Reconciling coral and marine-atmosphere oxygen isotope estimates of sea level change over the last climate cycle

Lead Research Organisation: University of Cambridge
Department Name: Earth Sciences

Abstract

Fundamental to understanding Pleistocene sea level variations is determining the history of changes in continental ice volume. Variations in the d18O of benthic foraminifera reflect global ice volume and sea level changes, but the signal is contaminated by temperature and local hydrographic effects. Here we propose to use the Mg/Ca of benthic foraminifera to estimate deep-water temperature and thereby derive a record of variations of d18O seawater over the past 130,000 years. Our d18O-based sea level proxy will be compared with other proxy records of sea level change from the literature to reconcile coral-based reconstructions with marine and atmospheric d18O records.
 
Description Since the time that deep-sea benthic oxygen isotope records were first produced, a discrepancy has existed between ice volume changes implied by benthic d18O and those derived from coral terraces. Most benthic d18O records are significantly contaminated by a temperature signal such that changes in the d18O of seawater cannot be inferred directly from d18O of calcite (Shackleton, 2000). Current understanding recognizes that changes in the d18O of benthic foraminifera reflect global ice volume (sea level), temperature and local hydrographic effects (Waelbroeck et al., 2002):

d18Ob = d18Ow + dtemp = dice vol + dlocal + dtemp [Eqn. 1]

where, d18Ob = change in benthic d18O; d18Ow = change in d18O of seawater; dtemp = change in temperature; dice vol = change in d18O of seawater due to ice volume; dlocal = change in d18O of seawater due to hydrographic change.

The objective of this project was to estimate changes in d18Ow by tandem measurements of foraminiferal d18O and Mg/Ca at IODP Site U1304 on the Gardar Drift in the North Atlantic. Contrary to sea level estimates from coral terraces indicating higher sea level during MIS 5.5 than 5.3 and 5.1, we found the d18O of seawater (_18Ow) was not substantially different during marine interstadial substages 5.1, 5.3, and 5.5. This finding is supported by independent estimates of d18Ow at ODP Site 1123 (Elderfield et al., 2010) and V19-28 (Shackleton, 2000). The results are also consistent with Dorale et al. (2010) who suggested sea level was ~1 meter above modern at 81 ka during MIS 5a. Our findings are significant because it challenges a long-held belief that sea level was lower by 14 to 18 m during MIS 5.1 and 5.3 relative to today.

Mg/Ca results indicate that deep-water temperature at 3069 m on Gardar Drift cooled by 3oC at the last glacial inception from MIS 5.5 to 5.4. The transitions between the substages of MIS 5 (5.4/5.3, 5.3/5.2, 5.2/5.1) are marked by abrupt changes in benthic d18O at Site U1304. We sought to determine whether these changes were due to temperature or seawater d18O. We found that the entire signal at 5.4/5.3 and about half at the 5.3/5.2 and 5.2/5.1 transitions is due to seawater d18O change. This evidence suggests that both accumulation and ablation of ice occurred very rapidly between the substages of MIS 5. Furthermore, estimated d18Ow during MIS 5.2 and 5.4 suggests substantial ice existed during these stadials when sea level may have been as much as 40-60 m below present. Our results are significant because they contradict previous studies that have assumed that ice buildups slowly and decays more rapidly. Rapid and abrupt ice growth and decay involving strong positive feedbacks appear to have been possible under the climate conditions of MIS 5.



Because of the scarcity of benthic foraminifera during MIS 3, we measured Mg/Ca and d18O on N. pachyderma to reconstruct changes in surface temperature and d18Ow. Between 35 and 65 ka, the Mg/Ca-derived temperature record resembles the benthic d18O signal and follows Antarctic, and not Greenland, temperature. Pisias et al. (2010) demonstrated two leading modes of low-frequency global climate variability during MIS 3: a Greenland or "northern" mode and an Antarctic or "southern" mode. Documentation of the Antarctic climate mode in the subpolar North Atlantic is highly significant for understanding interhemispheric coupling of climate change in the polar regions.
Exploitation Route The finding of abrupt changes in ice volume during substages of Marine Isotope Stage 5 challenges our understanding of rates of ice sheet growth.
Sectors Environment