A number of operations in the aviation industry, particularly in the terminal area, are weather-sensitive to local conditions in the short-term (3 to 18 hours). Often, they are reactive due to unavailability of appropriate predicted data at the required temporal and spatial scale. Hence, optimization that is applied to these processes to enable proactive efforts utilize either historical weather data as a predictor of trends or the results of synoptic- to meso-beta-scale weather models. Since this time range is beyond what is feasible with modern nowcasting techniques, near-real-time assessment of observations of current weather conditions may have the appropriate geographic locality, by its very nature is only directly suitable for reactive response.

Alternatively, meso-gamma-scale numerical weather models operating at higher resolution in space and time with more detailed physics may offer greater precision and accuracy within a limited geographic region such as a terminal area, for problems with short-term weather sensitivity. Such forecasts can be used to improve operational efficiency and safety by improving both the quality and lead time of such information. In particular, they appear to be well-suited toward improving economic and safety factors of concern to airport terminal operators. Among others, such factors relate to enabling air traffic controllers and dispatchers to develop more effective alternative flight paths to reroute aircraft around dangerous weather. Airline officials could initiate recovery plans before weather-induced disruptions actually occur, rescheduling passengers and aircraft to avoid congestion in affected areas and to improve safety. Airport operators could more efficiently schedule and staff aircraft deicing and snow removal operations during the winter.

To begin to address these issues, we build upon our earlier work, the implementation of an operational testbed, dubbed "Deep Thunder", which has been customized for transportation applications. This protoype provides nested 24-hour forecasts, which are typically updated twice daily, for the New York City metropolitan area to 1 km resolution utilizing explicit, bulk cloud microphysics. It has been recently extended to also provide forecasts for the Chicago and Kansas City metropolitan areas at 2 km resolution, at least once per day. All of the processing, modelling and visualization are completed in one to two hours on relatively modest hardware to enable sufficiently timely dissemination of forecast products for terminal area applications at reasonable cost.

In order to evaluate the potential utility of this class of numerical weather prediction for terminal operations, we pose a question: could the availability of such forecasts enable some ground delays to be avoided? We begin by determining particular days in 2002 during which there was a significant ground delay at any of the three large airports run by the Port Authority of New York and New Jersey, all of which are within our 1 km forecasting region. We then classify these days into two categories. The first is days when severe weather was reported in this geographic region by the local office of the National Weather Service (NWS) at Upton, NY and the NOAA Storm Prediction Center. We then need to determine what type of and when specific weather forecast information was available for potential use in airport operations. Since we do not have access to the direct forecasts utilized by either the Federal Aviation Administration (FAA) nor the Port Authority for such purposes, we utilize the zone forecasts produced by the Upton NWS office as a proxy. We address the questions, do these forecasts point to sufficiently severe weather or not? If the latter, then were the delays due to the FAA missing a traffic flow management alert?

The second category is days when there was a NWS forecast of severe weather, which resulted in the FAA issuing a ground delay, but the disrupting weather never materialized. For both categories, we compare forecasts produced for those specific days by Deep Thunder. In the first case, we consider if our forecasts indicated weather sufficiently severe to impact terminal operations. In the second, we determine if our forecasts illustrate a lack of severe weather in this geographic region. This work is on-going. We are in the process of extending this analysis to more recent incidents in 2003 and 2004.