Carbon cycle feedbacks and the initiation of Antarctic glaciation in the earliest Oligocene

https://doi.org/10.1016/j.gloplacha.2005.01.001Get rights and content

Abstract

The initiation of Antarctic glaciation in the early Oligocene (∼34 Ma) is represented by a distinct positive anomaly in the marine δ18O record designated Oi-1 and accompanied by positive excursions in the mean δ13C of oceanic dissolved inorganic carbon and biogenic sediment accumulation rates. Within 400 ky of the onset of Oi-1, the climate system settled into a more moderate but stable “glacial” state.

Here, through modeling, we investigate two of the principal biogeochemical processes involved in this response: silicate weathering and marine organic carbon cycling. We initiate the event with a rapid drawdown in atmospheric CO2 resulting from increased weatherability of the continents associated with Himalayan orogeny. This perturbation triggers the overshoot and adjustment of the δ18O record because of feedback among ice-sheet coverage, silicate weathering rates, and atmospheric CO2. The system is a damped oscillator, the strength of which depends on the sensitivity of chemical weathering rates to climate change and climate to changes in atmospheric CO2. Increased oceanic mixing associated with initial transition into a glacial world accelerates the rates of biological productivity and carbon burial, lowering atmospheric CO2 and accelerating global cooling and ice-sheet growth, and generating a carbon isotope response that crudely approximates that observed. The Oi-1 overshoot appears to require a rapid (<1 million year) application of the forcing (e.g., tectonic drawdown of atmospheric CO2). Although further investigation and more sophisticated models ultimately may show that other triggers and feedbacks prevailed during Oi-1, the modeling presented here demonstrates that simple feedbacks in the climate system can explain the overshoot and adjustment response to early Oligocene climate forcing.

Introduction

The most prominent step in the long-term transition from greenhouse to glacial conditions during the Cenozoic occurred in the early Oligocene, roughly 34 Ma. The sudden appearance of a large continental ice sheet on Antarctica is well-documented in oxygen isotope and glaciomarine sediment records (Ehrmann and Mackensen, 1992, Zachos et al., 1992, Hambrey et al., 1991, Wise et al., 1991, Barrett et al., 1989). Other significant changes include several degrees of high-latitude cooling, intensification of atmospheric and oceanic circulation, development and intensification of the polar front, and increased aridity in continental interiors (Bestland et al., 1997, Barrera and Huber, 1991, Baldauf and Barron, 1990, Keller et al., 1987). Isotopic and other evidence indicate that this reorganization, particularly the ice-sheet expansion, was abrupt, with much of the ice expansion occurring in less than 50 ky (Diester-Haass and Zahn, 1996, Zachos et al., 1996).

The effects of this rapid and extreme climatic transition on global environments were ubiquitous. Most marine and terrestrial ecosystems were affected, including the high latitude oceans where cooling and more vigorous upwelling elevated the biological production of opal at the expense of carbonates (Salamy and Zachos, 1999, Wei and Wise, 1990), and mid-latitude continents where increased aridity accelerated the expansion of scrubland at the expense of dense forests (e.g., Gregory and Chase, 1992, Wolfe, 1980, Wolfe, 1994). The changes in climate and global ecosystems, in turn, contributed to a fundamental shift in the character of global sedimentation. On passive margins, carbonate-rich facies were largely replaced by siliciclastic (Miller et al., 1998), and in the deep-sea the calcite compensation depth (CCD) deepened by more than a kilometer (Peterson and Backman, 1990, Van Andel, 1975).

The most unusual and, perhaps, important aspect of the early Oligocene transition is the manner in which the climate system reorganized itself. Rather than migrating along a direct path from one equilibrium climate state to the next, the climate system experienced an overshoot, or transient, in which it appears to have drifted beyond equilibrium into an extreme mode before settling into a more moderate stable mode. This pattern is best defined in high-resolution deep-sea oxygen isotope records from the South Atlantic and Indian Ocean (Diester-Haass and Zahn, 1996, Zachos et al., 1996). Each record exhibits essentially the same pattern, a 1.2‰ increase in δ18O between 33.7 and 33.5 Ma (Fig. 1), relatively elevated values for ∼200 ky, the duration of chron 13N, and then a partial return to less 18O-enriched values by 33.0 Ma. The rise in δ18O appears to have followed a non-linear path, increasing slowly at first, and then accelerating rapidly over the last 50 to 100 ky. Each record peaks at a mean value of 2.5‰ that is similar to values recorded in the Pliocene, implying similar ice volume and temperature perturbations. For the next 400 ky (33.4 to 33.0 Ma) values remain relatively high, oscillating between 2.3‰ and 2.5‰, before settling to lower values for the remainder of the early Oligocene. Given the isotope data and glaciomarine sediment evidence, this brief positive isotope anomaly, which has been designated Oi-1, must reflect an unusually “deep” glaciation characterized by ice volume close to present day.

