Intercomparison of large-eddy simulations of Arctic mixed-phase clouds

Description

An intercomparison of large-eddy simulations (LES) of mixed-phase stratus observed over the Arctic sea ice during Flight 31 of the Indirect and Semi-Direct Aerosol Campaign (ISDAC) has been conducted. Results from over a dozen model configurations submitted by nine research groups are analyzed to gain insight into the processes controlling the longevity of mixed-phase clouds and to investigate the origins for the diversity among LES results documented in previous studies. Three simulations are conducted with each model configuration: one without any ice processes included and two with different prescribed ice number concentration (1 and 4 per liter, respectively). In addition to prescribing ice particle number concentration, this study also constrains mass-size and fall speed-size relationships used in all models and employs a simple parameterization for the longwave radiative cooling rate. The imposed constraints reduce but do not eliminate the spread of LES results. Two principal sources of remaining inter-model differences are identified. First, even in the absence of any ice processes, simulations are shown to diverge due to differences in formulations of model dynamics (e.g., advection and turbulent mixing). The second major source of inter-model variability comes from differences in a representation of ice particle size spectrum. Although the rates for depositional growth and sedimentation on a single particle basis are uniformly constrained across models, variability in assumed (in bulk models) and predicted (in bin models) ice size distributions leads to different integral rates for cloud ice growth and removal. Simulations using two of the models that have both (bulk and bin) microphysics options available show that predicted ice water paths differ by as much as a factor of two, depending on whether the shape of ice spectrum is predicted or assumed to be exponential.