Land-change patterns and models 205of the mosaic sequence (e.g., from 70 % to 20 % of the initial more-suitable land
type), instead of progressively shrinking a single huge natural patch, a few large
natural patches are created (to spread risk), which are then sequentially rather
than synchronously shrunk and eliminated. Thus the overall jaws-and-chunks
mosaic sequence resembles wide-open jaws progressively moving from a corner
and consuming more-suitable land, then creating and sequentially removing
chunks of suitable land in the middle phase, and finally consuming the last
chunk and the scattered bits of suitable land in the last phase.
Attributes for evaluating patterns and models
Change is the norm, a process illustrated by sunrise, spring, the first
snowfall, and a warthog consumed by a crocodile. Urbanization, like the
warthog--crocodile transformation, has advantages and disadvantages. A town
in an urban region becomes jammed with traffic and loses its rural character,
yetconcurrently road access, cultural diversity, and economic opportunity grow.
Awetland that cleans streams of pollutants and maintains rich biodiversity is
drained, yet that loss eliminates malaria-carrying mosquitoes and reduces prob-
lems related to soggy soils in adjoining neighborhoods.
Thus evaluating an urbanization pattern, or model of it, requires both iden-
tifying the most important variables affected, and balancing the pros and cons.
The other key decision to make in choosing variables is how general or specific
theyshould be. Specific detailed variables are more numerous and individually
may provide more precise response patterns. But determining the right variables
to measure and interpreting large numbers of dissimilar specific responses nor-
mally are extremely difficult. Using broader indices that combine or integrate
specific ones reduces the complexity of detail and thus may lead to discovering
broad patterns. However, teasing apart the detailed causes or mechanisms may
be difficult. Nevertheless, a lucid big picture, as well as major variables and han-
dles for planning, will probably never appear with a detailed-variables approach.
Broader variables or indices are required for that.
The process used in identifying ecologically optimum landscape change (see
previous section) appears useful for analyzing urbanization. Rather than com-
paring the five mosaic sequences of landscape change using numerous specific
ecological characteristics, such as species richness, stream structure, groundwa-
terquality, interior species populations, and so forth, broader spatial attributes
were measured for evaluating landscape change (Forman1995). The ten spa-
tial attributes chosen included total patch-interior area, grain size of the land-
scape, ability to cross the landscape, total boundary length, and patch den-
sity. Conveniently the spatial attributes fell into three broad useful categories:
(1) patch size or landscape grain size; (2) connectivity; and (3) boundary length.