Comparison of Three Delineation Methods Using the Curve Number Method to Model Runoff

Abstract

Digital Elevation Models (DEMs) are spatial grids which are used to automate watershed boundary determination. Sinks are present within most DEMs. In order to easily process the watershed boundary, the sinks are reassigned to elevation equivalent to an adjacent cell. The derived DEM is called a “filled” DEM. Due to its relative simplicity, the use of the “filled” DEM is one of the most widely used methods to delineate watershed boundaries and works well in about 70 percent of the watersheds in the US. In landscapes with internal drainage, sinks may accurately represent these depressions. In this study, we compare two delineation methods that do not fill in sinks to another method that does fill in sinks. We examined ten gaged watersheds in Wisconsin and Minnesota. We found the spatial extent of the watersheds from the three methods were significantly different. To evaluate the delineation methods, we modeled ten runoff events using the Curve Number (CN) method and compared them to USGS gage discharge for each watershed. For small storms we found that there were no significant differences in the modeled runoff for three delineation methods. For large storms, we found the no-fill methods had a smaller error, but overall the difference was insignificant. This research suggests that capturing internal drainage by the delineation does not have much of an impact on the widely used CN model.