The aim of WP3 is to identify urban sprawl and to quantify the effects
of urban sprawl by statistical methods. In addition to existing measures,
a new measure for the quantification of urban sprawl, the H-measure, is
introduced and tested.
The methodology is based on an improved shift-share framework, which can
be applied to a nested spatial system, without searching for a comparison
region. The framework is applied to the six case cities of the SCATTER
project : Bristol, Brussels, Helsinki, Stuttgart, Milan and Rennes.
WP3 analyses, for each case city, the features and effects of urban sprawl,
by performing a statistical analysis of time series of socio-economic
data, concerning a period characterised by a sprawl of the urban activities
(households, jobs, retail shops and other services). When possible, the
related effects on the pattern of trip demand are also analysed.
The statistical analysis comprises 2 stages:
The design of the common analysis framework starts by selecting a common
definition for the metropolitan area (i.e. the functional urban region)
and for structural rings inside the metropolitan area, such as: the urban
centre, the outer urban ring and the rest, named as hinterland.
The issues to be considered are related to the aim of identifying urban
sprawl by statistical methods. However, as highlighted by WP2, qualitative
information is essential to understand the mechanisms, and the quantitative
data analysis by statistical methods can only partially disclose the interactions
between causes and effects, factors and consequences.
An appropriate modelling tool has been developed based on a generalised
shift-share analysis , which is rather robust and requires as less as
possible data input from the case studies. The main indicators considered
in the quantitative analysis of urban sprawl are:
Temporal mean growth rate (population, employment, commuters, etc.)
The concentration-measure H, which quantifies concentration or de-concentration
effects in the urban system
Global and Local Indicators of Spatial Autocorrelation (LISA)