A demand trend change early warning forecast model for the city of São Paulo multi-airport system

• Early warning forecast models are based on the combination of leading indicators and alarms.
• Variations on domestic number of air passengers are mostly related to changes on the economic activity.
• Two scenarios are related to a high growth rate and just one to a low growth rate.
• A trend change early warning forecast model is helpful not only to anticipate changes.

The need of accurate forecasts of air passenger numbers to assist managerial decision making for both short and long terms is well recognized and a central problem on both short and long term forecasting is how to handle future trend. The aim of this paper is to develop a demand trend change early warning forecast model (EWFM) for the city of São Paulo multi-airport system (SPMARs). For SPMARs the EWFM is based on the combination of leading indicators and alarms against possible occurrence of changes on trend component of the monthly number of domestic air passengers. A topdown induction procedure is employed to identify leading indicators to provide an interpretable prediction procedure to support the development of scenarios for future demand trend. Results show that changes on such demand trend are mostly associated to changes on the economic activity and six different scenarios were built combining the identified leading indicators. The EWFM was employed to assist managerial decision making for both short and long terms in order to evaluate different alternatives to prevent congestion delay occurrences and to support infrastructure planning.