Fig 1.
Location of the study area near the northern coast of the state of São Paulo (SP) (background image [25]).
The Serra do Mar State Park is outlined in white.
Table 1.
Landsat images characteristics summary used for LST calculation.
Fig 2.
Image classification workflow (Source: Adapted from Esri).
Fig 3.
(A) Classification of land cover and the human-modified forest fractional area (FAAI) index, where the black squares in (a) and (b) are magnified in (b) and (c), respectively. The degradation of spatial resolution is shown in (c), where square of 15 m × 15 m defined land cover (red circle), squares of 120 m × 120 m were used to calculate FAAI (large red square), and squares of 30 m × 30 m were used to calculate LST (small red square).
Fig 4.
(a) Altitude (m), (b) slope aspect (°).
Fig 5.
(a) Average FAAI at a resolution of 120 m (%). (b) Average LST at a resolution of 120 m (°C). (c) Histogram of FAAI occurrence by intervals of 0% to 5%, and then from 5% to 95% in 10% increments, and finally from 95% to 100%.
Fig 6.
(a) Relationship of raw LST (°C) with FAAI (%) (blue circles are average LST at each FAAI class; grey shaded area is bounded by standard deviation of LST per FAAI class; red line is linear regression of all LST data against FAAI, LST = 0.03 FAAI + 19.6; R2 = 0.6, significant p-value < 2.2e-16); and (b) histogram of the LST distribution at each FAAI class 0%, 50% and 100% and associated skewness coefficient (Ssk).
Fig 7.
(a) Forest vegetation patch to the south of a pasture area (perimeters outlined in light green) (source: Microsoft Bing Maps Services) near Catuçaba district, São Luiz do Paraitinga, SP, with associated information of (b) LST (°C) (August 1, 2010), (c) aspect (°), and (d) altitude (m).
Fig 8.
Average raw LST (°C) depending on the terrain aspect (in classes of 5°) and for FAAI classes (5–95%).
The gray scale is the median of each FAAI class (%).
Table 2.
Absolute maximum amplitude of LST (maximum LST minus minimum LST,°C) median (± SD) by FAAI class (%) for variation of aspect and altitude respectively.
Fig 9.
Average raw LST (°C) by altitude (m) among FAAI classes.
The gray shading represents the median of the FAAI class (%).
Fig 10.
(a) Scatterplot of raw LST and corrected LST (y = 4.7 + 0.83·x); R2 = 0.82, significant p-value < 2.2e-16). (b) Quantile–quantile distribution of raw LST and corrected LST. (c) relationship of corrected LST with FAAI (%) (blue circles are average LST at each FAAI class; grey shaded area is bounded by standard deviation of LST per FAAI class; red line is linear regression of all LST data against FAAI, LST = 0.038 FAAI + 20.9; R2 = 0.63, significant p-value < 2.2e-16).