Impact of the D infinity of the automatic extraction of drainage networks from Digital Elevation Models

Abstract

In the last years, a substantial effort has been dedicated to enhance the drainage network extraction from Digital Elevation Models. The direction of this enhancement has been varied. It has been suggested that more precise algorithms to compute the contributing area could help substantially to solve this problem. The current work proposes to asses the real benefit that this approach would have on the finally extracted network. For it, the algorithm with greatest impact in the last years has been choosed, which seems to offer more expectations in this sense (D¥), and has been compared with the most used one up to now (D8). To this purpose a study zone has been defined, composed of two contiguous basins, and the statistical techniques that was considered described better so much the differences as the similarities between both methods have been selected. Our results seems to confirm the hypothesis that the real impact that multi-flow methods, as D¥, have on realist networks extraction is fairly poor, which maybe it is showing the way for future lines of research.