Suitability of different Digital Elevation Models in the estimation of LS factor and soil loss

Abstract

Soil erosion is a global concern, and tons of fertile topsoil are lost worldwide. Topography significantly influences soil erosion patterns, shaping how soil loss varies across landscapes. In the Revised Universal Soil Loss Equation (RUSLE), the topographic factor (LS-factor) quantifies this impact, with Digital Elevation Models (DEMs) serving as key inputs for its derivation. The soil loss over Kerala, India, is estimated using different DEMs. The study also explored two methods for deriving the LS-factor, one based on flow accumulation and another based solely on the slope length. Among the approaches tested for LS factor estimation, the slope-based method proved more effective than one incorporating flow accumulation, as the study is for a region rather than a distinct hydrologic unit. Four freely available Digital Elevation Models, ALOS, ASTER, SRTM, and Cartosat-1 were selected for the study. The study showed that the general pattern of soil erosion can be captured by using any of these DEMs despite differences in individual elevation values. The mean potential soil loss estimated for the year 2020 was 215.91 t/ha/year, 205.70 t/ha/year, 203.99 t/ha/year, and 207.97 t/ha/year when using ASTER, ALOS, SRTM, and Cartosat-1, respectively. The ASTER DEM shows a slightly higher mean value but exhibited the least uncertainty, which was confirmed by bootstrap resampling uncertainty analysis. These findings emphasize the need for careful DEM selection based on terrain characteristics, enhancing the accuracy of soil erosion assessments and informing more effective land management strategies.