【作者】 何东进；

【导师】 洪伟； 胡海清；

【作者基本信息】 东北林业大学 ， 生态学， 2004， 博士

【摘要】 首次运用景观生态学理论与方法对世界文化与自然遗产——武夷山风景名胜区的景观空间格局、动态变化及其空间关系等方面进行研究；对景观生态学中的尺度、等级和干扰进行了较为深入的探讨；将分室理论与人工神经网络分别用于模拟人类干扰对景观格局的影响机制和建立景观多样性与景观类型数之间的关系；依土壤、植物物种多样性等环境因子分析了武夷山风景名胜区森林景观的异质性；提出了包含协调性、奇特性、社会性、梯度性等指标在内的景观生态评价指标体系，并对武夷山风景名胜区的生态环境进行了科学的评价。 1、在武夷山风景名胜区中，马尾松林、茶园与农田是3个主要景观类型，3者面积与周长之和分别占全区面积与周长的84.581％和81.669％，其中，马尾松林是该区的基质景观，在控制景观整体结构、功能和动态过程中起主导作用。 2、景观要素斑块特征分析表明，景观要素面积、周长和斑块数的分布极不平衡，其中各景观要素面积与周长的分布极为相似，而斑块数的分布却没有明显的规律性。斑块形状指数、景观破碎度指数、景观分离度指数等结果均显示河流、马尾松林景观斑块形状最为复杂、破碎度最小、分离程度最低，而居住地景观的斑块形状最简单、破碎度最大、分散程度最高，从而表明了人类活动是造成景观破碎化、分散化的重要原因 3、运用正态分布、对数正态分布、Weibull分布、Γ-分布、Bata-分布及负指数分布，分别在3种等级划分条件下对各景观类型斑块大小分布规律进行研究。结果表明多数斑块大小服从对数正态分布，少数服从Bata-分布和Weibull分布，而所有类型的斑块大小均不服从正态分布。斑块大小分布存在着等级效应，就整个风景区所有斑块大小分布来说，尚未有一种概率分布能理想地加以拟合。由具有相同分布规律的不同景观类型的斑块构成的整体，其斑块大小分布规律将基本保持不变，而由不同分布规律的不同景观类型的斑块构成的整体，其原有的斑块大小分布规律将发生变化或消失。 4、设计4种景观取样尺度，运用景观类型空间关联度指数，探讨武夷山风景区景观类型空间关系及其尺度效应。结果显示景观类型空间关系存在着尺度效应，尺度变化对破碎化程度高的景观类型空间关系的影响大，对破碎化程度低的景观类型空间关系影响小。进行景观类型空间关系研究时，取样尺度不宜太大。 5、景观空间格局与空间关系揭示了武夷山风景区景观空间格局为：在低海拔区呈现出以居民点为中心，通过农田、河流等向周围扩散的干扰景观格局；在中高海拔区则呈现出以马尾松近成熟、成过熟林占优势的天然马尾松林为主体、受人为干扰较小的天然景观格局。 6、景观空间格局动态分析表明，武夷山风景区在1996～2001间受到良好的保护，景观格局总体上未出现显著的变化。从局部上看，5a间武夷山风景区各景观类型除河流景观外均发生了一定的变化，其中杉木林、农田与马尾松林景观面积减少量居前3位，分别减少了0.5644、0.2833、0.2804(kmZ)，而茶园与居住地面积却分别增加了0.7285、0.0814(klnZ)。因此，毁林种茶、弃农种茶和发展旅游业是造成武夷山风景区景观格局变化的主要原因。 7、运用分室模型进行干扰模拟分析得出如下结论:在不改变景观类型丰富度的条件下，景观多样性随人类干扰强度的增大而增加;在改变景观类型丰富度的条件下，中等干扰将提高景观多样性，而强度或弱度千扰将降低景观的多样性，从而证实并丰富了前人有关干扰对景观多样性影响的论断。另外，人工神经网络方法在揭示景观多样性与景观类型数之间关系时效果理想。 8、武夷山风景区6类森林景观(马尾松林、杉木林、经济林、竹林、阔叶林和茶园)存在明显的土壤异质特征。其中，在土壤物理性质上表现为天然林(马尾松林、杉木林与阔叶林)景观结构层次复杂，土壤粒径分布的分形维数小，>0.25rn们。和>5~水稳定性团聚体含量高，非毛管孔隙发达，土壤水稳定性能良好，并具有良好的渗透性、自动调节及其抗逆性能;而人工林或半人工林景观(经济林、竹林和茶园)林分结构单一，其土壤粒径分布的分形维数较大，土壤渗透性能、>0.25111111和>srn们n水稳定性团聚体含量等方面均低于天然林。在土壤化学性质上表现为受人为经营措施(施肥、垦复等)的影响，人工林的土壤养分要普遍高于天然林。就同一森林类型而言，不同层次的土壤养分也存在着明显的异质性，除pH值外，其它土壤养分指标基本上呈现出随土壤深度的增加而减少的趋势。此外，土壤团粒结构的分形维数与土壤结构体破坏率、水稳定性团聚体含量之间存在极为显著的回归关系，因此，土壤团粒结构分形维数在揭示土壤物理性状方面具有明显的表征作用。 9、物种丰富度、多样性指数及均匀度指数的计算结果进一步表明上述6类森林景观在植物物种多样性方面也存在明显的异质特征。各森林类型物种丰富度的垂直结构为:灌木层>乔木层>草本层;不同森林景观植物物种多样性总体趋势为:阔叶林>马尾松林>杉木林>竹林>茶园>经济林;基质景观天然马尾松林植物物种多样性随演替进展而逐渐增加;不同研究区的物种多样性比较结果表明武夷山风景名胜区具?更多还原

