Analyzing Double Delays at Newark Liberty International Airport

Abstract

When weather or congestion impacts the National Airspace System, multiple different Traffic Management Initiatives can be implemented, sometimes with unintended consequences. One particular perceived inequity that is commonly identified is in the interaction between Ground Delay Programs (GDPs) and time based scheduling of internal departures by the Traffic Management Advisor (TMA) (now operationally superseded by the FAA's the Time-Based Flow Management system). Internal departures under TMA scheduling can take large GDP delays, followed by large TMA scheduling delays, because they cannot easily fit into the arrival flow at the runway. In this paper we examine the causes of these 'double delays' through an analysis of arrival operations at Newark Liberty International Airport (EWR) from June to August 2010. TMA scheduling delays are found to be generally higher than TMA airborne metering delays, regardless of prior GDP delays. Depending on how the double delay is defined, between 42% and 62% of all internal departures in GDP and TMA scheduling experienced double delays in this period. A deep dive into the data reveals that contributors to double delays include upstream flights departing before their Expect Departure Clearance Times (EDCTs); differences in the rates used for setting EDCTs and TMA Scheduled Times of Arrival; differences in the arrival demand expected based on EDCTs and the arrival demand entering TMA; and shorter en route times between takeoff and entry into TMA than assumed in the calculation of flight EDCTs, all of which undermine the sequencing and spacing underlying flight EDCTs. Double delays are also found to coincide with periods in which the virtual runway arrival queue being served by a TMA is large, there are periods of high demand relative to capacity, and there are high airborne metering delays. Data mining techniques are used to confirm that each of these factors contribute to the occurrence of double delay and/or high internal departure scheduling delay across three months of data from June to August 2010. Predictors of the occurrence of double delay and high TMA scheduling delay are built using logistic regression, providing prediction accuracies of 69% and 73%, respectively.

Predicting the operational acceptance of airborne flight reroute requests using data mining

“…These results are generally expected as ensemble learning techniques such as a random forest and AdaBoost typically outperform other machine learning algorithms because the group of classifiers trained performs more accurately than any single classifier, utilizing the strengths of the individual group of classifiers while at the same time circumventing the weaknesses of the individual classifier (Kotsiantis and Pintelas, 2004). However, this is not always the case (e.g., Evans and Lee, 2016). Beyond this observation, it is difficult to identify explicitly why the random forest performs better than, e.g., AdaBoost, for the particular data studied in this paper.…”

Effects of linear holding for reducing additional flight delays without extra fuel consumption

“…While the LH strategy proved to be efficient in delay recovery, as results suggest, one premise must be noted, which is time-of-arrival control in the trajectory in order to enforce the full assigned delay at the destination, as conveyed in the concept of trajectory based operations. Otherwise, airlines may be prone to depart intentionally earlier in order to compete for the reduced (and not enforced) available arrival slots, somehow aggravating traffic congestions, as has been identified in [2], as one of the main contributors to double delays.…”

“…Statistic results obtained from five airports of arrivals suffering the most pre-departure ground holding in 2015 were shown, suggesting that the additional delays of those EDCT (Expect Departure Clearance Time) affected flights were substantially larger in four of the five airports (about two to three times on average) than for arrivals that were not subject to ground holding. At the same conference, a similar analysis of "double delay" (or "double penalty"), due to the interaction between GDPs and arrival metering (terminal scheduling delays), was presented by Evans and Lee [2], providing a deep dive into the underlying causes of those double delays and the circumstances in which they occur in real operations.…”

Integrated Demand Management (IDM) - Minimizing Unanticipated Excessive Departure Delay while Ensuring Fairness from a Traffic Management Initiative

“…Newark Liberty International Airport (EWR) was selected as a test case because EWR experiences one of the highest arrival delays in the NAS. EWR also has a varying mix of international and domestic traffic, as well as departures from various origination airports that are both near and far from EWR [1,23,24]. In addition, aircraft departing for EWR from close-in (200-300 miles) airports (e.g.…”