This presentation focusses on the characteristics of persistent, thin (~300m) single-layer mixed-phase clouds such as altocumulus. Such clouds are easily identified using lidar by their strong reflection of the laser light and rapid attenuation of the beam in cloud. Doppler measurements from the liquid layer show that the net air velocity of the liquid droplets is close to zero and large scale ascent is minimal; but eddy dissipation rates from high resolution Doppler radar & lidar measurements show that the air is well mixed in the 500m below cloud top. This well mixed layer is bounded above and below by much drier, stable air. This inhibits entrainment of fresh air (and hence fresh ice nuclei) into the cloud layer, yet our radar and lidar observations show that a steady flux of ice crystals fall out over the course of many hours. Coincident aircraft observations have allowed us to quantify this flux, and we find that the supply of conventional ice nuclei would be depleted within ~1hr. We therefore argue that nucleation in these persistent supercooled clouds must be strongly time dependent (stochastic) in nature. Current ice nucleus counters are inadequate to quantify this mode of nucleation.
Once nucleated, the ice crystals are observed to grow primarily by vapour deposition rather than riming, because of the low liquid water path (microwave radiometer measurements show this is typically a few tens of g/m2), yet vapour-rich environment. Evidence for this comes from specular reflection of the lidar beam from their faceted surfaces, strong differential reflectivity signals from their radar returns, and in-situ imaging of the crystals. The flux of vapour from liquid to ice is quantified from in-situ measurements, and we suggest that this flux is offset by radiatively driven overturning of ice-saturated air below cloud base, allowing the cloud to persist in a quasi steady state.