EN中文

Journal of Geo-information Science >

2017 , Vol. 19 >Issue 6: 792 - 799

DOI: https://doi.org/10.3724/SP.J.1047.2017.00792

Orginal Article

Research on the Spatio-temporal Visualization of Multiple Time Series Land Use Change by the Self-organizing Map

  • QI Jianchao ,
  • LIU Huiping , * ,
  • GAO Xiaofeng
Expand
  • 1. State Key Laboratory of Remote Sensing Science, Beijing Normal University, Beijing 100875, China
  • 2. Beijing Key Laboratory of Environmental Remote Sensing and Digital City, Beijing Normal University, Beijing 100875, China
  • 3. School of Geography and RS, Beijing Normal University, Beijing 100875, China
*Corresponding author: LIU Huiping, E-mail:

Received date: 2016-10-11

  Request revised date: 2017-03-07

  Online published: 2017-06-20

Copyright

《地球信息科学学报》编辑部 所有
Fold

Abstract

Analysis of multiple time series land use of spatio-temporal evolution at fine scale is a hot and important research area currently. In this study, the self-organizing map （SOM） neural network was used to analyze the land use spatio-temporal distribution at township-level in Beijing. The study was based on 5 periods of land use classification data of Beijing in 2005, 2007, 2009, 2011 and 2013. We implemented spatio-temporal integrated expression and comparative analysis of multiple time series land use data at township-level. Through creating and training self-organizing map neural network, we could find out the distribution of different land use types （built-up land, farmland, forest land, grassland, garden, water, and unused land） on the SOM output plane. This represented the proportional relationship of different land use types in land use structure. By second-step clustering and building land use change trajectory, we got the spatio-temporal evolution rules of the land use in township of Beijing. The results revealed that there were five land use change trajectories and three spatio-temporal evolution patterns in Beijing at township level. The plain area is developing to the land use structure of high built-up land proportion. The mountainous area is developing to the land use structure of high forest land proportion, and the land use change of piedmont zone is complex.

Key words： self-organizing map; township-level land use change; spatio-temporal visualization; trajectory analysis; Beijing

Cite this article

QI Jianchao , LIU Huiping , GAO Xiaofeng . Research on the Spatio-temporal Visualization of Multiple Time Series Land Use Change by the Self-organizing Map[J]. Journal of Geo-information Science, 2017 , 19(6) : 792 -799 . DOI: 10.3724/SP.J.1047.2017.00792

1 引言

土地利用/土地覆被变化（简称LUCC）是全球环境变化研究的核心领域和土地可持续利用研究的重要课题[1]。城市土地利用变化的研究一直是地理学以及城市规划学等学科的核心内容之一[2],随着城市化进程的加快,关于城市土地利用变化的研究越来越多,并已取得了丰硕的成果[3-5]。当前的土地利用特别是城市土地利用研究多是从城市整体出发选用相关的指标或方法来刻画土地利用的空间格局及其演化[6-8],或是从地市级或区县级尺度来探索分析其时空演变规律[9-11],还有一些研究以栅格为研究对象对土地利用变化进行分析和预测[12-13],而从乡镇尺度上分析土地利用结构时空变化规律的研究比较缺乏[14]。由于乡镇是土地利用规划的最小单位,所以研究土地利用变化在乡镇尺度上表现出的时空规律对于精细的城市土地管理与规划具有重要的实际意义。
对于时间序列数据的处理也是土地利用变化分析中的热点与难点,当前对于多时间序列数据的分析多是基于单期或相邻2期数据的两两对比来揭示在多时间序列上土地利用变化的时空规律[15-16],另外,构建土地利用变化轨迹也是分析多时间序列土地利用变化规律的常用方法[4,17-18]。而当时间序列数据较多时,对其进行对比分析与可视化仍显得比较困难,因此需要一种可以实现对多时间序列土地利用数据的时空一体化表达及对比分析的方法。本文所采用的自组织映射（Self-Organizing Map,SOM）神经网络方法可以同时处理多时间序列土地利用数据,并利用其拓扑保持特性在其输出面板上进行二次聚类及构建土地利用变化轨迹,从而实现土地利用时空演变规律的定量分析与可视化表达。SOM方法已被广泛应用于各领域的时空数据分析、轨迹分析以及可视化研究中[19-21],而该方法在土地利用变化分析中的应用还比较少,因此本文引用自组织映射方法以为多时间序列土地利用时空演变规律与发展模式分析提供便捷,为实现土地利用变化动态过程的时空可视化表达提供支持。
本文基于北京市2005、2007、2009、2011、2013年5期土地利用分类数据,以乡镇为单元统计各乡镇的土地利用结构数据,采用自组织映射神经网络方法,实现了多时间序列土地利用变化数据的时空一体化表达与对比分析,揭示了2005年到2013年北京市乡镇级土地利用时空演变规律,为直观认识郊区长期的土地利用变化规律提供了有利的技术手段。

2 研究区与数据源

本文以北京市郊区的14个区下辖的286个乡镇为研究区。由于北京市中心城区以建设用地为主及密云水库区以水体为主的土地利用状况已经稳定,所以本文在研究中并未探讨这两个区域的土地利用变化规律。研究数据为北京市2005、2007、2009、2011、2013年5期由SPOT遥感影像解译获得的土地利用分类数据,该数据来自北京市农村经济研究中心的长期土地利用监测项目,其中2005年分类结果以SPOT4全色波段和多光谱波段融合影像为遥感解译的数据源,其余4期的分类结果均以SPOT5多光谱波段影像为遥感解译数据源,空间分辨率10 m。经随机撒点通过Google Earth高分辨率影像人工判读并结合部分地面实测样点验证,分类精度均在82%以上。土地利用分类体系参考全国农业区划委员会制定的《土地利用现状调查技术规程》中的8大类,并综合遥感数据分类特点将地类划分为建设用地、耕地、林地、草地、园地、水体和未利用地7类。

