中文摘要：

概述了利用特征向量统计回归反演算法，从EOS／MODIS的红外通道资料反演大气温湿度垂直分布过程，并与美国国家环境预报中心NCEP（National Centers for Environmental Prediction）等压面再分析场资料按照纬度和气压高度进行了真实性检验。结果表明：由MODIS资料反演得到的大气温湿度参数能够揭示大气温湿度的垂直分布。在各个等压面上均方根误差平均值在中纬度地区为3．39K，低纬度地区为1．40K，近地面层、对流层顶附近及下垫面地形复杂的区域误差较大，总体上低纬度地区要好于中纬度地区。反演的水汽误差也为低纬度地区小于中纬度地区，且随高度升高，中、高纬度误差均逐渐减小并接近。

英文摘要：

The retrieval algorithm based on an eigenvector regression method was summarized. The vertical distri- butions of the atmospheric temperature and moisture were retrieved using EOS/MODIS infrared data and were verified along the latitude and pressure altitude with isobaric surface reanalysis field data from NCEP （ national centers for environmental prediction）. The results indicate that the atmospheric temperature and moisture parameters retrieved by MODIS data can reveal its vertical distributions. The average of root mean square （RMS） errors at each isobaric surface is 3.39 K in middle latitude region and 1.40 K in low latitude region, respectively. The errors are significant near the ground and tropopause region as well as in complicated underlying surface region. In general, temperature retrieval results are better in low latitude regions than in middle latitude regions, so are vapor retrieval results. With the increasing of the height, the error decreases gradually in middle and high latitudes regions and is close to each other.