Surface Component Temperature Inversion Based On ASTER Multi-band Data

The land surface temperature (LST) is an important parameter for the landsurface physical processes on both regional and global scale, integrating the results ofinteraction between the land and the atmosphere, the good indicator of energy balanceand greenhouse effect of the earth, and has been extensively used in climatology,hydrology, biology, military affairs and bio-geochemistry, etc..During the past long period, much progress has been realized on the study of theretrieval of LST based on remote sensing, and many algorithms for the LST retrievalhave been developed. However, in these algorithms non-isothermal mixed-pixels areall treated as an isothermal and homogeneous surface, which caused directly theresults that the LST obtained by these algorithms is only an average temperature ofmixed-pixel. In fact,, it is desirable to obtain surface component temperature withclear physical meanings in many LST application fields.The vegetation-soil component temperatures play important roles in manydisciplines like climate change research, surface process simulation, crop yieldestimation and draught monitoring.Because of considering that multi-angular data set is more beneficial to surfacecomponent temperature retrieval than multi-channel data set，people have devotedthemselves to study the radiant directionality model of surface non-isothermalmixed-pixels and have put forward several models,on which a serial of studies oncomponent temperature retrieval algorithm have been carried out.However, valid spaceborne multi-angular data set is infrequent．With a viewofthe necessity of surface component temperature retrieval and exiguity of validspaceborne multi-angular data set, it is undoubtedly of great practical significance tostudy surface component temperature retrieval algorithm based on multi-channel dataset.Compared with thermal infrared data set from Landsat/ETM+, NOAA/AVHRRand MODIS, thermal infrared data set from ASTER are much better in both spatialand spectral resolution, which has made ASTER advantages to retrieve componenttemperature on land surface. For this reason, this dissertation carried out the following study on component temperature based on ASTER thermal infrared data set, takingthe mixed pixel with light vegetation, shadows vegetation, light soil and shadow soilas the research object:(1) Using the MonteCarlo method to simulate vegetation and soil effectiveemissivity, and area ratio of light vegetation, shadows vegetation, light soil andshadow soil, then calculate effective emissivity of light vegetation, shadowsvegetation, light soil and shadow soil;(2) By taking the directionality model of surface component effective emissivity,adding wavelength variable λ, and treating observing angle θ constant, the surfaceradiant wavelength variety model that use multi-band thermal infrared data as datasources is defined;(3) A method is proposed to solve the surface radiant wavelength variety modelby using ASTER5TIR bands establish temperature inversion equation;(4) Taking wheat growing areas in FuYang, AnHui as the study area, surfacecomponent temperature is obtained by the above-mentioned method based on ASTERdata. In the same pixel, light soil temperature is the highest, light vegetationtemperature and shadow vegetation temperature take second and third place, Shadowsoil temperature is the lowest.