Assessing and projecting land use land cover changes using machine learning models in the Guder watershed, Ethiopia

• GEE and ANN model effectively classify and project LULC analysis of Guder watershed. • Historical LULC analysis experienced a marked increase in cultivated land. • Projections indicate that there is substantial urban growth in the Guder watershed. • Forest may increasingly convert to farms and urban area due to human expansion. • The study discusses the environmental implications of projected LULC modifications. This study investigates the trends and frequencies of Land Use Land Cover (LULC) changes in the Guder watershed, located in the Upper Blue Nile Basin (Ethiopia), for the periods 1985 and 2021, with projections for 2039 and 2057. The research utilizes an integrated approach combining remote sensing (RS) and GIS for spatial analysis, Google Earth Engine (GEE) for cloud-based data processing, Random Forest (RF) machine learning for historical LULC classification, and an artificial neural network (ANN) model via QGIS's MOLUSCE tool for future LULC predictions. This innovative methodological approch allows for the examination of spatial and temporal LULC change patterns and future projections in the watershed. The results indicate that cultivated land increased from 54.8 % in 1985 to 72.9 % in 2021, and the built-up area experienced a significant increase of 227.5 % during this period. Percentage of land covered by forests fell from 35.9 % in 1985 to 9 % in 2021. By 2039 and 2057, shrubland, forest, and grassland are expected to decrease, while built-up and cultivated land will increase. Specifically, shrubland will decrease from 12.4 % in 2021 to 10.1 % in 2039 and 8.7 % in 2057, grassland from 4.8 % in 2021 to 1.9 % in 2039 and 1.1 % in 2057, and forest from 9.0 % in 2021 to 8.9 % in 2039 and 7.9 % in 2057. Meanwhile, the built-up area will rise significantly from 0.8 % in 2021 to 3.6 % in 2057. These shifts profoundly impact environmental management in the watershed. The motivation behind the present research was to provide a thorough understanding on LULC dynamics to improve land management practices using machine learning and ANN models for current and future environmental changes. Strategic interventions are crucial to mitigate adverse trends and promote sustainable land management based on current scenarios and future projections.