An Application of Principal Components Analysis to Vegetation Change in Permanent Plots

Publisher Website
Google Scholar
Abstract
Conventional ordination of repeated enumerations of the same set of stands over a period of time causes confusion between spatial and temporal trends. An alternative technique for isolating temporal trends is to eliminate the effect of initial differences between stands. Between-species correlations are calculated from the species .times. time matrix for each stand separately. The correlation matrices from individual stands are summed, and the summed matrix is subjected to principal components analysis. The method is illustrated with data from montane grasslands from which sheep have been excluded, in Snowdonia, North Wales [UK].