Toward an Empirical Estimation of Weather Impact on Airport Capacity

Abstract

The forecasting of NAS capacity in the one- to six-hour time frame is important for realtime traffic flow management planning. Airport and airspace capacity estimation is an ongoing area of research with much of the effort focused on model building. Here we report on empirical approaches to the problem of airport capacity forecasting in the presence of weather. Our results can be used to calibrate airport capacity models and to discover unmodeled influences. I. Introduction N order to use the National Airspace System (NAS) efficiently, it is important to know its capacity. Capacity limitations of both the NAS airspace and airports can cause bottlenecks in high- congestion scenarios. Planners can make efficient use of the NAS if they know the current, and more importantly the upcoming, airspace and airport capacities. But these capacities can be influenced by several factors, such as different meteorological phenomena. The precise influence of these factors on capacity is complex. Furthermore, forecasts of these factors are uncertain. Therefore, the NAS capacity often cannot be forecasted with high accuracy, and forecast uncertainties must be accounted for by planners. Not surprisingly, substantial effort has focused on the problem of understanding how these different factors influence capacity, and how forecasts of these factors can be improved. Much of this effort has produced improved models of both airspace and airport capacities under different conditions. Such modeldriven approaches provide the important relationships among NAS capacity and its different influences. There is also a need for data-driven approaches which compare forecasts with outcomes and NAS throughput with those capacity-influencing factors. These data-driven approaches can be used to calibrate the NAS capacity models and to discover unmodeled influences. Here we report on our data-driven analysis of airport capacity forecasting. We first evaluate airport weather forecast accuracies, and second, we develop new methods to deduce empirically the impact of weather on airport capacity.