Land–atmosphere coupling (LAC) is a critical process in the climate system, and has great effects on surface air temperature over the humid–dry transition zones. Using ERA5/ERA5‐land reanalysis datasets, we show that the summer temperature anomalies over Northwest China (NWC), known as one of the most important LAC hotpots in China, are significantly correlated with the LAC strength during 1981–2020. The stronger the coupling strength between air temperature and land surface, the more pronounced decrease in air temperature and extreme hot days over NWC. Among the four pathways of land–air temperature coupling, including soil moisture (SM)/soil temperature–latent heat flux/sensible heat flux–air temperature coupling, SM–sensible heat flux–air temperature pathway corresponds most closely with the temperature changes in summer. By analyzing the possible local and non‐local physical links in this relationship, we show that the wave trains composed of the intensification of Ural ridge and Central–Asia trough stimulate a meridional fluctuation of westerly airflow over the mid–high latitudes of Eurasia. Large moisture, therefore, converges into NWC, leading to excessive precipitation and deficient incident solar radiation reaching the land surface. In such case, wetter and cooler soil states accelerate the cooling of the environment through feedbacks between the lands, land–atmosphere exchanging fluxes and air temperature locally, accompanied by a significant enhancement in LAC strength. Our study provides insight into the physical processes causing variations in summer temperature over a typical LAC hotspot and suggests that the role of land surface–atmosphere interactions needs to be considered when studying temperature extremes under climate change.

No abstract

The summer atmospheric river (AR) frequency over East Asia and Western North Pacific (EA-WNP) is investigated by multiple AR detection algorithms based on the Atmospheric River Tracking Method Intercomparison Project (ARTMIP) Tier2 reanalysis dataset. The results show that AR frequency during the recent four decades experienced opposite interdecadal shifts, greatly contributing to the interdecadal equatorward trends of EA ARs and poleward trends of WNP ARs with a boundary around 135°E. The opposite variations are mainly influenced by a zonal dipole of integrated water vapor transport with cyclonic and anticyclonic anomalies centered over Taiwan and the ocean to the southeast of Japan, respectively. A major impact of the Pacific Decadal Oscillation and a reinforcement effect due to a zonal wave train from the North Atlantic jointly modulate the pattern. Considering ARs may curve their pathway over EA-WNP, the algorithms based on historical AR shapes should be cautiously used during AR detection.