Supplementary material to "Microphysical sensitivity of coupled springtime Arctic stratocumulus to modelled primary ice over the ice pack, marginal ice, and ocean"

Abstract

The influence of immersion-and contact-freezing within the Morrison et al. (2005) microphysics scheme was tested to quantify their contribution to N ice. Simulations with contact-freezing (Meyers et al. 1992-hereafter, M92) and immersion-freezing (Bigg 1953-hereafter, B53) switched either on or off are shown in Fig. S1. The addition of B53 and M92 produces a significantly larger ice crystal number concentration (up to 3 L −1 , 1.5 L −1 , and 10 L −1 in cases 1, 2, and 3 respectively) than the mean observed (0.47 ± 0.86 L −1 , 0.35 ± 0.20 L −1 , and 0.55 ± 0.95 L −1 respectively, Table 2). Modelled ice number concentrations with and without B53 and M92 active are similar in case 1. Both representations cause glaciation, and liquid water is not modelled at any point during the simulations. No improvement can be seen in the liquid water mixing ratio when both the B53 and M92 nucleation mechanisms are disabled. Modelled ice number concentrations for case 2 peak at ∼1.5 L −1 and ∼0.8 L −1 with and without both B53 immersion-and M92 contact-freezing nucleation active. Both scenarios allow for liquid water to form in the cloud, with ∼0.2 g kg −1 modelled. When B53 and M92 are active in case 3, high ice number concentrations are rapidly simulated at approximately 12 h-14 h. This event causes the evaporation of all simulated liquid water, and the region of high ice number concentration dissipates back to the original sustained concentration of ∼2 L −1 afterwards. This event is not simulated when B53 and M92 are disabled, suggesting these additional sources of ice number are the cause of this phenomenon.