Temperature calibration of Mg/Ca ratios in the intermediate water benthic foraminifer Hyalinea balthica

Core top samples from Indonesian and northeast Atlantic depth transects were used to calibrate Mg/Ca and delta O-18 in tests of the calcitic benthic foraminifer Hyalinea balthica to bottom water temperature between 4 degrees C and 13 degrees C. This shallow infaunal species is primarily abundant in neritic to upper bathyal sediments (< 600 m). Both linear and exponential calibrations suggest a temperature sensitivity of similar to 12% per degrees C that is similar to 4 times higher than observed in other species of deep-sea benthic foraminifera. Culture experiments support the core top calibration. We find no discernible effect of salinity and saturation on Mg/Ca. Comparison between the measured benthic foraminiferal delta O-18 and predicted equilibrium values suggests that on average H. balthica delta O-18 is 0.64 parts per thousand similar to 0.13 parts per thousand lower than predicted from the equilibrium composition. To test the reliability of using paired H. balthica Mg/Ca and delta O-18 measurements for reconstructing seawater delta O-18(sw) and salinity, we apply this calibration to another depth transect from Cape Ghir off NW Africa, which was not included in the calibration. Based on error analysis of the calibration data and this validation test, we show that the uncertainty of reconstructing bottom water temperature and salinity from paired Mg/Ca and delta O-18 measurements of H. balthica is better than +/- 0.7 degrees C and +/- 0.69 practical salinity scale, respectively. The small uncertainties allow for the reconstruction of seawater density to better than 0.3 sigma(theta) units, which is precise enough for the identification of specific water masses and reconstruction of changes in their properties. We propose that the relatively high Mg content and temperature sensitivity of H. balthica might be due to minor, biologically mediated contribution of high-Mg calcite to the primarily low Mg calcite test, which is influenced by the ambient temperature. This hypothesis, if correct, suggests that benthic species with relatively high Mg/Ca may be better suited for deepwater temperature reconstructions than species that have thus far been more commonly used.