An analysis of 12,000-year-old Antarctic ice revises our understanding of natural methane emissions to the atmosphere, and constrains estimates of the sensitivity of natural methane sources to abrupt climate-warming events. See Letter p.443
Methane is a potent greenhouse gas, but the interplay of its sources and sinks is poorly understood. After wetlands, geological methane sources are thought to be the main natural source to the atmosphere today1. On page 443, Petrenko et al.2 present measurements of methane trapped in air bubbles in ancient ice from Antarctica, and report that the contribution of geological emissions to that methane was much lower than expected. They also propose that geological sources made only a small contribution to the rapid increase in atmospheric methane levels that occurred during the most-recent abrupt climate-warming event. These findings call for a re-evaluation of the present-day methane budget, and imply that geological methane sources in the pre-industrial climate system were impervious to abrupt climate-warming events.
After carbon dioxide, methane is the largest contributor to greenhouse-gas effects associated with human activities. Its abundance is currently rising3, continuing a longer-term trajectory that has seen its concentration increase from 700 parts per billion by volume (p.p.b.v.) before industrialization to more than 1,800 p.p.b.v. today. This increase is responsible for about 15% of the total anthropogenic greenhouse-gas effect3.
Atmospheric methane concentrations also varied greatly in the past, before human activities had an effect (Fig. 1). Records recovered from gas bubbles trapped in deep polar ice cores show a positive correlation between methane levels and large-scale climate oscillations4. They also reveal5 a series of abrupt rises in atmospheric methane of up to 200 p.p.b.v., which coincided with Dansgaard–Oeschger events — abrupt warmings of the Northern Hemisphere of 5–16 °C (as reconstructed for central Greenland6) that occurred in a matter of decades.
But pinpointing the drivers of variations in atmospheric methane levels has proved difficult, because there are many natural and anthropogenic sources, and because the suite of chemical processes that control methane's residence time in the atmosphere are difficult to measure directly. One way to get a handle on methane cycling in the atmosphere is to measure the ratios of naturally occurring isotopes in the gas. Carbon-14 is particularly useful, because fossil-derived sources of methane, be they natural (geological) or anthropogenic (such as leaks from the extraction of natural gas), contain almost no 14C, whereas methane from other sources, such as that deriving from wetlands, have a 14C signature that is much closer to that of atmospheric CO2.
Petrenko et al. have made such measurements on methane preserved in ancient ice from Antarctica, and present a 14C record that covers the interval from 12,000 to 11,200 years before the present. This interval coincides with the beginning of the current interglacial period (the Holocene) and includes the last Dansgaard–Oeschger event, known as the Younger Dryas–Preboreal transition (Fig. 1).
The authors show that the 14C methane signature is relatively constant through this abrupt methane rise. They use a mass-balance calculation to show that this could not have happened if the rise was caused by emissions from geological sources, including hydrates — solid compounds in which methane is trapped in a crystal form of water, substantial quantities of which are deposited under sediments on ocean floors. This provides the best evidence so far that methane was not released from hydrates during this Dansgaard–Oeschger event (and, by implication, probably not during earlier such events), and adds to a growing body of evidence that suggests hydrates did not respond appreciably to climate variations during the most recent glacial period7.
The climate 12,000 years ago was globally about 2 °C cooler than today8. This, in turn, is as much as 5 °C cooler than that projected9 for the year 2100. Scenarios in which little is done to reduce greenhouse-gas emissions will therefore take the climate system into a state that is much warmer than during any of the ice-age Dansgaard–Oeschger events8,9. So, although the transition studied by Petrenko et al. was the most-recent abrupt polar warming event in the geological record, it may not provide a particularly good analogue of future warming. Further modelling and isotopic constraints for today's system, and for other time periods (perhaps including the interglacial that preceded the most-recent ice age), are required to better understand the probability of future geological methane change10.
The 14C ratio reported by Petrenko et al. is also useful in itself, because it allows the authors to provide the first quantitative constraint on 14C-free methane emissions (that is, those from geological sources) in the pre-industrial climate system — something that eluded workers from the same group in previous work11, but which is possible today because estimates of 14C production in ice can now be made more accurately. They estimate an upper limit of 15.4 teragrams of methane per year (Tg CH4 yr−1; 1 Tg is 1012 grams), which is much lower than the approximately 52 Tg CH4 yr−1 estimated for the present day1,2. Natural geological methane emissions are expected to have been higher during the past than in modern times12, and so the authors' estimate of emissions 12,000 years ago can be taken as an upper limit for today's climate system.
As the authors point out, if geological emissions today are indeed less than or equal to 15.4 Tg CH4 yr−1, rather than 52 Tg CH4 yr−1, then the difference of approximately 40 Tg CH4 yr−1 needs to be accommodated by revising our estimates of anthropogenic 14C-free emissions upward by about 25% — a substantial correction to our view of the contemporary methane cycle. Such a revision would imply that there is more scope to reduce human influence on climate than was thought, by reducing methane emissions associated with human activities.
Petrenko and colleagues' result comes at a crucial time for our understanding of atmospheric methane: unlike CO2 levels, methane's concentration is rising at a rate close to that of high-end projections13, and the total anthropogenic contribution has already been revised upward in more-recent estimates based on methane 13C isotopic data14. Further work is needed to understand whether these studies can be reconciled with each other, and with other methane constraints obtained for Earth's past and present.
The new study provides a compelling example of how studying the past helps us to better understand the present Earth system. Atmospheric methane levels seem capable of surprising us at every turn, from the rapid increases examined by Petrenko et al. to a period of puzzling stability observed at the start of this century15. If nothing else, we should heed warnings from the past if we are to understand the potential role of methane in future climate change.