【摘要】 景观边界特征与景观碎裂化过程之间的相互关系研究已经引起了广泛的关注。利用 1 987年和 1 997年两个时段遥感景观资料 ,编制了卧龙自然保护区的景观类型图 ,以验证景观碎裂化过程导致边界数量 ,特别是短边界数量增加这一假说。研究工作包括 4个方面的内容 :(1 )利用每年的 2 0个 2 0 0× 2 0 0像元的正方形样地 ,分析工作区内景观边界的一般特征 ;(2 )对每个样地边界数量的长度谱分布特征进行拟合 ,确定能够反映边界属性特征的拟合方程参数 ;(3 )将边界属性特征参数与样地的景观碎裂化指数进行比较研究 ,建立景观边界特征与景观碎裂化程度之间的量化关系模型 ;(4 )利用每个时段的 1 0个验证样地来检验关系模型的有效性。研究结果表明 ,所有样地边界数量的长度谱分布特征可以使用对数方程 y=bln(x) +c进行拟合。景观的碎裂化过程将同时导致拟合方程参数 b的绝对值和 c值线性增加 ,景观碎裂化水平的增加将导致景观边界数量和短边界数量均呈指数增长方式。从研究的结果中还可以推断出 ,生境碎裂化水平对于生物多样性的影响也将呈现出明显的放大效应。更多还原

【Abstract】 Extensive attention has been given to the interrelationship between the boundary characteristics and the fragmentation process of landscape. This paper employs the landscape maps obtained through satellite remote sensing data (TM images) in 1987 and 1997 respectively to explore the hypothesis that the landscape fragmentation process will result in an increase in the boundary number of landscape, particularly the number of short ones. The study covers four aspects: (1) 20 square samples of 200×200 cells are used to analyze the common characteristics of boundaries within the study area; (2) analysis of boundary number\|length distribution in each sample is conducted to identify the appropriate formula parameters that can reflect the boundary characteristics; (3) comparative study between the boundary parameters and the fragmentation index of the samples is made to establish a relationship model between the boundary characteristics and the fragmentation level; (4) 10 samples selected from each period of time are used to study the effectiveness of the model above. Analysis of Common Characteristics of boundary and fragmentation. Because of the significant variation of physical environment and human activities, the boundary and fragmentation attribute of the 20 sample exhibit remarkable differences. Variations of boundary number, average boundary length and fragmentation index of the 20 samples in 1987 were 228%, 110%, and 177%, and that in 1997 were 235%, 110%, and 164%. These results indicate that the boundary and characteristics of the same sample didn’t change significantly during the study period. The boundary number and the landscape fragmentation index of four samples have decreased while the average boundary length increased. The changing trend of the other samples was just the opposite. Study of boundary characteristics and the interrelationship between boundary characteristics and landscape fragmentation. The results of this study show that the boundary number\|length distribution characteristics of all samples, even though the values of their fragmentation level are dramatically different, can be appropriately expressed by the logarithm formula y = b ln( x) + c . Increases in landscape fragmentation lead to a linear increase of the value of parameter c and the absolute value of parameter b, indicating that both the total boundary number and the number of short boundary will increase in more fragmented samples. This result has illustrated the relationship hypothesis between boundary characteristics and landscape fragmentation. However, the relationship between the fragmentation process and the changes of boundary number is not simply a linear one. A quantitative conclusion that can be inferred from the model for the relationship between the landscape fragmentation and the boundary characteristics is that an increase in the fragmentation level will lead to a linear increase of the parameters of boundary characteristics. We can conclude form this result that while the landscape becomes more fragmented, the initial changes in boundary number are not obvious but will accelerate later on. Relationship between boundary characteristics and fragmentation level. The results of two tests have supported strongly the model of relationships between the marginal characteristics and the fragmentation level. For the fragmentation level, the error between the model results and the actual results ranges from 1% to 12%, the average value of prediction accuracy is 95%. In the test of using the fragmentation indices to deduce the parameter for the boundary number\|length distribution formula, the average error of parameter b ranges from 1% to 7%, with the accuracy is 97%. The error of parameter c ranges from 1% to 6% and the average rate of accuracy is 97%. The results of the tests have fully proven that the model out of this study for the relationship between the boundary characteristics and the fragmentation level is credible. A deduced conclusion about the edge effect. It can be inferred from the conclusion of this stu更多还原