Geography Reference
In-Depth Information
These, then, are the widely known entry points
to a generalised understanding of human
geography. Yet as students progress through the
discipline, so such models appear increasingly
redundant and irrelevant to structuring and
representing the real world. Over time, the
regularised geometry of urban land use becomes
too strained and implausible to sustain further
investigation and interest, and more 'advanced'
urban morphology research tends to refocus on
more ideographic ('one-off') studies of the
historiographies of particular urban forms. The
consequence is that some students find themselves
discouraged from seeking to draw generalisations
about urban morphologies across space and
through time.
The quest to generalise inductively has not
always been so readily ignored. In the heyday of
applied geography in the 1970s, the concepts of
social physics
and
spatial interaction
were widely used
by geographers and planners to model the
functioning of entire urban systems (Batty 1976;
Birkin 1996; Wilson 1974). This has been part of
an adherence to the basic tradition of
rational
planning
in which urban problems and their
solutions are conceived in largely physical terms
(Batty and Longley 1997). The urban models of
the 1970s were avowedly ambitious, in that they
attempted to model entire urban systems. They
were developed around the best technologies and
data sources that were available at the time, yet in
practice this inevitably meant 'making do' with
crude and inappropriate data at coarse scales of
spatial resolution, crude specifications of flows, and
analysis on mainframe computers less powerful
than the lap-top word processor on which this
chapter was written. In short, the quality of the
data and the computer technology of the time
were not commensurate with the power of the
underlying geographical concepts and ideas that
were developed. As the 1970s progressed,
increasing numbers of geographers came to the
view that such models were gross over-
simplifications of all that was interesting about
urban systems, and as such that the early urban
models were largely irrelevant to any but the most
trite understanding of them (e.g. Sayer 1979).
Disillusionment with the approach set in, and the
quest to develop quantitative depictions of the
state and dynamics of urban areas was all but
abandoned.
The tide has since turned. In this chapter, we
will illustrate some of the ways in which the
modelling and simulation of city systems is being
reinvigorated. We will focus upon two themes:
first, the implications of the revolution that is
taking place in the creation and availability of
digital data about urban areas; and second, the
ways in which these data are being reworked into
sophisticated data models of urban systems, which
in turn can inform further quantitative analysis.
These themes are each specific out-growths of the
'geographical information systems (GIS)
revolution' in geography, specific aspects of which
are described by Heywood and by Green
elsewhere in this topic.
GIS provide a framework for the orderly input,
storage, manipulation and output of geographical
data, and the field has developed rapidly over the
last two decades. There are now many excellent
textbooks (e.g. Burrough and McDonnell 1998;
Martin 1996) that provide guides to their use in
geography. The history of GIS is in some aspects
the history of an important and extraordinarily
successful branch of applied geography (Foresman
1998; Longley
et al
. 1999). GIS today is a huge
software industry probably worth $1 billion
world-wide, with about 1.6 million users. At the
current rate of expansion, there could be 8 million
'applied geographers' worldwide using GIS by the
year 2000! Over time, precipitous falls in the real
costs of computing have made it possible to
process larger data sets, much faster than ever
before, while developments in computer graphics
and net-working have encouraged the use of
information systems in an ever-wider range of
new and novel ways. GIS are thus central to any
discussion of applied geography at the end of the
twentieth century, and they provide an ideal
environment for a reinvigorated approach to
urban modelling and simulation. This is a software
environment in which niche applications of urban
modelling are once again flourishing—as in the
fine-scale analysis of retail location problems using
