Abstract
Flight delays and cancellations lower the service level of an airport through congestion and confusion. This paper provides a statistical method to grasp typical coincident/defusing patterns of flight delays and cancellations, which becomes a basement for within day prospecting for airport managers. Independent Component Analysis (ICA) is applied to the flight time data of 13 months at Sendai Airport, and picked up 20 patterns of flight delay occurrence and diffusion, due tobad climate, airport congestion and other accidents.