Improving Air Traffic Management During Thunderstorms

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner
Thursday, 2 February 2006: 2:00 PM
Improving Air Traffic Management During Thunderstorms
A301 (Georgia World Congress Center)
Mark E. Weber, MIT, Lexington, MA; and J. Evans, M. Wolfson, R. DeLaura, B. Moser, B. Martin, J. Welch, J. Andrews, and D. Bertsimas

Presentation PDF (455.6 kB)

Thunderstorms are the only weather phenomena that significantly affect both terminal and high altitude en route airspace capacity. Accurate forecasts of thunderstorms many hours in advance cannot currently be provided and, as a result, conservative air traffic management procedures are employed to minimize the possibility that aircraft will encounter thunderstorms along their route of flight. Based on FAA delay statistics, we estimate that thunderstorm related flight delays currently cost the commercial airline industry $2B annually in direct operating expenses. Aviation planners have called for two- to three-fold increases in the capacity of the air transportation system in the next twenty years. If this is to be maintained during thunderstorm outbreaks, significant improvements to the technologies and procedures used to cope with thunderstorm impacts on airspace capacity must be developed.

This paper discusses efforts underway that address ATM system needs for improved handling of thunderstorm related capacity constraints.

(i) Automated forecasts of the position, intensity, movement, height and trend (growing/decaying) of thunderstorms must be generated at fine time steps that span the zero to eight hour time window of interest for flight planning and development of air traffic management initiatives. These forecasts must include parameters that characterize their time-varying uncertainty in a way that allows for a corresponding estimate of the uncertainties in the capacity of the airspace they impact. The Regional Convective Weather Forecast [1] in use at major ATC facilities in the northeastern US is an example of an operational forecast product that is evolving in this direction.

(ii) The thunderstorm forecasts must be “translated” into time-varying estimates of the capacity reductions in affected en route sectors, terminal airspace and airports. These estimates must include uncertainty bounds for the future capacities. Initial work in this area has been accomplished under NASA support [2].

(iii) “Optimal” ATM strategies must be automatically developed that take into account the time varying capacity estimates to reroute traffic around thunderstorm-impacted airspace or, when necessary, impose minimally disruptive ground or airborne delay programs. An efficient optimization formulation has been developed for the situation when future airspace capacity is known exactly [3]. Work is commencing to extend this formulation to provide robust solutions when future capacities are not known with certainty [see 4].

(iv) The capacity of airspace not affected by adverse weather must be increased substantially relative to current limits so that planes rerouted away from thunderstorms can be accommodated. New approaches to aircraft separation assurance are being investigated [5] that may overcome current controller workload constraints and safely allow for a significantly higher density of aircraft, particularly in en route airspace.

In combination with methods for accurately forecasting airspace demand over multi-hour periods, the capabilities listed above provide a basis for area-wide flight planning that might substantially reduce the disruptive effects of thunderstorms.