Meteorological-hydrological Coupled Runoff Forecasting Based On WRF-Hydro Model

The coupled meteo-hydrological modelling is an effective way for conducting hydrological forecasting research in small and medium-sized basins with scarce data.The quality of meteorological data,especially precipitation data,is very important for providing reliable meteorological-hydrological forecasting.In this paper,the study on the real-time streamflow forecast of the coupled meteo-hydrological model based on precipitation bias correction was carried out over the Zhanghe River Basin in Hubei Province.The successive correction method was introduced to correct the bias of the WRF model precipitation,and a coupled WRF/WRF-Hydro meteo-hydrological forecasting system based on the numerical weather prediction(NWP)model and the high-resolution distributed hydrological model was finally constructed.The main conclusions are as follows:(1)Comparing the precipitation forecast results of eight combinations of different physical parameterization schemes in the WRF model,the combination of the Thompson microphysics scheme and the Kain-Fritsch(new Eta)cumulus convection scheme performs relatively better and more stably.Therefore,it is more suitable for precipitation forecast of the WRF model in the study area.However,the forecasted precipitation is quite different from the China Meteorological Administration(CMA)Multi-source merged Precipitation Analysis System(CMPAS),especially in the precipitation events with strong local characteristics,indicating that the prediction accuracy of the WRF model in small and medium-scale basins in mountainous areas needs to be further improved.(2)Parameters calibration and verification of the WRF-Hydro model.All the saturation soil moisture content,the Manning's roughness coefficient,the infiltration factor and the surface roughness factor have an effect on the volume of the simulated streamflow,while the Manning's roughness coefficient and the surface roughness factor also have a great impact on the shape of the simulated hydrograph.In general,the WRF-Hydro model with the input data from CMPAS performs well in the streamflow simulation in the basin.However,the streamflow results simulated by ERA5 reanalysis data are relatively poor.It indicates that the quality of the precipitation data is the more important to the performance of the WRF-Hydro model,and the calibration of the model parameters cannot completely compensate for the errors of streamflow forecast caused by the accuracy problems of precipitation data.(3)The bias correction of precipitation.The WRF precipitation corrected by the successive correction method is closer to the CMPAS in both temporal and spatial distribution,and the scores of 1 h accumulated precipitation after correction are generally higher.It indicates that the successive correction method is one of the effective ways to correct the bias of the WRF precipitation.(4)The streamflow simulation based on the coupled WRF/WRF-Hydro meteo-hydrological model.The accuracy of the streamflow simulated by the WRF precipitation after correction has been significantly improved.It proves that correcting the bias of forecasted precipitation by the successive correction method can greatly improve the performance of the streamflow forecast.In general,the coupled WRF/WRF-Hydro meteo-hydrological system with the bias correction of precipitation has the potential in real-time streamflow forecasting and reconstruction of historical meteorological datasets in the Zhanghe River Basin.