Classification of vegetational diversity in managed boreal forests in eastern Finland

Diversity of vegetation in managed forests is studied. A classification based on forest stand structure, the abundances of vegetation species and variations in these abundances is developed and diversity indices are calculated for the classes to describe the diversity of the vegetation within the classes. The classes were formed using detrended correspondence analysis (DCA), global nonmetric multidimensional scaling (GNMDS) and TWINSPAN classification. Discriminant analysis was used to determine the environmental variables differentiating between the classes, and Duncan's multiple range test was used to examine the ability of the diversity measures to distinguish the classes. Beta diversity was estimated with Okland's method based on DCA ordination of the sample plots. The results point to fertility and the successional stage of the stand as the main factors affecting species diversity, in addition to which soil type, the number of tree species, crown cover, basal area and certain variables describing the management of the stand were relevant to the classification. The most distinct diversity indices were the reciprocal of Simpson, Pielou's J' and species richness, while the highest alpha diversity was found in young stands with a low crown cover on herb-rich or mesic forest sites. Beta diversity was quite high, its highest values of all for the whole data being recorded along the fertility gradient.