Dynamics of land-use Change using Geospatial Techniques From 1986 to 2019: A Case Study of High Oum Er-Rbia Watershed (Middle Atlas Region)

Abstract

This work aims to expose the contribution of the use of the cloud google earth Engine (GEE) platform, in particular the capacity of optical monitoring by remote sensing to assess the impact of environmental changes on the evolution of natural resources in the Middle Atlas region. To achieve this goal, the dense time stacking of multi-temporal Landsat images and random forest algorithm based on the Google Earth Engine (GEE) platform was used. The spatial resolution of the images used is 30 meters for the TM 5 sensor (Thematic Mapper) and the OLI 8 sensor (Operational Land Imager). Further, the google earth engine platform is used primarily to download and prepare the images for the dates 1986, 2000, and 2019, then a supervised classification with the Random Forest (RF) algorithm to produce land use maps of selected dates with an overall accuracy exceeding 80%. This was followed by the production of maps and change matrices for the periods 1986-2000 and 2000-2019. The results obtained have shown a decline in grassland, forest land, and water body in parallel with an increase in the following classes: buildings, farmland, and arboriculture during the last 30 years. In addition, elevation was the most important characteristic variable for land-use classification in the study area. Obtained results provide theoretical support for adjusting and optimizing land use in the High Oum Er-Rbia watershed.