Observational constraints on the cloud thermodynamic phase in midlatitude storms

The conditions under which supercooled liquid water gradually gives way to ice in the mixed-phase regions of clouds are still poorly understood and may be an important source of cloud feedback uncertainty in general circulation model projections of long-term climate change. Two winters of cloud phase discrimination, cloud-top temperature, sea surface temperature, and precipitation from several satellite datasets (the NASA Terra and Aqua Moderate Resolution Imaging Spectroradiometer, and the Tropical Rainfall Measuring Mission) for the North Atlantic and Pacific Ocean basins are analyzed to better understand these processes. Reanalysis surface pressures and vertical velocities are used in combination with a synoptic storm-tracking algorithm to define storm tracks, create composite storm dynamical and cloud patterns, and examine changes in storm characteristics over their life cycles. Characteristically different storm cloud patterns exist in the Atlantic and Pacific and on the west and east sides of each ocean basin. This appears to be related to the different spatial patterns of sea surface temperature in the two ocean basins. Glaciation occurs at very warm temperatures in the high, thick, heavily precipitating clouds typical of frontal ascent regions, except where vertical velocities are strongest, similar to previous field experiments. Outside frontal regions, however, where clouds are shallower, supercooled water exists at lower cloud-top temperatures. This analysis is the first large-scale assessment of cloud phase and its relation to dynamics on climatologically representative time scales. It provides a potentially powerful benchmark for the design and evaluation of mixed-phase process parameterizations in general circulation models and suggests that assumptions made in some existing models may negatively bias their cloud feedback estimates.