| << | Profiling of Clouds | >> |
Abstract
We present an analysis of a 17 month of observations made with cloud radars and lidars. The study addresses two topics. Firstly, is there a large number of ice crystals smaller than 100 microns present in most ice clouds which would dominate the cloud albedo? Secondly, what nucleation mechanism operates to produce ice in thin but persistent supercooled liquid clouds layers?
Analysis of the velocities of ice crystals made with a vertically pointing 1.5 micron Doppler lidar shows that the average terminal velocity is consistent with that expected from ice particles having an exponential size distribution, and is much higher than would be expected if there were large numbers of small crystals. Observations of thin supercooled layer clouds made with Dopplerised radar and lidar indicate a continuous flux of small ice particles falling out of them which can persist for many hours. Most models assume that the ice nucleation occurs as a singular process, whereby at a particular temperature a given number of nuclei become active. Our observations suggests that an alternative 'stochastic' nucleation mechanism is operating, whereby many of the water drops contain an ice nuclei, but the nucleation is a stochastic statistical process so the fraction of nuclei which are activated in unit time is very low. Consequently if the clouds persist for many hours they can produce a much higher number of ice particles than predicted by the singular mechanism. Since such supercooled layer clouds are a frequent occurrence we suggest that this stochastic immersion nucleation process is quite common.