Abstract
The single-column model was used to evaluate candidate convection parameterizations in multiple IOP cases. Sensitivity simulations were also conducted to further investigate the impact of various details in the experimental design, e.g., specification of boundary conditions, nudging, and periodic reinitialization.
The analysis focused on the cloud and precipitation statistics, diurnal cycle, frequency distributions of convective events and precipitation fluxes, and the mass/number budget of cloud hydrometeors. In terms of precipitation statistics, all parameterizations did a reasonably good job, and no single parameterization outperformed the others in all test cases. However, the simulated cloud properties differed remarkably among the parameterizations. For example, UNICON simulated higher liquid water content (LWC) than the other schemes when convection was more active, and significantly lower LWC when convection is less active. Budget analysis was performed to help understand these differences.