Speaker
Description
Groundwater stands as a vital buffer against the growing impacts of climate change, especially in arid and semi-arid regions where surface water is ephemeral and rainfall patterns are becoming increasingly erratic. Understanding how recharge zones respond to climatic variability is crucial for ensuring long-term water security. This study conducts a comparative analysis between a knowledge-driven approach—the Analytic Hierarchy Process (AHP)—and data-driven machine learning models to delineate groundwater recharge potential in the Upper and Middle Drâa Basin, southeastern Morocco, a region highly sensitive to hydro-climatic shifts. The basin, extending over approximately 23,000 km², was analyzed using sixteen environmental and climatic factors derived from satellite datasets (Sentinel-2, ASTER-DEM) and geospatial modeling. The AHP method integrated expert judgment through pairwise weighting and identified high-recharge zones predominantly along low-slope alluvial corridors and near major drainage networks. In contrast, two machine learning models—Artificial Neural Network (ANN) and Support Vector Machine (SVM)—were trained using 80 % of the data and validated on the remaining 20 %. The ANN model achieved strong predictive performance (R² = 0.75, RMSE = 0.76, MAE = 0.51), indicating reliable spatial learning from complex environmental interactions. The SVM model, optimized with an RBF kernel (C = 300, γ = auto), produced an accuracy of 0.82 but a lower R² = 0.50, reflecting the inherent nonlinearity and climatic variability influencing recharge dynamics in desert basins. Despite these differences, both approaches revealed consistent spatial patterns, with the highest recharge potential concentrated in central basin corridors where borehole discharges exceed 6 L/s. The findings demonstrate that while AHP captures the conceptual structure of recharge processes, machine learning models better adapt to the intricate, climate-driven variability of groundwater systems. This comparative framework highlights the importance of integrating expert knowledge with data-based intelligence to enhance groundwater resilience assessments under changing climatic conditions. It offers a transferable methodology for decision-makers facing similar hydro-climatic challenges across arid and semi-arid regions worldwide.
