EnKF局地化算法对雷达资料同化的影响研究
投稿时间: 2015-05-29  最后修改时间: 2015-08-14  点此下载全文
引用本文:
摘要点击次数: 467
全文下载次数: 
作者单位E-mail
高士博 南京信息工程大学 shibogao@126.com 
闵锦忠 南京信息工程大学 minjz@nuist.edu.cn 
黄丹莲 南京信息工程大学  
基金项目:江苏省普通高校研究生科研创新计划项目(KYLX_0829);国家重点基础研究发展计划(973 计划)项目(2013CB43013)
中文摘要:分级集合滤波(Hierarchical Ensemble Filter ,简称HEF)和采样误差修正(Sampling Error Correction,简称SEC)局地化算法能够使采样误差取得极小值,且不需要给出距离的定义。为了检验其理论优势,基于集合卡尔曼滤波(Ensemble Kalman Filter,简称EnKF)方法同化模拟雷达资料,通过与Gaspari-Cohn(简称GC)局地化算法对比,分析不同局地化算法对 EnKF同化效果的影响。结果表明:HEF 和 SEC 局地化算法的雷达回波在水平和垂直方向上均强于GC 局地化算法。HEF局地化算法各个变量的离散度最高,均方根误差最低,SEC局地化算法离散度略低,均方根误差略高,GC 局地化算法离散度最低,均方根误差最高。相比于 GC 局地化算法,HEF 和 SEC 局地化算法的冷池强度减弱,面积减小,下沉气流的速度和范围增大,雹霰混合比的大小和覆盖面积增大。通过模拟发现, HEF局地化算法模拟的北侧对流中心最强,SEC局地化算法模拟的南侧对流中心最强,且模拟出(40km, 60km)处的强对流中心。HEF局地化算法模拟的冷池强度最强,HEF和SEC局地化算法基本上模拟出北侧的雹霰混合比高值区。这表明 HEF局地化算法有效地改进了基于GC局地化算法的EnKF雷达资料同化效果,SEC局地化算法减小了计算量,是HEF局地化算法较好地近似。
中文关键词:集合卡尔曼滤波  雷达资料同化  HEF局地化算法  SEC局地化算法  GC局地化算法
 
Research on the Impact of Localization Methods on the Radar Data Assimilation Using EnKF
Abstract:Hierarchical ensemble filter (HEF) and sampling error correction (SEC) localization methods can minimize sampling error without giving definition of physical distance. To examine advantages of the two methods, experiments of assimilating radar data are conducted using EnKF. Compared with Gaspari-Cohn (GC) experiment, the influence of localization methods on assimilation effect is investigated. Results show that analysis reflectivity of HEF experiment is bigger than GC experiment in horizontal and vertical directions. The spread of HEF experiment is largest while the root mean square error is the smallest. The spread of SEC experiment is larger while the root mean square error is smaller than GC experiment. Cold pools of HEF and SEC experiments are weaker and their areas are smaller. Areas of vertical wind and Graupel mixing ratio are bigger while their values are larger. Through simulation of analysis fields, it is found that the northern branch of the convective system of HEF experiment is stronger than SEC and GC experiments while the southern branch of the convective system of SEC experiment is stronger than HEF and GC experiments. SEC can simulate the new convective cell at (40km, 60km). The cold pool of HEF is coldest. Both HEF and SEC can simulate the center of graupel mixing ratio. Those results prove that HEF and SEC localization methods can improve the performance of EnKF based on GC localization method.
keywords:EnKF  radar data assimilation  HEF localization  SEC localization  GC localization
查看全文  查看/发表评论  下载PDF阅读器