【作者】 王天巍；

【导师】 蔡崇法； 刘高焕；

【作者基本信息】 华中农业大学 ， 资源环境信息工程， 2008， 博士

【摘要】 道路是人类生活和社会发展中最重要的物质和能量通道。作为人工引入的廊道系统,道路在带来巨大的经济和社会效益的同时,其对生态环境造成的影响却长期以来被人们所忽视。我国对道路生态的系统化研究起步较晚,研究对象也主要集中于西部高原和山区,且主要以高速公路为研究对象。平原区尤其是滨海平原的报道较少,也缺乏不同道路类型之间生态影响的差异性对比研究。因此,有必要在不同生态区以不同道路类型为研究对象,深入地探讨道路对生态环境的影响。作为我国保存最完整的河口三角洲湿地,黄河三角洲以其资源的稀有性、生态的脆弱性而成为研究者长期关注的热点。随着该区域经济的发展,越来越多的工程建设已经造成了原生湿地生境的损失和环境的衰退,成为三角洲经济社会发展与湿地生态保护的矛盾焦点之一。本研究在野外调查的基础上,通过室内分析、模式识别、模型模拟、空间演算等多种手段,在微观尺度和景观尺度两个水平上对黄河三角洲道路系统的特征和演化规律、道路对边际植物生境、草本植物种群、土壤污染和景观水平上的格局影响进行了深入的研究,以期对黄河三角洲地区自然资源的合理利用和区域可持续发展提供参考依据。(1)基于野外各类道路的调查结果,提出了包含形态检测、模式识别和逻辑推理3个认知层次,适合于滨海平原地区的半自动道路网络提取模型。构建了黄河三角洲1984～2006年间的逐年道路网络数据库。在图论理论的支持下,完成了道路网络的形态特征和空间拓扑结构分析。时序分析结果表明,不同类型道路的扩张特点存在差异。省道和县乡公路扩张的主要驱动力是居民出行需求,油田专用公路的修建主要以满足胜利油田石油生产为主要目的,高速公路的修建主要是国家宏观产业布局和交通布局的需要。空间网络分析和重心分析的结果表明,三角洲道路网络连通状况可以划分为网络发散扩张、回路缓速递减和网络相互连通三个阶段。道路网络空间发展形态呈现出由一轴两核扩张→两岸四区对峙→多核网络互通的发展演化模式。(2)以野外土壤样品的采集与分析为基础,结合区域环境背景资料,构建了基于区域环境因素推演道路边际重金属含量的MLP神经网络模型,并应用该模型对黄河三角洲道路边际50m范围内的土壤重金属含量进行了2.5m分辨率的逐栅格模拟。结果表明,该模型能够较好的模拟道路边际土壤重金属浓度的空间分布,所得结果具有较高的精度。现代黄河三角洲道路边际的土壤重金属含量略高于山东省土壤背景值,但低于国家环境质量一级标准(除Cd外)。受Cd大面积超标的影响,黄河三角洲道路两侧边际50m范围内的土壤80.69%为轻度污染,15.49%为尚清洁,仅有5.83km~2为清洁。表明道路是其边际土壤的重金属污染源,但对环境尚未造成广泛和显著的危害。不同的道路,其边际土壤的重金属累积规律不同。公路等级越高、车流量越大,重金属浓度越高。随着距道路距离的变化,重金属浓度发生改变。总体上来说,大多数重金属在30m以内达到峰值,其后逐渐降低接近背景值。同时,变化规律受区域环境条件、土地利用方式和重金属累积机制的影响。(3)在研究区开展了多次野外植被样方调查,首次获得了黄河三角洲道路边际植物群落的主要类别。应用二岐式划分方法,完成了草本植物群落的划分,并通过多样性分析和DCCA排序,研究了道路边际植物群落的梯度演变规律和主要环境驱动因子。研究结果显示,黄河三角洲道路边际草本植物种类计有20科、41属、48种,分别占现代黄河三角洲被子植物的51.3%,34.5%,25.5%,道路边际植被共涉及14个分布区类型,道路边际植被具有广域性,且以温带成分为主,热带性质的属起到补充成分的作用。道路15 m边际内的植被可分为4个群系,11个群丛。碱蓬、芦苇、獐茅、狗尾草是主要建群种。群落的形成和分布同时受到大尺度海陆环境演变和小尺度道路边际环境的双重影响。物种多样性从道路路肩向远处存在明显的递减梯度。物种更替速率形成以道路为中心的单峰分布格局。距道路距离,距海岸线距离和道路类型是决定道路边际植物种群分布的主要因素。省道边际物种更替速率最快,高速公路次之,县乡公路最低。道路作为典型的人工廊道,对现代黄河三角洲的湿生植物产生了较大的干扰,尤其是其两侧10m的带状区域,出现较为明显的湿生植物衰退和中生、旱生物种的入侵,是造成黄河口湿地萎缩的原因之一。(4)研究在野外调查的基础上,建立了包含1个景观类,3个景观亚类,12个景观系的道路边际景观分类体系。并应用多时相遥感影像,完成了主要道路两侧边际1km范围内的景观分类。应用多种景观格局指数,进行了不同道路类型,不同缓冲区范围和不同景观类别的动态演变特征分析。结果表明,道路边际景观的演变规律受道路发展过程、道路功能差异和区域发展背景的影响。距道路越近,受人为干扰影响越严重。人类对道路边际景观的开发利用降低了景观的多样性,增加了景观的破碎化,造成了自然湿地景观的减少,使景观边界形状趋于规则化发展。提出了基于波尔兹曼热力学的道路边际景观稳定性评价方法,结果表明人类对道路边际持续的负熵输入是决定非自然景观稳定性的主要原因,负熵输入的强度主要受道路功能和分布区域的影响。结合前述微观生态过程研究的结果,可以认为,道路对其边际植被、土壤的微观生态影响不足以导致景观尺度上的格局改变,水盐生态过程、流通过程、破碎化过程、以及接近效应和效益梯度场的存在是影响道路边际景观分布格局和演变的主要原因。更多还原

