Tiivistelmä |
Näytä lisätiedot
|
Artikkeli PDF-muodossa |
Tekijä
The relationships between 32 peatland and forest vegetation types described from southern Finland were studied by different multivariate techniques. The material included regional averages (by Kalela 1977) of the following forest types (reference numbers, cf. Figs. 1-3, mentioned in parentheses): (1-2) CT, Calluna type, (3-7) YT, Vaccinium vitis-idaea type, (8-13) MT, Vaccinium myrtillus type, (14-19) OMT, Oxalis - Vaccinium myrtillus type, (20) Pyrola type, (21) OT= OMaT, Oxalis-Maianthemum type, (22) FT, Filices type (for characterization of vegetation, see Kujala 1961, Kalliola 1973). The following forested mire types were included in this study (described by Eurola 1962, cf. also Heikurainen & Pakarinen 1982): (23-26) IR, low-shrub pine bogs, (27-28) KR, spruce-pine swamps, (29-31) VK, ordinary spruce swamps, and (32) herbrich hardwood-spruce swamps. Mean percentage cover of bottom and field layer species was used in numerical analyses.
A classification (Fig. 1) resulting from the agglomerative hierarchic clustering (Pritchard & Anderson 1971, Pakarinen 1976) shows four major groups: (A) dry heath forests - CT, VT, (B) mesic heath forests - MT, OMT, (C) spruce mires and herbrich forests, and (D) pine mires. Association analysis based on the presence-absence data (information-analytical method of Podani 1979) indicates that Luzula pilosa is primarily restricted to the upland (non-peaty) forests in the study area, and the reindeer lichen Cladonia stelloris (= C. alpestris) is limited in its occurrence to dry heath forests (Fig. 2); on the other hand Aegopodium podagraria and Anemone nemorosa appear to be differential species of herbrich forests. One should note, however, that the species mentioned are not constant in the primary material composed of site type means (not original sample plots). Therefore it is suggested that a threshold frequency go should be applied when developing identification keys to community types on the basis of individual species. An alternative approach, use of eigenvector methods (factor analysis, PCA, reciprocal averaging) is also briefly discussed in the paper (cf. Pakarinen 1979). Polar ordination (Bray-Curtis ordination, Fig. 3) based on percentage similarities (Jalas 1962, Gauch 1973) illustrates two major gradients in the study material: degree of paludification (x-axis), and trophic status (y-axis). While the material discussed in this paper represents the major types of mineral soil and peatland forests in their natural state in S Finland, description of different paludified forest types (with peat layer less than 30 cm) would complement the ordination diagrams. Much further research is still needed to elucidate the vegetational changes caused by nowadays common forestry practices, such as clear-cutting, peatland drainage, fertilization, etc.
-
Pakarinen,
Sähköposti:
ei.tietoa@nn.oo