Abstract Rapid snowmelt process, including rain‐on‐snow (ROS) events, can trigger flooding and related hazards in high mountains, yet snowmelt characteristics across the Southeast Tibetan Plateau (SETP) remain underexplored because of data scarcity. We utilized optimized 4 km convection‐permitting Weather Research and Forecasting (WRF) dynamical downscaling simulations, driven by ERA5 reanalysis, to investigate the variability of ROS water available for runoff (SM ROS , snowmelt plus rainfall on ROS days) over the SETP from 1995 to 2024. WRF reproduced daily mean air temperature, precipitation, and snow depth with reasonable accuracy (RMSE: 1.98°C/day, 1.32 mm/day, and 0.35 cm/day, respectively). Monsoon‐season accumulated SM ROS reached 623.84 mm (64.3% of the annual 970.91 mm), while extreme SM ROS days averaged 4.75 (85.59% of the annual 5.55 days), and activity was concentrated at 4,500–6,000 m. Accumulated SM ROS peaked in May, preceding the peak in extreme SM ROS frequency in June. The long‐term annual decline was driven primarily by monsoon‐season changes. Specifically, monsoon‐season accumulated SM ROS and extreme SM ROS frequency decreased by 44.88 mm/decade and 0.54 days/decade, compared with annual trends of 48.07 mm/decade and 0.52 days/decade, respectively ( p < 0.05). Net moisture supply during the monsoon season decreased by 1.78 × 10 6 kg s −1 /decade from weaker southern inflow and stronger northward export, contributing to declines in precipitation (−70.56 mm/decade) and rainfall (−60.77 mm/decade; p < 0.05). Meanwhile, near‐surface warming (0.29°C/decade, p < 0.01) reduced the snowfall fraction and snowmelt. These changes drove significant declines in the accumulated SM ROS amount and extreme SM ROS frequency, highlighting joint moisture‐supply and warming controls on regional ROS‐related runoff generation.
Yang et al. (Mon,) studied this question.