Evaluation of intermediate TanDEM-X digital elevation data products over Tasmania using other digital elevation models and accurate heights from the Australian National Gravity Database |
| |
| Authors: | M Rexer C Hirt |
| |
| Institution: | 1. Institute for Astronomical and Physical Geodesy &2. Institute for Advanced Study, Technische Universit?t München Arcisstrasse 21, D-80333, München, Germanym.rexer@tum.de;4. Institute for Advanced Study, Technische Universit?t München Arcisstrasse 21, D-80333, München, Germany;5. Department of Spatial Sciences &6. Western Australian Geodesy Group &7. The Institute for Geoscience Research, Curtin University of Technology, GPO Box U1987, 6845, Perth, WA, Australia |
| |
| Abstract: | The successful operation of the TanDEM-X satellite mission is the start of a new era of globally consistent and accurate digital elevation data for planet Earth. In this work available 12 m-resolution intermediate TanDEM-X products (DEM: digital elevation model; HEM: height error map; COV: coverage map; WAM: water indication mask) are evaluated over Tasmania. Elevations from the TanDEM-X intermediate digital elevation model (IDEM) are compared with (a) other global DEMs (30 m-resolution SRTM1 USGS v3 and 30 m-resolution Advanced Spaceborne Thermal Emission Reflectometer (ASTER GDEM2), (b) the local 25 m-resolution DEM made available by Tasmanian environmental authority (DPIPWE), and (c) over 15 000 accurate ground-control-points (GCPs) from the Australian National Gravity Database (ANGD). The comparison with ASTER and SRTM over the area of Tasmania involves over 500 million valid TanDEM-X IDEM elevations. The root-mean-square (RMS) of 8.8 m indicates a reasonable to good agreement of TanDEM-X IDEM and SRTM, while ASTER shows almost twice the disagreement in terms of RMS (～16.5 m). Both, ASTER and SRTM show a (mean) offset of –1.9 m and –2.3 m w.r.t. TanDEM-X IDEM, respectively. By comparisons with GCPs, we find that SRTM and ASTER overestimate the terrain height. The comparison with the AGND GCPs also allows an estimate of the absolute accuracy of the IDEM, which is found to be superior to that of SRTM or ASTER. The RMS error of 6.6 m shows that the IDEM is close to the officially denoted 4 m absolute vertical accuracy considering that the GCPs are not error free. The height error map information layer is found to a suitable first indicator of the (local) accuracy of the IDEM in a relative sense. However, we find that the HEM tends to underestimate observed differences to the GCPs. Terrain-type analyses reveal that the TanDEM-X IDEM is a very consistent elevation database over Tasmania. In conclusion, our study demonstrates that the new TanDEM-X elevation data sets provide improved high-resolution terrain information over Tasmania and beyond. |
| |
| Keywords: | TanDEM-X digital elevation model IDEM WorldDEM validation topography SRTM ASTER Tasmania |
|
|
