Revisiting the thermal bipolar seesaw model from the oceanic perspective by considering novel organic proxies from the Southern Indian Ocean

Paleotemperature records from Antarctica and the Southern Ocean show a millennial variability in addition to the glacial-interglacial variability over the last glacial cycle. This millennial variability is the Southern Hemisphere counterpart of the Northern Hemisphere abrupt variability via the thermal bipolar seesaw, a concept describing the meridional heat transport leading to opposite temperature changes between both hemispheres. However, the thermal bipolar seesaw is typically studied from the atmospheric perspective using temperature records from ice cores in Greenland and Antarctica. While the thermal bipolar seesaw model was recently revisited using oceanic temperature records from the Iberian Margin (Davtian and Bard, 2023 PNAS https://doi.org/10.1073/pnas.2209558120), oceanic temperature records from the Southern Ocean remain to be considered.

We generated oceanic temperature records over the last 160 kyr using novel organic proxies (e.g., RI-OH′ and TEX₈₆^OH) from three deep-sea sediment cores located in the Southern Indian Ocean (cores MD11-3353, MD11-3357, and MD12-3394). We assessed the paleothermometric potential of the novel organic proxies and accuracy of preliminary temperature reconstructions by comparing them with a more established organic proxy (TEX₈₆^L) at the same Southern Indian Ocean sites, as well as with Antarctic temperatures and modern oceanic temperatures. All novel organic proxies, except %OH, show a glacial-interglacial variability. TEX₈₆^OH best shows the Antarctic-like millennial variability, notably at the MD12-3394 site with the highest-resolution temperature records. At the MD11-3357 site north of the Subantarctic Front, global calibrations yield more realistic temperature reconstructions with TEX₈₆^OH (from 5 to 12 °C) than with RI-OH′ (from 0 to 7 °C), possibly due to a stronger water depth effect on RI-OH′ than on TEX₈₆^OH coupled to bottom waters colder by roughly 9–10 °C than surficial waters. At the MD11-3353 and MD12-3394 sites south of the Subantarctic Front, regional calibrations of RI-OH′ and TEX₈₆^OH yield more consistent and more accurate temperature reconstructions (from –2 to 5 °C for RI-OH′ and from –1 to 5 °C for TEX₈₆^OH) than do global calibrations of these proxies (from –2 to 4 °C for RI-OH′ and from 0 to 8 °C for TEX₈₆^OH), as RI-OH′ and TEX₈₆^OH show reduced thermal sensitivity below 5 °C. At the MD12-3394 site, millennial warming amplitudes based on TEX₈₆^OH reach 1.5–2.0 °C for the most pronounced Antarctic-like warming events.

We then revisited the classical thermal bipolar seesaw model by comparing reconstructed Southern Ocean temperatures with simulated Southern Hemisphere temperatures. We selected the TEX₈₆^OH record from core MD12-3394 and a Southern Hemisphere temperature record simulated with two independent organic proxies (RI-OH′ and U^K′₃₇) from the southern Iberian Margin (core MD95-2042; Davtian and Bard, 2023). Despite the limited amplitude of Southern Ocean millennial warming events, our data-model comparison shows that to revisit the thermal bipolar seesaw model from the oceanic perspective is feasible by considering temperature records from the Southern Ocean and Iberian Margin. However, our study also demonstrates the need for high-resolution (200 to 500 years) oceanic temperature records from multiple sectors and sites in the Southern Ocean, including with the tested novel organic proxies.