Sensitivity of Microphysical Parameterization on the Numerical Simulation of a Meiyu Front Heavy Rainfall Process

Microphysical processes in the Meiyu front rainfall have an important effect on the evolution of precipitation. Based on the WRF (version 3.4.1) model, one Meiyu front heavy rainfall case from 29 to 30 June is analyzed with three different microphysics schemes (Morrison, Thompson, and MY). The main findings are as follows. (1) The general large-scale circulation of the Meiyu rainfall case could be reasonably reproduced by all three experiments with different microphysics schemes, which was consistent with the ERA5 reanalysis data. The local circulation during the Meiyu front heavy rainfall was significantly influenced by microphysical processes and the differences in the local features between different experiments were evident. The local circulation and updraft in the Thompson experiment were stronger than those in the other two schemes. Precipitation in all the model output was overestimated and the hourly rain rate was always greater. The overestimation of the melting of ice phase hydrometeors or the accretion of cloud droplets by raindrops was one of the most important causes of the overestimation of the modelled precipitation. On the whole, the Morrison run performed relatively better. (2) The melting of ice phase hydrometeors and the accretion of cloud droplets by the raindrop were the key source terms to the growth of the raindrop. Moreover, the evaporation process was the most important sink term. On the whole, the raindrop collecting cloud droplet contributed more than the melting of ice phase hydrometeors to the growth of the raindrop. However, for each scheme, the differences of these microphysical process terms led to the difference of the modelled precipitation in distribution. (3) The Thompson run produced the largest amount of melting of ice phase hydrometeors and evaporation (especially in the low level). At the same time, it produced the largest amount of condensation that led to more collection of cloud droplets by raindrops. Therefore, the Thompson run produced the most raindrop and rainfall. The predominant product through the deposition and riming process was snow, and the largest amount of snow was produced. (4) Through the Bergeron process, the Morrison run produced more snow than graupel (ice particles nearly could be neglected), the Thompson run produced predominant snow, and the MY run produced more snow than ice particles (graupel nearly could be neglected). The largest amount of produced ice particles in the MY run through the process led to more ice particles than that in other schemes. (5) The cloud droplet contributed more than the raindrop in the riming process. In the Morrison and Thompson schemes, the amount of graupel collecting cloud droplet was larger than that through other riming processes. Other riming processes contributed to the growth of graupel in different degrees in the Morrison run, while other riming processes could nearly be neglected compared to the graupel collecting cloud droplet. Further, the MY run produced a larger amount of snow growth by deposition. Therefore, the differences of the Bergeron and riming processes in all three schemes led to the differences in the ice phase hydrometeors distribution.