面向台风暴雨的集合预报成员优选订正技术研究及应用试验

Research for the Ensemble Member Optimization Correction Technique on Typhoon Rainstorm Forecast and Its Application Experiment

本文以2013—2015年主要登陆台风暴雨过程为研究对象,利用ECMWF降水和台风路径集合预报以及中央气象台实时业务台风中心定位资料,在统计分析的基础上,提出一种业务上可用的针对单模式集合预报的台风降水实时订正技术（简称集合成员优选技术）。结果表明,在登陆台风暴雨过程预报中,集合成员优选技术对改进集合统计量降水产品有明显的帮助,并较ECMWF确定性预报产品有一定优势;该方法对改进短期时效预报产品的效果优于中期时效预报,对大暴雨评分的改进高于暴雨和大雨评分。另外,本文基于概率匹配平均（Probability Matching average,PM）和融合（FUSE）产品的计算原理,提出融合匹配平均（Fuse Matching average,FM）产品,结果表明,对36 h时效预报,优选10~15个成员的PM产品TS（Threat Scores）评分可达最优,大暴雨评分较确定性预报提高近10%;对60和84 h时效预报,FM产品大暴雨评分较确定性预报提高超过20%。

Based on several rainstorm processes of landed typhoons in 2013-2015 and statistics, we propose a real-time correction method for typhoon rainstorm forecast （also called ensemble member optimization method） using ECMWF precipitation and typhoon track ensemble forecasts and operational positioning of National Meteorological Certre （NMC） for the operational track forecast. The results show that ensemble statistic precipitation products are improved significantly by using the ensemble member optimization method, and the improved products give better performances than the ECMWF deterministic forecasts. Improvements of ensemble statistic products for short-range forecast are more significant than mediumrange forecast, and that for extra torrential rain forecast are more significant than torrential and heavy rain forecast. Moreover, Fuse Matching average （FM） product is proposed in this paper, based on the characteristics and principles of Probability Matching average （PM） and fusing products. The results also show that the PM products with 10 to 15 optimal selected members give the best performance for 36 h forecast, while the threat scores of PM products can be promoted by 10 % approximately for extra torrential rain forecasts. For 60 h and 84 h extra torrential rain ted by above 20%, compared with the ECMWF forecasts, the threat scores of FM products can be promo deterministie forecasts.