floodwall. Now the levee
only had about ‘half’ of its width and mass able to transmit the lateral forces to the underlying
soils. This combination resulted in lowering the lateral resistance with a commensurate lowering
of the factor of safety.
This development was incorrectly reported as “unforeseen and unforeseeable” by the
Interagency Performance Evaluation Task Force on March 10, 2006 (Marshall 2006; Seed and
Bea 2006). In 1985, the New Orleans district of the Corps of Engineers conducted a full scale
instrumented lateral load test of a 200foot long sheet pile / flood wall in the Atachafalaya basin
[the E99 sheetpile test] (U.S. Army Corps of Engineers 1988b). This particular location (south of
Morgan City, Louisiana) was chosen because of the close correlation of the soil conditions in the
New Orleans area with those at the test location. “The foundation soils are relatively poor,
consisting of soft, highly plastic clays, and would be representative of near worst case conditions
in the NOD (New Orleans District).” (U.S. Army Corps of Engineers 1988b).
Test data from the highly instrumented sheet pile wall and adjacent supporting soils
indicated a gapping behavior (separation of the sheet piles from the soils). The test was designed
to take an eight foot height of water (above the supporting ground level) with a factor of safety of
1.25. But, the wall was already in a failure condition (increasing lateral displacements with no
increase in loading) when the water level reached only 8 feet instead of the calculated 10 feet.
Strain gage readings on the sheet piles indicated that they were well below the steel yield point,
thus the yielding had to have been developing in the supporting soils. Two very important pieces
of information developed by the E99 sheet pile tests were that there was potential soil separation
from the sheet piles (allowing water to penetrate below the ground surface between the piles and
the soils) and that the calculated safety factor was not reached (it was overestimated due to
unanticipated deformations in the soils).
Additional reports and professional papers further developed the experimental
information and advanced analytical models that could be used to help capture such behavior
(U.S. Army Corps of Engineers Waterways Experiment Station 1989b). Later developments in
this work confirmed the gapping between the sheetpiles and the outboard side of the levee
embankment, and were published in USACE reports and were eventually reported in the open
professional engineering literature by Oner, Dawkins and Mosher (1997):
As the water level rises, the increased loading may produce separation of the soil
from the pile on the flooded side (i.e., a “tension crack” develops behind the
wall). Intrusion of free water into the tension crack produces additional
hydrostatic pressures on the wall side of the crack and equal and opposite
pressures on the soil side of the crack. Thus part of the loading is a function of
system deformations.
These developments in technology inexplicably were not reflected in the design
guidelines used (U.S. Army Corps of Engineers 1988a, 1989a, 1990). We also found no evidence
that questions regarding the adequacy of the design were raised after the design and construction
were completed. Loss of corporate memory, breakdowns in technology transfer, and abilities to
keep the design guidelines current with existing knowledge seemed to background these
developments.
A second suspect element in the development of the failure at the 17Th Sreet Canal
regarded characterizations of the soils that supported the earth levee and sheet piling in the
vicinity of the 17th Street canal breach. The processes used at the time of design to analyze the
12 15
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
soil types and engineering characteristics did not capture the unique characteristics of the soils.
Soil strengths based on samples from beneath the crest of the levee, with higher strengths
resulting from higher overburden loads and thus compression of these soils to a denser state,
were inappropriately used to characterize the strengths of the soils at and beyond the toes of the
levees (where lower overburden loads resulted in lower strengths). In addition, the spatial
averaging process (vertical and lateral) did not capture the unique soil characteristics in the
vicinity. Soils in Southern Louisiana and other parts of the Gulf Coast have very complex
histories due to past floods, hurricanes, the rise and fall of sea level, changes in vegetation, and
other events. Far from being uniform, they contain complicated and rapidly varying strata of
different materials with very different characteristics.
In 1964 1965 the Corps ran a full scale levee test in the Atachafalaya basin in which
advanced studies were conducted regarding characterizations of the soil strengths and
performance – stability characteristics of the levee (U.S. Army Corps of Engineers 1968;
Kaufman and Weaver 1967). The levee test sections were thoroughly instrumented and their
performance monitored during and after construction. Various analytical methods were used to
evaluate the usefulness and reliability of the various methods. These developments clearly
indicated the need to understand the geologic soil depositional processes and the associated
variations in soil strengths (horizontal and vertical) in order to understand the performance and
stability characteristics of levees. The importance of local soil conditions to performance of the
levee was clearly pointed out. The tremendous importance of overburden loads on soil strengths
was a major focus of this work; and led to awardwinning advancement of the Stress History and
Normalized Engineering Performance (SHANSEP) framework for evaluation and modeling of
the strengths of these types of soils. Additional reports and professional papers were published
that resulted in significant advances to the engineering knowledge (Duncan 1970, Ladd et al.
1972; Edgers et al. 1973; Foott and Ladd 1973, 1977). None of these vital principals, however,
were subsequently incorporated in the design and analysis of the 17th Street canal levees and
floodwalls.
Indepth background and understanding of the geologic and depositional environment
and history of vital importance to understanding the characteristics of the Mississippi Basin soils
were developed in the 1950s and 1960s (Fisk et al. 1952; Kolb and Van Lopek 1958; Krinitzsky
and Smith 1969) and the Corps of Engineers led in the development of this background. Of
particular importance was recognition that the marsh and swamp deposits were “treacherous”
and highly variable. It was repeatedly pointed out that “careful and detailed characterization of
the soil properties was required.” Further, the studies cited above led to the recognition that the
methods based on traditional Corps of Engineers soil characterization and stability analyses gave
factors of safety that were unconservative (too large); (Foot and Ladd 1977). As in the first
instance, these developments in technology inexplicably were not reflected in the design
guidelines and practices that were used in the actual design studies.
The safety factors used in design were not sufficient to accommodate the uncertainties
inherent in the design procedures and processes and inherent in the environment in which the
facility would exist. Important failure modes in the components were not recognized. When the
system was tested, it failed because of a confluence of intrinsic and extrinsic uncertainties. This
was not a design failure; this was a failure on the part of the organizations responsible for the
design and construction of the flood defense works to effectively use proven technology.
12 16
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
12.6
Life-Cycle Development of Flaws
Sources of flaws in the NOFDS developed during the lifecycle of the system starting
with its concept (e.g., SPH), then during design (e.g., Iwall configurations, strength and stability
guidelines, factors of safety), construction (e.g., normalized reports of excavation and forming
instabilities and seepage from canals) and operation (e.g., persistent reports of leakage from
canals and signs of ground instability), and finally during the maintenance (e.g., inground
construction, vegetation growth on and adjacent to levee toes) phases (Select Bipartisan
Committee to Investigate the Preparation for and Response to Hurricane Katrina 2006, Irons
2005). Similar lifecycle flaws were developed and propagated in the levee and flood protection
structures adjacent to the MRGO. Important flaws in the NOFDS were embedded in every stage
of the life-cycle. In many cases, these flaws were allowed to propagate and magnify. Early
warning signs were ignored or were ineffectively addressed. NOFDS component interface flaws
that developed throughout the lifecycle of the NOFDS were particularly evident.
When the NOFDS was challenged by hurricane Katrina, these flaws became evident. Had
these flaws not been present, it is likely that hurricane Katrina would not have developed into a
major catastrophe.
Design challenges not successfully addressed were traced to fundamental flaws that
became embedded in engineering design procedures and how these procedures were used. Tests
were performed and the results not properly utilized, and in several key cases, not utilized at all
(Seed and Bea 2006). Even though procedures for other similar facilities (e.g., dams, coastal and
offshore structures) existed and were highly developed, the design (also construction, operation,
inspection, maintenance, and repair) technology was not integrated into the design of the
NOFDS (rejection or misuse). In addition to flaws previously discussed, the design procedures
focused on individual components, with insufficient treatment given to the concepts of integrated
system performance, defenses in depth, and robustness (damage and defect tolerance). The
Member Scholars of the Center for Progressive Reform arrived at similar conclusions in their
report titled An Unnatural Disaster: The Aftermath of Hurricane Katrina (2005).
12.7
Findings - Looking Back
Failure of the NOFD was not caused by an overwhelming extreme natural event
(hurricane wind, waves, currents, surge). While portions of the NOFDS were overtopped by
hurricane Katrina's surge and waves, our studies indicate that the majority of the flooding
came from unanticipated and unintended breaches in the levees (many adjacent to other
structures), failures in the floodwalls, and water entering through gaps (floodgates not in
place) or low spots in the NOFDS. The roots of these unanticipated and unintended
developments were firmly embedded in Technology Delivery System flaws and malfunctions;
failures of organizations - institutions and their resource allocation processes.
ILIT identified eight categories of technology delivery system (TDS) malfunctions that
played primary roles in the failure of the NOFDS. Additional background on each of these TDS
malfunctions is provided in Appendices F and H.
Failures of foresight: Catastrophic flooding of the greater New Orleans area due to
surge from an intense hurricane had been predicted for several decades (Townsend 2006). The
consequences observed in the wake of hurricane Katrina were also predicted (Members Scholars
of the Center for Progressive Reform 2005). The hazards associated with the NOFDS were not
12 17
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
adequately recognized, defensive measures were not identified and prioritized, and effective
action was not mobilized to effectively deal with these hazards (Irons 2005; Senate Committee
on Homeland Security and Governmental Affairs 2006).
Failures of organization: The roots of the failure of the NOFDS are firmly embedded in
flawed organizational institutional systems (Select Bipartisan Committee to Investigate the
Preparation for and Response to Hurricane Katrina 2006). The organizational institutional
systems lacked centralized and focused responsibility and authority for providing adequate flood
protection (Government Accountability Office 2005a, 2005b; Carter 2005a, 2005b; ASCE
2006a; Senate Committee on Homeland Security and Governmental Affairs 2006). Dramatic
and pervasive failures in management existed, exemplified by ineffective and inefficient
planning, organizing, leading, and controlling to achieve desirable quality and reliability in the
NOFDS (Houck 2006, Braun and Vartabedian 2005). There were extensive and persistent
failures to demonstrate initiative, imagination, leadership, cooperation, and management
(Leonard and Howitt 2006).
Failures of funding: The failure of the NOFDS resulted in part from inadequate
provision of resources based primarily on recommendations provided by the Corps followed by
failure of the federal and state governments to fund badly needed improvements once limitations
were recognized (Members Scholars of the Center for Progressive Reform 2005; Houck 2006;
Braun and Vartabedian 2005). In several instances, State agencies pressured for 'lower cost'
solutions not realizing that these solutions would result in lowering the overall quality and
reliability of the NOFDS (Members Scholars of the Center for Progressive Reform 2005).
Important deficiencies existed in the costbenefit analyses used to justify the levels of protection
and their continued improvement as knowledge and technology advanced (Government
Accountability Office 2003, 2005; Heinzerling and Ackerman 2002).
Failures of diligence: Forty years after the devastating flooding caused by hurricane
Betsy, the flood protection system authorized in 1965 and based on the Standard Project
Hurricane (SPH) was still not completed (Government Accountability Office 2005a, 2005b). The
concept and application of the SPH was recognized to be seriously flawed, yet no adjustments
were made to the system before Katrina struck (Select Bipartisan Committee to Investigate
Preparation for and Response to Hurricane Katrina 2006). Early warning signs of deficiencies
and flaws persisted throughout development of the different components that comprised the
NOFDS and these signs were not adequately evaluated and acted upon (Houck 2006; Carter
2005a, 2005b).
Failures of trade-offs: A history of flawed decisions and tradeoffs proved to be fatal to
the ability of the system to perform adequately (Carter 2005a, 2005b). Compromises in the
ability of this system to perform adequately started with the decisions regarding the fundamental
design criteria for the development of the system, and were propagated through time as
alternatives for the system were evaluated and engineered (Houck 2006). Design, construction,
operation, and maintenance of the system in a piecemeal fashion allowed the introduction of
additional flaws and defects (Collins and Lieberman 2005). Efficiency was traded for quality,
reliability, and effectiveness. Superiority in provision of an adequate NOFDS was traded for
mediocrity and getting along (Collins 2005; Senate Committee on Homeland Security and
Governmental Affairs 2006).
Failures of management: Requirements imposed on the Corps of Engineers by
Congress, the White House, State and local agencies, and the general public have changed
12 18
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
dramatically during the past three decades. Defense, reconstruction, maintenance, waste
disposal, recreational, emergency response, and ecological restoration have served to divert
attention from flood control (Office of Management and Budget 2006, Vartabedian and Braun
2006). Public and Congressional pressures to reduce backlogs of approved projects, improve
project and organizational efficiency (downsizing, outsourcing), address environmental impacts
and develop appropriations for projects have served to divert attention from engineering quality
and flood control reliability (Carter and Sheikh 2003). Engineering technology leadership,
competency, expertise, research, and development capabilities appear to have been sacrificed for
improvements in project planning and controlling (Office of Management and Budget 2006;
Senate Committee on Homeland Security and Governmental Affairs 2006).
Failures of synthesis: While individual parts of a complex system can be adequate, when
these parts are joined together to form an interactive interdependent adaptive system,
unforseen failure modes can be expected to develop (Rasmussen 1997; Bea 2000). These
unforseen, but forseeable, failure modes developed in the NOFDS during hurricane Katrina. It is
evident that insufficient attention was given to creation of an integrated series of components to
provide a reliable NOFDS (ASCE 2006a). Synthesis was subverted to decomposition. As a
result, many failures developed at interfaces or 'joints' in the NOFDS (Committee on New
Orleans Regional Hurricane Protection Projects 2006; Seed et al. 2005).
Failures of risk assessment and management: The risks (likelihoods and
consequences) associated with hurricane surge and wave induced flooding were seriously
underestimated (Carter 2005a, 2005b). There was inadequate recognition of the primary
contributors to the likelihoods and consequences of catastrophic flooding. Sufficient defensive
measures to counteract and mitigate these uncertainties were not used. Safety factors used in
design of the primary elements in the NOFDS were insufficient (ASCE 2006a, 2006b). Quality
assurance and control measures invoked during the life of the system failed to disclose critical
flaws in the system (Vartabedian and Braun 2006). Inappropriate use was made of existing
engineering technology available to design, construct, operate, and maintain a NOFDS that
would have acceptable quality and reliability. Deficient risk management methods were used to
allocate resources and impel action to properly manage risks (Moteff 2004). Risk management
failed to employ continuing improvement, monitoring, assessment, and modifications in means
and methods which were discovered to be ineffective (Senate Committee on Homeland Security
and Governmental Affairs 2006).
12.8
References
ASCE External Review Panel (2006a). External Review Panel Progress: Report Number 1,
Letter to LTG Carl A. Strock, Chief, U.S. Army Corps of Engineers, Washington, DC.
ASCE External Review Panel (2006b). External Review Panel Progress: Report Number 2,
Letter to LTG Carl A. Strock, Chief, U.S. Army Corps of Engineers, Washington, DC,
Mar. 23, 2006.
Baecher, G. B., and Christian, J. T. (2003). Reliability and Statistics in Geotechnical
Engineering, Wiley, Chichester, UK.
12 19
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Bea, R. G. (1990). Reliability Based Design Criteria for Coastal and Ocean Structures. National
Committee on Coastal and Ocean Engineering, The Institution of Engineers, Australia,
Barton ACT.
Bea, R.G. (1996). “Reassessment and Requalification of Infrastructure: Application to Offshore
Structures.” J. Infrastructure Systems, (2)2, 2232.
Bea, R. G. (1998). “Reliability Characteristics of a Platform in the Mississippi River Delta.” J.
Geotech. and Geoenv. Engrg., 124 (8), 729738.
Bea, R. G. (2000). “Performance Shaping Factors in Reliability Analysis of Design of Offshore
Structures.” J. Offshore Mechanics and Arctic Engineering, (122) 1.
Bea, R. G. (2001). “Risk Assessment and Management of Offshore Structures.” Prog. Struct.
Engrg. & Materials, 3, John Wiley & Sons, Ltd., 180187.
Bea, R. G. (2006). “Reliability and Human Factors in Geotechnical Engineering.” J.
Geotechnical and Geoenvironmental Engineering, 132 (5), 631643.
Braun, S., and Vartabedian, R. (2006). “The Politics of Flood Control, Levees Weakened as New
Orleans Board, Federal Engineers Feuded.” Los Angeles Times, 12/25/06.
Brouwer, G. (2003). “The Creeping Storm,” Civil Engineering, ASCE, Reston, VA.
Carter, N. T. (2003). Army Corps of Engineers: Civil Works Reform Issues in the 107th
Congress, Congressional Research Service Report for Congress, Washington DC, April.
Carter, N. T. (2005a). Water Resources Development Act (WRDA): Army Corps of Engineers
Authorization Issues in the 109th Congress, Congressional Research Service Report for
Congress, Washington DC, March.
Carter, N. T. (2005b). Corps of Engineers Reform in WRDA 2005, Congressional Research
Service Report for Congress, N. T. Carter, Washington DC, April.
Carter, N. T. (2005c). New Orleans Levees and Floodwalls: Hurricane Damage Protection,
Congressional Research Service Report for Congress, Washington DC, September.
Carter, N. T. (2005d). Flood Risk Management: Federal Role in Infrastructure, Congressional
Research Service Report for Congress, Washington DC, October.
Carter, N. T., and Sheikh, P. A. (2003). Army Corps of Engineers Civil Works Program: Issues
for Congress, Congressional Research Service Report for Congress, Washington DC,
May.
Carter, N. T., and Hughes, H. S. (2005). Army Corps of Engineers Water Resources Activities:
Authorization and Appropriations, Congressional Research Service Report for Congress,
Washington DC, February.
Carter, N. T. et al. (2005). Army Corps of Engineers Civil Works Program: Issues for the 109th
Congress, Congressional Research Service Report for Congress, Washington DC, June.
Chatry, F. M. (1985). “NGS Benchmarks.” Letter to Commander, Lower Mississippi Valley
Division, U. S. Army Corps of Engineers from New Orleans District Corps of Engineers,
New Orleans.
12 20
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Christian, J. T. (2004). “Geotechnical Engineering Reliability: How Well Do We Know What
We Are Doing?” J. Geotechnical and Geoenvrionmental Engineering, 130(10), 985
1003.
Coastal Environments (1984). The Mississippi River Gulf Outlet: A Study of Bank Stabilization,
Report to St. Bernard Parish Police Jury, National Oceanic and Atmospheric
Administration and State of Louisiana Department of Natural Resources, Baton Rouge.
Collins, S. M. (2005). “Hurricane Katrina: Who's In Charge of the New Orleans Levees?”
Opening Statement, Homeland Security and Governmental Affairs Committee,
Washington, DC, 12/15/2005.
Collins, S., and Liberman, J. (2005). Senate Homeland Security Committee Holds Hurricane
Katrina Hearing to Examine Levees in New Orleans, Press Release, Washington DC,
Nov. 2.
Duncan, J.M. (1970). Strength and stress-strain behavior of Atchafalaya foundation soils.
Research Report TE701, Department of Civil Engineering, University of California,
Berkeley.
Duncan, M. J. (2000). “Factors of Safety and Reliability in Geotechnical Engineering.” J.
Geotech. and Geoenv. Engrg, 126 (4), 307316.
Edgers, L. et al (1973). Undrained creep of Atchafalaya levee foundation clays. Research Report
R7316, Soils Publication No. 319, Dept. of Civil engineering, Massachusetts Institute of
Technology, Cambridge MA.
Fischoff, et al. (1981). Acceptable Risk, Cambridge University Press, UK.
Foott, R., and Ladd, C. C. (1973). The behavior of Atchafalaya test embankments during
construction. Research report R7327, Dept. of Civil Engineering, Massachusetts
Institute of Technology, Cambridge MA.
