A classification of woody communities based on biological dissimilarity

Abstract

AbstractAimsTraditional quantitative approaches to forest classification are based on differences in species abundance or incidence among communities. In these approaches, all species are regarded as biologically equidistant regardless of the biological heterogeneity. The objective of the study is to evaluate the potential of the “Discriminating Avalanche” approach, which integrates species abundance and biological heterogeneity, as a new basis for forest classification.LocationChina, India, Iran, Ukraine, Germany, Italy, Mexico, Peru, and South Africa.MethodWe illustrate our approach using a set of 35 large tree‐mapped forest plots from various regions of the world. Our dissimilarity matrices, which integrate species abundance with biological heterogeneity, are compared with the standard Bray–Curtis and Whittaker dissimilarity indices, and provide the quantitative basis for a hierarchical cluster analysis.ResultsFour distinct groups of forests were identified using the proposed forest dissimilarity matrix. Afro‐montane forests from South Africa constitute a first group. A second group includes temperate deciduous broad‐leaved forests dominated by oak (Quercus) and beech (Fagus) from Europe and China. Conifer‐dominated forests constitute a third group. The remaining forests constitute the fourth group.ConclusionBiological heterogeneity provides a practical basis for vegetation classification. The results of this study, based on a variety of temperate and tropical forests, suggest that a measure of biological dissimilarity based on evolutionary and morphological differences among species is more effective than the traditional species abundance‐based approaches to classify an arbitrary set of plant communities. This approach promises greater refinement and consistency in ecological classification. In particular, it has advantage in classifying forests along large geographic scales in situations of high beta diversity and species turnover.