Geography Reference
In-Depth Information
geodemographics and other detailed sources as
described by Clarke in this topic.
Figure 44.1
The trade-off between model complexity and
fidelity to real-world representation.
REINVENTING URBAN MODELLING
AND SIMULATION
There is not space here to do justice to the
development of urban modelling within GIS,
although a wide-ranging overview is made by the
various contributions in Longley and Batty (1996).
These provide convincing evidence that GIS are
reinvig-orating our thinking about urban
modelling in an applied way, by virtue of the
increasingly data-rich depictions of reality that
they facilitate. An inherent characteristic of urban
modelling is that as a model becomes more
complex, so it more closely resembles the real-
world system that it seeks to approximate. We can
imagine 'complexity' here to be a composite
index of the amount and detail of data about
urban systems, and the nature of the way in which
relationships between the components of the
urban system are simplified. It has long been
recognised (Haggett 1978) that the rewards to
increasing complexity tail off, in the kind of way
illustrated in Figure 44.1, and the innovation of
GIS has undoubtedly made it possible to move
further up this 'learning curve' than hitherto. This
is because much of the complexity of urban
modelling arises out of the increased burden of
integrating, managing and processing data, and the
environment of GIS has begun to trivialise these
problems. In this way, new, firmer and more
substantial foundations to urban modelling are
being established. We will develop this idea by
developing two extended illustrative case
studies—the first in the context of predicting the
values of housing across an urban area, and the
second revisiting some of the classic concepts
underpinning micro-economic models of housing
demand. It is not possible in the space available
here to do more than review the principles and
methods used in urban modelling, although the
guide to further reading contains further details of
the ways in which the case study models can and
have been applied. A shared theme between our
two examples is that modelling (and by extension,
simulation) within GIS can reinvigorate the ways
in which we think about generalising the
characteristics of urban systems across space and
time.
These developments are timely, because urban
theory itself is being forced to change. The
demise of the 'first-generation' urban models was
also in no small part because the assumptions of
single centred cities (focused upon a dominant
central business district area, often dominated by
manufacturing industry, etc.) were increasingly
strained by the rapid changes in the structuring
of contemporary urban areas. Yet the
consequence, the relative demise of urban
modelling as an area of applied geography, has
also led to a slackening in the pursuit of the
rational planning goals that have dominated the
post-war era. It remains a largely uncontested
tenet of urban geography that the notions of
economic bid rents, urban hierarchies and such-
like remain of enduring relevance to the subject,
and they also provide a means of comparing
different urban systems. As cities have developed
and become more complex, varied and
specialised, so the old certainties about urban
structure are being swept away and new
approaches to the measurement of urban
structure and form must be developed. Batty and
Longley (1994) have written on this broader
theme at length, but here we will restrict our