Foott, R., and Ladd, C. C. (1977). “Behaviour of Atchafalaya levees during construction.”
Geotechnique, 27(2), 137160.
Government Accountability Office (1976). Cost, Schedule, and Performance Problems of the
Lake Pontchartrain and Vicinity, Louisiana, Hurricane Protection Project, Report to the
Congress, Washington DC, Aug. 31.
Government Accountability Office (1983). Water Project Construction Backlog - A Serious
Problem With No Easy Solution, Report to Committee on Public Works and
Transportation, House of Representatives, GAO/RCED8349, Washington DC.
Government Accountability Office (1997). Results Act: Observations on the U.S. Army Corps of
Engineers Civil Works Program's Draft Strategic Plan, Report to Committee on
Transportation and Infrastructure, House of Representatives, GAO/RCED9813R,
Washington DC.
Government Accountability Office (2003). Corps of Engineers, Improved Analysis of Costs and
Benefits Needed for Sacramento Flood Protection Project, Report to Congressional
Requesters, GAO0430, Washington DC.
Government Accountability Office (2005). Army Corps of Engineers, Improved Planning and
Financial Management Should Replace Reliance on Reprogramming Actions to Manage
12 21
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
funds, Subcommittee on Energy and Water Development, Committee on Appropriations,
House of Representatives, GAO05946, Washington DC.
Government Accountability Office (2005a). Hurricane Protection: Statutory and Regulatory
Framework for Levee Maintenance and Emergency Response for the Lake Pontchartrain
Project, Washington DC.
Government Accountability Office (2005b).Army Corps of Engineers, History of the Lake
Pontchartrain and Vicinity Hurricane Protection Project, GAO06244T, Washington,
DC.
Government Accountability Office (2005c). Army Corps of Engineers, Improved Planning and
Financial Management Should Replace Reliance on Reprogramming Actions to Manage
Project Funds, Report to Subcommittee on Energy and Water Development, Committee
on Appropriations, House of Representatives, Washington, DC, Sept.
Government Accountability Office (2006). Preliminary Observations Regarding Preparedness
and Response to Hurricanes Katrina and Rita, GAO06365R, Washington, DC.
Grahan, H. E. and Nunn, D. E. (1959). Meteorological Considerations Prtinent to Standard
Proejct Hurricane, Atlantic and Gulf Coasts of the United States, National Hurricane
research Project, U. S. Weather Bureau, Report No. 33, Washington, DC.
Harr, M. E. (1987). Reliability-Based Design in Civil Engineering, McGrawHill, New York.
Heinzerling, L. and Ackerman, F. (2002). Pricing the Priceless, Georgetown Environmental Law
and Policy Institute, Georgetown University Law Center, Washington DC.
Hallowell, C. (2001). Holding Back the Sea, The Struggle on the Gulf Coast to Save America,
Harper Perennial.
Houck, O. (2006). “Can We Save New Orleans?” Tulane Environmental Law Journal, 19(1),
<http://www.law.tulane.edu/tuexp/journals> (May 1, 2006).
Interagency Performance Evaluation Task Force (2006a). Performance Evaluation Status and
Interim Results, Report 1 of a Series, Performance Evaluation of the New Orleans and
Southeast Louisiana Hurricane Protection System, U. S. Army Corps of Engineers,
Washington, DC.
Interagency Performance Evaluation Task Force (2006b). Performance Evaluation Status and
Interim Results, Report 2 of a Series, Performance Evaluation of the New Orleans and
Southeast Louisiana Hurricane Protection System, U. S. Army Corps of Engineers,
Washington, DC.
Irons, L. (2005). “Hurricane Katrina as a Predictable Surprise.” Homeland Security Affairs, 1(2),
<http://www.hsaj.org> (April 1, 2006).
Jordan, I. (2005). Decisions Under Uncertainties, Probabilistic Analysis for Engineering
Decisions, Cambridge University Press.
Kammen, D., and Hassenzahl, D. (1999). Should We Risk It? Exploring environmental, Health,
and Technological Problem Solving, Princeton University Press.
Kaufman, R. I., and Weaver, F. J. (1967). “Stability of Atchafalaya levees.” J. Soil Mechanics
and Foundations Division, 93(4), 157176.
12 22
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Klotzbach, P.J. and Gray, W.M. (2006). Extended Range Forecast of Atlantic Seasonal
Hurricane Activity and U.S. Landfall Strike Probability for 2006, Dept. of Atmospheric
Science, Colorado State University.
Knoll, F. (1986). “Checking Techniques.” Modeling Human Error in Structural Design and
Construction, A. S. Nowak, ed, 2642.
Kolb, C.R., and Van Lopik, J.R. (1958). Geology of the Mississippi River Deltaic Plain,
Southern Louisiana. Technical Report No. 3483, U.S. Army Engineer Waterways
Experiment Station, Vicksburg, MS.
Krinitzsky, E. L., and Smith, F. L (1969). Geology of backswamp deposits in the Atchafalaya
basin, Louisiana. Technical Report S698, U.S. Army Engineer Waterways Experiment
Station, Vicksburg, MS.
Kulhawy, F. (1996). “From Casagrande's Calculated Risk to ReliabilityBased Design in
Foundation Engineering.” Civil Engrg. Practice, BSCE, 11(2), 4356.
Lacasse, S. (2004). “Risk Assessment for Geotechnical Solutions Offshore.” Proc. 23rd Int.
Conf. Offshore Mechanics and Arctic Engrg., OMAE200451144, ASME.
Ladd, C. C. et al. (1972). Engineering properties of soft foundation clays at two south Louisiana
levee sites. Research Report R7226, Dept. of Civil Engineering, Massachusetts Institute
of Technology, Cambridge MA.
Leonard, H.B. and Howitt, A.M. (2006). “Katrina as Prelude; Preparing for and Responding to
Future KatrinaClass Disturbances in the United States.” Testimony before the U.S.
Senate Homeland Security and Governmental Affairs Committee, Washington, DC.
Loosemore, M. (2000). Crisis Management in Construction Projects, ASCE Press, Reston, VA.
Marshall, R. (2006). “Floodwall failure was foreseen, team says.” Times Picayune, New Orleans,
LA, 03/14/2006, <http://www.nola.com> (May 1, 2006).
Members Scholars of the Center for Progressive Reform (2005). An Unnatural disaster: The
Aftermath of Hurricane Katrina, Center for Progressive Reform, Report Publication No.
512, <http://www.centerforprogressivereform.org> (May 1, 2006).
Molak, V. (Ed) (1997). Fundamentals of Risk Analysis and Risk Management, CRC Lewis
Publishers, New York.
Motef, J. (2004). Risk Management and Critical Infrastructure Protection: Assessing,
Integrating, and Managing Threats, Vulnerabilities and Consequences, Congressional
Research Service Report for Congress, Washington DC, 9/2/2004.
National Academy of Engineering (2006). Letter to Honorable John Paul Woodley, Assistant
Secretary of the Army, Civil Works, Committee on New Orleans Regional Hurricane
Protection Projects, Washington DC, Feb. 21.
National Research Council (1983). Safety of Existing Dams, Evaluation and Improvement,
Committee on Safety of Existing Dams, Commission on Engineering and Technical
Systems, National Academies Press, Washington, DC.
National Weather Service (1979). Meteorological Criteria for Standard Project Hurricane and
Probable Maximum Hurricane Windfields, Gulf and East Coasts of the United States,
NOAA Tech. Report NWS 23, Charleston, SC.
12 23
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Netherlands Water Partnership (2005). Dutch Expertise Water Management & Flood Control.
Delft, The Netherlands, <http://www.nwp.nl> (Dec. 1, 2005).
Office of Management and Budget (2006). Agency Scorecards, Corps of Engineers - Civil
Works, The Executive Office of the President, The White House,
<http://www.whitehouse.gov/omb> (May 1, 2006).
Oner, et al (1997). “SoilStructure Interaction Effects in Floodwalls.” Electronic Journal of
Geotechnical Engineering, <http://www.ejge.com/1997> (Jan,1, 2006).
Perdikis, H. S. (1967). “Hurricane Flood Protection in the United States.” J. Waterways and
Harbors Div., ASCE, Reston, VA.
Petroski, H. (1985). To Engineer is Human: The Role of Failure in Successful Design, St.
Martins Press, New York, NY.
Petroski, H. (1994). Design Paradigms, Case Histories of Error and Judgment in Engineering,
Cambridge University Press, Cambridge, UK.
Powers, K. (2005). Aging Infrastructure: Dam Safety, Congressional Research Service Report
for Congress, Washington DC, September 2005.
Rasmussen, J. (1997). “Risk Management in a Dynamic Society: A Modeling Problem.” Safety
Science, 27 (2), Elsevier, UK.
Schwert, et al. (1979). Meteorological Criteria for Standard Project Hurricane and Probable
Maximum Hurricane Windfields, Gulf and East Coasts of the United States, National
Weather Service, Washington, DC.
Seed, R. B., and Bea, R. G. (2006). Initial Comments on Interim (70%) IPET Study Report,
National Science FoundationSponsored Independent Levee Investigation Team (ILIT),
University of California, Berkeley, Mar. 12, 2006.
Seed, R. B., et al. (2005). Preliminary Report on the Performance of the New Orleans Levee
Systems in Hurricane Katrina on August 29, 2005, Report No. UCB/CITRIS05/01,
Berkeley, CA, <http://risk.berkeley.edu> (Nov. 17, 2005).
Select Bipartisan Committee to Investigate the Preparation for and Response to Hurricane
Katrina (2006). A Failure of Initiative. U.S. Government Printing Office, Washington,
DC, <http://www.gpoaccess.gov/congress/index.html> (Mar. 15, 2006).
Senate Committee on Homeland Security and Governmental Affairs (2006). Hurricane Katrina:
A Nation Still Unprepared, Washington DC.
Sevedung, I., and Rasmussen, J. (2002). “Graphic Representation of accident scenarios: mapping
system structure and the causation of accidents.” Safety Science, 40(1), Elsevier, UK.
Shapira, Z. (1995). Risk Taking, A Managerial Perspective, Russell Sage Foundation, New York.
Simpson, R. (Ed) (2003). Hurricane! Coping with Disaster, Progress and Challenges Since
Galveston, 1900. American Geophysical Union, Washington DC.
Sowers, G. F. (1993). “Human Factors in Civil and Geotechnical Engineering Failures.” J.
Geotechnical Engineering, 109(2).
Spouge, J. (1999): A Guide to Quantitative Risk Assessment for Offshore Installations,
Publication 99/100 CMPT 1999, ISBN I 870553 365, London, UK.
12 24
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Tekie, P. B., and Ellingwood, B. R. (2003). “Perspectives on probabilistic risk assessment of
concrete gravity dams.” Applications of Statistics and Probability in Civil Engineering,
Der Kiureghian, Madanat and Pestana (Eds.), Millpress, Rotterdam.
Townsend, F. F. (2006). The Federal Response to Hurricane Katrina, Lessons Learned, Report
to The President of the United States, The White House, Washington, DC.
URS Corporation (2005). Interim Report: Hydrodynamic Modeling Effort for MRGO Study,
Report to State of Louisiana Department of Natural Resources, Baton Rouge, LA.
U.S. Army Corps of Engineers, New Orleans District (1968). Field tests of levee construction,
test sections I, II, and III, EABPL, Atchafalaya Basin Floodway, Louisiana. Interim
Report, New Orleans, LA.
U.S. Army Corps of Engineers (1974). Final Environmental Statement, Lake Pontchartrain,
Louisiana and Vicinity Hurricane Protection Project I-2, New Orleans District, New
Orleans, LA.
U.S. Army Corps of Engineers (1988a). Lake Pontchartrain, LA., and Vicinity Lake
Pontchartrain High Level Plan, Design Memorandum No. 19 Orleans Avenue Outfall
Canal. Three Volumes, New Orleans District, New Orleans, LA.
U.S. Army Corps of Engineers (1988b). E-99 Sheet Pile Wall Field Load Test Report. Technical
Report No. 1, U.S. Army Engineer Division, Lower Mississippi Valley, Vicksburg, MS.
U.S. Army Corps of Engineers (1989a). Lake Pontchartrain, LA., and Vicinity Lake
Pontchartrain High Level Plan, Design Memorandum No. 19A London Avenue Outfall
Canal. Two Volumes, New Orleans District, New Orleans, LA.
U.S. Army Corps of Engineers (1989b). Development of Finite-Element-Based Design
Procedure for Sheet-Pile Wall. U.S.Army Corps of Engineers Waterways Experiment
Station, Tech. Report GL8914, Vicksburg, MS.
U.S. Army Corps of Engineers (1990). Lake Pontchartrain, LA., and Vicinity Lake Pontchartrain
High Level Plan, Design Memorandum No. 20, General Design, Orleans Parish,
Jefferson Paris, 17th. Outfall Canal (Metairie Relief). Two Volumes, New Orleans
District, New Orleans, LA.
U.S. Army Corps of Engineers (1994). Design of Sheet Pile Walls. Engineer Manual EM 11102
2504, <http://www.usace.army.mil/inet/usacedocs> (Apr. 1, 2006).
U.S. Army Corps of Engineers (1996). Risk-Based Analysis for Flood Damage Reduction
Studies. Engineer Manual EM 111021619, <http://www.usace.army.mil/inet/usace
docs> (Apr. 1, 2006).
U.S. Army Corps of Engineers (1999). Risk-Based analysis in Geotechnical Engineering for
Support
of
Planning
Studies.
Engineer
Manual
EM
111021902,
<http://www.usace.army.mil/inet/usacedocs> (Apr. 1, 2006).
U.S. Army Corps of Engineers (2000). Design and Construction of Levees. Engineer Manual EM
111021913, <http://www.usace.army.mil/inet/usacedocs> (Apr. 1, 2006).
U.S. Army Corps of Engineers (2003a). Slope Stability. Engineer Manual EM 111021902,
<http://www.usace.army.mil/inet/usacedocs> (Apr. 1, 2006).
12 25
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
U.S. Army Corps of Engineers (2003b). USACE 2012, Aligning the U.S. Army Corps of
Engineers for Success in the 21st Century, Washington, DC.
Vartabedian, R. and Braun, S. (2006). “Unfinished 1965 Project Left Gaps.” Los Angeles Times,
1/17/2006.
Versteeg, M.F. (1988). “External Safety Policy in the Netherlands: an Approach to Risk
Management.” J. of Hazardous Materials, Elsevier 17, Amsterdam, The Netherlands.
Vick, S. G. (2002). Degrees of Belief, Subjective Probability and Engineering Judgment, ASCE
Press, Reston, VA.
Weick, K. E., and Sutcliffe, K. M. (2001). Managing the Unexpected, JosseyBass, San
Francisco, CA.
Wenk, E., Jr. (1989). Tradeoffs: Imperatives of Choice in a High Tech World, The Johns
Hopkins University Press, Baltimore, MD
Wenk, E., Jr. (1995). Making Waves: Engineering, Politics and the Social Management of
Technology, University of Illinois Press, Urbana, IL.
Wenk, E., Jr. (1999). The Double Helix: Technology and Democracy in the American Future,
Ablex Publishing Corp., Stamford CT.
Whitman, R. V. (2000). “Organizing and Evaluating Uncertainty in Geotechnical Engineering.”
J. Geotech. Engrg, 126 (7), 583592.
Whitman, R. V. (1984). “Evaluating Calculated Risk in Geotechnical Engineering.” J. Geotech.
Engrg, 110 (2), 145188.
Wolff, T. (1999). Evaluating the Reliability of Existing Levees. ETL 11102556, U.S. Army
Corps of Engineers, <http://www.usace.army.mil/inet/usacedocs> (Mar. 1, 2006).
Wolff, T. (2005). “Factors of Safety for Levee Design”, Personal Communication.
12 26
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
CHAPTER THIRTEEN: ORGANIZED FOR SUCCESS
The excuse we have heard from some government officials throughout this
investigation, that Katrina was an unforeseeable ultra-catastrophe, has not only
been demonstrated to have been mistaken, but also misses the point that we
need to be ready for the worst that nature or evil men can throw at us. Powerful
though it was, the most extraordinary thing about Katrina was our lack of
preparedness for a disaster so long predicted.
This is not the first time the devastation of a natural disaster brought about
demands for a better, more coordinated government response. In fact, this
process truly began after a series of natural disasters in the 1960s and into the
1970s. One of those disasters was Hurricane Betsy, which hit New Orleans in
1965. The similarities with Katrina are striking: levees overtopped and
breached, severe flooding, communities destroyed, thousands rescued from
rooftops by helicopters, thousands more by boat, and too many lives lost.
Katrina revealed that this kaleidoscope of reorganizations has not improved our
disaster management capability during these critical years. Our purpose and
our obligation now is to move forward to create a structure that brings
immediate improvement and guarantees continual progress. This will not be
done by simply renaming agencies or drawing new organizational charts. We
are not here to rearrange the deck chairs on a ship that, while perhaps not
sinking, certainly is adrift.
This new structure must be based on a clear understanding of the roles and
capabilities of all management agencies. It must establish a strong chain of
command that encourages, empowers, and trusts frontline decision-making. It
must replace ponderous, rigid bureaucracy with discipline, agility, cooperation,
and collaboration. It must build a stronger partnership among all levels of
government with the responsibilities of each partner clearly defined, and it must
hold them accountable when those responsibilities are not met.
Senator Susan Collins
Opening Statement
Committee on Homeland Security and Government Affairs
Hurricane Katrina: Recommendations for Reform
Washington DC, March 8, 2006
13 1
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
We are Doomed to an Unacceptable Future - Unless …
What do the following accidents have in common? Torrey Canyon tanker (1967) and the
Exxon Valdez tanker (1989); the U.S.S. Greeneville (2002) and the U.S.S. San Francisco (2005);
the Challenger Space Shuttle (1986) and the Columbia Space Shuttle (2003); the Piper Alpha
Platform (1988) and the Petrobras P36 Platform (2001), Herald of Free Enterprise (1987) and the
Estonia ferry (1994), and failure of the NOFDS in the wake of hurricane Betsy (1965) and
hurricane Katrina (2005).
In each case someone, somewhere, understood that organizational and system processes
were as much the cause of the accident as were engineering design, construction and
maintenance errors (Appendix F). In each case this knowledge failed to prevent a second disaster
from happening in the same industry. This record suggests that we are doomed to a future in
which increasingly complex organizations and systems of organizations fail causing unnecessary
death and injury, large scale economic disruption, political haggling, and years of rebuilding.
We are doomed to this future despite growing evidence that preventing disasters is
always cheaper than recovery. We are doomed to this future despite the fact that we know that
technological failures virtually always occur within the context of management failures, and
there is a growing body of literature that describes management implementations designed to
reduce large scale failure (e.g. Roberts and Bea 2001a; 2001b; Dekker 2002; Weick and Sutcliffe
2001).
As an example of what doesn’t happen, the National Incident Management System
(NIMS) Integration Center issued this alert (Department of Homeland Security 2006):
All federal, state, local, tribal, private sector and non-governmental personnel
with a direct role in emergency management and response must be NIMS and ICS
trained. This includes all emergency services related disciplines such as EMS,
hospitals, public health, fire service, law enforcement, public works/utilities,
skilled support personnel, and other emergency management response, support
and volunteer personnel….
In mid March, 2005, Donald Hiett, Jr, Principal, Organizational Strategic Solutions
Group, was asked by Louisiana State University (LSU) to develop a NIMS training program
directed to the senior executive leadership in New Orleans to take place before June, 2006. On
March 28, 2005 he was informed there was no interest by these officials in taking this training
program (Hiett, personal communication).
