基于AMSR2数据的青藏高原地表温度反演

Retrieval of land surface temperature from AMSR2 data over the Qinghai-Tibetan Plateau

将青藏高原大致分为北部、西部和东部三个部分,利用2017年7月的AMSR2(Advanced Microwave Scanning Radiometer 2,即先进的微波扫描辐射计2)的一级亮温数据,结合NCEP(the National Centers for Environmental Prediction,即美国国家环境预报中心)FNL(Final)全球地表温度产品以及FY-3B(风云三号B星)上的MWRI(Microwave Radiation Imager,即微波成像仪)数据中的地表覆盖类型数据,通过AMSR2单通道亮温数据和NCEP/FNL表面温度产品得到了研究区域地表温度反演的经验算法,进而得到多通道反演算法.并且,对比了两种反演结果,并比较了对地表覆盖进行分类前后的多通道反演结果.结果表明:多通道反演结果和NCEP/FNL产品的相关系数远大于单通道反演结果,多通道反演结果的均方根误差比单通道反演结果小1~2 K;对地表覆盖进行分类后的多通道反演精度比不分类时的反演精度略高;降轨的反演结果普遍比升轨的反演效果好.

The Tibetan Plateau in the study area is roughly divided into three parts:the north,the west and the east.Using the brightness temperature data of AMSR2(Advanced Microwave Scanning Radiometer 2)in July 2017,combined with the global land surface temperature products of NCEP/FNL(the National Centers for Environmental Prediction Final)and the surface coverage type data in FY-3 B/MWRI(FY-3 B/Microwave Radiation Imager),inversion of the regional surface temperature empirical algorithm is obtained by AMSR2 single-channel brightness temperature data and NCEP/FNL surface temperature products,and then the AMSR2 multi-channel inversion algorithm is obtained.The single-channel inversion results were compared with those from multi-channel inversion algorithm before and after classification of the surface cover.The results show that the correlation coefficient of multi-channel inversion results is much larger than that of single-channel inversion results.The root mean square error of multi-channel inversion results is 1~2 K smaller than that of single-channel inversion results.The accuracy of multi-channel inversion after classification of surface coverage is slightly higher than that of non-classification;the inversion results from descending orbits data are generally better than those from ascending orbits data.