Toward an integrated approach to improve the utility of operational satellite data for diagnosing icing conditions

Accurate diagnoses and forecasts of the meteorological conditions associated with aircraft icing remain a significant challenge because these require identifying the location and vertical distribution of clouds with super-cooled liquid water (SLW) droplets, as well as characteristics of the droplet size distribution. The “current icing potential” (CIP) product is a valuable tool, available for operational use at the Aviation Weather Center. CIP is created with a data-fusion approach that combines satellite, radar, surface, lightning, and pilot-report observations with numerical model output to create an hourly three-dimensional diagnosis of the potential for icing and super-cooled large droplets (SLD). While some cloud top information from satellite data is incorporated in the CIP (and in the numerical model analyses), much of the information content provided by advanced satellite sensors remains under-utilized. Icing conditions often occur in non-precipitating clouds invisible to weather radar, in regions sparse with in-situ or ground–based observations, and in clouds that are often poorly characterized by numerical models on the scales needed for aviation. Icing conditions can also be highly variable, often occurring in small areas that cannot be resolved by model-based approaches. To respond to the need for improved icing diagnoses by the aviation weather forecasting community, NASA and NOAA recently sponsored research activities to (1) incorporate advanced satellite-derived cloud products in the CIP, and (2) derive flight icing threat estimates directly from the satellite data. Despite clear demonstrations of the benefits that the satellite–derived cloud products offered in both approaches, no path to operations was ever realized, nor is one currently planned. The purpose of this paper is to describe preliminary work toward a new integrated approach to the icing diagnosis problem that is more satellite-centric. To help accomplish this, a methodology is being developed to improve the vertical resolution of cloud properties from operational satellite data. Statistical techniques are employed to exploit the detailed information on the vertical structure of clouds, and the accurate characterization of cloud boundaries provided by active sensor data (i.e cloud radars and lidars) such as that obtained at the Department of Energy Atmospheric Radiation Measurement (ARM) facilities, and from the CALIPSO and CloudSat satellites. The vertical structure information is derived as a function of cloud type and constrained with cloud parameters derived from Geostationary Operational Environmental Satellite (GOES) data in order to derive cloud water content profiles at the GOES pixel level. The profiles are validated with ARM, CloudSat and CALIPSO data. Vertically resolved information on the partitioning of liquid and ice are also needed to estimate the potential for icing conditions. In our initial approach, this information is gleaned in a similar climatological, cloud type dependent fashion using the explicit microphysics scheme contained in the RUC and Rapid Refresh weather analysis system in order to derive profiles of SLW mass density, which are also constrained by the GOES retrievals. A procedure to scale these profiles to icing potential is developed and validated with icing PIREPS. The over-arching goal of this work is to develop a 3-dimensional approach to improve the utility of operational satellite data for constraining cloud properties in weather analyses and decision support systems that integrate cloud observations from multiple data sources.