Digital elevation model for flood hazards analysis in complex terrain: Case study from Jeddah, Saudi Arabia

Global Digital Elevation Models (GDEMs) have been increasingly used to assess the risk of flooding worldwide. However, their effectiveness and the performance of flood risk models in areas with complex terrain, such as Jeddah, Saudi Arabia, have not been comprehensively studied. This study aims to compare the performance of five distinct Global Digital Elevation Models (GDEMs) - 30-m SRTM, 30-m ASTER, 90-m MERIT, 10 m Sentinel-1 DEM, and 12.5 m ALOSPALSAR - in estimating the inundation extent and depth of the Jeddah watershed, including three dams. Both hydrological and hydraulic modeling approaches were utilized to achieve this objective. The study findings revealed that all Global Digital Elevation Models (GDEMs) produced similar watershed boundaries in the mountain area (Wadi Qaws), except for ALOS-PALSAR, which generated a different watershed boundary from previous reports and studies. However, all GDEMs failed to accurately delineate the drainage area of one of the dams, except for SRTM and ALOS-PALSAR. Moreover, ASTER and SRTM products provided the closest estimates to ground observations, producing peak discharges of 114.1 m3/s and 110 m3/s, respectively. For the entire watershed, encompassing mountain, urban, and coastal areas, GDEMs demonstrated significant differences in the watershed boundary, streams, outlet location, and peak discharge at the watershed outlet. In addition, each GDEM's flood inundation map was significantly distinct. Overall, the results suggest that SRTM outperformed all other GDEMs in the mountain area.