【作者】 徐敬敬；

【导师】 申广荣；

【作者基本信息】 上海交通大学 ， 园艺学， 2010， 硕士

【摘要】 随着人类社会的进步和经济的发展,生态环境及其安全问题愈来愈受到重视,生态安全的预测研究以及在预测研究基础上建立生态安全预警决策成为了人们关注的焦点。位于西太平洋沿岸中国海岸线中部、地处中国最大河流长江入海口的崇明岛,被称为“上海最后一块真正的生态净土”。依据最新规划和崇明的发展定位,崇明岛将“留足自然生态涵养空间”和“留足未来国际级项目的规划选址空间”,真正成为“21世纪生态岛”。因此,与生态安全问题紧密相关的土壤环境污染监测、安全监控以及肥力综合评价就成为了崇明生态安全预测预警研究的重要组成部分。土壤是一形态和演化过程都十分复杂的自然综合体,开展土壤养分空间变异性研究对于科学合理地制定农田施肥方案,提高养分资源利用率,减少由于过量施肥造成环境污染,实现精确施肥具有重要意义。本文根据崇明岛的土壤发育状况、土地利用特征,借助地统计空间变异分析和GIS技术在进行详细的土壤养分变异特征分析的基础上,探讨分析了土壤污染监测的采样点优化布局布设方案;并依据研究区成土过程的化学分类区域和近年的土地利用情况,采用隶属度函数和内梅罗指数结合的方法对崇明土壤进行了综合肥力评价,期望在分析了解崇明土壤养分空间分布情况和土壤肥力总体水平的基础上,为科学、合理施肥,保护环境及崇明生态建设提供依据。研究采用GIS与地统计学相结合的方法,首先,分析崇明土壤重金属的空间变异特征、结合反距离加权插值和克里金插值,对崇明土壤进行空间变异分析,分别获得其预插值污染范围和估计的重金属分布特征。通过研究发现,Zn、Hg无变异角度,其变程分别为2000m、2500m;As、Cu、Pb变异角度在50度左右,变程分别是55000m、35000m、5000m;而Cr、Cd、Ni变异角度在300度左右,变程是7500m、40000m、16000m。变程和变异角度将作为最基本的参数,在满足精度的条件下对崇明土壤做地形分析并针对不同区域进行采样布局设计,在初步采样的基础上,获得土壤二次采样的最优化布局,为进一步更准确地获取研究区土壤养分的空间变异特征奠定基础。其次,本文研究了上海崇明岛表层土壤有效铜、有效硼、有效铁、有效钼、有效锰、有效锌等微量元素的空间变异特征及分布规律。针对每个指标分别进行详细的试验分析,取得空间相关性最强的步长和分析尺度,并由此选择不同的拟合模型对各因子进行泛克里格插值.从而保证了进行筛选试验时理论插值模型的稳定性及其精度的可比性,同时获得相关微量元素的空间变异分布图。结果表明:6个指标均表现出强变异性。再次,本文基于对崇明土壤18项化学指标的精确的空间变异特征的研究,借助遥感图像处理软件ERDAS中对2002年的崇明ETM图像进行监督分类;利用主成分分析和因子转轴选择出6个主成分因子,在ERDAS中将6个主成分因子的栅格图进行波段融合获得组合后的三波段图,在ArcGIS中对融合后的图像进行重分类获得:钙质滨海盐渍土带,围垦冲击土带,高铁有机人为土带,半水成半人为土带,有机新成土带,有机人为土带,淤积新成土带等七个化学分类土带。导入选择成土过程中土壤发育的限制因子,以限制因子为评价对象,结合隶属度函数和内梅罗指数法对崇明土壤进行综合肥力评价,将土壤肥力分为五个等级,每个区都分别进行评价。根据以上研究,本文得到以下结论:(一)土壤具有高度的空间变异性,不同尺度下土壤养分及重金属均适宜采用地统计学模型进行空间分析与估测。结合GIS的地统计学技术作为研究土壤空间变异的重要工具,不同变量在不同取样尺度下插值分析空间估测结果是不一致的。本文基于原有S型采样分析数据,详细研究分析了土壤的空间变异特性,特别是土壤不同元素的变异角度和变程,并根据地形及土地利用情况进行了相应的坡度分析,在此基础上分析设计了二次采样方案。研究表明,依据二次采样分析数据的土壤空间变异特征更能反映研究区的土壤空间变异状况。因此,基于初步采样数据进行相应的空间分析再进行二次采样方案的设计,为最终选取适合的尺度和插值方法进行相应的区域土壤空间变异的分析是必要的,也是可行的。(二)土壤养分空间变异研究中,最核心的问题是尺度和插值模型及其参数的选择。在ArcGIS中自动计算选取的插值模型和参数不一定能最好地反应土壤数据的空间变异分布特征。本文通过不同的步长尝试,依据块金效应的原则来选择合适的模型参数。在模型的选择上考虑了多尺度的模型,引入双模型插值。结果表明,双尺度的模型提高了插值的精度,能更加准确的反应出土壤养分的空间变异特征。(三)崇明岛是一个冲击而成的岛,其发育过程相当复杂,每年都会发生不同程度的变化。依据成土过程进行分区,并结合土地利用类型进行肥力分区综合评价是一种新的思路,对了解崇明土壤肥力状况有十分重要的意义。通过研究每一类土壤发育限制因子,及其对崇明18项土壤化学指标的分析进行分类和分区,结合各评价区域的土地利用类型,进行不同区域的土壤肥力综合评价,更能客观地反映研究区土壤的肥力状况。与传统的肥力评价方法相比,本方法需要更多的数据及相应的处理技术支撑,但在现有的条件下,对本研究区域来说,这种方法是可行的且效果良好。更多还原

