DEM Resolution Effects on Hillslope Length and Steepness Estimates for Erosion Modeling

Abstract

The use of Geographic Information Systems (GIS) has greatly enhanced the capability of predicting soil loss using hillslope and watershed models. Selecting the appropriate Digital Elevation Model (DEM) resolution is important for estimating slope length and steepness values and subsequently estimating soil erosion and sediment delivery rates. Slope length and steepness estimates from 10-m and 30-m DEMs in ArcGIS were compared to those of observed hillslope rill patterns. The observed data were collected in the springs of 1973 through 1983 from several fall-seeded winter wheat fields in eastern Washington and northern Idaho, an area of steep and variable topography. Estimated erosion rates from the Revised Universal Soil Loss Equation, Version 2 (RUSLE2) model for the observed and DEM topographic data from ArcGIS were also compared. Statistical analyses resulted in the quantified rank of error for estimated slope steepness and slope length data in general ascending order as: DEM resolution (10-m, 30-m); and slope length estimation (Flow Length, Flow Accumulation). RUSLE2 soil loss estimates from 10-m DEM resolution underestimated 15%, and those from 30-m DEM resolution underestimated 22% as compared to RUSLE2 estimates from the observed topographic data. For the conditions of the study area, the impact of slope steepness on soil loss was more significant than that of slope length. The erosion estimates are site-specific, but rank of error should be regionally applicable to hillslope and watershed models using ArcGIS-based extensions to estimate slope steepness or flow path length. Given the lack of accuracy in erosion estimates with the better resolution 10-m DEM, consultants and agencies such as Natural Resources Conservation Service may need to develop a method to adjust GIS steepness estimates on steep, irregular topography in order to make realistic field-scale erosion estimates for planning purposes.