3 研究方法

3.1 自组织映射方法

自组织映射方法最早是在1982年由芬兰赫尔辛基理工大学的Kohonen[22]提出的,自组织映射方法是一种竞争型神经网络方法,与其他人工神经网络方法的不同之处在于它使用了一个邻近函数来保持输入空间的拓扑性质,从而可以直接在其输出面板上进行二次聚类以及轨迹构建。SOM算法的原理与流程如下：初始化各神经元权值向量;寻找输入数据对应的获胜神经元;调整优胜邻域内的神经元权值向量;重复寻找输入数据对应的获胜神经元及以后的步骤,直到迭代终止条件被满足（图1）。本文的输入数据为北京市5个年份286个乡镇的7种地类面积比例属性,即输入数据共1430条,然后设置SOM输出面板的尺寸为100×100,即输出10 000个神经元,要远大于输入的节点数,从而尽量保证在训练结果稳定后每一个输入节点只有单一的与其对应的获胜神经元。用获胜神经元代表每一条输入数据并在SOM输出空间中进行可视化,同时将聚类结果关联可视化到地理空间中以便于对比。由于是将5个年份的数据同时输入再训练输出,因此得到的聚类结果在时间和空间上具有一致的对比性。
Fig. 1 Flow chart of SOM algorithm

图1 SOM算法流程图

3.2 SOM各特征变量的成分平面图

SOM的输入数据为5个年份各乡镇的7种地类面积比例数据构成的7维向量,即将每个地类的面积比例属性作为一个特征变量,经过网络训练后可以在SOM输出面板上发现各地类的分布聚集状况。如图2所示的SOM输出成分平面图,每张图代表一种地类,该平面图只用来表示各地类特征变量值的分布,并不具有实际地理意义,图中红色代表高值聚集区,绿色代表低值聚集区,可以发现在SOM输出面板上建设用地（图2（a））、耕地（图2（b））、林地（图2（c））、草地（图2（d））、园地（图2（e））的分布比较聚集,而水体（图2（f））和未利用地（图2（g））的分布比较杂乱。同时,可以发现各地类的高值聚集区分布在SOM输出面板的不同位置,说明经网络训练后的输入数据在输入空间中的模式在SOM输出空间中得到了识别和表达。
Fig. 2 Output component planes of SOM

图2 SOM输出成分平面图

3.3 二次聚类与获胜神经元

当上述网络训练结果稳定后,采用K-means算法对10 000个输出神经元按其权值向量进行二次聚类,聚类数量通过多次试验确定为7时可以较好地解释研究区的状况。图3中7种颜色的面板区域划分即代表7种不同的二次聚类类型在SOM输出面板中的可视化。为了确定各聚类的含义,绘制二次聚类结果（土地利用结构类型）中各土地利用类型所占比例的折线图（图4）,据此将各聚类定义为不同的土地利用结构类型：耕地型、林地型、建设用地型、园地型等7类。其中,建设用地过渡型表示其建设用地比例相比于建设用地型为低,但仍为绝对优势比例地类,而其他地类像耕地比例较高一些,林地过渡型的定义同此理。
Fig. 3 Best matching unit and regional division of SOM output plane by second-step clustering

图3 获胜神经元与二次聚类SOM输出面板区域划分

Fig. 4 Area ratios of different land use types in second-step clustering results

图4 二次聚类结果（土地利用结构类型）中各用地类型所占比例

在属性空间中计算输入节点与所有输出神经元之间的欧氏距离,选择距离最小的输出神经元为该输入节点对应的获胜神经元,然后根据获胜神经元在SOM输出面板中的位置可以得到各乡镇在不同年份中分别属于哪种聚类类型,最后依据行政区划将各乡镇的聚类类型在地理空间中可视化（图5）,以直观地表达和对比不同年份之间各乡镇的土地利用结构类型的变化。图3-5对应共同的7个二次聚类类型,并采用相同的着色方案,即相同的颜色代表相同的土地利用结构类型。
Fig. 5 Geospatial visualization of the second-clustering results of township from 2005 to 2013

图5 2005-2013年各乡镇二次聚类结果在地理空间中的可视化
注：各图中心空白区域为北京市中心城区,东北部空白区域为密云水库区

3.4 轨迹分析

将每个乡镇在SOM输出面板上对应的各年份获胜神经元的位置按照时间先后依次连接,从而可在SOM输出空间中构建286条乡镇土地利用变 化轨迹,结合SOM输出成分平面图（图2）中各特征变量值的分布,可以发现各乡镇土地利用结构变化趋势。然后采用K-means算法对轨迹进行聚类, 输入数据为将5个年份各乡镇的7种地类的面积比例数据按时间顺序依次排列构成的35维向量,聚类数量经多次试验确定为5类时可以很好地揭示研 究区的土地利用变化模式,最后将轨迹聚类结果同时在SOM输出空间以及地理空间中进行可视化 表达。

