Land Surface Temperature/Emissivity Inversion Based On Satellite Thermal Infrared Data

In this paper,based on the temperature independent spectral index(TISI),a directional land surface emissivity extraction algorithm based on FengYun-2G(FY-2G)geostationary satellite data in China is developed.Aiming at the land surface and climatic characteristics of the desert area in northwest China,an improved split window algorithm is proposed to extract the land surface temperature.The main work of the article is as follows:1.Using the radiation transmission calculation software MODTRAN 5,the sensitivity of atmospheric transmittance,atmospheric upward radiance and atmospheric downward radiance for the infrared channels IR1,IR2 and IR4 of the FY-2G satellite to view zenith angle,water vapor content,atmospheric temperature and the type and concentration of aerosol was analyzed.2.Based on the temperature independent spectral index,the day/night observation data of the mid-infrared and thermal infrared channels of the FY-2G geostationary satellite are used to retrieve the land surface emissivity of the land area of China.During atmospheric correction,the input atmospheric profile(atmospheric temperature,humidity,ozone,etc.)comes from the reanalysis data of the European Centre for Medium-Range Weather Forecasts(ECMWF).Using multiple sets of data observed during the day/night,the land surface bidirectional reflectivity at multiple times during the day can be obtained,and then the land surface emissivity of the mid-infrared channel by the model parameters can be fitted with the land surface bidirectional reflectance distribution function(BRDF)model.The land surface emissivity of the thermal infrared channel can be calculated according to the temperature independent emissivity spectral index and the mid-infrared channel emissivity.Cross-validate the inversion results of the FY-2G thermal infrared two-channel land surface emissivity with the land surface emissivity products of the corresponding channels of MODIS.The verification results show that the day/night land surface emissivity results obtained from the inversion are in good agreement with MODIS land surface emissivity.3.According to the classification data of China's land surface types provided by the Resource and Environmental Science Data Center,six types of land surface types were classified in China's land surface.According to the inversion results of LSEs,the characteristics of the average land surface emissivity of different land surface types and their daily variation are analyzed.The results show that there is a difference between the land surface emissivity of different land surface types.The land surface emissivity changes with time throughout the day,and the variation range is around 0.01.4.Aiming at the desert area in northwestern my country,an improved split window algorithm is proposed for land surface temperature inversion.Based on the general split window algorithm,the inversion algorithm introduces the path correction items related to the view zenith angle to correct the inversion error caused by the long transmission path when the observation zenith angle is large.Using the clear sky atmospheric profiles of eight sounding sites near the northwest desert area of China for nearly 20 years provided by the University of Wyoming,the split window algorithm coefficients were fitted to establish a lookup table for different ranges of land surface temperature,water vapor content,land surface emissivity,and view zenith angle.The relationship between the atmospheric water vapor content,the SW algorithm two thermal infrared channel atmospheric transmittance ratio and the view zenith angle are constructed to extract the atmospheric water vapor content,which is used to determine the water vapor content partition in the split window algorithm coefficient lookup table.5.The algorithm coefficients are determined according to the view zenith angle,land surface emissivity,water vapor content,and land surface temperature estimates value of each pixel,and the land surface temperature inversion is carried out in combination with the brightness temperatures of the two thermal infrared channels of top of atmosphere.The land surface temperature inversion results are compared with the land surface temperature obtained from the atmospheric profile data in the global clear sky atmospheric profile library Seebor V5.0.The results show that there is a high degree of agreement between the two land surface temperatures and the error is around 1K.The land surface temperature results obtained by different split window algorithms are compared,and the results show that the inversion results of the land surface temperature between different algorithms are somewhat different.