Supporting the lifestyles of the populations of modem cities requires vast quantities of natural resources and leads to environmental stresses such as air and water pollution. Research into the metabolism of cities therefore aims at understanding the physical flows into, within, and out of cities with a view to reducing the use of resources and the environmental impacts. One important physical indicator is energy use. Most studies on cities only consider direct or end-user energy consumption. Since the function of cities is to serve the lives of their residents, indirect energy use in cities, or energy embodied in the consumption of goods and services by its residents, can be regarded as being as important as direct energy use. However, physical models of cities are extremely complex and have difficulty in dealing with boundary issues, and hence the indirect resource requirements. In this paper we use input-output analysis and detailed household expenditure data to yield comprehensive energy use breakdowns for the 14 Statistical Subdivisions of Sydney. Multivariate regression and structural path analysis (SPA) are used to interpret the results. Clear correlations can be drawn between energy use and income, household size, age, and degree of urbanity. The structural path analysis is used to demonstrate how significant differences in lifestyles between inner and outer areas of Sydney leads to different energy use characteristics. This final consumption-based approach to analysing the energy requirements for households has important implications for measures aimed at reducing energy use, which tend to only consider direct energy use. (C) 2004 Elsevier B.V. All rights reserved.
