The chance of freezing – a conceptional study to parameterize temperature-dependent freezing by including randomness of ice-nucleating particle concentrations

• Published in: Atmospheric chemistry and physics 2023-10, Vol.23 (19), p.10883-10900
• Main Authors: Frostenberg, Hannah C, Welti, André, Luhr, Mikael, Savre, Julien, Thomson, Erik S, Ickes, Luisa
• Format: Article
• Language: English
• Subjects: Aerosols; Atmosphere; Clouds; Contact angle; Datasets; Distribution; Dust; Freezing; Frequency distribution; Heterogeneity; Ice; Ice nucleation; Illustrations; Immersion; Large eddy simulation; Large eddy simulations; Meteorologi och atmosfärforskning; Meteorology and Atmospheric Sciences; Nucleation; Oceanic eddies; Parameterization; Phase transitions; Probability theory; Submerging; Temperature; Temperature dependence
• ISSN: 1680-7324; 1680-7316; 1680-7324
• DOI: 10.5194/acp-23-10883-2023

Ice-nucleating particle concentrations (INPCs) can spread over several orders of magnitude at any given temperature. However, this variability is rarely accounted for in heterogeneous ice-nucleation parameterizations. In this paper, we present an approach to incorporate the random variation in the INPC into the parameterization of immersion freezing and analyze this novel concept with various sensitivity tests. In the new scheme, the INPC is drawn from a relative frequency distribution of cumulative INPCs. At each temperature, this distribution describing the INPCs is expressed as a lognormal frequency distribution. The new parameterization scheme does not require aerosol information from the driving model to represent the heterogeneity of INPCs. The scheme's performance is tested in a large-eddy simulation of a relatively warm Arctic mixed-phase stratocumulus. We find that it leads to reasonable ice masses in the cloud, especially when compared to immersion freezing schemes that yield one fixed INPC per temperature and lead to almost no ice production in the simulated cloud. The scheme is sensitive to the median of the frequency distribution and highly sensitive to the standard deviation of the distribution, as well as to the frequency of drawing a new INPC and the resolution of the model. Generally, a higher probability of drawing large INPCs leads to substantially more ice in the simulated cloud. We expose inherent challenges to introducing such a parameterization and explore possible solutions and potential developments.