Time-dependent versus singular ice nucleation schemes: Estimated impact on mixed-phase stratiform clouds in ModelE3

Poster PDF

Description

In a companion study of North Slope of Alaska long-term measurements (Lamer et al. poster), we find that supercooled liquid water occurs most commonly in single-layer, shallow stratiform clouds that are fully overcast (where initial ice formation must proceed heterogeneously via activation of ice nuclei). We also evaluate the performance of the NASA GISS ModelE3 climate model with diagnostic stratiform ice crystal number concentration. In another companion study (Knopf et al. poster), we calculate the impact of K-feldspar dust emission size distribution on ice nuclei active in the immersion mode as predicted by a singular active-site scheme that does not treat time dependence. Here we attempt to address the question: is it important to use a time-dependent ice nucleation scheme in ModelE? Since ModelE3 ice formation will be prognostic only in the stratiform cloud scheme (whereas phase partitioning will remain diagnostic in the convective scheme), here we focus on stratiform clouds. Field experiments and modeling studies have commonly demonstrated that in single-layer, mixed-phase, boundary-layer clouds that are coupled with the surface (a relevant subset of the cloud type that most commonly hosts supercooled liquid water over the NSA site), ice number concentration can commonly be approximated as well mixed. Simulations that reproduce many aspects of observations also indicate that such clouds may commonly remain weakly desiccated by the ice present, in part explaining their longevity. Within such clouds, detailed simulations indicate that a characteristic size distribution may be achieved by cloud ice within the relatively well-mixed boundary layer, such that a change in the nucleation rate of new ice crystals simply scales the mean ice size distribution accordingly. Finally, a typical ice crystal lifetime of roughly one hour has been estimated based on detailed simulations and independent analytical calculations. Taken together, these conditions lead to a simplified model that is used to efficiently investigate the Lagrangian time evolution of boundary-layer cloud ice number concentration in response to heterogeneous ice nucleation over a wide range of simple and complex assumptions about ice-nucleating aerosol composition and behavior. We use this simplified framework to evaluate the likely importance of explicitly treating heterogeneous ice nucleation time dependence in the NASA GISS ModelE stratiform cloud scheme.