Parameterization of surface kinetic effects for bulk microphysical models: Influences on simulated cirrus dynamics and structure

Abstract

A parameterization for the influences of the deposition coefficient (αd) and, therefore, surface kinetics on ice crystal vapor growth in bulk microphysical models is derived. The parameterization is developed by considering three distinct growth regimes, each of which depends on the mean size of the ice crystals: inefficient growth in which surface kinetics dominate vapor growth, efficient growth in which diffusion dominates vapor growth, and an intermediate regime in which both surface kinetics and diffusion are important. Analytical solutions to the distribution‐integrated vapor growth equations are derived for the inefficient and efficient growth regimes, whereas a plausible approximation is suggested for the intermediate growth regime. Use in a numerical cloud model requires a method to choose between the growth regimes, and we show that the kinetic length scale can be used for this purpose. The parameterization is used in eddy‐resolving simulations of a warm and unstable cirrus case. Simulation results tend to match prior microphysical studies: cirrus microphysics are insensitive to αd if αd > 0.1, whereas lower values of αd produce relatively high ice concentrations and ice supersaturations. Our simulations suggest that large changes in cirrus structure and dynamics occur when αd becomes lower than ∼0.05. When growth is this inefficient, ice concentrations are high and precipitation rates are low, which leads to a cirrocumulus‐like structure over time. The dynamic motions in these clouds are driven primarily by infrared cloud top cooling and cloud base warming. At larger values of αd (>0.05), the more efficient vapor growth leads to lower ice concentrations and larger precipitation rates. This causes the initially layered cirrus to transition into cirrus uncinus with precipitation tails. The cloud fraction in this case is low, with lifetimes almost 6 hours less than in the cases with inefficient growth (αd < 0.05).