【作者】 吕永鹏；

【导师】 杨凯；

【作者基本信息】 华东师范大学 ， 环境科学， 2011， 博士

【摘要】 随着工业和生活污染源等点污染源逐步得到控制,非点源污染已成为影响城市水环境质量的最主要因素,得到各国政府和学者的重视。在这一背景下,非点源污染过程模拟以及其系统调控管理已经成为当前环境领域的热点问题和前沿领域,引起学界的广泛关注。独特的水文地貌特性使得城市化对平原河网地区城市非点源污染过程的影响深刻。论文以中国东部平原河网地区的上海市临港新城为例,从环境科学、生态学、水文学及管理学的交叉视角,运用理论研究与案例分析相结合、室内模拟与室外试验相结合、数学模型与计算机仿真相结合的方法,以平原河网地区集水区特征与划分为切入点,围绕非点源污染的“源-过程-汇”和“试验-模拟-调控”两条主线,阐释了平原河网地区城市集水区非点源污染的过程与特征,构建了平原河网地区城市集水区非点源污染调控模型,尝试提出了平原河网地区城市集水区非点源污染的系统调控方法、原则与核心策略。论文主要研究结论可概括为如下几个方面：1.基于平原河网地区所特有的且取得共识的水利分区/片管理体系,分析了平原河网地区城市集水区的特征,从空间尺度、汇流途径和调控强度三个维度,初步探讨了面向非点源污染调控的平原河网地区城市集水区四级划分体系：一级集水区从大尺度、高强度调控来体现研究区域与水利片的关系,重点关注重要外河；二级集水区从中尺度、中强度调控来体现功能分区和排水系统分布的关系,重点关注内河：三级集水区从小尺度、低强度调控来体现小区域地形与集水区出口的关系,重点关注小区域河流；四级集水区主要用于数据支持和微观机理阐释。以临港新城集水区为例,划分了1个一级集水区,13个二级集水区,137个三级集水区；二级集水区与现有规划的功能分区基本一致,可保证后续调控策略的落实。2.论文从理论上探讨了平原河网地区城市集水区非点源类型划分与过程概化方法；在此基础上,运行了大量室外试验和室内模拟的第一手数据,阐释了平原河网地区城市集水区非点源污染的过程与特征。(1)与国际上经典的“有效不透水面(Effective Impervious Areas, EIA)”相对应,从非点源调控角度提出“有效透水面(Effective Pervious Areas, EPA)’概念,即可接纳客地径流的透水面；进而引申出有效绿地(Effective Green Area)的概念。(2)分析探讨了城市不透水面的初始冲刷效应。临港新城集水区非点源污染总体输出水平较上海中心城区低；不透水面地表径流污染物中,溶解态污染物浓度低且初始冲刷效应不显著,可能与研究区域为新建区域有关；而颗粒态污染物浓度高且初始冲刷强度较大,与国内外相关研究结果基本一致。土地利用类型、降雨强度等因素对初期冲刷强度有显著影响。(3)重点阐释了城市透水面的源汇效应。理论上讲,城市透水面既是“源”也是“汇”,而现有研究主要聚焦于其汇效应,对其源效应探讨较少。a)就“源”而言：论文中的市区既有绿地地表径流污染物浓度总体较低,均低于国家地表水环境质量标准的Ⅴ类标准限值,而其初期效应不显著；而论文中的郊区新建绿地地表径流污染物浓度总体较高,大多劣于Ⅴ类标准限值,甚至部分点位COD浓度为上海市污水综合排放标准的二类标准限值的3.56倍,而其初期效应处于中度水平。因此,可以认为包括城市绿地在内的城市透水面是一种潜在的污染源,污染强度与绿地所处区域、绿地兴建时间、绿地土壤来源等因素有关。b)就“汇”而言：论文试验表明城市绿地对地表径流污染物具有显著的削减效果。COD的总体削减率(71.01±1.75%)略高于TP(69.00±1.78%),而显著高于氨氮(64.39±1.84%)；市区既有绿地显著高于郊区新建绿地(p<0.05)；不同覆被对地表径流污染物的削减效果显著(p<0.05),且总体削减率依次为麦冬>结缕草>无覆被；绿地对污染物的削减效应主要发生在深度60-80 cm的土层之上,并且以物理过程为主导。此外,污染负荷、水力负荷、停留时间、落干时间等因素对削减效应有一定影响。c)源汇转化条件：理论上讲,城市透水面既是“源”也是“汇”,非有效透水面以源效应为主,而有效透水面以汇效应为主；具体而言,设置路牙且客地径流难以汇入的绿地为源,而有客地径流汇入且对其起到削减作用的绿地为汇,路牙的设置与否是源汇转化的关键节点；减少路牙等因素的限制,促进客地径流的汇入是城市透水面由源向汇转变的关键途径；通过合理配置透水面的覆被植物、土壤类型和土层深度,合理引导径流汇入流量、流速和浓度,合理设置径流停留时间和落干时间,可以有效提升城市透水面对污染物的削减效应。d)以土壤盐度为例,分析探讨了降雨对城市透水面关键胁迫因子的淋溶效应。城市透水面土壤盐度在小尺度上具有时空异质性,其淋溶效应具有显著的城郊差异。本案例中的上海市区既有绿地盐化程度较低,704个表层土壤样点全部处于中度盐化以下水平；郊区滨海新建绿地盐化程度高,240个样点中,40.0%为盐土,且微域空间变异特性和斑块状盐渍化显著。市区和郊区绿地表层土壤平均变异系数分别约37.7%和67.9%。绿地土壤表层盐度季节分异显著(p<0.05),均值表现为春>冬>秋>夏,且季节间相关关系显著(p<0.01)表明存在季节协同变化趋势。中心城区既有绿地淋溶效应不显著,出水盐度处于淡水级别；而郊区新建绿地淋溶效应显著,处于混盐水级别；降雨强度和降雨量等因素对淋溶效果影响显著。e)临港新城集水区3类主要的非点源包括地表径流(又称土地利用污染源)、降雨以及分散排放的农村生活污水,地表径流是其最主要的非点源。近几十年来,集水区地表径流污染的产生负荷和输移通量均呈现显著增加态势,可初步划分为三个变化阶段。第一阶段为1965-1994年,非点源污染的产生负荷和输移通量总量较小,且变化较小；第二阶段为1995-2003年,非点源污染问题逐步显现；第三阶段为2004-2008年,非点源污染问题凸显。究其原因：一方面区域大规模、快速开发使得土地利用结构变化从而导致“源”产出负荷显著增加,另一方面城市汇的生态功能尚未有效发挥,绿地等“汇”对非点源的缓冲和去除效应平均仅为2.07%。3.论文整合系统动力学、多目标优化、情景规划、利益相关方调查、GIS等5种模型和方法,尝试构建了平原河网地区城市集水区非点源污染调控模型,并以上海市临港新城为典型案例进行实证研究。(1)指出平原河网地区城市集水区非点源污染调控具有五方面的不确定性：非点源产生的时空不确定性导致“源”调控的不确定性、非点源污染输移过程的不确定性导致“过程”调控的不确定性、非点源污染“汇”的不确定性导致“汇”调控的不确定性、城市集水区非点源污染的调控涉及的决策者偏好和利益相关方意愿的不确定性、城市集水区非点源污染的调控手段和途径本身具有较大的不确定性。(2)对于新开发区域而言,试图从土地利用结构方面对现有土地利用规划进行较大调整来实现非点源调控,可能难度极大。论文提出了从提升城市透水面“汇”的生态功能并辅以初期径流纳管措施的角度来进行非点源调控的思路。(3)初步探讨了系统动力学、多目标优化、情景规划、利益相关方调查、GIS等5种模型和方法的耦合机制,构建平原河网地区城市集水区非点源污染调控模型的4步耦合方法：子集水区划分和子系统分析、子集水区未来发展情景设定、集水区非点源时空模拟和预测、集水区非点源调控优化方案确定。(4)设定现状情景S1、功能提升情景S2-S3(考虑有效绿地和有效湿地)以及综合削减情景S4-S5(有效绿地、有效湿地和初期径流纳管的组合)等5种典型情景方案,并将上述调控模型应用于临港新城集水区,模拟结果表明：a)仅通过有限的功能提升措施,不辅以初期径流纳管措施,对非点源的调控效果并不显著。以2020年为例,相对于现状情景S1的COD输移通量15564 t/a和径流系数0.42,情景S2和S3分别削减了7.5%、7.8%的输移通量和1.3%、2.6%的径流系数。b)通过功能提升措施并辅以初期径流纳管措施,不仅实现非点源污染物输移通量的大幅度削减,还使得径流系数“零增长”。以2020年为例,相对于现状情景S1,情景S4和S5分别削减了33.3%、80.3%的输移通量和26.2%、50.6%的径流系数。c)土地利用结构微调优化。模型模拟不同情景方案下的2020年临港新城各子集水区用地结构与既有规划的平均相对误差较小,总体差异不显著(p>0.05),仅需对既有规划土地利用结构进行微调,从而在土地利用规划上具有可操作性。(4)基于系统论,从“高-中-低”三个调控等级、“源-过程-汇”三个调控程序、核心调控内容等三方面构建平原河网地区城市集水区非点源污染调控体系。针对不同区域提出差异性的平原河网地区城市集水区非点源污染的核心调控策略：以上海全市为例提出基于雨水利用的大型城市非点源污染总体调控策略,以上海城区为例提出基于增加有效绿地的老城区非点源污染调控策略；以上海临港新城为例探讨基于径流零增长的新城镇非点源污染调控策略。更多还原

