North Atlantic sea-level variability during the last half-millennium - salt-marsh proxy records of relative sea-level change along the coast of North America.

This dataset contains foraminifera abundance data and chronological information for samples taken from surface sediments and tidal marsh cores from sites at Chezzetcook (Nova Scotia), Sanborn Cove (Maine), and Barn Island (Connecticut). Relative sea-level (RSL) data has been determined by subtracting the elevation at which a fossil sample was originally deposited (calculated by a microfossil transfer function) from its present elevation (measured relative to national geodetic datums). Transfer functions were based on foraminiferal assemblages from previously collected data sets and, in one site (Connecticut), additional local foraminiferal data. The sampled depths cover age ranges of ~960 – 2010 CE (Chezzetcook), ~1741 – 2010 CE (Sanborn Cove), and ~1552 – 2010 CE (Barn Island). Foraminifera data were quantified in core CZ-25 from Chezzetcook, core SN-3.3 from Sanborn Cove, and core BI-47.5 from Barn Island using light microscopy. For the foraminifera abundance data from samples taken from the surface sediments, at Chezzetcook marsh we used the local transfer function developed by Gehrels et al. (2005) from 46 foraminiferal samples collected by Scott and Medioli (Scott and Medioli, 1980). Scott and Medioli grouped Balticammina pseudomacrescens and Jadammina macrescens together as Trochammina macrescens. We also combined these foraminifera in our training set because it was not possible to retrospectively split these taxa. For the transfer function applied at Sanborn Cove we used the regional Maine dataset of Gehrels (2000). We removed all low marsh and tidal flat samples, leaving 44 samples containing the six primary agglutinated salt-marsh foraminiferal taxa. For the Connecticut transfer function we used the dataset of Edwards et al. (2004) and included 26 new samples we collected from Barn Island. The chronological data were derived using a variety of dating methods, including 14C, 210Pb, 137Cs, 206Pb/207Pb, tephra, pollen markers and trace metal markers. The dataset includes 70 dated depths at Chezzetcook, 37 dated depths at Sanborn Cove, and 44 dated depths at Barn Island. The oldest uncalibrated radiocarbon age at Chezzetcook is 996 +/- 36 14C years, with the surface dating to 2004 CE. At Sanborn Cove, the oldest date is 2064 +/- 35 14C years, with the surface dating to 2010 CE. At Barn Island, the oldest date is 1020 +/- 23 14C years, with the surface dating to 2010 CE. The radiocarbon dates include high precision dates obtained by pooling the mean of multiple age determinations on individual plant fragments at 2 sigma precision. This significantly reduces the error of the 14C ages for samples younger than ~300 calendar years (to between +/-9 and +/-17 14C yrs). Measurements of 'bomb-spike' AMS14C in plant fragments are used to test model selection for 210Pb analyses. Age models are developed for cores using the Bayesian age-modelling software Bacon which takes into account the stratigraphical position of the sample, thereby reducing the number of possible solutions for the calibrated ages of the samples. The data were collected as part of an investigation into pre-industrial sea-level rise hotspots in the Northwest Atlantic Ocean and was funded by the Natural Environment Research Council. The grant (NE/G003440/1): North Atlantic sea-level variability during the last half-millennium, aimed to quantify sea-level changes in the recent past in order to provide more accurate and precise predictions for the future. W. R. Gehrels (University of York), S. Dangendorf (University of Siegen), N. L. M. Barlow (University of Leeds), M. H. Saher (Bangor University), A. J. Long (Durham University), P. L. Woodworth (National Oceanography Centre), C. G. Piecuch (Woods Hole Oceanographic Institution), and K. Berk (University of Siegen) were responsible for the data collection and interpretation.