The vegetation and number of physical and chemical soil properties were studied on a random sample of closed upland forest stands in Southern Finland. The material consists of a total of 410 sample plots. Two-way indicator species analysis (TWINSPAN) was carried out in order to produce a hierarchical clustering of samples on the basis of the vegetation data. Discriminant analysis and analysis of variance were applied in order to find environmental correlations of the vegetation clustering.
The vegetation was found to indicate the nutrient regime of the humus layer well, but to a less extent the properties of the sub-soil. The understorey vegetation was found to be jointly dependent on the site fertility and on the properties of the tree stand, especially on the tree species composition. Although the forest vegetation appears to be distributed rather continuously along an axis of increasing site fertility, relatively unambiguous classification can be based on the appearance of indicator species and species groups.
The results of the study were interpreted as indication that operational site classification done using the vegetation is rather good method for classification in closed forest stands. Different methods produce relatively consistent, natural and ecologically interpretable classifications. The results also imply that the vegetation cover and the humus layer develop concurrently during the development of the ecosystem, but the differentiation of the site type is regulated simultaneously by a number of interacting factors ranging from mineralogical properties of the parent material to the topographical exposition of the site. As the plant cover depicts all these primary factors simultaneously, only a relatively rough ecological site classification can be based on the vegetation.
The PDF includes a summary in Finnish.