A new approach to reduce inconsistency between MODIS and ASTER land surface temperature products

Land surface temperature (LST) is of importance in controlling most physical, chemical, and biological processes of the Earth system. Satellite-derive LST provides large-scale observation and is very useful to environmental studies. Among numerous satellite sensors, the MODerate resolution Imaging Spectroradiometer (MODIS) and the Advanced Spaceborne Thermal Emission Reflection Radiometer (ASTER) are onboard the same satellite platform TERRA. MODIS MOD11_L2 and ASTER AST_08 LST products have a spatial resolution of 1-km and 90-m, respectively. Our previous scaling study revealed 2.2K on-average differences between the MODIS and the upscaled ASTER LST over a heterogeneous semiarid area in the Loess Plateau of China (SPIE Proc., 5967: 58670O-1-8). Because the retrieval algorithm for MODIS 1-km LST product is subject to uncertainty in emissivity estimate over semiarid and arid areas, this paper uses ASTER emissivity data to reduce the LST inconsistency. Based on the MODIS LST retrieval algorithm, a new algorithm is derived. The algorithm does not rely on the coefficients used in the MODIS algorithm such that it can be implemented without acquisition of the raw MODIS datasets. The MODIS LST and band-31 emissivity as well as the upscaled ASTER emissivity are the necessary inputs to the proposed algorithm. Using the new approach, the rectified MODIS LST achieved the satisfied agreement with the upscaled ASTER LST. This study also suggested that the uncertainty in LST induced by retrieval algorithm could be larger than the scale induced uncertainty.