Modeling Seasonal Variation in Tourism Flows with Climate Variables

Abstract

The purpose of this article is to model and forecast the seasonal variation, the fluctuations in tourist numbers from season to season in Australian inbound holiday tourism, using climate variables such as maximum temperature, humidity, and hours of sunshine. For estimation purposes this study uses quarterly data on arrivals from the US, UK, Japan, and New Zealand to Australia from September 1975 to September 2009. Seasonal variation, which is the respective and predictable movement of visitation around the trend line, was first extracted from the quarterly holiday tourist arrivals time-series using the Basic Structural Model (BSM) approach. Subsequently, the influence of climate variables on seasonal variation in different seasons was identified using the average euclidean minimum distance (AD) statistics. The AD statistics show that climate variables shape the characteristic of seasonal variation of tourism flows but the effect tends to vary between seasons and countries. A time-series model was estimated with climate variables to forecast seasonal variation. The forecasting comparison result shows that climate variables improve the forecasting performance. The approach can be replicated to help destination managers and forecasters determine if climate variables influence tourism flows between other origins and destinations globally.