2010 年 9 巻 p. 123-134
This study assesses the applicability of combining the general standardization procedure (GSP) with principal component analysis (PCA) in ecological data analysis. For this purpose, we conducted a comparison of the performance of combined GSP and PCA with detrended correspondence analysis (DCA) in classifying the structure of species composition in natural forest stands at the Kwangneung Experimental Forest. Prior to analyzing the data using PCA, it was standardized using GSP to avoid the tendency of involutions accompanying PCA. Results showed that both approaches were successful in classifying stand species composition and provided ordination surfaces showing three separate groupsconiferous, deciduous and mixed forests. In terms of total explained variance, GSP-PCA seemed to demonstrate better performance than DCA. However, in classifying species composition structures by visual inspection of the ordination graphs, the ordination plots obtained using DCA were more effective than those generated using GSP-PCA.
