Fact-Check: is an Arctic “Methane Bomb” about to go off?

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.

One Percent Rule

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 +0.51°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.

A map showing the major permafrost regions bordering the Arctic ocean, and where the thaw is happening fastest. Source: Turetsky et al. (2019), Nature; data from Refs. 1 & 2

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, or around an extra 0.3°C of warming (including the boost from methane). 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. +~0.5°C) . In contrast, under a low emission scenario keeping within the 2°C Paris Target, the projected release is closer to 45 GtC (+~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 permafrost collapsing as it thaws suggests the lower warming figures might actually be doubled, as slumping could release CO2 and methane more rapidly.

Together, these studies tend to converge on permafrost thawing resulting in additional warming of up to 0.5-0.6°C by 2100 under business-as-usual high emission scenarios, and around 0.15-0.3°C by 2100 under low 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.

Short Shelf Life

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.

Map of the Arctic Ocean showing the East Siberian Arctic Shelf (ESAS). Source: NOAA/Skeptical Science

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.

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.

Exponential Claims

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:

Global average methane concentrations (at Earth’s surface in the remote marine troposphere) from 1998-2018 (top) and its isotopic signature (bottom). A longer record from the Arctic can be found here. Source: Nisbet et al. (2019) GBC

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.

Illustration of the Global Methane Budget for 2003-2012. Note that permafrost and hydrate emissions are counter under ‘Other natural emissions’, and the large sink in ‘chemical reactions in the atmosphere’ (mostly from free radicals). Source: Global Carbon Project

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.

Hard to Concentrate

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.

Methane concentrations in the mid-altitude atmosphere in January 2016. Note the higher concentrations across the Northern Hemisphere. Source: NASA Earth Observatory map by Joshua Stevens, using AIRS data.

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.

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).

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 ~0.15-0.3°C of extra warming by 2100 under low emission scenarios and up to ~0.5-0.6°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 the comment section for details. Apologies for the mistake!

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


  1. Really nice write up! However, I am confused about one statement in this article; “1.3°C rather than 5°C as often quoted. For such a big release of pure methane this is actually quite a low figure – based on our earlier sums 50 Gt of pure methane emitted all at once would lead to a whopping 9°C of warming!” What should be believed here? Should the modelling of 1.3°C be believed, or the authors own calculations of 9°C? For a layman like me, it becomes difficult to understand when all of these other alternatives get thrown around, especially with the regards of 50 Gigatonnes where I’ve heard statements of Wadhams 0.6°C, Shakhova’s 1.3°C, and now 9°C.


    1. Hi, good question. I had to double-check this before publishing too, as it was quite confusing. I believe it’s probably mostly due to different assumptions in how quickly the methane is released. It also turns out I actually made a commonplace carbon/CO2/methane mass conversion error in this section, so the pure methane warming would not be quite as much as 9°C (but it’s still more than in Shakhova’s paper). I’ve updated the article with the corrected figures, and explain the details below. Thank you for drawing my attention to it!

      My calculation was based on an instantaneous release of 50 Gt of pure methane – a very unlikely scenario, but a useful thought experiment that’s not far off what some people have proposed online. 50 Gt of CO2 would push up atmospheric CO2 by ~6.4ppm, which would drive around ~+0.03°C of warming (using TCS and a 415ppm baseline). But methane is c. 85 times more effective than CO2 in the short-term, so that’d boost this warming to ~+2.6°C (before declining). My mistake here was initially calculating 50 Gt of *carbon* rather than of specifically CO2, which as each 1ppm of CO2 equates to 2.12 Gt C or 7.81 Gt CO2 and global warming potentials are done on a tonne-per-tonne basis meant I accidentally boosted the numbers. Apologies for that… But 2.6°C is still higher than the 1.3°C claimed by Shakhova, as you point out.

      I’ve read through Shakhova’s paper and the methodology isn’t entirely clear, but if their peak radiative forcing (~+6.5 W/m^2) they model for their warmest scenario were maintained it would result in a far higher warming (~3.3°C, using a Transient Climate Sensitivity of 2°C per 2xCO2 – similar to my new figure of 2.6°C too) than what they actually show (1.3°C). However, this high radiative forcing falls soon afterwards as the methane decays, so the system is never in equilibrium and temperatures do not have enough time to catch up with the peak forcing. I haven’t checked their radiative forcing sums either, so can’t vouch for those figures.

