Abstract
The cloud model of Tripoli and Cotton was used to simulate a cumulonimbus cloud observed during the Cooperative Convective Precipitation Experiment (CCOPE). We tested the sensitivity of the precipitation pathways in the model to the initial concentration of cloud droplets above cloud base Nc (which is related to the concentration of cloud condensation nuclei). The results showed that for large Nc, Manton and Cotton's autoconversion parameterization properly suppressed supercooled rain formation via the “warm-rain” process in a cold-based, continental cloud, forcing ice processes (e.g., riming, aggregation and deposition of vapor) to produce graupel. With lower droplet concentrations, rain formed first through warm-rain processes, then graupel formed through freezing. The value of Nc, which determined the transition from graupel formation by freezing rain to graupel formation by ice processes was found to be sensitive to the parameter acm, which represents the critical mean radius at which collision and coalescence begin. The observed cloud was also compared with a cloud simulation which had approximately the correct initial cloud-droplet concentration (Nc). The simulated cloud base was somewhat lower than observed, indicating that the initial sounding was too moist in the subcloud layer. As a result, the modeled cloud was wetter than observed. In spite of this variation from the observed cloud base, other properties were represented rather well by the simulated cloud, including cloud top height, peak vertical velocities, and the growth stages in the development of the storm.