Linear and non-linear successional models have both been proposed for Bergen Swamp, NY, USA, a unique strongly mineotrophic mire with active marl deposition.  However, no taxonomic survey of the plant communities has been performed recently to provide suitable data for testing these hypotheses. In summer 2004, an array of randomly placed plots was established using a geographic information system (GIS) throughout Bergen Swamp to survey plant communities and measure associated edaphic factors. The survey included stem counts of herbaceous plant species and shrubs and associated abiotic factors (e.g. woody debris, leaf litter depth, pH, etc.). Landscape environmental data (e.g. distance to open water, depth to water table, soil type, etc.) were derived using a GIS. We performed a Raup and Crick clustering analysis at two different spatial scales to group plant communities.  We then used detrended correspondence analysis (DCA), an indirect gradient analysis, to infer and predict important local and landscape environmental gradients associated with identified communities. We also used a direct gradient analysis, canonical correspondence analysis (CCA), to test hypotheses of major environmental gradients and to compare them to those determined in past studies.
Results/Conclusions

Results of the clustering analysis demonstrated that there were five communities at the subplot level, and three communities at the plot level. Observed differences between spatial scales are possibly a result of micro-topological differences related to hummock and hollow formation. The major environmental gradients associated with plant communities were, in order of decreasing importance, depth to water table, hydrologic activity, and pH. These data combined with a map showing the arrangement of predominant community types were used to provide support for one of two proposed models of succession for the swamp. The results indicate that Bergen Swamp has undergone non-linear succession and plant communities in the swamp have overall been in a relatively stable state.