Análise geoespacial das perdas de solos no bioma Caatinga usando a plataforma Google Earth Engine e múltiplos produtos de sensoriamento remoto

Abstract

Soil erosion is a natural process that involves the detachment, transport, and deposition of sediments, influenced by both natural and anthropogenic factors. Considering the complexity of this phenomenon and the scarcity of studies on a biome scale, such as in the Caatinga, it is essential to develop research that contributes to understanding soil loss and environmental degradation over time and space. This issue is particularly relevant in the Brazilian semi-arid region, where erosion is one of the main environmental problems due to the natural characteristics of the area. In this context, the present study aimed to estimate soil losses in the Caatinga biome from 2001 to 2020, using the Google Earth Engine (GEE) platform and multiple remote sensing products. The Universal Soil Loss Equation (USLE) was applied, with factors derived from different sources: rainfall erosivity (R) from the CHIRPS dataset, soil erodibility (K) from OpenLandMap soil texture classes, topographic factor (LS) from MERIT Hydro, while cover and conservation practice factors (CP) were obtained from the MODIS vegetation index (MOD13A2). Due to the different spatial resolutions of the datasets, all inputs were resampled to 500 m. Results indicated that soil losses predominantly fell within the weak and very weak classes (0.001–0.0375 ton/ha), with higher values concentrated in areas of lower vegetation cover, possibly related to the reduction of arboreal and arboreal-shrub Caatinga, especially during the early part of the study period (2001–2010). In contrast, the later period (2010–2020) showed an increase in erosion, probably associated with higher erosivity and increased precipitation. Validation with observed data from the São João do Cariri Experimental Basin (PB) demonstrated a strong linear correlation (R² = 0.94), indicating a direct relationship between precipitation and erosive energy, reinforcing the key role of rainfall in water erosion. The application of the USLE processed in GEE from remote sensing data provided an improved understanding of erosion in the Caatinga biome. In addition to supporting land-use planning and environmental management, this approach can guide public policies aimed at soil and water conservation, especially in semi-arid regions.