4 结果及分析

4.1 土地利用结构类型演变分析

首先结合图2输出成分平面图中各地类比例值的分布及图3中获胜神经元的位置和二次聚类SOM输出面板区域划分,可以看出2005年获胜神经元的位置多分布在耕地比例较高的区域,即各乡镇的土地利用结构多以耕地型为主,而2013年获胜神经元的位置多分布在建设用地及林地比例较高的区域,即各乡镇的土地利用结构多以建设用地型及林地型为主,所以根据在SOM输出面板上不同年份获胜神经元位置的变化可以看出北京市各乡镇土地利用在整体上朝建设用地型和林地型2个方向发展。然后将二次聚类结果在地理空间中进行可视化（图5）,可以更直观地在地理空间中揭示上述土地利用变化规律,同时还可以发现更为精细的土地利用空间演变模式。由图5可以定量的确定建设用地型、林地型的增加以及耕地型的减少;园地型主要集中分布在平谷区,也有少部分分布在山前结合带,而且整体上园地型的数量有所减少;另外,山前混合型主要是林地和建设用地的比例相当,同时也包含其他地类,可以发现山前混合型在2005年分布较少,随后逐渐增加,在山前形成带状分布。其主要原因是在2005年与林地型或林地过渡型相接的是耕地型,而随着在城市扩张中耕地不断被建设用地侵占,耕地型减少而建设用地及其过渡型的增加,导致以林地为主导的用地结构和以建设用地为主导的用地结构在空间分布上直接相连,从而在二者之间逐渐形成带状分布的山前混合型。

4.2 土地利用变化轨迹分析

轨迹聚类结果如图6所示。其中,左图为在SOM输出空间的可视化（其中黑色和红色的点分别表示每条轨迹的起点和终点,即2005和2013年各乡镇获胜神经元的位置）;右图为在实际地理空间的可视化,且两图中相同的颜色代表同一聚类类型。将轨迹聚类结果结合轨迹方向和二次聚类结果以及SOM输出成分平面图中各地类比例的高值分布区可以发现：① 轨迹1表示向建设用地比例最高的土地利用结构发展,多表现为建设用地过渡型向建设用地型的变化,在地理空间中可以发现轨迹1主要在中心城区周边呈环状分布;② 轨迹2主要是从耕地比例很高的区域向建设用地比例较高的区域变化,在二次聚类结果中表现为耕地型向建设用地过渡型的变化,在地理空间中可以发现该聚类主要分布在北京市东部和南部,其土地利用的演变主要为耕地的退化以及建设用地的增长;③ 轨迹3在二次聚类结果中多表现为林地过渡型向林地型的变化,即向着林地比例较高的方向发展;④ 轨迹4表示向林地比例最高的土地利用结构发展,在二次聚类结果中表现为在林地型中朝着林地比例更高的方向变化,在地理空间中可以发现该聚类主要分布在山区植被最丰茂的乡镇;⑤ 轨迹5在SOM输出面板上其轨迹的变化表现的最为杂乱无章,在地理空间中主要分布在各地类混合分布的山前结合带,其土地利用变化轨迹常表现为从一种地类比例较高的区域向另一种地类比例较高的区域变化。结合图3所示二次聚类结果的SOM输出面板,可更加直观和精细地揭示上述北京市乡镇级土地利用变化模式。而且结合轨迹方向和SOM输出成分平面图中各地类比例的高值分布区（图2）,发现轨迹线不仅可以表现二次聚类类别间的变化,还可以表现某一类别中的变化,像西北山区的部分乡镇在本研究的5个时相中始终属于林地型,但其在林地型中仍然向着林地比例更高的方向发展。
Fig. 6 Land use change trajectories and visualization

图6 土地利用变化轨迹与可视化

综合土地利用结构类型演变以及土地利用变化轨迹分析,可将2005-2013年北京市乡镇级土地利用时空演变概括为3种主要模式：
（1）东南平原地区向建设用地型的发展模式。表现为建设用地型、建设用地过渡型的增加以及耕地型的减少,变化轨迹对应于1和2。由于城市扩展目前是以“摊大饼”式近域推进为主,扩展过程中首先侵占的土地利用类型为耕地,使耕地面积大幅减少,导致城市核心区周边的乡镇土地利用类型由耕地型向建设用地主导型转变。
（2）西北山区向林地型的发展模式,表现为林地型的增加以及林地过渡型的减少,变化轨迹对应于3和4。根据北京市“十二五规划”及《北京市土地利用总体规划（2006-2020）》等发展规划[23]将门头沟、平谷、怀柔、密云、延庆等区作为北京市生态涵养区。随着政策的实施,生态环境质量持续提高,山区森林覆盖率不断增加,使山区的乡镇土地利用结构表现为以林地为主导的类型,且越来越多的林地过渡型逐步转化为林地型。
（3）平原与山区相接的过渡地带向山前混合型的发展模式并逐渐形成带状分布,表现为山前混合型乡镇数量的增加,变化轨迹对应于5。由于山前地区地形较为复杂,土地利用方式多样化,随着靠近山前的城镇不断扩展,林地、耕地等其他土地利用类型逐步转换为建设用地,从而使山前地区土地利用结构变化更加复杂。

