Testing and improving a distributed rainfall-runoff model using slope surface flow as a new observation metric

This study proposes slope surface flow monitoring via time-lapse cameras as a novel validation metric for hydrological modeling in steep forested catchments. We evaluated a kinematic wave-based distributed hydrological model (1K-DHM) using both river discharge and slope surface flow duration observed during flood events. While 1K-DHM reproduced hydrographs well, it significantly overestimated slope surface flow duration, indicating structural limitations in representing vertical infiltration. To address this, we incorporated a Green-Ampt (GA) percolation module into 1K-DHM. The revised model (1K-DHM-GA) improved both hydrograph reproduction (the Nash-Sutcliffe Efficiency: 0.848, RMSE: 0.021 m3/s) and slope surface flow duration agreement (Jaccard Index: 0.745). Results demonstrate that integrating vertical percolation improves runoff simulation and that slope surface flow monitoring provides valuable constraints for evaluating model structure under equifinality.