基于滑动窗口的局部分维分析方法——元分维模型,采用2005年江苏省城镇基础数据,对江苏省1850个城镇及街道办的空间均衡程度进行定量测度,结果表明:①江苏城镇体系空间均衡程度差异明显,自南向北存在一定程度的空间梯度递减,其中江苏南部城镇分布最为密集且较为均衡,江苏中部次之,而江苏北部除徐州市区外,均衡程度较低;②江苏城镇分布上东部比西部均衡;③江苏城镇轴线集中发展的分布特征突出,沿长江分布的无锡、常州、苏州、镇江、扬州、泰州、南通、南京8个地区城镇分布密集,且均衡程度较高;④城镇体系的空间均衡程度与城市化水平基本一致。实验证明,元分维模型可以用于定量分析特定尺度下的城镇局部空间均衡程度的差异。
The spatial pattern of urban system in the result of the location selection is formed mainly through the economic activities and the population migration in history. It is an important part of the study of the urban system taking towns in Jiangsu Province as a point in a geographical space in order to discover the spatial distribution of the towns and to analyze the formation of the internal mechanism of this type. There are four methods of the urban system’s spatial structure research widely used currently. They are nearest neighbor point index, neighboring point’s average, variation coefficient of the Voronoi area and counting numbers of the individual goals in a grid using Keermagefu-Shimiernuofu formula and Lorenz curve. The methods above cannot describe the local spatial distribution of the point group. This paper uses a Meta Fractal Dimension Model (MFDM) that can be applied to describe the variation of local shape of map objects to measure the spatial structure of 1850 towns and blocks quantitatively using Jiangsu foundational data in 2005. The result shows that: ①Difference of the spatial distribution characteristic is significant and the evenness degree decreases from south to north. Towns in the South of Jiangsu Province distribute intensively and evenly, relatively slight in the middle region and not even in the north except downtown of Xuzhou; ②East region in Jiangsu Province is more even than the west; ③It is obvious that the distribution characteristic of towns, like Wuxi, Changzhou, Suzhou, Zhenjiang, Yangzhou, Taizhou, Nantong and Nanjing developing along the axis of the Yangtze River, is intensive; ④Evenness degree of the urban system is basically consistent with the urbanization level in Jiangsu Province. It is proved that the MFDM is effective to discover the variation of the local spatial distribution characteristic on a certain scale.
[1] 杨国安,甘国辉.中国城镇体系空间分布特征及其变化[J].地球信息科学,2004,6(3):12~18.
[2] 陈 涛,刘继生.城市体系分形特征的初步研究[J].人文地理,1994,9 (1):26~30.
[3] 刘继生,陈彦光.城镇体系空间结构的分形维数及其测算方法[J].地理研究,1999,18 (2):171~178.
[4] Herzfeld U C, Overbeck C. Analysis and Simulation of Scale-dependent Fractal Surface with Application to Seafloor Morphology [J].Computers and Geosciences,1999,25 (9):979-1007.
[5] Longley P A,Batty M. Fractal Measurement and Line Generalization [J].Computers and Geosciences, 1989,15(2): 167~183.
[6] 郭仁忠.空间分析[M]. 武汉:武汉测绘科技大学出版社,1997.
[7] 张 红,王新生,余瑞林.基于Voronoi图的测度点状目标空间分布特征的方法[J]. 华中师范大学学报(自然科学版),2005,39(3):422~426.
[8] 许学强,周一星,宁越敏.城市地理学[M].北京:高等教育出版社,2001.
[9] 龙 毅,毋河海,周 侗.地图目标局部分形描述的元分维模型的实现[J]. 武汉大学学报(信息科学版),2006,31(10):892~895.
[10] 顾朝林.江苏省城市现代化水平评价及预测[J].城市规划汇刊,2000,6:34~39.
[11] 马晓冬,徐建刚.江苏省城市化发展的区域差异分析[J].世界地理研究,2006,15 (3):68~74.
[12] 伍世代,王 强.福建省城镇体系分形研究[J]. 地理科学,2007,27(4):493~498.
[13] 陈彦光,周一星.豫北地区城镇体系空间结构的多分形研究[J].北京大学学报(自然科学版),2001,37(6):810~818.
[14] 崔世林,龙毅,周侗,王丽琴.一种地图点群空间分布特征的局部分维分析方法[J].地球信息科学,2007,9(6):25~30.
[15] CHENG Q.The gliding box methods for multi-fractal modeling[J].Computers & Geosciences.1999,25:1073-1079.
[16] PROKOPH A.Multi-fractal and sliding window correlation dimension analysis of sedimentary time series[J].Computers &Geosciences,1999,25:1009-1021.
[17] 胡丽娅.江苏城镇化现状特征分析[J].现代城市研究,2000,85:16~20.
[18] 江苏省统计局.江苏统计年鉴[M].中国统计出版社,2002.
