Global storm-resolving models (1–5km) provide new opportunities for explicitly simulating mesoscale convective processes, yet deep convection remains partly unresolved within the convective gray zone. Using global kilometer-scale simulations from the Digital Earth Global Hackathon 2025, this study develops a mesoscale vortex–MCS tracking framework to investigates the impacts of two deep-convection parameterization strategies, deep convection switched off (IFS-deepoff) and reduced cloud-base mass flux (IFS-rcbmf), on mesoscale convective systems (MCSs), mesoscale vortices, and precipitation during the 2020 Meiyu period over East Asia.  Systematic comparisons reveal that mesoscale vortex–convection coupling strongly modulates regional precipitation biases: concurrent overestimation of the occurrence frequency of vortices and MCSs in the Sichuan Basin enhances low-level convergence and upward motion, leading to severe precipitation overestimation, whereas underestimated vortex activity over the middle–lower Yangtze River (MLYR) suppresses convective organization despite abundant MCSs, contributing to dry biases. Stronger vortex circulation in IFS-deepoff promotes low-level moisture convergence and more rapid erosion of convective inhibition (CIN), facilitating convective initiation and organization. In contrast, weaker mesoscale vortex activity in IFS-rcbmf suppresses vortex–convection coupling and limits the development of precipitation extremes. These findings demonstrate that kilometer-scale global models remain fundamentally sensitive to subgrid convective assumptions through their regulation of vortex–convection coupling and multiscale convective organization. Accurately representing this coupling is essential for improving extreme rainfall prediction in the convective gray zone.

Global kilometer-scale models,Cumulus Convection Parameterization,Mesoscale vortex,Mesoscale convective systems,vortex–convection coupling,convective gray zone