ASTER and SRTM are two free available sources for digital elevation data covering the most of the world. In this paper vertical accuracy of such models over different terrain types compared with the accuracy of the local DEM produced from topographic maps of scale 1: 50,000 has been evaluated1 The northern Nile valley was chosen as a study area that extends from (30° to 31.5°) of longitudes and from (28° to 31.5°) of latitudes, where a number of 705 GCPs were available in that area-and used to accomplish the evaluation process. The results show that SRTM and the local DEMs are close to each other in terms of the RMSE, they have almost the same accuracy for the different terrain with minute differences, while ASTER DEM lies in the final ranking. The results showed also that there is 12.55 m downward (average shift between the GCPs and ASTER DEM). The vertical accuracy of this DEM was radically improved by 57.8% over steep terrain and 48.8% over flat terrain after eliminating such shift, so these removed values could be considered as systematic errors and such model is therefore considered a relative DEM. From the final results, it can be concluded that SRTM DEM can be used to update topographic maps of scale 1:50,000, since its accuracy was found to be less than half the contour interval of such maps. Simultaneously, the ASTER DEM can be utilized for the same purpose, but for maps of lesser scales after eliminating its vertical shift (vertical systematic errors). |
