Using Ensemble Weather Forecasts for Predicting Airport Arrival Capacity

Abstract

Strategic air traffic flow management requires predictions of airspace and airport capacities several hours into the future and particularly in response to forecasted constraints. Yet, forecasted weather at these longer time horizons is inherently uncertain, complicating the process both for traffic managers and the decision support automation being developed to assist in this process. Given the connectedness of the air traffic management system, automation is particularly well suited to capture the propagating effects of constraints under uncertainty; however, a generalized and adaptable approach for representing the impacts of both forecasted constraints and their underlying uncertainty is needed to populate these models. In this Paper, we examine the use of ensemble weather products to quantify uncertainty in airport capacity predictions. First, changes to an operationally structured runway configuration model are motivated and evaluated to improve model accuracy under observed data. Next, model performance is evaluated when using ensemble forecasts to determine whether the distribution of possible capacity outcomes and prediction confidence can be quantified. Thirty-five airports are examined, and results demonstrate that the model performance is improved by the changes proposed and furthermore that the forecast variation among ensemble members effectively captures the range of future airport arrival capacities. As such, the method proposed in this Paper demonstrates the viability of this generic method for use within future decision support systems in which reasonable multi-airport estimations are needed and development of detailed, airport-specific models would be prohibitive.