Taxi time variation at major U.S. Airports

Abstract

The objective of this study was to investigate aircraft taxi times at major U.S. airports. The authors of this investigation previously published a novel, statistical approach to characterizing taxi behavior. In that approach, we attempted to separate deterministic effects, such as departure queuing delays, from variations we asserted analysts could treat as noise. We produced a methodology for creating taxi time parameters for a known combination of airport, carrier, and equipment type. Analysts could use those parameters to simulate taxi times in NAS-wide models. We acknowledged that runway assignment introduced variation in taxi times that would be improper to consider noise. However, we resigned ourselves to ignoring runway assignment because runway data was not easily available. Additionally, we asserted analysts could ascertain a sufficient airport location from the aircraft's carrier and the equipment type. In this investigation, we continued those efforts. We used a similar approach to determine parameters suitable for describing taxi behavior by a particular aircraft, at a particular airport. However, in this investigation, we included some of those factors that we previously ignored. This time, we used the airport runway and the airport gate to help determine appropriate taxi time parameters. We demonstrated the benefits from using the increased information. The additional information used to determine taxi time parameters decreased the variation to the underlying taxi time distribution. However, the improved taxi time parameters yielded mix benefits regarding delay. Runway assignment proved very beneficial to improved delay calculations. However, airport gate only translated into a minimal benefit.