【Abstract】 As point pollution sources like industry pollution and domestic pollution have been under control, non-point sources pollution (NPS) has become the main factor on urban water quality, and received great attention from governments and researchers. In this context, process simulation and management and management of NPS have turned into a hotspot issue. Urbanization remarkably makes a great impact on the process of NPS in tidal plain with dense river network because of its special hydro-geomorphologic characteristics. Therefore, Lingang New City (LGNC) of Shanghai was taken as an example in this dissertation. From an intersection point of environment, ecology, hydrology and management science, combining theoretical research and empirical study, laboratory simulation and site monitoring, mathematical model and computer simulation, the dissertation takes the sub-watershed division as the breakthrough point, focuses on two main lines of non-point pollution, namely "source-process-sink" and "experiment-simulation-management". The dissertation illuminates the feature of NPS process, constructs a watershed-based management model and proposes methods, principles and key strategies for management of NPS in urban watershed. The main results are as follows,1. Based on typical water conservation zone management system in tidal plain with dense river network, the dissertation analyses characteristics of urban watershed in this area, and discusses the NPS-control-oriented zoning system from three dimensions of spatial scale, runoff path and controlling intensity. Zoning system is formed by four hierarchies:first-level watershed management is to balance relationship between study area and water conservation zones from large-scale and high-intensity by paying attention to main rivers:second-level watershed management is to regulate drainage system and function division from median scale and intensity by focusing on median rivers; third-levelwatershed management is to adjust the relation between small topographic area and watershed outlets from small-scale and low-intensity by concerning regional rivers; forth-level watershed management is for data support and mechanism exposition of NPS progress. As for LGNC, there are 1 first-level watershed,13 second-level sub-watersheds and 137 third-level sub-watersheds. Second-level watersheds are mainly in accord with the existing function zones, which may guarantee implement of management strategies.2. The dissertation discusses division method and process simplification method. On this basis, the dissertation represents process and features of NPS of urban watershed in tidal plain with dense river network using data by laboratory simulation and site monitoring.(1) The dissertation proposed the concept of Effective Pervious Areas (EPA) from aspect of NPS management compared to classical concept of Effective Impervious Area (EIA). EPA indicates the pervious area that receives and stores runoff from other areas. On this basis, the dissertation brings up the concept of Effective Green Area (EGA).(2) The dissertation analyses first flush effect of urban impervious area. Total level of NPS output of LGNC is lower than that of inner city of Shanghai. First flush effect of particulate pollutants is significant, but not dissolved pollutants. The concentration of particulate pollutants is much high than dissolved pollutants. It may because that LGNC is a newly-developed area. Land use type and rainfall intensity have great affect on first flush effect.