A Retrieval Algorithm for Passive Microwave-Based Land Surface Temperature Considering Spatiotemporal Soil Moisture and Land Scenarios

Land surface temperature (LST) is an important parameter for the study related to land–air coupled systems. Satellite-based passive microwave (PMW) sensor is a significant approach to retrieving LSTs, which can penetrate the atmosphere conditions. However, soil moisture is one of the variables affecting PMW brightness temperature, yet few existing methods have considered it in the procedure of retrieving LSTs. On this basis, we propose a retrieval algorithm for PMW-based LST that comprehensively considers soil moisture changes and land scenarios (RA-PLST-SM) to take the soil moisture into account. Additionally, two fusion strategies one of which fuses the landform and soil moisture, and the other further fusing with land cover are proposed to construct the land environment description. Besides, we also propose a fusion strategy to integrate the LSTs from the time-interval-based models of month, quarter, and year. The mean root-mean-squared error (RMSE) referring to MODIS LSTs shows that RA-PLST-SM LSTs gain 3.13 and 2.32 K at day/night time. Also, referring to advanced LST data product, RA-PLST-SM LSTs present a yearly mean STD of 4.12 and 2.48 K at day/night time. Furthermore, the metric results according to six in situ stations demonstrate that RA-PLST-SM LST is significantly enhancing the consistency of actual LSTs, obtaining a mean RMSE of 4.42 K/3.48 K at day/night time. Moreover, the LST fusion strategy proposed in this article can effectively improve the quality of results. The progress of this article can promote the research fields to acquire cloudy LSTs and even provide the technique reference for PMW-based LST retrieval.