一种利用卫星红外遥感资料反演晴空大气参数的物理统计方法

STATISTICAL METHOD FOR RETRIEVING THE CLEAR ATMOSPHERIC PARAMETERS FROM SATELLITE INFRARED MEASUREMENTS

给出一种新的由卫星辐射观测反演大气温湿廓线的物理统计方法,并用该方法分别进行了理想试验和实际反演试验,再将反演结果和一维变分反演方法的结果进行比较.结果表明:在理想试验中物理统计方法反演的温度廓线在大部分高度好于一维变分法;水汽廓线在模式中层的反演结果比背景场略差,在其他高度有所改进,而一维变分方法则对水汽廓线的背景值修正较小.在实际反演试验中,物理统计反演法对温度廓线的反演效果不够理想,只在400hpa以上高度和近地面对背景场有所修正,然而对水汽廓线的反演效果较好.

A new statistical method was presented to retrieve the atmospheric temperature and moisture profiles from the satellite radiative measurements. By using the method some ideal and actual retrieval tests were done, and then the retrieval results of the statistical-physical method were compared with that of 1DVAR method. The results indicate that the retrieved temperature is more precise than those of 1DVAR retrieval results at most heights; and the retrieval precision of moisture profiles is lower than the background at middle model layer, and relatively high at other layers, while the 1DVAR method can amend a little to the background of moisture profile. In actual tests, the retrieval results of the temperature profiles are not good enough; the background can be amended only above 400hpa and near the surface, while the retrieval effect of moisture profile is relatively good.