Multifractal detrended fluctuation analysis on air traffic flow time series: A single airport case

Abstract

Exploring multifractal characteristics of air traffic flow time series is helpful in understanding the self-similarity and the correlations embedded in the time series, and thus in obtaining insights into the evolution mechanism and the regular patterns of the air traffic flows, which may help to develop effective air traffic flow management measures. With the multifractal detrended fluctuation analysis method, we identify that the total, the arrival and the departure air traffic flow time series of Beijing Capital International Airport in the summer season of 2017 are of multifractality on the scales below the corresponding crossovers, and the primary cause of the multifractality is the long-range correlations of small and large fluctuations. Comparisons on the multifractality of the time series indicate that the total and the arrival air traffic flows are of the strongest and the weakest multifractality respectively, and that of departure air traffic flow is in-between. The comparison results also show that the total and the arrival air traffic flows are insensitive to large fluctuations and dominated by small fluctuations, whereas the departure air traffic flow is insensitive to small fluctuations and dominated by large fluctuations. In addition, an investigation of multifractal characteristics of the time series during the thunderstorm season and the non-thunderstorm season reveals that the impact of the thunderstorm season on the total air traffic flow is the strongest, and there are significantly essential differences in the multifractality of the total air traffic flow before and after the thunderstorm season. For the arrival air traffic flow, there is only a difference in the extreme fluctuation rate, whereas there is no difference in essence for the departure air traffic flow, except for some quantity differences.