【Abstract】 Located in the northwestern of Fujian Province, Wuyi Mountain is the most outstanding area for biodiversity conservation in southeast China and a refuge for a large number of ancient, relict species, many of them endemic to China. In December 1999, Wuyi Mountain was included on the World Natural and Cultural Heritage List by the 23rd Session of the World Heritage Committee of the UNESCO. Covering a total area of 999.75 km2, it is now the largest of China’s World Heritage sites, with the Biodiversity Preserve to the west, Nine-twist Stream Ecological Preserve in the center, Natural Beauty & Cultural Landscape Preserve to the east (i.e. Wuyishan Scenery District) and the ancient Chencun Minyue Relics to the southeast.In order to preserve Wuyi Mountain effectively and analyze the relationship between human activities and landscape structure, the Wuyishan Scenery District was chosen for the case study and was first researched by the theory and methods of landscape ecology. In this paper, the spatial pattern, dynamic and spatial relationships of the Wuyishan Scenery District were studied, several problems in landscape ecology included scale, hierarchy and disturbance were further discussed, the compartment theory and the artificial neural network were first introduced into landscape ecology to simulate the machanism between the landscape diversity and human disturbance and to set up the relationships between the landscape diversity and the number of landscape types, respectively, the heterogereities of forest landscapes were also analyzed in the view of soil and plant species diversity. Finally, the Wuyishan Scenery District was comprehensive evaluated by a new index system includes coordination, particularity, society, graident, etc.. The main conclusions showed as follow.-According to the classification principles of landscape ecology and community ecology, the two-grade landscape classification system of the Wuyishan Scenery District was first set up. The first-grade-landscape included Pinus massoniana forest, Cunninghamia lanceolata forest, economic forest, bamboo forest, broad leaved forest, Camellia sinensis plantation, farmland, river, residential area and bare area in all, and the second-grade-landscape (or element) included four types of Pinus massoniana forest and the other nine first-grade landscape types except Pinus massoniana forest . Among the ten first-grade landscape types, Pinus massoniana forest, Camellia sinensis plantation and farmland were three main components, whose area and perimeter were 84.581 % of the total area and 81.669% of the total perimeter, respectively, and Pinus massoniana forest was dominant in controlling landscape structure, function and dynamic processes.The results of patch characteristics showed that area, perimeter and the number of patches for different landscape elements were very uneven, and the distribution trend of perimeter was similar to that of area, while the relationships between the number of patches and area, perimeter could not be found. From the indices of patch shape, fragmentation, and separation, it could be concluded that river and Pinus massoniana forest were two most complicated landscapes in patch shape and least in fragmentation and separation, while residential area was on the contrary, which was the simplest in patch shape and highest in fragmentation and separation. This proved powerfully that human activity was one of the main factors breaking the landscape into fragments and separation.Six types of probability distribution, which include normal, lognormal, Weibull, Gamma, Beta and negative exponential were chosen to characterize the patches size distribution of the Wuyishan District. The results indicated that more landscape types could be characterized with lognormal distribution, fewer ones could be characterized with Beta distribution and Weibull distribution, and none of them could be characterized with the normal distribution. The patch size distribution was different under three different classifications, that is to say, th更多还原