Prediction Abilities of Subseasonal‐to‐seasonal Models for Regional Rainstorm Processes in South China

Based on the daily precipitation data and the forecast data from subseasonal‐to‐seasonal (S2S) models, the prediction abilities of four S2S models for regional rainstorm processes are assessed. The results show that: Compared with the observations, the precipitation amounts that predicted by CMA, KMA, NCEP and UKMO models are lower. If the rainfall area is more concentrated and the precipitation amounts are greater, the biases are larger. When the lead times are greater than 15 days, the models cannot predict the occurrences of regional rainstorms. Among the four models, NCEP and UKMO models have better prediction abilities. NCEP model can make real‐time forecast four times a day and initial time forecast results at 0000, 0600, 1200 and 1800 UTC are different. Compared with an initial time forecast, average of four initial time forecast results is closer to the observed values. The less precipitation predicted by NCEP, CMA, KMA and UKMO models may result of the underestimate of cyclonic circulation convergence intensity and water vapour flux amounts.