A new approach to the species classification problem in floristic analysis

Abstract

AbstractA new method of species (inverse) classification of vegetation data, i.e. classification of species into groups with similar ecological tolerances, is presented which overcomes the problems of species abundance distorting the results. The algorithm TWO‐STEP is based on the use of an asymmetric measure of dissimilarity: image where i, j are species, h is the stand, n is the total number of stands, and xih is the amount of species i in stand h.The algorithm uses the rows of the asymmetric dissimilarity matrix generated as above to form a second symmetric dissimilarity matrix using the measure: image where m is the number of species and k the species.Flexible sorting is applied to generate a species classification. Comparison of results after applying the TWO‐STEP algorithm and a standard alternative to an artificial data set demonstrates its efficacy. TWO‐STEP also shows considerable advantages over previous analyses for a Queensland rainforest data set (quantitative) and an English heath (qualitative) data set. Normalization of species data appears advantageous for quantitative data only.