The basis of quantitative analysis in fluvial geomorphology of Qinghai-Tibet Plateau—the processes and performance of SRTM3 void-fill.

Fluvial erosion is among the core components in tectonic geomorphology. The 90m-resolution SRTM3 DEM has been widely used in tectonic geomorphology, but the void data is the bottle-neck that hampers its application in quantitative fluvial geomorphology. In this paper, we compare the satellite-derived SRTM3 with the 1:250,000 scale topographical map-derived DEM of 90m pixel resolution (so called 1:250k DEM). 1:250k DEM results in artifacts and distortions is larger than the SRTM3 in generating river networks, particularly in relatively flat regions of flood plains or sedimentary basins. However, it is better than the SRTM3 in the steep and mountainous terrain, regions where SRTM3 has most of its data voids. Comparison of various studies on SRTM3 hole-filling suggests that high-resolution DEM patching is the key, and the choice of algorithm is in secondary importance in improving SRTM3 data quality. Using the complementary 1:250k DEM to fill the holes in SRTM3 is effective in quantitative studies in fluvial geomorphology, especially in dissected margins of the Tibe plateau. We also describe the procedure with SRTM3 data void-filling, for educational and practical purpose.