A biogeochemist weighs up the climatic influence of CO2.

Carbon dioxide constitutes a vanishingly small fraction of our atmosphere, but punches well above its weight in terms of greenhouse warming. So just how potent is it?

The geological record provides clues because, over time, Earth has oscillated between greenhouse and icehouse climates. But reconstructing coincident atmospheric CO2 concentrations is notoriously difficult. Modelling and proxy calculations are starting to converge on a single picture of atmospheric CO2 during greenhouse episodes, except for one fly in the ointment: estimates derived from the ratio of carbon isotopes in soil-precipitated carbonates are always higher than those derived from any other source.

Daniel Breecker, now at the University of Texas at Austin, and two co-workers confirm that these estimates are too large (D. O.Breecker et al. Proc. Natl Acad. Sci. USA 107, 576–580; 2010). The numbers relied on measurements of CO2 in soil pores, thought to reflect the growing-season mean. But the creation of soil carbonates is more likely during the driest and warmest parts of the growing season, when the release of CO2 from plant respiration is at a minimum.

To understand the implications of this, think of a gin and tonic. If you have less gin than you thought, you must lower the amount of tonic to get the same tasty ratio. Likewise, because there is less carbon than we thought from plant respiration, we lower our estimate of atmospheric carbon to accord with the observed ratio. The newly calculated values align beautifully with the emerging consensus. A mere 1,000 parts per million by volume (just two and a half times current atmospheric levels and similar to those predicted for AD 2100) is sufficient to induce the hottest greenhouse conditions — such as those of the Mesozoic period 251 million to 65 million years ago. CO2 truly is a heavyweight greenhouse gas.

Discuss this paper at http://blogs.nature.com/nature/journalclub