| Summary: | 碩士 === 國立臺灣大學 === 森林學研究所 === 88 === Spatial patterns of a plant community determine the intensity and mode of local competition, they may also induce habit heterogeneity through the differential modifications of soil and microclimate. As an on going project to understand the population biology and ecology of Taiwan sassafras (Sassafras randaiense Rhed.), this study focused on the spatial patterns of that species in a stand at Chilan Shan, northeastern Taiwan.
To examine the spatial patterns of Taiwan sassafras at different resolution, both quadrat and distance approaches were used in this study. For quadrat approach, which was used to examine the spatial patterns at a stand level, two-term local quadrat variance (TTLQV) method was adopted. Distance approach, which was used to examine the spatial patterns at an individual tree level, included nearest neighbor method, distance to second-nth nearest neighbors, a goodness-of-fit method, a nonparametric trend analysis method, and logistic regression.
The results showed that as a whole Taiwan sassafras in that stand had an aggregated pattern. However, different size classes had different patterns, with smaller- and medium-sized trees showing aggregated patterns, but large trees were randomly distributed throughout the stand. Trend analysis suggested that no particular relationship between the size of a Taiwan sassafras tree and its neighbors. Results from logistic regression further suggested that, based on the collected data, we were unable to predict the probability whether the nearest neighbor with a given size would be present at a particular distance or direction. However, the results from a main effect model did show that there was no association among distance, direction and size.
|
|---|
