Spatiotemporal characterization of land cover and degradation in the agreste region of Pernambuco, Brazil, using cloud geoprocessing on Google Earth Engine

Abstract

The characterization of land use and land cover (LULC), as well as the identification of its degradation factors, are necessary for the preservation of the agroecosystem and help in decision making. Therefore, the objective of this research was to evaluate the spatial-temporal dynamics of vegetation cover and soil degradation in the Agreste mesoregion of the state of Pernambuco, through the leaf area index (LAI), soil-adjusted vegetation index (SAVI), and pasture vegetation cover index (PVC) from Sentinel-2/MSI images from the Multispectral Instrument (MSI) sensor and, develop software for automatic cloud processing on the Google Earth Engine (GEE) platform. Were used 8 Sentinel-2/MSI satellite images (2017–2020), processed in GEE, to obtain the SAVI, LAI, and PVC. A pluviometric survey was carried out for the month before the registration of each image, from the Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS). The images were subjected to descriptive statistics and principal component analysis (PCA). The months that comprise the highest rainfall exceeded the accumulated 100 mm. The SAVI values ranged from −0.2 to 0.80, of the LAI between 0 and 3 m2 m−2 (with small regions greater than 3). Regions of the low spectral response of LAI and SAVI, correlated with areas of low rainfall in the study region. The total variance showed 100% representativeness and significance of the data for the accumulated in PC2, this denotes how significant the correlations between SAVI, LAI, and PVC are. Because of the new climatic and environmental scenarios, this study presented a relevant contribution to characterizing the potential areas in soil degradation processes, mainly in the Caatinga biome. Between the years 2017–2020, there was the occurrence of soil degradation processes in the Agreste region of Pernambuco, where the state's dairy basin is located, caused by poor management, use of arable land, and long periods of drought.