There are significant differences in the density, both between different cities, and within a given city. Kenworthy and Laube (2000) undertook a detailed analysis of some 40 major cities in economically advanced countries, including in:

• Australia (Adelaide, Brisbane, Canberra, Melbourne, Perth and Sydney)
• The United States (Boston, Chicago, Denver, Detroit, Houston, Los Angeles, New York, Phoenix, Portland, Sacramento, San Diego, San Francisco and Washington)
• Canada (Calgary, Montreal, Ottawa, Toronto, Vancouver and Winnipeg)
• Western Europe (Amsterdam, Brussels, Copenhagen, Frankfurt, Hamburg, London, Munich, Paris, Stockholm, Vienna and Zurich)
• Asia (Hong Kong, Singapore, Tokyo)

They developed the measure of activity density (jobs plus population per hectare), since both population and jobs tend to be key determinants of travel patterns, and examined how this varied across and within cities. Australian cities had an overall activity density of 18 across the metropolitan area, lower than the average for the US cities (22), and well below that in Canada (43), Western Europe (85) and the Advanced Asian cities (239) in the sample. Sydney's activity density was significantly higher than other Australian cities (24), broadly comparable to the US cities, but still low on a world scale. Activity densities in the inner suburbs were generally two to three times higher than the metropolitan averages (see fig. 1).

Relationship between Transit Systems and Density

The size and density of cities is closely linked with the sort of mass transit systems which tend to be adopted.

All cities have bus-based public transport at a minimum since this is the cheapest and most flexible mode. A relatively small number of cities (eg Ottawa, Brisbane) have built bus transitways to provide higher capacity and speed on key corridors.

However nearly all large cities, as well as many smaller cities, also have some form of rail-based public transport, and in many cases more than one form. The type of rail system(s) used tends to depend on the characteristics of the city:

• Larger cities tend to have surface rail systems, possibly with parts of the system underground in the central business district. This is the case for all the Australian cities (except Canberra) and most other cities above one million population.
• Higher density cities such as Tokyo, Paris, New York and Vienna also have extensive underground (metro) systems with dense networks of lines and large numbers of stations, to handle the high volumes of passenger movement. However there appears to be an effective density thresh-hold for such systems. For example, virtually all the cities in the sample with overall metropolitan densities above 50, and inner suburb densities above 100, had extensive underground metro type systems, while very few cities with densities below these thresholds had such systems (See Figure 2A and 2B).
• Both large cities (Like Paris) and small cities (like Zurich) often have light rail systems (See figures 3A and 3B). However the density threshhold is much lower than for metro style rail systems, with many cities having a metropolitan-wide activity density of less than 25, or an inner suburbs density of less than 50, having light rail systems.

  This follows from the capital cost of building surface based versus underground systems, and the capacity of the different types of systems. As shown in the table below, underground rail systems typically cost three to ten times as much to build as surface light rail systems, but typically provide around four times the capacity. Where the capacity is required, they are justified and are the most efficient solution to mass transit, however where the capacity is not required, it is difficult to justify them.

  Table 1: Capacity and Cost for Underground versus Light Rail

  Type of System	Typical Capital Cost A$ / km, double track including stations / stops	Typical maximum capacity (passengers per hour per direction in peaks, including standees)
  Light Rail	$10 - $40 million	4,000 - 8,000
  Underground Rail	$100 - $150 million	15,000 - 30,000

  
  Implications for Sydney

  The comparison with other cities indicates that Sydney is at the lower limit of where extensive underground metro rail style systems are likely to be viable, even in the future with 30% growth in population and employment density. On the other hand, Inner Sydney in particular already exceeds the densities where light rail is common overseas.

  The actual choice of mode for particular applications depends on many factors, including patronage potential; density; the length of the corridor; topography, heritage or other constraints; and urban form and potential for development. For example:

  • The total potential patronage on a mass transit system will depend on the length and density of a corridor, the extent to which the corridor serves key desire lines (such as to a Central Business District), and competition from other alternative routes or corridors. Because of their high cost, fully underground metro rail systems are generally only justified where very high patronage can be developed (15,000 passengers per hour per direction in the peak hour or above). Partially underground or surface heavy rail systems can be viable at somewhat lower patronage levels.
  • In addition, the longer the corridor the more important it is that higher speeds are achieved in order to keep travel times competitive. Thus for a 5-10 km corridor, on-street based systems such as light rail or buses which tend to achieve average speeds around 20 kph can be appropriate, while for a 20 - 40 km length corridor, some form of grade separation (either dedicated guideway, or above ground or underground installation) is necessary to achieve higher average speeds (30 kph or above).
  • Where topography is undulating, partial use of tunnels or above ground installation may be necessary because of gradients. In other cases, heritage constraints may make above ground installations undesirable, so that systems requiring grade separation will need to be underground.
  • Urban form is also important. Heavy rail and metro rail systems with stations 1- 2 km apart (or more) tend to be associated with intense development around a relatively small number of nodes examples such as Bondi Junction and Chatswood in Sydney illustrate this. By contrast light rail systems tend to have more frequent stops (typically every 400-800metres), which support linear development along the whole corridor, as evidenced in Sydney with strip shopping centres and medium density housing along the original tram routes.

  Consideration of all the above factors, together with analysis of current travel patterns suggests that, for Sydney:

  • the inner suburbs south of the harbour are highly suited to a light rail network, which would complement the existing land use pattern, provide appropriate capacity, and support transit oriented development.
  • Extension / upgrading of the suburban heavy rail system is likely to be more appropriate to the much longer corridors linking the CBD to the outer suburbs (eg north-west and south-west sectors, Central Coast and Warringah peninsula. This could however involve some tunnelling and use of metro style rollingstock.
  • For the secondary corridors linking lower density areas to secondary CBDs (eg NW sector Parramatta) busways may be the most appropriate and cost-effective mode.