This is the fourth post in a new climatetippingoints.info series fact-checking claims that various climate tipping points have been crossed, and that sudden catastrophic warming is now inevitable. See the Introduction post for an overview.
Fact-Check: is an Arctic “Methane Bomb” about to go off?
Claim: A huge amount of methane is trapped in permafrost and methane hydrates in the Arctic and is starting to leak out, and even a partial release could at any time trigger a sudden shock increase in global warming of up to 5°C within 5 years.
Reality: Methane levels have recently increased but so far have a mainly tropical or fossil fuel source. Methane release from permafrost and hydrates will happen as a gradual chronic leak acting as an unwelcome but modest feedback on warming, rather than being a sudden, catastrophic release.
In recent years there have been repeated claims (e.g. 1,2,3,4,5,6,7,8) that a huge amount of methane trapped in Arctic permafrost and hydrate deposits will imminently be released once global warming passes a critical temperature threshold, which will in turn trigger abrupt and catastrophic warming.
In this post of climatetippingpoint.info‘s new Fact-Check series, we investigate how much methane is locked up in the Arctic, how much could be released, and how soon it might happen.
A common claim is that there’s so much of the potent greenhouse gas methane stored in the Arctic permafrost (permanently frozen soils rich in organic matter) that even if only 1% of this methane escaped then we’d see a +0.5–1°C global warming jump in just a year.
Let’s look at the numbers. Our current best estimate is that there’s up to around 1600 gigatonnes (Gt – a billion tonnes) of organic carbon locked up in Arctic permafrost soils. One percent of this would be 15 GtC, which if it were released as CO2 would lead to only a modest warming (=55 GtCO2, so CO2 would go up by ~7 ppm, which would drive +~0.03°C over a few decades). However, methane is a far more potent greenhouse gas than CO2, but it also decays into CO2 with a half-life of only ~10 years. Together, this means that methane produces about 84-86 times more warming than the same mass of CO2 on a 20 year timescale, but decays to 28-34 times more warming on a 100 year timescale. Based on this, if all of that 15 GtC was released from Arctic permafrost as pure methane (=20 GtCH4), it would lead to around 1.1°C of warming in the short-term and around 0.4°C in the longer term (figures updated).
This would seem to support the claim that a release of only 1% of the permafrost methane would trigger a sudden warming of ~1°C. But here’s the key question: can 20 GtC worth of pure methane actually leak out in one sudden event? Here the reality is less dramatic.
The main issue is that not all of the carbon in permafrost would be released as methane. When we talk about organic carbon in permafrost, what we mean are the remains of tundra plants that in warmer climates would rapidly decay and much of its carbon returned to the atmosphere. But in polar climates, these remains can become permanently frozen into deep soil layers known as permafrost, stopping the decay process. This gradually leads to a big build up of undecayed organic carbon, making it a big global sink of carbon. But with warming, these organic-rich permafrost soils begin to thaw out, and lots of this organic material can begin to decay and release gases.
Different types of organic matter decays at different rates, with only some decaying rapidly – in one experiment, 50-75% of the organic matter in permafrost decayed when left unfrozen over 12 years. Most of the bacteria and fungi driving this decay release CO2 as a by-product of respiration, but in waterlogged soils anaerobic bacteria that release methane can dominate instead. But how much methane gets out into the air also depends on the soil and vegetation above, with some being broken down and turned into CO2 on the way. Overall, the proportion of carbon released as methane has been estimated at ~2% of future emissions – nowhere near pure methane, but enough to boost the warming effect of permafrost carbon release by ~40% over the next century. Some CO2 will be taken up by warming-driven vegetation growth in these regions as well, partially counteracting the CO2 release from permafrost thawing.
