LAND SURFACE TEMPERATURE ESTIMATED BY REMOTE SENSING DATA FROM FY-1D POLAR METEOROLOGICAL SATELLITE

The split window method was widely used to retrieve the land surface temperature (LS T) from satellite remote sensing data in the 1980s and the 1990s.The scientists demonstrated theoretically or statistically that there exists a linear or non-linear relationship between the LST and AVHRR channel 4, 5 brightness temperatures.2818 clear sky radiosonde profiles were used to simulate the FY-1D HRPT channel 4, 5 brightness temperatures for different ground emissivities.The derived no n-linear regression equation for each ground type is :T_LS=T₄ +A(T₄-T₅)+B(T₄-T₅)² +D.Meanwhile, the calculating method of LST by using FY-1D HRPT data and the regression equation were described.Finally, the accuracy of the calculated LST in main land of China was given :compared with the 0 cm soil temperatures observed by Chinese meteorological ground stations, the LSTs calculated by FY-1D data are very co-ordinate with the station observations.For most ground stations, the deviations between the satellite LST and stations observations are less than 3.0 K.