A forested wetland data set from northeastern Ontario, consisting of species cover and environmental measures in 43 stands, was analyzed with canonical correspondence analysis. Results showed two main vegetational gradients related to factor complexes of peat depth ndash; moisture (mire margin to mire expanse) and pHndash;calcium. Stands within each of the vegetation types were positioned closely, and gradients of types were similar to those from earlier analyses, suggesting the validity of a previous classification. Axis I of the ordination was highly related to peat depth, several elements (Al, Fe, and Cu), loss on ignition, bulk density, and water content in peat. Axis II was highly related to loss on ignition, depth of fibric layer, pH, and several elements (Ca, Mg, Mn, and N). The number of species in a plot was strongly correlated to the pHndash;calcium gradient, whereas vegetation cover was strongly correlated to the peat depth ndash; moisture gradient. Analysis with detrended correspondence analysis gave results very similar to canonical correspondence analysis, suggesting that there was a relatively high correspondence between vegetational and environmental gradients. Environmental measures were partitioned into physical and chemical attributes, to detect the relative contribution to vegetational variation. Both physical and chemical variables were important, and 81 of the variation in vegetation was explained by the environmental measures.Key words: boreal forest, multivariate analysis, Ontario, wetlands, vegetation pattern, diversity.
展开▼
