Integrating urban economics and cellular automata to model periurbanisation : spatial dynamics of residential choice in the presence of neighbourhood externalities

This thesis addresses the issue of urban expansion and aims to explore the emergence and morphology of a periurban zone at the periphery of a city, where residents mix with agricultural activities. In Western Europe, these periurban areas have experienced rapid land use change and concern an increasing number of people. Thus, they challenge social and environmental spatial policy goals. Periurban morphologies vary from compact clusters to highly scattered developments. This thesis tests whether this diversity of spatial patterns can result from changes in individual residential choices. The thesis aims to understand the morphogenesis of residential patterns within the commuting periphery of a monocentric city through the development of a spatio-dynamic model. The model integrates two different modelling paradigms of urban land use change: Urban Economics and Cellular Automata. In this way, emerging periurban structures are explicitly spatial and founded on residential micro-economic theory. Moreover, the framework is dynamic and combines both short- and long-run equilibria through a utility adjustment process. The role of past history in shaping settlement patterns is emphasized because of the irreversibility of urban land conversion and the sequential location of utility maximizing households. Households are assumed to make residential location choices based on contrasting perceptions of their neighbourhoods that reflect preference for open space and social interactions. These externalities act as local agglomeration and dispersion forces at the scale of neighbourhoods and affect the monocentric structure that would result from the trade-off between commuting and housing costs. The model shows that a mixed periuban belt can emerge at the periphery of a city in stripped or scattered forms. The compactness of residential settlements within this belt varies according to the assumed preferences. 2D and 1D theoretical configurations are analysed, including sensitivity of spatial morphologies and land values to change in income and commuting cost. The model is also used to explore spatial segregation processes for two residential groups with different income and social preferences. A pattern matching procedure is finally proposed to explore the possibility of calibrating behavioural parameters from observed spatial patterns.Doctorat en sciences (sciences géographiques) (GEOG 3)--UCL, 200