QUANTIFYING THE RELATIONSHIP BETWEEN OSTRACODE ASSEMBLAGES AND THE ENVIRONMENT IN LAKES ON RUM CAY AND SAN SALVADOR ISLAND BAHAMAS

Transfer functions are commonly used methods of paleoenvironemntal reconstruction which convert preserved fossil assemblages into an environmental variable. Transfer functions are constructed by measuring the strength of modern assemblage-environment and applying this relationship to preserved assemblages. Here we tested whether the precision of a transfer function can be improved by expanding the modern dataset used to measure the assemblage-environment relationship. We sampled sixty-four modern assemblages of ostracodes from lakes on two Bahamian islands: Rum Cay and San Salvador and five associated environmental parameters. Bahamian islands provide an ideal site for this study since they can be analyzed at nested spatial scales: island, interisland, and archipelago. Consequently, we measured the strength of the assemblage-environment relationship within both islands separately and as a combined interisland dataset.

When analyzed separately at the scale of individual islands, assemblages change independently of space with conductivity (salinity), but with a stronger relationship on San Salvador than on Run Cay. This indicates that within islands, species are able to track their preferred habitats, so preserved assemblages should be faithful records of the environment. However, when analyzed at the interisland scale, spatial factors correlate more strongly than any environmental parameters measured. But, when spatial factors are accounted for, conductivity does independently explain changes in assemblages. That spatial factor explain why modern assemblages are different suggests that not all species are able to track their preferred habitat between the two islands. But the persistence of conductivity in explaining differences in assemblages at the interisland scale in controlling differences in assemblages suggests that the abiotic environment still drives changes in assemblages, albeit to a lesser extent than within islands. Adding samples from San Salvador to those of Rum Cay strengths the assemblage-environment relationship, but the reverse is not true- adding Rum Cay assemblages to those of San Salvador weakens the relationship. Thus, paleoenvironmental records should be generated at the island scale for San Salvador, but at the interisland scale in the case of Run Cay.