Comparison of hydrological model ensemble forecasting based on multiple members and ensemble methods

Abstract Ensemble hydrologic forecasting which takes advantages of multiple models has made much contribution to water resource management. In this study, four hydrological (the Xin’anjiang model (XAJ), Simhyd, GR4J, and artificial neural network (ANN) models) three ensemble methods simple average, black box-based, binomial-based methods) were applied compared simulate the process during 1979–1983 in representative catchments (Daixi, Hengtangcun, Qiaodongcun). The results indicate that for a single model, XAJ GR4J performed relatively well with averaged Nash Sutcliffe efficiency coefficient (NSE) values 0.78 0.83, respectively. For models, show method (dynamic weight) outperformed volume error reduced by 0.8% NSE value increased 0.218. best performance on runoff occurs Hengtang catchment integrating based binomial method, achieving 2.73% 0.923. Finding would provide scientific support engineering design resources management study areas.