Land use/land cover changes and bare soil surface temperature monitoring in southeast Brazil

Abstract

The land surface temperature (LST) provides important information about energy exchange processes, which are influenced by land use/land cover (LULC). Thus, our objective was to evaluate LST patterns driven by LULC changes, detected over a time series of Landsat images. The study area of 2990 km2 is located in the Piracicaba region, state of São Paulo, Brazil. We acquired Landsat images from 1985 to 2019, in dry and moist seasons. Six LULC classes (agriculture, bare soil, straw, forest, water, and pasture) were identified by maximum-likelihood supervised classification every five years and then LST was estimated using the inversion of Planck’s function in the thermal band. Spectral indices representing vegetation, water, bare soil, and straw were calculated and correlated to LST in specific years. Bare soil images and their respective LST in both seasons were used annually to approach the influence of bare soil areas on the LST, considering soil class, time and rainfall. LULC alterations over 1985–2015 were an important factor on the LST change, which varied on average from 21.46 °C to 41.31 °C in the moist season and 17.05 °C to 31.67 °C in the dry one. Water bodies and vegetation had the lowest LST values, whereas bare soil and straw had the highest ones. The correlation between LST and spectral indices somewhat agreed with such patterns. Arenosols presented the highest LST mean values in both seasons and differed from Acrisols in the dry season, which is probably related to their texture and mineralogical composition. In the moist season, LST was negatively correlated to rainfall, suggesting the influence of soil moisture content on its surface temperature. In the dry season, the LST of bare soil areas increased by an average of 0.13 °C per year, indicating a warming trend. In general, LST increased in the studied period, probably due to the increase of anthropic activity, such as the expansion of agricultural areas. These findings can assist future studies on the influence of soils and land use on climate alterations.