Comparing the Complete Hierarchical Structuration of Species Abundances in Reef Fish Communities According to Coral Morphology, Using the Numerical Extrapolation of Only Incomplete Inventories

Numerous investigations have been conducted trying to unveil some possible correlations between coral-associated fish communities and different gross types of coral morphology that are assumed being differentially attractive to various kinds of fishes. Investigations in this respect are, yet, strewn with pitfalls and exposed to unexpected source of bias. One potential cause of bias comes from having to deal with substantially incomplete samplings, a situation practically unavoidable with species-rich communities having very uneven distribution of abundances. To overcome this difficulty, the numerical extrapolation of incompletely sampled communities offers a convenient solution, serving as a surrogate, when having no reasonable possibility to pursue samplings until reaching completion.

As an example, previous work came to the conclusion that relatively faint differences only distinguish the fish communities respectively associated with two distinctly different stony-coral morphologies, namely the regularly convex Diploria strigosa and the deeply lobed Orbicella annularis. This reported conclusion may be considered questionable, however, as it has been drawn from partial inventories only. Accordingly – given the practical impossibility to further improve the levels of completeness of the available samplings – a more soundly established analysis would require the implementation of a procedure of numerical extrapolation of the Species Abundance Distributions of both communities. After completion of the Species Abundance Distributions by numerical extrapolation, it then become clear that, in fact, substantial differences actually distinguish both fish communities, in terms of true (total) species richness and taxonomic distinctiveness. And, rather unexpectedly, it is the coral having lesser gross rugosity (Diploria strigosa) that hosts the most species-rich fish community. Also, the degree of unevenness of species abundances substantially differs between both communities while, however, the genuine intensity of the process of hierarchical structuring of species abundances remains practically unchanged. It is argued that this discrepancy between the unevenness pattern and the intensity of the underlying process actually results from the influence of species richness upon the degree of unevenness of species abundances – a purely mathematical influence, regrettably too often ignored and, indeed, devoid of biological significance.

On a more general methodological ground, this case study demonstrates, once more, the usefulness of numerical extrapolation when dealing with incomplete samplings, in order to avoid the potential risk of deriving erroneous conclusions.