Abstract
The principal purpose of this study is to investigate the uncertainty of Soil and Water Assessment Tool (SWAT) model in simulating outputs of runoff and sediment yield due to changes in Digital Elevation Model (DEM) resolution, sources, and resampling methods. The Upper Blue Nile catchment was selected as the study area. DEM scenarios were generated based on DEMs differing in resolution (30–1000 m), sources (ASTER GDEM2, CartoDEM, and SRTM 1-Arc Global DEM), and resampling technique (nearest neighbor, bilinear interpolation, cubic convolution, and majority). In this phase of study, the SWAT model uncertainty with DEM scenario based on resampling technique was assessed. The key findings of this study are: (1) for DEM obtained from bilinear interpolation, the simulated runoff and sediment yield was higher in rainy months as compared with other techniques, (2) the SWAT model was found to be more sensitive to resampling techniques in simulating sediment yield than in simulating runoff, (3) based on a few selected statistical measures (R2, NSE, and RMSE), it was found that use of bilinear interpolation technique is the most efficient while simulating runoff and sediment yield, (4) majority and nearest neighbor techniques were found to be the least efficient while simulating sediment yield and runoff, respectively.