【作者】 徐雁南；

【导师】 薛建辉；

【作者基本信息】 南京林业大学 ， 生态学， 2011， 博士

【摘要】 本文系统总结了生态园林城市建设与发展过程,生态园林城市评价方法和3S技术在生态园林城市评价中应用等相关研究进展；以常州市为例,在已有国家生态园林城市评价体系的基础上,采用统计方法构建并优化生态园林城市评价指标体系。以该指标体系为指导,借助于3S技术,拓展和深化城市绿地的空间分析与定量评价方法,形成了一套生态园林城市综合评价理论和方法；应用该指标体系对常州市生态园林城市建设水平进行综合评价,并提出绿地建设的规划对策。主要研究结果如下：(1)在已有国家生态园林城市评价标准(试行)的基础上,采取主观与客观、定性与定量相结合的指标筛选方法初选出72个生态园林城市的评价指标,并应用聚类分析、主成分分析和相关性分析等方法分别对初选指标进行筛选与优化,最终选取了25个指标,并运用层次分析法赋予每个指标相应的权重,构建了一套生态园林城市评价指标体系。(2)对2005年常州市建成区1：10000的航片影像进行目视解译,提取建成区绿地空间信息。由斑块面积分析可知,建成区内绿地斑块较为破碎,中、小型斑块占绿地斑块总数的98.25%,占总面积的72.17%；由绿地类型分析可知,居住绿地、其他附属绿地和生产绿地的斑块数较多,占总斑块数的87.81%,占总面积的81.41%；从植被组成角度看,乔木和灌木数量占绿地总数的92.54%,面积占绿地总面积的83.24%。同时,利用2005年TM影像,采用Isodata无监督分类方法提取市域范围内城市绿地、农田和建设用地等5类地物信息,分类精度为91.12%,kappa系数为0.88,其中市域范围内较大面积的绿地斑块主要分布在武进区西部和南部以及新北区西部。(3)采用城市绿地连通性指标对绿地斑块的敏感度进行空间分析与评价,以此来确定城市敏感区的范围和等级。分别选择10hm2、5hm2、20hm2和25hm2四个最小斑块面积等级和200m、400m、800m、1200m和1600m五个连通阈值,计算不同参数组合下9个连通指数的重要值,并利用Spearman秩相关系数和PIOP值优选AWF (Area-weighted flux)作为分析景观连通敏感度指数,同时确定了400m连通阈值和10 hm2最小斑块面积的最优组合。分析表明,胡家山-中大湾和西湾里-丁家坞区均为连通高敏感区域。(4)着重对工业污染指数、路网布局状况和人口密度分布等指标进行空间扩展,确定了道路交通、水系分布、古树名木等7个生态因子,采用层次分析法AHP(Analytic Hierarchy Process)模型和频率比FR (Frequency Ratio)模型对研究区绿地进行适宜性分析,引入ROC (Receiver Operating Characteristic Curve)曲线对两种模型的计算结果进行检验和优选。3组AHP模型和FR模型所计算的ROC曲线面积分别为：0.604、0.612、0.597和0.724,其中FR模型预测精度较高；利用该模型对常州市绿地适宜性进行分析评价,高适宜区主要分布在城市的西部以及北部,不适宜区主要分布城市中心城区以及周边地带；根据面积统计,高适宜区占总面积的13.14%,不适宜区占总面积的17.75%。(5)根据常州生态园林城市建设的特点和已构建的指标体系中每个指标的加权评价值获得常州生态园林城市建设综合得分,其中2005年常州生态园林城市建设水平综合评分为0.6741,按照生态园林城市建设分级表,常州市的城市生态园林建设水平等级为：Ⅳ级,处于生态园林城市建设中下等水平,与2008(0.7216)、2009(0.7224)年进行比较,确定指标体系的可行性；由绿地斑块空间布局综合分析可知,研究区内高适宜地区的绿地斑块聚集度较高,斑块面积较大、破碎化程度较低,其最大斑块指数(LPI)、聚集度指数(AI)和平均斑块面积指数(MPS)明显高于低适宜地区；大于10 hm2以上的高敏感和较高敏感的绿地斑块主要分布在高适宜区域,而低适宜区域受干扰较为严重,绿地斑块的连通敏感度较低,是城市绿地系统规划中需重点关注的问题之一。更多还原

