Performance assessment of global open-access digital elevation models in China mainland coastal region

Abstract

Digital elevation models (DEMs) are the fundamental datasets for coastal ecosystem monitoring, and several global open-accessed DEMs have recently been reported. In coastal regions, a comprehensive vertical accuracy assessment of these DEMs has not yet been carried out. In this study, eight open-access DEM datasets, including SRTM-3, SRTM-1, TanDEM-X, ASTER GDEM v3, MERIT DEM, AW3D30, NASADEM, and CoastalDEM, were investigated across the coastal region of the Chinese mainland using high accuracy ICESat-2 data as a reference. Statistical tools including mean absolute error (MAE) and root mean square error (RMSE) were selected to describe the data error/uncertainty and spatial distribution. Moreover, the effects of elevation ranging, slope degree, geomorphogenesis and landuse on vertical accuracy were further analyzed to assess their applicability. The assessment results revealed that the CoastalDEM and NASADEM datasets had the highest accuracy, with MAE values of 1.68 and 1.88 m and RMSE values of 2.55 and 2.61 m, for 3-arc second and 1-arc second resolution DEMs, respectively. Other DEMs with close accuracies include AW3D30, SRTM-1, MERIT, and SRTM-3 DEM. The results proved that the CoastalDEM outperformed other datasets, indicating its applicability in coastal regions.