Analysis of Uncertainties and Associated Convective Processes in Simulations of Extreme Precipitation Over Cities With a Regional Earth System Model: A Case Study

Abstract This study utilizes the Weather Research and Forecasting model coupled with an atmospheric chemistry model, a multi‐layer urban canopy model (UCM), and a building energy model to simulate the extreme rainfall event influencing the Guangzhou city in South China on 7 May 2017. By employing small variations in the longwave emissivity of buildings within the UCM, 11 convective‐permitting experiments are conducted. The maximal 18‐hr and hourly rainfall accumulation vary from a 2‐year return period to as high as a 20 or 40‐year return period with notable spatial differences. Comparisons between the more accurate group (GOOD) and less accurate group (POOR) simulations highlight that some minor differences in the near‐surface air thermodynamic conditions in urban area could lead to substantial differences in local convection and its impacts on subsequent convective systems. With persistent transportation of warm, moist airflows from the northern South China Sea, formation of a slow‐moving mesoscale outflow boundary to the north of the urban agglomeration leads to the development of a quasi‐stationary, compactly structured meso‐γ‐scale rainstorm in the GOOD simulations. The stronger low‐level to near‐surface convergence and mid‐level cyclonic shear within this system substantially enhance low‐level updrafts, leading to increased microphysical production and stronger horizontal advection of rainwater within the system. These findings offer some process‐based understanding about the uncertainties in simulating urban extreme rainfall and underscore the need to develop ensemble forecasting methods for convection‐permitting numerical models that incorporate increasingly complicated representations of anthropogenic influences.