4-D cloud properties from passive satellite data and applications to resolve the flight icing threat to aircraft

The threat for aircraft icing in clouds is a significant hazard that routinely impacts aviation operations. Accurate diagnoses and forecasts of aircraft icing conditions requires identifying the location and vertical distribution of clouds with super-cooled liquid water (SLW) droplets, as well as the characteristics of the droplet size distribution. Traditional forecasting methods rely on guidance from numerical models and conventional observations, neither of which currently resolve cloud properties adequately on the optimal scales needed for aviation. Satellite imagers provide measurements over large areas with high spatial resolution that can be interpreted to identify the locations and characteristics of clouds, including features associated with adverse weather and storms. This thesis develops new techniques for interpreting cloud products derived from satellite data to infer the flight icing threat to aircraft in a wide range of cloud conditions. For unobscured low clouds, the icing threat is determined using empirical relationships developed from correlations between satellite imager retrievals of liquid water path and droplet size with icing conditions reported by pilots (PIREPS). For deep ice over water cloud systems, ice and liquid water content profiles are derived by using the imager cloud properties to constrain climatological information on cloud vertical structure and water phase obtained apriori from radar and lidar observations, and from cloud model analyses. Retrievals of the SLW content embedded within overlapping clouds are mapped to the icing threat using guidance from an airfoil modeling study. Compared to PIREPS, the satellite icing detection and intensity accuracies are found to be about 90% and 70%, respectively. Mean differences between the imager IWC retrievals with those from CloudSat and Calipso are less than 30%. This level of closure in the cloud water budget can only be achieved by correcting for errors in the imager retrievals due to the simplifying but poor assumption that deep optically thick clouds are single-phase and vertically homogeneous. When applied to geostationary satellite data, the profiling method provides a real-time characterization of clouds in 4-D. This research should improve the utility of satellite imager data for quantitatively diagnosing and predicting clouds and their effects in weather and climate applications.