The role of CO2 change in driving the ‘greenhouse-to-icehouse’ transition?

Recent research suggests that this change in Cenozoic climate state from the greenhouse of the early Cenozoic to the icehouse of the late Cenozoic was driven by a decline in the CO₂ content of the atmosphere. It is thought that this was caused by a very subtle imbalance between long-term carbon fluxes from the mantle and upper crust to the atmosphere: either a decline in CO₂-sources (volcanism/metamorphism), or an increase in CO₂-sinks such as silicate weathering.  Although the role of CO₂ decline in triggering the rapid transition between climate states at the Eocene/Oligocene (E/O) boundary (~34 Ma) has recently been confirmed [1],[2]  there are currently two crucial limitations to the hypothesis that the more long-term Cenozoic climate change was also CO₂ driven. These are:

1) Continuous records of pCO₂ only indicate a broad decline over the last 50 million years that is only loosely correlated with climate variations (Fig. 1).  For example, pCO₂ undergoes relatively rapid and large scale oscillations in the early Cenozoic, whilst climate shows a gradual cooling.  If these records are reliable, this implies that other factors such as ocean circulation may influence climate more than pCO₂ during this time interval[3].

2) There is also very little correlation between the records of silicate weathering and/or mantle outgassing (reflected by oceanic ⁸⁷Sr/⁸⁶Sr and mean spreading rate, respectively) with these pCO₂ records (Fig. 1).  Indeed, although weathering may have increased during the early Cenozoic in response to Himalayan uplift[4],[5] , spreading rates are, if anything, lower during the early part of the Cenozoic compared to the latter half[6], implying an increasing, not decreasing, flux of mantle CO₂ during the course of the Cenozoic (Fig. 1).

Thus, we neither understand the cause of pCO₂ change nor its influence on the Earth’s climate and biota, during even this most recent and best studied greenhouse-icehouse transition. These limitations strike at the heart of our understanding of what is controlling the long-term evolution of Earth’s climate and, in detail at least, challenge the widely accepted view that changes in the atmospheric CO₂ concentration are crucial in driving climate on these geological timescales. Reconstructing the climate system in the past is a difficult task and the methods used undergo continuous assessment and improvement.  Of particular relevance here is that the methods used to reconstruct pCO₂ beyond the reach of the ice cores has matured greatly in the last decade (see below).  Consequently, before the role of pCO₂ in Cenozoic climate change can be fully assessed, better and more complete pCO₂ records need to be generated with state-of-the-art techniques and understanding.

Figure 1. (a) TEX86 based SST estimates61, (b) δ18O of benthic foraminifera, (c) Oceanic 87Sr/86Sr, (d) averaged mid ocean ridge half-spreading rate, (e) continuous pCO2 estimates using δ13C of alkenones and δ11B of foraminifera. Also shown are relevant changes in ocean gateways and the cyrosphere.

[1] Pearson, P.N., Foster, G.L., and Wade, B.S., 2009, Atmospheric carbon dioxide through the Eocene–Oligocene climate transition: Nature, v. 461, p. 1110-1113.