Quantitative vertical zonation of salt-marsh foraminifera for reconstructing former sea level; an example from New Jersey, USA.

We present a quantitative technique to reconstruct sea level from assemblages of salt-marsh foraminifera using partitioning around medoids (PAM) and linear discriminant functions (LDF). The modern distribution of foraminifera was described from 62 surface samples at three salt marshes in southern New Jersey. PAM objectively estimated the number and composition of assemblages present at each site and showed that foraminifera adhered to the concept of elevation-dependent ecological zones, making them appropriate sea-level indicators. Application of PAM to a combined dataset identified five distinctive biozones occupying defined elevation ranges, which were similar to those identified elsewhere on the U.S. mid-Atlantic coast. Biozone A had high abundances of Jadammina macrescens and Trochammina inflata; biozone B was dominated by Miliammina fusca; biozone C was associated with Arenoparrella mexicana; biozone D was dominated byTiphotrocha comprimata and biozone E was dominated by Haplophragmoides manilaensis. Foraminiferal assemblages from transitional and high salt-marsh environments occupied the narrowest elevational range and are the most precise sea-level indicators. Recognition of biozones in sequences of salt-marsh sediment using LDFs provides a probabilistic means to reconstruct sea level. We collected a core to investigate the practical application of this approach. LDFs indicated the faunal origin of 38 core samples and in cross-validation tests were accurate in 54 of 56 cases. We compared reconstructions from LDFs and a transfer function. The transfer function provides smaller error terms and can reconstruct smaller RSL changes, but LDFs are well suited to RSL reconstructions focused on larger changes and using varied assemblages. Agreement between these techniques suggests that the approach we describe can be used as an independent means to reconstruct sea level or, importantly, to check the ecological plausibility of results from other techniques.