      I believe Wadham’s figure is lower because they modelled a decade-long pulse of 50 GtCH4, which is long enough that most of the initial methane released will have decayed by the end of the pulse, resulting in a much lower peak methane concentration and radiative forcing overall. Shakhova’s peak scenario was a release within 1-5 years, allowing a bigger methane buildup before it’s reduced by decay. My sum is for instantaneous less-than-a-year release, so is the extreme case. This fits with the slowest release rate (Wadham) getting the lowest warming (0.6°C) and the fastest (me) the highest warming (2.6°C), with Shakhova’s 5y release in the middle (1.3°C). But this all has the caveat of a rapid release of 50 GtCH4 being relatively unlikely, and a more longwinded leakage being most likely.


      1. Thank you for a really detailed reply, David! I’ve been confused around the subject of a 50 Gigatonne release for a while, so it’s nice to finally see some numbers and details around it all.

        Liked by 1 person

      2. Upon re-reading I have a few more questions I’d really like answered. This must be annoying, sorry!

        From reading your article and comment it seems that the generic “C” has a higher value than CH4 and CO2? As in, C appears to have more potential than both of those. Not sure if that statement makes sense. Specifically it’s about this: “as each 1ppm of CO2 equates to 2.12 Gt C or 7.81 Gt CO2” statement. How much CH4 is needed for 1 PPM of CO2?

        50 Gigatonnes of methane would be the equivelant of 4000 Gt of CO2 and boost the methane burden in the atmosphere to ~20-21 PPM, how much is this in CO2e PPM? It seems that it would more than double the CO2e temporarily, before it decays.

        Third; Would the peak of 2,6°C even be reached? There’s a multitude of negative feedbacks to hold such an immense and quick warming down and make it more gradual, if I’ve understood a few things correctly.

        Thank you for any response, I know it must be annoying to have someone pester you about things like this.


      3. No problem, comments and questions help to clarify future posts!

        So by generic C I’m referring to the amount of “pure” carbon locked up in various forms, but it is rarely stored as pure carbon in the Earth (unless it’s graphite or diamond) or released as pure carbon into the atmosphere (except soot). Carbon in places like permafrost is locked up within organic matter, which is normally come combination of carbon, hydrogen, and oxygen atoms in various different molecules but generically represented as CH2O. Carbon has an atomic mass of 12 while hydrogen is 1 and oxygen is 16, so a molecule of CH2O weighs 2.5 times a single carbon atom (as CH2O has a combined atomic mass of 30, versus 12 for C) and means organic matter containing 50 GtC weighs 125 Gt itself in total. When this organic matter is then degraded and respired, if all 50 GtC was released within CO2 then it’d be 183 GtCO2 (as CO2 atomic mass = 12 + 16 + 16 = 44 – the extra mass is from atmospheric oxygen), but if it were all degraded anaerobically and released as methane it’d be 67 GtCH4. Similarly, 1 ppm of CO2 in the atmosphere is equivalent to 7.81 Gt CO2, of which 2.12 Gt is just the carbon without the O2 (as 12/44=0.272). When measuring carbon storage climate scientists tend to measure it in terms of pure C to make conversions between different forms easier, but you have to remember to turn it back into CO2 or CH4 when calculating warming impact – which is what I initially forgot to do!

        If 50 Gt of pure methane were released in one go (unlikely as discussed before, but lets work on it as a useful hypothetical), then it’s global warming potential on a timescale of over a couple of decades is equivalent to 4250 GtCO2 (as it’s ~85 times more powerful). This would translate to a temporary ~+544 ppm CO2 in the atmosphere. So yes, a more than doubling and would translate to a +2.6°C warming (assuming 415 ppm baselined and a transient climate sensitivity of 2°C) as calculated before.

