风暴尺度天气下利用集合卡尔曼滤波模拟多普勒雷达资料同化试验Ⅰ. 不考虑模式误差的情形

Storm-Scale Ensemble Kalman Filter Data Assimilation Experiments Using Simulated Doppler Radar Data.Part Ⅰ:Perfect Model Tests

本文在假定模式无偏差的情况下,利用一次风暴过程的模拟多普勒雷达资料进行一系列风暴天气尺度的集合卡尔曼滤波资料同化试验,检验集合卡尔曼滤波在风暴天气尺度资料同化方面的效果,并验证各集合卡尔曼滤波参数对同化效果的影响。试验结果表明,集合卡尔曼滤波能有效地应用于风暴尺度的资料同化;40个集合成员以及6km的局地化尺度能较好地滤除采样误差造成的虚假相关,同时可以将观测信息传递到无观测的模式格点;利用背景场加上空间平滑的高斯型随机扰动生成初始成员的方式较未经过平滑的方式有更好的分析效果;背景场扰动方法能够提高样本的离散度;只同化反射率的同化试验表明,反射率的同化效果较明显,也证明了集合卡尔曼滤波在非常规资料同化中的作用;增加径向风资料同化的效果优于只进行反射率同化的结果。

Spatial and temporal characteristics of the Indian Ocean Dipole（IOD） and the relationship between the variabilities of IOD and ENSO at several interannual time scales are addressed.It is found that the interannual variability of IOD is associated with that of ENSO significantly at the time scales of quasi-1-2 years,quasi-3 years,and quasi-4 years.Besides,IOD also exhibits a robust variation at the quasi-8-month time scale which appears to be independent on the ENSO.At the three interannual time scales mentioned above,the IOD contributes to the interannual variations of ENSO through the westward shifts of zonal wind anomalies associated with the IOD over the tropical Indian Ocean into the tropical Pacific Ocean.The similar process is not found for the quasi-8-month mode of IOD.In addition,the ocean-atmosphere interactions in the tropical Indian Ocean responsible for the interannual variations of IOD are different at the four different time scales.