Fuzzy Components in Community Level Comparisons

Abstract

The conventions of traditional plant community analysis assume an arrangement of the community components in discrete populations. While this assumption is in line with the notion of absolute discreteness in classical taxonomy in the sense that no organism may belong to more than one taxon, it is not on all fours with reality. The assumption of overlapping populations is indeed more realistic by virtue of the fact that organisms will show affinities to other organisms not in their taxa and not in their community. Starting from this fundamental fact, and observing that by practicing the assumption of absolute discreteness leads to an over-abundance of zeros (absences) in the data and through this to increased levels of indeterminacy in community comparisons, it makes sense to us to pass from an absolutely discrete taxonomy to one which has provisions for similarity-based population overlap. In one case, the field records use discrete taxa which are replaced in the analysis by their fuzzy set equivalents; the latter which emphasize a posteriori taxon similarities. Fuzzy sets of this kind form the basis of a new similarity measure which we propose for community level comparisons. We describe this measure and illustrate it by example.