An Evaluation of Convection‐Permitting Ensemble Simulations of Coastal Nocturnal Rainfall Over South China During the Early‐Summer Rainy Season

Abstract

AbstractConvection‐permitting ensemble Weather Research and Forecasting simulations were successively conducted on each day during the early‐summer rainy season (April‒June) in 2018 over South China. These simulations were configured by eight sets of parameterization schemes associated with the planetary boundary layer and surface layer. The performance of the ensemble products in reproducing the nocturnal rainfall in the coastal region of South China was evaluated using observations from densely distributed rain gauges and wind profiling radars. The simulations at a lead time of 12 hr present a relatively better performance than those at 36 hr lead time, but still exhibit a low skill especially after midnight. The degradation of forecast skill after midnight is closely associated with the worse underestimation of the low‐level onshore flow and land surface air temperature. The large‐scale onshore flow at low levels from the South China Sea (SCS) interacted with land/mountain breezes, enhancing the lower‐troposphere moisture pooling and convergence for the rainfall production over the coastal mountains/cities. In the longer lead‐time simulations, the low‐level onshore flow presented a deceleration tendency and thus indirectly amplified the effects of land/mountain breezes in nighttime, leading to more serious underestimation in rainfall amount and larger southward displacement error than the shorter lead‐time simulation. The findings highlight the dependence of the nocturnal rainfall forecast on accurately predicting the upstream low‐level onshore flow over the SCS and the thermal conditions in the urban region surrounded by mountains.