Land Surface Temperature Retrieval From AMSR-E Data

Land surface temperature is a key parameter of surface physical process and plays an important significance to the energy balance in the world.Since the end of 70's in 20 centuries, the domestic and international scholars have made a great deal of research on how to get land surface temperature quickly by thermal infrared remote sensing .In theory, thermal infrared remote sensing can provide accurately high spatial resolution of land surface radiation and temprature information. However, thermal infrared image is influenced easily by atmospheric water vapor, and it can't get through cloud. Microwave remote sensing is different from thermal infrared remote sensing.It is influenced little by atmosphere, that can get through cloud. Microwave remote sensing is particular with different polarization and working mode in all day. Therefore, it is superior to the earth's surface temperature retrieval and the soil humidity measurement.Along with microwave radiometers SMMR and SSM/I, AMSR-E is aboard on different satillites. And the inversion of surface temperature from passive microwave remote sensing is paid more and more many attention. But the microwave data resolution is limit. In this study, the traditional empirical statistics method is chosed to retrieve land surface temperature.Because the spatial resolution of the passive microwave data resolution is very low and the surface situation is very complex, it is extremely difficult to measure the surface temperature synchronally. Thus, land surface temperature (LST) products of MODIS are used to replace the AMSR-E surface measured data, which is onboard the same satellite to AMSR-E. the relationships between the LSTs and Bright Temperature (BT) at different frequency ofAMSR-E are first analyzed. Meanwhile, different types of land cover, such as the Tibet area (plateau frozen earth), Jilin the western area (saline land), the Inner Mongolian area (prairie), the Changbaishang mountain area (forest vegetation), the Lake Qinghai area (water body), the tower carat Masurium area (desert), and so on, are taken as examples. The results shows that the relationships are different for different land cover. Thus, the equations of land surface temperature at different land cover should be different.Based on the relationship between LST and BT at different frequency of AMSR-E, their linear relationships are built respectively by regression analysis for different land cover. Finally, the West of Jilin Province is taken to be a experimental area, where the land cover is classified by the bright temperature data of AMSR-E. For different land cover, the bright temperature data of AMSR-E is changed to land surface temperature by the paramenters of the linear equations between the LST and BT of AMSR-E at different land cover. And the retrieval land surface temperature from AMSR-E in the west of Jilin Province are thus mapped.