Ice crystal dynamics: in the lab and the atmosphere

The fall behaviour of atmospheric ice crystals plays a critical role in determining cloud lifetime, precipitation formation, and radiative effects. Yet, their motion and orientation remain poorly represented in weather and climate models. This thesis investigates the descent dynamics of complex ice crystals through three complemen tary experiments: controlled laboratory studies of analogues in free-fall in glycerine solutions, in still air, and in-situ natural snowfall observations. In the first set of experiments, 3D-printed ice crystal analogues were dropped through glycerine solutions to replicate atmospheric Reynolds numbers. Their tra jectories were captured using three orthogonal cameras and reconstructed using the Trajectory Reconstruction Algorithm implemented through Image anaLysis (TRAIL). The second set of experiments were designed for this thesis, using micro-machined ice crystal analogues falling through quiescent air, using a similar methodology. Finally, the Video In-Situ Snowfall Sensor (VISSS) instrument captured high-resolution, multi-frame observations of natural snowflakes. Each of the experiments revealed stable and periodic (unstable) behaviours, in cluding a spectrum of periodic behaviours between zigzagging and spiralling. The in-air and VISSS experiments revealed a tumbling mode of behaviour. Motion type was found to depend on particle shape, Reynolds number, Re, and dimensionless moment of inertia, I∗. The onset of unsteady motion occurred at lower Reynolds numbers for simpler particles, such as hexagonal plates, and at lower Re with in creasing I∗. Unstable particles were found to have orientation distributions with a non horizontal mode, deviating from common model assumptions. Unstable particles had broader distributions of vertical and horizontal velocity, implying greater ag gregation potential and sensitivity in remote sensing signals. Collectively, these findings provide new constraints on ice particle fall velocity and orientation, with relevance for microphysical growth, remote-sensing retrievals, and cloud optical properties. Such findings may improve the representation of clouds in climate models, where representation of cloud microphysical processes remains the largest source of uncertainty.