Forecasting of an extreme flood in mountainous area of North China based on WRF-Hydro with distributing parameters

The flood events in the mountainous area of northern China has the characteristics of high intensity and strong sudden occurrence, and atmospheric-hydrological coupling system can improve the forecast accuracy and prolong the lead time. This paper discusses the simulations of the enhanced WRF-Hydro model on a historical flood that occurrs in a mesoscale catchment of Taihang mountain on July 21, 2012. Firstly, the precipitation accuracy of WRF, WRF data assimilation, co-kriging merging method of radar QPE data are as three different input sources for WRF-Hydro. The results show that the rainfall of merging QPE can achieve better simulations in time and space. In addition, the rainfall of WRF assimilation data is obviously better than that of WRF, but still underestimates the rainfall values. The extreme event rainstorm mainly proceeds in 5 hours, and for the assimilation data, the spatio-temporal simulations of the rainfall data in the first 2 hours are slightly poor. Hence we compare the combination of the first few hours to use the merging QPE and following by assimilation precipitation as the model input. In addition, according to the parameters of the WRF-Hydro model, a gridding parameter calibration method based on topographic index is constructed.