Convective weather is responsible for large delays and widespread disruptions in the U.S. National Airspace System (NAS), especially during summer months when travel demand is high. This has been the motivation for Air Traffic Flow Management (ATFM) algorithms that optimize flight routes in the presence of reduced airspace and airport capacities. These models assume either the availability of reliable probabilistic weather forecasts or accurate predictions of robust routes; unfortunately, such forecasts do not currently exist. This paper adopts a data-driven approach that identifies robust routes and derives stochastic capacity forecasts from deterministic convective weather forecasts. Using techniques from machine learning and extensive data sets of forecast and observed convective weather, the proposed approach classifies routes that are likely to be viable in reality. The resultant model for route robustness can also be mapped into probabilistic airspace capacity forecasts.
Keywords: Air traffic management, Convective weather, Weather-ATM integration, route robustness
/ Other authors: