Assessment of vertical accuracy of open source 30m resolution space-borne digital elevation models

Abstract

Highly accurate air-borne Digital Elevation Models (DEMs) are not commonly available in most regions of the world, particularly developing countries like Nepal. Better topographic data are required to assess catchment hydrology, flood/inundation propagation, and sediment transport/deposition. Ground-based surveys (i.e., topographic and bathymetry) and Light Detection and Ranging (LIDAR) are expensive in terms of cost and time. We rely mostly on space-borne DEMs when the area of interest is quite large. And, for river cross-sections at specific regions, ground-based surveys are conducted at some regular intervals. The ALOS World 3 D − 30 m (AW3D30), ASTER Global DEM Version 3 (ASTER GDEM v3), SRTM-30, and Carto DEM v3 are some of the popular and latest space-borne DEMs that have been made available to the general public free of charge at a spatial resolution of 30 m. The availability of multiple space-borne DEMs may likely puzzle the practitioners on the selection of appropriate products. Here in this study, we evaluate the vertical accuracy of different space-borne DEMs with respect to the ground-based survey points (2891 points) at river cross-sections and ground-based indicative spot heights (8102 points) across the West Rapti River basin, Nepal. Performance metrics, like mean error and root mean square error (RMSE), were used to quantify deviation with respect to observed elevation. The AW3D30 showed the lowest RMSE value (i.e., 3.41 m) in the lower region of the study area. We numerically simulated the inundation propagation based on four selected DEMs using a calibrated two-dimensional runoff inundation model, and we find the AW3D30 outperforms other DEMs. All four DEMs showed a similar tendency on a catchment scale with respect to the information of indicative spot heights, suggesting interchangeable applications of all four DEMs in hydrologic analyses. However, precise ground-based observations across the catchment are needed for assuring the applicability.

Keywords

• digital elevation model
• error
• inundation space-borne
• west rapti river basin