中文摘要：

本文采用一维变分（1D—Var）方法对卫星反演相对湿度资料进行质量控制试验。结果表明，卫星反演相对湿度在质量控制后，其500hPa及以上等压层的绝对平均误差有所减小，其中400hPa层减少最明显，同时绝大部分层次的均方根误差减小。1D—Var质量控制能改善中高层干区偏湿和低层湿区偏干的卫星反演相对湿度。

英文摘要：

A one-dimensional variational （1D-Var） method was applied to the quality control of the satellite relative humidity retrievals, with the 6-h forecasts of the MM5 mesoscale model as the background profiles and the retrievals as the first guess profiles. It shows that due to the 1D-Var quality control, the absolute mean biases between the re- trievals of the levels above 700hPa and the collocated radiosonde observations are reduced, and so are the root mean square errors in the almost levels. 1D-Var quality control improves the humidity retrievals, especially the unreal wet retrievals in the dry areas of the mid and upper troposphere.