On the fidelity of oxygen isotope-based sea level reconstructions: A case study of Marine Isotope Stage 3

Although foraminiferal oxygen isotope records (δ¹⁸O_F) have long been used to develop time-continuous estimates of past global ice volume, growing evidence indicates that the ice volume and temperature contributions to δ¹⁸O_F diverge considerably over the last glacial cycle. Across Marine Isotope Stage 3 (MIS 3; 57 to 29 ka), estimates of peak global sea level based on δ¹⁸O_F are ∼30 m lower than reconstructions based on recent glacial geology, geologic sea-level records, and glacial isostatic adjustment (GIA) modeling approaches. Here we address this mismatch by quantifying uncertainties in converting δ¹⁸O_F to global sea level during MIS 3. We review the geochemical tools to assess the contributions of ocean temperature, diagenesis, and the isotopic composition of ice sheets to δ¹⁸O_F records across MIS 3. We show that recent advances in our understanding of these and other factors that contribute uncertainty (e.g., global sea level during the Last Glacial Maximum) help to explain the MIS 3 gap between global sea level estimates based on δ¹⁸O_F and GIA analyses of geologic sea-level records. We find that the accuracy and precision of MIS 3 ocean temperature reconstructions is likely a key contributor, and various MIS 3 temperature reconstructions can reduce the “gap” by 30%–70% (∼10–20 m). Future improvements in the precision of temperature reconstructions will better constrain the ice volume component of δ¹⁸O_F records. Ultimately, assessing the fidelity of δ¹⁸O_F as a proxy for global ice volume is key to disentangling the feedbacks between climate and ice sheets over the Pleistocene.