Abstract:
Using the Observation precipitation data and objective statistical verification method based on the MET system,the 24h precipitation forecasting capabilities of the SWCWARMS model,GRAPES model and ECMWF model are verified in the Sichuan region during the rainy season (May to October) from 2017 to 2013.The main results are: (1) Precipitation forecasted by SWCWARMS model is greater than observation from light rain to rainstorm,Precipitation forecasted by GRAPES model is greater than observation from light rain to hard rain,rainstorm is often underreported,ECMWF model light rain and moderate rain is larger than the observation, from heavy rain to rainstorm are more underreporting. (2) ECMWF model is superior to SWCWARMS model and GRAPES model in predicting rainy season from light rain to heavy rain in Sichuan,Part of hard rain and rainstorm prediction by the SWCWARMS model is better than the ECMWF model,TS score of GRAPES model is slightly lower. (3) The GRAPES model improved more than the SWCWARMS and ECMWF model in the forecast of moderate rain and above in 2018,and the false prediction of the SWCWARMS model in 2019 decreased significantly compared with 2017 and 2018;the capability of 3 models are equal in light rain and moderate rain,GRAPES advantage on the forecast of heavy rain and rainstorm in 2019,the advantage of ECMWF is in the forecast of heavy rain in autumn 2017 and early summer of 2018,the ability of SWCWARMS to forecast heavy rain and rainstorm is between the two models. (4) The ECMWF and SWCWARMS model forecast in eastern Sichuan are superior to western Sichuan,and the GRAPES model forecast in western Sichuan is superior to eastern Sichuan. Among the three models,there are more false prediction in East Sichuan than in West Sichuan,and the ECMWF model has more false prediction than the others.