Retrieval Of Long Time Series Land Surface Temperature For China Territory With NOAA-AVHRR Satellite Data

Land surface temperature (LST) is one of the important parameters of the earth’ssurface energy balance. It plays a very important role in the process of interactionbetween the atmosphere and the earth’s surface. With the deepening of the researchon global change and sustainable development, LST is widely concerned because ofits key role. On the basis of the atmospheric radiative transfer model MODTRAN4,using NOAA-AVHRR cloud-free data, this thesis builds the LST database with longtime series in China territory. The work is listed as below:1) The thesis analyzed and compared the common retrieval algorithm of LST,including the typical single channel algorithms and multichannel algorithms,especially general split-window algorithm, day and night algorithm based on theinfrared and thermal infrared channel, the algorithm of separation of land surfacetemperature and emissivity.2) Based on the atmospheric radiative transfer theory, using NOAA-AVHRRcloud-free data, the land surface temperature retrieval algorithm, land surfaceemissivity estimation model and atmospheric water vapor content inversion modelapplicable to series of NOAA meteorological satellite were found:a) A simulated database including the thermal infra-red channels (Channel4with spectral range10.3-11.3μm and Channel5with spectral range11.5-12.5μm) radiance observed at the satellite level was established withthe atmospheric radiative transfer model MODTRAN.b) It covers various land surface types and atmospheric conditions. Then thecoefficients of the GSW were determined by grouping the LST, theatmospheric water vapor content (WVC), the land surface emissivity (LSE),and the viewing zenith angle (VZA)by several sub-ranges.c) Using ASTER spectrum library, combined with the classification of16kindsof IGBP surface coverage, based on NDVI threshold method, the surfaceemissivity estimation model was built.d) The WVC was estimated with the transmittance ratio method proposed by Liet.al (2003).3) The data preprocessing of NOAA meteorological satellites includes radiativecorrection and re-projection. According to the NOAA user guide, we realized theradiative calibration of remote sensing data. The LST, LSE and WVC were retrievedby proposed methods. 4) Combined with the Stefan-Boltzmann law, using the2010full-year fieldmeasured data from AmeriFlux station which was30minutes time scales, using LSEsprovided by MODIS, the validation of the retrieved LST was performed.5) The proposed algorithms were applied to remote sensing data. We realized theretrieval of LST, multiband image mosaic and tailoring by IDL programming. Finally,we generate the day cycle LST database with long time series in China territory.