Mapeamento de áreas susceptíveis à degradação no semiárido brasileiro no contexto de BIG EO DATA

Abstract

Environmental degradation in semi-arid regions represents one of the greatest challenges to ecological and socioeconomic sustainability on a global scale. In Brazil, the political semiarid region encompasses a significant portion of the national territory and concentrates a considerable population that is vulnerable to climate variability, water scarcity, and the anthropic impact in the local ecosystems. This dissertation aims to contribute to the analysis of susceptibility to land degradation in the Brazilian semiarid by integrating climatic and spectral data within a Big Earth Observation Data framework. The specific objectives are: (i) to automate the integration of spectral and climatic variables in Google Earth Engine through the creation of replicable codes based on time series; (ii) to develop a spatial model capable of mapping susceptibility to degradation using NDVI, LST, and accumulated precipitation; and (iii) to analyze the regional degradation scenario by integrating orbital data with meteorological station records. The adopted methodology is structured around three main axes: temporal analysis of MODIS (LST and NDVI) and CHIRPS variables; weighted overlay modeling in a GIS environment; and integration with historical temperature and precipitation series from INMET (1993–2024). The results revealed heterogeneous spatial patterns of susceptibility, with emphasis on three critical nuclei: northern Bahia, Angicos-Seridó, and the backlands of Ceará— areas consistent with previously identified desertification hotspots in the literature. The analysis of climate trends showed that over half of the meteorological stations analyzed exhibit both increasing mean temperatures and decreasing precipitation in recent decades, indicating a trend toward aridification. The use of remote sensing tools under a time-first paradigm proved effective in identifying temporal patterns of degradation, while the developed spatial model contributes to the diagnosis and continuous monitoring of the study area. The approach adopted advances methodological frameworks for the use of Big EO Data in Geography and provides technical support for the development of public policies aimed at mitigating environmental degradation in semiarid regions of Brazil and beyond.