We are doomed – unless. This chapter deals with “unless.” It first discusses the
assessment of safety, and the usual engineering responses to risk. It then asks that the reader
adopt a new perspective regarding the USACE and the contextual issues it needs to consider. It
then discusses preventing the “next Katrina”, and offers recommendations.
13 2
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
13.1
How Safe is Safe Enough?
The hurricane Katrina catastrophe exposed a technological failure of inadequate defenses
against a predictable, risky and potentially lethal event. Recent studies have tended to focus
primarily on death and destruction from flood waters released by collapse of the NOFDS.
Studies of cause acknowledge the extreme forces of nature, but also cite human and
organizational errors (HOE) that now occur more conspicuously because the engineering
parameters are fairly well understood. HOE failures far exceed mechanical sources in the overall
Katrina catastrophe.
Because protection against human weaknesses is more art than science, the study of the
causes and remediation of HOE require a context for risk analysis. Nonspecialists with policy
and management responsibilities should be helped by a perspective that points to the systems
based and interdisciplinary requirements for the NOFDS. Such a perspective can help us answer
the enigmatic question, “How Safe is Safe Enough?” In other words, what level of risk is
acceptable when making decisions about public safety and security?
Risk is usually defined as a condition in which either an action or its absence poses
threats of socially adverse and sometimes extreme consequences. Risk happens from acts of
nature, from weaknesses of human nature, and from side effects of technology, all situations that
mix complex technical parameters with the variables of social behavior. Although each risk
event is unique, all display commonalities that permit systemic analysis and management. These
recurring properties lead to certain principles.
To begin, the acceptability of risk cannot be extracted from science or mathematics; it is a
social judgment. The spectrum of risk thus embraces both the physical world defined by natural
laws, and the human world loaded with beliefs instead of facts, and with values, ambiguities and
uncertainties. Among other features, the physical world may be thought of as a mechanism
whose behavior follows principles of causeandeffect. The human world performs more like an
organism whose components are not fixed but may grow, and which may be altered by the thrust
of events and their interplay with other elements.
Following a notion that what you can’t model you can’t manage, a systems model is
needed to represent the processes by which both physical and societal factors are defined,
interconnected and interact. Such technologybased human support systems are labeled by their
intended social functions: food production, shelter, military, homeland security, etc. In our
modern era, these and other functions are enormously strengthened by applications of scientific
knowledge, applied through engineering.
It helps to think of technology as more than the hardware of planes, trains and computers.
Rather, it is a social system comprising many organizations, synchronized by a web of
communications for a common purpose. It is energized by forces of free market demand, of
popular demand for security and quality of life, and by forces of scientific discovery and
innovation. It is best understood as a Technological Delivery System (TDS) that applies scientific
knowledge to achieve society’s needs and wants.
Technology then acts like an amplifier of human performance. Like the water wheel, the
steam engine and the bomb, it amplifies human muscle. With the computer it amplifies the
human mind and memory. It also amplifies social activity, mobility, quality and length of life.
A paradox arises when technologies introduced for specific benefits also spawn side
effects. These can induce complexity, conflict and even chaos. Most of these are unwanted by
13 3
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
some sector of stakeholders, now or in the future. This paradox is dramatized when technologies
are introduced to defend against violence of nature, or against human and organizational error,
but themselves spring unintended and possibly dangerous consequences.
The investigation of risk and of measures to contain it within safe limits requires both
hindsight and foresight. The past can illuminate failures, their causes and their control as lessons
for engaging new issues and threats. The future commands the exercise of foresight, an
imaginative preparation of scenarios stirred by such questions as, “what might happen, if ?” or
“what might happen, unless?” Those inquiries should then examine the timing of impacts
(immediate or hibernating), the identities of players on the risk horizon who may trigger risk, and
those parties responsible for risk abatement and those who may be adversely affected now or in
the future.
Modeling then becomes essential to represent a full cast of stakeholders and their inter
relationships, including both the private and the public sectors. The concept of a technology
delivery system (TDS) is simply an attempt to model how the real world works.
The responsibility to manage risk stems from the American Constitution, from custom,
and from a growing body of public law. Federal, state, and local governments are heavily
involved in all of the technologies previously discussed and many more. With waterways, for
example, the Army Corps of Engineers (USACE) has a predominant statutory responsibility.
That accords with the historic federal stewardship of national infrastructure, from roads, shipping
channels, harbors and canals to airplane routes and the Internet.
That achievement carries significant but subtle implications. For one thing, safety costs
money. The federal budget is constantly challenged to meet a rainbow of different demands, the
total of which always exceeds Congressional appropriations. The mismatch must then be
reconciled through tradeoffs at the highest policy levels stretching all the way to the President of
the United States and the Congress.
Often, a focus on power of the Federal Government misses a major premise of
democratic governance. As the Declaration of Independence states, those who govern should do
so only with the consent of the governed; we would say the informed consent. This notion is
reflected in such regulatory legislation as the National Environmental Policy Act (NEPA),
Section 102(2) c. It requires estimates of harm that could result from technological initiatives,
along with alternatives to accomplish the same goals but with less harm. After preparation, these
environmental impact statements (EIS) are made available for public comment and possible
amendment. The point is that this process makes every citizen a part of government process to
negotiate the question of how safe is safe enough and thus provide citizens the levels of safety
and security that they desire.
Implied is a prospective national policy that those put in harm’s way have a voice in what
otherwise could be involuntary exposure to risk. This principle leaves implementation of the
concept to the responsible federal agencies, subject to Constitutional safeguards. Despite a
tendency to flare the sensational, the media can enrich understanding with a backstory because
disasters so agitate a functioning system as to reveal the full cast of stakeholders, their roles in
increasing or decreasing risk and their degree of injury.
In this modern era, society demands better protection against threats to life, peace,
justice, health, liberty, lifestyle, private property and to the natural environment. These
challenges are not new, but two things have changed—the increased potency of technology and
13 4
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
increased media coverage. Technological factors are more robust in speed of delivery and in
potential harm. Media covers events live, 24/7, and worldwide. Events anywhere have
repercussions everywhere. The better informed public tends increasingly to be risk averse.
Apprehension and fear peak after a calamity with demands for better protection through better
governance. Higher expectations are legitimate because so many threats just itemized are due to
human and organizational errors either in catering technologies to meet market demand or in
guarding against hazards. This current study shows that the Katrina event fits that pattern.
Government at all levels failed to provide security to citizens before and during the catastrophic
flooding. Victims are justified in asking how this pathology of a mundane levee technology
developed; How can that knowledge be applied to prevent a reoccurrence?
13.1.1 The Engineering Response to “How Safe is Safe Enough?”
The engineering profession has long practiced social responsibility by a technique of
overdesign, to compensate for uncertainties in loading, in materials, in quality of construction
and maintenance, etc. This may be accomplished by adopting some multiple of loading as a
margin of safety ranging from 1.4 to 5.0 and even greater. How these margins are set, and by
whose authority, is of critical importance; especially where tradeoffs with cost or other
compelling factors such as deadlines may compromise the intended reduction of risk.
This method of safety assurance is more applicable to design of mechanisms not subject
to human and organizational errors. The term “errors,” incidentally, is shorthand for a broad
spectrum of individual and societal weaknesses that include ignorance, blunder, folly, mischief,
pride, lack of competence, greed and hubris. Protecting structures against violence of nature
such as with earthquakes, volcanic eruptions, tsunamis, floods, landslides, hurricanes, pestilence,
droughts and disease may utilize the concept of overdesign, based on meteorological,
hydrological, seismic and geophysical data of past extreme events.
Learning from documented failures is a powerful method for reducing risks of repeated
losses. Another method is to learn from close shaves. Many dangerous events fortunately
culminate in only an incident rather than an accident, but the repetition of similar incidents can
serve as early warnings of danger. Indeed, the logging and analysis of such events on the nation’s
airways partially accounts for commercial aviation’s impressive safety record. A system for
reporting close encounters of aircraft was installed decades ago. Anticipating the possibility that
perpetrators of high risk events might be reluctant to blow the whistle on themselves, many years
ago the Federal Aviation Administration arranged for NASA to collect incident data and to
sanitize it to protect the privacy of the incident reporter. NASA also screened reports to identify
patterns as early warning of dangerous conditions. Similar systems are in place for reporting
nuclear power plant incidents.
With the growing recognition of human factors in accidents or in failures to limit
damage, a class of situations entailing uncommonly high risks but conspicuously good safety
records was examined. In the Navy, for example, high risks are a part of daily operations of
submarines and aircraft carriers. Yet accident rates are paradoxically low. Careful analysis of
these situations showed that certain qualities of leadership and organizational culture foster
integrity, a sense of responsibility among all participants, a tolerance by authority figures for
dissent, and consensus on common goals of safe performance. High safety performance is
associated with an institutional culture that is bred from the top of the management pyramid. The
most critical element of that culture is mutual trust among all parties (e.g., Roberts 1990).
13 5
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Long experience with military and paramilitary organizations such as first responders
proves the value of rehearsals to reduce risks and control damage. Of special virtue is proof of
satisfactory communications. Evaluation of dry runs has repeatedly turned up serious problems
in communication. So has postaccident analysis of real events when delays or blunders in
communication of warnings and rescue operations cost lives.
13.1.2 Insights from Addressing These Issues
To sum up, the context for analyzing the levee failures from Hurricane Katrina illustrates
several realities. The most compelling imperative of life is survival. Yet the experience of living
teaches that there is no such thing as zero risk. Some exposures must be tolerated as “normal,”
whether in rush hour traffic or when coping with nature, with human nature or with unintended
consequences of technology. The preceding situation analysis opens a window on a number of
issues treated in more detail in subsequent sections and Appendix H, including the following:
ク
ク
ク
ク
ク
ク
ク
ク
ク
ク
ク
ク
The design of precautionary measures requires inspired foresight, to imagine or foresee
alternative futures.
Tradeoffs are inevitable between short and longrange events and consequences, between
safety and cost, between special interests and social interests, between who wins and who
loses, and who decides.
All human support systems entail technology, and all technologies project unintended
consequences.
Society embraces a spectrum of values that often conflict, as with the goals of efficiency in
the private sector and of sustainability and social justice in the public
Key decisions regarding citizen safety and security are made by government through public
policies to manage risk. These policies dominate the legislative agenda.
This mandate imposes a heavy burden on the President and on Congress, both bodies
requiring access to authentic and immediate information.
Making decisions and assuring implementation draws on political capital in the structure of
authority by the exercise of political power and political will.
In our democracy, this authority should flow from citizens following the principle that those
who govern do so at the informed consent of the governed.
The quality of risk management can best be judged by the effects on future generations.
The geography of risk crosses boundaries between federal, state and local entities, and also
between the United States and other nations.
Different cultures have different risk tolerances, including attitudes distinguishing voluntary
from involuntary risk.
Analysis of risk and its control extracts lessons from past failures, although the most
catastrophic events are so rare as to often frustrate projections.
This portfolio of issues illustrates the anatomy of risk and the complexity of its
management. They sound a wakeup call for deeper understanding by those responsible for risk
13 6
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
management, and by those attentive citizens who are exposed and are entitled to a voice in the
decision process.
13.2
Maximizing How Safe is Safe in the U.S. Army Corps of Engineers (Context)
We know a finite number of precursors lead to major disasters. But in order to understand
what they are we must place a focal first responder organization into its context. For example,
the Corps of Engineers (USACE) is nested within a large number of organizations that should be
interdependent with one another. The social science literature addresses this problem by using
such concepts as interstices (Grabowski and Roberts 1999), “interdependencies” (Heath and
Staudenmayer 2000) or the “space between” (Bradbury and Lichtenstein 2000; Buber 1970).
Failure to consider the processes that operate both within any one unit and across multiple units
is failure to be ready for the next large scale catastrophe. This discussion focuses on contex,t and
asks the reader to take a new perspective of the Corps of Engineers.
Hurricane Katrina provided an interesting, if devastated setting for understanding what
not to do in a quickly changing potential disaster. The organizational liquefaction that occurred
after the Hurricane (the heart of the disaster as opposed to the storm), laid bare the skeletons of
the organizations that should have had flesh and muscle to respond. It laid bare for the public to
see, not only skeletons, but complete organizational disregard for the interdependences so
necessary to a coordinated response. As Houck (2006) observes:
So What Do We Do? Here is what we know. It is not just the tire, it's the car. And
it's not just the car, it's the driver. Nothing in the system has made a numero uno
priority either of protecting New Orleans from hurricanes or to restoring or even
hanging onto - the Louisiana coast. We have a flood control program, a
navigation program, a permitting program, a coastal management program, a
flood insurance program, a coastal restoration program - just for openers - and
they do not talk to each other. They are riddled with conflicts, basically headless,
basically goal-less, weakened by compromises and refuse outright to deal with
first causes and first needs.
The key phrases here are “and they do not talk to each other” and “They are riddled with
conflicts, headless, basically goalless…and refuse outright to deal with first causes….”
In reaction to the organizational liquefaction that developed during hurricane Katrina the
Senate Committee on Homeland Security and Governmental Affairs recommended (2006):
The Corps and local levee sponsors should immediately clarify and memorialize
responsibilities and procedures for the turn-over of projects to local sponsors,
and for operations and maintenance, including, but not limited to procedures for
the repair or correction of levee conditions that reduce the level of protection
below the original design level (due to subsidence or other factors) and also
emergency response. It must always be clear - to all parties involved - which
entity is ultimately in charge of each state of each project. The Corps should also
provide real-time information to the public on the level of protection afforded by
the levee system. A mechanism should be included for the public to report
potential problems and provide general feedback to the Corps.
13 7
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
13.2.1 The Office of the President, the Congress, and the Corps
Other things happen at interstices. Figure 13.1 shows the Presidential and Congressional
budget requests and Congressional recommendations for Corps of Engineer funding for 1975
through 2005 for the Lake Pontchartrain and vicinity hurricane flood defense projects.
Several hypotheses can be gleaned from this information. First, it appears that while the
president was trying to reduce Corps funding Congress was trying to protect Corps funding.
With the Lake Pontchartrain projects only about sixty percent complete as of 2005 (40 years
after authorization) it may be that Congress, in its wisdom, decided to fund only what it thought
needed to be completed. The graph shows other interesting issues about interdependencies. The
Corps of Engineers is interdependent with both the Office of the President of the United States
and Congress. Congressional members bring pressure to bear on the Corps for new large
projects. Faced with these pressures the Corps, then, defers maintenance. For over a decade
Congress has funded the Corps at higher levels than recommended by the President. The Corps,
then, has to devote time to currying favor with Congress. Currying favor with Congress is not
supposed to be a main task of the Corps.
Yet another interesting hypothesis can be derived from these data. When multiyear
projects are funded annually an interesting dilemma is created for the funded organizations. The
funding oscillation level is at one level, but organizations struggling under that oscillation
oscillate at a higher frequency. It is hypothesized that this is because the funded organization
operates under a considerable amount of ambiguity and uncertainty. This suggests that the
unpredictability of the Congressional process creates unintended and negative consequences for
its funded agencies. The processes and responses to them are both schizophrenic.
This is almost surely the same as the case for NASA. The Columbia Accident
Investigation Board (CAIB) report said (Columbia Accident Investigation Board, 2003):
The White House and Congress must recognize the role of their decisions in this
accident and take responsibility for safety in the future.… Leaders create culture.
It is their responsibility to change it.… The past decisions of national leaders –
the White House, Congress, and NASA Headquarters – set the Columbia accident
in motion by creating resource and schedule strains that compromised the
principles of a high risk technology organization.
Diane Vaughan reports that both economic strain and schedule pressure still exist at
NASA. She notes that it is unclear how the conflict between NASA’s goals and the constraints
upon achieving them will be resolved but that one lesson from Challenger and Columbia is that
system effects tend to reproduce (Vaughan 2005). This also happens to military installations
every time a Base Reallocation and Closing (BRAC) list is formed. From the day of its
publication until the day of decisions, the installations on this list spend considerable time trying
to get off the list, distracting them from their principle tasks.
In the Katrina case, will Congress and the Office of the President take a sweeping look at
their own behaviors in concert with those of the Corps of Engineers? They probably will not
because there is not yet a stated strong incentive for them to do so. One incentive might be that
the cost of cleanup is always more than the cost of prevention. Money is not limitless. But since
we’ve observed many costly past disasters that were not prevented, and many instances in which
they could have been mitigated or prevented, the reality is they probably will do nothing. Thus,
the challenge is to find incentives that will encourage both disaster prevention and emergency
13 8
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
response organizations, from the President on down, to examine their own organizational
skeletons, muscle, and flesh, as well as to look at the “spaces between.”
13.2.2 Additional External Interstices for the Corps
Three additional sorts of interfacing between the USACE and its constituents need to be
thought about. The first are the interfaces mandated by Emergency Support Function # 3 of the
National Response Plan NRP (Department of Homeland Security, 2004a).
ESF #3 is structured to provide public works and engineering-related support for
the changing requirement of domestic incident management to include
preparedness, prevention, response, recovery and mitigation actions. Activities
within the scope of this function include conducting pre- and post-incident
assessments of the public works and infrastructure; executing emergency contract
support for life-saving and life-sustaining services; providing technical assistance
to include engineering expertise, construction management, and contracting and
real estate services; providing emergency repair of damaged infrastructure and
critical facilities; and implementing and managing the DHS/Emergency
Preparedness and Response/Federal Emergency Management Agency
(DHS/EPR/FEMA) Public Assistance Program and other recovery programs.
To accomplish these goals, USACE can draw on the resources 15 federal government
agencies. In addition, state, local and tribal governments are “fully and consistently integrated
into EFS #3 activities.” (Department of Homeland Security, 2004a). All of this occurs, of course,
when an incident or potential incident overwhelms state, local, and tribal capabilities.
The NRP concept of operations states that the DOD/USACE is the primary agency for
providing ESF #3 technical assistance. It further states that close coordination is to be maintained
with federal, state, local, and tribal officials to determine potential need for support. In addition it
spells out the organizational structures for providing support, naming the Interagency Incident
Management Group (IIMG) as the resource for providing oncall subjectmatter experts to
support IIMG activities.
Regional and field level mechanisms of support are clearly defined. ESF #3 activities are
also spelled out and include such processes as:
coordination and support of infrastructure risk and vulnerability assessments,
participation in pre-incident activities, such as pre-positioning assessment
teams,… participation in post-incident assessments of public works and
infrastructures to help determine critical needs and potential workloads,
implementation of structural; and non structural mitigation measures, including
deploying protective measures to minimize adverse effects or fully protect
resources, prior to an incident.
In the wake of hurricane Katrina, neither the USACE nor any other agency was fully
successful in rolling out the NRP. If the integration required by this plan is too difficult for
agencies to implement, then it is the duty of the agencies and their oversight agencies (e.g.:
DOD, DHS, HHS, etc.) to indicate this and to develop strategies to revise the NRP to create a
workable plan and document. Lee Clarke (1999) discusses at length “fantasy plans” and that
looks to be exactly what we have here. Thus, a last word on integration across agencies (Lakoff
2006):
13 9
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
From the vantage of preparedness, the failed response to Hurricane Katrina did
not undercut the utility of “all-hazards” planning. Rather, it pointed to problems
of implementation and coordination. This suggests that in the aftermath of the
event, we are likely see the redirection and intensification of already-developed
preparedness techniques rather than a broad rethinking of the security question.
Given our experiences with accident response, without substantial leadership and
reorganization it is this team’s conclusion that neither comprehensive technical nor social
reforms will likely soon be developed to address future natural or man made catastrophes.
