Synthesis of a global daily average gapless land surface temperature dataset from 2003 to 2018.

Land Surface Temperature (LST) plays a crucial role in research in the fields of energy balance, hydrology, meteorology, geography, and ecology, serving as a significant input indicator of widespread interest. Remote sensing LST is one of the most widely used and applied methods. However, due to factors such as cloud, remote sensing LST data often exhibit a significant number of missing values, which can pose challenges to its application. This study proposes a gapless LST product with a spatial resolution of one kilometer and a temporal resolution of one day, generated by integrating two remote sensing datasets (1 km daily mean land surface temperature (DMLST) dataset by Zhan et al. and 1 km DMLST dataset by Liu et al.) and two reanalysis datasets (GLDAS and ERA5-Land). The validation results show that the Root Mean Square Error (RMSE) of the synthesized product is about 2 K, which is better than the other four products, the accuracy of RMSE has increased by 0.4-0.6 K, and it have shown significant improvements in arid and cold regions.