干旱区生态环境知识资源中心

A multiregression technique was developed for deriving infrared land surface emissivity from a set of microwave emissivities. By removing the atmospheric contribution to Special Sensor Microwave Imager/Sounder (SSMIS) observations, the microwave land surface emissivity can be obtained by using the SSMIS window channel signatures and the surface temperature. High correlation between the Moderate Resolution Imaging Spectroradiometer (MODIS) land surface infrared emissivity product and the set of microwave emissivities was found. The experimental results show that single-channel microwave-infrared regressions do not work but multiple microwave channels can describe much of the infrared variability. The derived infrared land surface emissivity agrees with the MODIS land surface emissivity in general. The method can be used to compute infrared land surface emission at both clear and very cloudy conditions. Large variability in the infrared surface emissivity is found over desert regions. The change in the infrared emissivity over the Saharan, Mongolian and Australian deserts is due to the different materials of the deserts.