Classifying natural forests by multivariate analysis (I)

概要

論文の詳細を見る
For the management of natural forests, it is impracticable to have too many forest types. This paper analyzes the relationships between data describing each of the natural selection forests that make up the Tokyo University Forest in Hokkaido. Multivariate analysis is applied to the survey results, and the possibility of subdividing the forest into different natural forest types is considered. The density distribution of the number of trees in six tree-species groups and six diameter classes are used as the basic data. The analysis proceeds as follows: 1) Natural selection forests are classified using cluster analysis, and a dendrogram for each is drawn. 2) Using principal component analysis the score of each of the survey results is plotted on three scatter diagrams. 3) The relationships between each natural forest stand and the trends of changes in each stand are considered using both a dendrogram and three scatter diagrams. 4) Finally, using the results of the analysis, the suitability of different combinations of aggregation in the classification of natural selection forests is discussed from the point of view of the forest manager. The results were consistent with experience and showed the advantages of using multivariate analysis for the classification of natural forest stands.
日本森林学会の論文