Revisiting the Mid-Pleistocene Transition

Online seminar by Parker Liautaud from Harvard University

14 April 2021
KST 10:00 – 11:00

Join us online:

Present descriptions of Quaternary climate assert that after major Northern Hemisphere ice sheets first appeared three million years ago, a steady rhythm of obliquity-driven, 40-ky glacial cycles dominated global climate for two million years before giving way, in an abrupt middle-Pleistocene transition, to large-amplitude ~100 ky ice ages whose origins remain mysterious. In this talk, I will discuss evidence that early- and late-Pleistocene glacial cycles are more qualitatively similar than presently believed. First, it has been argued that climatic precession accounts for negligible variance in early-Pleistocene 𝛿18O despite having a large influence on summer insolation and being clearly present in late-Pleistocene records. New analyses of benthic 𝛿18O records indicates a larger precession contribution than previously recognized in the early Pleistocene, in qualitative agreement with the late Pleistocene (Liautaud et al. (2020), EPSL). Second, it is thought that a singular transition to ~100 ky cycles occurred at ~1 Ma, but previous studies have identified intermittent earlier intervals of ~100 ky variability (e.g. Lisiecki (2010), Nat. Geosci.). We show that statistically-significant ~100 ky amplitude in benthic 𝛿18O consistently arises under large-amplitude orbital forcing. This suggests ~100 ky cycles are externally-driven and that the Mid-Pleistocene Transition is just one of multiple transitions into ~100 ky cycles. Third, CO2 estimates from foraminiferal 𝛿11B indicate that sea-level became more sensitive to CO2 forcing after ~1 Ma, engendering proposals of a shift in ice-sheet dynamics (Chalk et al. (2017) PNAS). We use a hierarchical Bayesian method to infer the relationship among orbital variations, sea level, and atmospheric CO2 over the late Pleistocene, and use the inferences to predict early-Pleistocene CO2 from a sea-level reconstruction. Temporal shifts in sea-level sensitivity are reproduced on the basis of ice-albedo feedbacks and ice-sheet geometry without a change over time in model parameters or equations. Our findings to date suggest that early Pleistocene glacial variability is qualitatively consistent with that of the late Pleistocene, but with the response to orbital forcing gradually amplifying over the past 3 Ma.