A record of bottom water temperature and seawater δ¹⁸O for the Southern Ocean over the past 440 kyr based on Mg/Ca of benthic foraminiferal Uvigerina spp.

Abstract

The sensitivity to temperature of Mg/Ca ratios in the shallow-infaunal benthic foraminifera Uvigerina spp. has been assessed. Core-top calibrations over ∼1–20 °C show a range in sensitivity of 0.065–0.084 mmol/mol/°C but few data are available spanning the temperature range anticipated in deep-sea records over glacial–interglacial cycles. In contrast to epibenthic foraminiferal species, carbonate ion saturation appears not to affect Mg/Ca significantly. A method based on estimating the ratio of the temperature sensitivity of foraminiferal Mg/Ca to that of δ¹⁸Ocalcite shows that sensitivity for Mg/Ca at the high end of the observed core-top range (∼0.1 mmol/mol/°C) is required for consistency with LGM–Holocene differences in each property as constrained by independent proxy data. This is supported by a Mg/Ca record for Uvigerina spp. generated for the Southern Ocean over the past 440,000 years from Ocean Drilling Program Site 1123 (Chatham Rise, New Zealand). The record shows variability that correlates with climate oscillations. The LGM deep ocean temperature derived from the Mg/Ca record is −1.1 ± 0.3 °C. Transformation to temperature allows estimates to be made of changes in bottom water temperature and seawater δ¹⁸O and comparison made with literature records. Analysis reveals a ∼2.5-kyr lead in the record of temperature over calcite δ¹⁸O and a longer lead over seawater δ¹⁸O. This is a reflection of larger phase offsets at eccentricity periods; phase offsets at tilt and precession are within error zero.

Introduction

Mg/Ca ratios of foraminifera have the potential to determine the temperature of the seawater in which they calcified and, in combination with foraminiferal δ¹⁸O, provide an estimation of seawater δ¹⁸O (δ¹⁸O_seawater; Mashiotta et al., 1999, Elderfield and Ganssen, 2000). This proxy has been extensively applied to planktonic foraminifera (see Barker et al., 2005, for a review) but the application of Mg/Ca thermometry to investigate deep-sea temperature (and, with δ¹⁸O, global ice volume) using benthic foraminifera has been more limited (see Elderfield et al., 2006, and Bryan and Marchitto, 2008, for reviews).

Because Mg/Ca ratios of foraminifera are biologically mediated (Erez, 2003), empirical calibrations studies are essential to determine temperature sensitivity and the roles if any of other contributing factors. Unlike the situation for planktonic foraminifera where calibrations from culture experiments, sediment traps and core-top samples can be compared (most results define temperature sensitivities of 9–10% increase in Mg/Ca per °C), benthic Mg/Ca calibrations have relied almost exclusively on core-top studies. Recent reviews (Elderfield et al., 2006, Bryan and Marchitto, 2008) show a range in temperature sensitivities with values between about 6 and 15% per °C.

An important additional factor recognised in recent work is the influence of carbonate ion saturation ($\Delta \left[{\text{CO}}_3^{2-}\right]={\left[{\text{CO}}_3^{2-}\right]}_{\text{measured}}-{\left[{\text{CO}}_3^{2-}\right]}_{\text{saturation}}$) on Mg incorporation into benthic foraminifera (Elderfield et al., 2006, Rosenthal et al., 2006) as inferred for other trace elements, Cd, Ba, Zn, Sr, Li and B (McCorkle et al., 1995, Elderfield et al., 1996, Marchitto et al., 2000, Marchitto et al., 2005, Rosenthal et al., 2006, Lear and Rosenthal, 2006, Yu and Elderfield, 2007). At present there are differences of opinion as to whether the carbonate ion effect is minor (Lear, 2007) or for some species dominates over temperature (Yu and Elderfield, 2008), or is limited to ocean waters near or below saturation or is more extensive (Elderfield et al., 2006, Bryan and Marchitto, 2008).

One previous study has estimated bottom water temperature from benthic Mg/Ca for the past ∼330 kyr (Martin et al., 2002), and highlighted the potential influence of dissolution (or saturation state) as an important secondary influence requiring further study. In the present study we have followed the approach of Martin et al. (2002) with the purpose of assessing how well bottom water temperature history may be derived from benthic foraminiferal Mg/Ca using Uvigerina spp. This genus has a shallow infaunal habitat and, as such, may show a weaker (or perhaps absent) sensitivity to carbonate ion than epifaunal species (Elderfield et al., 2006). Firstly, the approach taken has been to summarise new and literature core-top calibration results for Uvigerina spp. followed by two alternative methods of calibration, one of which provides a promising approach. Secondly, we present new benthic Uvigerina spp. Mg/Ca data from Ocean Drilling Program (ODP) Site 1123 located in the southwest Pacific sector of the Southern Ocean over the past 440 kyr and convert data into a bottom water temperature record. We assess the fidelity of the record through sensitivity analysis, comparison with other records, and via estimation of δ¹⁸O_seawater.