        But as you point out, this is very much the maximum warming, with many factors limiting the actual likely warming. The main issue is that even an instantaneous release of methane would begin to rapidly decay, which combined with the warming lag (the delay between increase in radiative forcing and temperature catching up) means that by the time temperatures are increasing the total radiative forcing would already be falling again. This would make the peak warming a bit later, lower, and smoother than the methane spike itself. This can be seen in the projections of Wadhams and Shakhova, who (as well as using a multiyear methane release) project substantially lower than 2.6°C peaks delayed by a few years after peak methane. To get the full likely warming from my hypothetical instantaneous 50GtCH4 release we’d have to use a climate model that resolves methane decay and warming lag, but it would likely not reach the maximum 2.6°C in practice, although probably still a bit more than Shakhova’s 1.0-1.3°C as they used a 1-5 year release rather than instantaneous.


      4. You’ve been all too kind answering all my questions in such a great manner! The only thing I’m left wondering is if this calculation accounts for indirect forcings (water vapour, ozone etc.)? Or is that where the TCS comes in?


      5. No problem! And yes you’re right, the Transient Climate Sensitivity automatically includes the effect of fast feedbacks like water vapour on the direct CO2-driven warming (with ~1°C from CO2 and ~1°C extra in fast feedbacks for a doubling of CO2 equivalent), so those warming calculations include the known predictable feedbacks. For longer timescales of around a century the slower feedbacks push up the sensitivity by another degree or so, and even more on a millennium-plus timescale.


      6. From reading your replies and a few other sources I think I’ve come up with a relatively reasonable conclusion relating to this. It seems that it would be roughly as bad as a continued business as usual. Actually, it’s probably not as bad long-term because of the short atmospheric life of CH4, while CO2 remains virtually forever in a human sense.

        Yet again, thank you so much for answering all my silly questions and allowing me to learn more about the science and scenarios, I’ll make sure to keep following your blog for other potential posts!


      7. Yes, I think that’s a reasonable way of looking at it. In the Intro post [climatetippingpoints.info/2019/04/01/climate-tipping-points-fact-check-series-introduction/] I put it in terms of risk, with climate tipping points and feedbacks loading the dice a bit more against us by making high warming scenarios slightly more likely than otherwise (for example, in one study the upper 5% probability temperature shifts from 4.9 to 6.5°C due to tipping point risks). This means even if we stuck to an rcp 4p5 or 6p0 scenario (which represents current commitments), the chance of warming overshooting and going closer to rcp 8p5 (i.e. business-as-usual) might be somewhat higher than we originally thought.

        Thanks for your questions and feedback!


    1. Hi Yinn, yes all the warming estimates implicitly include things like the water vapour feedback. This is because I’ve used the Transient Climate Sensitivity of ¨2°C per doubling of CO2 (for timescales of the next few decades), which includes around 1°C for direct CO2 warming and 1°C of fast feedbacks, most of the latter being driven by water vapour. For longer timescales (e.g. around a century) the Equilibrium Climate Sensitivity of ~3°C per CO2 doubling is used, which adds on the effect of the slower feedbacks as well.


  2. Hi, glad to see you debunking the methane bomb myths. I wonder if you know about this recent 2018 models roundup type review – which found that conclusions for the melting of the terrestrial permafrost are much less than these estimates. If I use their RCP 4.5 figures, whihc is about what we are on, it is less than a tenth of a degree by 2199 and it can remain carbon negative all the way through to 2199. It seems to be currenlty carbon negative.

    I did a general debunk here

    The section on the 2018 review paper is as follows:

    For 2.4°C by 2100 (RCP 4.5), the models range from 66 gtons loss of CO2 to a 70 gtons gain BY 2299. (A petagram is the same as a gigaton, a billion metric tons).

    The reason it can go carbon negative is because of the growth of peat and ground cover which offsets the methane emissions.

    For 4.9°C by 2100 (RCP 8.5 or “business as usual”), then it is between 74 and 641 gigatons emissions by 2299.

    As we saw, RCP 8.5 is not a realistic futue. There is no way we are gong to do that even with existing policies – and to get to that scenario means not just reversing all our policies now, but no reductions at all through to 2199.

    To get a rough idea of what that means, then at 420 gigatons for half a degree, then 642 gigatons is about 0.76 C and the 70 gigatons of the 2.4 C path is about 0.083 C or about a twelfth of a degree (and 66 gigatons is a net reduction by a fourteenth of a degree).

