No abstract

Understanding the processes of rainfall extremes and their response to anthropogenic climate change is pivotal for improved adaptation of unprecedented flood hazards around the world. Here we take the record‐breaking 20 July 2021 storm over central China as an example. We investigate the response of this particular storm to atmospheric warming (i.e., increase in air temperature) and wetting (i.e., increase in atmospheric moisture content) based on a series of convection‐permitting model simulations. Our results show non‐monotonic changes of the space‐time rainfall variability to either increased temperature or atmospheric moisture content. The most extreme rain rate is produced when relative humidity is increased by 20%–40% or temperature is increased by less than 2°C. The non‐monotonic rainfall response is more clearly revealed at fine spatial (100–1,000 km2) and temporal scales (less than 6 hr) rather than over the entire domain (∼104 km2) and aggregated over the storm duration (around 2 days). This is mainly attributable to the distinct feedbacks from atmospheric dynamics (i.e., moisture convergence and interaction with regional topography) rather than regulated by thermodynamic changes alone. Atmospheric warming poses notable changes in the vertical structure of storm cells, contributing to reduced areal reduction factors at small spatial scales and short durations, while atmospheric wetting additionally modifies storm evolution properties. Our modeling analyses challenge the existing practices for hydrologic designs under a changing climate, highlighting particular vulnerability for cities or small basins to short‐duration rainfall extremes and the resultant flash flood hazards.