Many studies have indicated that convective bursts (CBs) are closely related to tropical cyclone (TC) intensification, but few studies have been conducted on the mechanisms that control the formation and evolution of CBs. In this study, the 1‐min output data of a simulated TC are used to understand the convective extreme updrafts of CBs in the TC eyewall. Three different reference states including the local square‐area mean (LAM), the lower wavenumber‐components mean (WNM), and the arc‐area mean (ArcM) are used to calculate the driving forces of convective extreme updrafts. Contrary to the WNM and ArcM reference states, the LAM reference state struggles to capture realistic basic state structures. The LAM reference state can be modified through the inclusion of the mean hydrometeor mixing ratio in the basic state to produce physically consistent forcing in relation to the simulated convective extreme updrafts. The simulated convective extreme updrafts in the TC eyewall exhibit two peaks at middle and upper levels, respectively, since the effect of hydrometeor loading, decelerates the air parcels between the updraft maxima. While the positive buoyancy makes air parcels in the CBs accelerate at middle levels, in agreement with previous studies, it is found that, at the upper levels, both the positive buoyancy and the upward vertical perturbation pressure gradient force accelerate the air parcels. This study suggests that the vertical perturbation pressure gradient force also plays an important role in the formation of CBs in the TC eyewall.

Significant variability of raindrop size distributions (DSDs) has been observed in the “21·7” Henan extremely heavy rainfall event (the “21·7” Henan EHR event), while the capability of model to reproduce such complicated heavy rainfall DSDs is yet unclear. This study primarily evaluates the simulated DSDs of multiple microphysics schemes by comparing with the observations from a network of 50 disdrometers. Constrained DSD variability is identified in most schemes that the simulated raindrop mean sizes are gradually restricted around specific values as the growth of heavy rainfall intensity. The schemes are also incapable of reproducing the different raindrop mean sizes from deep convection and shallow convection. Moreover, simulations show unrealistic evolutions of raindrop mean size standard deviations as height declines. By investigating the empirical formula and performing sensitivity experiment, the constrained DSD variability in heavy rainfall is largely blamed on the insufficient parameterizations of the self‐collection (breakup) processes.