How does a snowflake form? A new generation of ice cloud chambers is set to give us the first detailed insightinto this delicate process and could even help weather forecasters better predict when snow is likely to fall.
Next year, Paul Connolly from the University of Manchester, UK, will study ice crystal aggregation in the Manchester Ice Cloud Chamber. This giant three-storey stainless steel cylinder can be cooled to temperatures as low as -50 °C to simulate the conditions that produce snow.
Ice crystals form when a burst of compressed air enters the tank. The air expands and cools, triggering water vapour in the chamber to freeze. As the ice crystals drift down the chamber they merge to create snowflakes.
Connolly's set-up includes a pair of laser beams that trigger a burst from a third laser when a flake crosses their path. This third laser projects a shadow of the flake onto a digital camera. Taking different images of the flakes as they fall will allow Connolly to gauge how their size changes during their journey down the cloud chamber.
He can also study the shape of the flakes by placing oil-covered glass plates at different levels within the chamber - which record the outline of the flakes as they settle.
By running experiments at different temperatures, Connolly hopes to see how the speed of growth of the snowflakes changes. Heavier snowflakes fall faster but are difficult to forecast, so this will allow his collaborators at the UK Met Office to better understand the relationship between snowfall and temperature.
Kenneth Libbrecht at the California Institute of Technology in Pasadena is using a similar but smaller chamber to crack another problem: why the structure of ice crystals changes so much with temperature.
Hexagonal crystals dominate when the temperature is just below zero. But as it drops, the crystals become columnar and then switch back to hexagonal as it gets even colder.
Libbrecht says the answer may lie with impurities in the air, which end up on the surface of a crystal and slow its growth. "This is something that people have been speculating about for 50 years," says Chris Westbrook, a meteorologist at the University of Reading, UK. He says Libbrecht's model is a potential solution.