Dynamic Downscaling Simulation and Projection of Precipitation Extremes over China under a Shared Socioeconomic Pathway Scenario

In this study, Weather Research and Forecasting (WRF) -based dynamic downscaling simulation was performed over China at a horizontal resolution of 25 km. To reduce the systematic large-scale biases from lateral boundary conditions, a dynamic blending (DB) technique was introduced in downscaling, and its performance was compared with downscaling without DB (NO_DB) as well as the driving global climate model (GCM, i.e., HadGEM3). In the present-day simulation for verification, added values were found in DB, relative to the GCM and NO_DB, in simulating the precipitation extremes, especially over southeastern China. Possible causes responsible for this improvement were further analyzed. In the GCM, excessive moisture and atmospheric heating in the upper troposphere in conjunction with abnormally strong deep convection resulted in excessive extreme precipitation over southeastern China, while in NO_DB, insufficient moisture and atmospheric heating in the whole troposphere in conjunction with abnormally weak convection suppressed extreme precipitation there. In comparison, DB showed a closer representation of observations in terms of vertical velocity and vertical profiles of atmospheric moisture and heating, accounting for the improved simulation of extreme precipitation. In the future projection under the SSP5-8.5 scenario, both the GCM and DB predicted that extreme precipitation would increase in most parts of China; however, DB indicated a larger increase over southeastern China relative to the GCM. Stronger moisture flux convergence over southeastern China in DB accounted for this larger increase, and in addition, the thermodynamic effect associated with more precipitable water dominated the stronger moisture flux convergence.