Modelling secondary ice production in Arctic mixed-phase clouds

Ice number concentration is a critical parameter for Arctic mixed-phase clouds. Several observations have shown that such relatively warm clouds can have orders of magnitude higher ice concentrations than expected based on typical Ice-Nucleating Particle (INP) concentrations. The most common explanation is that Secondary Ice Production (SIP) such as rime splintering (Hallett-Mossop ice multiplication) process causes the increase in ice concentration. Here we use observations from two field campaigns. In both campaigns the observations indicated that one or more SIP processes were actively producing ice. Due to the high temperatures around 265 K, the focus is on Hallett-Mossop process. Observed meteorological conditions and aerosol size distributions were used to initialize high-resolution large-eddy model UCLALES-SALSA simulations. Primary ice formation was modelled based on fixed INP concentrations so that the observed ice concentration was at least ten times larger than the INP concentration. Hallet-Mossop ice multiplication factors due to rime-splintering did not reproduce observed rates of secondary ice production. An increment of about one order of magnitude was needed to find agreement between modeled and observed ice number concentrations. This highlights the urgent need of laboratory and model studies that unveil the variable dependencies controlling SIP mechanisms. Secondary ice production can be increased by adjusting the simulated cloud temperature towards the optimal value and by increasing cloud water content. Extending simulation time up to 10 hours or more will also help. Although high ice concentrations can be obtained simply by increasing the INP concentrations, details such as vertical ice distribution and spatial variability will be different than in the case where SIP is used. Although this difference has a small impact on cloud dynamics during these 10-hour simulations, long-term impacts are likely.