利用全极化微波辐射计资料反演台风境内海面风场

作为一种新兴的被动遥感技术，全极化微波辐射计不仅可以提供海面风速产品，还可以提供海面风向产品.以往利用全极化微波辐射计观测亮温进行海面风场反演仅在晴空条件下进行，本文通过对观测亮温结合台风区域海面风场的分布特征进行分析，验证了全极化微波辐射计具有在台风等恶劣天气条件下进行海面风场观测的能力.基于敏感性分析实验，确定使用6.8 GHz和10.7 GHz等低频通道组合可进行台风区域内海面风场反演.其中，海面风速反演使用基于统计的多元线性回归算法，同时对海面温度、大气水汽含量、云中液态水含量及降水强度等物理量进行反演计算，为海面风向反演做准备.海面风向反演使用物理统计法进行，借鉴散射计风向反演使用的最大似然估计法.通过在全极化辐射传输前向模型中加入降水对大气透过率的影响、设计第三和第四Stokes通道亮温环境影响修正函数，在实现台风区域内海面风向反演的同时减小了反演误差.通过对“云娜”台风境内海面风场进行数值计算，验证了本文反演算法的可行性，并对反演误差的空间分布特征进行了分析.将2004年各台风过程的海面风场反演结果与散射计风场产品进行对比，海面风速和海面风向反演的均方根误差分别为1.64 m·s^-1和18.02°.

Sea surface wind retrieval from polarimetric microwave radiometer in typhoon area

As a new technology in the field of passive remote sensing, polarimetric microwave radiometer can provide wind direction products as well as wind speed. In the past, the sea surface wind vector retrievals from polarimetric microwave radiometer are only in clear sky conditions. By analyzing the distribution characteristics of polarimetric microwave radiometer brightness temperature combined with wind vector distribution characteristics in typhoon area, we confirmed that polarimetric microwave radiometer is capable of observing sea surface wind under inclement weather conditions such as typhoon. Through sensitivity analysis experiment, we choose the low frequencies group (6.8 GHz and 10.7 GHz) to use in sea surface wind vector retrievals. Sea surface wind speed retrieval uses the multiple linear regression algorithm based on statistics, and other physical quantities like sea surface temperature, atmospheric water vapor content, cloud liquid water content and precipitation intensity should be retrieved at the same time, preparing for sea surface wind direction retrieval. Sea wind direction retrieval uses the physical statistical method-maximum likelihood estimation, learned from scatterometer. By adding precipitation influence in polarimetric microwave radiative Forward Model, which comes from Naval Research Laboratory for the U.S. Navy, and designing environmental influence empirical correction function for the third and fourth Stokes channels, we not only achieved wind direction retrieval but also reduced retrieval errors. Through numerical calculation of sea wind vector retrievals under Rananim, we verified the feasibility of the retrieval algorithm, and analyzed the spatial distribution characteristics of wind speed and direction retrieval errors. Comparing sea surface wind vector retrieval results with scatterometer wind field products from every typhoon process of the year 2004, the root mean square error is 1.64 m·s^-1 for sea surface wind speed and 18.02°for sea surface wind direction.