    So, roughly, we are headed for at most 0.76 C by 2199 with “business as usual” and if we can get down to 2.4 C by 2100 then it’s at most a twelfth of a degree by 2199, in both cases most of that after 2100.

    Their conclusion says:

    >”Despite model uncertainties, the results of this study indicate that, under climate change trajectories resulting from little or no mitigation effort, such as the RCP8.5 climate we considered, the northern permafrost region would likely act as a source of soil carbon to the atmosphere, but substantial net losses would not occur until after 2100. Under climate change trajectories resulting from more aggressive mitigation, such as the RCP4.5 climate we considered, our analysis indicates that the northern permafrost region could act as a net sink for carbon. These results have significant implications for climate mitigation policies, as they indicatethat effective mitigation policies could attenuate the negative consequences of the permafrost–carbon feedback that are likely to occur under policies that result in little or no mitigation.”

    >Dependence of the evolution of carbon dynamics in thenorthern permafrost region on the trajectory ofclimate change (https://www.pnas.org/content/pnas/115/15/3882.full.pdf)

    In that paper they say that there would be more of a carbon sink if it weren’t that the plants are reduced in growth due to lack of nitrogen in the soil. With more nitrogen in the thawing permafrost there would be much more of a carbon negative effect here

    [If you spot any errors be sure to say!]


    1. Hi Robert, thanks for your comment.

      So the permafrost release scenario of 130-160 PgC by 2100 (based on a model range of 37-174 PgC) I quote is from Schuur et al (2015), and is based on rcp8p5. This range broadly ties in with the study you link to (McGuire et al, 2018), who show a model range of soil carbon change of approx. +40 to -180 PgC by 2100 for rcp8p5. However, they also project an increase in vegetation carbon due to boreal forest expansion in these regions (although with a very large range as dynamic vegetation models disagree a lot in these regions, with modelling of nitrogen availability being particularly limited – most indications are that global vegetation gain will be less when accounting for N changes), which leads to a net ecosystem carbon change of +200 to -40 PgC by 2100 for rcp8p5. Schuur et al (2015)’s numbers don’t include this vegetation change, so could be seen as overestimating total carbon loss for the tundra region. However, it should also be remembered that boreal forest expansion in the tundra region is also at the expense of boreal forest loss further south to steppes, and it’s not clear that the boreal forest northward shift leads to a net global carbon storage gain (whereas permafrost loss is definitely a net global loss). This means the difference between the modelled net carbon change ranges depends on whether one calculates the net change for just the tundra region itself or globally. Personally, I prefer that permafrost soil carbon loss is reported separately to the overall boreal forest shift carbon changes to make sure the net global impact is clear, but it’s worth bearing in mind that on a regional scale the ecosystem may temporarily gain carbon.

      Either way, the different studies here definitely agree that carbon loss from permafrost acts a long-term feedback with most of the impact after 2100 with much of the initial loss probably offset by vegetation gain. There are large uncertainties though, with for example the recent paper by Turetsky et al (2019) [https://www.nature.com/articles/d41586-019-01313-4] suggesting that thaw craters could drive carbon loss a bit faster than considered in models so far. The evidence on its current status is also unclear – at least one review found evidence of tundra being a net CO2 source over the last few decades: http://doi.wiley.com/10.1111/ele.12164. The total permafrost carbon pool size tends to be underestimated in models by as much as half as well, which might downwardly bias current emission projections.

