【作者】 刘勇；

【导师】 尹京苑；

【作者基本信息】 上海大学 ， 通信与信息系统， 2009， 博士

【摘要】 由于城市的扩张、工业的发展,城市周边土地承受着越来越大的压力。土地类型、结构不断转化,农用地大幅度减少;同时,工矿企业生产排放的化学品和农药化肥的施用造成土壤污染,土地质量下降;人口密度增大,动植物栖息地丧失,生物多样性减少等等都造成区域土地生态系统的风险,已引起了研究者、管理者和公众的广泛关注。然而如何在实践中辨识土地生态风险的压力因子?如何进行有效的生态风险评价并用之于土地管理?还没有成熟的方法,目前我国对于土壤风险评价大多数是针对土壤中重金属和农药污染,还没有形成完善的方法和指标体系来表征或评价包含物理压力、化学压力和生物压力在内的土地生态风险。本文在研究土地生态风险理论的基础上,将土地生态风险压力因子分为物理因子、化学因子、生物因子,选择能源重化工基地山西省太原市为研究区域,利用空间信息技术、建模技术,结合土壤采样分析,研究了太原市土地生态风险的物理压力和化学压力以及压力因子的时空变化和分布特征,并进行综合评价。本文对土地生态风险评价的理论和实践进行了有益的探索,为土地生态风险评价方法的研究提供了新的思路和可借鉴的经验。本文的主要研究内容及结果如下:1、以土地生态风险分析为出发点,对重工业城市周边土地生态风险的压力因子、暴露水平、影响因素及其作用机理进行了研究,构建了多压力因子的区域土地生态风险评价概念模型,并主要对物理压力和化学压力的影响进行评价。2、以太原市1990和2004年LandsatTM遥感影像为数据源,在RS和GIS技术支持下,分析太原市土地利用格局的时空动态变化和景观格局变化,选取土地利用指数和景观指数分析了土地生态系统的风险。结果表明:1)在城市化的作用下,原城镇周边的土地类型如耕地、草地、林地等被大量占用,城镇用地不断增加并集中。2)区域土地利用类型发生了复杂的相互转换,尤以城市周边的林草交错带和农草交错带变动剧烈。3)区域景观的多样性增加,优势度减小,土地利用趋于破碎化,土地生态系统的风险程度增大。3、对太原市周边土壤进行布点采样,分析测定土壤中的重金属元素含量和土壤理化性质,将GIS技术与地统计学方法相结合,探讨了该地区土壤重金属元素含量的影响因素及其空间分布特征。结果表明:重金属元素的空间相关性较弱,N、P、K的空间相关性中等;不同的重金属元素和养分元素的分布格局不同,这是由于结构性因素(如地形、母质等)和随机因素(如工农业生产、交通等)的作用引起的。城市工农业的区划与土壤元素的空间分布有一定的影响。4、应用土壤污染单因子指数法、综合指数法、Hankanson生态风险指数法等多种方法对研究区土壤重金属污染的生态风险进行综合评价,结果表明:八种重金属元素中,Hg、Cd的风险程度较大,其他重金属元素的风险较低;Hankanson综合风险指数评价表明土壤重金属的风险基本在“中”到“较高”,个别点的风险程度高;研究地区的北部和中部风险程度较高,这两个地方是太原市工业比较集中的地区,人口密集,交通发达,工业污染、农业污染和交通污染共同导致了土壤重金属的高风险。5、选择能表征土地质量的性状、功能和立地条件的综合指标,包括土壤区位条件、立地植被类型、土壤pH、理化性质(N、P、K)和重金属元素(Pb、Cu、Ni、As、Cd、Cr、Hg、Zn),运用支持向量机(Support Vector Machine,SVM)方法对各项指标进行集成,形成综合表征土地质量的模型,对研究区土地质量的生态风险进行评价。结果表明:SVM评价法是基于多个土地生态系统指标的综合评价,可以弥补其他评价方法的不足,用于土地质量综合评价有着广阔的前景。6、在土地生态风险分析的基础上,构造多元数组结构模型进行太原市土地生态风险综合评价。本文的创新之处在于:从城市周边土地由于土地利用变化、土壤污染以及土地承载力造成的生态风险分析出发,构建了土地生态风险评价的概念框架,从多尺度、多压力的角度评价土地生态风险。将空间信息技术、地统计分析技术、景观生态学理论综合运用于生态风险评价研究中,对太原市土地利用覆被变化的结构风险和土壤重金属污染的生态风险进行了研究,丰富了土地利用覆被变化和土壤污染评价的研究成果。基于支持向量机SVM理论,构建一种新的土地质量综合评价方法,克服了传统的单因子评价、综合指数评价和生态风险指数评价的不足,为土地质量综合评价提供了新的研究方法。如何对多尺度、多压力的生态风险进行综合是一个难题,本文首次构建了多元数组结构模型,为多压力因子的土地生态风险综合评价在方法上提供了新思路,并在实践中进行了有益的探索。更多还原

