Land use change simulation and prediction using integrated cellular automata models : A case study of Ipswich City, Australia
Abstract
Land use change models are important tools for understanding geographical phenomena and their dynamic interactions with other systems. They play a key role in the fields of land resource management, spatial science and urban planning. Among all land use change models, cellular automata (CA) have been widely explored and used in the literature, with considerable contributions to theories, methods and corresponding applications all around the world. An important issue with CA models is the unavoidable precision-loss with traditional raster-based CA and the demand of high-accuracy modelling tools. In this thesis, a vector-based CA model is proposed, implemented as prototype, and then integrated with artificial intelligence algorithms and a planning support system. Taking Ipswich City, South East Queensland (SEQ) Region, Australia as the study region, the thesis provides a comprehensive exploration of vector-based CA modelling, with macro and micro spatial variables, qualitative and quantitative evaluation. Key topics assessed are the comparison, simulation and evaluation of vector and raster-based CA models; the effects of spatial heterogeneity and partitioned transition rules at multi-level spatial scales; the integration of vector-based CA model with planning support system for predicting future scenario developments. In general, the thesis has explored multiple research points of vector-based CA modelling, with a mixture of data sources. The characteristics and mechanisms of historical and future residential expansion, as well as its connection with government policies and social-economic factors, have been analysed and illustrated. In conclusion, vector-based CA model could reduce the sensitivity of cell size and computation time effectively, identify the nonlinear connections between spatial variables and land use patterns, as well as forecasting the future development in a logical and coherent way.
Problem melden