Premium
Vegetation‐environment relationships of old‐growth spruce forest vegetation in Østmarka Nature Reserve, SE Norway, and comparison of three ordination methods
Author(s) -
Rydgren K.
Publication year - 1996
Publication title -
nordic journal of botany
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.333
H-Index - 33
eISSN - 1756-1051
pISSN - 0107-055X
DOI - 10.1111/j.1756-1051.1996.tb00254.x
Subject(s) - ordination , vegetation (pathology) , spatial variability , gradient analysis , canopy , environmental science , ecology , plant cover , environmental gradient , physical geography , biology , geography , mathematics , statistics , medicine , pathology , habitat
The relationship between vegetation and environmental variables has been studied in 100 sample plots, each 0.25 m 2 , in old‐growth spruce forest at Høgkollen, ØOstmarka Nature Reserve, SE Norway. Each sample plot was supplied with measurements of 13 environmental and 5 biotic variables. Parallel application of three ordination techniques, PCA, DCA and LNMDS, resulted in different sample plot configurations. PCA performed poorest due to strong influence of outliers and circumstantial evidence indicated better performance of LNMDS than DCA. Statistical analyses of the relationships between vegetation and ecological data revealed a parallel gradient in soil moisture (decreasing) and canopy closure (increasing) as the most important for differentiation of the vegetation. Species number and field layer cover decreased, while bottom layer cover increased, due to increasing cover of Dicranum majus , with decreasing soil moisture and increasing canopy closure. Constrained canonical correspondence analysis (CCA) was used to partition the variation of the species‐sample plot matrix into spatial, environmental and unexplained variation, and combinations. The fraction of unexplained variation was high (80.9 %), most likely due to small sample plot size and short gradient lengths. Most of the explained variation was attributable to environmental factors alone (54.5%). Only 6.3% was shared between environmental and spatial variation, which indicated minor importance of broad‐scale and geographically structured environmental variation. Strictly spatial variation constituted 39.3%. However, the spatially structured environmental variation was low, so the causes of spatial variation were likely not to be found among the measured environmental variables.