Sensitivity of RUSLE to Data Resolution: Modeling Sediment Delivery in Upper Little Tennessee River Basin
Gardiner, Edward P.
Meyer, Judy L.
MetadataShow full item record
Data resolution influences the magnitude and variability of erosion and sedimentation estimates from geographic information system (GIS) representations of the Revised Universal Soil Loss Equation (RUSLE). We implemented a GIS version of RUSLE for the Upper Little Tennessee River Basin in Macon County, North Carolina. When raster layers depicting elevation, land cover, and soil erodibility were coarsened in resolution, sediment load estimates differed by two orders of magnitude. Model output from low resolution data (285m x 285m pixels) captured less than 10% of the variability of model output from 30m data. We further quantified the importance of soil and land cover data resolution on model output. State Soil Geographic (STATSGO) soil erodibility factors were, on average, 34% (range -23% to 662%; s.d. 43%) greater than the same factors taken from the Soil Survey Geographic (SSURGO) database. Model results based on STATSGO were not tightly correlated with estimates based on SSURGO data. Linear regressions were weak but significant (R 2 approximately 0.50 in all cases) for comparisons of model output based on lowered cover factor resolutions. Before applying similar models in a management context, analysts should calibrate model output with in situ sediment loading or soil loss measurements because input data resolution and accuracy have significant influences over model output.