5 结论与讨论

本研究将自组织映射方法应用到土地利用时空演变分析中,利用2005、2007、2009、2011、2013年5个时相的土地利用遥感分类数据,分析了北京市乡镇级土地利用时空变化特征。将SOM方法应用于多时间序列土地利用变化数据的时空一体化表达和演变规律分析中的思路,可以为其他多时间序列时空数据分析及可视化提供借鉴。
研究中通过将SOM输出面板空间与实际地理空间进行关联可视化,实现了多时间序列土地利用变化数据的时空一体化表达和对比分析,得到了北京市乡镇级土地利用时空演变规律,概括为3种主要模式：东南平原地区向建设用地型的发展模式、西北山区向林地型的发展模式、平原与山区相接的过渡地带向山前混合型的发展模式。其中,平原区由于城市化及城市扩张整体向建设用地比例更高的土地利用结构方向发展,并伴随着耕地的减少;山区作为生态涵养的功能整体向林地比例更高的土地利用结构方向发展;而处于二者过渡带的山前结合带,其土地利用变化轨迹较为复杂。
本研究中的分析单元为北京市各乡镇,由于乡镇作为土地利用规划的最小单位,从乡镇尺度来分析北京市的土地利用时空演变规律具有很好的解释性和实际意义,后续的研究将分析不同自然网格尺度划分下的土地利用时空演变规律,以探究统计面元的变化对土地利用时空演变分析结果的影响,更全面地揭示与理解土地利用变化的内在规律。

The authors have declared that no competing interests exist.

References

Publishing order | Descend order by publishing year | Descend order by cited within
[1]
李秀彬. 土地利用变化的解释[J].地理科学进展,2002,21(3):195-203.用途转移和集约度变化构成土地利用变化的两种基本类型。土地特性自身的变化、土地使用者个体经济行为分析及社会群体土地管理行为分析 ,构成土地利用变化解释的理论框架。从土地特性考察 ,多宜性和限制性是土地利用发生变化的基本条件。人类对土地利用的结果 ,总是趋向于使土地的多宜性降低和功能类型减少。竞租曲线、转移边际点以及打破土地利用空间均衡的条件分析 ,是土地利用变化经济分析的理论基础 ;“土地利用 -环境效应 -体制响应”反馈环的作用机制 ,构成社会群体土地管理行为分析的理论框架。土地利用变化的机理模型 ,目前主要以新古典经济学和地租理论为基础。多视角的探索可能是土地利用变化机理综合分析的有效途径。

DOI

[ Li X B.Explanation of land use changes[J]. Progress in Geography, 2002,21(3):195-203. ]

[2]
鲁春阳,杨庆媛,靳东晓,等.中国城市土地利用结构研究进展及展望[J].地理科学进展,2010,29(7): 861-868.城市土地利用结构一直是城市规划学和城市地理学研究的核心内容之一.合理的城市土地利用结构能促进城市功能发挥及用地效益提升,促进区域经济与环境的和谐发展.本文在总结中国城市土地利用结构演变、驱动机制、结构优化及动态模拟等方面研究进展的基础上,发现在研究方法与研究内容方面仍存在不足,有待深化.本文认为,未来中国城市土地利用结构研究应从城市功能结构合理配置的角度,探究不同产业阶段城市土地利用结构特征;不仅关注城市数量结构研究,更注重城市土地利用空间布局的研究;研究方法从以静态为主向动态化转变,建立科学的土地利用决策支持系统;构建不同类型城市发展必须的功能性用地及用地比例的理论框架,为城市规划编制和优化用地空间布局提供依据.

DOI

[ Lu C Y, Yang Q Y, Jin D X, et al.Research progress and prospects of the researches on urban land use structure in China[J]. Progress in Geography, 2010,29(7): 861-868. ]

[3]
Sun J, Zhang L, Peng C, et al.CA-based urban land use prediction model: A case study on orange county, Florida, U.S.[J]. Journal of Transportation Systems Engineering and Information Technology, 2012,12(6):85-92.For a long time,interactions between land use and transportation have been one of the research hotspots in urban planning,which however,has not been investigated in much detail until recently.This paper started from spatial changes of regional land use,with an objective of understanding the relationship between trip generation and urban land use.Cellular automata and geographical information system techniques were used to store and update the spatial data dynamically.In addition,MATLAB software was adopted to conduct the logistic regression using land use/land cover data from Orange Country,Florida,U.S.(1990 and 2000).With fully utilizing the advantages of cellular automaton models,simulation results indicate the enhenced reliability of the model,which consequently assists to understand evolution of urban land uses.

DOI

[4]
Liu H, Zhou Q.Developing urban growth predictions from spatial indicators based on multi-temporal images[J]. Computers, Environment and Urban Systems, 2005,29(5):580-594.Landuse change in metropolitan areas is largely focused on the dynamic nature of urban landuse change. In this research, a spatial statistical model was used to support decision-making with regard to urban growth predictions in the urban fringe of Beijing, China. The model adopted in this study was based on the integration of remote sensing, geographical information systems, and multivariate mathematical models. The model emphasises the spatial distribution of the landuse/cover units and the spatio-temporal patterns, which were modelled by landuse/cover change trajectories over a series of observation years. The main trajectories for the landuse/cover change model were based on five sets of multitemporal landuse/cover data derived from remotely sensed images. Using the integrated GIS, several spatial variables were derived, including the proximity to major roads and built-up areas. A multivariate model was established to establish relationships between urban expansion and above spatial variables. The landuse/cover change trajectories and the multivariate model were then integrated to construct a multivariate spatial model that is capable of estimating the spatial probability of the urban expansion.

