Evaluation of an ice nucleation parameterization based on classical nucleation theory using three mixed-phase cloud events observed during ISDAC

The parameterization of heterogeneous ice nucleation continues to be a large source of uncertainty when simulating cold clouds in models ranging from the cloud-resolving to the global scale. In this presentation, a new heterogeneous ice nucleation parameterization based on classical nucleation theory is described and evaluated. The parameterization represents the nucleating ability of three different externally mixed aerosol compounds (namely dust, black carbon and organics/bioaerosols, each described by two size modes) in the immersion, condensation and deposition nucleation modes. A contact angle distribution representative for each aerosol population is used to describe the nucleation efficiency of each particle type. The contact angle distributions are parameterized based on recent laboratory measurements. The ice nucleation parameterization is implemented in the Large-Eddy Simulation solver MIMICA in order to test the ability of the parameterization to properly model ice formation under realistic cloud conditions. The evolution of three mixed-phase clouds developing under different aerosol and thermodynamic conditions and observed during the Indirect and Semi-Direct Aerosol Campaign (ISDAC, North Slope of Alaska, April 2008) are simulated. ISDAC provides detailed aerosol size/composition measurements performed for more than 30 cloud events on-board a research aircraft. These observations, together with available meteorological data, are used to initialize the model. The three cases span a range of different temperatures and aerosol concentrations. In the first case (flight 16), the cloud top temperature was -17ºC and relatively few aerosol particles (and also ice nuclei) were observed (≈175 cm^-3). In the second case (flight 31), the cloud top was slightly warmer (-15ºC) and more aerosols were present (≈210 cm^-3). The last case (flight 25) was quite different compared to the other two with a cloud top temperature close to -10ºC combined with high aerosol concentrations (Na ≈750 cm^-3) originating from the advection of a bio-mass burning plume. The model is compared with available data gathered during ISDAC, in particular with respect to the simulated cloud ice properties. In addition, sensitivity simulations are performed to examine the most crucial parameters of the parameterization.