盐城自然保护区丹顶鹤生境选择影响因素分析

发布时间：2018-03-08 06:38

本文选题：丹顶鹤　切入点：生境选择　出处：《南京林业大学》2015年硕士论文　论文类型：学位论文

【摘要】：识别影响濒危物种生境选择的关键环境变量是进行物种生境保护工作的前提。然而环境变量存在于不同的空间尺度上,只选取单一尺度的环境变量不能全面反映物种的生境特征;另外,环境变量存在的空间自相关现象(空间上越靠近越相似)和多元共线性问题(变量之间线性相关)分别违背了统计检验中样本点之间相互独立的基本假设以及自变量之间相互独立的基本假设,这些问题可能会导致推断出的影响物种生境选择的环境变量不够准确。为解决尺度、空间自相关和多元共线性这3个问题,本文以盐城湿地珍禽国家级自然保护区为研究地,以越冬期的丹顶鹤为研究对象,结合野外调查和遥感信息采集微生境和景观2种尺度上的环境数据,同时在环境变量中加入空间变量以描述空间自相关现象,将环境变量按照2个等级水平进行划分,在第一等级水平上将所有环境变量分为微生境、景观和空间3组变量,在第二等级水平上将微生境变量分为植被因子和遮蔽物因子,将景观变量分为植被因子和干扰因子,然后利用方差分解法计算第一等级水平上微生境、景观和空间的独立效应和联合效应,以及第二等级水平上各生境因子对丹顶鹤分布变异的独立解释及联合解释,最后利用等级划分法验证方差分解法的结果,并且计算各环境变量对丹顶鹤分布变异的独立贡献。通过研究,主要得到以下结果:(1)第一等级水平的方差分解结果表明:所有环境变量共解释了67.4%的丹顶鹤分布变异。景观变量总共解释了52.9%的丹顶鹤分布变异,微生境变量总共解释了32.2%的丹顶鹤分布变异,空间变量总共解释了33.7%的丹顶鹤分布变异。微生境、景观和空间3组变量的独立解释都很小,分别独立解释了7.0%、4.5%和5.7%的丹顶鹤分布变异,而它们联合解释了50.2%的丹顶鹤分布变异,其中景观变量与微生境变量、空间变量分别联合解释了22.2%和25.0%的丹顶鹤分布变异。(2)第二等级水平的方差分解结果表明:微生境变量对丹顶鹤分布变异总的解释方差比例(32.2%)中,遮蔽物因子和植被因子分别独立解释了9.8%和8.6%的丹顶鹤分布变异,联合解释了13.8%的丹顶鹤分布变异。景观变量对丹顶鹤分布变异总的解释方差比例(52.9%)中,干扰因子和植被因子分别独立解释了17.4%和39.9%的丹顶鹤分布变异,联合解释了-4.4%的丹顶鹤分布变异。(3)等级划分结果表明:微生境变量、景观变量和空间变量分别独立解释了9.1%、14.9%、3.1%的丹顶鹤分布变异。景观变量和微生境变量中的植被因子分别解释了7.1%和6.1%的丹顶鹤分布变异,高于其它生境因子的解释量。500 hm2范围内的碱蓬面积百分比、50 hm2范围内的水源面积百分比、遮蔽物距离对丹顶鹤分布变异的独立解释最高,它们分别解释了4.12%、3.34%和3.04%的丹顶鹤分布变异。研究结果表明:景观尺度是影响丹顶鹤生境选择的最重要尺度。景观尺度上的植被因子是影响丹顶鹤生境选择的最重要生境因子。碱蓬面积是影响丹顶鹤生境选择的最重要环境变量。空间变量与景观变量的联合解释很高(25.0%),表明丹顶鹤分布的空间结构与景观变量的空间结构有很强的相关性,因此景观生境因子的空间分布特征能够准确反映丹顶鹤喜好或厌恶的生境类型。方差分解法和等级划分法对丹顶鹤分布变异的分解结果基本一致,都得出景观尺度和植被因子是限制丹顶鹤分布的最重要尺度和生境因子。
[Abstract]:Identifying key environmental variables selection of endangered species habitat is a prerequisite for habitat protection. However, environmental variables exist in different spatial scales, environmental variables can not only choose a single scale fully reflect the habitat characteristics of species; in addition, the environmental variables are the spatial autocorrelation phenomenon (space closer and more similar) multicollinearity problem (linear correlation between variables) are contrary to the basic assumption between the sample points in the statistic test independent basic assumptions and variables independent of each other, these problems will lead to environmental variables of species habitat selection effect inferred is not accurate enough. In order to solve the problem of scale, spatial autocorrelation and multicollinearity of the 3, taking the Yancheng wetland national reserve for the study, to the wintering Red Crowned Crane as the research object, combined with the field The environmental survey data and remote sensing information collection and microhabitat 2 landscape scale, and spatial variables added in environment variables to describe the spatial autocorrelation phenomenon, environment variables are divided according to the 2 level in the first grade level, all environment variables are divided into microhabitat, landscape and space in 3 groups of variables. Second grade level, microhabitat variables are divided into vegetation and cover factor, the landscape variables are divided into vegetation factor and disturbance factor, then decomposition method to calculate the first level of microhabitat use variance, independent effect of landscape and space and combined effect, and the second level of the habitat factors of independent interpretation of the Red Crowned Crane Distribution the variation and joint interpretation, finally by using grading method verify the variance decomposition results, and calculate the environment variables on the distribution of