Comparing metropolitan regions

Making comparisons between metropolitan regions around the world is fashionable. Newman and Kenworthy's Cities and automobile dependency (1989) was one of the first attempts to collect data from cities around the world and make systematic comparisons (although highly criticized: see, for example, Gomez-Ibanez, 1991). Recent books (Simmond and Hack, 2000; Susteren, 2005) and even a summer 2007 exhibit at the Tate Modern continue the tradition. A further attempt, a poster prepared by Sorensen and me, has been sponsored by the Neptis Foundation, a Toronto based non-governmental organization (Sorensen and Hess, 2007). Its data and methods are described at The aim was to promote discussion, showing simple comparisons (of regional form, density, and the use of private motorized vehicles) between the built-up area of Toronto and selected metropolitan areas in Canada, the United States, Europe and Australia (Figure 1). In the course of the project major difficulties were encountered. Scheduled to take less than 2 months, the project ended up taking 2 years and required significant efforts and technical support by greatest difficulties were locating comparable data and arriving at similar definitions for how the built-up areas were to be measured and defined. These issues were not fully resolved. The transportation data, for example, came from Millennium cities database by Kenworthy and Laube (2001). This provides little information on sources or accuracy but yielded the best data that could be found for international comparisons. We had more control over the definition of built-up areas. In the United States and Australia we used the national census, which in each case defined the built-up or urbanized area of each metropolitan region based on a population density of about 4 persons per hectare for small census units (blocks in the United States) and provided rules on their contiguity. In Canada, national census data allowed us to develop a similar method. For European cities, comparable small census units were not available and the built-up area was defined using a combination of population density, based on administrative units, and land cover data for 'urban morphological zones' from the European Environmental Agency. This meant that some non-urban population was ascribed to built-up areas, thus raising density. Despite the difficulties, we believe this is some of the most careful comparative work to date and allows for informative comparisons between the selected metropolitan regions. Differences in patterns of development at the edges of the metropolitan regions are striking. The …