Long-term Traffic Forecasts and Operating Pattern for a European Airport

Long-term Traffic Forecasts and Operating Pattern for a European Airport


N P S Dennis, University of Westminster, UK


This paper presents a methodology for forecasting new routes and frequencies for an airport and allocating them to airlines and schedule slots. A simulated pattern of operations is created and the output for Aberdeen Airport presented as an example.


A common problem facing airport planners and policy makers is to estimate the detail of the flight schedule that will be operated at an airport in future years. This is important to assess issues such as runway, apron, terminal and airspace capacity requirements, noise and emission outputs as well as connectivity with other flights and surface modes of transport. For airline commercial analysts it is important to understand the competitive pressures and opportunities that will develop in different markets.

This paper uses an approach adopted for the VANTAGE project, conducted for the UK Department of Trade and Industry during the period 2005-07 to forecast the detail of airline operations at UK regional airports in 2015.

The paper considers the initial production of a generalised forecast at the route level (based on historic CAA traffic data) and how this can then be disaggregated into detail such as aircraft size, airline and flight schedule. The experience of UK regional airport network development over the period 1995-2005 has been used as a basis for the addition of new destinations and the growth of frequencies. Separate models have been developed for domestic routes and European routes (with Ireland, Channel Islands etc included with the domestic market).

The existing routes and frequencies are used as a basis for building the future service pattern using a distance-based weighting algorithm. The detailed history of service to each international destination from the UK is used to model the future service. For example, a destination such as Malaga will be served primarily by UK airlines with a strong seasonal bias towards the summer while a destination such as Dubai will be served primarily by Emirates with services clustered at certain times of the day to feed their hub. The extra frequencies created by the traffic model are then allocated to airlines and time windows (of one hour) using the D?Hondt method of highest averages. This simulates the schedule effectively by clustering flights at the most popular times while generating some services at other times that currently see operations. Separate models are used for Monday-Friday flights, Saturday and Sunday. A cleaned database of over 10 500 flights is used to calibrate the model. An exact schedule to a division of 5 minute intervals is then created, with a conflict resolution procedure.

An example peak week schedule for 2015 is presented for Aberdeen airport accompanied by a commentary on the key strategic and policy implications for the industry.


Association for European Transport