用多光谱卫星信息分析白昼云天条件下的湿度场

TOVS资料的晴空湿度场反演已进行多年 ,如何利用卫星遥感信息实现云天条件下的湿度场反演 ,仍是一个值得研究的问题。多光谱卫星图像上不同云类的云层高低分布蕴涵着三维湿度场的分布信息 ,文中用相关分析、最小二乘拟合分析和多元线性回归分析分别讨论了多光谱卫星信息与各标准层相对湿度的关系。标准回归分析的复相关系数R在地面最小 ,约为 0 6 4,随高度增加 ,R逐层增大 ,70 0hPa以上大于 0 7,2 5 0hPa以上大于 0 8。相应显著性检验的结果F值从地面的 30 71,也逐层增大 ,至 2 0 0hPa时达 119 36。逐步回归分析在剔除了不显著的变量因子后 ,回归自变量减少到 2～ 4个 ,反映显著水平的F值有较大提高。在此基础上建立了一组云天条件下的多元线性回归方程 ,由此可得到具有卫星图像像素分辨率 (8～ 10km)的 9层湿度场。

HUMIDITY FIELDS ANALYSIS IN DAYTIME CLOUDY SKY WITH MULTI-SPECTRAL SATELLTE INFORMATION

Despite that the retrieval study of hum idity fields in clear sky with TOVS data has been many years,how to obtain the retrieved humidity fields in cloudy sky with satellite remote sensing inform ation is also a problem.This paper addresses the relationship between multispectral satellite information and probed relative humidity in each of standard isobaric surfaces using correlation analysis,least-square fitting and multivariate linear regression separately.Complex correlation coefficient R(≈0.64) of the standard regression scheme is minimal at ground level and then grows with height,reaching>0.7 above 700hPa and>0.8 above 250 hPa,with significance test results (F value) increasing from 30.71 at ground to 119.36 at 200 hPa.When a stepwise regression scheme is adopted,most of insignificant variables have been removed in such a way that only 2-4 variables remain,and F value showing significance level is raised quite considerably.Accordingly,we have established a system of multi-variant linear regression form ulae that can be used in cloudy sky with the aim to determine standard-level RH fields of satellite imagery pixel resolution(8-10 km) directly from multispectral satellite information.