【作者】 张振国；

【导师】 温家洪；

【作者基本信息】 上海师范大学 ， 环境科学， 2014， 博士

【摘要】 随着全球气候变暖和城市化进程的加快，城市暴雨内涝已引起各国政府和学者的高度关注。社区作为组成现代城市的基本单元，在城市减灾降险中具有重要的基础作用。因此，以社区为基础的灾害风险管理成为近年来国际社会普遍认可并被实践证明是行之有效的管理灾害的理念与手段，而风险评估作为社区灾害风险管理的基础和前提则成为各国学者探讨的热点问题之一。本文以国家自然科学基金重点项目“沿海城市自然灾害风险应急预案情景分析（40730526）”和教育部人文社科青年基金项目“基于PGIS的社区灾害风险管理模式研究（11YJCZH241）”为依托，在实地考察和调研典型城市社区（上海普陀区金沙居委地区）暴雨内涝灾害及其风险管理现状，并获得大量文献资料和一手数据的基础上，综合运用PGIS方法、情景分析方法和概率统计方法开展了典型城市社区暴雨内涝灾害风险评估的实证研究。主要研究工作和结论如下：1.提出了城市社区暴雨内涝灾害风险评估的理论框架与方法体系。在对城市社区、暴雨内涝和灾害风险等基本概念进行分析与界定，并对灾害风险评估的相关理论进行梳理与分析的基础上，构建了城市社区暴雨内涝灾害风险评估的理论框架，认为城市社区暴雨内涝灾害风险评估就是在风险评估专家和当地不同利益相关者共同参与下，综合运用PGIS（参与式GIS）、情景分析和概率统计等方法，对研究区不同情景下的内涝灾害风险进行模拟、分析、评价与表达的全过程，主要包括风险识别、风险模拟和风险评价与表达等3个主要内容，分8个步骤完成。2.构建了城市社区暴雨内涝灾害风险评估模型。在全面分析城市暴雨内涝成因和参考已有内涝模型的基础上，构建了城市社区暴雨内涝灾害风险评估模型，主要包括内涝灾害数据库、城市社区暴雨内涝模型和城市社区风险模型等三个模块，并通过重建研究区2012年8月台风“海葵”的内涝灾情，验证并修正了模型参数，使模拟结果与实际情况较为吻合。3.对上海城市暴雨内涝灾害历史灾情进行了分析。基于历史灾情数据，对上海市近60年（1951-2010）来暴雨内涝的历史灾情进行了分析，结果表明：（1）城市自然环境受人为影响较大，主要表现为不透水面增大，地面沉降和水域面积减少，加上经济发达、人口众多和密布的城市基础设施，大大增加了涝灾风险；（2）从时间上看，暴雨内涝灾害发生数量总体呈现增长的趋势，且主要集中在6-8月，从空间上看，暴雨内涝灾害以中心城区最多，沿江沿海次之，内陆区县相对较少；（3）暴雨内涝灾害具有较强的扩散效应，其扩散的通道主要有城市生命线系统、产业链和生态环境系统等，且主要通过直链式扩散、直链发散式扩散、循环式扩散和发散式扩散等四种扩散方式对城市不同类型的承灾体产生影响。4.选取上海普陀区金沙居委地区为例，运用PGIS方法对研究区居民的暴雨内涝灾害风险认知进行了全面的分析。运用基于PGIS的半结构式问卷调查、访谈、实地测量等方法，对社区居民的暴雨内涝灾害风险认知进行了分析，主要包括风险知识、风险态度和风险行为等三方面的内容。分析结果表明：（1）社区居民认为暴雨内涝灾害每年平均发生3-4次，且多集中在7-9月份，主要影响集中在研究区中部区域；（2）社区居民认为当积水深度在脚踝及以下时（0-15cm），对家庭有一定的影响，在小腿肚及以下时（15-30cm），对家庭影响较大，超过小腿肚时（＞30cm），对家庭影响最大；（3）一半以上居民认为内涝灾害形成是自然原因，所有居民在灾前、灾中和灾后均采取不同的措施，但均没有认识到或不重视排水管道的清理与疏通问题。5.运用情景分析方法对研究区暴雨内涝灾害风险进行了分析。基于ArcGIS软件，运用情景分析方法对研究区4种排水情景下（0mm/h，18mm/h，36mm/h和50mm/h）的8种重现期（5a，10a，30a，50a，100a，200a，500a和1000a）暴雨的内涝风险进行了模拟与分析，主要包括危险性分析，暴露分析和脆弱性分析等三部分内容。危险性分析表明：（1）所有内涝情景下，暴雨内涝危险性均集中在研究区中部的金沙新村和解放村，积水面积占研究区总面积的比例在0%－47.8%，最大积水深度在0-0.52m之间，但均随着暴雨重现期的增长而增加，随着排水量的增大而减小；（2）32个内涝灾害情景中，高危险性占21个，中危险性占9个，而低危险性仅占2个；（3）4种排水情景下，200a及以上重现期暴雨的内涝危险性均具有高危险性；（4）随着排水量的增大，暴雨内涝的中危险性年超越概率由100%减小为6.8%，高危险性年超越概率由26%减小为0.8%。暴露性分析表明：（1）各种内涝情景下，暴露的家庭数量最多的是791户，最少的为0户，但均位于研究区中部，建筑结构主要以砖木结构为主；（2）暴露家庭数量随着暴雨重现期的增长而增多，随着排水量的增大而减少，但减少的程度与重现期大小成反比；（3）在4种排水情景下，200a、500a和1000a一遇暴雨分别有30户、137户和216户家庭均暴露在高危险性区域，是防涝减灾中需要重点关注的对象。脆弱性分析表明：（1）研究区暴雨内涝的脆弱性主要体现在居民家庭的墙面、地板和室内财产等三个方面；（2）各种内涝情景下，绝大多数暴露家庭的的墙面损失率在0-0.06之间，地板损失率均为1；(3）只有高危险性内涝情景下部分暴露家庭的室内财产会遭到损失，其损失率分布在0-0.12之间。6.对研究区多种情景的内涝灾害风险进行了评估。在风险分析的基础上，对各种情景的内涝灾害风险进行了分析与评价，结果表明：（1）绝大多数家庭的内涝损失均在0-1000元之间，且主要位于研究区中部的金沙新村；（2）各类损失均随着暴雨重现期的增大而增长，随着排水量的增大而减小，其中室内财产损失增长和减小均最快，其次是地板损失，墙面损失增长和减小均最慢；（3）根据四种排水情景下8种重现期建立的超越概率－损失曲线，得出研究区年平均内涝损失分别为：0mm/h排水情景下为25895元，18mm/h排水情景下为10672元，36mm/h排水情景下为6781元，50mm/h排水情景下为1937元。城市暴雨内涝灾害风险评估是一个复杂的系统工程，涉及灾害学、风险学、地理学、环境学和社会学等多个学科的理论与方法。本文只是从社区的角度对城市暴雨内涝灾害风险评估做了部分有益的探索，相对于本研究领域科学理论的拓展和现实问题的解决还仅仅是一个开端。今后，还需要在居民参与和当地风险知识采集与应用、城市不同尺度风险评估方法与范式、风险动态评估和实证研究等方面进行深入的研究与探讨。更多还原

【Abstract】 With global warming and the acceleration of urbanization, urban storm waterlogging has aroused great concerns of governments and experts. As the basic unit tomake up the modern city, the communities play an important role to reduce urbandisaster risk. Therefore, community-based disaster risk management has beenrecognized generally and is proven to be effective disaster management conceptsand means in recent years. Meanwhile, as the basis and premise ofcommunity-based disaster risk management, risk assessment has been one of thehotspots by scholars from various countries. In this paper, based on National NaturalScience Foundation of China “scenario analysis of natural disaster risk contingencyplans in coastal cities (40730526)” and Humanities and Social Sciences Youth FundProject of the Ministry of Education “PGIS-based community disaster riskmanagement model research (11YJCZH241)”, with the case of field study andresearch on typical urban communities (Jinsha Community Putuo District in Shanghai)and access to a large number of documents and first-hand data, an Empirical Study ofthe disaster risk assessment on typical urban community water logging are carriedout with use of PGIS methods, scenario analysis and probabilistic method. The mainresearch