Estimation of distributed Grid-Xin′anjiang Model parameters based on SoilGrids

To finely simulate the rainfall runoff process in small and medium-sized watersheds, rational estimation of the spatial parameters of the distributed hydrological model is of great importance. Utilizing the new version of the Global Digital Soil Mapping System (SoilGrids), this study developed a scheme to obtain spatial parameters for the Grid-Xin′anjiang Model (GXM). Sixteen selected flood events in the Chenhe watershed of Shaanxi Province from 2003 to 2012 were simulated using the GXM model, which were compared with those of the Xin′anjiang model (XAJ). The sensitivity of free water storage capacity and its spatial distribution characteristics were quantitatively analyzed in various flood phases. The results indicate that the GXM simulation reduced peak time error by approximately 0.31 h, with high accuracy of peak discharge and runoff depth, and that dynamic spatial distribution of hydrological elements, such as soil moisture, can be reasonably simulated. Free water storage capacity significantly impacted the deterministic coefficients of flood rise, peak phases, and relative runoff volume error during flood rise, but insignificant influence on flood recession. Free water storage capacity in valley and ridge areas is greater than that in middle segments of slopes.