<p>An accurate description of microphysical processes is essential for the prediction of severe convective weather. To better understand the microphysical processes of severe convective weather induced by the northeast cold vortex, observed and simulated microphysical characteristics were evaluated for two types of severe convective weather events (a short-term heavy precipitation event in June 2020 and a thunderstorm gale event in April 2021) based on observed polarimetric radar data. Each microphysical scheme could be used to suitably simulate the difference in the number of hydrometeor particles between the two cases. Thus, the simulation characteristics of each scheme, as reflected by the radar reflectivity, were relatively similar. However, significant differences were observed in the polarimetric radar variables. For both events, the Weather Research and Forecasting (WRF) double-moment 6-class (WDM6) scheme produced significantly lower differential reflectivity (Z_DR) values than the observed values, whereas the WRF single-moment 6-class (WSM6) scheme yielded Z_DR values that were systematically higher. The Milbrandt–Yau (MY) and Morrison schemes produced excessively high Z_DR values at lower levels. The Thompson scheme yielded Z_DR values that were the closest to the observed values in both cases, particularly for vertical structural features. The Z_DR–Z_H probability distributions of the Thompson scheme were more consistent with the observed physical patterns. The Thompson scheme best captured the differences in wet and dry environmental conditions between the two convective weather events. Improvement in terms of high-level hydrometeors is likely to be an important factor in the performance of the Thompson scheme.</p>