Evaluate and Enhance the Efficiency of Freely Available Global DEMs for Flood Modeling in High-Mountain Environments: A Case Study of Glacial Lake Outburst Floods (GLOFs)

Digital Elevation Models (DEMs) are among the most critical factors influencing the performance of flood modeling. In many regions worldwide, freely available satellite-derived global DEMs are often the only accessible source of topographic data. Extensive research has focused on improving freely available DEMs to support catchment-scale flood modelling, particularly in low-lying areas. However, relatively little attention has been given to high-mountain and rugged terrains, such as the Himalayas. In these environments, the low resolution of open-access DEMs often fails to capture key hydrological features, such as narrow valleys and streams, leading to suboptimal performance of hydrodynamic models. This study uses Glacial Lake Outburst Floods (GLOFs) — widely recognised as one of the most devastating natural hazards in the Himalayas — as a case study. We evaluate the performance of five contemporary 1 arc-second (~30 m) DEMs: FABDEM, Copernicus DEM, NASADEM, AW3D30, and SRTM. The evaluation is conducted by analysing differences in simulated inundation areas, water depths, flow velocities, and flow arrival times for GLOFs using a GPU-based high-performance hydrodynamic model. To address the limitations of freely available DEMs, this study proposes a novel method for hydrological correction in DEMs to improve the accuracy of GLOF modelling. GLOF events are simulated using the original and hydrologically corrected DEMs, followed by a comparative analysis to assess the simulation accuracy and performance of the different DEMs. The results demonstrate that the corrected DEMs yield significant improvements in modelling accuracy, highlighting the potential of this approach for more reliable flood hazard and risk assessments in high-mountain environments.