【作者】 李丽娜；

【导师】 达良俊； 由文辉；

【作者基本信息】 华东师范大学 ， 生态学， 2010， 博士

【摘要】 城市化问题,一直是我国经济发展与现代化进程中需要解决的重大理论与政策问题。本文旨在从脆弱性和区域复杂系统分析的角度切入,建立区域城市化脆弱性研究的理论体系,并以此理论为基础评估区域发展中的脆弱度,提出了区域在城市化进程下的脆弱性评估模型,并应用于上海市城市化的实践,进行了实证分析。全文可分为三大部分,研究内容如下：第1章和第2章描述脆弱性研究内容。笔者概括地综述了脆弱性研究的主题与最新发展,梳理其理论发展的脉络,并从区域复杂系统分析的角度提出了城市化背景下脆弱性的内涵。同时,突出了对自然-人工复合生态系统脆弱性研究的新见解,并确定了本研究的研究重点以及和其他领域脆弱性研究的区别。第3章和第4章,首先笔者根据脆弱性的评估内容的研究,建立了评估的框架结构；这一框架以暴露／压力,敏感性,和适应性作为核心概念,阐述了物理脆弱性与社会脆弱性的相互作用及其动力学过程。并运用主成分分析法来确认城市化的主要动因,按照分析结果,城市化主要驱动力可分为4类：(1)土地利用影响；(2)交通影响；(3)人口影响；(4)经济影响。依据这4驱动力综合集成GIS、蒙特卡罗法(Monte Carlo method)和元胞自动机(Cellular Automata)构建了城市化进程中脆弱性评估模型。为了检测UVA算法的实用性,笔者在第5章和附录对上海进行了个案研究。通过UVA算法,CA模型和地理信息系统(GIS)联合使用,对这一区域城市化进程中的生态脆弱性进行了分析。利用UVA算法,分别计算出1985,1990,1995,2000的脆弱性后,通过遗传算法校准,用CA模型对2005年的生态系统脆弱性进行模拟。再将模拟结果和实测2005年生态系统脆弱性进行对比。针对城市区域自然环境和社会环境具体实际情况,本论文既注重对脆弱性自然属性的研究,更加强对其社会经济属性等的综合研究,主要在以下几个方面取得了一些新的研究进展：1.根据城市区域自然环境和社会环境的特征,构建了暴露/压力,敏感性,适应性三要素为核心概念的UVA模型,该模型科学合理、简洁易操作。模型的权重和公式均由历史数据回归生成,人为影响小,便于城市间进行对比研究。2.为减少专家意见差异造成的影响,本研究运用主成分分析法和蒙特卡罗算法科学合理地确定了个驱动力和参数。3.在CA模型中,选择均一自然单元来进行宏观因子和微观因子的数据提取及量化,能真实反映研究对象的本质特征。4.对比实测和拟合结果,R2=0.94,拟合情况十分理想。结果表明,UVA算法结合CA模型的方法,真实可靠,有较强的实际应用价值,可用于长期预估。这个项目的长期目标是建立一种可以运用于各个大城市的脆弱性评估方法。通过这种方法。可以(1)描述和分析城市化进程给生态系统和社会系统带来的影响,(2)为后续城市设计开发和政策措施评估,提供指导和计量工具,以防止或减轻不良影响。更多还原

【Abstract】 Economic development and modernization in our country have created significant urbanization problems in both theoretical fronts and policy fronts that need to be solved. The purpose of this dissertation is to develop a model that can measure region urbanization vulnerability quantitatively and unambiguously, and demonstrate its application by applying it to Shanghai. This dissertation can be divided into three parts as follows:Chapters 1 and 2 describe the pre-existing research content on vulnerability. After reviewing the footprint of vulnerability, this dissertation will present a theoretical reconceptualization of vulnerability from a region system complexity angle. At the same time, it will highlight new insights into the conceptualization of the vulnerability of human-environment systems and identify critical points of convergence of what otherwise might be characterized as disparate fields of research.In Chapters 3 and 4, based on the reconceptualization of vulnerability, the heart of this dissertation is introduced:the urban vulnerability assessment (UVA) framework. This framework will include the elements of exposure, sensitivity and adaptability as the core of the reconceptualization and dissect the dynamics of and mutual interactions between social and biophysical vulnerabilities. By principal components analysis, four key drivers have been identified in this framework:(1) the self propensity driver, (2) transportation driver, (3) demographic driver, and (4) economic driver. The UVA model will be developed by integrating GIS, Monte Carlo method and Cellular Automata (CA) to assess vulnerability by these four drivers in human-environment system under urbanization processTo examine the applicability of UVA, Chapter 5 presents a case study of Shanghai, in which the concepts and models developed are applied using UVA model as a quantitative tool for measuring vulnerability. GIS and other real data from 1985,1990,1995, and 2000 are used to produce UVA results for those years, with which the CA model is run with Monte Carlo calibration to simulate a 2005 prediction. This prediction is then compared with the real 2005 vulnerability map of Shanghai, validating the UVA model. Pre-research provided an in-depth look into the social and natural characteristics of Shanghai. Based on the integrated studies we did, we emphasize social-economic attributes to vulnerability, not only the natural attributes. We have made some improvements in this field as follows:●According to the social and natural characteristics of Shanghai, we established an urban vulnerability analysis model (UVA) on the basis of exposure, sensitivity and adaptability. The UVA is scientific, and operational. Because most of the weights and equations are regressed from statistical data, without human impact, results between cities will be more comparable.●To avoid the effect of measuring differences between experts, in this dissertation we didn’t use any scoring method. Instead we used cumulative contribution from principal components analysis to identify driving forces, and decided parameters by GA method.●In CA model we adopt the smallest meaningful cells to derived and standardized data from micro and macro scale. This cell size can reflect the essential characteristics of the studied objects.●The predicted results compared to real vulnerability data very closely (R2=0.94). Although some details were missing, the general patterns were captured fairly well. This means that long-term predictions can be put into use.The long term objectives of this research are to develop a methodology and operational procedure applicable on the level of the individual metropolitan (1) to quantitatively and unambiguously describe and analyze the impacts of urbanization on its territory, ecosystems, and social systems, and (2) to provide instruments or tools to design, explore and evaluate policy measures and policy interventions to prevent or alleviate undesirable impacts.更多还原