多参数背景场误差模型在散射计资料台风风场反演中的应用  

Application of the multi-parameters error model in cyclone wind retrieval with scatterometer data

作　　者：钟剑[1] 费建芳[1] 黄思训[1] 黄小刚[1] 程小平[1] 

机构地区：[1]中国人民解放军理工大学气象海洋学院,南京211101

出　　处：《物理学报》2013年第15期575-581,共7页Acta Physica Sinica

基　　金：国家公益性行业专项(批准号:GYHY201106004);国家自然科学基金(批准号:41175025;41005029;41105012;41105065)资助的课题~~

摘　　要：利用散射计资料反演海面风场时,台风区域普遍存在降雨使得风场反演误差很大,引入降雨地球物理模型函数(GMF+Rain)及多解方案(MSS),结合二维变分(2DVAR)模糊去除思想风速反演误差很大程度减小,但风向反演误差仍有待进一步改善,如何进一步减小风向反演误差有待进一步研究.文章介绍了2DVAR模糊去除方法的基本思想,针对背景场误差较大时,2DVAR模糊去除风向误差较大,引入包含若干参数的背景场误差模型.基于台风个例数值试验结果,着重从理论分析角度讨论各参数关于2DVAR模糊去除效果的敏感性,进而提出最优参数设置方案以改善风向模糊去除效果.2006年"摩羯"台风QuikSCAT数据风场反演数值试验结果结合理论分析表明:引入多参数误差模型,通过设置粗糙误差概率等于0,2DVAR风向模糊去除效果明显改善;同时,背景场的影响可通过增大背景场误差方差,减小背景场误差相关尺度和减小粗糙误差概率而减小,进而减小在背景场误差较大情况下的风向反演误差.Combined with the multiple solution scheme （MSS） and the rain considered Geophysical model function （GMF＋Rain）, the two-dimensional variational （2DVAR） ambiguity removal technique is applied to the cyclone wind retrieval under rain condition with QuikSCAT scatterometer data. With the GMF＋Rain model, the retrieved wind speed is effectively improved, but large wind direction error still exists when the background is in large error. In this paper, a changeable multi-parameter error model is introduced in the 2DVAR to reduce the wind direction error, and the sensitivity experiments of 2DVAR to its error model parameters are studied with cyclone Yagi QuikSCAT data, to choose the best parameters setting for cyclone wind retrieval with theoretical explanation. Numerical results show that 2DVAR is more effective in wind direction ambiguity removal with the proposed multi-parameter error model when the gross error probability in the multi-parameter error model is set to zero in comparison of the standard setting. The influence of the background is decreased with increasing backround error variance, decreasing the background error correlation length, or decreasing the gross error probabilities in multi-parameter error model.

关 键 词：台风风场反演 二维变分 多参数误差模型 散射计资料 

分 类 号：P444[天文地球—大气科学及气象学]