The second set of interfaces that need to be thought about are those created by the Corps’
needs to “outsource” (the hiring of outside, private firms and/or individuals to perform work,
including engineering design and construction.) The requirement for the Corps to do this has
been imposed by the federal government; specifically through the White House Office of
Management and Budget and through Congressional actions. Input from current Corps of
Engineers personnel in multiple settings and private briefings with our team, clearly has
indicated that through outsourcing and diversion of efforts the USACE has lost “engineering”
(Figure 13.2). Core engineering (practicing, research, development) competencies have been
sacrificed to pressures to outsource, to improve project management, and to develop
environmental restoration and mitigation capabilities, all within a finite overall budget and
resources.
Partnering has a number of advantages and disadvantages. Some operational benefits
accrue from partnering. One can learn new things from partners, perhaps through access to best
ofclass processes. Perhaps partnering competitors can learn technology secrets from one
another. Where industry benchmarks aren’t well known, partnering with a competitor can offer
insights on a company’s productivity, quality, and efficiency.
But there are also obvious disadvantages. Lack of control is a critical disadvantage. The
demise of ValuJet, for example, happened because the company outsourced cargo handling to a
company it had no control over in terms of quality standards. In another form of outsourcing,
competitors learn from each others’ operations, which may be detrimental to one or more
partners. Or a “coopetition” (combination of cooperation and competition) may selfdestruct
before the renewal option dates arrive. A new company board for one of the partners may not
approve of the other partner. The strategic aims of partners may change midstream, causing
failure. These are just some of the reasons for outsourcing failures. (Roberts and Wong, 2006).
The Corps needs to examine its partner relationships, asking itself if it has lost too much.
One of the Corps sister agencies in time of chaos, FEMA, has also created problems
through outsourcing its disaster response efforts (Perrow 2005):
For example, when the Nisqually earthquake struck the Puget Sound area in
2001, homes that had been retrofitted for earthquakes and schools with FEMA
funds were protected from high-impact structural hazards. The day of that quake
was also the day that the new president, G. W. Bush, chose to announce that
Project Impact would be discontinued (Holdeman 2005). Funds for mitigation
were cut in half, and those for Louisiana were rejected. Disaster management was
being privatized, with the person who was to be promoted to head the agency,
Michael Brown, saying at a conference in 2001, “The general idea—that the
business of government is not to provide services, but to make sure that they are
provided—seems self-evident to me” (Elliston 2004). The administration tried to
13 10
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
cut federal contribution for large-scale natural disaster expenditures from 75
percent to 50 percent.
13.2.3 The Corps Internal Interstices
Two other organizational processes also result in lost institutional memory and loss of
control. They are downsizing and retirements. Table 13.1 shows that in recent years the USACE
has lost employees. Figure 13.3 shows that the Corps is also losing employees through
retirements. Recently, we were told by a high ranking official of the Corps that during the next 5
years, the Corps expects to loose approximately 40 to 50% of its civilian workforce through
retirements.
In 2002, between 35 and 40 percent of architecture and engineering work was outsourced
to private firms, while all construction projects were outsourced (U.S. Army Corps of Engineers
2002). The simultaneous operation of the three processes (outsourcing, downsizing and
retirement) have been and will be disasters for the Corps. Retirements, downsizing, and
outsourcing are interdependent in terms of the problems they cause for organizations. Again, the
causes are probably buried in not only the Corps activities, but in the Corps' relationships with its
external constituencies.
New approaches to looking at organizational failure examine the degree to which
organizations are internally stovepiped. Figure 13.4 shows that the Corps organizational
structure might lend itself to this. It appears regions and districts act pretty autonomously.
In addition Houck (2006) observes:
…restoring coastal Louisiana is a national issue and will require remedies
beyond this state. We lie at the receiving end of a large watershed, and some of
what we need has been turned off and other stuff that is hurting us has been
turned on. The Corps districts need to talk to each other. The EPA has to step up
to the plate, upstream states have to change some habits too. If the nation’s
taxpayers are going to be asked to spend more money than America spent on the
Marshall Plan to fix all of post-war Europe, then they have a right to expect a
national effort.
McCurdy (1993) discusses how stovepiping existed when NASA was created. Today the
adverse results of NASA’s stovepiping are excessive unit independence, specialization and
neglect of mutual coordination in a situation that should be characterized by just the opposite
(Roberts et al., 2005).
All in all, the Corps ability to do its job has been organizationally handicapped. It has lost
engineering and research and development muscle and flesh, it has lost its ability to maintain old
projects, it fails to be appropriately interdependent with various constituencies, and it fails to act
effectively on issues of internal interdependence. And, it cannot get well on its own.
13.4 Preventing the Next Katrina
In virtually all human affairs, risk is normal. The consequences of neglect may be grave,
if not now, in the future. As we indicated in the beginning of this chapter we are skeptical that
13 11
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
those with power and resources to prevent the next Katrina will take the steps necessary to do so
and we provided evidence for this assertion
From our larger discussion about defining safety and including all stake holders in
definition and response, three recommendations emerged:
ク
ク
ク
Responsibilities for vigilance and decision making at the tip of the authority structure
should be clarified and strengthened to enhance management of all modes of risk.
Additional technical Congressional staff should be appointed to assure adequate revenues
to manage risk and to monitor performance of the Executive Branch in its duties of care.
New processes should be authorized at a local level to foster informed consent and
dissent, and to function as early warnings in disasterprone areas, and to reflect that
citizens at risk are entitled to information regarding their exposure and opportunities to
participate in governance.
One central purpose should animate all the entities involved, separately and in tandem.
They should address the question, “How Safe is Safe Enough?” That investigation demands
foresight in the spirit of the injunction, “Without vision, the people perish.”
In addition to this larger purview, specific attention needs to be given to the Corps and
the organizations with which it is interdependent. We know a great deal about how to fix
problems of this nature, and there are growing bodies of engineering, legal, public policy,
organizational, and other literatures that address such issues. There is also a growing body of
experts from different areas who know how to talk about such issues. The problem is that
stakeholders have huge incentives not to pay any attention to this. They are no more likely to fix
this problem than they were likely to prevent the Challenger problem from becoming the
Columbia problem or the Betsy problem from becoming the Katrina problem.
Fixing the problem will require a set of processes that affected stakeholders do not want
to engage in:
ク
ク
ク
ク
ク
ク
They must come together to decide exactly what they want (clear and consistent goals) in a
politically complex and charged world.
They must be willing to spend many years addressing such problems in a world in which
incentives result in attention spans that more typically run the gamut of minutes to weeks.
Agencies must work together and trust one another.
They must recognize the interdisciplinary nature of their problems.
They must be willing to spend money and make recipients of that money accountable for
their spending.
They must develop oversight programs and agencies with real teeth.
13 12
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
13.5 Re-Engineering the USACE
Fixing the USACE’s technical problems will have only limited impact unless we also fix
the organizational problems. The USACE must strive to become a High Reliability Organization
HRO (emulating the Rickover Navy; see Appendix G). Four recommendations that would go a
long way toward repairing the Corp’s ability to design and build effective flood control projects:
ク
ク
Rebuild the USACE’s engineering and R&D capability,
ク
Develop a National Flood Defense Authority,
ク
Restructure the federal/state relationship in flood control,
Create effective disaster planning.
Three years before Katrina, the National Research Council concluded that the “Corps’
more complex planning studies should be subjected to independent review by objective, expert
panels.” (National Research Council 2002). This is an obvious point – which makes it all the
more urgent to implement. Although the need for independent project review has been apparent
for years, none of the past proposals have yet been implemented.
13.5.1 Rebuilding USACE Technical/Engineering Capacity
The USACE’s engineering and R&D capabilities were degraded over the past twenty
years as a result of streamlining and budget cuts (downsizing and outsourcing). As a nation, we
cannot afford the loss of this expertise. Although outsourcing can be efficient in some instances,
it cannot be allowed to deplete USACE’s own core expertise. As the National Research Council
concluded, “Shifting analytical tasks to the private sector, however, has its limits, as core, “in
house” competence is necessary for the Corps to commission, manage, and comprehend the
advice of external experts.” (National Research Council, 2004)
The Army Corps of Engineers must be, first and foremost, the nation’s premiere expert in
flood control engineering. Through no fault of its own, the Corps has been stripped of much of
what it needs to perform this role. Congress must adopt a plan and allocate the necessary funds to
“put the ‘engineers’ back into the Corps of Engineers.” It must remake the Corps into the
organization that new, “wet behind the ears” civil engineers will want to join to sink their teeth
into their new profession. It must retain and perform sufficient challenging engineering work as
to encourage these engineers to develop their careers within the USACE. It must define and
perform sufficient R&D work to help support the activities of these engineers. And it must pay
them adequate salaries as to be suitably competitive with private industry.
The Working Group for PostHurricane Planning for the Louisiana Coast has advanced
some complimentary recommendations for Corps staffing in their report A New Framework for
Planning the Future of Coastal Louisiana after the Hurricanes of 2005 (2006):
An essential element in enhancing the credibility and soundness of planning and
implementation is an agency's internal staff capabilities. The Corps of Engineers
is facing a significant loss of staff numbers and capability through retirement, just
at the time that the demands for its skills are increasing. Indeed, the integrated
planning process will demand a wider array of skills form the engineering,
hydrologic, geological, biological and social sciences than is currently available
in the agency or in federal or state agencies generally. Also, the effectiveness of
13 13
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
the long-term program requires the institutional memory that develops within a
permanent and professional staff..
13.5.2 Restructuring the Federal/state Relationship in Flood Defense
The USACE’s relationship with local flood control entities in Louisiana is dysfunctional.
Some of the issues relate to the fragmentation of the local entities, which the state has begun to
address. However, a number of the issues are broader.
Often, water planning activities involve not only multiple federal agencies, but also state
and local governments. In the blunt words of one observer, “The first consequence is that flood
defense has no head . . . . Whatever the merits of this diffusion of authority, it does not produce
coherent flood control.” (Houck, 2006). One useful model may be what has been called
“modularity” a concept which involves provisional and functional rearrangement of units in
terms of alternative configurations of tools, structures and relationships. (Freeman and Farber,
2005).
13.5.3 Developing a National Flood Defense Authority
A National Flood Defense Authority (NFDA) might be instituted and charged with
oversight over the construction and maintenance of flood control systems. Each state would have
an equivalent organization that could foster cooperation and developments between and within
the states. The Corps of Engineers, state flood control authorities, and technical advisory boards
would work with the NFDA to foster application of the best available technology and help
coordinate development and maintenance efforts and planning. Federal and state governments
would provide reliable and sustainable funding for the lifecycle (design, construction, operation,
maintenance) of specific flood defense systems. To facilitate coherent funding, Congressional
authorization and financing would be separated from the traditional Water Resources
Development Act process.
The Corps of Engineers, in cooperation with other qualified agencies and industrial
partners, would have the responsibility to design and construct, and if directed and authorized,
operate and maintain flood defense systems. The NFDA would be based on a continuous and
integrated process of flood risk assessment and management for specified flood defense systems,
with each of these systems being integrated with other allied flood defense systems. Flood risk
assessment and management processes would include proactive, reactive, and interactive
(adaptive) approaches based on the best available proven technology. Flood defense system
planning and development would engage public and industrial stakeholders and responsible
federal and state agencies in a cooperative and vigilant Technology Delivery System.
The Interagency Floodplain Management Review Committee in 1994 advanced similar
concepts as a result of their indepth evaluation of the performance of existing floodplain
management programs following the disastrous 1993 Midwest flooding. The Working Group for
PostHurricane Planning for the Louisiana Coast has advanced similar recommendations for
organization and funding in their report A New Framework for Planning the Future of Coastal
Louisiana after the Hurricanes of 2005. This group observed (2006):
Organizational and funding barriers that have inhibited the adoption of an
integrated planning and adaptive decision making process persist. Both new
organization and funding reforms are needed to support coastal planning and
project implementation by the Corps and the State.
13 14
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
This group proposed a model that involves proposals for Federal Intragovernmental
Coordination, development of working processes with the new Louisiana Coastal Protection and
Restoration Authority, the development of a Coastal Assessment Group and Coastal Engineering
and Science Program. This model includes recommendations for programmatic authorization and
funding including formation of a new Louisiana Coastal Investment Corporation and major
revisions in the Water Resources Development Act appropriations process.
13.5.4 Creating Effective Disaster Planning
Research on organizational learning finds that practices and routines in organizations
develop incrementally through feedback from the organization’s environment. Organizations
generally tend to be inert, adapting less than perfectly to and falling in and out of alignment with
their environments (Nelson and Winter, 1982).
This stagnation is especially dangerous for organizations that deal with major
emergencies such as floods, fires, and other natural and manmade disasters. Organizations that
await major failures before adapting tend to enter crisis mode and find learning and response
even more difficult (Staw et al., 1981; Turner, 1976). For example, following the demise of the
space shuttle Challenger, NASA faced political pressures, inertia, and resource constraints that
expedited some organizational changes but made other structural and cultural adjustments more
difficult (McCurdy, 1993). Furthermore, in the absence of a significant environmental change or
destabilizing event, lessons learned in organizations often tend to be forgotten or misapplied (de
Holan and Phillips, 2004; March et al., 1991).
Even worse, because of the infrequency with which major disasters occur, trial and error
organizational learning processes may lead organizational members to forget lessons from past
disasters. Levitt and March (1988) argue that in the case of disaster preparedness, trial and error
processes lead to “pernicious learning” – organizational leaders conclude that resources
designated for disaster preparedness are left idle and should be applied elsewhere. Disaster
preparation calls for a different form of learning in which organizations draw on not only their
own experiences but also those of other organizations. Such network effects exist for a variety of
learning processes (e.g. Argote et al., 1990; Baum and Ingram, 1998; Beckman and Haunschild,
2002).
Over the past few decades, scholars from many disciplines have advocated relational or
systems approaches, as opposed to reductionist approaches that study particular events and
entities in isolation (Miller, 1972; Wolf, 1980; Kastenberg et al., 2003). Taking a relational
approach will help us identify and examine learning processes as they affect and are influenced
by organizations responding to major catastrophes. The issues we discuss may occur at several
different levels in organizations – the interpersonal level, the subunit level, or the inter
organizational level.
Fortunately, we have learned a great deal about how to overcome these organizational
barriers. What is needed is to instill “mindfulness” toward risks. We suggest three ways of doing
so:
ク
ク
ク
Create a National Disaster Advisory Office in the White House.
Create a Catastrophic Risk Office in Congress.
Make FEMA into a High Reliability Organization (HRO).
13 15
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
13.5.4.1 Creating a National Disaster Advisory Office in the White House
No one in the White House has the job of disaster response. Yet, federal disaster response
requires action by many agencies – not just FEMA but also DOD, EPA, CDC, and others. White
House coordination of these executive branch activities is crucial. Just as the White House has a
National Security Advisor, it needs to have an official charged with national disaster oversight.
This official would also be in charge of monitoring organizational problems in the line agencies
in charge of disaster response. Moreover, a natural part of the official’s portfolio would be
disaster prevention efforts, where the aim should be to avoid ever again being taken unawares by
a “predictable surprise” like Katrina.
13.5.4.2 Creating a Catastrophic Risk Office in Congress
An integrated approach to catastrophic risk is lacking. One lesson from Katrina is that
disasters are not just engineering failures, they are social system failures and failures of
government. Societal and physical infrastructures can collapse. Consequently, disaster
prevention cannot be considered in isolation from disaster response, mechanisms for
compensation and risk spreading, and reconstruction planning. All of these issues are tightly
coupled, yet the linkages receive little attention.
Under the Constitution, Congress bears the primary responsibility for developing national
policy and setting national priorities. Congress authorizes and controls FEMA, the Army Corps,
flood control projects, the flood insurance program, and other aspects of our nation’s response to
catastrophic risks. Yet Congress lacks the expertise needed to accomplish these tasks in a
systematic way.
13.5.4.3 Making FEMA an HRO
Some organizations cannot afford to fail (Appendix F). Accidents can be disastrous on
nuclear submarines, aircraft carriers, in air traffic control, and in hospital emergency rooms.
Successful organizations of these kinds have learned to attain high reliability. By studying these
organizations, experts have learned the ingredients to creating a High Reliability Organization
(HRO). And there is a growing body of research on high reliability organizations (Weick 1987;
Roberts 1990; Madsen et al., in press) and on high reliability systems of organizations (Roberts
and Grabowski, in press; Roberts et al. 2005). Until organizations representing various aspects of
disaster preparedness and disaster management seriously see themselves as systems of
organizations, they cannot adequately address the problems they face.
13.6 Recommendations – Organizing for Success
The primary requirement for reconstitution of a Technology Delivery System that can
and will provide an adequate and acceptable NOFDS is mobilization of the 'will' to provide such
a system. If the United States decides that the catastrophe of Katrina will not be repeated, then
the necessary leadership, organization, management, resources, and public support must be
mobilized to assure such an outcome. One of the primary challenges is time; the clock is ticking
until this area of the United States is again confronted with a severe challenge of flooding.
Recommendation 1: Seriously consider defining risk within the framework of federal,
state, and local government responsibilities to protect their citizens.
Recommendation 2: Exploit the major and unprecedented role that exists for citizens,
who should be considered part of governance in the spirit that those who govern do so at the
13 16
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
informed consent of the governed. This is the population exposed to catastrophic risks, and the
people that will be protected by the NOFDS. Authorities for catastrophic risk management
should ensure that those vulnerable have sufficient and timely information regarding their
condition and a reciprocal ability to respond to requests for their informed consent especially
regarding tradeoffs of safety for cost. The public protected by the NOFDS need to be encouraged
to actively and intelligently interact with its development.
Recommendation 3: Intensify, focus, and fund Corps of Engineers modernization
efforts; increasing inhouse engineering capabilities and project performance, increasing in
house research and development capabilities, increasing inhouse engineering performance of
technically challenging projects, developing an organizational culture of high reliability founded
on existing cultural values of Duty, Honor, Country, and developing a leadership role and
responsibility for technical and management oversight of all phases of development of a
NOFDS. Technical superiority must be reestablished. Outsourcing must be balanced with in
sourcing to encourage development and maintenance of superior technical leadership and
capabilities. This will require close and continuous collaboration of federal legislative, executive,
and judicial agencies. This will require that the Corps of Engineers reconceptualize itself as a
pivotal part of a modular organization developing partnerships with other federal agencies, state
and local governments, enterprise interests, and private stake holders.
Recommendation 4: Restructure federal/state relationships in flood control. One
possible model is what has been called “modularity” a concept which involves provisional and
functional rearrangement of units in terms of alternative configurations of tools, structures and
relationships. Enhancing cooperation and collaboration, reducing confusion as to overlapping
areas of operation and responsibility, and mutually supportive crosschecks and communication
should all be advanced.
Recommendation 5: Develop a National Flood Defense Authority (NFDA) charged
with oversight over the design, construction, operation and maintenance of flood control
systems. Each state would have an equivalent organization that could foster cooperation and
developments between and within the states. The Corps of Engineers, state flood control
authorities, and technical advisory boards would work with the NFDA to foster application of the
best available technology and help coordinate development and maintenance efforts and
planning. In cooperative developments, federal and state governments would provide reliable
and sustainable funding for the lifecycle of specific flood defense systems. This development
should be accompanied by development of an integrated and coherent Louisiana Flood Defense
Authority representing state, regional, local, city, and public stakeholders that can focus and
prioritize stakeholder interests and requirements and collaborate with the Corps of Engineers in
development of a NOFDS.
