Research on retrieving land surface temperature from MODIS thermal infrared data

Abstract

In Earth Observing System (EOS) plan, MODerate-resolution Imaging Spectroradiometer (MODIS) is loaded on both the two polar-orbit satellites: Terra (EOS-AM1) and Aqua (EOS-PM1). MODIS data has 16 Thermal InfraRed (TIR) channels (3.5~14.5 μm) among all of its 36 channels. Land Surface Temperature (LST) is an important indicator of earth surface energy balances and climate changes, as well as a key parameter in physical processes of land surface on both global and regional scale. LST is widely applied in the research of disciplines such as meteorology, hydrology, ecology, biochemistry, etc1. Therefore, retrieving LST from appropriate MODIS TIR bands is one of the important applications. First of all, this paper introduces theoretical foundations of retrieving LST from remote sensing data, such as the method of selecting appropriate TIR bands by conditional analysis of atmospheric window. Then, this paper provides an overview of LST retrieving algorithms up to now, including Single Window Algorithm, Split Window Algorithm, Improved Split Window Algorithm, Generalized Split Window Algorithm and Day/Night Algorithm. And at last, towards to the limitations of LST retrieving algorithm, the authors indicates their specific perspectives on the directions of further correlative research in two aspects: improving LST retrieving algorithm and increasing LST retrieving accuracy.