【Abstract】 In the footsteps of the industrialization and urbanization, agricultural societies have been transformed into urbanized landscapes; land quality has decreased by the advancing soil degradation processes caused by industrial chemicals and agrochemicals; and land ecosystem has been threatened by loss of biodiversity and habitats. Land ecological risk assessment has been gain more and more concern by researchers, managers and publics. To achieve a sustainable development, impacts on land ecosystem of land use changes, chemicals and other stresses must be considered in land ecological risk assessment on regional scales. Many studies on soil risk assessment have been conducted about chemicals such as heavy metals and pesticides but there is as yet not a well-defined universal methodology to characterize land risk and to define a set of clear indicators includes physical, chemical and biological factors.This paper defines land ecological risk stress as physical factors which are land use change and loss of habitats, chemical factors which are heavy metals and soil degradation, and biological factors which are loss of biodiversity etc. This paper, selected Taiyuan city, Shanxi Province as a case study area, focus on a joint analysis of the two definitions of soil quality and land use/cover change, analyzes the distribution characters and spatio-temporal of these factors, leading to comprehensive assessment of land ecological risk that have relevance for sustainable land management.The main contents and results are as follows:Ecological risk assessment is a process that evaluates the likelihood that adverse ecological effects are occurring as a result of exposure to one or more stressors. Based on identification of stressors and reorganization of adverse effects of potential ecological concern in heavy industry city, this chapter developed conceptual model for integrating multi-stressors exposure to land ecosystem in an urbanizing environment on regional level. . Land-used/Cover Change in Taiyuan City is analyzed based on remote sensing images in 1990, 2004. GIS technology, land used dynamic degree model and landscape theory are used to research the spatio-temporal trend and land ecological risk of LUCC. From the study, the following conclusions are drawn: 1)the area of Cultivated land, woodland and grassland were decreased with the time, while, the area of constructed land was increased. 2) The land-use types of this area had changed frequently in the past 14 years, especially those interlaced areas with fragile ecological environment between cultivated land, woodland and grass land. 3)the landscape diversity and fragmentation had increased while the landscape dominance had reduced. The land ecosystem structure risk had increased.Soil samples were collected from the research area and analyzed to determine heavy metals concentrations and nutrients concentration. The spatial distribution of Hg and Cr in soils in this area was explored using the technologies of Geographic Information System and Geostatistic. Results of spatial distribution analyses showed that 1) the spatial self-correlation of heavy metals in soils in this area were low dependent, while nutrients were moderately interdependent. 2) the spatial interpolation map revealed different elements had a different spatial distribution, which are affected by both structural (including topography factor, soil type, soil native matter and soil texture factors) and random (including fertilizer, crop, management and human actions) factors; 3) Various spatial distributions suggested that urban planning and soil pollution control should focus on soil spatial character and local condition.Single Factor index, comprehensive index and Hakanson potential ecological index were used to assess the ecological hazards of heavy metals in urban soils of research field. The results showed that Hg and Cd had a higher risk level and the other heavy metals had a lower risk. Hankanson potential risk assessment showed that most of soil samples in this area had moderately risk, while few samples had extremely risk level. Risk of heavy metals in the north part and central part of research area were higher than that of other parts. Industrial pollution, agricultural pollution and traffic pollution lead to the high risk of these soils.An comprehensive assessment model of land quality was established based on Support Vector Machine (SVM), The model is experimented with indicators which include soil heavy metals data, soil nutrients data, soil pH, sampling locations and land cover type data of research area, The results demonstrate that Support Vector Machine can resolve problems of classification with small sample more effectively and accurately. SVM method can reach an ideal comprehensive evaluation results accurately and objectively for land quality assessment.To integrate the land physical risk and chemical risk, multi-variable array model was developed to put landuse/cover structure risk and soil hankanson risk into together.更多还原