Near-Real-Time Estimation of 1-km All-Weather Land Surface Temperature by Integrating Satellite Passive Microwave and Thermal Infrared Observations

A widely used approach for all-weather land surface temperature (LST) estimation is the integration of satellite passive microwave (MW) and thermal infrared (TIR) remote sensing observations. However, there are still few methods for estimating near-real time (NRT) all-weather (AW) (NRT-AW) LST. Besides, estimation of the LST within the swath gap of the satellite MW images is still greatly limited. This letter proposes a so-called NRT-AW method for the estimation of NRT-AW LST. NRT-AW firstly fills up the brightness temperatures (BT) inside the AMSR2 swath gap. Then, the NRT AW LST is estimated by learning the mapping between the time series of AMSR2 BT and MODIS LST on the annual scales. The results of the application of NRT-AW in the Heihe River Basin (HRB) show that the NRT-AW LST is spatially continuous and highly consistent with the original MODIS LST with a standard deviation (STD) of 1.27-1.77 K. Validation based on in situ LST indicates that the NRT-AW LST estimate has a root mean square error (RMSE) of 2.46-4.62 K. This method is beneficial for rapid mapping of all-weather LST over large areas and, thus, can satisfy associated applications.