Forecasting air transportation demand and its impacts on energy consumption and emission

Abstract

With the increasing demand of passenger and freight air transportation and their key role in energy consumptions, this study developed a hybrid framework integrating machine learning algorithms to predict passenger and freight demand in air transportation and analyzed the impacts of the air transportation demand on energy consumption and CO2 emission. Critical variables are identified through correlation analyses between air transportation characteristics and economic, social, and environmental components, which are further selected to predict the air passenger and freight transportation demand. A hybrid framework is then developed where pairwise machine learning algorithms are developed to enhance prediction accuracy with an optimization model. Furthermore, sensitivity analyses are conducted to assess the effects of passenger and freight demand on energy consumption and CO2 emission. The framework is employed using Canadian air transportation as a case study. The pairwise machine learning algorithms are compared to single algorithm which increases prediction accuracy by 24.5% for passenger demand and by 25.85% for freight demand. It is predicted that the air passenger will increase by 94% to 94.25% and freight transportation demand will increase by 28.12% to 33.97% in 2050 relative to 2019 level. This surge can raise energy consumption and CO2 emission by 8.91% to 9.46% and 11.32% to 12.12% respectively. A 50% decrease in both air passenger and freight transport demand will result in a 5.43% reduction in energy consumption in Canada's transportation sector. Conversely, if demand increases by 50%, CO2 emissions will rise by 3.42% for air passenger transport and 3.67% for air freight transport. The predicted air transportation demand growth and trends can be used for capacity development and emission mitigations for sustainable planning and controls in passenger and freight demand in air transportation.