Investigating a New Ground Delay Program Strategy for Coping with Sfo Stratus*

Integration of weather information into the air traffic flow control decision process is a key component of the planned Next Generation Air Transportation System (NexGen). The NexGen Joint Program Development Office (JPDO) has designated a Weather Integration sub-team to specifically address the issue of translating weather information into air traffic decisions (Andrews et al., 2007). The presumed source of weather information is an authoritative database which contains a distilled representation of weather parameters derived from a variety of external sources. The future concept is that this information will be represented probabilistically, recognizing that this is most suitable for an objective optimization of air traffic flow decisions on a systemwide scale. Many weather products emerging from the research community include a probabilistic representation of weather information, most notably those associated with convective weather which occurs on time and space scales that are particularly challenging for providing reliable deterministic forecasts beyond very short time frames relative to what is required for strategic traffic flow planning. Translating probabilistic information into decisions that represent optimized system wide benefits has long been a subject of operations research, but is a rare practice in the National Airspace System (NAS) where operational decisions are primarily made based on deterministic information, including forecasts of expected weather conditions. Changes to operational procedures in recent years that have attempted to integrate probabilistic weather information, particularly location and intensity of convective weather, have typically resulted in overly-conservative decisions to restrict traffic flow, failing to exploit improved forecasts provided by the currently emerging suite of convective weather products. One of the difficulties in applying the probabilistic information in the operational setting is the complex multi-dimensionality associated with both the impact and the prediction of convective weather. In contrast, the prototype San Francisco (SFO) Marine Stratus Forecast System (Clark and Wilson, 1997; Ivaldi et al. 2007). provides a probabilistic representation of weather