Observational constraints on cloud thermodynamic phase in midlatitude storms

Uncertainties in cloud feedbacks estimated from general circulation models can be related in part to assumptions used for cloud phase parameterization. Current parameterizations are often based on simple relationships between temperature and glaciation that rely on in-situ aircraft measurements obtained in mid-latitude clouds which were found to be fully glaciated near -15°C. However, more recently, ground based and space borne lidar observations and satellite observations revealed that supercooled water can be present at cloud top down to temperatures as low as -30°C. These conflicting measurements were obtained for spatially and temporally limited datasets and their differences may point to other factors that intervene in cloud glaciation, such as the concentration in ice nuclei, the vigor of vertical motions and precipitation and the age and depth of the clouds. We use, for 2 winters, cloud-top temperature and phase retrievals from MODIS over the North Atlantic and Pacific mid-latitude storm tracks, in addition to information on precipitation rates from the Tropical Rainfall Measurement Mission and pressure and vertical velocity profiles from the NCEP reanalysis dataset. Storms that occurred over both oceans during both winters are tracked and composites centered on the storms pressure minimum created for all parameters. The mean cloud-top temperature for clouds composed of 50% ice is used as a diagnostic quantity to evaluate spatial variations of glaciation behavior in storms. We will show how the 50% ice cloud top temperature relates to vertical velocities and sea-surface temperatures and provide a tentative explanation of the processes involved. Differences between both ocean basins will also be discussed. Finally, comparisons with ground-based measurements will be presented for a discussion on how MODIS observations relate to previous measurements.