work and conclusions are as follows:1. Disaster risk assessment theory and methodology of urban communities’water logging are presented. Based on the analysis and definition of fundamentalconcepts of urban communities, water logging and disaster risk, and relative theoriesof disaster risk assessment, the framework of urban Community storm water loggingdisaster risk assessment is constructed which is the whole process of simulation,analysis, evaluation and expression of water logging disaster risk under differentscenarios in the study area with the participation of risk assessment experts and localstakeholders and integrated use of PGIS (participatory GIS), scenario analysis andprobability statistics. It mainly includes three major content points of riskidentification, risk modeling and analysis, and risk assessment and the expression,and is fulfilled by eight steps. 2. Urban communities’ water logging disaster risk assessment model isconstructed. Based on comprehensive analysis of the causes of urban water loggingand reference of existing water logging model, urban communities water loggingdisaster risk assessment model is built. It mainly includes three modules of waterlogging disaster database, urban communities’ water logging model and the urbancommunity risk models. Meanwhile, model parameters is verified and correctedthrough reconstruction of typhoon "anemone" water logging disaster in the studyarea in August2012which make the simulation results more consistent with theactual situation.3. Historical waterlogging disasters were analyzed in Shanghai. Based onhistorical data, historical waterlogging disasters were analyzed in Shanghai for nearly60years (1951-2010), which showed that:(1) urban natural environment was greatlyinfluenced by human activities, mainly representing as the impervious surfaceincreasing, ground subsidence and water area reduction. Meanwhile economicdevelopment, big population and densely urban infrastructure greatly increase therisk of floods;(2) from the perspective of time, the overall number of waterloggingdisasters was a growing trend and concentrated in June to August. From the spatialpoint of view, waterlogging disasters were the most at the central city, followed bythe coast, and inland counties were relatively the fewer;(3) Waterlogging disasterhas a strong diffusion effect. The diffusion channels are mainly composed of urbanlifeline systems, industrial chain and the ecological environment system, and etc. Andthe diffusion has influence on a variety of urban hazard-bearing body throughstraight diffusion, linear divergent diffusion, circulation diffusion and divergentdiffusion and so on.4. A comprehensive analysis of the residents’ perception on water loggingdisaster risk of the study area of Jinsha Community Putuo District in Shanghai withthe use of PGIS is conducted. In the detail, based on semi-structured questionnaires,interviews and field measurement methods of PGIS, community residents’ awarenessto water logging disaster risk is analyzed. It mainly includes three aspects of riskknowledge, risk attitudes and risk behavior. The results show that:(1) community residents believe that rainstorm water logging disasters each year occur3-4times onthe average, more concentrate in July to September and the main impact is in thecentral region of the study area;(2) community residents believe that there is someinfluence on families when the water depth is in the ankle and below (0-15cm), thereis greater influence on families when the water depth is in the calf and below(15-30cm) and there is the greatest influence on families when the water depth isover the calf (>30cm);(3) more than half of residents believe that water loggingdisasters is caused by nature. All residents take different measures in thepre-disaster, disaster and post-disaster, but don’t recognize or don’t attachimportance to the drains cleaning and clearing.5. Rainstorm water logging disaster risk in the study area is analyzed by use ofscenario analysis method. In the