To address the objectives of this project, we selected a comprehensive calibration dataset from a rich archive of stratified plankton samples, which was paired with stable oxygen and carbon values of the water column taken at the time the plankton was removed from the water. The final dataset covered a large gradient in temperature, salinity, and carbonate chemistry to test the application of our geochemical proxies to a large range of hydrographic parameters. Next, we combined established and new analytical techniques in trace element and isotope geochemistry to derive and isolate carbonate system parameters from the temperature signal (e.g. Mg/Ca-palaeothermometry) recorded in NPS. This included three separate analytical techniques. Fist, we prepared samples for stable oxygen (δ18O) isotope analysis to determine the geochemical signature of the calcification environment (e.g. calcification depth) of NPS. This was followed by the preparation of sample material for boron isotope δ11B and finally trace element (El/Ca) analysis.
To estimate the influence of seawater carbonate ion concentration [CO3] on Mg/Ca values for our calibration dataset, we derived the relationship between Mg/Ca ratios (Mg/Ca[meas]/Mg/Ca[pred]) and water column [CO3]. As hypothesized our results describe a strong sensitivity of Mg/Ca to [CO3] at low concentrations that is best described by a power regression (r2=0.81). We note that the sensitivity described by our data is stronger than equations presented in Evans et al. [2016] and Morley et al. [2017] for other planktonic foraminifera and propose that the higher sensitivity can be explained by a non-linear response of [CO3] on Mg incorporation at various ambient temperatures. The steeper response supports the hypothesis that at lower temperatures the [CO3] effect is greater than at higher temperatures.
The δ11B data from NPS are lower than that of ambient borate B(OH)-4 ion and plot below the 1:1 line similar to G. bulloides (also non-symbiont baring foraminifera) and O. universa. The slope of the regression of 1.41 (±0.40) indicates that open ocean NPS display a pH sensitivity that is similar (within uncertainties) to that of aqueous borate ion with a slight negative offset from the 1:1 line. Similar to other non-symbiont baring species, the negative offset of NPS δ11B from the 1:1 line could be explained by changes in microenvironmental pH due to respiration since the effect of respiration could lower pH in the microenvironment and thus δ11Bborate [Henehan et al. 2016]. We note that the slope of the regression is slightly larger than 1 (within uncertainties), which could suggest that the effect of respiration on the microenvironment is greater at low pH, which increases the sensitivity of the δ11B-pH proxy for NPS. The regression supports the hypothesis that δ11B measured on NPS are sensitive to changes in pH and that δ11B can thus be used to reconstruct past changes in pH at high northern latitudes.
For the elemental incorporation of Boron over Calcium (B/Ca) into NPS we find that there is a departure of the predicted B/Ca to borate over bicarbonate relationship predicted by Hemming and Hanson [1992] with B/Ca values apparently unrelated to carbonate system parameters. We hypothesize that exceptionally low or constant calcification rates that are associated with NPS, prohibit a dependency of B/Ca values to carbonate system parameters as seen during inorganic calcite experiments [Uchikawa et al. 2017]. Practically, our findings thus prevent us from defining an empirical relationship between B/Ca and carbonate ion concentrations. For past reconstructions of sea surface temperatures an alternative proxy must therefore be used. A potential solution to this issue may be to define temperature, salinity, and carbonate chemistry sensitivities in culture experiments and to correct for carbonate chemistry sensitivities using δ11B similar to the approach developed by Evans and Gray (2019) for other planktonic foraminifera.
