Land surface thermal characterization of Asian-pacific region with Japanese geostationary satellite

Land Surface Temperature (LST) is a significant indicator of energy balance at the Earth's surface. It is required for a wide variety of climate, hydrological, ecological, and biogeochemical studies. Although LST is highly variable both temporally and spatially, it is impossible for polar-orbiting satellite to detect hourly changes in LST, because the satellite is able to only collect data of the same area at most twice a day. On the other hand, geostationary satellite is able to collect hourly data and has a possibility to monitor hourly changes in LST, therefore hourly measurements of geostationary satellite enables us to characterize detailed thermal conditions of the Earth's surface and improve our understanding of the surface energy balance. Multi-functional Transport Satellite (MTSAT) is a Japanese geostationary satellite launched in 2005 and covers Asia-Pacific region. MTSAT provides hourly data with 5 bands including two thermal infrared (TIR) bands in the 10.5-12.5 micron region. In this research, we have developed a methodology to retrieve hourly LST from thermal infrared data of MTSAT. We applied Generalized Split-window (GSW) equation to estimate LST from TIR data. First, the brightness temperatures measured at sensor on MTSAT was simulated by radiative transfer code (MODTRAN), and the numerical coefficients of GSW equation were optimized based on the simulation results with non-linear minimization algorithm. The standard deviation of derived GSW equation was less than or equal to 1.09K in the case of viewing zenith angle lower than 40 degree and 1.73K in 60 degree. Then, spatial distributions of LST have been mapped optimized GSW equation with brightness temperatures of MTSAT IR1 and IR2 and emissivity map from MODIS product. Finally, these maps were validated with MODIS LST product (MOD11A1) over four Asian-pacific regions such as Bangkok, Tokyo, UlanBator and Jakarta , It is found that RMSE of these regions were 4.57K, 2.22K, 2.71K and 3.92K, respectively. Large RMSEs of Bangkok and Jakarta in the tropical zone can be result from unsuitable parameters used in the MODTRAN simulations and remained haze or cirrus cloud. However, comparison between MTSAT LST and MODIS LST showed linearity and consistency, therefore MTSAT LST contribute to a better understanding in a wide variety of the surface energy balance research. Finally, thermal characterizations such as the rate of LST change or diurnal LST range and so on have been mapped by using constructed MTSAT LST database.