城市扩张经典CA模型模拟精度的时空衰减效应--以广州市2000-2010年城市扩张为例

摘要

Cellular automata (CA) is one of the most efifcient models for simulating the urban sprawling. However, simulation accuracy of the CA model based on only the first law of geography is obviously inclined to decrease at the spatial-temporal scale. It indicates that the simulation accuracy is decreasing with the time extended. In this research, the urban sprawling of Guangzhou city from the year 2000 to 2010 was discussed in detail. According to the simulation results, it can be found the simulation accuracy of classical CA model decreased with the time variation and showed obvious spatial difference. Further analysis veriifed that classical CA model is just only able to forecast the urban sprawling in a very short time and unavailable for simulation of urban growth at the long time scale. Results demonstrate that combination of the following factors should be considered to better explain complex characteristics of urban growth: the first law of geography, the spatial attenuation effect and the geographical spatial differentiation and so on.%城市快速增长是我国当前城市化过程的典型特征，而利用地理模拟工具预测未来城市空间形态有助于城市规划政策制定。元胞自动机（CA）是一种模拟城市扩张的有效建模工具，但由于城市是一个自然人文复合系统，仅根据地理学第一定律构建的CA模型模拟精度存在时空衰减效应。研究以广州市2000-2010年的城市扩张为案例，通过对经典C A模拟结果时间和空间两个维度的精度分析，探讨了经典CA模型模拟精度的时空衰减规律。