So how much overall carbon release might we expect from permafrost in the future? Models of future permafrost change are imperfect but improving, and project that around 92 GtC could be gradually released by 2100 under business-as-usual high emission scenario, and up to 200 GtC by ~2300. The former would equate to a CO2 increase of around +43 ppm in the atmosphere, which produces an extra ~0.15°C of warming [including the boost from methane, but with ~50% carbon drawn down]. Another study suggested around 10% of permafrost carbon could possibly be released by 2100 under our current warming trajectory, which is around 130-160 GtC (i.e. +0.2-0.25°C) . In contrast, under a lower emission scenario keeping within the 2°C Paris Target, the projected release is within 28-92 GtC (+0.05-0.15°C) by 2100. These warming figures are similar to or slightly higher than the last IPCC report’s scenario range of 0.13-0.27°C by 2100. However, recent research on abrupt permafrost thaw in areas with large ice pockets or unstable slopes suggests the warming impact might actually be doubled compared to the average permafrost response in the models of the last IPCC report (CMIP5).
Together, these studies converge on permafrost thawing resulting in additional warming of around 0.5°C by 2100 under a burn-all-the-carbon high emission scenario, and around 0.1-0.3°C by 2100 under more moderate emissions scenarios. This extra warming represents an unwelcome climate feedback and makes keeping under the 1.5°C or 2°C Paris Targets that much more challenging, but a sudden permafrost methane release triggering 0.5-1.0°C of warming within only a few years is not supported by current research. And contrary to claims that climate models ignore permafrost, gradual permafrost thaw was included in several models in the last IPCC report, although their responses varied widely and didn’t include abrupt thaw.
Another big claim is that a region in the Arctic Ocean known as the East Siberian Arctic Shelf (ESAS) is leaking methane out of vast undersea methane hydrate deposits (a pressurised mixture of ice and methane trapped in sediments – see our previous Arctic methane post for more information). The ESAS is said to be poised to release 50 Gt of this methane into the atmosphere, triggering a rapid and catastrophic warming of 5°C in only 5 years (the “ESAS-bomb” scenario). But this claim doesn’t entirely stack up either.
This ESAS-bomb scenario started out with a trio of papers. The first paper found a lot of methane dissolved in the water of the ESAS, which was calculated to represent a surprisingly high release of around 8 Teragrams (Tg – equivalent to a Megatonne, 1000 times less than a Gigatonne) of methane (CH4) a year, and in a more recent paper was updated to 17 TgCH4/year. It was hypothesised that this is the result of permafrost under the seabed thawing and letting methane from hydrates trapped underneath to escape. From this, it was then posited that there could be up to 1400 Gt of methane under a subsea permafrost lid – a massive value for only one regional sea compared with the global methane hydrate estimate of up to 2000 GtC (including ~200 GtC in the Arctic) and 1600 GtC stored in all Arctic permafrost. Finally, a third paper then speculated that 50 Gt of this methane is poised to be released imminently, and could happen in as little as 1-5 years. This would cause extreme and rapid warming, quantified as $60tn worth of economic damage in one paper and which another commentator described as being worse than anything since the Permian-Triassic mass extinction 250 million years ago.
However, other Arctic and methane hydrate researchers have been sceptical of many of these claims, pointing out several major issues.
First of all, while the evidence of high levels of methane was clear, there was no conclusive proof of this being a totally new leak. As detection methods have become more sensitive and widely used, more previously unknown sources of methane have been discovered. Some of these – for example the famous methane bubbling off of Svalbard – on further examination turned out to be going on naturally for thousands of years, mostly doesn’t make it to the atmosphere, and may even draw down CO2. Modelling also suggests that even under a shallow sea permafrost has a lag time of hundreds of years behind global warming before significant destabilisation occurs. And although 8-17 TgCH4/year sounds like a lot, on a global scale it’s not huge compared to the average global release of ~558 TgCH4/year – a fact noted in the first ESAS-bomb paper (“…the current estimate is not alarmingly altering the contemporary global CH4 budget“). Subsequent in-depth modelling and statistical analysis of the ESAS data also suggests that winter emissions have been over-estimated, bringing the emissions down to a more modest 0-4.5 TgCH4/year from a mostly biogenic permafrost source. This fits with a recent research cruise using accurate methane detectors, which estimated ESAS emissions of ~3-4.7 TgCH4/y.
