The Study Of Algorithms For Land Surface Temperature Retrieval From Feng Yun Satellite Data

Land surface temperature (LST) is a key factor in the physics of land surfaceprocesses on regional and global scales, playing an important role in the landand atmosphere interactions and energy exchange. Thermal infrared remote sensingdata can be used to retrieval LST continuous in space and time, which has greatapplication value in areas such as climate, agriculture, environment anddisaster monitoring. FY3A/MERSI and FY2/VISSR thermal infrared data have theadvantage of high spatial and temporal resolution, so their LST products wouldhave a good prospect of application. Because no algorithm can be used directly,it is necessary to study inversion algorithms based on the characteristics ofthese two sensor data.First, choose MOTRAN4to simulate the impact of atmospheric and geometricfactors on the atmospheric transmittance、upwelling and downwelling radianceof FY3A/MERSI and FY2/VISSR thermal infrared bands. The results indicate thatview zenith angle and water vapor column play major roles in clear sky conditions.And these two factors can be only corrected in LST retrieval.Second, fast downwelling radiance measurement in surface emissivitymeasurements, which is studied by simulation and verified by field measurements，can catch accurate instantaneous downwelling radiance to ensure the emissivitymeasurement accuracy. Sereval comparative analysis of experimental emissivitymeasured in western China show the reliability of the measurement results, therepresentativeness of the targets, the uniformity of the measuring surface, oneof which was stable in time. These measurement data are valuable on surfaceemissivity retrieval algorithm development and validation.Third, study single-channel algorithm for FY3A/MERSI sensor channel,including doing simulation to fit the coefficient of the atmospheric andobservation angle expressions, and using experimental data to develop andvalidate emissivity retrieval algorithm. Simulation verification shows thatretrieval accuracy improvement is extremely significant in the angle of greaterthan20°by adding the angle factor. Verification by comparison of star-earth data indicates that the MERSI retrieval LSTs have-1.985K bias in average withtemperature measured on surface.Fourth, select the split window algorithm considering the water vapor factorto inversion FY2F LST. Choose the fitting of water vapor content expression withhighest accuracy through the simulation experiment for the split windowalgorithm. Then fit the coefficients for observation angles to establish lookuptables. Verification experiments by simulation and remote sensing data indicatethat the algorithm of this paper has obvious advantages for its high precisionLSTs.