Remote sensing techniques via Google Earth Engine for land degradation assessment in the Brazilian semiarid region, Brazil

Abstract

Land degradation is a major challenge worldwide, limiting socio economical development and jeopardizing biodiversity, being associated to changes in land use and to climate variability. FAO shows that the world's soils have environmental degradation in large scale, particularly in semiarid areas, reducing environmental services, and impacting agriculture and livestock. Therefore, the objective was to evaluate the spatio-temporal dynamics of vegetation cover and land degradation in the microregion of Vale do Ipojuca in the Pernambuco State (Caatinga biome), in the Brazilian semiarid region, through vegetation indices. Land use and land cover maps (LULC), Normalized Difference Vegetation Index (NDVI), Leaf Area Index (LAI), and Normalized Difference Water Index (NDWI) were determined based on the annual average between the years 2014–2021. The water bodies showed remarkable reductions between 2014 and 2016, when the values were 1226, 789, and 558 ha, respectively. As observed in the NDVI, from the year 2017 onwards, the LAI values were resilient. The determination of vegetation indices associated with LULC maps and data, allowed us to accurately validate the responses of the indices studied in the semiarid region. The approach highlights the strong resilience of the Caatinga Biome. A robust methodology is presented in this study, being compatible for existing arid regions in the world.