*2.2. Data Processing*

The processes for obtaining geospatial images and data were performed using the Google Earth Engine (GEE) tool [46]. Cloud computing platforms, such as GEE, facilitate the processing of satellite images globally. With access and analysis of several petabytes of images and a broad set of geospatial data, dating from 40 years of Earth observation data, updated and expanded daily, the tool allows viewing, manipulating, editing, and

creating spatial data [46]. The EVI vegetation indices derived from the Moderate Resolution Imaging Spectroradiometer (MODIS) sensor reflectance data were used to represent the variability of native vegetation, investigating some biophysical mechanisms that involve soil–plant–atmosphere relationships in the natural dynamics of the Caatinga. The MODIS sensor images and the TerraClimate and CHIRPS data sets' environmental drivers were evaluated from 2000 to 2019. The environmental drivers were made up of monthly precipitation, air temperature, soil moisture, and water deficit time series for the three experimental areas.
