摘要
提出用变分方法和正则化方法相结合反演双多普勒雷达资料的低层二维风场的方法,即在目标泛函中加入正则化项(先验信息项)且合理选择正则化参数。数值试验结果表明:正则化项的引入能有效滤除反演风场中因奇点和观测资料引起的误差,提高反演精度,得到最优的全局反演风场。
The method of direct solving equations is often applied in the retrieval of lower layer 2-D winds from the dual-Doppler radar data ,but there are many limitations：singular points shall pollute their circumambient retrieval results and the observations＇ errors would affect the accuracy of retrieval, etc. A variational method in combination with a regularization method is introduced in this paper, wherein a regularization term is added to the target functional and regularization parameters are reasonably selected. Numerical experiments show that the introduction of the regularization term can efficiently filter the errors of retrieval wind field, resulted from singular points and observations, improve the retrieval accuracy, and yield the optimal global retrieval wind field.
出处
《南京气象学院学报》
CSCD
北大核心
2008年第4期453-459,共7页
Journal of Nanjing Institute of Meteorology
基金
国家自然科学基金资助项目(40775023
40575052)
江苏省气象灾害重点实验室资助项目(KLME0606)
关键词
双多普勒雷达反演
变分方法
正则化方法
二维风场
retrieval of dual-Doppler radar
variational method
regularization method
2-D wind field
