Ancient clues help quantify modern methane

Assessments of the current CH₄ budget fall into two broad categories. Top-down studies measure the concentration, distribution, and isotopic composition of the CH₄ already in the air and use models to figure out where it came from. Bottom-up studies measure emissions from known CH₄ sources, extrapolate them globally, and add them all up.

Both classes of studies are subject to large uncertainties, and even estimates of the same type can disagree substantially. Total global emissions are somewhere in the range of 500 to 700 teragrams per year, with bottom-up studies giving values some 20% higher than top-down ones.² The lion’s share of that amount is biogenic, mostly from wetlands and agriculture. Only a minority—perhaps 100 to 150 Tg/yr—comes from fossil sources. But the CH₄ budget includes not only sources but also sinks, and CH₄ is generally removed from the atmosphere at more than 90% of the rate at which it’s added. So even a small reduction in emissions could make a big difference.

The isotopic measurements in top-down studies often focus on carbon-13 and deuterium. Although those stable isotopes aren’t created or destroyed on Earth in significant amounts, they’re fractionated by various physical, chemical, and biological processes, so different sources of CH₄ have different isotopic makeup. But stable isotope signatures are imperfectly known, overlap for different sources, and can even change over time for a single source.

Radioactive ¹⁴C is different. It’s created by cosmic rays in the stratosphere and upper troposphere, and it decays with a half-life of 5700 years. The atmosphere therefore has a steady-state concentration of ¹⁴C, as do living organisms, which are constantly taking in and expelling carbon. When an organism dies, the carbon exchange stops, and the ¹⁴C content declines.

That ¹⁴C decay is the well-known basis for carbon dating. It also offers a way to distinguish between CH₄ sources. Biologically sourced CH₄ has the same ¹⁴C level as the organisms that created it; the atmospheric lifetime of CH₄, about 10 years, is too short for any appreciable decay to happen in the atmosphere. On the other hand, fossil CH₄, trapped underground for millions of years, has long lost all its ¹⁴C. In principle, then, measuring the ¹⁴C in a sample of CH₄ should reveal how much of it came from biological and how much from fossil sources.

In today’s atmosphere, that measurement is hampered by the appreciable amount of ¹⁴CH₄ produced by nuclear power plants, and the biological–fossil breakdown of the modern CH₄ budget is still a matter of some controversy. The preindustrial atmosphere, as preserved in glacial ice, presents no such complication.

Petrenko and colleagues’ first attempt to measure ¹⁴CH₄ in ancient ice was in 2009, when Petrenko was a graduate student with Jeffrey Severinghaus (also an author on the new paper) at the University of California, San Diego.³ They had several reasons for focusing on the period from 11 000 to 12 000 years before the present. It’s well before the rise of any anthropogenic emissions, but recent enough—just two half-lives ago—that enough ¹⁴C remains to measure. And it was a period of rapid change, both globally and regionally. The world was just emerging from the Younger Dryas period, which saw much of the Northern Hemisphere covered in ice, and entering the more hospitable Preboreal period. During the transition, many parts of the globe experienced abrupt climate change—parts of Greenland, in particular, warmed by as much as 10 °C in just 20 years. And the global CH₄ level rose sharply—from 500 to 700 ppb—at an annual rate comparable to what we’re experiencing today. The researchers were interested in what sources drove that increase and how it was related to the regional warming.

Because CH₄ is a trace atmospheric gas and ¹⁴C is a trace isotope, an accurate measurement requires about 1000 kg of ancient ice—a bit more than a cubic meter (see figure 1). That’s too much to be feasibly obtained by the traditional method of drilling deep into a stratified ice sheet. Fortunately, certain locations in Greenland and Antarctica expose plenty of old ice at their surfaces. (Because the mixing time of the global atmosphere is about a year, an order of magnitude less than the CH₄ atmospheric lifetime, both Arctic and Antarctic ice record the same long-term trends.) The ice is dated through measurements of quantities, such as total CH₄ concentration and various stable-isotope levels, that changed in known ways over time.

In their 2009 study, Petrenko and colleagues analyzed several samples from across their period of interest. Surprisingly, all their ¹⁴C values were far too high to be explained even by attributing all the ancient ¹⁴C to biological sources. The discrepancy, they determined, was because ¹⁴C creation by cosmic rays isn’t limited to the upper atmosphere; the rays also penetrate near-surface ice to create ¹⁴C in situ.

It took the researchers until 2016 to figure out how to correct for the cosmogenic ¹⁴C. They travelled to Antarctica’s Taylor Glacier, an extraordinary region where surface ice ranges from 8000 to more than 100 000 years old. They analyzed samples of 50 000-year-old ice—old enough that all its original ¹⁴C was gone, and the only ¹⁴CH₄ remaining was cosmogenic.⁴ 

Then, for the present study, they moved 300 m along the glacier to collect samples from the Younger Dryas–Preboreal period of interest. They used the 50 000-year-old samples to correct for cosmogenic ¹⁴CH₄ and estimate the true amount of ¹⁴CH₄ present in the atmosphere 11 000–12 000 years ago.

Figure 2 shows their results, along with a few model calculations based on different assumptions. As indicated by the light blue curve, the data are most consistent with the scenario in which all the CH₄ throughout the period was biological, not fossil. Even the case of 10% fossil CH₄, shown by the green curve, is outside one standard deviation. Far outside the uncertainty limits is the red curve, which assumes a constant 53 Tg/yr of fossil CH₄, a common estimate of today’s nonanthropogenic fossil CH₄ from bottom-up and ¹³C studies.⁵ In fact, Petrenko and colleagues found that at the 95% confidence limit, at most 15 Tg/yr of ancient CH₄—about 7% of the total—came from fossil sources.

It’s not impossible that geological CH₄ seepage could have increased threefold over 11 000 years. But it would be a puzzling rate to explain. Geological processes don’t usually change that much over mere millennia, and if anything, the emissions should have decreased over time. Sea levels have risen since then and covered up some potential CH₄ seeps (and it appears that CH₄ emitted deep underwater may be gobbled up by bacteria before it reaches the surface; see Physics Today, August 2017, page 21). And as oil fields are drained of their natural gas, less remains to seep out naturally.

Petrenko and colleagues’ work therefore suggests, but doesn’t prove, that modern estimates of natural fossil CH₄ emissions may need to be revised downward—and thus that estimates of anthropogenic fossil CH₄ may need to be revised upward. The researchers plan to make similar measurements on ice from 200 to 250 years ago—much more recent but still largely preindustrial—to solidify their case.