珠三角中心镇城市化对区域城市空间结构的影响——基于CA的模拟和分析

THE INFLUENCE OF URBANIZATION OF CENTRAL TOWN ON THE SPATIAL STRUCTURE OF REGIONAL CITY BASED ON THE STIMULATION AND ANALYSIS OF CELLULAR AUTOMATA

在珠三角的城市化过程中,中心镇的城市化对区域城市空间结构的形成和发展具有重要的影响,定量地研究中心镇的城市化对揭示珠三角地区城市化过程与城市空间结构的变化具有重要意义。本研究以东莞市为例,运用城市元胞自动机模型,模拟市域内1988—2004年的城市扩张过程,分析中心镇城市化对整个城市空间结构的影响。主要结论:1988—1993年,中心镇的城市化对区域城市空间结构影响较小,城市主要沿道路扩张,城市在空间布局上相对松散;1993—1997年,中心镇对区域城市空间结构影响增强,城市以中心镇为中心的扩张非常明显,城市空间沿道路的扩张,基本形成了渐进扩散式的“点—轴”空间发展模式,形成沿道路和中心镇扩展的格局;1997年以后,中心镇的辐射作用进一步增强,形成了“点—轴—集聚区”的城市空间扩展模式,城市空间扩展区在空间上更集中于中心镇。

Urbanization of central town plays an important role in the spatial structure of regional city. The traditional way of studying the influence is description and simple statistics. It has limitation to research on the relationship of central town＇s urbanization and urban spatial structure. With the development of geography information system and remote sensing, geographical spatial analysis model should be used to study the relationship between them. This paper discusses how the central town influences the spatial structure of city based on urban cellular automata model. This paper presents a new way to study the relationship by the integration of urban cellular automata and GIS, etc. Cellular automata can be used to simulate the complex systems by diffusion and dispersion effect. The diffusion and dispersion effect can be reflected by the weight of parameters in transition rules. Transition rule is the core of urban cellular automata. The weight of parameters, which can be obtained from the transition rules, can explain and analyze how the central towns influence the spatial structure of city. The weight of parameters in transition rule will be optimized by experience data with genetic algorithm firstly. Genetic algorithm can optimize the combination of all parameters that relate to urban development. This method is applied to the simulation and analysis of urban expansion in Dongguan, a fast developing city in the Pearl River Delta in south China. The weights of parameters of transition rule in urban cellular automata are obtained from experience training data from 1988 to 1993. The land use class data are captured from TM remote sensing image. Other data are captured from geography database by spatial analysis methods in ArcGIS. The urban cellular automata are then used to simulate urban development from 1988 to 2004. Some conclusions are obtained from the simulation results.