Ice formation at moderate supercooling in mixed-phase clouds and its link to precipitation

Ice formation in the atmosphere is important for the generation of precipitation and the radiative properties of clouds. An integrated understanding of ice formation processes is still missing. This is especially true for ice formation at low to moderate supercooling. In this mixed-phase cloud temperature regime, primary ice is formed via heterogeneous ice nucleation, where ice-nucleating particles (INPs) promote freezing. Such INPs are mainly of biological origin and are present in relatively low concentration in the atmosphere. If the ice particle concentration is higher than the INP concentration, this indicates that secondary ice formation processes are active in addition to heterogeneous nucleation. Secondary ice formation processes can multiply the primary ice by up to several orders of magnitude. However, these processes are diverse and difficult to quantify. After ice formation, various other processes may occur before surface precipitation is observed. The complex chain of intertwined microphysical mechanisms that ultimately lead to precipitation can take different paths.
Here we applied different approaches to obtain information on primary and secondary ice formation at moderate supercooling. In particular, we present observations of INPs active at around $-$15 °C in more than 120 aerosol samples and 220 individual dendritic ice crystals that were collected and analysed at mountain stations in the Swiss Alps during winter months of 2018 and 2019. Aerosol particle concentrations, air mass origin and precipitation history were combined to parameterise INP concentrations measured at Weissfluhjoch (2671 m a.s.l.). Primary dendritic ice crystals were quantified at Jungfraujoch (3580 m a.s.l.) using an approach that makes use of their particular and narrow growth temperature range. In addition, precipitating snow particles captured at ground level and coinciding radiosonde ascents were analysed to investigate whether mixed-phase clouds were relevant for snowfall at an Arctic site throughout a total of eight cold months in 2019 and 2020.
We found that it is more promising to parameterise atmospheric concentrations of INPs active at $-$15 °C measured at Weissfluhjoch using aerosol particle number concentrations of a size $>$ 2 µm as compared to smaller aerosol particles (Chapter 2). Differentiating between air masses that were precipitating, non-precipitating, and carrying Saharan dust and non-precipitating improved the prediction. The ratio of INP to aerosol particle $>$ 2 µm was larger in precipitating air masses than in non-precipitating air masses. Through freezing assays of individual dendritic ice crystals sampled within clouds at Jungfraujoch, we found that on average one out of eight dendrites contained an INP active at moderate supercooling (Chapter 3). Therefore, the multiplication factor for dendrites was on average only around one order of magnitude. At a site in Northern Finland, observations of often small, unrimed snow particles and the matching relative humidity profiles indicated that probably one quarter of the precipitating clouds were mixed-phase and the remainder were fully glaciated (Chapter 4).
The simultaneous investigation of different microphysical ice processes in clouds can provide information about intertwined processes. Field observations of heterogeneous ice nucleation, secondary ice formation and precipitation are limited in space and time. In this thesis, we have used experimental top-down approaches to determine quantities related to ice formation at moderate supercooling. These results can be incorporated into atmospheric models, which in turn can place the measurements in a larger perspective. Therefore, experimental research in cloud physics is critical for the development of models that can be used to understand and simulate the driving forces and effects of microphysical ice processes and their changes in future climates. This will require close collaboration between researchers working in situ and in silico.