Bathymetric species-diversity patterns and boundary constraints on vertical range distributions

Abstract

A system where a set of species is randomly distributed within two spatial boundaries produces parabolic species diversity patterns, where diversity is lowest in the vicinity of the boundaries, and highest in the mid-zone between the boundaries. We demonstrate this phenomenon using simulated data. We tested whether parabolic depth trends in species diversity (Rex, 1981, 1983) can be explained by invoking only this phenomenon. We analyzed data sets for northwest-Atlantic gastropods and polychaetes previously used to document parabolic diversity patterns. Data sets are matrices of species’ abundance where rows index depth, and columns species. Simulations where species vectors were randomly rearranged within the shallowest and deepest stations generally produced parabolic diversity patterns, with higher diversity at intermediate zones, and lowest diversity closest to the shallowest and deepest stations. For the gastropods, the location of the peaks for the observed and randomly rearranged taxa coincided at similar depths. Random rearrangements, however, did not match the original patterns in curvature (peakedness of the diversity curve) or in magnitude (height of the diversity peak). Highest values were for observed taxa, implying that the original distribution of species is highly non-random and that other factors not assumed in the simulation influence bathymetric species-diversity patterns. For the polychaetes, randomly rearranged data sets matched the magnitude of the observed data set. The original parabolic diversity curve, however, peaked at a shallower location than the random rearrangements, and the magnitude of the peak was higher for the observed taxa. Overall, we find that the random rearrangements cannot explain most characteristics of the parabolic diversity patterns of gastropods and polychaetes. We also explored the influence of species with large vertical range in influencing parabolic species diversity patterns. Removal “experiments”, where a portion of the species with the largest vertical range was removed, also removed parabolic bathymetric diversity patterns, for both observed and simulated taxa, suggesting that species with large vertical range are disproportionally important in determining such patterns.