From my colleague Bob Poole:
Much of the debate about urban land use and mass transit is driven by misconceptions. All too many elected officials (and a declining fraction of transportation planners) still have the mental model of the mono-centric city–a single huge downtown where the majority of jobs are, able to be served effectively by radial transit lines from the suburbe. Think New York and Chicago–or more accurately Manhattan and the Loop.
One of the first popular challenges to this mono-centric view was Joel Garreau's 1991 book Edge City, which identified the growing phenomenon of large-scale office and retail developments in the suburbs. More recently, Robert Lang of Virginia Tech provided a more sophisticated look with his 2003 book Edgeless Cities. Lang found that a large and growing fraction of commercial development actually exists in smaller and more scattered forms across the whole urbanized area. And that pattern, of course, makes transit (and especially rail transit) a far more difficult proposition.
Lang and two colleagues recently published (via the National Center for Real Estate Research) a further analysis, which I commend to your attention. "Beyond Edgeless Cities: Office Geography in the New Metropolis" presents the results of analyzing 13 large urban office markets, using a geographical information system (GIS). They describe what they call the new suburban metropolis, which is urban in function but not in form. "Many suburbs now have essentially all the elements that make a place urban," but arrayed in a form that differs considerably from the mono-centric model. The new metropolis "is mostly low-to-mid density, automobile dependent, and dispersed."
The new analysis finds that edgeless cities account for almost 40% of the total office space in those areas, while their downtowns averaged 33% and their edge cities just 14%. The balance fell into "urban envelopes" (5.2%), "corridors" (3.8%), and "secondary downtowns" (1.2%).
The most fascinating aspect of the report is not the averages but the differences among these large urban areas. Atlanta and Miami have the smallest fraction of office space downtown, at just 6.7% and 8.7% respectively. The edge-city champions are Houston (33.3% edge-city space), Detroit (27.1%), and Washington, DC (23.3%). And the metro areas with the largest fraction of their office space in edgeless cities are Miami (72.1%), Detroit (54.1%), Philadelphia (54.3%), and Denver (50.8%).
The final sections of the report discuss policy implications. Citing a 1977 study by Pushkarev and Zupan which found that 8,000 people per square mile was the minimum threshold for rail transit, Lang and colleagues parsed their data to determine the amount of edgeless city office space located in neighborhoods with at least that much density. Atlanta had zero, and another seven had only single-digit percentages. The most promising were Los Angeles (36.7%), San Francisco (27.8%), and Miami (13.8%). Those portions of those metro areas were suggested as feasible candidates for rail transit–and all three of them do have various rail projects in being and others under way.
Note: the Boris Pushkarev and Jeffrey Zupan book is Public Transportation and Land Use Policy: Indiana University Press, 1977.