WRF-Hybrid背景误差协方差调整在台风同化及预报中的应用研究

WRF-Hybrid background error covariance adjustment in typhoon assimilation and forecasting

利用雷达径向风单点试验和2006年8月超强台风"桑美"个例,首先研究了静态背景误差协方差的尺度化因子和方差在台风系统雷达资料同化中对台风路径和强度预报的影响。结果表明:在高时空分辨率的雷达资料同化中,较小的尺度化因子能显著改进对台风路径的预报;尺度化因子的影响比方差的影响更为显著。基于上述实验结果,进一步研究了WRFDA-Hybrid系统中"流依赖"控制变量的水平局地化和垂直局地化对台风预报的影响。试验结果表明:当"流依赖"的水平局地化距离与静态背景误差协方差的尺度化因子具有等效影响范围的时候,WRFDA-Hybrid能够得到比较合理的分析结果。同时针对雷达观测资料的空间分布特征,本文提出了一种新的基于雷达探测高度的垂直局地化方案,对台风的强度和路径预报均有显著的改进。

By using the single observation test case in August, 2006, the influences of length scale on radar radial wind and the super typhoon Saomai factor and covariance static background error covari- ance in typhoon data assimlation on the typhoon track and intensity forecasting are studied. Results show that in high spatial and temporal resolution typhoon radar data assimilation, smaller scale factor can sig- nificantly improve the typhoon track prediction, and the impact of scale factor is larger than that of covari- ance. Furthermore, based on the above test results, the influence of the horizontal and vertical localiza- tion coefficient of the ＂flow dependent＂ in WRFDA-Hybrid system on typhoon prediction is investigated. Results indicate that reasonable analysis can be obtained when the equivalent influence radius occurrs in both horizontal localization of ＂flow dependent＂ and the length scale factor of static background error co-variance. Meanwhile, aimed at the spatial distribution features of radar observation data, a new vertical localization scheme for the radar detection height is proposed, which is able to better forecast typhoon in- tensity and track.