      As for the appropriateness of rcp8p5, many studies used rcp8p5 as a reporting baseline as it’s the scenario that our emissions were roughly tracking in 2014 [https://rdcu.be/bFDEd], although now it’s more in the middle [https://www.carbonbrief.org/analysis-why-the-ipcc-1-5c-report-expanded-the-carbon-budget] (although technically rcp 4p5 & 2p5 is greater than rcp6p0 CO2 ppm at this point due to scenario details – they’re all pretty close though). However, as you point out there’s some debate as to whether rcp8p5 is a realistic future. It’s been argued that there simply isn’t enough extractable fossil fuels to provide this much carbon, although this assumes there aren’t many more large undiscovered deposits left to find. Governments have also made commitments towards the Paris agreement that if kept to would bring us down to between rcp4p5 and rcp6p0, although this assumption of course depends on if they fulfil their promises. Overall, I think rcp8p5-level human emissions are unlikely (but not impossible) with reality likely to be more like rcp4p5-rcp6p0, but that long-term natural feedbacks are likely to somewhat amplify anthropogenic emissions more than accounted for in IPCC AR5’s rcp scenarios. As a result, rcp8p5 should be seen as the worst-case feasible scenario – not necessarily likely, but worth studying anyway as motivation to avoid it! The next IPCC round (AR6) will have new more realistic scenarios, and its models will better resolve permafrost and vegetation changes, so hopefully will help resolve some of these questions.

      Also a check on your warming estimates – by my estimate, an example 70 GtC release would lead to +33 ppm CO2 (based on 2.13 GtC per 1 ppm CO2), which assuming a current CO2 baseline of 415 ppm CO2 would lead to a short-term warming of ~0.15°C and a longer-term warming of ~0.24°C (based on a Transient sensitivity and Equilibrium sensitivity of 2°C and 3°C per CO2e doubling respectively).


  3. Hello! I highly appreciate these articles, and they are an amazing way to help me not freak out over overexaggeration.
    However, what do you think of the idea of the freeform gas underneath the ESAS, as stated in https://envisionation.co.uk/index.php/nick-breeze/203-subsea-permafrost-on-east-siberian-arctic-shelf-now-in-accelerated-decline ? The place I usually get these articles from is not particularly helpful, and pushes the fact that we’re doomed.

    Forgive me if this is already covered, I’m not too well versed in the world of climate science.


    1. Hi Tyler,

      This topic is mentioned in the “Short Shelf Life” section in this post, in which we discuss the claims that there’s a huge amount of methane ready to be released from the East Siberian Arctic Shelf (ESAS). In short, while they definitely have detected some reasonably significant methane emissions in the region (although the yearly flux is probably less than they estimated), they’ve provided no direct observations or evidence that this is from modern hydrate destabilisation or a vast free gas reservoir. No-one has actually found hydrates or free gas down there yet, but it’s treated like it’s certain and quantified as a huge reservoir.

      Firstly, the leak they detected is likely to have been going on for millennia already, with several similar sites having long-term leaks as a result of ice sheet retreat. Better data collection and instruments means we’re now finding it, but extra evidence would be needed to establish an acceleration due to modern warming. The 1400 Gt of hydrates and 50Gt of free methane figures often quoted was originally a speculative hypothetical scenario based on several significant assumptions and extrapolations (detailed in the post), but has now been quoted so often as to seem like a real estimate of observed gas. [When I tried to source the first figure I came to two offline russian papers from the 80s I couldn’t source, so it’s hard to check]. This also goes for their 1 to 5 year release of free methane scenario, which was a speculative scenario used for a worst-case modelling scenario based on extrapolation rather than significant evidence. Lastly, if this shelf was so capable of massive free methane releases it likely would have also erupted during previous similarly warm periods in the Arctic (like the Eemian or early Holocene), but no evidence of this has been found in the geological or ice records.

      In the linked post Shakhova does make a case for why low pressure hydrates (a key debated assumption) could be below ESAS, but the reasons given are again fairly speculative (with no actual pressure or hydrate measurements) and still don’t justify for example why they assume things like the whole sedimentary column holds hydrates. A key paper in response to these claims (https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2011JC007218) found that under current warming trajectories it will take centuries for the subsea permafrost to degrade significantly (because of thermal intertia), which means even if there was lots of methane trapped underneath it’s still fairly well trapped in the meantime (unless we assume it can come out of just a few holes, but that is highly speculative too).

      For more details and links check out the “Short Shelf Life” section, and if anything there isn’t clear let me know and I can clarify it!


      1. Thanks for the reply! I greatly appreciate it.

        I guess I didn’t make the full connection to it there, really. If I remember right (could also be horribly wrong), they’re treating it as such due to how the permafrost/ice/??? of the region acts a certain way, from what I’ve read.