(3) The dissertation illuminates the "source-sink" effect of urban pervious area. In theory, urban pervious area is the source as well as the sink of NPS. However, existing studies mainly focus on its sink effect.a) In terms of"source":surface runoff pollutant concentrations from existing green spaces in downtown are low, better than Grade V of national standard; however, surface runoff pollutant concentrations from newly-developed green spaces in suburban area are higher, inferior to national standard of Grade V, even reach up to 3.56 times of Shanghai Sewage Discharge Standard. Thus, pervious surface areas including urban green spaces can be regarded as potential pollution source, and pollution intensity is related to the location, completion time and soil origin.b) In terms of’sink’":urban green spaces are efficient for surface runoff pollution removal. Removal rate of COD(71.01±1.75%) is higher than TP(69.00±1.78%) and NH4+-N(64.39±1.84%). Removal effect of urban existing green spaces is more significant than suburban newly-developed green spaces (p<0.05). For different covers, effect of Ophiopogon japonicas is best, then Zoysia japonica, and exposed soil. Removal of pollutants happens mainly in soil horizon of 60cm to 80cm underground, and physical process is predominant. Moreover, pollution load, hydraulic load, detention time and timing of drainage have affect on pollution removal rate.c) Transformation of source and sink:In theory, urban pervious area acts as both source and sink. Non-effective pervious area functions as source, while effective pervious area functions as sink. Concretely, green spaces which have curbstones and cannot take in external runoff are sources, and green spaces which absorb and reduce pollution of external runoff are sinks. In summary, setting of curbstones is key point to change source to sink. By constructing curbstones at proper sites, disposing appropriate plant cover, soil type and soil depth, leading external runoff into green spaces in a proper velocity and volume, installing enough detention time can improve the reduction efficiency of urban pervious area.d) The dissertation takes soil salinity as an example to analyze leaching effect of precipitation to key element of urban pervious area. Soil salinity of urban pervious area has spatial and temporal variation in small scale, and the leaching effect differs from urban to suburban. In this study, salinity of existing urban green spaces is low. and salinity values of all 704 samples are under moderate salinity. In contrast, newly-developed green spaces are in highly salinization. Of all 240 samples, forty percent samples are saline soil, and plaque salinization is notable. Average variation coefficients of urban green spaces and suburban green spaces are 37.7%and 67.9%. Result of variance analysis shows that seasonal variance of four green-spaces is significant (p<0.05), the average value of topsoil salinity is spring>winter>fall>summer, and there is significant correlation between seasons of topsoil (p<0.01). Leaching effect of existing green spaces in center city is not significant, but situation of newly-devolped green spaces in suburban area is significant. Effluent of the former is plain water, while effluent of the latter is mixohaline. Rainfall intensity and precipitation are main factors.e) Three main non-point pollution sources are surface runoff (pollution from land use), rainfall and disperse rural domestic sewage. In recent decades, pollution load and transform flux