| dc.description.abstract | During the glacial cycles of the Pleistocene (2.6 Ma–present), large ice sheets expanded and retreated on timescales of 40,000 to 100,000 years, driving sea level change and affecting the worldwide climate. This history is recorded in the oxygen isotope ratio (δ¹⁸O) of benthic foraminifera, a proxy commonly used for paleoclimate reconstruction. Benthic δ¹⁸O reflects a combined signal of deepwater temperature and global ice volume. Accurate interpretation of δ¹⁸O records requires a deconvolution of these two influences. This double thesis explores two approaches to benthic δ¹⁸O deconvolution: one based on clumped isotope thermometry, and one using ice sheet modelling.
In the clumped isotope study, deepwater temperatures were reconstructed at ODP Site 1267 in the South Atlantic. Measurements from benthic species indicate that during glacials MIS 2, 6 and 10, the average bottom water temperature was 4.4 °C (3.0-5.8 °C), two degrees warmer than the present day temperature. This contrasts with other reconstructions, which typically find that glacial deepwater was colder than today. Combined with measured carbonate δ¹⁸O, these temperatures imply a 1.8‰ increase in the δ¹⁸O composition of seawater, nearly twice the global change. Together with previous reconstructions form ODP Site 1264, these findings suggest that the glacial South Atlantic deepwater consisted of a relatively cold upper layer and a remarkably warm and saline abyssal layer. Potentially, the high abyssal temperature can be explained by larger retainment of geothermal heat in the strongly stratified glacial ocean. To validate these results, more measurements form nearby South Atlantic and Southern Ocean sites are essential. If confirmed, these findings showcase major regional variability in bottom water conditions, and caution against interpreting single site records as globally representative.
The modelling chapter focuses on the calculation of δ¹⁸O composition in ice sheets, aimed to improve the estimates of seawater δ¹⁸O. In benthic δ¹⁸O deconvolution attempts, δ¹⁸Osw is commonly assumed to be driven solely by ice volume changes, although composition of δ¹⁸O in ice sheets changes over glacial cycles. To adress this, a new Lagrangian framework was developed to simulate isotope transport and calculate the (changing) δ¹⁸O composition in ice sheets. Application to the present day Greenland ice sheet shows that average englacial δ¹⁸O lags surface forcing with a characteristic response time of 11 kyr. When extended to a simulation of Greenland over the last glacial cycle, the results reveal that that englacial δ¹⁸O varies by several per mille as a delayed response to changes in δ¹⁸O precipitation. In Greenland’s contribution to benthic δ¹⁸O, a a distinct compositional signal emerged, clearly seperable from the ice volume component. Preliminary estimates for other Pleistocene ice sheets suggest that compositional effects could account for ~10% of the δ¹⁸Osw signal. These results imply that ice sheet δ¹⁸O composition change should not be overlooked in benthic δ¹⁸O deconvolution, and call for a more thourough quantification of their magnitude. | |