        This is all incredibly helpful for the doom and gloom chaos frequently found- it’s bad enough already, let alone with misused sources. Thanks!


      2. No problem, glad it’s helpful!

        The crux of their argument that this permafrost/hydrate region is special (at least as summarised in a comment in https://envisionation.co.uk/index.php/nick-breeze/203-subsea-permafrost-on-east-siberian-arctic-shelf-now-in-accelerated-decline) is that a combination of the permafrost forming a lid at the seafloor, plus some sort of pore-freezing effect when the shelf was above sea level, helps pressurise the sediment underneath to preserve some of the original hydrates and trap any methane from hydrates that have dissociated. Otherwise, under normal conditions you only get hydrates forming in sediment under 100s of metres of sea, because it requires both low temperature and high pressures. But these justifications remain fairly hypothetical and speculative – the article has an awful lots of telltale “coulds” here – and it remains the case that there’s little direct observational evidence of any very shallow hydrates or large volumes of free gas in this region. The pressure figures given in the article here are interesting too – 5-8m of permafrost needed to pressurise versus 10m of water isn’t actually that big a difference in depth terms, and to me that doesn’t justify hydrates being that much shallower… I also remain very sceptical about assuming the whole ~2000m sedimentary drape could host hydrates (which is what makes it a vast reservoir), when everywhere else hydrates have a lower bound some 10s to a few 100 of metres deeper for clear thermomechanical and chemical reasons [I helped do some hydrate modelling a while back to do with interactions with climate, and we used very typical thicknesses of 30m in shallower water and 150m in deeper, colder water: https://agupubs.onlinelibrary.wiley.com/doi/full/10.1002/2016GL069676%5D.

        I think the case remains that simple subsea permafrost degradation (much of it relict from post-glacial changes, with an increasing modern climate component over time) is more than sufficient to explain their methane readings, and that extraordinary claims beyond that require extraordinary evidence to justify them. It’s fine for them to hypothesise and speculate about these things if they then go out and test and justify them empirically, but it shouldn’t be stated and treated as such certainties in the meantime when they’re not.


    1. Hi Tyler, thanks for the link, I hadn’t seen this paper yet (been on holidays). – it’s nicely done and the point of accounting for processes like compaction is sound.

      What’s interesting is how most of lost carbon ends up laterally transported by water rather than emitted directly as CO2 to the atmosphere. The implication is that previous experiments observing permafrost CO2 emissions got that about right, but missed even more carbon being seeped away by water under the surface, either in dissolved or particle form. How this then affects total CO2 emissions gets complicated though – it depends on what happens to that dissolved /particulate carbon once it gets flushed out into rivers and the sea, where it can be buried and lost or respired and slightly enrich the seawater carbon stock (which might in turn affect CO2 drawdown in the ocean). As they put it, “the fate of the transported C is highly uncertain”. The caveat to all this though, as ever with experimental/observational studies, is how much this can be extrapolated across all the permafrost region, as soil types and dynamics can widely vary and this process may happen in some soil types more than others.

      Overall I think the main takeaway is that more permafrost C might well be lost than I summarise in this article, but that this extra C probably won’t be directly emitted as CO2 (it may end up emitted some point later depending on seawater chemistry, or buried at sea – it’s hard to say). This paper also backs up the point about how very little of permafrost C ends up as methane during degradation, with them finding it almost negligible in this experiment.


  4. So the whole Arctic permafrost melting 70 years early and your tweets about it (https://twitter.com/ClimateTipPoint/status/1141654407936917504?ref_src=twsrc%5Etfw%7Ctwcamp%5Eembeddedtimeline%7Ctwterm%5Eprofile%3Aclimatetippoint%7Ctwcon%5Etimelinechrome&ref_url=https%3A%2F%2Fclimatetippingpoints.info%2F2019%2F04%2F15%2Ffact-check-is-global-dimming-shielding-us-from-catastrophe%2F) don’t change anything in this article? Meaning we’ll still have modest feedback on warming (~0.5-0.6°C under high emission scenarios)? Just somewhat confused because you say melting permafrost is “bad” and a “worrying trend”, but it doesn’t contain much of the total carbon?


