Enhancing summer atmospheric water cycle simulations in the Three-River Headwaters Region via dynamical downscaling

Characterizing and understanding the evolving water cycle in the Three-River Headwaters (TRH) region of the Tibetan Plateau, where data scarcity poses significant challenges to climate and atmospheric water cycle research, heavily depends on model simulations. However, advanced global climate models and reanalysis datasets frequently overestimate precipitation. To address this, we employ high-resolution (9 km) regional climate simulations (WRF9km) to examine atmospheric water cycle variables in the TRH region, comparing them with in-situ observations and ERA5 reanalysis. Our study demonstrates that WRF9km substantially reduces the overestimation of summer precipitation (by 24.0 %) and evapotranspiration (by 52.7 %) compared to ERA5, thereby improving its alignment with observational data. The reduced biases in precipitation are attributed to diminished moisture influx from the southern boundary and local evapotranspiration, coupled with increased moisture export from the eastern boundary. Summer precipitation recycling (PR) accounts for approximately 20 % of total precipitation in the TRH region. Despite divergent trends in PR between a water accounting model (WAM) and a bulk method, our findings support the reliability of WAM, indicating a slight decrease in summer PR (−0.4 %/10a for ERA5 and − 0.6 %/10a for WRF9km). While WRF9km accurately captures the spatial pattern of summer PR, ERA5 appears to overestimate it, likely due to biases in evapotranspiration and moisture inflow. In conclusion, WRF9km provides a more accurate representation of the atmospheric water cycle in the TRH region.