Would Earth's early ocean have been a frozen wasteland had levels of atmospheric methane not been sky high? Maybe. Or maybe, according to a new view of an old idea, the main warming agent was carbon dioxide.
For the first three-and-a-half billion years of Earth's history, the Sun burned only about 70–90% as brightly as it does today. A famous paradox of ancient climate demands that we reconcile the persistence of a life-sustaining liquid ocean with these less-warming rays from a faint young Sun. The solution lies with greenhouse gases1, which trap heat near the planet's surface. In recent years, methane has become the favoured greenhouse agent, overcoming the earlier popularity of carbon dioxide in climate models spanning most of the Precambrian eon — that is, during all of Earth's history up to about three-quarters of a billion years ago, after which life emerged on a large scale.
On page 395 of this issue, Ohmoto and his colleagues2 dispute the arguments against CO2. Their challenge has additional significance because the assertion of inadequate CO2 is the most frequently cited evidence for the existence of high levels of atmospheric methane. And given the incompatibility of methane and oxygen, this debate speaks more broadly to the oxygenation history of the atmosphere and its link to the evolution of early life.
Today, most of us worry about the 30% rise in levels of CO2 since the Industrial Revolution and how that may drive global warming. But a generation of models of Earth's atmosphere invoked CO2 concentrations as high as a thousand times greater than those of today to explain the unfrozen early ocean1. About a decade ago the idea of CO2 as the dominant greenhouse agent was dealt a blow when Rye and his co-workers3 used the absence of siderite, an iron-rich carbonate mineral, in ancient soils to set a maximum for CO2 levels in the atmosphere 2.2 billion to 2.75 billion years ago (Fig. 1). This maximum, although still many times the concentration observed today, was well below that necessary to offset the faint young Sun.
Rye et al. were compelled to suggest that another greenhouse gas — methane — must have taken up the slack. But their story was based on only a handful of data and assumed, among other things, that original mineral constituents and chemical properties can be inferred from the rocky remnants of highly altered ancient soils. Also, chemical reactions in the natural world frequently deviate from the equilibrium behaviour assumed in Rye and colleagues' thermodynamic model.
Nonetheless, the absence of siderite in ancient soils has become the foundation of arguments for lower CO2 and thus for a methane-rich early atmosphere. Precious little direct evidence backs up that view, which is based mainly on the CO2 estimates, but a body of independent, circumstantial evidence lends compelling support. By most accounts, for example, the atmosphere was deficient in oxygen before about 2.3 billion years ago4. Anoxia favours methane preservation, as well as methane production by methanogens — anaerobic, or oxygen-intolerant, microorganisms that first appeared early in the Precambrian.
Supplies of sulphate in the ocean also figure prominently in the production and longevity of methane. Sulphate is the second most abundant anion in sea water today. But because it is produced mostly by weathering on the continents in the presence of oxygen, the early ocean would have been sulphate poor5. Bacteria that reduce sulphate compete with methanogens for metabolizable organic compounds, and anaerobic decomposition of methane requires the presence of sulphate6. Conditions that were comparatively favourable to methane, including low concentrations of sulphate in sea water, may have persisted until about 0.75 billion years ago7,8, when oxygen concentrations in the atmosphere and ocean once again took a big step upward.
Ohmoto et al.2 have a different take on the presence or absence of iron-rich carbonate minerals in ancient soil horizons. By taking a broader thermodynamic view of these so-called palaeosols, they suggest that siderite is absent largely because oxygen levels were too high. Ohmoto et al. are quick to point out that even minuscule levels of oxygen could have been enough to disfavour siderite formation. As with the model of Rye et al., this conclusion requires assumptions about equilibrium chemical conditions, ambient temperatures, and gas partitioning between the atmosphere and subsurface soil environments — among other poorly known parameters. If the assumptions of Ohmoto et al. are correct, this additional constraint tells us that CO2 levels in the atmosphere could have been higher than those suggested by Rye and his colleagues.
Despite the absence of siderite in palaeosols, which Ohmoto et al. attribute to factors other than CO2, the mineral was common in marine settings before 1.8 billion years ago. Under these conditions, Ohmoto et al. argue, the controlling factor was CO2, implying that levels of this gas in the atmosphere must have been high — high enough to minimize the need to invoke methane as a major greenhouse gas.
As with the palaeosols, however, factors other than CO2 availability can control the formation of siderite in marine environments. For example, the high concentrations of dissolved iron that most workers imagine existed in the early, oxygen-deficient ocean9 would reduce the amounts of CO2 required for the mineral to form. Given such uncertainties, it is difficult to view Ohmoto and colleagues' estimates of high CO2 levels as the only interpretation of early distributions of siderite. Furthermore, although concentrations of CO2 higher than those proposed by Rye et al. may weaken the principal argument for high methane concentrations, they don't preclude both CO2 and methane from being much more plentiful than they are in today's atmosphere10,11.
A universal theme in studies of the early Earth is that big stories are told with little data and lots of speculation, and interpretations of Precambrian siderite are no exception. Nevertheless, Ohmoto et al. open our eyes to another way of looking at the presence and absence of this diagnostic mineral. In doing so, they show us that exploratory steps can lead in many directions, but with each step the ancient atmosphere becomes a little more transparent.
Kasting, J. F. et al. Science 259, 920–926 (1993).
Ohmoto, H., Watanabe, Y. & Kumazawa, K. Nature 429, 395–399 (2004).
Rye, R., Kuo, P. H. & Holland, H. D. Nature 378, 603–605 (1995).
Holland, H. D. Geochim. Cosmochim. Acta 66, 3811–3826 (2002).
Habicht, K. S., Gade, M., Thamdrup, B., Berg, P. & Canfield, D. E. Science 298, 2372–2374 (2002).
Boetius, A. et al. Nature 407, 623–626 (2000).
Pavlov, A. A., Hurtgen, M. T., Kasting, J. F. & Arthur, M. A. Geology 31, 87–90 (2003).
Kah, L. C., Lyons, T. W. & Chesley, J. T. Precamb. Res. 111, 203–234 (2001).
Canfield, D. E. Nature 396, 450–453 (1998).
Pavlov, A. A., Kasting, J. F., Brown, L. L., Rages, K. A. & Freedman, R. J. Geophys. Res. 105, 11981–11990 (2000).
Hessler, A. M., Lowe, D. R., Jones, R. L. & Bird, D. K. Nature 428, 736–738 (2004).