Verification and Analysis of Precipitation Forecast during the Meiyu Period of 2021 in Anhui Province
Mitigating substantial loss of life and damage to personal property from Meiyu rainfall calls for the accurate and reliable numerical weather predictions(NWP).A general approach to improve the accuracy of Meiyu rainfall forecast over a region of interest involves verifying and intercomparison multiple NWP models'results.This work focuseed on verifying and analyzing the Meiyu rainfall forecasts in Anhui province,China,based on the outcomes from seven NWP models,including three regional models(i.e.,CMA-MESO,CMA-SH9,Anhui WRF),three global models(i.e.,CMA-GFS,ECMWF,NCEP-GFS),and the Anhui Intelligent Grid(AIG),during the Meiyu period of 2021(from June 10th to July 10th).The traditional verification and MODE methods and TS(Threst Score)were employed to verify and evaluate above models quantitively.The results showed that AIG and regional models perform better than global models using the traditional verification method regarding clear-sky and rainfall accuracy,among which CMA-MESO outperforms others.For the heavy rain and above magnitude of heavy precipitation forecast,AIG has the highest TS of 23.83,followed by ECMWF(20.12)and CMA-SH9(19.34).Moreover,the MODE method findings indicate that different NWP models can predict Meiyu rainfall location and area differently.Specifically,compared to the observation,ECMWF and three regional models'Meiyu rainfall location have a large discrepancy,especially in latitude,and NCEP-GFS predicts a much smaller rainfall area.Furthermore,it is found that all models'heavy rain and above forecasts exhibit an apparent diurnal variation,with higher TS are observed in the period before midnight and during the morning and lower scores are in the afternoon to early evening.This phenomenon may be attributed to the dominant occurrence of short-duration convective rainfall triggered by surface heating from the sun in the afternoon during the Meiyu period.
verificationMODE methodMeiyu rainfallnumerical weather prediction model
国家科技期刊平台
NSTL
万方数据
维普
