Large-scale Analysis of Trajectory Interaction Networks in Europe

A challenging and open topic in air traffic management is the understanding of how aircraft interact between them to avoid separation losses, consequently creating downstream effects that go well beyond single pairs. We here present a concise modelling technique of such interactions and of their propagations, based on complex network representations of the same. By relying on some basic hypotheses, we show how this method can efficiently scale and be used to represent all flights crossing the European airspace in one day. We then use this methodology to study the resulting structures in three case studies: the normal dynamics in Europe, the impact of COVID-19 on it, and an hypothetical scenario involving free flight trajectories. Among other results, we show how these structures are a function of the main traffic flows, but not of the individual flights, and are hence an emergent property of the system; how COVID-19 impacted traffic beyond what expected from a simple reduction in traffic; and how geodesic trajectories actually dampen the propagation of interactions. We further discuss open questions and possible venues for future research.