Sensitivity analysis of hydrodynamic models to storm, DEM, and dam configurations: the Bu Mansour Dam Basin in Derna, Libya

Climate change is affecting flood-vulnerable areas in the Mediterranean, highlighted by the 2023 collapse of Bu-Mansour Dam. This study used a 2D hydrodynamic model to estimate the effect of Daniel Storm’s spatiotemporal distribution on maximum pool water elevation behind Bu-Mansour Dam. The study assessed the model feasibility by comparing results to other literature. The study also evaluated how DEM accuracy, rainfall, and dam geometry influenced 2D model results. ALOS DEMs predicted higher overtopping depths than SRTM. Rainfall interpolation increased water levels, while storm temporal distribution affected peak flow but not maximum elevation. Mesh alignment caused elevation differences up to 7.4 m, emphasizing the role of precise model geometry. Discrepancies with prior overtopping estimates were found, suggesting data preprocessing flaws. The study concluded that 2D models are reliable with high-resolution data, though computationally demanding. Revealing the need for accurate hydraulic data and DEMs to ensure accurate dam safety assessment.