DOI

[5]
冯健. 杭州城市形态和土地利用结构的时空演化[J].地理学报,2003,58(3): 343-353.根据分形理论研究杭州1949～1996年间城市形态和土地利用结构的演化特征,发现杭州城市具有明确的自相似规律.研究表明,杭州城市形态和土地利用结构的分形性态逐渐变好,这与国外学者"演化的城市分形"观相互印证.各类土地形态的维数都小于整个城市形态的维数,从而证实了国内学者"城市化地区的分维大于各职能类土地空间分布维数"的理论推断.从时空变化来看,杭州城市形态的分维呈上升趋势,1996年接近Batty等提出的理论预期维数D=1.71;居住用地、工业用地和对外交通用地的分维近20年来趋于增大,而教育用地和绿化用地的分维则有所减小.杭州市的分形演化和分维变化总体上揭示了城市自组织演化的特征,但工业用地维数的大幅度上升和绿化用地维数的下降显然暗示该城市在进化过程中的局部退化倾向.

DOI

[ Feng J.Spatial-temporal evolution of urban morphology and land use structure in Hangzhou[J]. Acta Geographica Sinica, 2003,58(3):343-353. ]

[5]
Estoque R C, Murayama Y.Spatio-temporal urban land use/cover change analysis in a hill station: The case of Baguio city, Philippines[J]. Procedia - Social and Behavioral Sciences, 2011,21:326-335.

[7]
Wu J, Jenerette G D, Buyantuyev A, et al.Quantifying spatiotemporal patterns of urbanization: The case of the two fastest growing metropolitan regions in the United States[J]. Ecological Complexity, 2011,8(1):1-8.Urbanization is the most drastic form of land use change affecting biodiversity and ecosystem functioning and services far beyond the limits of cities. To understand the process of urbanization itself as well as its ecological consequences, it is important to quantify the spatiotemporal patterns of urbanization. Based on historical land use data, we characterize the temporal patterns of Phoenix and Las Vegas the two fastest growing metropolitan regions in the United States using landscape pattern metrics at multiple spatial resolutions. Our results showed that the two urban landscapes exhibited strikingly similar temporal patterns of urbanization. During the past several decades, urbanization in the two desert cities resulted in an increasingly faster increase in the patch density, edge density, and structural complexity at both levels of urban land use and the entire landscape. That is, as urbanization continued to unfold, both landscapes became increasingly more diverse in land use, more fragmented in structure, and more complex in shape. The high degree of similarity between the two metropolitan regions may be attributable to their resemblance in the natural environment, the form of population growth, and the stage of urban development. While our results corroborated some theoretical predictions in the literature, they also showed spatiotemporal signatures of urbanization that were different from other cities. Resolving these differences can certainly further our understanding of urban dynamics. Finally, this study suggests that a small set of landscape metrics is able to capture the main spatiotemporal signatures of urbanization, and that the general patterns of urbanization do not seem to be significantly affected by changing grain sizes of land use maps when the spatial extent is fixed. This landscape pattern analysis approach is not only effective for quantifying urbanization patterns, but also for evaluating spatial urban models and investigating ecological effects of urbanization.

DOI

[8]
徐丽华,王欢欢,张结存,等.近15年来杭州市土地利用结构的时空演变[J].经济地理,2014,34(7):135-142.以杭州市为例，基于Landsat的TM遥感影像，解译土地利用专题信息，计算土地利用转移矩阵、土地利用动态度和各地类重心的转移，分析1995-2010年城市土地利用变化的时空格局特征及其影响因素。结果表明：①近15年来各类土地在数量上呈现四增一减一平的趋势，空间上朝着复杂化、多样化的方向发展。②各类用地之间相互流转的过程较为复杂，变化最剧烈的是城市绿地，转移量最大的是耕地。③建设用地持续增加，增速逐渐降低，土地利用集约度提高。④各类用地重心均有偏移，建设用地和湿地的重心偏移距离最大，林地偏移距离最小。⑤1995-2000年，杭州土地城市化进程相对较慢，出现向钱塘江以东区域扩张的态势；2000--2005年是土地城市化进程最快、旧城改造力度最大、耕地锐减的时期；2005--2010年则表现出城市规模扩张与内涵挖潜的双重特征。

[ Xu L H, Wang H H,Zhang J C, et al.Spatial-temporal dynamics of land use in the Hangzhou city during the recent 15 years[J]. Economic Geography, 2014,34(7):135-142. ]

[9]
鲁春阳,文枫,杨庆媛,等.地级以上城市土地利用结构特征及影响因素差异分析[J].地理科学,2011,31(5):600-607.

[ Lu C Y, Wen F, Yang Q Y, et al.Characteristics and driving factors of urban land use structure of cities at provincial level and above[J]. Scientia Geographica Sinica, 2011,31(5):600-607. ]

[10]
Deng J S, Wang K, Hong Y, et al.Spatio-temporal dynamics and evolution of land use change and landscape pattern in response to rapid urbanization[J]. Landscape and Urban Planning, 2009,92(3-4):187-198.Analyzing spatio-temporal characteristics of land use change is essential for understanding and assessing ecological consequence of urbanization. More importantly, such analysis can provide basic information for appropriate decision-making. By integrating historical high spatial-resolution SPOT images and spatial metrics, this study explored the spatio-temporal dynamics and evolution of land use change and landscape pattern in response to the rapid urbanization process of a booming-developing city in China from 1996 to 2006. Accurate and consistent land use change information was first extracted by the change detection method proposed in this study. The changes of landscape pattern were then analyzed using a series of spatial metrics which were derived from FRAGSTATS software. The results indicated that the rapid urbanization process has brought about enormous land use changes and urban growth at an unprecedented scale and rate and, consequently, given rise to substantial impacts on the landscape pattern. Findings further revealed that cropland and water were the major land use types developed for urban sprawl. Meanwhile, the landscape pattern underwent fundamental transition from agricultural-land-use dominant landscape to urban-land-use dominant landscape spanning the 10 years. The results not only confirmed the applicability and effectiveness of the combined method of remote sensing and metrics, but also revealed notable spatio-temporal features of land use change and landscape pattern dynamics throughout the different time periods (1996–2000, 2000–2003 and 2003–2006).

