INTRALAKE VARIABILITY IN OSTRACODE ASSEMBLAGES FOR USE IN PALEOENVIRONMENTAL RECONSTRUCTION: A CASE STUDY FROM SAN SALVADOR ISLAND, BAHAMAS

Quantitative paleoenvironmental models are based on the collection of modern assemblages at various sites and measurement of environmental variables that drive the distribution of those assemblages. Past environments can be inferred by transforming abundances of fossil species into the environmental variable controlling their distribution. Many models rely on a few samples of fossil assemblages per site, normally lakes. This method assumes that a few samples capture an entire lake’s community. This assumption could be violated if one lake is made of many habitats or if taphonomic processes obscure environmental gradients. This study tests this assumption using Ostracoda (Phylum: Arthropoda) in lakes on San Salvador Island, Bahamas. 16 living and dead ostracode assemblages were quantified in 7 lakes. Living and dead assemblages were highly and significantly rank-correlated in most lakes. Within individual lakes, assemblages quantified at specific locations were highly and significantly rank-correlated to assemblages quantified at other locations in the same lake. Finally, non-metric multidimensional scaling on the Bray-Curtis dissimilarity matrix with analysis of similarity indicated significant differences in community structure between and not within most lakes The only exceptions to these results occurred in deep, tidally influenced blue holes with seasonal subaerial exposure, where living and dead assemblages were less correlated and intralake variability was more pronounced. The correlation between living communities and death assemblages as well as the lack of intralake variability make ostracode on San Salvador a good system to use for paleoenvironmental reconstruction.