Selecting cities, determining boundaries, mapping regions 115(2) Good- or adequate-quality satellite image of urban region available at
time of city selection (thus Bogota, Lagos, Xi’an not considered).
(3) Formof metropolitan area seemingly can be extrapolated to many other
cities (thus Johannesburg, Naples, New York not considered).
(4) Considerable useful ecological and planning information on the region
readily available (thus Berlin, Chicago, Ottawa included).
(5) Author familiarity with the urban region through residency, work, or
travel.Population of a city rather than of the metropolitan area or urban region was
used as the indicator of size, because somewhat consistent and reliable data are
available for all of the cities (Turner2005).Population data for the cities chosen
were for 1999--2002, with four exceptions: Iquitos (Peru) was 334 000 in 1998,
Abeche (Chad) 188 000 in 1993, Samarinda (Indonesia) 335 000 in 1990, Rahimyar
Khan (Pakistan) 234 000 in 1998. All were adjusted to the year 2000 using a 5 %
annual growth rate. For convenience in comparing major worldwide cities, small
cities have 0.25 to 1 million, medium cities 1 to 4 million, and large cities 4 to
16 million inhabitants. Satellite cities located around a major city are smaller,
with <250 000 inhabitants.
Geographic areawasused instead of continent to better capture the combi-
nation of location and culture, and so that each geographic area included a
similar number (five to eight) of urban regions studied (Table 5.1). The six geo-
graphic areas were: Europe, North America, Latin America, Africa, West to East
Asia [Moscow and Erzurum (Turkey) to Sapporo], and South Asia to Australia
[Rahimyar Khan (Pakistan) to Samarinda (Indonesia) and Canberra].
Ihavelived in or near nine of the urban regions (Atlanta, Barcelona, Canberra,
Chicago, London, Mexico City, Ottawa, Philadelphia, Tegucigalpa), and visited
another ten (Bangkok, Kuala Lumpur, Beijing, Portland (Oregon), Edmonton, San
Diego, Stockholm, Bucharest, Rome, Berlin). This general familiarity helped espe-
cially in understanding the urban region beyond the city. However, I have never
seen half of the 38 regions analyzed, which highlights the importance of the
three primary analytic approaches used: measuring remotely sensed images, lit-
erature survey, and consultations with knowledgeable persons.
An initial set of cities was selected because they are well known in the lit-
erature for illustrating key planning and ecological patterns, or I had accu-
mulated especially useful ecological information on them (Barcelona, Berlin,
Stockholm, San Diego, Portland, Chicago, Atlanta, Ottawa, Mexico City, Brasilia,
Canberra, Beijing). Large satellite images were printed for 120 urban regions
under consideration. From these, several additional distinctive, but widely appli-
cable, patterns stood out, and those cities were selected accordingly (Santiago,