台风暴雨数值预报中ATOVS资料的变分同化试验

THE USE OF ATOVS DATA IN NUMERICAL FORECAST ON HEAVY RAIN OF TYPHOON

利用ATOVS卫星辐射率资料和NCEP资料,采用中尺度数值模式WRF(V3.3)及其同化系统WRF-3DVar,设计了几组同化试验方案,对由0908号台风"莫拉克"导致的台湾南部2009年8月8—9日特大暴雨过程进行数值预报。结果表明,直接同化ATOVS辐射率资料对初始温度场、湿度场和风场均有较明显的改善,其中连续循环同化试验的改善效果最显著,同化后初始场增加了暴雨中心上空对流不稳定性,同时风场也表现出更强的低层辐合、高层辐散的中尺度特征。对比不同的辐射率资料发现,同化MHS资料对湿度场改进效果较好,同化AMSUA资料对温度场和风场改进效果较好,同时同化MHS和AMSUA资料改进效果整体上要好于前两者。在对降水预报定量检验中,同化试验的整体评分要高于控制试验,特别是降水临界值超过300 mm后评分提高最为显著。连续循环同化试验对这次台风暴雨的预报与实况最为接近。

A heavy rain caused by typhoon Morakot（0908） occurred in the south of Taiwan during 8—9 August 2009.By using ATOVS Satellite Radiance data and NCEP data,the process is simulated by the mesoscale numerical model WRF（V3.3） with different assimilation schemes designed by the WRF-3DVar.Results show that direct assimilation of ATOVS Satellite Radiance data can improve the distribution of temperature,humidity and wind fields efficiently,and the continuous loop assimilation experiment has the most significant effect.The adjusted initial field enhances the convectively instability and the wind filed displays the mesoscale trait with enhancement of lower wind convergence and upper wind divergence in the center of rainstorm.By contrasting the three experiments,it can be concluded that MHS radiance data is better to adjust humidity field,AMSUA radiance data is better to adjust temperature field and wind field,and synchronous uses of these both two kinds of data have better result.Both the ETS and the HSS experiments are proved better than control experiment during the quantitative verification of precipitation forecast,especially to the case when the critical value of precipitation is larger than 300 mm.The continuous loop assimilation experiment has the best forecast on the heavy rain of typhoon Morakot among the assimilation experiments.