Second, there is even less proof of the existence of such a huge methane hydrate reservoir directly under the subsea permafrost. The claim in the second ESAS-bomb paper is based not on their own observations as sometimes reported, but on two old and hard-to-find sources and from considerable assumptions and indirect extrapolations. Although hydrate deposits have been found across large areas of the Arctic (around 200 GtC), none of the studies claimed in support of the ESAS-bomb actually prove the existence of deposits specifically under the ESAS as massive as 900 Gt (less than 1400 Gt as 500 Gt of the headline figure was general organic carbon rather than methane). Given that methane hydrates only forms under high pressure (unless it is in a rare metastable state, for which clear evidence in the ESAS is still lacking), it’s also likely that most of it lies further below the seafloor and will take a while for warming to reach and for the gas to then travel up. It’s notable too that more recent work from the same research group no longer specifically posits a 1400 GtCH4 reservoir ready to rapidly release 50 Gt.
Thirdly, the evidence for a rapid 1-5 year release is also limited. The initial scenario in the third ESAS-bomb paper was based on extrapolating a 5% emissions growth rate over 50 years, leading to an eventual emission of 50 Tg/year. However, a typo in the original paper stated this as 50 Gt instead, which has since been clarified. The paper then adds a 1-5 year release scenario based on the stated assumption that ~3-5% of the seabed is vulnerable to instability, and so a similar percentage of the assumed massive methane reservoir (stated as 50 GtCH4) would be immediately released if this area slumped. But as we’ve explained, the evidence for such a large reservoir of free methane is shaky, and no justification is given for 3-5% of the ESAS seafloor simultaneously collapsing. Supporters of the 1-5 year scenario cite various papers stating that methane release this rapid is possible, but these papers actually only support releases over several decades rather than only a few years. Even so, the warming modelled for the rapid 5 year release was only an additional 1.3°C rather than 5°C as often quoted. Based on our earlier sums, even if 50 Gt of pure methane were emitted all at once it would lead to 2.6°C of short-term warming rather than 5°C (figures updated – see the comments for calculation details).
Finally, there’s no evidence that a dramatic methane release occurred from Arctic methane hydrates during past episodes of Arctic warmth and rapid warming, most notably during the warm Eemian interglacial, deglaciations, or Holocene climatic optimum. Some have suggested this is because ice bubbles in cores form too slowly to capture rapid methane releases. But these processes (known as diffusion and smoothing) only loses some of the annual signal and can be mathematically recovered, and a large part of rapid spikes would still be captured anyway. The resulting 5°C warming within a few years would also definitely be clear – but is clearly lacking – in many other types of palaeoclimate records. Claims that the Eemian had too different an orbital seasonality to be relevant ignore that the Arctic was still 2-4°C warmer than today, with forests reaching the Arctic Ocean and hippos living in the River Thames. And while records of Eemian permafrost melt show that a significant part thawed when warmth hit 1.5°C above pre-industrial, temperatures didn’t go up much higher as a result. This indicates that this isn’t an abrupt tipping point for rapid warming. The thawing also occurred on land, which responds a lot quicker to warming than under even a shallow sea.
Overall, since 2010 evidence has accumulated that the ESAS is releasing more methane to the atmosphere than previously realised, and so constitutes a fairly significant global methane source. But there’s little evidence that all of these emissions are entirely new and driven by global warming, and claims of vast stores of free methane gas lying just under the seabed ready to be abruptly released are highly speculative and unproven. Claims that critics have insufficient expertise for this specific sea also ignore the critics’ wide-ranging and relevant experience across both Arctic and methane hydrate science.
