High precision paleosalinity determination from measured porewater density

Document Type


Date of Original Version



We have developed a density-based method for determining porewater salinity that can be performed shipboard on small volume samples with greater efficiency and precision than the currently available shore-based chloride titration technique. This approach is based on a recently developed water column method that determines salinity at the precision of a conductivity measurement through density measurements and the seawater thermodynamic equation of state. Diagenesis causes deviations in porewater composition from standard seawater values, affecting the density salinity relationship, that we correct for through precise measurements of each ion's concentration before converting measured density to chloride concentration. We account for the diffusive change in porewater chloride that occurs over time independent of diagenesis by optimizing diffusion modeled, sea-level determined bottom water chloride as a function of time to measured modern porewater and converting the best fit to salinity. We applied our density method to porewater samples extracted from adjacent long cores collected from the deep western North Atlantic, determining Last Glacial Maximum (LGM) bottom water paleosalinity in a region critical to understanding deep water mass distribution. High uncertainty is associated with current LGM bottom water salinity characterizations and their implications for LGM overturning circulation and climate. Density was determined to a precision of 2.3 × 10−6 g/mL, which translates to a relative uncertainty of 0.03% for LGM salinity. We compare the high precision chloride concentration profiles determined using our method to profiles determined from chloride titrations of parallel samples. Salinity change at our site between the pre-industrial and LGM is 3.07 ± 0.03 % and 3.65 ± 0.06 % when determined from density and 2.96 ± 0.12 % and 1.96 ± 0.21 % when determined from titrated Cl− for the two co-located cores analyzed. This is consistent with nearby deep Atlantic paleosalinity data (Adkins et al., 2002) and global sea-level-change determined salinity change (Clark and Mix, 2002). By comparing these uncertainties we demonstrate that porewater salinity can be determined to a higher precision and with increased reproducibility through our density protocol compared to titration-determined salinity. Application of our shipboard method at further locations will increase the resolution, precision, and accuracy of available LGM bottom water salinity reconstruction, improving the characterization of glacial deep water masses and overturning circulation.

Publication Title, e.g., Journal

Marine Chemistry