CMA-GEPS对中国超强梅雨天气过程的预报能力分析

Assessment of CMA-GEPS prediction capability for the extreme Meiyu process over China

基于中国气象局自主研发的全球集合预报系统CMA-GEPS,针对2020年6—7月的长江中下游地区超强梅雨天气过程,开展了中国梅汛期强降水的预报能力分析。结果表明,西太副高的稳定维持及夏季风的持续增强为梅汛期强降水的发生提供了有利的动力和水汽条件。CMA-GEPS对西太副高各指数的快速调整趋势预报,可提前7~10 d在预报中呈现;对西太副高强度和面积预报技巧与NCEP集合预报相当,表现为偏弱的估计,脊线和西伸脊点预报技巧与ECMWF集合预报相当,表现为脊线位置偏南、西伸脊点偏东的偏差。CMA-GEPS对东亚夏季风指数预报技巧在9 d以上,较控制预报提前2 d。CMA-GEPS的控制预报存在强度偏弱、雨带位置偏南的系统性偏差和漏报现象;基于空间和时间权重修正的邻域方案TSWNP,明显提高了大雨预报的技巧,减少了漏报的发生,优于控制预报和传统的单点概率预报,从而表明,TSWNP方案对梅汛期大雨过程的预报是有效且合理的。

This study evaluates the prediction capability of the Global Ensemble Prediction System in the China Meteorological Administration(CMA-GEPS)for the extreme Meiyu process over China in 2020.Results show that during the Meiyu season,the strong and stable western Pacific subtropical high(WPSH)and the gradually strengthened East Asian summer monsoon provide favorable dynamic and moisture conditions for strong rainfall.The CMA-GEPS can skillfully forecast the evolution trend of the WPSH index with 7—9 leading days.The prediction skills of the WPSH strength and area are about the same level as the results from the NCEP ensemble prediction system,and the WPSH strength presents a weaker bias compared with the observation.The CMA-GEPS prediction skills for the ridge line and the western ridge point index are comparable to the results from the ECMWF ensemble prediction system,and the forecasting bias is mainly attributed to the more southward location of the ridge line and the more eastward center position of the western boundary.For the East Asian summer monsoon,CMA-GEPS can skillfully predict the index with 9 leading days,which is two days earlier than the control forecast.The CMA-GEPS control forecasting bias is mainly attributed to the weaker precipitation intensity and more southward location of the strong rainfall belt;it also fails to predict heavy rain in some regions in the middle and lower reaches of the Yangtze River.The Time-Spatial Weight Ensemble Probability(TSWNP)scheme improves the prediction skills related to heavy rain with CMA-GEPS,and the scheme reduces the occurrence of missing precipitation predictions.The results of precipitation probability prediction are verified by observation and Brier scores,indicating that the TSWNP scheme is superior to the original single-point ensemble probability forecast method and the control forecast,and also has good application value for heavy storm prediction in the Meiyu period.