South Atlantic deep-sea temperature evolution across the Pliocene-Pleistocene from clumped isotope thermometry

The deep ocean comprises a large and relatively stable heat reservoir in the climate system. Hence deep sea temperature (DST) records constitute an important pillar of knowledge in the study of past climate states. Furthermore, a promising link exists between DSTs and global mean surface temperatures (GMSTs) [1]. DST reconstructions are nevertheless hampered by idealistic assumptions surrounding the interpretation of widely-used proxies like the ratio of oxygen isotopes on benthic foraminifera (δ¹⁸O_b). One such uncertain variable key in the analysis of δ¹⁸O_b is the isotopic composition of paleo-seawater (δ¹⁸O_sw). Deconvolving this signal into temperature and ice-volume components across the last 5 Ma is non-trivial. Attempts to resolve it have often relied on independent temperature constraints from foraminiferal Mg/Ca ratios that also involve added uncertainties regarding variations in ocean chemistry through time, as well as the need for species-specific calibrations.

Carbonate clumped isotope thermometry (Δ47) can aid in overcoming these limitations since it is based on thermodynamic principles that govern the abundance of ¹³C-¹⁸O bonds within the crystal lattice, therefore granting it independence from estimations surrounding the composition of the precipitating fluid [2]. Recent clumped isotope DST records of the Eocene [3] and the Miocene [4] reveal significantly warmer ocean temperatures than traditionally accepted [5], suggesting a re-evaluation of the interpretation of δ¹⁸O_b records in the geological past.

Here we present Δ47-based DST constraints of the last 5 Ma obtained from benthic foraminifera of ODP Site 1264 in the South Atlantic Ocean. Reconstructions covering the Pliocene-Pleistocene transition shed light on the global climatic change that followed the mid-Pliocene warm period and culminated in full glaciation of the Northern Hemisphere. In particular, our Pliocene DST data lies in closer agreement with recent GMSTs estimations [6] than δ¹⁸O_b-based DST would imply.

 

[1] Valdes, P. J., et al. (2021). Climate of the Past, 17(4), 1483-1506.

[2] Eiler, J.M. (2007), Earth Planet. Sci. Lett. 262, 309-327.

[3] Agterhuis, T., et al. (2022), Commun Earth Environ 3, 39

[4] Modestou, S. E., et al. (2020) Paleoceanography and Paleoclimatology 35, e2020PA003927.

[5] Westerhold, T., et al. (2020), Science, 369, 1383–1387

[6] McClymont, E. L., et al. (2020) Climate of the Past, 16(4), 1599-1615.