Relação espaço-temporal entre o uso e a cobertura da terra e a temperatura de superfície

Abstract

Land Surface Temperature (LST) is one of the main physical parameters used to assess the impacts of changes in Land Use and Land Cover (LULC). These changes may vary over time and allow a diagnosis of the anthropic impacts responsible for modifying the natural landscape. This study aims to evaluate if types of LULC affect the LST variation over the last 30 years in a region on the west of São Paulo state, Brazil. This region is composed of 32 municipalities and marked by intense agricultural cultivation. We evaluated three dates during the winter station (July), using images from the Landsat 5, TM (Thematic Mapper) sensor, for the years 1987 and 2000, and Landsat 8, OLI (Operational Land Imager) and TIRS (Thermal Infrared Sensor) sensors for the 2017 year. Maximum Likelihood classification was applied to data from TM sensor bands (5, 4, 3) and OLI sensor bands (6, 5, 4) to make LULC maps for each dates of analysis. We extracted the LST from the Landsat thermal images using the band 6 for the years 1987 and 2000 of the TM sensor and the band 10 for the 2017 year of the TIRS sensor. We verified that the LST variation for each class of LULC (vegetation area, crop fields, water body, pasture and bare soil) at a significance level of 5%. The pasture and bare soil class had a higher LST over the 30-year period (1987 to 2017) analysis, while the water bodies and the vegetation areas contributed to the reduction of the LST in the same period. We found an average increase of 5 Celsius degree of LST for the studied area during the last 30 years. We conclude that the type of LULC implies a variation of LST