西北太平洋(含南海)热带气旋路径集成预报分析

OPERATIONAL ENSEMBLE FORECASTING AND ANALYSIS OF TROPICALCYCLONES OVER THE WESTERN NORTH PACIFIC(INCLUDING THE SOUTH CHINA SEA)

基于2004—2009年中国中央气象台、日本气象厅、美国联合台风警报中心、欧洲中心对西北太平洋和南海编号热带气旋主客观预报资料,利用算术平均、多元回归以及历史平均误差等三种集成方法,建立了热带气旋路径集成预报业务化系统。通过2007—2009年的业务运行结果分析发现,欧洲中心客观预报参与的24、48和72 h集成比主观预报三个成员集成预报水平分别提高约2%、3%～5%和3%～5%,减小误差2.5 km左右、6～9 km和10～12 km。技巧分析发现,24～72 h集成预报有正技巧,多元回归集成技巧相对稍低,而算术平均和以各成员平均误差的平方倒数为权重系数的集成技巧对于各集成成员来说技巧差异不大。96 h集成预报对欧洲中心的客观预报没有正技巧。

Based on TC data from CMO,JMA,JTWC,ECMWF over the years 2004—2009,three ensemble methods-arithmetic mean,multivariate regression,and historical mean error-are used to build an operational ensemble system for TC track forecasting.Results of operational runs for the years 2007—2009 showed that the ECMWF objective forecasts improved the 24 h,48 h,and 72 h forecast skill of three members of subjective forecast by 2%,3%～5%,3%～5%,respectively and reduced track bias by about 2.5km,6～9 km,10～12 km,respectively.Skill analysis indicated that the 24～72 h ensemble forecast had positive skill.For individual ensemble members,the multivariate regression shows relatively low skill while the arithmetic mean and the ensemble that takes as weighting coefficients the reciprocals of the squares of mean errors of all members have comparable effects.96 h ensemble forecast had no positive skill for the ECMWF objective forecast.