Urbanisation is driven by the complex interactions of many physical and human factors where human actions and decisions, individually and collectively, ultimately shape the patterns of urban landscapes. Agentbased modelling is an emerging technique in land use science that is designed to study multiple heterogeneous and locally interacting active entities within a system. An example of a local interaction is the request made by residents to planners for building permits. The decisions of planners in response to this request leads to emergent properties at an aggregate level such as city growth, assuming no equilibrium conditions. This thesis develops a framework for investigating in space and in time future residential land use change over a polycentric region using a case study of East Anglia, UK. Conceptually, the framework views the complexity of housing development in a system of cities (macrogeographical level) as the visible and concrete outcome of interactions between household demand for new dwellings (micro-geographical level) and the supply of building permits by local planners (meso-geographical level). Demand and supply are driven by household location preferences, as well as local planning, and evolve over time, leading to future land use change at speci c locations. The IPCC socio-economic scenarios are adapted to describe plausible evolutions in these preferences and strategies in order to evaluate di erent urban land use change pathways and the associated potential consequences for people (e.g. ooding risks) and the environment (e.g. biodiversity loss from land fragmentation). Simulation of new housing scenarios is undertaken within the agent-based modelling paradigm using a new computer programme developed in NetLogo. Issues of sensitivity analysis, validation, calibration and system complexity are addressed throughout the thesis. The thesis contributes to the eld of landscape and urban ecology by exploring urban complexity with a spatio-dynamic model of residential location behaviour driven by human and natural variables. As land use and land cover change is known to strongly a ect ecological landscape functions and processes, understanding the relationships between social and natural systems within changing landscapes helps to highlight hotspots of potential pressure and their e ects on the natural environment as part of an assessment of the possible ecological impacts of new urban development.
|Publication status||Published - 2010|
- agent-based modelling
- land use science
- urban planning
- urban development