Cloud-phase sensitivity of a stable Arctic mixed-phase cloud during ARTofMELT to microphysical factors

Arctic low-level clouds are highly sensitive to microphysical processes, which can either sustain or break down the cloud-phase state and thereby determine the longevity of the clouds and their radiative impacts. They are influenced by aerosol particles, which can act as ice nuclei or cloud condensation nuclei, and simulating these clouds is additionally influenced by the parameterization schemes used for the aerosol-cloud interactions and the microphysical processes in the cloud. In the presented study, we simulate a stable mixed-phase stratocumulus cloud case observed during the ship-based ARTofMELT campaign (Atmospheric rivers and the onset of Arctic melt) on 7 June 2023 with the large-eddy simulation model MIMICA-LES. The simulation is initialized by radiosoundings and constrained by ground-based remote sensing (liquid water path (LWP) and ice water path (IWP)) and aerosol measurements (aerosol size distributions, hygroscopicity, and aerosol type). We perturb the total aerosol number concentration, aerosol type, initial liquid water content (LWC), prescribed ice crystal number concentration, and ice habit to estimate the relative importance of these aerosol and microphysical parameters with respect to the modeled LWP/IWP using a factorial analysis as a statistical approach. Through factorial analysis, we can quantify the variance contribution of all parameters to LWP/IWP and quantify the interaction between different parameters. We find that ice crystal number concentration has the greatest impact on LWP and IWP, followed by the ice crystal habit, which can determine whether a cloud glaciates or not, given a fixed ice crystal number concentration. The ice habit is relatively less important, but it can determine whether a cloud glaciates or not, given fixed aerosol type and ice crystal number concentration. The results from our study can help to constrain and improve future closure studies between observations and small-scale modeling.