Characterization of delay propagation in the US air transportation network

TL;DR

This paper analyzes flight delay patterns in the US air transportation network, revealing robust delay distributions across various conditions and highlighting the influence of destination airports on long delays.

Contribution

It provides a detailed characterization of delay propagation and distribution in the US air network, focusing on long delays and their dependence on airport location and destination.

Findings

Delay distributions are consistent across different operational conditions.

Remote airports exhibit unique delay patterns with longer delays.

Destination airports significantly influence the occurrence of long delays.

Abstract

Complex networks provide a suitable framework to characterize air traffic. Previous works described the world air transport network as a graph where direct flights are edges and commercial airports are vertices. In this work, we focus instead on the properties of flight delays in the US air transportation network. We analyze flight performance data in 2010 and study the topological structure of the network as well as the aircraft rotation. The properties of flight delays, including the distribution of total delays, the dependence on the day of the week and the hour-by-hour evolution within each day, are characterized paying special attention to flights accumulating delays longer than 12 hours. We find that the distributions are robust to changes in takeoff or landing operations, different moments of the year or even different airports in the contiguous states. However, airports in remote areas (Hawaii, Alaska, Puerto Rico) can show peculiar distributions biased toward long delays. Additionally, we show that long delayed flights have an important dependence on the destination airport.