Recommendation 6: Because of the importance of emergency response in the NOFDS,
FEMA should be developed as a high reliability organization (HRO) and returned by the
executive branch to Cabinet level status. A new Council for Catastrophic Risk Management
should be appointed within the White House and given oversight of disaster preparation and
response. A similar body should be appointed within Congress. Incentives must be created to
encourage all levels of government to deal proactively and effectively with potential national,
regional, and local catastrophes.
13 17
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
13.7 References
Argote, L., et al. (1990). “The Persistence and Transfer of Learning in Industrial Settings.”
Management Science, 36(2), 140154.
Baum, J. A. C., and Ingram, P. (1998). “SurvivalEnhancing Learning in the Manhattan Hotel
Industry, 18981980.” Management Science, 44(7), 9961016.
Beckman, C. M., and Haunschild, P. R. (2002). “Network Learning: The Effects of Partners’
Heterogenity of Experience on Corporate Acquisitions. Administrative Science Quarterly,
47(1), 92124.
Bierly, P.E., and Spender, J.C. (1995). “Culture and high reliability organizations: The case of
the nuclear submarine.” Journal of Management, 21, 639656.
Bradbury, H., and Lichtenstein, B.M.B. (2000). “Relationality in Organizational Research:
Exploring the Space Between.” Organization Science, 11, 551564.
Buber, M. (1970). I and Thou trans. Walter Kaufman, New York: Scribner’s Sons
Clarke, L. (1999) Mission Improbable: Using Fantasy Documents to Tame Disaster, University
of Chicago Press, Chicago, IL.
Collins, S. (2006). Opening Statement, Hurricane Katrina: Recommendations for Reform,
Committee on Homeland Security and Government Affairs, Washington
DC,
03/08/2006.
Columbia Accident Investigation Report (2003). Columbia Accident Investigation Report, U.S.
Government Printing Office, 1, Washington DC.
De Holan, P., and Phillips, N. (2004) “Remembrance of Things Past? The Dynamics of
Organizational Forgetting.” Management Science, 50, 16031613.
Dekker, S. (2002) The Field Guide to Human Error Investigations, Ashgate, Aldershot, England.
Department of Homeland Security (2006). “Our Top Five Most Frequently Asked Questions.”
NIMS Alert, No. 04006, Washington DC..
Department of Homeland Security (2004a). National Response Plan, U.S. Government Printing
Office, Washington DC.
Department of Homeland Security (2004b). National Incident Management System, U.S.
Government Printing Office, Washington DC.
Freeman, J. and Farber, D. (2005). “Modular environmental regulation.” Duke Law Journal, 54,
795912.
Grabowski, M., and Roberts, K. H. (1999). “Risk Mitigation in Virtual Organizations.
Organization Science, 10, 704721.
ALSO IN JOURNAL OF COMPUTER MEDIATED COMMUNICATION, 1998, 3, 4.
Heath, C., and Staudenmayer, N. (2000). “Coordination Neglect: How Lay Theories of
Organizing Complicate Coordination in Organizations.” Research in Organizational
Behavior , B.M. Staw and R.I. Sutton (Eds.), Elsevier, New York, 22, 153191.
Houck, O.A. (2006). “Can We Save New Orleans?” Tulane Environmental Law Journal, 19(1),
<http://www.law.tulane.edu/tuexp/journals> (Apr. 25, 2006).
13 18
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Kastenberg, W.E., HauserKastenberg, G., and Norris, D. (2003). “Shifting the Paradigm of Risk
Analysis.” Proceedings of the International Seminar on Nuclear War and Planetary
Emergencies, 29th Session, Erice, Italy, May 1015.
Lakoff, A. (2006). “From Disaster to Catastrophe: The Limits of Preparedness.” Understanding
Katrina: Perspectives from the Social Sciences, The Social Science Research Council,
<http://understandingkatrina.ssrc.org/Lakoff> (May 15, 2006).
Levitt, B., and March, J. G. (1988). “Organizational Learning.” Annual Review of Sociology, 14:
319340.
Madsen, P., Desai, V., and Roberts, K. H. (in press). “Designing for High Reliability: The Birth
and Evolution of a Pediatric Intensive Care Unit Organization.” Science
March, J. G., Sproull, L., and Tamuz, M. (1991). “Learning from Samples of One or Fewer.
Organization.” Science, 2, 113.
McCurdy, H. E. (1993). Inside NASA: High Technology and Organizational Change in the U.S.
Space Program, Johns Hopkins Press, Baltimore MD.
Miller, J.G. (1972). Living Systems, McGraw Hill, New York.
National Research Council (2002). Review Procedures for Water Resources Project Planning.
National Academies Press., Washington DC.
Nelson, R. R., and Winter, S. G. (1982). An Evolutionary Theory of Economic Change,
Cambridge MA.
Perrow, C. (2005). Using Organizations: The Case of FEMA, Social Science Research Council
Forum, <http://understandingkartina.ssrc.org> (Apr. 21, 2006).
Roberts, K. H. (1990) “Some Characteristics of One Type of High Reliability Organization.”
Organization Science, 1, 160176.
Roberts, K. H., and Bea, R. G. (2001a). “Must Accidents Happen: Lessons from High Reliability
Organizations.” Academy of Management Executive, 15, 7079.
Roberts, K. H., and Bea, R. G. (2001b). “When Systems Fail.” Organizational Dynamics, 29,
179191.
Roberts, K. H., Madsen, P. M., and Desai, V. M. (2005). “The Space Between in Space
Transportation: A Relational Analysis of the Failure of STS 107.” Organization at the
Limit: Lessons from the Columbia Disaster, W. H. Starbuck and M. Farjoun (Eds.),
Blackwell, Malden MA, 8198.
Roberts, K. H., and Wong, D. (2006). “Outsourcing and Temporary Workers.” Encyclopedia of
Industrial/Organizational Psychology, S. Rogelberg (Ed.), Sage, London, England.
Roberts, K. H., and Grabowski, M. (in press). “Risk Mitigation in Healthcare Organizations and
in Aggregations of those Organizations.” Human Error in Medicine, M.S. Bogner (Ed.),
2nd Ed., Erlbaum, Hillsdale NJ.
Senate Committee on Homeland Security and Governmental Affairs (2006). Hurricane Katrina:
A Nation Still Unprepared, Washington DC.
Staw, B. M., Sandelands, L. E., and Dutton, J. E. (1981). “ThreatRigidity Effects in
OrganizationalBehavior a Multilevel Analysis.” Administrative Science Quarterly,
26(4), 501524.
13 19
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Tate, B., and Halford, G. (2002). “New moves will streamiline gov't.” Engineer Update, U.S.
Army Corps of Engineers, 26(7),
<http://www.hq.usace.army.mil/cepa/pubs/jul02/story3.htm> (Mar. 20, 2006).
Turner, B. A. (1976). Organizational and Interorganizational Development of Disasters.
Wykeham Publishers, London, United Kingdom.
United States Army Corps of Engineers (2002). “USACE Strategic Workforce Planning.”
USACE Command Brief, <http://www7.nationalacademies.org/ffc/usace.pdf> (Mar. 30,
2006).
Vaughan, D. (2005). “Systems Effects: On Slippery Slopes, Repeating Negative Patterns and
Learning from Mistakes.” Organization at the Limit: Lessons from the Columbia
Disaster, W. H. Starbuck and M. Farjoun (Eds.), Blackwell Publishing, Malden MA, 41
59.
Weick, K. E. (1987). “Organizational Culture as a Source of High Reliability.” California
Management Review, Winter, University of California, Berkeley CA.
Weick, K. E., and Sutcliffe, K. (2001). Managing the Unexpected, Jossey Bass, San Francisco
CA.
Wolf, F. A. (1980). Taking the Quantum Leap, Harper & Row, New York.
Working Group for PostHurricane Planning for the Louisiana Coast (2006). A New Framework
for Planning the Future of Coastal Louisiana after the Hurricanes of 2005, University of
Maryland Center for Environmental Science, <http://www.umces.edu/larestore> (Apr.
23, 2006).
13 20
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Dollars in Thousands
100,000
80,000
Congressional
Budget Request
60,000
Presidential
Budget Request
40,000
20,000
1975
1980
1985
1990
1995
2000
2005
Fiscal Year
Figure 13.1: Lake Pontchartrain and Vicinity Project Construction Appropriations Over the
Past 30 years [2005 dollars]; President’s Budget Request (grey) and the Amount
Recommended by Congress (black).
Figure 13.2: Artwork by Jan Fitzgerald illustrating the debate surrounding President Bush’s
initative to streamline the federal government (Tate and Halford 2002).
13 21
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 13.3:
Human Capital Planning Projected Retirement (USACE 2002).
Washington,
Headquarters
Division A
District A1
Division X
District AX
District X1
District XX
Figure 13.4: Conceptual Organizational Chart of the U.S. Army Corps of Engineers Civil
Works Program. 1
1
The Corps Civil Works Program is composed of 8 Divisions and 38 subordinate districts. Prime Power, ERDC,
Centers, and FOAs are not shown for clarity. In addition, a 9th provisional division with four districts was activated
January 25, 2004, to oversee operations in Iraq and Afghanistan. A more complete organizational chart is available
in USACE 2012 – Appendix G, Resource Analysis, page 1.
13 22
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
CHAPTER FOURTEEN: ENGINEERING FOR SUCCESS
The tragedies of Hurricanes Katrina and Rita in 2005 have revealed to the
world the enormous challenge Louisiana now faces. South Louisiana appears
to have entered a period when the convergence of two powerful forces is
working against its survival. Since the 1950’s, the processes driving coastal loss
have continued only slightly abated. Since 1990, meteorological and oceanic
processes driving tropical systems have more frequently generated category 4
and 5 hurricanes. More destructive hurricanes are predicted for coming
decades. ~ South Louisiana’s ongoing peril is the continued overlap of
weakened hurricane protection with more frequent and intense hurricanes.
In light of this predicament, how can the coast and culture of south Louisiana
survive? The survival of a culture and a region is at stake. Hurricanes Katrina
and Rita may have narrowed the field of discussion from what we might want,
down to what we absolutely need. There is a growing consensus that what is
needed is a pragmatic and effective strategy to integrate both coastal habitat
restoration and engineered flood protection, such as levees. This strategy must
be established soon and while under duress.
John Lopez (2006).
The Multiple Lines of Defense Strategy to Sustain Louisiana's Coast
Report to Lake Pontchartrain Basin Foundation, New Orleans.
14.1 Introduction
At the present time, the federal government is just completing a significant effort to re
establish the New Orleans Flood Defense System (NOFDS) to “preKatrina conditions” by a
target date of June 1, 2006. The federal government has proposed to further improve the NOFDS
to meet “100year flood conditions” by 2010. Studies are currently underway by the Corps of
Engineers to define an expanded and more reliable NOFDS (see Appendix G). In this Chapter
we explore options for the engineering elements that could be provided in an improved longterm
NOFDS.
The first question to be addressed in going forward is: “what should we do about
providing adequate flood protection for the greater New Orleans area?” To the people who lived
and continue to live in this area, this is not a question. These people are in the process of
rebuilding their homes and lives. A majority of people who live in this area are committed to re
building and continuing the development of this area. Some have and will decide not to return;
they will rebuild elsewhere.
The real question is about the ‘we’. The following thoughts on this question were
advanced by former Speaker Newt Gingrich (2006):
Shortly after Hurricane Katrina devastated New Orleans, Speaker of the House
Dennis Hastert wondered aloud whether the Federal Government should help
rebuild a city much of which lies below sea level. The most tough-minded answer
to that question demonstrates that rebuilding and protecting New Orleans is in
14 1
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
the national interest. Reason: The very same geological forces that created that
port are what make it vulnerable to Category 5 hurricanes and also what make it
indispensable.
If engineering the Mississippi made New Orleans vulnerable, it also created
enormous value. New Orleans is the busiest port in the U.S.; 20% of all U.S.
exports and 60% of our grain exports, pass through it. Offshore Louisiana oil and
gas wells supply 20% of domestic oil production. But to service that industry,
canals and pipelines were dug through the land, greatly accelerating the washing
away of coastal Louisiana. The state's land loss now totals 1,900 sq. mi. that land
once protected the entire region from hurricanes by acting as a sponge to soak up
storm surges. If nothing is done, in the foreseeable future an additional 700 sq.
mi. will disappear, putting at risk port facilities and all the energy-producing
infrastructure in the Gulf.
…Washington also has a moral burden. It was the Federal Government's
responsibility to build levees that worked, and its failure to do so ultimately led to
New Orleans' being flooded. The White House recognized that responsibility
when it proposed an additional $4.2 billion for housing in new Orleans, but the
first priority remains flood control. Without it, individuals will hesitate to rebuild,
and lenders will decline too invest.
How should flood control be paid for? States get 50% of the tax revenues paid to
the Federal Government from oil and gas produced on federally owned land.
States justify that by arguing that the energy production puts strains on their
infrastructure and environment. Louisiana gets no share of the tax revenue from
the oil and gas production on the outer continental shelf. Yet that production puts
an infinitely greater burden on it than energy production [from] other federal
territory puts on any other state. If we treat Louisiana the same as other states
and give it the same share of tax revenue that other states receive, it will need no
other help from the government to protect itself. Every day's delay makes it
harder to rebuild the city. It is time to act. It is well past time.
For us it is not a question of if we go forward to provide an adequate and acceptable
NOFDS. It is a question of how we go forward. Going forward will demand a lot of all involved
including vision, commitment, responsibility, respect, organization, cooperation, leadership,
knowledge, resources, preparations, time, and some good luck. While this Chapter examines the
engineering aspects of providing longterm hurricane flood protection for the greater New
Orleans area, it should be clearly understood that a PREREQUISITE to a successful venture
must be reengineering the Technology Delivery System (see Chapter 13) needed to develop
such a system. History has clearly shown that without an effective and sustainable TDS, we can
expect a deficient and defective longterm NOFDS. History will repeat itself if we let it.
During the next several decades, hurricane seasons are expected to produce greater
numbers of more severe storms. Unnecessary delays in embarking on development and
realization of a longterm NOFDS only increase our chances of failing. We learned this lesson
during the 40year period between the disastrous flooding of New Orleans in 1965 (hurricane
Betsy) and the catastrophic flooding of 2005 (hurricane Katrina). Now is the time for careful and
deliberate thought followed by effective and timely action. Another disastrous flooding of the
greater New Orleans area should not be an option.
14 2
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
14.2 Engineering Considerations
The ILIT addressed two key aspects associated with the engineering considerations of
going forward: (1) the NOFDS physical facilities, and (2) the engineering criteria and guidelines
for these facilities.
14.2.1 Physical Facilities
Evaluation of the options for NOFDS physical facilities requires a basic understanding of
the natural environmental geological ecological setting of this area, the commercial
industrial complex established in this area, and unique cultural social institutional political
elements. This is a very complex system whose future is shadowed by its past.
A systematic and integrated study needs to be performed of the options for provision of
physical facilities so that informed choices can be made about how best to provide longterm
flood protection for the greater New Orleans area. The NOFDS is part of an even larger
challenge that involves other parts of the Gulf coast and the floodplain of the Mississippi River
(Dean, 2006). The real threats of increased hurricane activity and intensity, coastal degradation,
subsidence, and climate change (rise in sea level, increase in rainfall and flood potential) must be
recognized and appropriate and effective preparations put in place to help protect life and
property in this area.
The Mississippi River and the Gulf of Mexico have been interacting in this part of the
United States for millions of years (Kelman, 2003). As a result of sediments transported and
deposited by the Mississippi River during the past 100,000 years, a vast complex of delta lobes
have developed where a succession of different river channels meet the Gulf of Mexico
(Coleman, 1988). Sixteen of these lobes have been developed and abandoned during the past
20,000 years. The sediments deposited by these delta lobes dominate the geology of this area,
and the recently deposited sediments reach thicknesses exceeding 500 feet (U.S. Army Corps of
Engineers, 2004).
The Mississippi Delta is a broad wedgeshaped floodplain whose top is about where the
Atchafalaya River branches off from the Mississippi River and whose broad curved base is the
Gulf of Mexico coastline (about 150 miles wide) (Sparks, 2006). The coastline is delineated with
a long line of barrier islands. The shape of this delta is determined by sediment accumulation,
compaction, subsidence, growth faulting, changing sea level, and most recently by man's
activities. Recent information indicates that since the sea reached its present level (about 6,000
years ago), six major lobes including a developing new one at the mouth of the Atchafalaya
River have existed. The modern Birdsfoot Delta that lies to the southeast of New Orleans
(Plaquemines parish) has existed for only about the last 1,000 years.
The river has been trying to change its course to the Atachafalaya River (100 miles to the
west) as the length of the Mississippi River to the Gulf of Mexico has increased (now more than
200 miles). In order to maintain New Orleans as a deepwater port in the 1950s, the Corps of
Engineers constructed the Old River Control Structure to help divert about 30% of the
Mississippi River water down the Atachafalaya and keep the remainder flowing to the Gulf
through its present course. In 1973, a flood on the Mississippi River almost caused failure of the
Old River control Structure. The Corps completed a new auxiliary structure in 1985 to take some
of the pressure off the Old River control Structure. At the present time, the Atachafalaya lobe is
actively building toward the Gulf of Mexico and the lobe south of New Orleans is regressing.
14 3
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
A variety of processes have altered the natural process of land building by the Mississippi
River and its tributaries (Hallowell, 2005; Committee on the Restoration and Protection of
Coastal Louisiana, 2006; Zinn, 2004, 2005a, 2005b). These include the building of upstream
dams and flood control structures (decreased sediment supply), building of levees (which do not
permit sediment transport to adjacent areas), building of canals and pipelines (oil and gas
exploration and production), building of navigable waterways (e.g.: Gulf InterCoastal Water
Way, Mississippi River Gulf Outlet), and configuration of the current mouth of the Mississippi
river to “shoot” sediments out into the Gulf of Mexico, over the edge of the continental shelf, in
order to reduce the need for active dredging to maintain navigability of the main river channel
for shipping. All have had their effects on reducing replenishment of sediments to keep up with
subsidence, on the balance coastal transport processes, and on providing nutrients to sustain
freshwater wetlands.
With population and industrial growth along the Mississippi River and its tributaries (it
drains about 40% the United States), influx of byproducts and waste products have also taken
their toll on the wetlands. Exploration for and production (extraction, transport) of hydrocarbons
have also taken their toll on wetlands and contributed to land loss. The rise of sea level has also
taken its toll. The result is a rapidly degrading and regressing coastline. This coastline is
projected to loose about 10 square miles of land per year during the next 50 years (Dean, 2006;
Sparks, 2006). The rapidly regressing coastline has had important effects with regard to the
increase in hurricane risk affecting the NOFDS.
The NOFDS is faced not only with the challenges associated with potential hurricane
surges and waves, but also with potential floods from the Mississippi River, with subsidence and
compaction, with reduction of the stormbuffering provided by coastal barrier islands and
wetlands, and with potential water and saltwater ingress provided by manmade waterways.
Oliver Houck (2006) addressed these challenges:
So here is the starting point: exactly what we do want the Louisiana coast to look
like, to do for us, for say, the next century? …Earth to Louisianans: you really
can't have this cake and eat it too. With all due respect, it is not just a matter of
doing everything we want 'smarter.' It is a matter of getting straight what we
want, and that comes first. What comes next is the hardest step for any American
community to take, and shall be heresy in South Louisiana. A plan. The mere
mention of planning raises blood pressures and brings on cries of Godless
Communism. What we have had in the city of New Orleans and along the entire
gulf coast is planning by default (local attorney Bill Borah calls it 'planning by
surprise'). Planning takes place. It's just that we haven't taken part in it. Where
water resources are concerned, it starts with real estate developers, port
authorities, levee boards and other outside-the-ballot-box enterprises, their
projects facilitated and funded by the Army Corps of Engineers. In their minds,
the only question is a technical one: what kind of engineering do we need to get
our project done? The system has produced the expected results: more rip-rap
here, more drainage there, and levees to the horizon. The goal is - although it is
never stated anywhere - to develop as much of the coast as possible. When you
add the projects up, they determine the destiny of the city and South Louisiana.
