Modelling Infiltration Based on Source-Responsive Method for Improving Simulation of Rapid Subsurface Stormflow

In humid hilly regions, macropore preferential flow in soils dominates the distribution of event water, thereby influencing the generation and development of runoff. However, the mechanism of how soil functions on macropore drainage and matrix absorption remains poorly understood due to complex soil water dynamics in a multi-porosity subsurface network. In this study, based on the source-responsive method that divides the soil into source-responsive and diffusive domains, the allocation ratio of infiltrated water in macropores recharging the matrix were derived and it was coupled with PIHM (Penn State Integrated Hydrologic Model) as PIHM-SRM (PS). By simulating the soil moisture process at profile scale and the runoff process at catchment scale, it was found that the PS overcame the difficulty of most hydrologic models in describing the process of replenishing moisture in dry soil. This leads to more satisfactory performance for flood peaks at the outlet (CCC > 0.84) and soil moisture peaks at three profiles (CCC = 0.97) compared to original PIHM models. Moreover, the separate channel of film flow in the PS further improves the simulation accuracy of peak response speed in subsurface floods under rainstorms (TP > 40 mm). Additionally, sensitivity analysis shows that the storage-discharge capacity of soil profiles dominates torrential flood forecasting in humid headwaters when considering the influence of macropores. Finally, considering the parameter-predictive property in the PS, field-based parameterized strategies are vital for distributed catchment modeling. This will enable the PS to improve flash torrent predictions in headwaters and be applied at catchment scales.