A Comparison of Clustering Techniques Using Biogeographical Data

Geographers and ecologists frequently employ clustering methods to achieve meaningful ecological groups based on similarity measures. Four popular clustering techniques, and resulting similarities and differences produced by each method when applied to biogeographical data are reviewed. Data were collected and classified within a riparian site on a mid-river island in a tornadodamaged area of northwestern Pennsylvania and on upland sites on either side of the river. Results illustrate that similar environmental clusters are produced by the four methods under investigation. However, more complex relationships between upland sites and the riparian sites also emerged that would not have been discovered had only the ordination technique been used.