晴空时大气红外遥感及其反演问题研究 II. 反演试验研究

前文给出了卫星红外观测资料的反演处理方法，并对方法作了理论上的分析，在本文中，我们将此方法应用到TOVS、GOES-8实际资料及AIRS模拟观测资料处理中并重点对TOVS资料进行试验。在反演中，我们将大气温度廓线和水汽混合比（lnq）廓线用各自的经验正交函数（EOF）表示以提高反演的稳定性并缩短计算时间，同时采用一种过滤操作以加速迭代的收敛速度。除了用本文提出的方法对TOVS资料进行处理外，还用国际TOVS处理软件包（ITPP-4.0）对同样的资料进行反演。试验表明，本文方法的反演结果优于ITPP的结果。AIRS模拟反演结果表明，高分辨率红外垂直探测器的遥感精度能够达到温度1 K、水汽10%。

Infrared Remote Sensing of Clear Atmosphere and Related Inversion Problem. Part II: Experimental Study

The inversion method was described in part I of this paper for retrieval of the atmospheric parameters from satellite observed infrared radiances under clear conditions. In this part, the method is applied to the real satellite observed and simulated data, especially to the TIROS N Operational Vertical Sounder (TOVS) data. The atmospheric temperature and water vapor mixing ratio profiles, are expressed as a linear combination of their Empirical Orthogonal Functions (EOFs) in order to improve the stability of inversion procedure and reduce the computation time. Also, a filtering operation is used to speed up the iterative convergence. The Root Mean Square Deviation (RMSD) between retrieval and radiosonde observations is compared with that between International TOVS Processing Package (ITPP4.0) retrieval and radiosonde observations. Experiments show that the new method is superior to the ITPP version 4.0 schemes.