GPS/PWV资料在梅雨锋暴雨个例中的同化试验

Assimilation Experiment of GPS/PWV Data in The Rainstrom Case of Meiyu Front

基于WRF(Weather Research and Forecasting Model,天气预报模式)及其三维变分同化系统3DVAR,利用江苏省GPS/PWV(PWV:Precipitable Water Vapor,GPS反演得到可降水量)资料,并将其与探空资料比对订正,针对2011年6月18日梅雨锋暴雨进行3 h循环同化模拟。在降水参数化方案敏感性试验与单点同化试验基础上,设计多组试验对6 h降水量进行TS(Threat Score)评估。结果表明:(1)同化订正GPS/PWV资料对降水预报能力显著提高,特别是大雨、暴雨量级以上的预报能力;(2)降水量的RMSE(Root Mean Squared Error,均方根误差)相比控制试验均减小,CC(Correlation Coefficient,相关系数)均增大,最显著试验RMSE从19.1 mm下降到12.6 mm,CC从0.45上升到0.74;(3)NMC方法统计的背景误差协方差条件下中雨至暴雨量级TS评分均有一定程度提高,默认的背景误差协方差在大雨以上量级TS评分大幅提高。

Based on WRF （Weather Research and Forecasting Model） and its three-demensional variational data assimilation system 3DVAR, three hours cycling assimilation experiments of Meiyu front rainstorm occured on June 18, 2011 are carried out with GPS/PWV （ PWV ： Precipitable Water Vapor） data which are corrected by comparing with sounding data. Based on the sensitivity researches of parame- terization schemes for precipitation forecast and single point assimilation tests, several experiments are de- signed to assess 6 hour accumulated rainfall by TS scores （Threat Score）. The results show that ： （ 1 ） the capacity of precipitation forecast is significantly increased with the assimilation of the corrected GPS/ PWV data, especially in heavy rain and torrential rain. （2） RMSE（ Root Mean Squared Error） are re- duced and CC （correlation coefficient） are increased by comparing with control tests. The RMSE drops from 19. 1 mm to 12. 6 mm and the CC increases from 0. 45 to 0.74 in the most significant tests. （3） The TS scores are improved in the magnitudes from middle rain to torrential rain under the background error covariance derived from NMC method, while TS scores are improved significantly in the magnitudes be- yond heavy rain under the default background error covariance.