What is apparent is that these levees, designed by engineers and approved by
Congress, are the basic planning documents for the future of South Louisiana.
What is north of these levees will be developed. What is south of them will be
14 4
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
anyone's guess, although not for long; the map on global warming shows these
coastal marshes gone within a century. De facto, we end up with a wall. Not all
that adequate a wall, by the way. Only Category three, if that. Can you imagine
the costs of maintaining even a Category three levee system winding back and
forth to the Gulf from New Orleans to Texas” Can we imagine what will happen
when development piles in behind it, and then gets flooded? Do we already know,
from Lakeview and New Orleans East, what happens to land elevations behind
levees once they are drained and paved?
Our choice is to start this process from the other end. If we do, another range of
options open. There are a dozen major towns across the southern tier with
thousands of homes and residents, and they deserve protection. But the way to
provide it may be with the same kind of ring levee systems that protects (or
should) New Orleans and its surrounding parishes, supplemented by flood gates
at the mouths of the main canals. Or, it may mean peninsular levee systems down
the historic ridges of the bayous, protecting what has always been the high
ground. …Problem is, we have lacked the process - we have lacked even the
language - for such a discussion. In addition to scientists and engineers, we may
need some social workers. In saying this, I am most serious.
The ILIT examined two basic alternatives to develop a longterm NOFDS. The first was
constructing levees, floodwalls, and pump stations capable of providing a long term NOFDS. At
the present time, efforts are underway to provide “100year” flood protection. But, the question
is why “100year” protection? Why not 1,000 year or 10,000 year protection (frequently posed as
Category 4 or 5 hurricane protection)?
Our studies of economic costbenefit guidelines, and historic and current standards of
practice for public facilities in the United States and elsewhere indicated that protection against
disastrous flooding of the greater New Orleans area should be for conditions having average
return periods much more demanding than the present goal of “100year” flood protection. This
issue was addressed by another very similar region that must defend its population and
commercial enterprises at elevations up to 23 feet below sea level the Netherlands (Netherlands
Water Parternership, 2005):
Our standards are accepted risks related to the design-criteria of our dikes. Those
standards are laid down in the Flood Defense Act. For the economically most
important and densely populated part of the country, we design our dikes and
dunes to be strong enough to withstand a storm-situation with a probability of 1
to 10,000 a year. That means, that a Dutchman - if he should live a 100 years has a chance of 1 percent to witness such an event. For our parliament, these
odds became the acceptable standard. For the less important coastal areas we
calculate the probability of 1 to 4,000 and along the main rivers 1 to 1,250.
This background was developed largely after the Netherlands suffered catastrophic
flooding of the country in 1953. This flooding was comparable to the flooding of the greater
New Orleans area in the wake of hurricane Katrina (approximately 1,800 dead, 50,000 destroyed
homes, 350,000 acres of flooded land). It was also preceded by a history that included a large
number of malfunctions that included poor organization, bad maintenance, warnings not heeded,
poor communications, underestimation of the danger, negligence, lack of preparedness (Jurjen
14 5
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Battjes, personal communication; Dec. 30, 2005). This same history was repeated in the
catastrophic failure of the NOFDS.
Following the 1953 catastrophe, the Dutch vowed “never again” and developed a system
that is today a model of advanced engineering and water resource management. It also provides a
model for the organizational reengineering required to realize the system they have in place
today, and that they continue to maintain and improve. This organization is a centralized
Rijkswaterstaat which is the national public works department in charge of all flood defense
works. This department has direct ties and interfaces with the local agencies responsible for
continued development, maintenance, and improvement of flood defense work (including
evacuation and disaster recovery). However, the Dutch have learned the sad lessons of trying to
overwhelm nature with engineered works. They have seen many unintentional consequences
from such an approach surface as very severe negative environmental and quality of life impacts.
And, they learned from these mistakes and gone on to remediate the mistakes and develop new
strategies (Netherlands Water Partnership, 2005):
Climate changes are increasing the likelihood of flooding and water-related
problems. In addition population density continues to increase, as does the
potential for economic growth, and consequently, the vulnerability to economic
and social disaster. Two undesirable developments that, in terms of safety,
exacerbate one another - a grown risk with even larger consequences. As such,
the safety risk is growing at an accelerated pace (safety risk - chance multiplied
by consequence).
The Netherlands is changing its approach to water. This change involves the idea
that the Netherlands will have to make more frequent concessions. We will have
to relinquish open space to water, and not take back existing open spaces, in
order to curb the growing risk of disaster due to flooding, We will also need to
limit water-related problems and be able to store water for expected periods of
drought. By this we do not mean space in terms of the height of ever taller levees
or depth through continued channel dredging, but space in the sense of flood
plains. This approach will require more area, but in return we will increase our
safety and limit water related problems. Safety is an aspect that must play a
different role in spatial planning. Only by relinquishing our space can we set
things right; if this is not done in a timely manner, water will sooner or later
reclaim the space on its own, perhaps [in a] dramatic manner.
The Dutch continue to be challenged by their countrymen not to become conceited or
complacent they are devoted to a culture of continuous improvements in their flood protection.
Our consideration of this background indicated that the most attractive option for
provision of an acceptable and sustainable longterm NOFDS is one of re-establishing and
enhancing selected natural defenses supplemented with engineered works as necessary to
provide long-term flood protection. Guidelines and many useful insights are provided by John
Lopez (2006) in the report The Multiple Lines of Defense Strategy to Sustain Louisiana's Coast
about how such an option might be developed. Additional background for development of this
option is also provided in the reports Coast 2050: Toward a Sustainable Coastal Louisiana
(Louisiana Coastal Wetlands Conservation and Restoration Task Force 1998), Ecosystem
Restoration Study (U.S. Army Corps of Engineers 2004), Drawing Louisiana's New Map
(National Research Council 2006), and A New Framework for Planning the Future of Coastal
14 6
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Louisiana after the Hurricanes of 2005 (Working Group for PostHurricane Planning for the
Louisiana Coast 2006). Results contained in these studies provide a coherent and substantial
basis for development of a longterm NOFDS. Lopez (2006) proposes eleven Lines of Defense
(see Figure 14.1):
1st Offshore shelf within the Gulf of Mexico: The offshore shelf ranges in depth
from 300 feet at the shelf edge to zero depth at the gulf shoreline. Its width vanes
from a few miles to hundreds of miles. The primary benefit of the shallow shelf is
to dramatically reduce wave height and wave energy from an approaching
tropical system. A negative aspect of the shelf is that it will promote higher storm
surges inland. The variable influences storm surges due to the geometry of the
shelf needs to be considered for storm surge analysis. Also, dredging activities on
the shelf should avoid increasing shoreline erosion by wave refraction around
dredge holes. The gulf fisheries and the oil and gas industry are key economic
aspects of the shelf. Examples: Narrow shelf at the mouth of Mississippi River &
Wide shelf offshore from Cameron Parish
2nd Barrier Islands: The Louisiana barrier island shoreline is characterized by
fragmented barriers or shoals with low vertical profiles and low sand content.
However, barrier islands provide an important wave barrier for interior sounds
and coastal marsh. The primary benefits of barrier islands are the near-complete
reduction in wave height and the slight reduction in storm surge further inland. A
negative aspect of barrier islands is their ephemeral nature and unpredictable
local impacts to them from hurricanes. Barrier islands also have significant
recreational aspects such as fishing and birding. Examples: Chandeleur Islands
and Grand Isle
3rd Sounds: The primary benefit of the sounds is to provide a relatively shallow
water buffer to deep water currents. Sounds do have a negative aspect during
storms by allowing waves to re-generate on the sound side of barrier islands.
Also, sounds may cause storm surge and wave erosion on the back side of barrier
islands.
4th Marsh Landbridges: Marsh landbridges are areas of emergent marsh with
relative continuity compared to adjacent bays, sounds or areas of significant
marsh/land loss. Ideally, landbridges connect other elevated landforms such as
natural ridges. Since some ridges are developed and have adjacent levees, marsh
landbridges may also bridge adjacent levee systems and economic corridors.
Marsh landbridges compose much of the residual internal framework of the coast
which reduces fetch and shoreline erosion of interior marshes and lagoons.
Landbridges impede storm surge movement inland and protect other emergent
marsh areas that may perform the same function. Some landbridges are
threatened themselves by various processes of marsh loss and need to be
sustained through restoration and maintenance. The landbridges represent an
increasing fraction of the remaining emergent marsh of the coast and provide
typical high productivity and fishery benefits typical of coastal wetlands.
Examples: East Orleans landbridge, Biloxi Marsh landbridge, Barataria Basin
landbridge, Upper Terrebonne Bay landbridge, Grand Lake-White Lake
landbridge, Western Marsh Island landbridge, south Calcasieu Lake landbridge
14 7
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
5th Natural Ridges: In southeast and central Louisiana, most natural ridges are
the natural levees of abandoned distributary channels. These channels now act as
tidal channels and are often colloquially named bayous or rivers. In southwest
Louisiana, most natural ridges are chenniers running parallel to the Gulf
coastline. Natural ridges may have continuous elevation of several feet and,
therefore, will impede overland flow across the ridge and potentially reduce
storm surge. Natural ridges often define (at least historically) the hydrologic
basins of the coast. Natural ridges are most effective when they have at least 6
feet of elevation and well drained soils to maintain upland forests. Forests will
also slow the movement of overland flow and may also provide a wind barrier.
Natural ridges tend to be the economic corridors across the coast including
primary state highways and coastal communities. These highways are also likely
to be evacuation routes. Examples: Bayou la Loutre, Bayou Lafourche
6th Manmade Soil Foundations: Manmade soil foundations for transportation
may provide incidental benefit to storm surges. Railroads, highways and spoil
banks may run parallel to the coast and locally provide a manmade ridge several
feet [high]. These foundations may have settled and may need improvement to
provide reliable transportation routes without chronic flooding. If highway
improvements are contemplated, the effects 011 storm surge may be considered.
Examples: Highway 90, Hwy 82
7th Flood Gates: Flood gates are typically designed to withhold flood water and,
therefore, remain open under most conditions. Flood gates are generally open so
as not to impede navigation or natural ebb and flow of tides and aquatic
organisms. Flood gates would be closed during a threat of flooding and to reduce
flood tides in channels. Because of the generally low elevation of the coast, the
effectiveness of flood gates may depend on the nearby topography or constructed
features such as levees or spoil banks. Examples: Bayou Bienvenue, Bayou Dupre
8th Flood Protection Levees: Flood protection levees are designed and
constructed for flood protection of municipalities or other coastal infrastructure
features. Levees are generally designed to be an absolute barrier defining a flood
side and a protected side. The intent is to have zero storm surge flooding on the
protected side, but an unintended consequence may be to increase water levels on
the flood side. Levees are generally not designed to be overtopped or to withstand
significant wave erosion. Exceptions include “potato levees” or other low relief
levees designed to reduce flooding from non-storm tides. Typical hurricane
protection levees protect limited portions of the coast with intense economic
development. Examples: St. Bernard levee, Jefferson and Orleans Parish levees
on Lake Pontchartrain
9th Flood protection pumping: Pumping stations are generally within leveed
areas and are used to reduce flood risk from rainfall and are not designed to
pump out flood water in the case of a levee breach. Most pumping stations are not
prepared with fuel, staff or other requirements to be effective to pump out flood
water from a significant levee breach. Generally, these are large capacity pumps
which displace water vertically above the water level on the flood side of the
levee. Pumping stations are generally to protect areas of intense development.
Examples: Orleans and Jefferson Parish’s pumping stations.
14 8
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
10th: Elevated Homes and Businesses: All homes and businesses in south
Louisiana are subject to being flooded if they are not elevated above the normal
land elevation. Even those behind levees are not 100% safe. Hurricanes Katrina
and Rita made this painfully clear. All attempts to reduce storm surge height or
its extent are limited by the intensity and attributes of particular storm events.
Since there will always be the potential of a storm exceeding the limits of
protection from storm surges, immovable assets such as homes and businesses
should be elevated to the appropriate flood elevation risk. This is the last line of
defense for immovable assets. Elevated homes also provide important side
benefits such as improved protection from termites and more economic capacity
to re-level or raise the houses due to settlement or increased flood risk. Example:
pre-1940 housing in New Orleans, LUMCON, Marina del Ray in Madisonville
11th Evacuation: Evacuation routes are typically highways, but could also
include other means of transportation such as railroads, air transportation, etc.
Evacuation routes are the last line of defense for people or moveable assets.
Evacuation routes and procedures should be established for the coast. Ideally,
evacuation routes may also serve as re-entry routes for first responders and as
routes to re-populate after a storm event. Evacuation routes are generally
selected based capacity to move a large number of people to safer areas as a
storm approaches the coast. Some routes may be subject to flooding quickly and
need to be improved. Examples: Regional contra-flow evacuation plan for
southeast Louisiana.
The Corps of Engineers and other organizations are continuing work to develop an
advanced and more reliable NOFDS that is more compatible with the natural, industrial and
social environments of southern Louisiana. The Working Group for PostHurricane Planning for
the Louisiana Coast recently concluded (2006):
In the long term, hurricane protection for larger population centers, including the
New Orleans region, can only be secured with a combination of levees and a
sustainable coastal landscape. This will require adapting to changing conditions
by re-establishing the constructive processes associated with distributing
Mississippi River water and sediments across the coastal landscape, as well as
alleviating the other destructive effects of past or future human activities.
With presently observed subsidence rates and anticipated acceleration of sealevel rise, most - although not all - of the coastal landscape could be maintained
through the 21st century. And with efficient management of the river's resources,
this landscape could be expanded in some places. However, this result can only
be achieved with very aggressive, strategic, and well-informed restoration efforts,
varying in size and objective but integrated within a landscape management plan.
The challenges associated with rehabilitation and improvement of the NOFDS need to be
addressed in an integrated way combining public and social, organizational and institutional,
natural and environmental, and commercial and industrial considerations. This is a “systems
problem” that has many parts which are interactive, interdependent, and highly adaptive. We
need to understand potential impacts, positive and negative, on the parts of this system so that
wise choices and informed decisions can be made on how best to proceed. This is a different
kind of “engineering problem” in which the Technology Delivery System used to address that
problem is of utmost importance. Gerald Galloway (2006) summarized these issues:
14 9
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Since 1983, when the Water Resources Council was effectively abolished, there
has been no central direction to or coordination of federal water efforts, among
the many departments that deal with water issues. Congress remains locked in a
turf-conscious committee system that does not encourage coordination. Except for
enforcing water quality standards there is little federal guidance, other than
budgetary or ad hoc initiatives, on other water issues.
Given the present policy vacuum and the reluctance on the part of Congress and
the administration to support comprehensive planning, New Orleans and coastal
Louisiana will have to develop, in coordination with federal agencies, their own
vision for the future and move ahead in a way that brings together solutions to the
many water challenges facing the region. this comprehensive plan must address
all aspects of coastal Louisiana's water challenges.
Each of the alternatives for development of a longterm NOFDS has its pluses and
minuses, costs and benefits. It is clear that these alternatives need to be continually examined in
an integrated and systematic way. The fundamental technology exists to develop an adequate
longterm NOFDS. The question is not “can we do it?” The question is “will we do it?”
14.3 Engineering Criteria and Guidelines
The basic technology exists to develop an effective and efficient NOFDS. A major
challenge is timely and proper application of this technology. The following recommendations
are made to facilitate such application.
Recommendation 1: Develop an integrated and coherent Flood Defense System for the
greater New Orleans area (NOFDS) to provide desirable and acceptable levels of flood
protection throughout its lifecycle. Particular attention must be paid to interfaces and
interdependencies in this system. The NOFDS should be balanced, complete, cohesive, clear,
consistent, and have controls and continuity. The NOFDS should be based on the best available
and safest technology and most uptodate legal standards. Risks should be properly identified,
contained and compartmentalized. The system must recognize the unique natural environmental
setting including its geology, meteorology, oceanography, the Mississippi River floodplains,
deltas and wetlands, subsidence, and the rise in sea level and frequency and intensity of
hurricanes. The system must also recognize and accommodate the unique societal and cultural
environments of this area.
Recommendation 2: Develop a NOFDS based on enhancing natural defenses
supplemented with engineered defenses that incorporate concepts of defenses in depth,
robustness or resilience, and failsafe performance. Selective reestablishment of natural coastal
defenses and wetlands and restored floodplains to provide for river floods should be
supplemented with engineering works that together have the capabilities of providing desirable
and acceptable levels of flood protection. Coastal management must be focused on providing
safety from flooding and environmental protection. Water should be given space. Some areas
will have to be returned to nature and judicious and wise decisions must be reached on which
areas will be populated and developed and the levels of protection that will be provided to these
areas. Engineering works should include raising, strengthening and defending levees, providing
floodgates and storm surge barriers, positioning and defending modern pump stations.
Engineering must also address compartmentation to limit potential flooding and adequate and
14 10
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
effective evacuation measures to help limit effects on people and their possessions. A robust
NOFDS will require a combination of appropriate configuration of engineered elements and
components, ductility or an ability to deform and stretch and not loose important performance
characteristics, excess capacity so that if some elements or components are overloaded or do not
perform desirably, desirable protection can be maintained, and appropriate correlation or mutual
relationships so that desirable protection is realized. Fail safe characteristics should be provided
in all of the important elements of the NOFDS so that when the design and ultimate performance
conditions are exceeded, the performance characteristics are not appreciably compromised.
Recommendation 3: Develop a NOFDS founded on advanced Risk Assessment and
Management principles for all phases in the lifecycle including concept development, design,
construction, operation, and maintenance. These principles should address natural, analytical
modeling, human and organizational performance, and knowledge acquisition and utilization
uncertainties and be based on proactive, reactive, and interactive risk assessment and
management approaches. These approaches should be based on reductions in likelihoods of
failure, reduction in the consequences associated with potential failures, and increases in
detection and correction of developments that can lead to failures. Advanced Risk Assessment
and Management approaches should be used to provide decision makers with information to
define what levels of protection should be provided for which areas to be protected and how
much can and should be spent for those purposes.
Recommendation 4: Develop updated engineering guidelines and procedures for all
elements and components to be incorporated in the FDS for all lifecycle phases based on proven
stateofpractice and stateofart technology. Where technology gaps are identified, substantial
development programs should be implemented to fill them with existing research results. Where
technology gaps can not be filled with existing research results, research should be undertaken or
sponsored to enable their timely filling.
Recommendation 5: Develop, implement, and enforce advanced Quality Assurance and
Quality Control methods and procedures for all lifecycle phases of the NOFDS. Quality
Assurance (proactive) and Quality Control (interactive) measures are of particular importance to
help disclose 'predictable surprises' and variances in the desirable quality characteristics of the
elements and components in the NOFDS. These methods and procedures should be used in all
lifecycle phases of the NOFDS including concept development, design, construction, operation,
maintenance, and continued improvement. These procedures and measures need to assure that
the best available and safest technology is used and used properly.
14.4 References
Coleman, J. M. (1988). “Dynamic Changes and Processes in the Mississippi River Delta.”
Geological Society of American Bulletin, 100 (7).
