广州市农业表层土壤重金属污染及其与景观格局的多尺度关系

发布时间：2018-01-01 09:32

本文关键词：广州市农业表层土壤重金属污染及其与景观格局的多尺度关系　出处：《广州大学》2015年硕士论文　论文类型：学位论文

【摘要】：随着快速城市化和工业化,我国农业土壤污染尤其重金属污染问题日益突出。人类活动干扰下的土壤污染极其复杂,具有点源和非点源污染特征。重金属迁移扩散并累积到土壤的过程往往受到从样点到区域一系列尺度上景观要素时空配置的影响,将两者联系起来有助于理解土壤污染机制,但目前研究多局限于单一尺度。鉴于土壤和景观特征的空间异质性,亟待从多尺度角度探讨土壤重金属污染分布及其与景观格局的关系。本文以广州市农业表层土壤为研究对象,基于2005年641个重金属(Cd、Pb、As、Cr、Ni)样点数据、土地利用/覆盖图,采用多尺度嵌套模型、Pearson/Spearman相关分析、地理加权回归和偏冗余分析方法,揭示土壤重金属污染特征及其多尺度空间结构,并在此基础上叠加不同景观格局因子,解析不同尺度上景观格局与土壤重金属污染的关系,识别特定尺度起决定作用的景观格局因子。主要结论如下:(1)广州市农业土壤多处于安全水平,只有5-18%的土壤受到Ni、Cds和A污染,并且以轻度污染为主。不同农业用地类型上重金属污染水平多按照农药化肥施用量递减:果园/菜地水田/旱地林地,但是多数情况下不同地类间的差异并不显著,可能与本研究区土壤重金属元素主要受到成土母质的影响有关。(2)农业土壤重金属的空间变异具有尺度依赖性,多尺度嵌套模型较单尺度的普通克里格插值,更能充分地挖掘和利用土壤Cd和Pb的空间变异结构,插值精度明显优于粗粒度和中粒度尺度,但与细粒度尺度相当。多尺度嵌套模型插值结果显示,广州市农业土壤Cd和Pb含量普遍高于广东省土壤背景值,但多数未超过中国农业土壤最大允许标准,Cd污染区主要位于中心城区,少量位于从化区、南沙区以及花都区与白云区交界,可能与交通活动释放、污水灌溉和过量施肥有关。(3)Pearson/Spearman相关分析、地理加权回归(GWR)和偏冗余分析显示,土壤重金属污染与景观格局的关系因重金属元素和空间尺度而异。调整的决定系数、校正的赤池信息准则(AICc)和模型残差的自相关性检验(Moran’s I)显示,GWR模型相比OLS模型能更好地揭示土壤重金属污染与景观格局的关系,且最佳分析尺度为5km×5km。GWR模型和相关分析结果显示,3个尺度上土壤重金属污染与景观格局的关系存在空间变异性,两者的关系可正可负,多数土壤重金属污染水平与成土母质组成(沉积物或花岗岩的比例)和土地利用格局(土地利用类型的比例,尤其是水体的比例)呈显著相关,显著相关区域主要位于白云区、番禺区和南沙区,表明成土母质、污染物随水体迁移或者污水灌溉可能是土壤重金属输入的主要途径。而偏冗余分析表明,景观格局可解释土壤重金属污染变异的12.8-29.1%,解释力度随尺度递增:10km×10km5km×5km2km×2km。不同尺度上,三组景观格局因子(成土母质,距离-密度变量,土地利用)对土壤重金属污染变异的贡献不尽相同,50%以上可解释的土壤污染变异是由两组或两组以上的景观格局因子共同决定。总体上,成土母质和距离-密度变量对土壤污染的影响随研究尺度增加,而土地利用的影响则随研究尺度递减。
[Abstract]:With the rapid city and industrialization of agricultural soil pollution in China especially heavy metal pollution problems have become increasingly prominent. Soil pollution under the interference of human activities is extremely complex, with point source and non-point source pollution characteristics of heavy metal migration diffusion and accumulated into the soil samples from the process is often subject to regional scale on a series of temporal and spatial distribution of landscape elements effect of linking the two help to understand the mechanism of soil pollution, but the current research is limited to a single scale. In view of the spatial heterogeneity of soil and landscape features, to investigate the soil heavy metals pollution distribution and its relationship with the landscape pattern from the multi-scale perspective. Based on the agricultural soil surface in Guangzhou city as the research object, in 2005 641 based on the heavy metals (Cd, Pb, As, Cr, Ni) sample data, land use / land cover map, using multi-scale nested model, Pearson/Spearman correlation analysis, geographical weighted back Return and partial redundancy analysis method, and reveals the multi scale spatial structure characteristics of heavy metal pollution in soil, and the factors of different landscape pattern based on the superposition relation analysis on different scales of landscape pattern and soil heavy metal pollution, identify the specific scale determining factor of landscape pattern effect. The main conclusions are as follows: (1) Guangzhou City Agricultural Soil in the security level, only 5-18% of the soil by Ni, Cds and A pollution, and light pollution. Different types of agricultural land on the heavy metal pollution level in accordance with the amount of chemical fertilizer pesticide decline: Orchard / vegetable / paddy upland forest, but not significantly difference between different land types in most cases, may be related to soil heavy metals the elements in the study area are mainly affected by soil parent materials. (2) the spatial variability of heavy metals in agricultural soils with scale dependent, multi-scale nested model with single scale The ordinary Kriging interpolation, to fully excavate and utilize the soil Cd and Pb spatial structures, the interpolation accuracy is better than the coarse grain size and scale, but the scale is quite fine-grained and multi-scale nested model interpolation. The results showed that Guangzhou agricultural soil Cd and Pb content is higher than the general background values of soils in Guangdong Province, but most did not exceed the maximum allowable standard of agricultural soil Chinese, Cd pollution areas are mainly located in the city center is located in Conghua District, a small, at the junction of Nansha District and Huadu District and Baiyun District, may release and transport activities, sewage irrigation and excessive fertilizer. (3) Pearson/Spearman correlation analysis, geographical weighted regression (GWR) and partial redundancy analysis show that the relationship between soil heavy metal pollution and landscape pattern due to heavy metals and different spatial scales. The adjusted coefficient of determination, Akaike information criterion (AICc) self correlation correction and model residuals (Moran s I) test showed that the GWR model compared with the OLS model can better reveal the relationship between soil heavy metal pollution and landscape pattern analysis, and the best scale of 5km * 5km.GWR model and correlation analysis showed that the 3 scale relationship between soil heavy metal pollution and landscape pattern in the spatial variability of the relationship between the two can be positive or negative, most of the soil heavy metal pollution level and soil composition (sediment or granite ratio) and land use pattern (land use type proportion, especially water ratio) were significantly correlated, significant correlation area mainly located in Baiyun District, Panyu District and Nansha District, shows that soil parent material, with water pollutant migration or sewage irrigation may be the main way of soil heavy metal input. And partial redundancy analysis showed that the landscape pattern can explain the variation of 12.8-29.1% of soil heavy metal pollution, interpretation efforts with scale of delivery Add: 10km * 10km5km * 5km2km * 2km. three groups of different scale, landscape pattern factor (soil, distance and density variables, land use) of soil heavy metal pollution variation contribution is not the same, the soil pollution variation of more than 50% can be explained is decided jointly by two or more than two groups of landscape factors overall, soil density and distance variables on the soil pollution research with the scale increasing, and the influence of land use along with the scale of decline.

【学位授予单位】：广州大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：X53;P901

【相似文献】