大气风温湿垂直观测网资料快速更新混合同化试验研究

Assimilation experiments of the Rapid Refresh Hybrid scheme with wind,temperature and humidity data in the vertical observation network

基于WRF预报模式、WRFDA Hybrid集合变分同化系统和ETKF方法,构建了面向城市气象观测网数据的快速更新混合同化系统。针对北京地区地基微波辐射计和风廓线雷达组网观测资料数据同化,开展了静态背景误差调整因子(特征长度尺度因子和方差因子)、局地化距离和集合权重系数4个重要参数敏感性试验研究。试验结果表明:当温度、相对湿度、u风和v风的特征长度尺度因子和方差因子分别调整为0.7/1.0、1.0/1.0、0.7/1.0和0.7/1.0,局地化距离和集合权重系数分别调整为11.2 km和0.5时,快速更新混合同化系统的分析场均方根误差最小。为对比三种常用同化方案,开展了默认参数混合同化、最优参数混合同化、三维变分同化对比试验,试验结果表明:在针对北京地区地基微波辐射计和风廓线雷达组网观测资料的快速更新同化预报试验中,混合同化方案表现优于三维变分,同时相对于默认参数混合同化方案,最优参数混合同化方案的风场、温度及湿度的分析场和预报场得到了进一步改善:风温湿的分析场均方根误差分别最大降低了13%、19%和5%,12~24 h预报场的均方根误差分别最大降低了2%、12%和5%。

In this study,a Rapid Refresh Hybrid system was constructed based on the Weather Research and Forecasting(WRF)model,WRF Hybrid Data Assimilation system,and Ensemble Transform Kalman Filter(ETKF),while assimilating both Wind Profile Radar Detection(WPRD)and Microwave Radiometer(MWR)data.Experiments were performed on the impact of four important parameters on the system(that is,two tuning factors of static background error,localization scale and ensemble weighting factor),and contrast research was carried on to the results of the hybrid and 3DVAR schemes.Some encouraging conclusions were reached:Tuning these four parameters could improve performance of the Rapid Refresh Hybrid system,the analysis and forecast of the hybrid scheme with parameters not tuned were superior to those of 3DVAR,and the best results were those of the hybrid schemed with parameters tuned.