单通道算法地表温度反演的若干问题讨论——以Landsat系列数据为例

Some Issues in Land Surface Temperature Retrieval of Landsat Thermal Data with the Single-channel Algorithm

在地表温度反演中,Jiménez-Muoz和Sobrino开发的单通道(SC)算法因其需要的实时大气参数少而被广泛应用。由于SC算法在2003年提出时只提供了针对Landsat TM的大气参数,导致许多后续基于ETM+数据的地表温度计算也都采用TM的大气参数。即使SC算法于2009年提出了改进版,但这一混用现象仍未改观。因此,基于实测地表温度,以Landsat 5/7/8号卫星的热红外数据为例比较了2003和2009版的算法,探讨了Planck函数、λ取值和大气参数等常见问题。结果表明,2009版的算法明显提高了ETM+地表温度反演的精度,且以采用Planck函数和USGS提供的λ值时所获得的平均精度最高。在当前,如果要用SC算法来反演Landsat 8的地表温度,可在大气水汽含量较低时选用TIRS 10波段来单独进行。

The single-channel algorithm（SC）developed by Jiménez-Muoz and Sobrino in 2003 has been frequently used in retrieving land surface temperature（LST）.However,many studies employed the Landsat TM＇s atmospheric parameters for the LST retrieval of ETM＋ thermal data because the parameters for ETM＋ were not provided in 2003.Although Jiménez-Munoz and Sobrino revised the SC algorithm and provided the atmospheric parameters specialized for ETM＋in 2009,such misusing still not changes.Therefore,this paper compared the LSTs retrieved from Landsat TM/ETM ＋/TIRS thermal data with the SC＇s 2003 and 2009versions,discussed the correct usage of the Planck function and wavelength value,and explored the potential of the SC for retrieving LST from Landsat8 TIRS data.Results show that the 2009 version can achieve higher accuracy averagely for ETM＋thermal data than the 2003 version,when calculated using wavelength value provided by USGS and the brightness temperature computed with Planck function.This study also reveals that at this stage the LST of Landsat 8could be retrieved using the SC with TIRS band 10.This could be achieved when the atmospheric water vapor content is relatively low.