Land-change patterns and models 203thecauses, mechanisms, and processes of urbanization. An example is the set of
spatial processes,i.e., perforation, dissection, fragmentation, shrinkage, and attri-
tion/disappearance, that act on the existing land mosaic to produce changing
landscape patterns (Forman1995,McIntyre and Hobbs1999,Lindenmayer and
Fischer2006). In contrast, a prime interest here is to provide a foundation and
framework to understand the implications of different urbanization patterns.
This will facilitate analyses of urban regions worldwide later in the chapter.
Furthermore, we will attempt to identify the optimum urbanization pattern for
people and nature.
Concentric zones around a population center are the usual starting point for
modeling land-use pattern and change (Christaller1933,Losch1954,Covich 1976 ,
Haggettet al.1977,Antrop2000), an approach based on central-place theory and
its different-width zones of influence. Population centers or nodes are intercon-
nected by transportation routes and a hierarchy of nodes (village to city) and
routes (local roads to major highways) develops. Competition for space among
population nodes produces a regular pattern sometimes reminiscent of, and
modeled as, a multi-scale hierarchy of hexagons.
Elaborations and competing models have inevitably evolved (De Blij1977,
Formanand Godron1986), including: a sector model of land-uses organized like
wedge-shaped pieces of a pie; a multi-nodal model with separated growth nodes
superimposed on sectors or concentric zones; and multiple star-shaped nodes in
ahierarchical pattern produced by heterogeneity of the natural environment,
regional history, and communication networks (Antrop2000). A ‘‘rotating-sector
model” where land uses rotate or alternate in pie-shaped slices around a point
mimics the open-field landscape in Europe’s Iron Age (Orwin and Orwin1967,
Rackham1980,Formanand Godron1986), as well as a sequence for forest cutting
(Harris1984,Peterkenetal.1992). Many other spatial models of change have been
used in landscape ecology (Baker1989,Sklarand Costanza1990,Zonneveld and
Forman 1990, Mladenoff2005,Verboom and Wamelink2005).
Several early types of urban-growth models, some mathematical, link trans-
portation and land use (Formanet al.2003,Berling-Wolff and Wu2004): (1) grav-
ity models for interaction between cities; (2) location-of-work models; (3) journey-
to-work models; and (4) 1960s transportation models for Detroit and Chicago,
still used in the USA to evaluate potential regional air-pollution effects. Newer
approaches offer a richness of ways to think about changing land (Berling-Wolff
and Wu2004,Wiens and Moss2005), though few have been directly used to
model urbanization: (a) dynamic modeling techniques; (b) cellular automata;
(c) spatial-statistics models; (d) GIS and visualization techniques; (e) ecologi-
cal process models; (f) fractals (Milne1991a, 1991b); (g) ecological energetics;
(h) fuzzy-logic theory; and (i) neural-network theory. Another approach used in