Uncertainty Assessment in Soil Erosion Modelling Using RUSLE, Multisource and Multiresolution DEMs

Abstract

Soil erosion is a key concern for the environment and natural resources since it leads to a decline in-field productivity and soil quality, resulting in land degradation. In this study, assessment of uncertainty in soil erosion modelling of the Karso watershed, India, was carried out by employing the revised universal soil loss equation (RUSLE) and geospatial technologies to evaluate the effect of multi-source digital elevation models (DEMs) [Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER), Cartosat and Shuttle Radar Topography Mission (SRTM)] with resampled multi-resolution grids. The rainfall erosivity factor (R) was computed using the mean monthly Tropical Rainfall Measuring Mission rainfall estimates for 1998 to 2012. The slope length factor was derived using the ASTER and Cartosat DEMs at grid sizes of 30 m, 50 m, 100 m, 150 m, 200 m, and 250 m, and for the SRTM DEM at 100 m, 150 m, 200 m and 250 m resolutions for the Karso watershed, Jharkhand, India. Significant differences were obtained in the soil loss estimates across the different DEM sources and resampled grid sizes. The Cartosat DEM with a 200 m grid was found to estimate the soil loss the best out of all the DEM combinations considered. The Cartosat DEM proved to be more reliable than the ASTER and SRTM DEMs. The results indicated that the RUSLE is a scale-dependent model since the model estimates were affected not only by the DEM source but also by its resolution. The prediction of erosion potential by employing the multisource, multiresolution DEMs and the RUSLE helped to identify the soil erosion's spatial pattern within the watershed. The study provided an impact analysis of the uncertainties when selecting the multisource, multiresolution DEMs for soil erosion modelling.