The Concept of Taxon Invariance in Ecology: Do Diversity Patterns Vary with Changes in Taxonomic Resolution?

Diversity patterns cannot be properly interpreted without a theory providing criteria for their evaluation. We propose a concept to prevent artifictions caused by improper consideration of changes in observed patterns due to variation in taxon delimitation. Most biodiversity patterns concern assemblages of species of given higher taxon (e.g. class). Some patterns seem to be universal, e.g., body size distribution, species-abundance distribution, species-area relationship, or the relationship between diversity and energy availability. However, truly universal patterns should not change when we change taxonomic scope by focusing on subtaxa or when we merge several sister taxa together and analyze patterns in resulting higher taxon. Similarly, some patterns may not change when changing the basic unit of the study e.g., when replacing species by genera or families (or any monophyletic clades), although other patterns may not be invariant against the variation of the basic unit. In fact, there are only two possibilities: biodiversity patterns are either taxon-invariant or they vary systematically with taxonomic resolution, which would indicate some fundamental taxonomic level with interesting implications for biological processes behind those patterns. Here we develop the concept of taxon invariance of diversity patterns and apply it on the abovementioned patterns. We show that simple theoretical considerations markedly constrain the set of possible patterns, as some of them cannot be simultaneously valid for both a taxon and its subtaxa - frequency distributions of abundances cannot be simultaneously lognormal for a given taxon and all its subtaxa, the taxa-area relationship cannot follow a power-law for all levels of taxonomic resolution, and energy availability cannot affect diversity of all taxonomic units in the same way. Analyses of the variation in the form of biodiversity patterns with changing taxonomic resolution thus provide an extremely useful tool for revealing properties of respective patterns, their universality and logical consistency.