Major Air Traffic Flow Identification with Fractal-Based Graph Simplification

Major air traffic flows are concentrated streams of flights that follow similar trajectories between specific geographical regions or airport pairs, and they play a key role in tasks such as workload evaluation and demand–capacity balancing. How- ever, real-world flows exhibit strong interconnectivity, including merging, overlapping, and splitting, which makes flow boundaries difficult to distinguish. This complexity limits traditional methods based on trajectory similarity or predefined origin–destination pairs, which rely on a top-down perspective to directly extract flow segments from the global traffic structure. This paper reconceptualizes major flow identification as a bottom-up process. Major flows are identified by deriving and consolidating local flow structures, thereby capturing flow interconnectivity to offer a more interpretable and robust representation. First, flow- shaping areas are detected through density-based clustering of trajectory waypoints, capturing potential origins, terminations, and transit points. Second, local flow trees are constructed through fractal-based simplifications to preserve the essential flow hierarchy while smoothing small-scale deviations. Third, a stability metric is introduced to extract major flows that exhibit both high traffic volume and spatial persistence. Finally, an optimization framework consolidates these local major flows into a coherent set of global major flows across the airspace. The methodology is tested on northwest–southwest traffic in European airspace with flight plan data on 14th July 2023. Results show that the majority of flights can be organized into major flows, yielding stable representations of dominant traffic corridors. In addition, the framework highlights alternative routing options and abnormal trajectory patterns, supporting applications in traffic flow management, re-routing, and anomaly detection.