Improving Soil Erosion Estimation: Modifying Soil Erodibility and Vegetation Factors in the RUSLE Model

Abstract

Soil erosion is a serious threat to food security and ecosystems, especially in dry regions where it is difficult to estimate accurately. This study aims to spatially estimate soil erosion rates across Iran using the modified Revised Universal Soil Loss Equation (RUSLE) (CARUSLE) model integrated with remote sensing and geographic information systems (GIS). Main factors—including rainfall erosivity (R), soil erodibility (K), topography (LS), vegetation cover (C), and soil conservation management (P)—were derived from digital elevation model (DEM), HYDROSHEDS, and satellite data over the 2024 year. In CARUSLE, the K factor was refined by incorporating soil carbon and pH, whereas the C factor was adjusted using normalized difference vegetation index (NDVI) and land-use data. Model validation was performed using 300 expert-interpreted reference points derived from Google Earth, and performance was assessed through statistical metrics including correlation coefficient, coefficient of determination, mean square error, root mean square error, mean absolute error, and mean absolute percentage error. Results indicate a mean annual soil erosion of 1700–2000 t km⁻² year⁻¹ across Iran, with the highest rates in the Alborz and Zagros mountains. The finding of soil erosion intensity within the CARUSLE illustration is that the erosion classes are in the very low, low, moderate, high, and very high classes with percentages of 45.67, 11.35, 11.59, 10.90, and 20.49. Steep slopes, high rainfall, and sparse vegetation are the main factors of erosion, and critical areas are concentrated in the western and northern regions. Model validation generated an overall accuracy of 78% and a kappa coefficient of 0.71, confirming the reliability of CARUSLE in reproducing erosion severity patterns.