GIS-based spatial analysis, modeling, and simulation: A case study on the changing spatial structure of Hong Kong, 1966-2006

This dissertation investigates the possibility and feasibility of integrating spatial analysis and modeling with Geographic Information Systems (GIS) to examine the dynamics of urban spatial structure.;This study found that the current generation of GIS is functionally limited as a research tool and needs to be further integrated with various spatial analytical modeling techniques. Only through such an integration that can GIS be used to address questions raised in a theoretical context and simulate policy impacts in a decision-making environment. This dissertation has demonstrated that the cross-fertilization of GIS with spatial analysis can overcome the inherent constraints of each technique, and has produced insights that would otherwise have been missed in geographic research.;By incorporating three new models of urban spatial structure--the Erickson model, the Hartshorn-Muller model, and Bourne's hypotheses--this study examined the changing spatial structure of Hong Kong between 1966 and 1986. The future development patterns by 2006 were also simulated at the census district level. It was found that the spatial distribution of population in Hong Kong has dispersed from the central urban areas around the harbor to urban peripheral areas as well as rural areas in the New Territories as a result of the government's public housing and new town programs. Despite these changes, Hong Kong's suburban space economy was still confined to the Spillover/Specialization stage in the Erickson model and Bedroom Community stage in the Hartshorn-Muller model, because most new towns still have to depend on central cities for job opportunities, particularly in the high-paid service sectors. Of Bourne's five hypotheses on the new urban spatial structure, three--Dispersed City Hypothesis, Social Mosaic Hypothesis, and Spatial Mismatch Hypothesis--have found substantial support in the context of Hong Kong; two--Reversed Gradient Status Hypothesis and Multinucleated City Hypothesis--were rejected. It was also found that vertical zonation is a prominent feature of residential segregation in Hong Kong. It is suggested that the Hong Kong government should design explicit policies to decentralize jobs from the central urban areas to the new towns for future development. (Abstract shortened by UMI.)...