P1.27 Laboratory studies of the fall speeds and interactions of complex ice particles

Monday, 28 June 2010
Exhibit Hall (DoubleTree by Hilton Portland)
Christopher David Westbrook, Univ. of Reading, Reading, United Kingdom; and C. Roberts and A. J. Heymsfield

The fall of ice crystals and snowflakes through the atmosphere has been identified as a key sensitivity in climate models, affecting the development of precipitation, the lifetime of clouds, and the distribution of water vapour in the troposphere. State-of-the-art cloud microphysics schemes increasingly choose to link ice particle size, shape and density explicitly to fall velocity. But currently, the basic characteristics of how snowflakes fall are not well understood, largely due to the difficulty in measuring all the relevant parameters at once for tiny, fragile ice crystals falling through the air. To better understand the fall characteristics of complex ice particle shapes we have performed experiments in the laboratory using scaled-up models falling slowly through fluids of different viscosities. In this controlled environment we have accurately measured the fall speeds, weight, area ratio and Reynolds number, along with their fall patterns and other characteristics. Using the principle of dynamic similarity, the results are applied to real snowflakes in the atmosphere, and used to test common formulae for calculating ice particle fall speeds. For particles with an open geometry (low area ratio) we find that the particle fall speed is significantly overestimated using present methods, and a new correlation to fix this bias is proposed.

In addition to studying the behaviour of single particles, the interaction of pairs of ice crystals falling in proximity to one another has also been studied. We find that wake capture is ubiquitous, with interaction over many particle diameters typical. Jayaweera and Mason observed this behaviour for two identical circular discs: we demonstrate that the effect is much more general. Initially the two particles fall steadily; then the upper crystal begins to oscillate and accelerate; finally the upper particle comes to rest on top of the lower particle, and the pair move on together steadily through the fluid. We also find that particles of very different fall speeds can often repel one another, suggesting that the collision kernel for ice crystal collection is very different to the standard geometric sweep-out which is normally assumed. The implications for explicit modelling of the aggregation process in numerical models is discussed.

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner