The Research On Numerical Simulation Of A Warm Region Rainfall Occurring In South China

The torrential rain that occurs during the pre-summer flood season in South China has attracted increasing attention.Such heavy rainfall in the warm region is especially difficult to resolve in numerical and operational forecasts because of its abrupt and particular characteristics.Based on observational and ECMWF_FINE data,a diagnostic analysis was undertaken on the warm region rainstorm that occurred in South China on May 19–20,2015.Subsequently,using the GRAPES and WRF models,simulations of the rainstorm were performed to study the models' capability in predicting rainstorms in the warm region.To evaluate the forecasting skill of the GRAPES model,the 1D+3D var assimilation was applied to obtain a realistic water vapor profile and to produce better initial conditions for the high-resolution model.Furthermore,the impact of the 1D+3D var assimilation of rain rate upon the initial conditions and short-term precipitation forecast was estimated.The “5.19” warm region rainstorm showed characteristics of precipitation concentration,strong intensity,and localization.The spatiotemporal distribution of precipitation showed obvious daily variation and pronounced mesoscale structural features.Compared with the GRAPES model,the WRF model produced better prediction results,especially regarding the center of heavy rain in the coastal area of East Guangdong.The WRF model was good at predicting middle–low-level wind disturbances and surface temperature,which are very important in triggering mesoscale convective systems.The matching of vertical motion and the distribution of water vapor makes the area and magnitude of the precipitation more consistent with the observations.Conversely,the precipitation results of the GRAPES model were remarkably overestimated.The output of the GRAPES model had a strong southern jet and largescale vertical motion.In addition,the GRAPES model demonstrated poor skill in simulating both the convective motion over the ocean and the distribution of low-level temperature.These findings might offer guidance regarding the improvement of the GRAPES model's capability for predicting warm region rainstorms.Based on the diagnostic analysis and numerical simulations,we concluded that the low-level shear line is the important driving force for convective systems,and the intensity and location of the southern jet determines the distribution of precipitation.However,factors such as rich moisture content,strong environmental atmospheric thermodynamic instability,the complex boundary-layer triggering mechanism,land–sea thermal difference,and the external forcing effect of particular terrain all combine to make it difficult to simulate and forecast this type of warm region rainfall.Based on the GRAPES₃DVAR system,we estimated the impact of the 1D+3D var assimilation of rain rate.Compared with the control test,the distribution of initial specific humidity was modified reasonably by adjusting the water vapor profile,while the A index and the K index indicated heavy rain.Thus,convective instability in the area of heavy rain increased.The assimilation of reflectivity data adjusted the wind field during the integration procedure and it strengthened the low-level convergence markedly.Simulation experiments indicated that the assimilation of reflectivity data could significantly improve nowcasting and short-term precipitation forecasting skills,the TS scores of which were also obviously improved.However,the impacts of assimilating reflectivity data are limited to within six hours,and a disadvantage is that the rainfall intensity is slightly overestimated.In addition,the results showed that the experiments achieved improved simulation of hourly rainfall intensity and location,and that the development and decay of the mesoscale convective system was clearer.