DOI

[11]
周生路,朱青,赵其国.近十几年来南京市土地利用结构变化特征研究[J].土壤,2005,37(4):394-399.该文以1985、1992、1996、2000、2002年5个典型年份南京市各土地利用类 型面积的调查数据为基础资料,运用景观生态学结构定量分析方法对南京市十几年来土地利用结构特征的变化进行了研究,并在此基础上分析了造成这种变化的主要 驱动力及其驱动机制.研究表明:①该方法能够很好地揭示南京市十几年来土地利用结构特征的变化规律;②南京市土地利用结构变化特征为:土地利用结构更趋于 多样化和平均化,土地利用类型的异质性加强,且在2000年到2002年该变化更加明显;③南京市土地利用结构特征变化受到经济发展、人口增加、基础设施 建设、产业结构调整、土地政策波动等因素的强烈影响.

DOI

[ Zhou S L, Zhu Q, Zhao Q G.Variation of land-use structure in Nanjing over the last decade[J]. Soils, 2005,37(4):394-399. ]

[12]
Tan R, Liu Y, Zhou K, et al.A game-theory based agent-cellular model for use in urban growth simulation: A case study of the rapidly urbanizing Wuhan area of central China[J]. Computers, Environment and Urban Systems, 2015,49:15-29.Accurate modeling of urban growth is an extremely important component of urban geographic studies and is also vital for future urban planning. The trajectories of urban growth can be monitored and modeled by the use of geographic information system techniques, remote sensing data, and statistical analysis. In this study, we couple game theory with an integrated agent-cellular method to develop a model of the major determinants controlling urban development, which not only accounts for socioeconomic driving forces but also captures human actions. Wuhan, the largest city in central China, is undergoing rapid urbanization and is facing uncontrolled urban expansion. The city proper region of Wuhan is selected as the case study area to simulate urban growth during the period between 2003 and 2023. The results indicate that the social conflicts between the different stakeholders in urban development can be identified by utilizing a game tree. The game-theory based agent-cellular model is shown to be more effective than a pure cellular automata model in urban growth simulation. The results also show that, from 2013 to 2023, the urban area of the Wuhan city proper region is predicted to grow to 442.77 km 2 , which is almost two times the area in 2003. This research is the first study to use empirical data and game theory to analyze the decision-making process in urban development in the Wuhan area.

DOI

[13]
Yang X, Zheng X, Lv L.A spatiotemporal model of land use change based on ant colony optimization, Markov chain and cellular automata[J]. Ecological Modelling, 2012,233:11-19.This paper proposes a spatiotemporal model of land use change based on ant colony optimization (ACO), Markov chain and cellular automata (CA). These three methodologies have previously been used separately or in pairs to simulate land use change. In this paper, we apply them in combination, using ant colony optimization and cellular automata to manage the spatial distribution of land use, and applying Markov chain and cellular automata to manage the total amount of land use coverage. We first describe the principle and implementation of the model. Then a land use map of an experimental area (Changping, a district of Beijing) based on land use maps from 1988 and 1998 is simulated for 2008 using the model. By analyzing with real situation, accuracy of the simulation result manifests that the model is useful for land use change simulation. And compared with the other two models (CA–Markov model and ACO–CA model), the model is more appropriate in predicting both the quantity and spatial distribution of land use change in the study area. Therefore the model proposed by this paper is capable of simulating land use change.

DOI

[14]
王谦. 基于乡镇尺度的大冶市LUCC及其景观格局变化研究[D].南宁:广西师范学院,2014.

[ Wang Q.Studies on land use/land cover change and landscape pattern change of town level in Daye city based on RS and GIS[D]. Nanning: Guangxi Teachers Education University, 2014. ]

[15]
Long H, Tang G, Li X, et al.Socio-economic driving forces of land-use change in Kunshan, the Yangtze River Delta economic area of China[J]. Journal of Environmental Management, 2007,83(3):351-364.This paper analyzes characteristics, major driving forces and alternative management measures of land-use change in Kunshan, Jiangsu province, China. The study used remote sensing (RS) maps and socio-economic data. Based on RS-derived maps, two change matrices were constructed for detecting land-use change between 1987 and 1994, and between 1994 and 2000 through pixel-to-pixel comparisons. The outcomes indicated that paddy fields, dryland, and forested land moderately decreased by 8.2%, 29% and 2.6% from 1987 to 1994, and by 4.1%, 7.6% and 8% from 1994 to 2000, respectively. In contrast, the following increased greatly from 1987 to 1994: artificial ponds by 48%, urban settlements by 87.6%, rural settlements by 41.1%, and construction land by 511.8%. From 1994 to 2000, these land covers increased by 3.6%, 28.1%, 23.4% and 47.1%, respectively. For the whole area, fragmentation of land cover was very significant. In addition, socio-economic data were used to analyze major driving forces triggering land-use change through bivariate analysis. The results indicated that industrialization, urbanization, population growth, and China's economic reform measures are four major driving forces contributing to land-use change in Kunshan. Finally, we introduced some possible management measures such as urban growth boundary (UGB) and incentive-based policies. We pointed out that, given the rapidity of the observed changes, it is critical that additional studies be undertaken to evaluate these suggested policies, focusing on what their effects might be in this region, and how these might be implemented.

