The sustainability of vegetation is closely linked to anthropogenic practices, climate, and investments in environmental management. The objective is to help prioritize the impacts of anthropogenic practices and rainfall on ecosystem evolution. Following the rainfall gradient, a diachronic analysis of Landsat images from November or December for the years 1986, 1999, 2013, and 2018 was carried out. These cloud-free satellite images from Landsat 7 to 8 bands were downloaded from the USGS website: http://www.earthexplorer.usgs.gov/. Processed according to supervised classification, a Pearson correlation matrix relating to land cover types (rock outcrop, irrigated crop, rain-fed crop, fallow land, water body, forest reserve, shrub savanna, shrub steppe) was established. The results show that anthropogenic practices influencing vegetation change are highly diverse.