【Abstract】 With the development of economic, ecological environment and security issues are paid more attentions, as well as the prediction of ecological security and the construction of ecological security-warning on the base of predict research. Located in the west-pacific coastline and the estuary of China’s largest river, Chongming Island considered as the last ecological land in Shanghai, and it plans to become a real ecological island in 21 century for leaving space to the ecological protection and sampling sites selection to the world level project, according to the latest planning and development orientation. So soil pollution monitor, surveillance and comprehensive evaluation of fertility which are closely related to the ecological security are become one of the most important part in ecological security-warning program. Soil is a nature complex that has complex morphology and evolution, carried out spatial variability research of soil nutrition has great significance to the farmland fertilization, nutrition efficiency and reduction of environmental pollution which induced by over fertilization.Soil was a natural synthesis with complicated configuration and evolvement process. The study on soil spatial variability of soil nutrients is very important to promote the development of soil science. It is not improving the resource utilization ratio of soil nutrients, but also reducing the environmental pollution.Based on the soil land use and development conditions,on the one hand ,the study discussed how to optimize the distribution of sampling sites by using geostatistics and GIS technology. On the other hand, according to the process of soil formation and land use, the research also analysis fertility evaluation with method of combining Nemero index. On the basis of spatial distribution of soil nutrient, we hope that we can provide evidence for rational fertilization, environmental protection and ecological construction in Chongming.In the research, first of all, by analyzing spatial variability of soil heavy metals, spatial interpolations were made by universal kriging and inverse distance weighted methods. The results are as following:The range of Zn is 2000m which has no variation of the angle; the anisotropy of Hg is not significant, its range is 2500m; the anisotropy of As, Cu and Pb are the same 50°, their range are 55000m, 35000m and 5000m; the Cr, Cd, Ni variation of angle of 300 degrees, variable range is 7500m, 40000m, 16000m. As variable-range and anisotropy are the most important parameters, it is necessary to develop the second sample plan in order to obtain a more accurate spatial variability of soil nutrients. Secondly, this paper studied the spatial variability and distribution of available Cu, Mn, Zn, B, Mo, Fe in the soil surface of Chongming island. According to the analysis of soil trace elements, the best lag size and scale were chosen. The results showed that: six indicators all shown strong variability. Thirdly, based on the Spatial Variability research on the 18 chemical indexes, this paper performed Supervised Classification to the ETM image produced in 2002 by ERDAS. 6 Principal component factors were selected by using PCA and factor axis; 3-band graphic was obtained by composing bands; and 7 classes were eventually reclassified via ArcGIS. After introducing the pedogenic development factors which were selected as evaluated objects to conduct comprehensive evaluation of the soil by combining Membership function and Nemero Index, five levels of soil maturity were discerned by the evaluation in each zone separately.Based on the above, this paper obtained the following conclusions:1. The feature of soil behaves as an obvious spatial heterogeneity and different observational scales of soil nutrient and heavy metal elements’distributions are suitable to be analyzed and predicted by geostatistical models. As a magnificent method in analyzing the spatial variation of soil, difference among the interpolations of different elements that being estimated at different sampling scales come up under the help of geostatistics with GIS. In this article, according to the sampling data of traditional“S”-sampling method, detailed analysis on the spatial distribution was performed, especially on the anisotropy and range of each element, with the gradient of terrain and land use factors, to present a reformative schema of sampling. The result shows a discernable advantage of the second sampling schema that the variations are more precise to reflect the status of interesting area. Therefore, the method that to draw up a schema of second sampling referring to the results observed from the first sampling sites is feasible and indispensable.2. In the research on soil nutrients, the crucial issue is to choose an appropriate model, scale and relevant parameters for interpolation. The model and parameter suggested automatically by ArcGIS software itself could hardly provide a satisfactory portray of the features of soil data. By iterate the lag sizes of different scales, the effective parameters and model could be found out with the minimum value of nugget effect. Introduced multi-scaled models to interpolation in this research, And these models are proved consequently to improve not only the accuracy of interpolation but also the similarity of the variation features of soil nutrients.3. Chongming island is formed by rush, and its development process is quite complicated. The changes at different degrees take place every year. It is important to understand Chongming’s soil fertility. With the investigation to the limiting factors of soil development, classification and division to 18 chemical index of soil, as well as did comprehensive evaluation to the soil fertility; soil fertility of research area is objectively reflected. This approach needs more data and related technology by comparing with traditional method, but under the existing condition, this approach is feasible and has good results to this research area.更多还原