Geospatial mapping and multi-criteria analysis of groundwater potential in Libo Kemkem watershed, upper blue Nile River basin, Ethiopia

Abstract

This study aimed to delineate groundwater potential zones in the Libo Kemkem watershed, Northwestern Ethiopia, utilizing an integrated approach combining Remote Sensing (RS), Geographic Information Systems (GIS), and the Analytical Hierarchy Process (AHP). Thematic layers such as slope, rainfall, drainage density, lineament density, soil, land use/land cover, distance from rivers, lithology, and the Normalized Difference Vegetation Index (NDVI) were used to assess groundwater potential. The weighted overlay analysis revealed that approximately 40% of the study area exhibited high groundwater potential, while 27% showed low to very low potential. Areas with flat terrain, high rainfall, and dense lineaments were identified as the most favorable for groundwater recharge, whereas regions with steep slopes and poor soil permeability had limited potential. The results were validated using field data from 11 wells, yielding an overall accuracy of 81.8%, supported by Receiver Operating Characteristic (ROC) curve analysis, which produced an AUC value of 60.4%, indicating satisfactory model performance. The study demonstrates the effectiveness of RS, GIS, and AHP as a cost-effective and efficient method for groundwater potential mapping. These findings provide critical insights for sustainable water resource management, guiding the development of groundwater extraction strategies in high-potential areas and conservation efforts in low-potential zones. Future research should focus on integrating machine learning techniques, expanding field validation with more well data, and investigating the long-term impacts of climate change on groundwater recharge.