Chasing ice crystals: Lagrangian trajectories in ICON-LES for investigating liquid and ice phase interactions

The ice phase is a major contributor to precipitation formation over continents, as ice hydrometeors efficiently grow to large enough sizes for them to sediment. Several mechanisms underlie the growth of ice crystals, with one being the growth through vapor deposition onto the ice crystal. The speed of this growth depends next to temperature also on the availability of water vapor, one source being cloud droplets in mixed-phase clouds that locally may experience water-subsaturated conditions. This is referred to as the Wegener-Bergeron-Findeisen (WBF) process and describes the ability of ice crystals to grow at the expense of cloud droplets. While the presence of the WBF process is established, the actual growth rates of ice crystals in such conditions remain ambiguous. We conducted field experiments within the CLOUDLAB project with the goal to infer ice crystal growth rates in naturally occurring supercooled clouds through local perturbations from cloud seeding. A unique dataset was collected describing the characteristics of cloud droplets and ice crystals in the probed clouds, including their sizes, number concentrations, and ice and water contents.

Here, we combine large-eddy simulations (LES) in 65 m horizontal resolution with online Lagrangian trajectories to achieve a more straightforward comparison to our observations. We show that the model simulations can reproduce the field experiments in terms of ice crystal number concentrations. However, both the simulated changes in the liquid phase as a consequence of the WBF process and the ice crystal growth rates are underestimated compared to the observations. We perform a series of sensitivity studies on the vapor depositional growth equation of ice crystals given the uncertainty and simplification of two parameters of that equation.  We find that an increase of the vapor deposition efficiency up to a factor of three achieves comparable growth rates. However, matching growth rates in the model and observations does not lead to coinciding changes in the liquid phase, i.e., the WBF process remains too slow. We identify two limitations of our approach: (i) the simulated and actual water vapor fields may differ and (ii) our LES are still too coarse to fully capture the small-scale interactions between the liquid and ice phases. This study highlights the synergy of high-resolution model simulations and field observations for investigating a fundamental cloud process. Our results provide insights for future mixed-phase cloud modeling studies.