基于遥感和图论的绿地空间演变和连通性研究——以郑州为例

Dynamic changes of green-space connectivity based on remote sensing and graph theory: a case study in Zhengzhou,China

绿地空间是人类和自然生态环境的重要组成部分,其连通性是绿地有效发挥生态和社会服务功能的重点,对促进生物多样性保护、人类户外空间游憩和城乡空间的健康可持续发展具有重要意义。以郑州为例,基于图论的原理和算法,以4个时期30m分辨率的Landsat 5和Landsat 8影像为基本数据源,在Arc GIS 10.2、ENVI 5.0和Graphab 1.2.1软件中,提取郑州市域范围内的绿地空间信息,分析了绿地斑块规模、数量和空间分布特征的动态变化过程,揭示了绿地空间连通性动态变化的规律和原因,探讨了图论指数在诊断和优化绿地空间连通性上的应用。结果表明:2000年到2013年间,随郑州市域绿地空间总量和斑块规模的增加,绿地空间整体连通性指数增加了4.7倍,且在2009年到2013年间增加幅度最大;绿地斑块规模与绿地空间整体连通性指数、最大连通单元体规模和绿地斑块节点度指数呈显著正相关关系;绿地空间连通性的区域差异特征与绿地空间分布特征一致,均呈现西高东低的分布规律,且随时间变化呈现出由西向东扩展的趋势;图论指数可以用来反映各种尺度上绿地空间的连通性,其中绿地斑块的节点度指数和中间度核心性指数可以用来诊断和筛选对整体绿地空间连通性起关键性和脚踏石作用的绿地斑块,在今后城乡绿地系统规划和绿地生态网络优化上具有重要的应用前景。

Green space is the main part of human and natural ecological environmental systems. The connectivity of green space is fundamental for the survival of flora and fauna, human outdoor recreation, and urban sustainability. The objective of this study was to quantify the dynamic changes of green-space connectivity of Zhengzhou, including the rapid urbanization processes, determination of external and internal causes, and analysis of the graph theory applied to the study of green- space connectivity diagnosis and optimization bu using remote sensing data and graph metrics. The results indicated that green-space connectivity in Zhengzhou increased by 4.7-fold with the growth of the green-space area and patch size from 2000 to 2013, with the largest increment in 2009--2013. The green patch size was positively correlated with the integral index of connectivity, size of the largest component, and node degree. The characteristics of green-space connectivity in different urban districts were consistent with the distribution of green space. Graph metrics were calculated using Graphab software and to reflect the green-space connectivity in different scales. Node degree and betweenness centrality can be used to reflect the location and changes of the keystone green patches, which need more consideration in the next step of green- space planning. Therefore, the application of graph theory combined with remote sensing is important for the future green- system planning and optimization.