Quantitative paleoenvironmental models are used to transform past communities into the most important abiotic environmental factor driving their assembly. They are created by utilizing modern communities to determine the relationship between a taxonomically-defined metacommunity and its abiotic environment. By directly correlating modern communities with current abiotic factors in their environment, these models assume that the metacommunity dynamics conform to a species sorting model in which the abiotic environment is the main determinate of community β-diversity.
We examined whether species sorting is the most appropriate metacommunity framework for the creation of quantitative paleoenvironmental models by evaluating four metacommunity models: species sorting, mass effects, patch dynamics, and neutral. We collected species abundance data on twenty communities of lacustrine ostracodes (Phylum: Arthropoda) on San Salvador Island, Bahamas, and measured nineteen associated physio-chemical variables. Metacommunity models were evaluated by examining the degree of spatial autocorrelation in community structure, individual species abundances, and the abiotic environment. Furthermore, we determined the relationship between β-diversity and the environment. If the metacommunity dynamics of ostracodes most closely conforms to a species sorting model, then β-diversity would not be spatially autocorrleated, but would correlate with the abiotic environment. We would expect to see more dissimilar communities in more dissimilar environments, and this relationship would be independent of the distance between communities.
Results/Conclusions
We found no abiotic factor was spatially autocorrelated and then assessed whether space or the abiotic environment was more important in determining community structure. β-diversity of communities and individual species abundances were not spatially autocorrelated. Detrended correspondence analysis (DCA) and canonical correspondence analysis (CCA) indicated that community β-diversity was strongly and significantly correlated with three environmental factors: lake area, conductivity, and latitude. These three factors comprise a multivariate hydrological gradient, with latitude probably corresponding to an unmeasured precipitation gradient across the island. β-diversity of ostracode communities did not correlate with the distance between communities, but did correlate with the hydrological environment. This lack of spatial autocorrelation in community structure together with the significant change in community composition across a complex hydrological gradient supported a species sorting metacommunity model. Therefore, the ostracode metacommunity on San Salvador Island can reliably be used in the creation of a paleoenvironmental model to infer past hydrological conditions from preserved ostracode communities.