Several sources go on to say that methane concentrations in the atmosphere have already started to go up exponentially, and as it’s highest in the Arctic that this proves that Arctic methane is the source.
Here’s the recent record of methane concentrations in the atmosphere:
There clearly has been an increase in global methane concentrations since around 2007, and at a faster rate since 2015. But the increase is not exponential, either in its narrow mathematical definition or in the broader sense of continuously accelerating increases. The long-term trends in methane are far more erratic than the smooth increase in CO2, with for example an 8 year hiatus (with uncertain causes, but possibly from quicker methane breakdown) before growth resumed in 2007. Some sources try to project exponential increases based on this, but these are based on extrapolating from a few selectively chosen data-points with poor trend-fitting techniques that ignore the actual dynamics of what’s driving the increase (as explained in our previous Arctic sea ice fact-check). But where is this recent increase in methane concentrations coming from?
Methane sources can be tricky to pinpoint, but researchers can use tracers such as its isotopic signature (the balance of heavier and lighter atoms of the same element) and global patterns to narrow it down. Modelling using these isotopes and concentration patterns indicate that the recent increase is mostly from either tropical or sub-tropical sources (e.g. from farming or wetlands), fossil fuels (e.g. natural gas leaks), and/or a slowdown in how quickly methane breaks down in the atmosphere. This modelling also indicates that after 2007 it’s mostly not been coming from polar regions, which would rule out Arctic permafrost or methane hydrates as the driver of the recent increase.
This also matches with global methane budgets, which try to measure sources and sinks of methane around the world and have found no significant increase for Arctic methane (at about 1 TgCH4/y from permafrost and 2 TgCH4/y from hydrates). Although some of these estimates are more uncertain than for CO2 this doesn’t mean there’s not a reasonable consensus (as some claim otherwise) on the overall trend, especially on the gap between sources and sinks that’s driving the observed increase. As there’s quite a bit of uncertainty for each source but more certainty on the total source/sink difference, discovering that one source is bigger than previously estimated would mean that the others would likely be smaller instead, rather than this counting as a definite overall increase in emissions. And as with climate skepticism, the presence of uncertainty and debate over the details of the mainstream understanding of atmospheric methane does not make the opposite alternative just as likely and worthy of equal credibility.
Some people question the relevance of the global average methane concentration record though, and point to some higher concentrations detected high up above the Arctic as proof of an Arctic methane leak. This is justified by claims that the Arctic Ocean is under-sampled at the surface, and that the difference between surface and mid-altitude concentrations supports a methane hydrate source causing methane plumes.
But while Arctic methane concentrations are often higher than the global average, this doesn’t actually make it likely that’s where methane is coming from. Most methane emissions comes from the northern hemisphere as it has greater land area than the south and so contains more key methane sources like wetlands (and humans!). As the Arctic lies at the northern hemisphere’s centre this effect is felt greatly there. Methane concentrations also heavily depend on other factors, including the rate of its breakdown by chemicals known as free radicals of which there is less at the poles and more towards the tropics. Together, the overall effect is that slightly more methane tends to naturally accumulate in the northern hemisphere anyway, and especially in the Arctic.
Another problem with these claims is that the data shown is often an extreme outlying value cherry-picked from raw unprocessed data on one day or week (e.g. 1,2,3). These values are often far above times before and after, making the situation seem far worse than it actually is. For data to be useful though it has to be interpreted within its full context, much like one week of intense cold weather does not disprove the long-term global warming trend. It’s the global trend that determines methane’s total impact on global warming, rather than temporary local fluctuations. These can have many causes (e.g. weather patterns and local decay rates) other than an increase in methane sources.
Finally, claims that patchy data collections from the surface ocean means that we under-estimate Arctic methane concentrations also ignore the data provided by satellites across all altitudes, as well as the overall constraints provided by global methane budgets that use multiple different data types and methods. Methane tracking and budgets are not solely dependent on surface station data, which complement the other methods. Recent research cruises have helped fill some of these gaps as well.