Committee on the Restoration and Protection of Coastal Louisiana (2006). Drawing Louisiana's
New Map: Addressing Land Loss in Coastal Louisiana, Ocean Studies Board, National
Research Council, The National Academies Press, Washington, DC,
<http://www.nap.edu> (May 1, 2006).
Dean, R. G. (2006). “New Orleans and the Wetlands of Southern Louisiana.” The Bridge,
National Academy of Engineering, Washington DC, <http://www.national
academies.org> (May 1, 2006).
14 11
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Galloway, G. E. (2006). “Restoring Coastal Louisiana: Planning without a National Water
Policy,” The Bridge, National Academy of Engineering, Washington DC,
<http://www.nationalacademies.org> (May 1, 2006).
Gingrich, N. (2006). “Why New Orleans Needs Saving.” Time Magazine, New York,
03/06/2006.
Hallowell, C. (2005). Holding Back the Sea, The Struggle On the Gulf Coast to Save America.
Harper Collins Publishers, New York, NY.
Houck, O. A. (2005). “Chapter 5: Environmental Protection and Sustainable Development.”
Report to Mayor Nagin's Bring New Orleans Back Commission: An Alternative Vision
for Rebuilding, Redevelopment & Reconstruction, From the Lake to River Foundation,
New Orleans, LA, <http://www.fromthelaketotheriver.org> (May 9, 2006).
Houck, O.A. (2006). “Can We Save New Orleans?” Tulane Environmental Law Journal, 19(1),
<http://www.law.tulane.edu/tuexp/journals> (Apr. 25, 2006).
Kelman, A. (2003). A River and Its City, The Nature of Landscape in New Orleans, University of
California Press, Berkeley, CA.
Lopez, J.A. (2006). The Multiple Lines of Defense Strategy to Sustain Louisiana's Coast, Lake
Pontchartrain Basin Foundation, Metairie, LA, <http://www.saveourlake.org> (May 9,
2006).
Louisiana Coastal Wetlands Conservation and Restoration Task Force (1998). Coast 2050:
Toward a Sustainable Coastal Louisiana, Louisiana Department of Natural Resources,
Baton Rouge LA.
National Research Council (2006). Drawing Louisiana's New Map, Addressing Land Loss in
Coastal Louisiana, Committee on the Restoration and Protection of Coastal Louisiana,
Ocean Studies Board, The National Academies Press, Washington DC.
Netherlands Water Partnership (2005). Dutch Expertise Water Management & Flood Control,
Delft, the Netherlands, <http://www.nwp.nl> (Feb. 3, 2006).
Sparks, R. E. (2006). “Rethinking, Then Rebuilding New Orleans.” Issues in Science and
Technology, National Academy Press, Washington DC, Winter, 3339.
U.S. Army Corps of Engineers (2004). Louisiana Coastal Area (LCA), Louisiana Ecosystem
Restoration Study, Final, Vol. 1, LCA Study Main Report, New Orleans District.
Working Group for PostHurricane Planning for the Louisiana Coast (2006). A New Framework
for Planning the Future of Coastal Louisiana after the Hurricanes of 2005, University of
Maryland Center for Environmental Science, <http://www.umces.edu/larestore> (Apr.
23, 2006).
Zinn, J. (2004). Coastal Louisiana: Attempting to Restore and Ecosystem, Congressional
Research Service Report for Congress, Washington DC, October.
Zinn, J. (2005a). Coastal Louisiana Ecosystem Restoration: The Recommended Corps Plan,
Congressional Research Service Report for Congress, Washington DC, April.
Zinn, J. (2005b). Hurricanes Katrina and Rita and the Coastal Louisiana Ecosystem Restoration,
Congressional Research Service Report for Congress, Washington DC, September.
14 12
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Figure 14.1: Eleven Lines of Defense (Lopez, 2006; graphic provided by the New Orleans
Times Picayune)
14 13
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
CHAPTER FIFTEEN: FINDINGS AND RECOMMENDATIONS
15.1 Overview
This report presents the results of an investigation of the performance of the New
Orleans regional flood protection system during and after Hurricane Katrina, which struck the
New Orleans region on August 29, 2005. This event resulted in the single most costly
catastrophic failure of an engineered system in history. Current damage estimates at the time
of this writing are on the order of $100 to $200 billion in the greater New Orleans area, and
the official death count in New Orleans and southern Louisiana at the time of this writing
stands at 1,293, with an additional 306 deaths in nearby southern Mississippi. An additional
approximately 300 people are currently still listed as “missing”, and the death toll is expected
to continue to rise a bit further. More than 450,000 people were initially displaced by this
catastrophe, and at the time of this writing more than 200,000 residents of the greater New
Orleans metropolitan area continue to be displaced from their homes by the floodwater
damages from this storm event.
This investigation has targeted three main questions as follow: (1) What happened?,
(2) Why?, and (3) What types of changes are necessary to prevent recurrence of a disaster of
this scale again in the future?
In the end, it is concluded that many things went wrong with the New Orleans flood
protection system during Hurricane Katrina, and that the resulting catastrophe had it roots in
three main causes: (1) a major natural disaster (the Hurricane itself), (2) the poor performance
of the flood protection system, due to localized engineering failures, questionable judgments,
errors, etc. involved in the detailed design, construction, operation and maintenance of the
system, and (3) more global “organizational” and institutional problems associated with the
governmental and local organizations responsible for the design, construction, operation,
maintenance and funding of the overall flood protection system.
15.2
Performance of the Regional Flood Defense System During Hurricane Katrina
As Hurricane Katrina initially approached the coast, the resulting storm surge and
waves rose over the levees protecting much of a narrow strip of land on both sides of the
lower Mississippi River extending from the southern edge of New Orleans to the Gulf of
Mexico. Most of this narrow protected zone, Plaquemines Parish, was massively inundated
by the waters of the Gulf.
The eye of the storm next proceeded to the north, on a path that would take it just
slightly to the east of New Orleans.
Hurricane Katrina has been widely reported to have overwhelmed the eastern side of
the New Orleans flood protection system with storm surge and wave loading that exceeded
the levels used for design of the system in that area. That is a true statement, but it is also an
incomplete view. The storm surge and wave loading at the eastern flank of the New Orleans
flood protection system was not vastly greater than design levels, and the carnage that resulted
15 1
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
owed much to the inadequacies of the system as it existed at the time of Katrina’s arrival.
Some overtopping of levees along the eastern flank of the system (along the northeastern
frontage of the St. Bernard and Ninth Ward protected basin, and at the southeast corner of the
New Orleans East protected basin), and also in central areas (along the GIWW channel and
the IHNC channel) was inevitable given the design levels authorized by Congress and the
surge levels produced in these areas by the actual storm. It does not follow, however, that this
overtopping had to result in catastrophic failures and breaching of major portions of the levees
protecting these areas, nor the ensuing catastrophic flooding of these populous areas.
The northeast flank of the St. Bernard/Ninth Ward basin’s protective “ring” of levees
and floodwalls was incomplete at the time of Katrina’s arrival. The critical 11 mile long levee
section fronting “Lake” Borgne (which is actually a Bay, connected directly to the Gulf of
Mexico) was being constructed in stages, and funding appropriation for the final stage had
long been requested by the U.S. Army Corps of Engineers (USACE), but this did not arrive
before Katrina struck. As a result, large portions of this critical levee frontage were several
feet below final design grade. In addition, an unfortunate decision had been made to use local
dredge spoils from the excavation of the adjacent MRGO channel for construction of major
portions of the levees along this frontage. The result was that major portions of these levees
were comprised of highly erodable sand and lightweight shell sand fill.
When the storm surge arrived, massive portions of these levees eroded
catastrophically and the storm surge passed through this frontage while still on the rise,
crossed an open swamp area that should have safely absorbed most of the overtopping flow
from the outer levees (if they had not catastrophically eroded), and it then crossed easily over
a secondary levee of lesser height that had not been intended to face a storm surge largely
undiminished by the minimal interference of the too rapidly eroded outer levees fronting Lake
Borgne. The resulting carnage in St. Bernard Parish was devastating, as the storm surge
rapidly filled the protected basin to an elevation of approximately +12 feet above sea level;
deeply inundating even neighborhoods with ground elevations well above sea level in this
area.
The stormsurgeswelled waters of Lake Borgne also passed over and then through a
length of levees at the southeast corner of the New Orleans East protected basin. Here too,
the levees fronting Lake Borgne had been constructed in part using materials dredged from
the excavation of an adjacent shipping channel (the GIWW channel), and these levees also
contained significant volumes of highly erodable sands and lightweight shell sands. These
levees also massively eroded, and produced the principal source of flooding that eventually
inundated the New Orleans East protected area. Here again, there was an area of undeveloped
swampland behind the outer levees that might have helped to absorbed the brunt of any
overtopping flow, and a secondary levee of lesser height was in place behind this swampland
that might then have prevented or at least greatly slowed and reduced the catastrophic
flooding of the populous areas of New Orleans East. This secondary levee was not able to
resist the massive flows resulting from the catastrophic erosion of the highly erodable sections
of the Lake Borgne frontage levee, however, and some of the eroded and breached frontage
levees allowed waters to bypass the secondary levee line. As a result, the floodwaters from
the breaches and eroded sections of levee at the southeast corner of the New Orleans East
protected area passed inland and began the filling of the New Orleans East protected basin.
15 2
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
The catastrophic erosion of these two critical levee frontages need not have occurred.
These frontages could instead have been constructed using well compacted clay fill with good
resistance to erosion, and they could have been further armored in anticipation of the storm
surge and wave loading from Lake Borgne. The levee at the northeast edge of St. Bernard
Parish could have been completed in a more timely manner. The result would have been
some overtopping, but not catastrophic erosion and uncontrolled breaching of these critical
frontages. Some flooding and damage would have been expected, but it need not have been
catastrophic.
The storm surge swollen waters of Lake Borgne next passed laterally along the east
west trending GIWW/MRGO channel to its intersection at a “T” with the northsouth oriented
IHNC channel, overtopping levees along both banks to a limited degree. This produced an
additional breach of a composite earthen levee and concrete floodwall section (at a transition
to a full earthen levee section) along the southern edge of New Orleans East, adding
additional uncontrolled inflow to this protected basin. This failure could have been prevented
at little incremental cost if erosion protection (e.g. a concrete splash pad, or similar) had been
emplaced along the back side of the concrete floodwall at the levee crest, but the USACE felt
that this was precluded by Federal rules and regulations regarding authorized levels of
protection.
The surge next raised the water levels within the IHNC channel, and produced a
number of failures on both the east and west banks. Two major failures occurred on the east
side of the IHNC, at the west edge of the Ninth Ward. Overtopping occurred at both of these
locations, but this was not the principal cause of either of these failures. Both failures were
principally due to underseepage flows that passed beneath the sheetpile curtains supporting
the concrete floodwalls at the crests of the levees. Like many sections of the flood protection
system, these sheetpiles were too shallow to adequately cut off, and thus reduce, these
underseepage flows. The result was two massive breaches that devastated the adjacent Ninth
Ward neighborhood, and then pushed east to meet with the floodwaters already rapidly
approaching from the east from St. Bernard Parish as a result of the earlier catastrophic
erosion of the Lake Borgne frontage levees.
Several additional breaches also occurred farther north on the east side of the IHNC
fronting the west side of New Orleans East, but these were relatively small features and they
just added further to the uncontrolled flows that were now progressively filling this protected
basin. These breaches occurred mainly at junctures between adjoining, dissimilar levee and
floodwall sections, and represented good examples of widespread failure to adequately
engineer these “transitions” between sections of the regional flood protection system.
Several breaches occurred on the west side of the IHNC, and these represented the
first failures to admit uncontrolled floodwaters into the main metropolitan (downtown)
protected area of New Orleans. These features did not scour and erode a path below sea level,
however, so they admitted floodwaters for a number of hours and then these inflows ceased as
the storm surge in the IHNC eventually subsided. Only 10% to 20% of the floodwaters that
eventually inundated a majority of the main (downtown) New Orleans protected basin entered
through these features.
15 3
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
These failures and breaches on the west side of the IHNC all appear to have been
preventable. One failure was the result of overtopping of an Iwall, with the overtopping flow
then eroding a trench in the earthen levee crest at the inboard side of the floodwall. This
removal of lateral support unbraced the floodwall, and it was pushed over laterally by the
water pressures from the storm surge on the outboard side. Here again the installation of
erosional protection (e.g. concrete splash pads or similar) might have prevented the failure.
The other failures in this area occurred at “transitions” between disparate levee and
floodwall sections, and/or at sections where unsuitable and highly erodable lightweight shell
sand fills had been used to construct levee embankments. Here, again, these failures were as
much the result of design choices and/or engineering and oversight issues as the storm surge
itself.
Particularly frustrating were a pair of failures on the east and west banks of the IHNC
where the CSX railroad crossed the IHNC. These two sites both breached as a result of
improper detailing of the intersections between the railroad tracks and their support gravel
ballast, and adjacent roadways also crossing the federal levees at these same two locations.
These represented additional examples of repeated problems associated with coordination,
design, and oversight of complex “intersections” wherein multiple agencies and utilities
(including roadways, rail lines, etc.) intersect the protective levee system. Frustratingly, it is
noted that these same two rail crossings at the east and west sides of the IHNC had also failed
and breached in 1965, during hurricane Betsy.
As the eye of the hurricane next passed to the northeast of New Orleans, the
counterclockwise swirl of the storm winds produced a storm surge against the southern edge
of Lake Pontchartrain. This produced additional temporary overtopping of a long section of
levee and floodwall at the west end of the lakefront levees of New Orleans east, behind the
old airport, adding further to the flows that were progressively filling this protected basin.
The surge against the southern edge of Lake Pontchartrain also elevated the water
levels within three drainage canals at the northern edge of the main metropolitan (downtown)
New Orleans protected basin, and this would produce the final, and most damaging, failures
and flooding of the overall event.
The three drainage canals should not have been accessible to the storm surge. The
USACE had tried for many years to obtain authorization to install floodgates at the north ends
of the three drainage canals that could be closed to prevent storm surges from raising the
water levels within the canals. That would have been the superior technical solution.
Dysfunctional interaction between the local Levee Board (who were responsible for levees
and floodwalls, etc.) and the local Water and Sewerage Board (who were responsible for
pumping water from the city via the drainage canals) prevented the installation of these gates,
however, and as a result many miles of the sides of these three canals had instead to be lined
with levees and floodwalls.
The lining of these canals with levees topped with concrete floodwalls was rendered
very challenging due to (a) the difficult local geology of the foundation soils, and (b) the
narrow right of way (or available “footprint”) for these levees. As a result of the decision not
15 4
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
to install the floodgates, the three canals represented potentially vulnerable “daggers” pointed
at the heart of the main metropolitan New Orleans protected basin. Three major breaches
would occur on these canals; two on the London Avenue Canal and one on the 17th Street
Canal. All three of these breaches eroded and scoured rapidly to well below sea level, and
these three major breaches were the source of approximately 80to 85% of the floodwaters that
then flowed into the main (downtown) protected basin over the next three days, finally
equilibrating with the still slightly elevated waters of Lake Pontchartrain on Thursday,
September 1.
The central canal of the three, the Orleans Canal, did not suffer breaching, but a
section of floodwall topping the earthen levee approximately 200 feet in length near the south
end of the canal had been left incomplete, again as a result of dysfunctional interaction
between the local levee board and the water and sewerage board. This effectively reduced the
level of protection for this canal from about +12 to +13 feet above sea level (the height of the
tops of the floodwalls lining the many miles of the canal) to an elevation of about +7 feet
above sea level (the height of the earthen levee crest along the 200 foot length where the
floodwall that should have topped this levee was omitted). As a result of the missing
floodwall section, flow passed through this “hole” and began flowing into the heart of the
main New Orleans protected basin. This flow eventually ceased as the storm surge subsided,
and so was locally damaging but not catastrophic.
The three breaches on the 17th Street and London Avenue canals were catastrophic.
None of these failures were the result of overtopping; surge levels in all three drainage canals
were well below the design levels, and well below the tops of the floodwalls. Two of these
breaches were the result of stability failures of the foundation soils underlying the earthen
levees and their floodwalls, and the third was the result of underseepage passing beneath the
sheetpile curtain and resultant catastrophic erosion near the inboard toe of the levee that
eventually undermined the levee and floodwall.
A large number of engineering errors and poor judgments contributed to these three
catastrophic design failures, as detailed in Chapter 8. In addition, a number of these same
problems appear to be somewhat pervasive throughout other areas of the New Orleans
regional flood defense system(s), and call into question the integrity and reliability of other
sections of the regional flood protection system that did not fail during this event. Indeed,
additional levee and floodwall sections along the drainage canals appear to have been
potentially heading towards failure when they were “saved” by the occurrence of the three
large breaches (which rapidly drew down the canal water levels and thus reduced the loading
on nearby levee and floodwall sections.)
15.3 Engineering Issues
The New Orleans regional flood protection system failed at many locations during
Hurricane Katrina, and by many different modes and mechanisms. This unacceptable
performance can in many cases be traced to engineering lapses, poor judgments, and efforts to
reduce costs at the expense of system reliability. These, in turn, were to a large degree the
result of more global underlying “organizational” and institutional problems associated with
the governmental and local organizations jointly responsible for the design, construction,
15 5
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
operation, and maintenance of the flood protection system, including provision of timely
funding and other critical resources.
Our findings to date indicate that no one group or organization had a monopoly on
responsibility for the catastrophic failure of this regional flood protection system. Many
groups, organizations and even individuals had a hand in the numerous failures and
shortcomings that proved so catastrophic on August 29th. It is a complex situation, without
simple answers.
It is not without answers and potential solutions, however, just not simple ones. There
is a need to change the process by which these types of large and critical protective systems
are created and maintained. It will not be feasible to provide an assured level of protection for
this large metropolitan region without first making significant changes in the organizational
structure and interactions of the national and more local governmental bodies and agencies
jointly responsible for this effort. Significant changes are also needed in the engineering
approaches and procedures used for many aspects of this work, for the standards used in such
design, in the conceptual approaches considered, and in the conceptualization and engineering
treatment of potential modes of failure and poor performance during design, construction and
operation. There is also a need for interactive and independent expert technical oversight and
review as well. In numerous cases, it appears that such review would have likely caught and
challenged errors and poor judgments (both in engineering and in policy and funding) that led
to failures during Hurricane Katrina.
There are many detailed engineering lessons developed within this report, but a
number of overarching engineering issues have been identified, and a number of the most
important of these are presented below. These are a somewhat urgent set of issues, as the
USACE and the IPET investigation are currently working to assess the level of risk associated
with the now largely reconstructed system, and these issues impact that assessment.
1. Overall levels of safety and reliability targeted during engineering design and analysis
were inappropriately low for a critical system protecting a major metropolitan area.
Factors of safety and analysis methods and procedures used in design calculations for
the “transient” loading conditions associated with hurricaneinduced storm surge,
coupled with the design surge elevations employed, provided levels of risk that were
on the order two to three orders of magnitude higher than the standards generally used
in U.S. dam practice where similarly large populations are at potential risk. This left
too little room for error, uncertainties, or surprises.
2. The difficult and complex geology of the region posed design challenges that were
not adequately addressed. Insufficient site investigation and characterization of
foundation soil conditions at many sites produced minor shortterm project savings,
but these pale against the massive losses that ensued. More attention needs to be paid
to the geology, and more detailed site investigation and site characterization is clearly
warranted given the potential consequences of failures.
15 6
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
3. There was a persistent pattern of attempts to reduce costs of constructed works, at the
price of corollary reduction in safety and reliability. This represented a policy that
has now been shown to be massively “penny wise and pound foolish”.