detail, rainstorm water logging risk in the study areain8return periods (5a,10a,30a,50a,100a,200a,500a and1000a) under4drainagescenarios (0mm/h,18mm/h,36mm/H and50mm/h) is simulated and analyzedbased on ArcGIS software and with the use of scenario analysis method. It mainlyincludes three sections of risk analysis, exposure analysis and vulnerability analysis.Risk analysis shows that:(1) under all scenarios, water logging risks are allconcentrated in Jinsha New Village and Jiefang Village which are in the central studyarea. The water area account for0%-47.8%of the total study area. The maximumaccumulation of water depth is between0and0.52meters which increases alongwith rainstorm growth in the return period and decreases along with the increase ofdisplacement.(2) Among32water logging disaster scenarios,21ones are high risk,9are middle risk and only2are low risk.(3) Among4drainage conditions, waterlogging risks at the return periods with200a and over are all high.(4) As thedisplacement increases, annual probability of exceedance of middle risk of waterlogging is reduced from100%to6.8%, annual probability of exceedance of high riskis reduced by26%to0.8%. Exposure analysis shows that:(1)Among a variety ofwaterlogging scenarios, the most households which is exposed to the risk are791and the least are0which are all located in the central study area with brick andwood houses.(2) The number of exposure households increases as the return periods of rainstorm increase and decreases as the displacement increases. But thedegree of reduction is inversely proportional to the size of the return periods.(3) atthe4kinds of drainage conditions,30,137, and216families in200a,500a and1000aare exposed to high-risk areas when there were rainstorms which we should focus onin water logging prevention and mitigation. Vulnerability analysis shows that:(1)waterlogging vulnerability is mainly reflected in three aspects of the households ofwalls, floors and indoor property;(2)under a variety of waterlogging scenarios, wallsurface loss rate of the majority of exposed families is between0-0.06, floor loss rateis1;(3) only under high risk water logging scenarios, indoor property of a fewexposed families would suffer a loss and the loss rate distribution is between0-0.12.6. Waterlogging disaster risk under a variety of scenarios in the study area isexpressed and evaluated. In the detail, based on risk analysis, Waterlogging disasterrisk under a variety of scenarios is expressed and evaluated from the perspectives oftime and space. research results show that:(1) waterlogging loss of the vast majorityof families is between$0-1000and these families are mainly located in Jinsha NewVillage in the central study area;(2) All types of loss increase with the increase of thereturn period rainstorm and decrease with increasing displacement, of which indoorproperty loss grows and reduces the fastest, followed by the loss of floor and wallloss grows and reduces the slowest;(3)According to the probability of exceedance-loss curve established on8return periods under4drainage scenarios, it is concludedthat average annual loss of waterlogging:25,895yuan under0mm/h drainagescenarios,10,672yuan under18mm/h drainage scenarios,6781yuan under36mm/h drainage scenarios and1937yuan under50mm/h drainage scenarios.Urban water logging disaster risk assessment is a complex systematic projectinvolving theories and methods of disasters, risk, geography, environmental studies,sociology and other disciplines. In this article, urban storm water logging disaster riskassessment is did some useful exploration from the perspective of community whichis only the beginning with respect to the expansion of scientific theory and realisticsolution to the problem. In the future, some works need to be studied and discussedin-depth in these aspects such as residents participation and local risk knowledge acquisition and application, risk assessment methods and paradigms at the citydifferent scale and dynamic risk assessment and empirical research.更多还原