Evaluation of multi-scale model forecasts for subtropical high edge-type precipitation in the Haihe River Basin
Based on six kinds of subjective and objective precipitation forecast data,including ECMWF(EC),GRAPES-GFS(CMA-GFS),GRAPES-3 km(CMA-3 km),GRAPES-SH9(CMA-SH9),GRAPES-BJ(CMA-BJ)and National Guided Forecast(NWGD),hourly precipitation data from automatic stations and gridded precipitation fusion product data provided by the National Meteorological Information Center,by using the conventional inspection and scoring method,MODE(Method for Object-based Diagnostic Evaluation)method and SAL(Structure Amplitude Location)method,the forecast results of four subtropical high system-influencing precipitations in the Haihe River Basin in 2020 were evaluated to compare and analyze the advantages and disadvantages of these six products.Results are as follows:the TS of the six precipitation forecast products showed a decreasing trend with the increase of the precipitation magnitude in the four precipitation processes.For the forecast of light rain,CMA-SH9、CMA-3 km and EC model had a slight advantage,but as the precipitation magnitude increasing,the advantage of the mesoscale model became more prominent.At the same time,the large-scale models were high-altitude for heavy rain forecasting.Compared with other numerical models,the forecasts of CMA-3 km and CMA-BJ werwmore stable due to they both ranked the top three in the rate of empty and false alarms of rainstorms.Compared with the other five models,the correlation coefficient between the NWGD and the actual precipitation was the highest,and the accuracy rate of rain or shine ranked first,while the standard deviation of the CMA mesoscale model forecast was smaller.Using the MODE method and the SAL method to conduct a spatial inspection of heavy rain,it was found that the CMA-3 km and EC models could be used for rainband shape forecast.For the prediction of structure,strength and location,the CMA mesoscale model could be comprehensively referred to.
万方数据
维普
