Urbanization options evaluated 215These apparent mismatches between metro-area and urban region may be
harbingers of challenges ahead for the three cities.
Forboth of the case studies, Mexico City and Nantes, biodiversity sites in the
region are most affected or degraded in the dispersed-sites model (Table 8.1). They
are least covered by development in the satellite-cities model. Thus the satellite-
cities model is deemed best, and dispersed-sites model worst, for biodiversity
sites. Results for rivers/streams are different than those for biodiversity sites, and
indeed effects on rivers/streams differ between regions. For Nantes the effects on
biodiversity sites and rivers/streams are quite similar (Table 8.1). Based on these
two variables, in this urban region the satellite and transportation models are
better, and the dispersed and concentric ones worse strategies for urbanization.
The reasons for some of these results are revealed in characteristics on the
land. Differences for biodiversity-site effects seem related to a concentration of
sites by water close to the Nantes metro area, and a scarcity of sites in the mid
and outer agricultural portions of the Mexico City Region (Color Figures 22 and
25). Differences in river/stream effects appear related to the location of Nantes
at the confluence of rivers, in contrast to the lowered water-tables and dried-out
river beds near Mexico City. Indeed cities began where dispersed natural systems
were accessible in a day’s walk, many now being threatened by urbanization.
In brief, the case studies have detected intriguing patterns in response to the
alternative spatial models. Changing geometry is a central theme for understand-
ing urbanization, but also specific physiographic and land-use patterns may help
explain variations in response pattern. Yet how representative are the results
based on two extremely different cities and two dissimilar variables? Broader
response patterns at a global scale may transcend these detailed case studies.
Urbanization options evaluated with 18 attributes and 38 regions
Several land-cover types and >40 site types represented by pictograms
are mapped on the 38 urban regions (Chapter5 and Color Figures2--39). From
these, 18 diverse, particularly informative attributes are selected for evaluating
thefour urbanization models applied to all the regions. Most of the other
promising attributes are excluded because of limited sample sizes. Many of the
attributes chosen are direct measures of suitable conditions for natural systems
or their human uses (e.g., biodiversity sites, recreation sites, flood hazard).
Others are included as indirect measures known to correlate with direct factors
(e.g., drainage area protection, development on nearby slopes, distance from
major highway).
The big picture emerges by examining a summary of the best-to-worst urban-
ization models for the 18 attributes of 38 urban regions (Table 8.2). Certainly