抄録
This paper presents a method to estimate the land surface temperature (LST) from Himawari-8 data. The Advanced Himawari Imager (AHI) onboard Himawari-8 has three thermal infrared bands in the spectral range of 10–12.5 μm. We developed a nonlinear three-band algorithm (NTB) that makes the best use of these bands to estimate the LST. The formula of the algorithm includes ten coefficients. The optimum values of these coefficients were derived using a statistical regression method from the simulated data, as obtained by a radiative-transfer model. The simulated data sets corresponding to a variety of values of LST, surface emissivity, type and season of temperature and water vapor profiles. Viewing zenith angles (VZAs) from 0° to 60° were considered. For the coefficients obtained in this way, we checked the root-mean-square error (RMSE) in terms of the VZA, LST and precipitable water dependence. We showed that the NTB can accurately estimate the LST with the RMSE less than 0.9 K by comparison with the nonlinear split-window algorithm developed by Sobrino and Romaguera (2004). Moreover, we evaluated the sensitivities of the LST algorithms to the uncertainties in input data by using the dataset independent of dataset used to obtain coefficients. Consequently, we showed that the NTB has the highest robustness against the uncertainties in input data. Finally, the stepwise LST retrieval method were constructed. This method includes a simple cloud mask procedure and the LSE estimation. The LST product was evaluated using in-situ data over the Tibetan Plateau, and then the validity was confirmed.
