Evaluation of Global Digital Elevation Model (GDEM) over Peninsular

Abstract

Open-source Digital Elevation Models (DEMs), such as Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER), Advanced World 3D (AW3D30), Forest and Buildings removed DEM (FABDEM), Global Land One-km Base Elevation (GLO-30), Multi-Error-Removed Improved-Terrain (MERIT), NASA Digital Elevation Model (NASADEM), Shuttle Radar Topography Mission (SRTM), and TanDEM-x, are widely used in environmental modelling and geospatial studies, where spatial resolution and vertical accuracy are critical. This study evaluates the accuracy of these DEMs by comparing their elevations with orthometric heights derived from Global Navigation Satellite System (GNSS) observations. Accuracy was assessed using Root Mean Square Error (RMSE) and the correlation coefficient (R²). The results indicate that FABDEM achieves the highest performance, with an RMSE of 1.681 m and the strongest correlation with GNSS-derived heights 0.998, largely due to its effective removal of vegetation and building artefacts. These findings suggest that FABDEM is the most reliable DEM among those assessed, making it particularly suitable for coastal and flood-prone areas where precise elevation data are essential for risk assessment, hydrological modelling, and geoid determination.