2019年影响中国台风的降水预报误差及其来源研究

Research on Rainfall Forecast Errors and Their Sources of Typhoons Influencing China in 2019

针对2019年影响中国的8个台风,利用面向降水对象的CRA(contiguous rain area)方法研究了欧洲中期天气预报中心确定性预报的降水误差来源,及其在不同预报时段和降水量级下的变化趋势,分析了台风路径预报误差与降水对象的CRA位置误差之间的相关性,对比计算了台风路径修正与CRA shifting方法对于改进降水预报的作用,并评估了台风降水概率分布、径向分布和非对称分布的预报误差。结果表明:总体而言,台风降水预报的主要误差来自于位置误差和形态误差;除特大量级降水以外,台风降水对象的CRA位置误差与路径误差显著相关,通过修正台风路径能改进降水预报,但其效果要逊于CRA shifting方法;预报的台风降水概率密度分布形态与观测总体上较为一致,但台风核心区内的预报降水强度均大于观测;台风登陆或靠近我国沿海前后,预报降水较观测更靠近台风中心,且略滞后于观测,预报降水的非对称性明显弱于观测。

By using the object-oriented method of contiguous rain area(CRA),this paper investigates sources of the rainfall forecast error in deterministic forecasts by the ECMWF,and their changes with different rainfall levels and forecast periods when typhoons influenced China in 2019.Then,the correlations between typhoon track errors and displacement errors of rainfall events are analyzed and the performance improvements of rainfall forecasts calculated with track correction or CRA shifting are compared.Finally,forecast errors of rainfall probability distribution,radial and asymmetric rainfall distribution are verified and analyzed.The results are as follows.In general,the main forecast errors come from displacement error and shape error.Except for enormous amount of rainfall,track error is significantly correlated to the displacement error of CRA rainfall events.The improvement of rainfall forecasts made by track error correction is less than that made by CRA shifting correction.The shape of probability density distribution for rainfall forecasts resembles the observed one,but the forecasted rainfall intensity in typhoon core area is stronger than the observed one.Before and after typhoons’making landfall or approaching the coast,the forecasted rainfall is much closer to typhoon center than the observed one,and it lags behind the observation.The asymmetric structure of forecasted rainfall is significantly weaker than the observed one.