4. A pattern of optimistic engineering assessment with regard to a number of potential
sources of risk and of potential modes of failure was endemic to the detailed design of
a number of major system elements. This included:
(a) The risks associated with underseepage flows during “transient” storm surges
were systematically underestimated. This led to the use of sheetpile curtains that
were extended to inadequate depths at a number of locations, and it led directly to
a number of the major failures and breaches during hurricane Katrina.
Appropriate consideration and analysis of underseepage issues (including
potential embankment instability due to pore pressure induced strength reduction,
and potential erosion and piping) for transient storm surge conditions was
routinely missing, and the overall system should now be reevaluated with regard
to these underseepagerelated potential modes of failure.
(b) The use of highly erodable sand and even lightweight shell sand fills in levee
sections also figured prominently at numerous locations of breaching and
catastrophic erosion. Use of such materials should henceforth be disallowed in
this system that protects a major metropolitan region. Here again, the overall
system should be reevaluated for their presence, and the levels of risk posed by
the presence of these unsuitable materials, both in levee embankments and at
shallow depths within the underlying foundation soils; and this risk should be
mitigated.
(c) Similarly, design procedures did not include consideration of the potential failure
mode that involves formation of a ‘gap’ at the outboard side of the floodwalls,
between the outboard section of the earthen levee embankment ant the sheetpile
curtains supporting the floodwalls. Formation of such gaps occurred at a number
of sites as pressure increased on the outboard sides of the floodwalls, and water
then intruded into the gaps and greatly increased the lateral “push” of the storm
surge (water) against the sheetpile/floodwalls. A number of failures occurred as a
result. In the future, such gapping should be “assumed” during analysis and
design. Many of the “Iwall” type concrete floodwalls are currently being
removed and replaced by the more robust “Twall” type floodwalls (which have
additional battered piles to help then resist overturning and lateral displacement.)
These Twall systems will have somewhat increased capacity, but they too will
need to be analyzed with regard to this potential failure mode. It cannot simply
be assumed that “Twalls” are intrinsically completely safe.
5. Design review was generally inadequate, and there was an institutional failure to catch
and challenge unconservative design assumptions and interpretations of data,
misconceptions, poor judgements, and errors. Instigation of interactive consultation
and review by consulting panels of leading outside experts is common practice in dam
engineering. It should be common practice in levee engineering as well; especially
when the levee systems protect significant populations. In addition, it would be wise
15 7
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
for local interests (e.g. the State and/or the City) to mount an additional unbiased
expert review panel (again including leading outside experts) to provide a second
check and opinion. At many failure sites it appears likely that suitable expert review
would have caught and challenged errors and questionable judgments that contributed
to the failures observed.
6. Improved advantage needs to be taken of ongoing technical advances related to the
engineering, design and construction of these types of regional flood defense systems.
Engineering design concepts and analysis approaches employed at many locations
were sorely “outdated” at the time of their use, and there was a lack of movement
towards embracing new and improved methods and tools. “This is how we have
always done it” is a potentially dangerous concept here, and inertia in terms of
embracing technical advances was a troubling issue. Failure to embrace their own full
scale field testing and research led the Corps to neglect the “waterfilled gap” as a
potential failure mode to be addressed during design. And it is time to relegate the
“Method of Planes” to its place in history and to adopt more modern and more flexible
stability analysis methods capable of addressing a wider range of geometries and
potential failure modes.
7. The USACE is the lead oversight agency with regard to engineering and construction
of the regional flood defense system. The Corps needs to be allocated adequate
funding and support, given the ability to perform research, and granted adequate
freedom and support to facilitate the continuing professional development (and
retention) of highly qualified engineers within the Corps in order to ensure an adequate
inhouse supply of engineering expertise for their critical role.
15.4 Looking Back - Organized for Failure
The ILIT mandate at the outset of this investigation study was to include study of
historical and organizational institutional issues, political and budgetary considerations,
decision making, utilization of technology, and the evolving societal, governmental, and
organizational priorities over the life of the Flood Defense System for the Greater New
Orleans Area (NOFDS). One cannot understand the failure of the NOFDS without
understanding both the underlying engineering and organizational mechanics that were
interwoven in the evolution of this failure.
ILIT's view of the importance of these organizational, institutional, resource and
technology delivery factors increased during the course of this study. These factors are
grouped into what is termed a Technology Delivery System (TDS). A TDS can be represented
as system that has organizational components, inputs, outputs, and information linkages that
are interactive, interdependent, and adaptive. Three primary organizational components
comprise a TDS for a system such as the NOFDS. These are: (1) society (the public), (2)
government (federal, state, local), and (3) enterprise (commercial, industrial, private). These
components are embedded in and interact with their natural and cultural environments. Inputs
comprise knowledge plus human, natural, and fiscal resources. Outputs include desired goods
or services and undesired outcomes or unintended consequences.
15 8
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Eight principal categories of TDS malfunctions were identified that played major roles
in the catastrophic failure of the NOFDS, and these are as follow:
Failures of foresight: Catastrophic flooding of the greater New Orleans area due to
surge from an intense hurricane was predicted for several decades. The consequences
observed in the wake of hurricane Katrina were also predicted. The hazards associated with
the NOFDS were not adequately recognized, defensive measures were not identified and
prioritized, and effective action was not mobilized to effectively deal with these hazards.
Failures of organization: The roots of the failure of the NOFDS are firmly embedded
in flawed organizational institutional systems. The organizational institutional systems
lacked centralized and focused responsibility and authority for providing adequate flood
protection. There were dramatic and pervasive failures in management represented in
ineffective and inefficient planning, organizing, leading, and controlling to achieve desirable
quality and reliability in the NOFDS. There were extensive and persistent failures to
demonstrate initiative, imagination, leadership, cooperation, and management.
Failures of resource allocation: Contributing to the failure of the NOFDS was
provision of inadequate resources based primarily on recommendations provided by the Corps
of Engineers. This was followed by failure of the federal and state governments to fund badly
needed improvements once limitations were recognized. In a number of instances, State
and/or local agencies pressured for 'lower cost' solutions not realizing that these solutions
would result in lowering the overall quality and reliability of the NOFDS. There were
important deficiencies in the cost benefit analyses used to justify the levels of protection
(and reliability) provided, and also the continued improvement in these levels of protection
(and reliability) as knowledge and technology advanced.
Failures of diligence: Forty years after the devastating flooding caused by hurricane
Betsy, the flood protection system authorized in 1965 and based on the Standard Project
Hurricane (SPH) was still not completed when hurricane Katrina arrived. In addition, the
concept and application of the SPH was recognized to be seriously flawed, yet there were no
adjustments made to the system to address this before Katrina struck. Early warning signs of
deficiencies and flaws persisted throughout progressive development and construction of the
different components that comprised the NOFDS, and these warning signs were not
adequately evaluated and acted upon.
Failures of decision making: The history of this system was marked by a series of
flawed decisions and tradeoffs that proved to be fatal to the ability of the system to perform
adequately. Compromises in the ability of this system to perform adequately started with the
decisions regarding the fundamental design criteria for the development of the system, then
were propagated through time as alternatives for the system were evaluated and engineered.
Design, construction, operation, and maintenance of the system in a piecemeal fashion
allowed the introduction of additional flaws and defects. Efficiency was traded for
effectiveness. Superiority in provision of an adequate NOFDS was traded for mediocrity,
lower expenditures, and getting along.
15 9
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
Failures of management: Requirements imposed on the Corps of Engineers by
Congress, the White House, State and local agencies, and the general public have changed
dramatically during the past three decades. Defense, reconstruction, maintenance, waste
disposal, recreational development, emergency response, and ecological restoration have
served to divert attention from flood control. Public and Congressional pressures to (1) reduce
backlogs of approved projects, (2) improve project and organizational efficiency (e.g.: down
sizing, outsourcing, etc.), (3) address environmental impacts, and (4) develop appropriations
for projects have served to divert attention from engineering quality and reliability of flood
control. Engineering technology leadership, competency, expertise, research, and
development capabilities appear to have been sacrificed for improvements in project planning
and controlling.
Failures of synthesis: While individual parts of a complex system can be adequate,
when these parts are joined together to form an interactive interdependent adaptive system,
unforseen failure modes can be expected to develop. These unforseen, but forseeable, failure
modes did develop in the NOFDS during hurricane Katrina. It is evident that insufficient
attention was given to creation of an integrated series of components to provide a reliable
overall NOFDS. Synthesis was subverted to decomposition, as projects were engineered and
constructed in piecemeal fashion to conform to incremental appropriations. As a result, many
failures developed at interfaces or 'transitions' in the NOFDS.
Failures of risk assessment and risk management: The risks (likelihoods and
consequences) associated with hurricane surge and wave induced flooding were seriously
underestimated. There was inadequate recognition of the primary contributors to the
likelihoods and consequences of catastrophic flooding. Sufficient defensive measures to
counteract and mitigate these uncertainties were not employed. Factors of safety used in
design of the primary elements in the NOFDS were not sufficient; and represented implicit
levels of system reliability that were inappropriately low for a system protecting a major
metropolitan region. Quality assurance and quality control measures invoked during the life of
the system failed to disclose critical flaws in the system. Inappropriate use was made of
existing engineering technology available to design, construct, operate, and maintain a
NOFDS that would have acceptable quality and reliability. Deficient risk management
methods were used to allocate resources and impel action to properly manage risks. Risk
management failed to employ continuing improvement, monitoring, assessment, and
modifications in means and methods which were discovered to be ineffective.
15.5 Looking Forward - Organizing for Success
The following recommendations are offered for consideration in developing a NOFDS
that will have desirable and acceptable quality and reliability. These recommendations are
divided into two categories: engineering developments and organizational developments. It
will take both, working together, to realize the desired goals of an appropriately improved
NOFDS. The primary challenge is timely mobilization of inspired and inspiring leadership,
adequate resources, existing technology, and high reliability organizations.
15 10
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
15.5.1 Strategic and Engineering System Issues:
The technology exists that can be used to develop a NOFDS that will be effective and
efficient. A major challenge is timely and proper application of this technology.
Recommendation 1: Develop an integrated and coherent Flood Defense System for
the greater New Orleans area (NOFDS) that will provide desirable and acceptable levels of
flood protection throughout its lifecycle. Particular attention must be paid to interfaces and
interdependencies in this system. The NOFDS should be balanced, complete, cohesive, clear,
consistent, and have controls and continuity. The NOFDS should be based on the best
available and safest technology and most uptodate legal standards. Risks should be properly
identified, contained and compartmentalized. The system must recognize the unique natural
environmental setting including its geology, meteorology, oceanography, the Mississippi
River floodplains, deltas and wetlands, subsidence, and the rise in sea level and frequency and
intensity of hurricanes. The system must also recognize and accommodate the unique societal
and cultural environments of this area.
Recommendation 2: Develop a NOFDS based on enhancing natural defenses
supplemented with engineered defenses that incorporate concepts of defenses in depth,
robustness or resilience, and failsafe performance. Selective reestablishment of natural
coastal defenses and wetlands, and restored floodplains to provide for river floods, should be
supplemented with engineering works that together will have the capabilities of providing
desirable and acceptable levels of flood protection. Coastal management must be focused on
providing safety from flooding and environmental protection. Water should be given space.
Some areas will have to be returned to nature, and judicious and wise decisions will have to
be reached regarding which areas will be populated and developed and the levels of protection
that will be provided to these areas. Engineering works should include: (1) raising,
strengthening, improving the reliability, and improvement of the erosion resistance of levees,
(2) provision of floodgates, and storm surge barriers, (3) improved positioning and defense of
modern pump stations, (4) compartmentation to limit potential flooding consequences, and (5)
adequate and effective evacuation measures to help limit effects on people and their
possessions. A robust NOFDS will require a combination of appropriate configuration of
engineered elements and components, ductility or an ability to deform and stretch and not lose
important performance characteristics (e.g. the ability to overtop for some limited period of
time without catastrophic breaching), and provision of excess capacity so that if some
elements or components are overloaded or do not perform desirably then desirable protection
can still be maintained. Fail safe characteristics should be provided in all of the important
elements of the NOFDS so that when the design and ultimate performance conditions are
exceeded, the performance characteristics are not excessively compromised.
Recommendation 3: Develop a NOFDS founded on advanced Risk Assessment and
Risk Management principles for all phases in the lifecycle including concept development,
design, construction, operation, and maintenance. These principles should address natural
processes, analytical modeling, human and organizational performance, and knowledge
acquisition and utilization uncertainties and be based on proactive, reactive, and interactive
risk assessment and management approaches. These approaches should be based on
reductions in likelihoods of failure, reduction in the consequences associated with potential
15 11
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
failures, and improvements in detection and correction of developments that can lead to
failures. Advanced Risk Assessment and Risk Management approaches should be used to
provide decision makers with information to define what levels of protection should be
provided for which areas, and how much can and should be spent for those purposes.
Recommendation 4: Develop updated engineering guidelines and procedures for all
elements and components to be incorporated in the FDS for all lifecycle phases based on
proven stateofpractice and stateofart technology. Where technology gaps are identified,
then substantial development programs should be implemented to fill these gaps with existing
research results. Where technology gaps cannot be filled with existing research results, then
research should be undertaken or sponsored to enable timely filling of the technology gaps.
Upgrading the technical capabilities of the engineers responsible for oversight and design, and
the use of interactive boards of consultants as well as expert external review boards, would
likely greatly improve the ability to deliver reliable flood protection.
Recommendation 5: Develop, implement, and enforce advanced Quality Assurance
and Quality Control methods and procedures for all lifecycle phases of the NOFDS. Quality
Assurance (proactive) and Quality Control (interactive) measures are of particular importance
to help disclose 'predictable surprises' and variances in the desirable quality characteristics of
the elements and components in the NOFDS. These methods and procedures should be used
in all lifecycle phases of the NOFDS including concept development, design, construction,
operation, maintenance, and continued improvement. These procedures and measures need to
assure that the best available and safest technology is being used and used properly.
15.5.2 Technology Delivery System Developments - Organizing for Success
It will not be feasible to create an adequately reliable regional Flood Defense System
without addressing the organizational, institutional, political and resources issues that
adversely affect the current process. Simply changing engineering procedures, design
manuals, and the review process will not suffice.
The primary requirement for reconstitution of a Technology Delivery System that can
and will provide an adequate and acceptably reliable NOFDS is mobilization of the 'will' to
provide such a system. If the United States decides that the catastrophe of Katrina will not be
repeated, then the necessary leadership, organization, management, resources, and public
support must be mobilized to assure such an outcome. One of the primary challenges is time;
the clock is ticking until this area of the United States is again confronted with a severe
challenge of flooding.
Recommendation 1: Seriously consider defining risk in the framework of federal,
state, and local government responsibilities to protect their citizens.
Recommendation 2: Exploit the major and unprecedented role that exists for citizens
who should be considered part of governance in the spirit that those who govern do so at the
informed consent of the governed. This is the population exposed to catastrophic risks and the
people that will be protected by the NOFDS. Authorities responsible for catastrophic risk
management should ensure that those vulnerable have sufficient and timely information
regarding their condition, and a reciprocal ability to respond to requests for their informed
15 12
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
consent especially regarding tradeoffs of safety for cost. The public protected by the NOFDS
need to be encouraged to actively and intelligently interact with its development.
Recommendation 3: Intensify, focus, and fund Corps of Engineers reorganization
and modernization efforts directed toward (1) increasing and maintaining inhouse
engineering capabilities and project performance, (2) increasing inhouse research and
development capabilities, (3) increasing inhouse engineering performance on technically
challenging projects, (4) developing an organizational culture of high reliability founded on
existing organizational cultural values of Duty, Honor, Country, and (5) developing a
leadership role and responsibility for technical and management oversight of all phases in
development of a NOFDS. Technical superiority must be reestablished. Outsourcing must
be balanced with insourcing to encourage development and maintenance of superior
technical leadership and capabilities within the USACE. This will require close and
continuous collaboration of federal legislative, executive, and judicial agencies. This will
require that the USACE reconceptualize itself as a pivotal part of a modular organization
developing partnerships with other federal agencies, state and local governments, enterprise
interests, and private stake holders. This will require additional funding; in the end the nation
will get only what it is willing to invest and pay for.
Recommendation 4: Restructure federal/state relationships in flood control. One
possible model is what has been called “modularity” a concept which involves provisional
and functional rearrangement of units in terms of alternative configurations of tools, structures
and relationships.
Recommendation 5: Develop a National Flood Defense Authority (NFDA) charged
with oversight over the design, construction, operation and maintenance of flood control
systems. Each state would have an equivalent organization that could foster cooperation and
developments between and within the states. The Corps of Engineers, state flood control
authorities, and technical advisory boards would work with the NFDA to foster application of
the best available technology and help coordinate development and maintenance efforts and
planning. In cooperative developments, federal and state governments would provide reliable
and sustainable funding for the lifecycle of specific flood defense systems. This development
should be accompanied by development of an integrated and coherent Louisiana Flood
Defense Authority representing state, regional, local, city, and public stakeholders that can
focus and prioritize stakeholder interests and requirements and collaborate with the Corps of
Engineers in development of a NOFDS.
Recommendation 6: Because of the importance of emergency response in the
NOFDS, FEMA should be developed as a high reliability organization and returned by the
executive branch to Cabinet level status. A new Council for Catastrophic Risk Management
should be appointed in the White House and given oversight of disaster preparation and
response. A similar body should be appointed to Congress. Incentives must be created to
encourage all levels of government to responsibly deal with potential national, regional, and
local catastrophes.
15 13
New Orleans Levee Systems
Hurricane Katrina
July 31, 2006
Independent Levee
Investigation Team
15.6 Conclusion
The performance of the New Orleans regional flood protection system during
hurricane Katrina was unacceptable. Detailed study has now led to understanding of the
physical causes and mechanisms of most of the many failures and breaches, and this in turn
provides a basis for development of improved conceptual and engineering design methods, as
well as improved review and overview paradigms.
Simply addressing engineering design methods, standards and procedures is unlikely
to be sufficient to provide a suitably reliable level of protection, however. There is also a
need to resolve difficult issues intrinsic in the operations and relationships between (1)
Federal and more local government as they serve as decisionmaking, policy and funding
sources, (2) the Federal and local agencies responsible for the actual design, construction,
operation and maintenance of such flood protection systems, and (3) private enterprise that
must assist in construction. Some of these groups need to enhance their technical capabilities;
a longterm expense that would clearly represent a prudent investment at both the national and
local level, given the stakes as demonstrated by the massive losses in this recent event.
Steady commitment and reliable and sustainable funding, shorter design and construction
timeframes, clear lines of authority and responsibility, and improved overall coordination of
disparate system elements and functions are all needed as well.
The overall philosophy and basis for design of these types of expensive and vital
systems warrants reconsideration. Improvements such as (1) conceptual design strategies that
involve working in conjunction with natural barriers and other favorable features, (2) system
based risk assessment, analysis and design, (3) allocation of appropriate resources, (4)
embracing research and appropriate technological advances, and (5) maintenance of a
deliberate culture of diligence in seeking overall system reliability would all represent
significant steps forward.
And there is some urgency to all of this. The greater New Orleans regional flood
protection system was significantly upgraded in response to flooding produced by Hurricane
Betsy in 1965. The improved flood protection system was intended to be completed in 2017,
fully 52 years after Betsy’s calamitous passage. The system was incomplete when Katrina
arrived. As a nation, we must manage to dedicate the resources necessary to complete
projects with such clear and obvious ramifications for public safety in a more timely manner.
New Orleans has now been flooded by hurricanes six times over the past century; in
1915, 1940, 1947, 1965, 1969 and 2005. It should not be allowed to happen again.
15 14