Abstract
With the advantage of anaerobic soil carbon storage, wetlands are widely recognized as potential nature‐based solutions for climate mitigation. While their role in the carbon–climate system has been extensively studied from a biogeochemical perspective, their biophysical effects on land surface temperature (Ts) remain poorly understood. Here, we investigate the biophysical mechanisms of Ts regulation by wetlands by integrating physically derived variables with machine learning. Globally, 76% of wetlands exhibit an annual cooling effect relative to adjacent forest ecosystems, whereas 34% show net warming. Latitudinally, boreal wetlands exhibit the strongest cooling. In the mid‐latitudes, wetlands display a distinct diurnal pattern, characterized by daytime cooling and nighttime warming. In tropical regions, wetlands tend to exert cooling in spring and warming in summer, with a net annual cooling effect. In the northern hemisphere and during spring, wetland Ts is predominantly controlled by albedo, whereas in the Southern Hemisphere and during summer, variations in Ts are primarily driven by evapotranspiration and aerodynamic resistance. The findings provide deeper insight into the complex interactions between wetlands and the climate system. They also enhance the accurate characterization of the climate regulation functions of global ecosystems, providing a scientific basis for the development of future nature‐based climate solutions.