【Abstract】 In this thesis, the author systematically summarizes the research progress of ecological garden city’s construction and development, evaluation method for ecological garden city and 3S methodology’s application in evaluating the ecological garden city, etc. Based on the existing evaluation system of the national ecological garden city, the statistic approach has been used to build and optimize the evaluation index system of the ecological garden city. This index system is applied as the guidance along with the 3S methodology to explore and deepen the space evaluation system of the urban green space, which forms a set of theories and methods to comprehensively evaluate the ecological garden city. Furthermore, the index evaluation system mentioned above is used to comprehensively assess the construction of Changzhou city, and some strategic recommendations are also suggested for the green space construction. The main results were described as follows:(1) Based on the characteristic of Changzhou’s ecological garden city construction and the evaluation standard for existing national ecological garden city, it uses the index selection method by means of combining subjective and objective, qualitative and quantitative models, to edit, reinforce and optimize 72 ecological garden city evaluation indices, and uses the principal component analysis, factorial analysis and cluster methods respectively to optimize the original indices and choose 25 indices derived by the principal component analysis, and then uses the AHP method to establish the ecological garden city evaluation index system and assign corresponding weight to each index. This index system consists of 3 first level indices and 25 second level indices (Including landscape connectivity, distribution status of road network and distribution status of population density, etc).(2) Using some indices from the index system constructed by the space data collection, such as remote sensing images, the author conducts visual interpretation for the 1:10000 aerial images of the built-up area in Changzhou city in 2005 and selects the green space information of the built-up areas. According to the patch areas, green space patch of Built-up areas is relatively fragmentized, the total number of medium and small size green space patches accounts for 98.25% of the total number of green space patches, but their total area only takes up 72.17% of the total green space area. According to the green space type, occupancy green space, other afflicted green space and manufactured green space which accounts for 87.81% of the total number of green space patches, but their total areas only takes up 81.84% of the total green space area; According to the composition of vegetation, the number of trees and shrubs account for 92.54% of the total number of the green space and their areas occupy 83.24% of the total area of green space; meanwhile using the TM images from 2005, the thesis adopts the Isodata unsupervised classification to select 5 ground feature information from the area of the city, namely urban green space, farmland, construction area, water body and other sites. The total precision degree after classification is 91.12%; the kappa coefficient is 0.88. The green space patches that have relatively large areas in the studied area mainly are located in the western and southern areas of Wujin district, western areas of Xinbei district.(3)City green space connectivity indice are used to analyze and evaluate sensitivity of green space patches, and identify city sensitivity areas and levels. Selecting 4 minimum levels of patch areas including 10,15,20 and 25 hm2, and also 5 connectivity thresholds including 200,400,800,1200 and 1600 meters, the node importance of 9 connectivity indices as calculated, and Spearman Rank Correlation Coefficient and PIOP value are used to optimize AWF index (Area-weighted flux) as the index of analyzing the connectivity sensitivity of landscape. Meanwhile, the best combination of the 400m connectivity threshold and the 10hm2 smallest patch area is determined. The results show that Hujia Mountain-Zhong Da Wan and Xi Wan Li-Dingjiawu District are the most sensitive areas, the green space sensitive areas with relatively high sensitivity are mainly located at the Northwestern area of the city, and that the green space patches in the northern and eastern areas of the city are relatively scattered and have the lowest connectivity sensitivity.(4) In this paper ten levels of single factor analysis are conducted for the seven factors such as traffic condition, water body distribution, old and rare trees in the research area. According to the grading forms from 3 groups of specialists, the author constructs green space suitability distribution pattern based on AHP model and FR model by 3 Scoring tables, uses ROC curve to test the result of 4 types of suitability. The results show that 3 groups ROC linear areas calculated by the AHP model and FR model are:0.604,0.612,0.597 and 0.724. Among them, the FR model’s prediction accuracy has a relatively high reliability. The research uses this model to analyze and evaluate the suitability of Changzhou’s green space. Highly suitable districts are mainly located at the western and northern areas of the city and unsuitable districts mainly located at the central and surrounding areas of the city. According to the area statistics, the highly suitable districts account for 13.14% of the total studied area and the unsuitable districts take up 17.75% of the total area.(5) According to the characteristics of Changzhou ecological garden city construction and the weighted evaluation value of each indicator the comprehensive score of Changzhou ecological garden city construction is obtained. The composite score of Changzhou’s construction level of ecological garden city in 2005 was 0.6684. According to the classification table of ecological garden city construction, Changzhou’s construction level of ecological garden city is at levelⅣ, which is the middle level of the ecological garden construction, compared to the data of 2008 (0.7160) and 2009 (0.7168) in order to determine the feasibility of the index system. Among them, the evaluation values of the urban ecological environment, urban living environment and urban infrastructure are 0.6297,0.7212 and 0.7115 respectively. The index of the urban ecological environment takes the largest weight and has a relatively large impact on the index system. In this index level, several factors, such as the green coverage in the built-up area, landscape’s connectivity, public green space area per capita in the built-up area and native plants index, have relatively large impacts on it. According to the comprehensive analysis of the green space distribution, the green_space patch density of highly suitable districts in the studied area is relatively high, and the patch area is relatively large with the relatively low level of fragmentation. Its largest patch index (LPI), aggregation degree index (AI) and mean patch area (MPS) are clearly higher than those of the low suitable districts. The green space patches with high sensitivity and relatively high sensitivity greater than 10hm2 are mainly located at highly suitable areas, but low suitable districts are seriously interfered and the connectivity sensitivity of the green_space patches are relatively low, which would become one of the mainly focused regions in the urban green space system planning.更多还原