Abstract
Abstract Soil erosion poses a significant environmental threat in the Bilate sub-basin, where steep terrain, heavy rainfall, and unsustainable land use practices intensify the problem. The innovativeness of this study lies in systematically integrating the Soil and Water Assessment Tool (SWAT) model with Multi-Criteria Decision Analysis (MCDA) to overcome the limitations of standalone methods, thereby providing a more robust and spatially explicit erosion risk assessment. The MCDA framework, using the Analytic Hierarchy Process (AHP), weighted eleven erosion factors (Land Use/Land Cover (LULC), soil, slope, Normalized Difference Vegetation Index (NDVI), elevation, aspect, Topographic Wetness Index (TWI), Stream Power Index (SPI), drainage density, distance to roads, curvature) to generate a susceptibility map. While NDVI does not have a direct effect on erosion, it serves as a proxy for vegetation cover density, which influences surface protection, root cohesion, and runoff dynamics, thereby indirectly affecting soil erosion risk. Concurrently, the SWAT model was calibrated (NSE = 0.72, R2 = 0.78) and validated (NSE = 0.68, PBIAS = 12.4%) using observed streamflow and sediment data to simulate sediment yield. Key findings indicate that 34.5% of the sub-basin is covered by erosion-prone Leptosols and Luvisols, while 29.9% has slopes exceeding 15%. The SWAT-MCDA integrated analysis identified that 53.46% of the sub-basin faces severe erosion and 9.12% faces extreme erosion, particularly in downstream and upstream areas, threatening agricultural productivity and Lake Abaya. This integration corrected individual model biases: MCDA alone underestimated extreme erosion (0.21%), while SWAT alone overestimated it (11.9%). The combined result of 9.12% extreme erosion is therefore more balanced. In conclusion, anthropogenic activities, especially agricultural expansion covering 91.61% of the sub-basin, are the primary drivers of soil erosion. The study recommends agroforestry, terracing, and check dams in high-risk areas, alongside policy measures for sustainable land use.