Background

Christmas Island is an Australian Territory located northwest of Western Australia and administered by the Commonwealth Government. The Island has an unusual mix of visitors attracted to a rich marine life and an unusual ecosystem. There is an additional seasonal demand from tourists to see the annual red crab migration.

Recently, the island attracted intense media attention as asylum seekers became a point of public concern. The presence of the asylum seekers has stimulated the Christmas Island economy, with guards being located on the island. The other important event currently for the island is the proposal to construct a space centre by Asia Pacific Space Centre (APSC), which will involve six to eight satellite launches per year and the refurbishment of the Casino to provide accommodation for visitors.

Westralia Airports Corporation has the contract to manage and develop the Christmas Island airport. The airport is unique, as it is regarded as an international airport for all passengers, including Australians, and has restrictions associated with this. (Another feature is the location of the airport - on top of a hill - a challenging runway for pilots.)

The Problem

Data Analysis Australia and Dr Paul McLeod of Economic Research Associates prepared the previous forecasts of passenger numbers on the island in 1995. Westralia Airports Corporation required an investigation of how major events on the island since the previous forecasts would affect passenger numbers in the Airport Master Plan.

The radical changes in the island economy required an assessment of the impact on the airport, and an update of the model incorporating new statistical and qualitative information.

The Data Analysis Australia Approach

The approach of the consultants was to conduct an intensive study of the current and prospective economy of the island on a week's study tour, including consultations with all important stakeholder groups. This qualitative information was combined with quantitative data to produce estimates of the future economic activity on the island. The mix of skills in the project team was essential to balance the statistical understanding and economic knowledge required.

With such a small airport, an individual event such as a tourism development can have a significant impact on the demand for runway time and space and resources. Unlike a larger airport, which can rely on average numbers for forecasting purposes, one event can strain the capacity of a small airport if the contingency is not properly planned for. In this environment, the most appropriate method of quantifying 'spikes' in demand is to assign probabilities.

Probabilities were assigned to each event or activity, as well as distributions for their possible size and timing and the previous model was altered to incorporate the different outlook for the island. The simulation program @RISK was again used for the modelling. This program uses Excel as a base, but extends it's capabilities to include probability distributions in models. It then runs simulations to give forecast distributions.

This flexibility was particularly important for this project as the model had to be able to incorporate different scenarios and probability distributions relatively easily (ie, an event might have a probability of 80% of occurring, with between 200 and 300 passengers assigned to that event). In this way, the uncertainty around future events occurring could be incorporated into the model. 

The forecasts needed to consider reasonable upper limits on demand for the long term component of the master plan as well as shorter term reasonable values that are used for immediate construction projects. The idea of a good master plan is that it should not be necessary to move or vacate a facility because it is too small (and can't be extended) or in the wrong position. In this event, the results of @RISK produced upper, lower and median numbers for each scenario.

This forecasting model was different to the previous model, which almost exclusively related passenger numbers to the casino. The probabilities of the casino staying closed or reopening on a small/large scale had to be incorporated. These were the three major scenarios developed as a result of the consultations.

The Result

A master plan must be able to cope not only with what will probably happen, but also with what might happen. Some facilities may not be immediately constructed, but the plan must set aside appropriate space and resources for the eventuality that the demand does crystalise. The estimates produced by the model showed the best, worst, and median case scenarios for passenger numbers for the next twenty years. The model was able to incorporate the unique mix of features on the island that attract different types of visitors (events such as the annual crab migration or the impact of asylum seekers).

The model then was used to forecast the numbers of aircraft for the next twenty years for the airport master plan.