DOI PMID

[16]
Fan Q, Ding S.Landscape pattern changes at a county scale: A case study in Fengqiu, Henan Province, China from 1990 to 2013[J]. CATENA, 2016,137:152-160.Human activities and natural factors drive landscape pattern changes and limit provision of ecosystem services (ES) for human well-being. The analysis of landscape pattern change is one of the most important methods to understand and quantify land use and land cover change (LUCC). In this study, a series of satellite images (1990, 2002, 2009, 2013) of Fengqiu County of Henan province in China and some social data were used for analyzing landscape pattern changes and driving forces. Our results showed that landscape pattern and indices of Fengqiu County have serial changes during 1990–2013. From 1990 to 2013, the unused land (UL) nearly disappeared (the area of UL changed from 19.1 to 1.0602km 2 ) and the area of water area (WA) dramatically decreased (from 71.41 to 11.402km 2 ). The mutual transformations among cultivated land (CL), forest land (FL) and settlements and mining sites (SMS) were relatively frequent. By further analysis of the number of patches (NP), largest patch index (LPI), perimeter–area fractal dimension (PAFRAC) and Shannon's evenness index (SHEI) both at class and landscape scale, we found that anthropogenic influence increased gradually, intensity of land use is strengthened, and landscape heterogeneity reduced. Human activity, especially population growth was the main driving force which impacted the landscape changes in studied area. The natural factors (temperature and precipitation) make a large impact on WA area. At last, we firstly introduce “Entropy model” to analyze the whole land use change. All the quantifications of LUCC and driving forces can reasonably provide basic information for government to guide the land use and ecological protection.

DOI

[17]
黄勇,王凤友,蔡体久,等.环渤海地区景观格局动态变化轨迹分析[J].水土保持学报,2015,29(2):314-319.对变化轨迹的目标、特征等进行了总结、归纳,提出变化轨迹斑块的意义,通过轨迹斑块将复杂的时空动态现象用二维图谱表示出来.同时,基于变化轨迹方法,对环渤海地区土地覆盖时空动态变化进行分析,并选取占所有发生变化面积91.25％的30种主要的变化轨迹,生成变化轨迹空间分布图谱,并结合格局指数进行分析.结果表明,(1)“耕地”属于人类干扰最为剧烈的地类,其变化主要表现在耕地的流失上;(2)“耕地”向“人工表面”的转换是研究区最大的变化类型,其中2005-2010年变化面积远大于2000-2005年,占用耕地进行开发建设现象更加严重;(3)“海域”向“人工表面”的转换主要发生在2005-2010年间,主要由人工围填海造成;(4)湿地面积总体上变化不大,但其转入、转出比例均较大,证明湿地内部变化剧烈,湿地的破坏和人工湿地的补充同时进行;(5)研究区内主要变化轨迹类型大都来源于人类的开发利用活动,人类干扰起了至关重要的作用.研究结果为研究区景观格局提供了清晰的时空动态刻画与分析,为有关部门在土地合理利用、耕地和生态用地保护等方面提供了基础数据和理论支撑.

DOI

[ Huang Y, Wang F Y, Cai T J, et al.Landscape pattern dynamic analysis based on change trajectory method in Bohai rim area[J]. Journal of Soil and Water Conservation, 2015,29(2):314-319. ]

[18]
Feng H, Zhao X, Chen F, et al.Using land use change trajectories to quantify the effects of urbanization on urban heat island[J]. Advances in Space Research, 2014,53(3):463-473.This paper proposed a quantitative method of land use change trajectory, which means the succession among different land use types across time, to examine the effects of urbanization on an urban heat island (UHI). To accomplish this, multi-temporal images from Landsat 5 Thematic Mapper (TM) and Landsat 7 Enhanced Thematic Mapper Plus (ETM+) of Xiamen City in China from 1987 to 2007 were selected. First, the land use change trajectories were extracted based on the classified images from different years. Then the land surface temperatures (LST) were retrieved and the magnitudes of the UHI were evaluated using the UHI intensity (UHII) indicator. Finally, the indices of the contribution to UHI intensity (CUHII) were constructed and calculated to quantify the effects of each land use change trajectory on the UHI during urbanization. The results demonstrated that the land use change trajectories and CUHII are effective and useful in quantifying the effects of urbanization on UHI. In Xiamen City, a total of 2218 land use change trajectories were identified and 530 of them were the existing urban or urbanization trajectories. The UHII presents a trend of continuous increase from 0.83 C in 1987 to 2.14 C in 2007. With respect to the effects of urban growth on UHI, the contribution of existing urban area to UHI decreased during urbanization. Prior to 2007, the existing urban area of trajectory NO. 44444 had the most significant effect on UHI with the greatest CUHII, while the value has decreased from 55.00% in 1987 to 13.03% in 2007 because of the addition of new urbanized area. In 2007, the greatest CUHII was replaced by a trajectory from farmland to built-up area (NO. 22224) with the CUHII of 21.98%, followed by the existing urban area of trajectory NO. 44444 with the CUHII of 13.03%. These results provide not only a new methodology to assess the environmental effects of urbanization, but also decision-supports for the planning and management of cities.

