Using the boron isotope-pH proxy to investigate CO₂ driven retreats of the Antarctic Ice Sheet in the geological past

Abstract

Understanding the relationship between atmospheric CO₂ levels and ice sheet stability is of great importance given the anthropogenic emissions of CO₂ over the past 250 yrs. Since the inception of a permanent Antarctic ice sheet at the Eocene-Oligocene transition, fluctuations in global temperature and ice volume are reconstructed to have taken place in the absence of large changes in atmospheric CO₂. However, the output from ice sheet modelling experiments suggest that once large ice sheets have grown on East Antarctica they are inherently stable as a result of a powerful hysteresis effect and consequently higher levels of CO₂ are needed in order to initiate deglaciation. To further investigate this paradox, boron isotope (δ¹¹B) records have here been used to reconstruct CO₂ across two intervals of the Miocene- the middle Miocene climatic optimum (MCO) and the Mi-1 glaciation- both of which coincide with major deglaciation events. In order to calculate CO₂ from δ¹¹B a number of additional parameters are needed. Here, changes in the boron isotope composition of seawater (δ¹¹B_sw) and the life habit of the planktic Foraminifera Globigerina bulloides (G. bulloides) have been investigated through time.

The most useful species of foraminifera for generating δ¹¹B records across the Mi-1 glaciation is G. bulloides, however, its life habit may have been different from the modern. The life habit of G. bulloides is investigated using carbon and oxygen isotope (δ¹³C and δ¹⁸O) data of different size fractions to show that, unlike the modern species, Oligocene-Miocene G. bulloides had symbionts. A compilation of published δ¹³C data for the Neogene suggests that G. bulloides acquired its modern non-symbiotic lifestyle in the late Miocene. Another essential constraint is the boron isotope composition of seawater. Changes in δ¹¹B_sw through time are investigated here using the δ¹¹B and δ¹³C compositions of planktic and benthic foraminifera. This approach suggests that δ¹¹B_sw was ~ 38‰ at the Oligocene-Miocene boundary before decreasing to ~37 ‰ during the middle Miocene. δ¹¹B_sw then increased by ~ 2.5 ‰ to modern day values (39.61‰). A comparison of this record to the isotopic composition of other major and minor ions in the ocean suggests that variations in riverine input may have been one of the key drivers of the change in δ¹¹B_sw between 14 Ma and the present.

Armed with this improved understanding of the proxy a new δ¹¹B record from across the Mi-1 glaciation suggests that CO₂ is low (~ 250 ppm) and invariable prior to, and across the Mi-1 glaciation event and, increases to ~ 400 ppm during the deglaciation. A new δ¹¹B record from between 15.5-17 Ma is also presented that shows CO₂ is elevated to maximum levels of ~ 500 ppm between 15.5 and 16.5 Ma but that it is also extremely variable – regularly oscillating between ~300 ppm and ~500 ppm on orbital timescales. When both of these new records are compared to ice volume and climate records it appears that a dynamic Antarctic ice sheet existed even with a relatively modest CO₂ forcing across the Miocene. Assuming the Miocene can be used as an analogue for the future, this study suggests that the Antarctic ice sheet can be surprisingly dynamic even at relatively low CO₂ (<500 ppm).

More information

Identifiers

ORCID for P.A. Wilson: orcid.org/0000-0001-6425-8906

Catalogue record

Export record