Reconciling the Apparent Discrepancy Between Cenozoic Deep-Sea Temperatures From Proxies and From Benthic Oxygen Isotope Deconvolution

Understanding past deep-sea temperature and sea-water oxygen isotope ratios is fundamental to environmental Earth science. For example, it provides crucial insight into past ice-volume variations, an important climate system feedback. Moreover, deep-sea temperature is important to deep-sea ecology and biogeochemical cycling. Here we compare deep-sea temperature and sea-water oxygen isotope ratios from model-based deconvolution of benthic foraminiferal carbonate δ¹⁸O with clumped isotope-based deep-sea temperature data in 1000-year timesteps over the Cenozoic. To assess wider implications of the observed differences, we quantitatively evaluate a range of potential explanatory hypotheses—such as diagenetic overprints, carbonate ion effects, ice-sheet morphology changes, and warm saline deep-water admixture—but find that, individually, none can explain the observed differences satisfactorily. We then evaluate the implications of possible combined effects and recent advances in clumped isotope temperature calibration. We find that combined consideration of a recently proposed cool clumped isotope calibration and possible carbonate ion or pH influences can provide results that approximate deep-sea temperature reconstructions based on conventional δ¹⁸O_c deconvolution. The match can be further improved if modest warm saline deep-water contributions are considered during past warm periods. This contrasts with ice-volume and ice-sheet morphology changes, which appear unrealistic or insignificant, respectively. Our quantitative comparison offers a means toward formulation of a comprehensive and internally consistent understanding of Cenozoic variability in sea level (ice volume), GIA-corrected ice-sheet heights and mean ice δ¹⁸O, sea-water δ¹⁸O, sea-water δ¹⁸O_w, deep-sea temperature, and deep-sea [CO₃²⁻] variations.