Latent Heat Release Drives the Vertical Evolution of Seeded Ice Plumes in Supercooled Stratus Clouds

The interactions among aerosol perturbations, cloud droplet freezing, and atmospheric dynamics are a critical source of uncertainty in our understanding of mixed-phase clouds. In the context of glaciogenic cloud seeding, for example, it is unclear whether the vertical transport of newly nucleated ice crystals is passively controlled by pre-existing turbulent flow or actively controlled by latent heat release associated with ice crystal growth. To address this question, we present a comprehensive analysis that bridges the gap between Eulerian field observations and Lagrangian process understanding. We use high-resolution large-eddy simulations coupled with the habit-resolving bin microphysics scheme SCALE-AMPS. These simulations are constrained by in situ measurements from a targeted seeding experiment in supercooled stratus during the CLOUDLAB campaign in Switzerland.

Our sensitivity analysis, which systematically varies the vertical wind conditions at the time of seeding, reveals a fundamental decoupling between the initial vertical wind speed and long-term plume evolution. Although the ambient vertical velocity determines the trajectory of the ice plume during the initial minutes, we identify a "crossover point" at which latent heat release begins to dominate. The growth of the seeded crystals through vigorous vapor deposition releases substantial latent heat, generating a localized buoyancy flux. This thermal perturbation is strong enough to terminate and eventually reverse the descent of plumes that form in downdrafts. Plumes seeded into updrafts rise rapidly, yet they are stopped vertically as they reach the cloud-top inversion layer. Conversely, plumes initiated in downdrafts undergo a delayed, buoyancy-driven ascent, resulting in a deeper vertical spread and enhanced mixing. Although downdraft plumes temporarily lose liquid water when approaching the drier cloud base, they recover and persist within the mixed-phase layer due to self-generated lift.

These results demonstrate that seeded ice plumes actively influence their development and always rise in our simulations independent from the vertical velocity within the cloud. This provides new constraints for modeling aerosol-cloud interactions in weakly forced stratiform systems.