Deep convective storms can overshoot the tropopause, thus altering the composition of the stratosphere by vertically transporting tropospheric air. The transport of water vapor and ice particles into a sub-saturated environment can hydrate the stratosphere, with implications for radiative forcing and ozone chemistry. Cloud-resolved models, if employed at high spatial resolutions, are used to probe process-level questions about cross-tropopause deep convective hydration and its controls. There is considerable diversity in model representations of processes associated with water transport and transformation, and the choice of a microphysics scheme affects model skill in simulating deep convective events. This motivates our evaluation of state-of-the-art, as well as widely used standard schemes, in a high spatial- and temporal-resolution framework. Six bulk microphysics schemes were employed in a WRF-LES setup, initialized with a sounding profile representative of a tropopause-overshooting storm. We used an idealized framework to isolate the effect of microphysics on the dominant processes that control the reach of deep convection and stratospheric hydration. All schemes produced the highest reaching updrafts 8–12 hr into the simulation but the strength and persistence of updrafts varied across the schemes; maximum storm heights ranged 9.1–12.6 km across the schemes. Varying microphysics produced large differences in the vertical extent and horizontal aggregation of convection, and an order of magnitude spread in above-tropopause water vapor concentrations. Plain Language Summary Strong thunderstorms often result in deep convection, by generating vertical winds that transport water and air from the troposphere into the stratosphere. The resulting moistening of the stratosphere can impact the radiative budget in that region and impact ozone chemistry. Satellite and aircraft based measurements have provided evidence of increased moisture in the lower stratosphere due to deep convective storms but there remain many gaps in our process level understanding of such observations. Informative simulations of convective storms require an understanding of how the dynamic representation of transformations of water in vapor, liquid and frozen forms in cloud-resolving models affects model skill in generating convection. This study uses an idealized setup to investigate the performance of six different microphysical schemes and finds that microphysics mediate large differences in storm dynamics and associated stratospheric moistening.