Abstract
Land surface temperature (LST) is a key variable governing land–atmosphere energy and water exchanges. Despite the importance of LST, satellite observations and reanalysis products often differ in how they define the effective LST depth and in the assumptions underlying their estimates, making comparisons and interpretation challenging. In this study, we present a detailed comparison of LST from GOES-16 (satellite) and MERRA-2 (reanalysis) across the contiguous United States for 2022 and 2023. The results reveal systematic diurnal and seasonal differences: GOES-16 tends to be warmer than MERRA-2 in the afternoon and at night, but cooler in the morning. The magnitude of these differences varies by season. At night, GOES-16 is warmest relative to MERRA-2 for forests; in the morning, it is coolest for croplands and grasslands; and in the afternoon, it is warmest for barren and shrublands. Within individual land cover types, variability in surface conditions—such as soil moisture and elevation—modulates the differences at night and in the morning, with GOES-16 LST being warmer at night and cooler in the morning for wetter soils and at higher elevations. Our analysis also indicates that Leaf Area Index plays a role during spring and autumn, likely due to the association of temperature with leaf emergence and senescence. These findings provide new insights into the mechanisms underlying LST differences between these datasets, and highlight the importance of accounting for surface condition variability when developing LST fusion and assimilation workflows.