are in significant increase. Change of pollution load and transform flux can be divided into three phases:the first phase (from 1965 to 1994) when load produced was not many and grew at low speed, the second phase (from 1995 to 2003) when problem of non-point pollution appeared, and the third phase (from 2004 to 2008) when the problem of non-point pollution emerged as a major issue. Rapid and large-scale development in LGNC results in change of land use structure and more pollution to be produced. On the other side, function of green space as sink should be promoted (average removal rate of 2.07%).3. The dissertation builds a management model combining systematic dynamics, multi-objective optimization, scenario planning, stakeholder survey and GIS, and takes LGNC as a case.(1) The dissertation points out that there are five aspects of uncertainty in management of NPS in urban watershed:Spatial and temporal uncertainty, source management uncertainty, process management uncertainty, sink management uncertainty, uncertainty of stakeholders and decision maker’s preference and uncertainty of models themselves.(2) For newly-developed zone, to manage NPS by changing land use structure is really difficult. In consequence, the dissertation proposes a thought of management by raising ecological function of urban pervious area as "sink" and using sewer net to receive first flush.(3) The dissertation investigates coupling mechanism of systematic dynamics, multi-objective optimization, scenario planning, stakeholder survey and GIS, and proproses the four steps coupling process:sub-watershed and sub-system analysis, sub-watershed development scenarios setting, spatial and temporal simulation and prediction of watershed non-point pollution, and optimal proposal determination.(4) The dissertation sets five typical scenarios and applies them into simulation of LGNC. The five scenarios are current situation SI, function promotion situation S2 and S3 (considering effective green space and effective wetland), and comprehensive reduction situation S4 and S5 (considering combination of effective green space, effective wetland and sewer construction). The main results are as follows,a) Through limited function promotion without sewer construction, management effect is not satisfied. For example, compared to current COD transform flux of 15564t/a and runoff coefficient of 0.42, in 2020, COD transform flux and runoff coefficient decrease by 7.5% and 7.8% by S2 and 1.3% and 2.6% by S3.b) Through function promotion and sewer construction, NPS transform flux is drastically cut, and runoff coefficient realizes "zero growth". Compared to SI, in 2020, COD transform flux and runoff coefficient decrease by 33.3% and 80.3% by S4 and 26.2% and 50.6% by S3.c) Relative error between land use structure of LGNC simulated in 2020 and current planning is tiny (p>0.05). Thus, to adjust land use structure planning fractionally is operable for NPS management.(5) The dissertation constructs a system for NPS management of urban watershed from’high-median-low" management levels, "source-process-sink" management procedures, and key strategies based on system theory. In accordance with different urban scales, the dissertation raises differentiated non-point pollution management strategies:take Shanghai as a case to show rainfall utilization is essential at at the city or regional level, take inner-city of Shanghai as a case to show effective green area is essential for the developed area, take LGNC of Shanghai as a case to show zero-growth of runoff is essential for the newly-developed area.更多还原