A COMPARATIVE STUDY OF HILLSLOPE AND SLOPE SEGMENT APPROACHES IN A SEMI-DISTRIBUTED INFLOW PREDICTION MODEL

Accurate inflow prediction models are essential decision-support tools for effective reservoir management and optimal water resource utilization.Among semi-distributed hydrological models, the hillslope-based approach has traditionally offered the finest spatial resolution.However, this study explores the integration of a higher-resolution spatial discretization method-slope segments-originally introduced by Tantasirin et al. ( 2016) for soil erosion modeling, into inflow prediction modeling for reservoir systems in the Lam Phra Phloeng watershed, Thailand.The Lam Phra Phloeng watershed was subdivided into 4,839 hillslope units and 54,378 slope segments.Compared to hillslopes, slope segments provided more homogeneous spatial units, reducing internal variability in land use and soil properties.This refinement allowed for more representative Curve Number (CN) values, enhancing the accuracy of runoff estimation using the SCS-CN method.Model performance was evaluated using three metrics-Percentage Error in Peak Flow (PEPF), Root Mean Square Error (RMSE), and Nash-Sutcliffe Efficiency (NSE)-during the validation years 2015, 2019, and 2020.The slope segment-based model consistently outperformed the hillslope-based model.For example, PEPF values were significantly lower (0.18%, 0.18%, and 2.02% vs. 1.25%, 1.47%, and 7.01%), RMSE values were reduced (0.61, 1.23, and 4.32 MCM vs. 0.66, 1.50, and 4.48 MCM), and NSE values were higher (0.94, 0.36, and 0.44 vs. 0.90, 0.30, and 0.40).These results demonstrate that slope segment-based spatial subdivision enhances the predictive performance of inflow models and offers a promising approach for improving hydrological modeling in reservoir systems.Future applications in diverse watersheds could further validate its utility for water resource planning and management.