Runoff Forecast And Uncertainty Analysis Coupled With TIGGE Rainfall Information

With the continuous development of numerical weather prediction technology,the accuracy and reliability of rainfall forecast products have been significantly improved,providing the possibility for hydro-meteorological coupling forecast.Coupling quantitative rainfall forecast products with hydrological models is conducive to improving the accuracy of river basin hydrological forecasting and extending the effective forecast period of hydrological forecast,and also has important guiding significance for flood warning,flood prevention and disaster relief,water resources management and allocation,etc.However,in the practical engineering application of hydro-meteorological coupling forecast,there are inevitably problems of the applicability of numerical rainfall forecast and the uncertainty analysis of hydrometeorological coupling model.Therefore,a challenging and urgent problem that needs to be solved is how to carry out hydro-meteorological coupling forecast research in view of the applicability of numerical rainfall forecast and the uncertainty of hydrometeorological coupling models,so as to improve the accuracy of hydrological forecast and extend the forecast period.For this reason,this paper takes the section from Xiluodu to the Three Gorges on the mainstream of the Yangtze River as the research object,and carries out runoff forecast and uncertainty analysis research coupled with TIGGE(THORPEX Interactive Grand Global Ensemble,TIGGE)rainfall information,the applicability of the ensemble precipitation forecast products of different forecast centers in the upper reaches of the Yangtze River was explored,a hydro-meteorological coupling forecast system was constructed in the upper reaches of the Yangtze River,and a hydro-meteorological coupling forecast model uncertainty analysis method was proposed.The research results can provide forecasters and decision makers with two kinds of runoff forecast information,deterministic and uncertain,and strongly support the runoff forecast of the river basin and reservoir operation.The main research contents and innovative achievements of this paper are as follows:(1)Focus on the applicability of the ensemble forecast data from different forecast centers in the upper reaches of the Yangtze River,this paper used the measured precipitation data as a reference,and used TS score,sunny/rainy threat score,dimissal alarm rate,false alarm rate and Talagrand distribution map to carry out the quantitative and qualitative tests on the ensemble precipitation forecast results of each forecast centers.In addition,in order to analyze the error distribution characteristics of each forecast centers,the spatial distribution of the average daily precipitation in the flood season was tested.The research results show that,overall the applicability of ECMWF forecast center data in the upper reaches of the Yangtze River is better than NCEP and JMA,and ECMWF can be used as the precipitation input for the forecast period of the hydrological model to extend the forecast period of runoff forecast in the basin.(2)Focus on the construction of the hydro-meteorological coupling model in the upper reaches of the Yangtze River,this paper introduced the MOSCDE multi-objective algorithm to optimize and calibrate the hydrological model parameters of each interval at first,and used the improved TOPSIS non-inferior solution optimization method to optimize model parameters of each interval;Then,the ECMWF rainfall information was used as the forecast period rainfall input of the hydrological model to obtain the predicted flow process to verify the application effect of ECMWF in hydrological forecast;Finally,on the basis of the above theoretical research,the corresponding realtime runoff forecasting system function module was developed.The research results show that the parameters optimized by the improved TOPSIS non-inferior solution optimization method have good forecasting effects;the Xin'anjiang model has good applicability in the Yangtze River Basin;the hydro-meteorological coupling model can extend the forecast period of hydrological forecast while ensuring the forecast accuracy during a certain forecast period.(3)Focus on the hydrological uncertainty of the hydro-meteorological coupling model,the hydrological uncertainty processor PD-HUP(Precipitation-Dependent Hydrological Uncertainty Processor,PD-HUP)that considers the rainfall of forecast period was introduced,and the GMM model was used to fit the flow marginal distribution,and then the posterior probability density function under different rainfall scenarios was obtained,which was compared with PI-HUP(Precipitation-Independent Hydrological Uncertainty Processor,PI-HUP)forecast results and deterministic forecast results.The research results show that when the rainfall of the forecast period is heavy,the hydrological uncertainty increases significantly,and hydrological uncertainty increases with the increase of forecast flow,and also as the forecast period extends,the hydrological uncertainty increases;in terms of the accuracy index and physical meaning,the PD-HUP forecast results are the best,followed by PI-HUP,and the deterministic forecast results are slightly inferior to the first two;and with the increase of the forecast period,the accuracy of the three forecast results also decreases.