The remnant of Typhoon Haikui (2311) caused extreme precipitation in the Guangdong-Hong Kong-Macao Greater Bay Area, breaking multiple precipitation records in Shenzhen and Hong Kong. In this paper, three microphysical schemes, LIUMA, WSM6, and THOMPSON, are selected for simulation based on CMA-MESO V6.0. The forecast performance, errors in the prediction of various meteorological elements, microphysical variables, and their changes are compared across the different schemes. Additionally, latent heat sensitivity tests are conducted for further analysis. The LIUMA run shows a slightly higher threat score (TS) but exhibits higher Bias than other 2 runs. In the region experiencing heavy precipitation, all three runs showed peaks of significant snow and graupel totals near the 0°C layer, and peaks of rainwater near 3 km. The contribution of ice-phase particles varied among the schemes, with WSM6 contributing mostly graupel, and THOMPSON and LIUMA mostly snow. The LIUMA run demonstrated a significantly larger net latent heat release near the 0°C layer compared to the WSM6 and THOMPSON runs, and also forecasted stronger warm core structures and the remnant of a Landfalling Tropical Cyclone (LTC). All three runs predicted a warm-core structure located to the east of the low-level center of the LTC remnant, which slowed the westward movement of the remnant and facilitated the maintenance of heavy precipitation. Compared to the WSM6 and THOMPSON experiments, the LIUMA experiment releases significantly more latent heat near the 0°C layer and forecasts a stronger warm-core structure and LTC remnants, leading to more false alarms. Latent heat sensitivity tests revealed the importance of latent heat feedback from microphysical processes, indicating that a decrease in latent heat could lead to a rapid westward weakening of the remnant of LTC and an increase in forecast error. The model also exhibited significant uncertainty in its forecasts of extreme precipitation, which could impact the overall forecast field through latent heat release.
Chen et al. (Wed,) studied this question.