Increasing the Detail of Analysis to Understand Better the Determinants of Residential Energy and Water Consumption

As the world struggles to improve the sustainability of human endeavours, a large focus has been on reducing resource consumption. The number of people living in urban areas has, for the first time, overtaken the number living in rural areas. In Australia 89% of the population is urbanised. Methods of limiting resource consumption or the waste from their use, such as CO2, have largely focused on the city. A common approach has been to make changes to the distribution and nature of activity within the city, particularly urban form. This approach stems from the 19th century initiatives when physical changes were made to cities in an attempt to improve health outcomes. Improving sustainability of our cities now lies within similar planning controls. The justification that changes to a city's form will improve sustainability is largely untested. It is often based on comparisons between cities with little regard for other differences within those cities. This research asked "How can our understanding of the determinants of residential resource consumption be improved using more detailed data?". This research examined differences in direct consumption of operational energy (electricity and gas) and water by households in residential dwellings in Canberra, Australia. Differences in dwellings and households were explored to understand the differences in consumption. The research suggests that energy and water consumption cannot be analysed using simple univariate correlations that have been used to compare aggregate consumption at the city level. This thesis presents a sequential approach to analysing the problem. It begins with novel methods of inexpensively obtaining data on the physical aspects of each dwelling. It demonstrates methods of cleaning such data. A new method of estimating occupancy rates in different types of buildings was shown to improve estimates of per capita sustainability compared with the current practice of assuming 'average' occupancy. The propagation of errors and the creation of new errors as many datasets are joined had not been previously investigated. Without a Census, it was not possible to accurately determine the size of the joined dataset and the subsequent proportion of erroneous values. A new method of quantifying this propagation of erroneous data was developed. To answer the research question new statistical methods of analysis were required. These allowed for multi-factor and non-linear effects. This research applied statistical techniques not previously used with this type of data. The sensitivity of mathematical models of the relative sustainability in transport to small changes in the assumptions about how different energy types are combined was also explored. Only after considering these problems could a robust analysis of the determinants of energy and water consumption be undertaken. The determinants of residential energy (electricity and natural gas) and water consumption were explored based on quarterly bill data for the period 2006 - 2012 and covering 29,555 dwellings. The major findings included: a) Per capita water and energy consumption was highly variable - some households used 10 times the median. b) The lowest consumption was in high density detached houses, townhouses and semi-detached houses. Very low density and very high density forms of development used the most per capita. c) The greatest predictor of household energy consumption was water consumption (hot water). d) Wealthier households had higher consumption. e) Households in newer dwellings consumed less. f) Dwellings with gas used twice as much energy as those with electricity alone. g) In high density dwellings, solar orientation did not appear to affect energy consumption but servicing of communal areas added 17% to average dwelling consumption. h) Temperature, rather than rainfall, was a major determinant of water consumption. i) Household behaviour is likely to be the most significant determinant of consumption