The onset of Oligocene glaciation is generally attributed to the regional cooling effects brought on by either the gradual widening of the oceanic passages surrounding Antarctica (Kennett and Shackleton, 1976), although the deep-circumpolar current probably was not established until the opening of the Drake Passage in the Miocene (Beu et al., 1997), and/or the reduction of greenhouse gas concentrations from peak levels in the early Eocene (DeConto and Pollard, 2003, Pearson and Palmer, 2000). In either scenario, a threshold involving temperature is exceeded initiating rapid accumulation of ice, and/or an abrupt change in ocean circulation. Such behavior is common in complex dynamical models which often exhibit multiple equilibrium states for the same sets of boundary conditions (Crowley and North, 1988). The pattern of the Oligocene transition, however, is complicated by the step into an extreme but temporary state, Oi-1. The most likely candidates for generating strong but temporary feedbacks on climate change are large-scale biogeochemical processes, specifically those involved in controlling the level of CO2 in the atmosphere (Broecker, 2000).

The purpose of this study is to identify and evaluate potential large-scale biogeochemical feedbacks on the Oligocene climate–ocean system as it rapidly shifted from a non-glacial to glacial regime. We have elected to focus on two feedbacks that can influence the concentration of carbon dioxide in the atmosphere. The first is a negative feedback: the chemical weathering of continental silicates, which is a principal sink for atmospheric CO2 on geologic time scales. In theory, ice-sheet growth and cooling inhibit silicate weathering thereby reducing the rate of CO2 drawdown (e.g., Broecker and Sanyal, 1998, Gibbs and Kump, 1994, Kump and Alley, 1994) ultimately allowing CO2 to rise. The second is a positive feedback: the rate of oceanic overturn, which controls primary production and carbon burial. High latitude cooling and increased wind stress would tend to accelerate rates of upwelling and productivity, and increase the burial rate of organic carbon relative to carbonate carbon, thereby lowering CO2 (e.g., Salamy and Zachos, 1999, Zachos et al., 1996). This mechanism is often invoked to explain glacial to interglacial variations in CO2, although the linkages between climate and biological productivity are far from simple (e.g., Broecker, 2000, Sigman and Boyle, 2000). Benthic foraminiferal δ13C and opal accumulation rates show brief positive excursions coeval with Oi-1 (Salamy and Zachos, 1999, Diester-Haass and Zahn, 1996, Zachos et al., 1996). The downward flux of organic carbon increases as well in several regions (Latimer and Filippelli, 2002, Diester-Haass and Zahn, 2001). Also, Oi-1 marks the initiation of an extreme global deepening of the CCD (Zachos et al., 1996, Peterson and Backman, 1990).

In order to quantify the potential effects from each on ocean δ13C as well as CO2, we have devised a simple model to examine how the global carbon cycle might respond to variations in (1) silicate weathering rates as influenced by ice-sheet growth and decay, and (2) organic and inorganic carbon burial as influenced by abrupt changes in the rate of oceanic overturn (Kump and Arthur, 1999, Kump, 1991). We keep the formulation of feedbacks simple, as the functional dependencies of these processes are still not well known. Our goal is not to produce definitive climate simulations supported by exhaustive sensitivity analyses. Rather, we use the toy climate model described below to demonstrate that first-order features of the proxy isotope records can be generated by internal dynamics of the global carbon cycle. Proper modeling of the problem awaits the development of fully coupled Earth systems models of intermediate complexity with ocean–atmosphere–ice sheet–terrestrial weathering components and capable of continuous simulations over millions of years.

Section snippets

Model description

We used Stella® (http://www.hps.com) to model the Eocene–Oligocene climate transition with a fixed time step of 1 ky and the forward Euler method for solution of the set of ordinary differential equations that define the system. The model results are insensitive to reductions in time step size (as must be the case for valid model results). The model has four modules: ice-sheet mass balance, oceanic carbon balance, oceanic phosphate balance, and oceanic carbon isotope balance. Each of these

Simulation results

Although we ran many suites of simulations to explore the behavior of the model, only two are described here (Stella® users can obtain the model from the authors upon request). In the first suite of simulations (“silicate-only”), we removed the organic carbon, carbon isotope, and phosphate modules so that we could isolate the potential effect of the ice-sheet and silicate weathering feedback on the nature of the Eocene–Oligocene climate transition. We also tested the sensitivity of the model to

Discussion

Our model simulations of the early Oligocene glaciation indicate that biogeochemical feedbacks have the potential to influence climate in a way that is consistent with observation. Of the two major feedbacks tested, the effects of the silicate weathering feedback were most pronounced in scale and duration. As a negative feedback with a long time constant, the weathering feedback creates an oscillating response in climate as the system attempts to restore a balance between CO2 sources and sinks.

