Assessing Spatiotemporal Land Dynamic Using Google Earth Engine and Random Forest: Trends and Drivers of Change in Ethiopia’s Fragile Jema River Basin

Quantifying land use land cover (LULC) dynamics in vulnerable ecosystems of the Ethiopian highlands is crucial for understanding the drivers of environmental degradation and informing sustainable land management to protect ecosystem integrity. The Jema River Sub-basin, a critical contributor to the Upper Blue Nile, is highly vulnerable to agricultural expansion, the loss of native vegetation, and landscape fragmentation. This study quantified the spatiotemporal dynamics of LULC change and identified their potential drivers between 1994 and 2024, to support evidence-based and climate-resilient land-use planning in this basin. A machine-learning framework was implemented on the Google Earth Engine (GEE) platform to classify LULC. We utilized Landsat 5 TM, Landsat 7 ETM+, Landsat 8 OLI, and Sentinel-2 MSI imageries, which were classified using the Random Forest (RF) algorithm. The classification was optimized by integrating spectral indices, including the Normalized Difference Vegetation Index (NDVI), Normalized Difference Built-up Index (NDBI), and Modified Normalized Difference Water Index (MNDWI) as well as topographic variables (slope and elevation). Classification accuracy, assessed using 3,100 ground-truth samples, exceeded 87% overall accuracy with Kappa coefficients above 83%. Temporal change detection, transition matrices, and trend analyses quantified alterations across seven LULC classes. The results show that Farmland (FL) expanded by 37.41%, and built up area (BA) by 230.45%, while Natural Vegetation (NV) and Grazing Land (GL) declined by 55.98% and 67.42%, respectively. Focus group discussions and key informant interviews identified population pressure, agricultural intensification, and weak governance as the dominant socio-economic drivers. The integration of multi-temporal geospatial analysis with local knowledge indicates substantial anthropogenic modification of the landscape and provides empirical evidence to inform sustainable land-use planning and ecosystem restoration initiatives in the Jema Sub-basin