Suburbanisation Vs Reurbanisation: Population Distribution Changes in Australian Cities

Australia: State of the Environment Second Technical Paper Series (Human Settlements), Series 2
Emma Baker, Neil Coffee and Graeme Hugo
Department of the Environment and Heritage, 2000
ISBN 0 642 54791 2

Contents

This report was prepared by Emma Baker, Neil Coffee and Graeme Hugo of the National Key Centre for Social Applications, Geographical Information Systems (GISCA). The views it contains are not necessarily those of the Commonwealth, State or Territory Governments. The Commonwealth, which includes the Australian State of the Environment Committee, does not accept responsibility in respect of any information or advice given in relation to or as a consequence of anything contained herein. The copyright for the use of the images has been granted to the authors.

Introduction

In Australia as in most OECD nations, the last two decades have seen a focus in urban development policy on increasing the population density of cities. This has involved a range of policies encouraging urban redevelopment, the subdivision of land parcels within the built up urban area etc. It has been bolstered in the Australian context by some indications of a shift in the balance of housing preferences away from a total dominance of low density, outer suburban living toward one in which a larger proportion of the population favours living in higher density inner and middle suburban locations. The question remains however to what extent have these developments produced changes in Australian cities in

  1. The spatial patterns of population growth.
  2. The spatial pattern of population density.

In the present paper an attempt is made to utilise geographical information systems to provide answers to these questions by analysing changes in patterns of population growth and population density over the 1981-1996 period.

The Issue

The Australian Bureau of Census uses a population density criteria to distinguish urban census collectors districts (CCDs) from non-urban areas to delimit major urban areas. The low population density of Australian cities is underlined by the fact that the entire island of Java in which 50.8 percent of the population are classified as rural, has a population density of 891 persons per square km. For much of the post-war period Australia's major cities have exhibited a "doughnut" pattern of population growth whereby the inner cities have experienced population declines due to residential land uses being replaced by higher rent activities and the bulk of population growth has occurred on the peripheries. The archetypical pattern of a low-density city with growth occurring at the periphery is exemplified by the situation at the 1976 census. This was analysed fully by The Atlas of Australian Resources Volume 2: Population (Division of National Mapping 1980) and Figure 1 which is taken from this analysis indicates a pattern of low density of population declining out from the city centre and some variation in cities with the highest densities being in Sydney and Melbourne.

Figure 1: Residential Densities 1976, Large and Medium Cities

	Residential Densities 1976, Large and Medium Cities

Source: Division of National Mapping 1980, p. 14

The pattern in 1976 with respect to population change is depicted in Figure 2 and there is a clear trend of increase in population growth toward the periphery. In recent decades however along with all the OECD nations Australia has attempted to move toward more "sustainable" compact cities by changing land use policy to encourage increases in population density in built up areas (Parham and Konvitz 1996). Australian cities have been criticized as the least environmentally sustainable in the world because of their low density and reliance on motorcar ownership (Newman and Kenworthy 1992, 1999). Arguments for encouraging urban consolidation in Australia have been on the basis of reducing unnecessary suburbanisation of valuable agricultural land, reducing fossil fuel usage, reducing greenhouse gas emissions, reducing reliance on motorcars, making more efficient use of infrastructure and improving equity. As a result each major Australian city and its component level governments have been pursuing urban consolidation policies. This has not been without opposition especially by Stretton (1991) and Troy (1992, 1995, 1996) who argue that consolidation policies fly in the face of the preferences of the bulk of Australians and that the perceived benefits of consolidation are exaggerated. There has even been debate about whether or not a change is occurring in the population distribution in Australian cities as a result of these shifts in policy and this is where the present paper seeks to make a contribution.

Figure 2: Rates of Population Change 1971-76, Large and Medium Cities

	Rates of Population Change 1971-76, Large and Medium Cities

Source: Division of National Mapping 1980

Methodology

The present study seeks to analyse changes in population density and population growth patterns in Australia's largest five cities over the 1981 to 1996 periods to precisely establish the extent to which reurbanisation and suburbanisation are occurring. To do this, Geographical Information Systems are employed. GIS are complex computer based technologies and methodologies which facilitate storage, analysis, retrieval, visualisation and modelling of spatially referenced information. In order to analyse changing patterns of population density and population growth in the five cities, GIS is utilised to establish changes occurring in concentric rings around the Central Business District of each city. These concentric rings are depicted in Figure 3.

Figure 3: Australian Major Capitals, Urban and Peri-Urban Areas, 5km Concentric Distance Zones

	Australian Major Capitals, Urban and Peri-Urban Areas, 5km Concentric Distance Zones

Source: ABS Census of Population and Housing 1981, 1986, 1991 and 1996

The boundary of analysis for each of the five cities was the urban and peri-urban area as defined by McKenzie (1996). These areas were calculated using a combination of journey to work and distance from the CBD to reflect the functional extent of the city for residential uses. Population data from the 1981, 1986, 1991, and 1996 Censuses was assembled at the smallest unit of collection, the Census Collector District (CCD). Each CCD describes the characteristics of roughly 200 households, and provides the most accurate Australia-wide portrayal of population characteristics. The spatial centroid of each CCD was calculated, and the Census population data was allocated to this centroid. Five kilometre concentric rings, beginning at the CBD and continuing to each cities' outer boundary were produced using the GIS software ArcView. To these concentric rings, the centroid population data was spatially assigned. The result was five spatial surfaces representing the large Australian cities, containing population information by concentric distance from the city. This data could then be compared, manipulated and analysed.

This generalisation of population data across the major Australian cities by distance from the CBD allows population change to be compared between cities and over time. The analysis focuses on change in population number and density.