This is linked to a claim that measurement stations and satellites fail to spot intense methane plumes rising from Arctic methane hydrates, and that this is supported by methane concentrations being higher at mid-altitudes after spreading out. But no evidence is provided for the existence of super-concentrated plumes that hardly spread out in the lower atmosphere (in spite of clear evidence of polar regions having more effective horizontal than vertical methane transport), and there are many other mechanisms driving higher concentrations further up. Measurement and models of atmospheric methane show that methane decays far more quickly at the surface (within ~1 year) than the mid-upper troposphere (i.e. mid-altitude atmosphere, where it lasts ~12 years), and especially during polar winter in the latter (~1000 years). A recent cruise across the whole ESAS found a few bubbling hotspots, but these weren’t large enough to drive significant emissions and were nowhere near as big as superplume proponents have suggested.
As well as this, surface and mid-altitude methane have opposite yearly cycles, with surface concentrations lowest (due to more rapid decay) and mid-altitude concentrations highest (due to increased upwards convection and a higher tropopause) in the Arctic summer and vice versa. This naturally leads to higher methane concentrations in mid-altitudes than at the surface. The changing gradient in March cited as proof of methane plumes is likely to be due to this summer contrast between surface and mid-altitude concentrations arriving earlier in the year with warmer springs. Using mid-summer methane profiles would no doubt have shown a constant long-term pattern instead.
More methane – a potent greenhouse gas – is likely to be emitted from Arctic permafrost than the IPCC originally anticipated, but modern models have started to catch up and quantify them better. Permafrost emissions are likely to drive ~0.1-0.3°C of extra warming by 2100 under low emission scenarios and up to ~0.5°C under high emission scenarios. But there’s no evidence for massive releases of pure methane over only a few years either now or in the past. Similarly, while evidence has emerged of greater than anticipated methane emissions from subsea Arctic permafrost in the East Siberian Arctic Shelf, there’s scant direct evidence of a huge reservoir of metastable methane hydrates just below the surface that could suddenly leak and trigger extreme warming.
Methane concentrations in the atmosphere have begun to increase again after 2007 and more rapidly since 2015, but the increase is not exponential and the likeliest drivers are towards the tropics rather than the Arctic. While there’s still large uncertainties on individual methane sources and sink estimates, there’s still a reasonable consensus on the overall global balance and trends. And higher mid-altitude methane concentrations above the Arctic (with occasional very high local readings) are mostly driven by regional atmospheric dynamics rather than undetectable methane plumes.
The most likely situation then is one of a gradually growing chronic leakage of additional methane and CO2 from the Arctic over the coming decades and centuries, rather than an abrupt “methane bomb”. This will act as a gradual amplifier of human-driven global warming, making staying within the Paris Targets of 1.5-2°C even more challenging and urgent.
Updated on: 17/5/19 to correct the methane warming calculations for permafrost and the ESAS-bomb (see comment section for details); 17/10/19 to clarify the permafrost warming estimates (no significant change overall); 31/1/20 with new study supporting modest ESAS CH4 emissions; 17/2/20 to add new study giving more details on abrupt thaw potentially doubling permafrost emissions versus the CMIP5 average.
This post was written by Dr. David A. McKay, currently a Postdoctoral Researcher at Stockholm Resilience Centre (Stockholm University), where he is part of the Earth Resilience in the Anthropocene Project (funded by the European Research Council) and is researching non-linear climate-biosphere feedbacks. This post was written in his spare time with no funding support for this site.
Featured Image: Permafrost thaw ponds in Hudson Bay Canada near Greenland. By Steve Jurvetson – http://www.flickr.com/photos/44124348109@N01/2661598702 en:Flicker.com, CC BY 2.0, https://commons.wikimedia.org/w/index.php?curid=34436593