    1. Hi Joe, yes I think the feedback calculations in this article are still about right. The worrying-ness of that early-melting article was mostly just that it goes to show how quickly parts of the Arctic are warming up because of polar amplification and how something we assumed was safe turned out not to be so much, even if it doesn’t boost the overall carbon feedback very much itself (it will a little of course, but I’d guesstimate it to probably be within the current uncertainty range). The 0.5-0.6°C (by 2100 on top of high emission rcp8p5) I collated is mostly driven by the C-rich permafrost regions, which as summarised in that twitter thread have differing thaw dynamics to the low-C / ice-rich stuff that’s unexpectedly melting. Of course that’s still a fairly significant amount, and is only modest compared to some of the bigger claims out there! Interestingly another recent but less-noticed paper (discussed in the comments above) might have bigger ramifications for my estimates, but as discussed it’s not a direct feedback on atmospheric CO2 so it gets a bit complex.


  5. Hi, doctor McKay!

    I’ve recently read this article and it’s great! This site has helped a lot with my anxiety over climate change and made me try to look for the real science amongst the different “groups” that have came up recently.

    However, from reading the article and comments I’d like to know why a GWP of 86 was used for methane? I’m not a scientist, so I have some trouble finding the very near-term GWP, but I’ve heard estimates ranging from 120-150 when it comes to GWP1! As a layman I of course know very little about the physical science of climate change, but if I were to take your calculations of 0.03°c and multiply it by 150, it would be more like 4.5°c, would it not?


    1. Hi John, thanks for the positive feedback! Re. GWP over very short timescales, it’s theoretically true that methane is even more powerful when it’s first released with a GWP of around 120xCO2, which would seem to make abrupt methane-induced warming much worse than estimated. But one of the big issues here is that even though the methane technically has a GWP at this point, it takes time for it to spread out and mix into the atmosphere, and during this time a significant fraction of the methane is broken down before it even gets to doing any warming. It also depends on where and when the methane is emitted, with methane only lasting around a year if it sticks near the surface but a 1000 years in the cold high atmosphere above the Arctic. And then there’s the warming lag, as once your radiatively forcing gas makes to the right part of the atmosphere it still takes several decades for this to translate into a temperature rise – much like turning the temperature up on an oven doesn’t make it hot straight away – and in the meantime some more of the methane has decayed. Because of this issue using the 20 year methane GWP for short-term impact can be used as a simplified heuristic is OK enough as a rough estimate, as it roughly takes account of the fact that not all of the emitted methane even makes it far enough to cause any warming but doesn’t entirely ignore the short-term impacts like the 100 year GWP. But GWPs are only a rough guidelines anyway, with analysing methane releases in full models the best way of establishing the likely impact of methane releases.


      1. Thank you for the quick reply and explanation! If I’m understand this correctly, it seems like if a large and abrupt (very unlikely as you describe it!) release were to occur, it would initially be an abrupt impact on radiative forcing/energy imbalance (?), but not an abrupt temperature rise?

        Really makes me wonder why certain groups and individuals are so keen on using the initial GWP.


      2. Yes indeed – the analogy I use is that it’s like turning up an oven. Turning up the dial is like increasing radiative forcing, as it controls how much extra energy is entering the oven, but it still takes a little while for the heat to actually build up inside as a result. Whacking up the dial loads would trigger some rapid-ish warming, but it will still take some time to have an effect (and by which time the dial will likely be turned back down somewhat).

        Re. GWP usage, it was initially introduced to enable simple back of the envelope calculations to compare the relative strength of different greenhouse gases (and therefore how much impacts cuts in each may have) for policymakers. But as with all handy simplifications, the devil is in the detail, and the thing that gets left out or confused is the implication of what timescale is picked. Picking GWP over 20y (GWP20) illustrates the power of short-lived gases but over-emphasises their long-term impact, but while using GWP100 illustrates long-term impacts better it under-emphasises their short-term power. It makes it all very easy to spin – for example some vegan pressure groups took to using GWP20 for methane to make it look like animal-agri caused the majority of climate change, but they only used GWP20 for animal-agri methane and left the other two thirds of methane emissions (including fossil fuel methane leaks, landfill, etc.) at GWP100! In the end it’s models that best account for this stuff, and when GWP estimates are used it should be provided with healthy caveats.