DOI

[19]
Lee A C D, Rinner C. Visualizing urban social change with self-organizing maps: Toronto neighbourhoods, 1996-2006[J]. Habitat International, 2015,45:92-98.Change in the socio-economic status of urban neighbourhoods is a complex phenomenon with multiple space, time, and attribute dimensions. The objective of this research was to explore the use of a Self-Organizing Map (SOM) to visualize patterns of urban social change. In a case study, we collected, organized, and joined data from the 1996, 2001, and 2006 Canadian Census for the City of Toronto. Urban neighbourhoods were represented by Census tracts. The SOM translates multi-dimensional data into two-dimensional graphical patterns of neighbourhood socio-economic status. These were associated with patterns in geographic space. Spatio-temporal change was represented by trajectories in the SOM. The study identified trends of decreasing neighbourhood diversity and shifts in the dynamics of urban social change in Toronto. The proposed methodology could assist with strategic planning of urban development and efficient resource allocation that fits with local needs.

DOI

[20]
Augustijn E, Zurita-Milla R.Self-organizing maps as an approach to exploring spatiotemporal diffusion patterns[J]. International Journal of Health Geographics, 2013,12(1):1-14.Abstract BACKGROUND: Self-organizing maps (SOMs) have now been applied for a number of years to identify patterns in large datasets; yet, their application in the spatiotemporal domain has been lagging. Here, we demonstrate how spatialtemporal disease diffusion patterns can be analysed using SOMs and Sammon's projection. METHODS: SOMs were applied to identify synchrony between spatial locations, to group epidemic waves based on similarity of diffusion pattern and to construct sequence of maps of synoptic states. The Sammon's projection was used to created diffusion trajectories from the SOM output. These methods were demonstrated with a dataset that reports Measles outbreaks that took place in Iceland in the period 1946-1970. The dataset reports the number of Measles cases per month in 50 medical districts. RESULTS: Both stable and incidental synchronisation between medical districts were identified as well as two distinct groups of epidemic waves, a uniformly structured fast developing group and a multiform slow developing group. Diffusion trajectories for the fast developing group indicate a typical diffusion pattern from Reykjavik to the northern and eastern parts of the island. For the other group, diffusion trajectories are heterogeneous, deviating from the Reykjavik pattern. CONCLUSIONS: This study demonstrates the applicability of SOMs (combined with Sammon's Projection and GIS) in spatiotemporal diffusion analyses. It shows how to visualise diffusion patterns to identify (dis)similarity between individual waves and between individual waves and an overall time-series performing integrated analysis of synchrony and diffusion trajectories.

DOI PMID

[21]
Chen N, Ribeiro B, Vieira A, et al.Clustering and visualization of bankruptcy trajectory using self-organizing map[J]. Expert Systems with Applications, 2013,40(1):385-393.Bankruptcy trajectory reflects the dynamic changes of financial situation of companies, and hence make possible to keep track of the evolution of companies and recognize the important trajectory patterns. This study aims at a compact visualization of the complex temporal behaviors in financial statements. We use self-organizing map (SOM) to analyze and visualize the financial situation of companies over several years through a two-step clustering process. Initially, the bankruptcy risk is characterized by a feature self-organizing map (FSOM), and therefore the temporal sequence is converted to the trajectory vector projected on the map. Afterwards, the trajectory self-organizing map (TSOM) clusters the trajectory vectors to a number of trajectory patterns. The proposed approach is applied to a large database of French companies spanning over four years. The experimental results demonstrate the promising functionality of SOM for bankruptcy trajectory clustering and visualization. From the viewpoint of decision support, the method might give experts insight into the patterns of bankrupt and healthy company development.

DOI

[22]
Kohonen T.Self-organized formation of topologically correct feature maps[J]. Biological Cybernetics, 1982,43(1):59-69.CiteSeerX - Scientific documents that cite the following paper: Self-organized formation of topologically correct teaturc maps

[23]
Ju H, Zhang Z, Zuo L, et al.Driving forces and their interactions of built-up land expansion based on the geographical detector - a case study of Beijing, China[J]. International Journal of Geographical Information Science, 2016,30(11):1-20.Scientific interpretation of the driving forces of built-up land expansion is essential to urban planning and policy-making. In general, built-up land expansion results from the interactions of different factors, and thus, understanding the combined impacts of built-up land expansion is beneficial. However, previous studies have primarily been concerned with the separate effect of each driver, rather than the interactions between the drivers. Using the built-up land expansion in Beijing from 2000 to 2010 as a study case, this research aims to fill this gap. A spatial statistical method, named the geographical detector, was used to investigate the effects of physical and socioeconomic factors. The effects of policy factors were also explored using physical and socioeconomic factors as proxies. The results showed that the modifiable areal unit problem existed in the geographical detector, and 4000 m might be the optimal scale for the classification performed in this study. At this scale, the interactions between most factors enhanced each other, which indicated that the interactions had greater effects on the built-up land expansion than any single factor. In addition, two pairs of nonlinear enhancement, the greatest enhancement type, were found between the distance to rivers and two socioeconomic factors: the total investment in fixed assets and GDP. Moreover, it was found that the urban plans, environmental protection policies and major events had a great impact on built-up land expansion. The findings of this study verify that the geographical detector is applicable in analysing the driving forces of built-up land expansion. This study also offers a new perspective in researching the interactions between different drivers.

DOI

Options
Outlines

/

〈 〉