A Double-Moment Multiple-Phase Four-Class Bulk Ice Scheme. Part I: Description

Abstract

A detailed ice-phase bulk microphysical scheme has been developed for simulating the hydrometeor distributions of convective and stratiform precipitation in different large-scale environmental conditions. The proposed scheme involves 90 distinct microphysical processes, which predict the mixing ratios and the number concentrations of small ice crystals, snow, graupel, and frozen drops/hail, as well as the mixing ratios of liquid water on wet precipitation ice (snow, graupel, frozen drops). The number of adjustable coefficients has been significantly reduced in comparison with other bulk schemes. Additional improvements have been made to the parameterization in the following areas: 1) representing small ice crystals with nonzero terminal fall velocities and dispersive size distributions, 2) accurate and computationally efficient calculations of precipitation collection processes, 3) reformulating the collection equation to prevent unrealistically large accretion rates, 4) more realistic conversion by riming between different classes of precipitation ice, 5) preventing unrealistically large rates of raindrop freezing and freezing of liquid water on ice, 6) detailed treatment of various rime-splintering ice multiplication mechanisms, 7) a simple representation of the Hobbs-Rangno ice enhancement process, 8) aggregation of small ice crystals and snow, and 9) allowing explicit competition between cloud water condensation and ice deposition rates rather than using saturation adjustment techniques. For the purposes of conserving the higher moments of the particle distributions, preserving the spectral widths (or slopes) of the particle spectra is shown to be more important than strict conservation of particle number concentration when parameterizing changes in ice-particle number concentrations due to melting, vapor transfer processes (sublimation of dry ice, evaporation from wet ice), and conversion between different hydrometeor species. The microphysical scheme is incorporated into a nonhydrostatic cloud model in Part II of this study. The model performed well in simulating the radar and microphysical structures of a midlatitude?continental squall line and a tropical?maritime squall system with minimal tuning of the parameterization, even though the vertical profiles of radar reflectivity differed substantially between these storms.