218Urbanization models and the regions
results vary from attribute to attribute (AppendixII also shows variability from
region to region) as expected in a relatively comprehensive analysis. Neverthe-
less, the totalnumberof 1 s(best models) suggests that the satellite-cities and
concentric-zones models are better than the other two (12 and 5 versus 1 and
0, respectively) (Table 8.2). The overall averages furthermore suggest that the
satellite cities approach is the optimum (2.00 vs. 2.44, 2.64, and 2.92).
Analogously the dispersed sites approach is apparently the worst for urban-
ization (no best-model 1 s, six worst-model 4 s, and the highest overall average,
2.92) (Table 8.2). Of the two intermediate cases, the concentric-zones model (five
best-model 1 s, two worst-model 4 s, and overall average 2.44) is noticeably better
than the transportation-corridors model (one best-model 1 ,sevenworst-model
4 s, and overall average 2.64). In essence, this broad analysis points to a rather
clear ordering of the urbanization models from best to worst: (1) satellite cities;
(2) concentric zones; (3) transportation corridors; and (4) dispersed sites.
These results suggest two guiding principles:
(1) Regional urbanization in dispersed sites surrounding a metropolitan
area, and to a lesser extent along transportation corridors, appears
tocause extensive nature-and-human resource degradation, and thus
should be avoided or minimized.
(2) Urbanization focused around satellite cities, which causes the least over-
all resource degradation, appears to be the best regional development
pattern, though factors specific to a region may indicate a preference for
combining satellite-city development with concentric-zone development
adjacent to a metropolitan area.
While the worldwide patterns provide a framework or foundation for under-
standing and planning, differences from the central pattern can be detected
forcertain cities. Based on the distribution of best and worst urbanization
models (sums of 1 sand 4 sfor each region in Table 8.2), no urban region is
strongly at variance with the worldwide pattern. Half of the regions closely fit
thebroad pattern (Barcelona, Cairo, Canberra, East London, Edmonton, Erzurum,
Kuala Lumpur, Moscow, Nairobi, Portland, Rahimyar Khan, Samarinda, Santiago,
Sapporo, Seoul, Tegucigalpa, Tehran, and Ulaanbaatar).
Tworegions (Chicago, Bucharest) only vary from the core pattern by having
slightly more 1 sfor the concentric-zones than the satellite-cities model, though
this isnot inconsistent with the second guiding principle above. Four regions
(Mexico City, Nantes, London, Brasilia) diverge from the central pattern by hav-
ing slightly more 4 sintheconcentric and satellite category than in the trans-
portation and dispersed category (though 1 sareconcentrated in the concentric
and satellite category, as expected). Philadelphia perhaps diverges most from the