Weather is a major driver for the efficiency of aircraft operations in the National Air Space (NAS). During peak periods, 4,000 to 6,000 aircraft operate in the NAS, which equals approximately 50,000 aircraft operations per day (NBAA, 2007). Forecast tools that link weather forecasts to airport operation impacts can inform the Airspace Flow Program (AFP) of capacity restrictions, in advance, allowing for planned, rather than reactionary, airspace management. The AFP will in turn establish appropriate Flow Constrained Areas (FCAs) based on the coverage and severity of expected weather events. The FAA Air Traffic Control System Command Center (ATCSCC) can then issue AFP advisories to meter flow through and around the FCAs in anticipation of expected events, in addition to responding to real-time weather. But forecasting the weather, particularly convective weather, is full of uncertainty. How does uncertainty in weather forecasts in the next 6-12 hours, or 1-5 days, translate into airport impacts?

Ensemble weather forecasts have been adopted widely in the meteorological community in the last decade. Ensembles produce multiple weather forecasts that can be used, among other things, to assess forecast uncertainty. In our study, we link key weather parameters from European and US ensemble weather forecasts to airport operations. We use forecasts from the European Center for Medium-range Weather Forecasts' (ECMWF) Ensemble Prediction System (EPS) and the U.S. National Centers for Environmental Prediction's (NCEP) Global Forecast System (GFS) as sources of ensemble weather forecasts. Operations at specific airports for selected weather events are assessed from the ATCSCC Advisory Database. Given the airport operations profile and corresponding ensemble weather forecasts, we will find the statistical relationships between ensemble weather forecast parameters and airport delays. Does weather forecast uncertainty correlate to uncertainty in actual airport impacts? What is a significant amount of certainty in the weather forecast, from the point of view of airport efficiency? While our study uses global weather models whose predictive skill extends out several days to a week, similar applications can be made to shorter-range, 0-24 hr, ensemble weather forecasts.