Red Crowned Crane variant of the independent contribution. Through research, the main results are as follows: (1) the first level of the variance decomposition results show that: all the environmental variables explained 67.4% of the variation in the distribution of red crowned cranes. The landscape variables altogether explained 52.9% of the variation in the distribution of Red Crowned Crane, microhabitat variables explained 32.2% of the total variation of Dan crane top distribution, spatial variables altogether explained the Red Crowned Crane 33.7%. Spatial variation of microhabitat, independent landscape and 3 sets of variables are very small, respectively explained 7%, 4.5% and 5.7% of variation in the distribution of Red Crowned Crane, and they jointly explained 50.2% of the variation in the distribution of red crowned crane, the landscape variables and microhabitat variables, combined with spatial variables explained the Red Crowned Crane distribution variation of 22.2% and 25%. (2) the second level variance decomposition results show that microhabitat variables on the distribution of Red Crowned Crane explained variance proportion (32.2%) of the total variation, shelter and vegetation factor Independent factors were explained 9.8% of the variation and distribution of Red Crowned Crane 8.6%, jointly explained 13.8% of the variation in the distribution of red crowned cranes. The landscape variables on the distribution of Red Crowned Crane explained variance proportion the total variation (52.9%), interference factor and vegetation factor respectively explained 17.4% and 39.9% of the top Dan crane variation distribution, joint interpretation of red crowned the distribution variation of -4.4%. (3) classification results show that microhabitat variables, landscape and spatial variables respectively explained 9.1%, 14.9%, 3.1% of the variation in distribution of Red Crowned Crane. The landscape variables and microhabitat variables in the vegetation factor explained 7.1% and 6.1% of variation in the distribution of Red Crowned Crane, the area percentage of Suaeda salsa higher than that of other ecological factors explain the amount of.500 in the range of Hm2, the area percentage of water within the range of 50 Hm2, covering the distance independent distribution variation of Red Crowned Crane is the highest, which accounted for 4.12%, 3. The Red Crowned Crane distribution variation of 34% and 3.04%. The results showed that the landscape scale is the most important influence on scale selection of Red Crowned Crane Habitat. Vegetation landscape scale is the most important influencing habitat selection of Red Crowned Crane. Salsa area influence the selection of Red Crowned Crane and the most important environmental variables. Combined with spatial variables and interpretation the landscape variables is very high (25%), showed a strong correlation between the spatial structure of the Red Crowned Crane Distribution spatial structure and landscape variables, habitat types so landscape habitat factors can reflect the spatial distribution of Red Crowned Crane preferences or disgust. Variance decomposition method and classification method of decomposition of the variation of the basic distribution of Red Crowned Crane get consistent, landscape scale and vegetation factor is to limit the distribution of red crowned crane scale and the most important habitat factors.

【学位授予单位】：南京林业大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：Q958.1

【共引文献】