【Abstract】 Yellow River delta (YRD) is one of the biggest deltas that there is a large area of wetland in the world. Thanks to soil (sands) sediment carried by the Yellow River, there was averagely the newly formed land 21.3 km~2 in YRD. During the development of petroleum industry and urban expansion, wetlands were degraded due to population growth, irrational land use and so on. Roads, appeared as major artificial corridors, have been proved to result in habitat loss and environmental decay. However, the important ecological effects were ignored. So, in this paper, based on fields survey, the characteristics and Evolution rules of the road system in the Yellow River delta, and also the ecological effect, such as plant’s habitat and Species nearly the road, soil heavy metal pollution and landscape patterns were study on micro and macro scale by various methods, and the results as flowed(1) Based on analysis results of the fields survey, the semi-automatic road extraction model adapt to Coastal Plain , including morphology detection, pattern recognition and Logical Reasoning was put forward, and the yearly road network database in the YRD from 1984 to 2006 were built by the model. The results from morphological characteristics and spatial topological analysis of the road network showed that, the expansion characteristics were different for road types. The driving force of the expansion for provincial roads and county roads was resident trip demands, while to meet the need of petroleum industry development was the most important reason for expansion oilfleld highways, and the construction of highway mainly affected by the national industrial distribution and traffic distribution. And for the result of network and barycenter analyses, the processes of network development can be divided into three stages: network divergent expanding, loops of network decreasing slowly, and network mutual connecting. The spatial form of road network development mode was from ’expanding around one axis-two cores’, ’four cores confronting in two banks’, to ’network connecting with multicores’.(2) Based on the analysis results of soils near the road, and combining with regional background information, a MLP artificial neural network was constructed, which could deduce heavy metals content of soil in road verges from regional background information. Heavy metals content of Soil within 50m road verges in the YRD were simulated grid by grid with 2.5m resolution, and the results showed that, this model provided a better simulation of the spatial distribution heavy metal content of soil, and the simulation results had high accuracy. The heavy metal content of soil in road verges were higher than the background values of those in Shandong province slightly, but not exceed national environmental quality standard for soils except the slight pollution of Cd. Affected by the Cd pollution, for the soil in 50m of road verges in the YRD, there are 80.69% soil are light pollution and 15.49% are almost non-pollution, and the left are non-pollution. The results indicated that though Road in study area played an important role in heavy metal accumulation, but did not bring broad and notable environment contamination yet. For different types of road, the concentrations of the tested metals were in order of Express Way (R_g)> Provincial Road (R_s)> Rural Road (R_x), except Pb. The concentrations of heavy metals besides road verge reach to peak value in 30m, and the distribution was impacted mainly by the factors such as wind direction, soil properties and vegetation cover. But the different change pattern was observed in individual metal elements because of their different diffusion mechanism.(3) The herbaceous communities beside road verges were firstly obtained based on field investigations and Twinspan classifications. The gradient change law of these plant communities and mainly environmental driving factors then were studied by using diversity analysis and DCCA ordination methods. The results showed, herbaceous plant of road verges in modern YRD are belonged to 20 families, 41 genera and 48 species, which account for 51.3%, 34.5% and 25.5% of angiosperm in YRD separately. Results of areal types analysis shows that, the plant of road verges are eurychoric, and the temperate type was dominant, while the tropical type was supplementary. Vegetation in road verges can be divided into 4 formations and 11 associations, and the constructive species in most adjacent area of road verges were mesophytes and xerophytes while hygrophyte distribute a little farer. Mainly constructive species are Suaeda glauca, Phragmites australis, Aeluropus sinensis and Setaria viridis. The plant community distribution patterns in road verges are affected by salt gradient in land-sea scale and by water gradient, nutrient gradient and microclimate in local scale. The species turnover rate in the community was exhibited a unimodal pattern and distributed symmetrically with the road. And turnover rate of provincial road was significantly greater than that of rural road. Distance from road verges, distance from seashore and road type are the mainly determinant factors. As a typical artificial corridor, roads disturbance the habitat of hygrophyte. Especially in strip within 10m of road verges, mesophytes and xerophytes invasion are observed. It is one of the most important factors which conducted wetland degraded.(4) Classification system of landscape in road margins was constructed based on field investigation, which include 1 landscape class, 3 subclass and 12 series. Multi temporal remote sensing image then used to identify landscape with 1km of mainly road verges. And dynamic evolution characteristics of landscape with different road types, different buffers and different units were analyzed by several landscape pattern indices. The result showed, dynamic evolution laws of landscape in road verges is impacted by development process of road network, function of roads, and regional development policies. The nearer of road verges was, the more serious of disturbance by human being was. The development and utilization by human decreased landscape diversity, increased landscape fragmentation, caused wetland vanishing, and regularized the boundary of landscape. Different landscape appeared different spatial pattern. Tidal flat、grass and shrub、reed field and salt field were largest in area, dry land and settlement area were largest in patch number, meantime tidal flat and reed field were highest in aggregation degree. Fractal dimension of natural wet landscape was higher than that of agricultural and artificial landscape. Combining with the study results of micro ecology process, so it can be considered that the micro ecology effect of road is not strong enough to changes the landscape pattern at landscape scale, and the main reason caused landscape changes were the persistent negative entropy input to road margin by human. hydro-ecological processes, geographical processes, biological processes and urbanization process was the dominate ecological processes impact the dynamic evolution of road verges.更多还原