Evaluation of a Stand-Alone WRF-Hydro Modeling System Using Different Rainfall Forcing Data: Case Study Over the Godavari River Basin, India

Among all the variables that relate to the water cycle, precipitation is considered the most important variable for streamflow modeling and forecasting. Uncertainty in the precipitation measurement is one of the chief limitations in the hydrological modeling where the hydrology model is run in a stand-alone mode. In this study, the WRF-Hydro model is used for modeling streamflow over Godavari basin. This study explores the effect of uncertainty in the precipitation forcing in simulating the streamflow by the uncoupled WRF-Hydro model. For the present study, satellite-based [Global Precipitation Measurement (GPM), Tropical Rainfall Measuring Mission (TRMM), Precipitation Estimation from Remotely Sensed Information Using Artificial Neural Networks-Climate Data Record (PERSIANN-CDR)] precipitation products are used. Furthermore, the present study also utilized gauge-based [Indian Meteorological Department (IMD) gridded] and reanalysis-based [Global Land Data Assimilation System (GLDAS)] precipitation products. The uncertainty in the satellite-based and reanalysis-based precipitation is evaluated quantitatively by using the gauge-based IMD gridded dataset as the reference dataset. The results indicate that the IMD gridded rainfall forced model output streamflow is in agreement with the observed streamflow at four stations having positive Nash–Sutcliffe coefficient of efficiency (NSCE) values. The results indicate the importance of utilizing precipitation datasets (satellite/reanalysis) having similar spatial and temporal variability of precipitation as the observed rainfall for accurate simulation of streamflow using hydrological models.