Summary

The potential for clear expression of biogeochemical feedbacks is greatest when the global climate system is undergoing a major reorganization from one mode to another. The early Oligocene cooling and onset of Antarctic glaciation represents one such reorganization for which we have examined two of the larger potential sources of climate-sensitive feedbacks, silicate weathering and the ocean carbon cycle. Through modeling of the most basic responses, we have found that both processes are

References (59)

  • W. Wei et al.

    Biogeographic gradients of middle Eocene–Oligocene calcareous nannoplankton in the South Atlantic Ocean

    Palaeogeography, Palaeoclimatology, Palaeoecology

    (1990)
  • J.A. Wolfe

    Tertiary climates and floristic relationships at high-latitudes in the northern hemisphere

    Palaeogeography, Palaeoclimatology, Palaeoecology

    (1980)
  • J.A. Wolfe

    Tertiary climatic changes at middle latitudes of western North America

    Palaeogeography, Palaeoclimatology, Palaeoecology

    (1994)
  • D. Archer et al.

    Effect of deep-sea sedimentary calcite preservation on atmospheric CO2 concentration

    Nature

    (1994)
  • Baldauf, J.G., Barron, J.A., 1990. Evolution of biosiliceous sedimentation patterns-Eocene through Quarternary:...
  • E. Barrera et al.

    Paleogene and early Neogene oceanography of the southern Indian Ocean; Leg 119 foraminifer stable isotope results

  • P.J. Barrett et al.

    Synthesis

  • W.A. Berggren et al.

    A revised Cenozoic geochronology and chronostratigraphy

  • R.A. Berner

    3geocarb Ii—a revised model of atmospheric Co2 over Phanerozoic time

    American Journal of Science

    (1994)
  • R.A. Berner et al.

    GEOCARB III: a revised model of atmospheric CO2 over Phanerozoic time

    American Journal of Science

    (2001)
  • R.A. Berner et al.

    The carbonate–silicate geochemical cycle and its effect on atmospheric carbon dioxide over the past 100 million years

    American Journal of Science

    (1983)
  • E.A. Bestland et al.

    Stepwise climate change recorded in Eocene–Oligocene paleosol sequences from Central Oregon

    Journal of Geology

    (1997)
  • W.S. Broecker et al.

    Tracers in the sea

    (1982)
  • W.S. Broecker et al.

    Does atmospheric CO2 police the rate of chemical weathering

    Global Biogeochemical Cycles

    (1998)
  • T.J. Crowley et al.

    Abrupt climate change and extinction events in earth history

    Science

    (1988)
  • R.M. DeConto et al.

    Rapid Cenozoic glaciation of Antarctica induced by declining atmospheric CO2

    Nature

    (2003)
  • L. Diester-Haass et al.

    Eocene–Oligocene transition in the Southern Ocean; history of water mass circulation and biological productivity

    Geology

    (1996)
  • B.P. Flower

    Middle Miocene deep-water paleoceanography in the southwest Pacific—relations with east Antarctic ice-sheet development

    Paleoceanography

    (1995)
  • P.N. Froelich et al.

    The marine phosphorus cycle

    American Journal of Science

    (1982)
  • Cited by (134)

    • Paleozoic carbon cycle dynamics: Insights from stable carbon isotopes in marine carbonates and C<inf>3</inf> land plants

      2021, Earth-Science Reviews
      Citation Excerpt :

      Carbon cycle mass balance models show that the long-term variations of δ13Ccarb are due to the carbon exchange between inorganic and organic reservoirs (Veizer et al., 1980) through adjustment in the proportion of inorganic and organic carbon burial (Kump, 1991; Kump and Arthur, 1999). It has also been suggested that the geologic carbon cycling may have been inherently oscillatory (Bachan et al., 2017), and can yield secular changes of δ13Ccarb without external forcings or minimally affected by them (Wallmann, 2014; Zachos and Kump, 2005). To summarize, the long-term δ13Ccarb shifts can be caused by both internal oscillation and external forcings through alterations of the intensity and timing of the forcings, which are largely attributed to changes in ocean redox conditions and long-term sea level changes caused by plate tectonics (Hannisdal and Peters, 2011).

    View all citing articles on Scopus
    View full text