    1. Hi Joe, thanks for the link – this is the same study that I replied to Tyler on a few comments up. My short takeaway is that the estimates of the direct CO2 release to the atmosphere by permafrost that I collate in this article still stand, but that this paper shows how there can be more lateral transport of carbon by water directly in to the ocean. What happens to the carbon then is the big unresolved question – some of the organic carbon particles will likely be buried in sediments carried by the same big rivers, while the dissolved carbon will have a complex knock-on effect on ocean carbonate chemistry that would take modelling to unravel and trace the impact of. So it’s important, but doesn’t change everything!


  6. Are wildfires included in estimates of permafrost thaw/release? I’ve read a few papers:





    It seems they do affect thaw, but sometimes it affects vegetation, active layer or the permafrost itself, and it takes several years to reach its peak subsidence. But it seems the permafrost can grow back? I’m just confused. Wildfires always happen and I never see much talk of it when it comes to permafrost thaw/release. So I’m assuming they are incorporated into the models with some variability?

    I guess what I’m wondering is could they affect the estimates of methane feedback at all that you presented here for later in the century (I know obviously they affect CO2 release)?


    1. Mmm, good question (and topical) – as far as I can tell the main review I used for this section doesn’t specifically mention wildfires as components of the constituent models, but I’d have to check the individual models to be sure (a few I’ve checked so far suggests not). My guess would be they either mostly don’t include it yet or only rather simply, but most permafrost models are now in the process of being upgraded and incorporated into the next big climate model round for the next IPCC report.

      The linked studies do seem to converge on wildfires triggering subsidence/thermokarst growth especially in ice-rich permafrost, but that under ‘normal’ conditions the permafrost eventually recovers. [Recovery here specifically refers to the active layer thickness, which is the top layer of soil that unfreezes and refreezes each year and so is where most of the action is – increasing ALT allows more soil organic carbon to thaw and degrade each year. Recovery implies the ALT decreasing again, so the annual thaw reaching less deep into the permafrost. Historically these wildfire cycles of permafrost damage and recovery would’ve most likely balanced out across the Arctic, and so ignorable in simple models.] But they’re all quite local studies and I’m not aware of any work that extrapolates this regionally, so the big question is whether they’ll continue to recover similarly in abnormal times like ours! Ecosystem-Permafrost interactions are notoriously tricky to understand and model (with things increasing emissions often unexpectedly countered by some other component like fertilised above-surface growth), so it’s unsurprisingly confusing. Northerly boreal forest fires could even help preserve permafrost in an odd way, as dark forests boost local temperatures and so their northward expansions acts as a minor positive feedback (& so losing some forest could keep temperatures a bit lower locally) – it gets complicated!

      As for impact on this post’s estimates, assuming that wildfire-permafrost interaction is currently poorly captured in models I would think that overall it would somewhat boost the total emission numbers (by expanding the thawing and slumping affected areas fasted than anticipated) but not so much the CH4/CO2 ratio (as the decay processes remain the same post-fire) – so the 0.5-0.6^oC by 2100 under high emissions estimate might be increased by some uncertain but probably smallish amount. I’ll monitor the literature and update it if it becomes clearer…


    1. These calculations are all for CO2/CH4, so don’t include N2O. It’s true that N2O is itself a potent short-lived greenhouse gas, and recent studies suggest permafrost emits more of it both in general and during thawing than anticipated. However, it’s also not very well studied, has complex interactions with ecosystem changes (more be made, but more may also be broken down by extra bugs & veg), and has a lot of uncertainties on what even the normal baseline emissions are, let alone recent increases. Overall though the global contribution of N2O to warming so far has been fairly well constrained as having a small but not insignificant impact (with a radiative forcing of ~0.17 W/m^2, compared to net total of ~2.3 W/m^2, so about 7% – https://climatetippingpoints.files.wordpress.com/2019/04/ipcc_ar5_rf_aerosol.png?w=1140). Future model projections (e.g. http://dx.doi.org/10.1038/nature10176) suggest that N2O emissions will indeed increase with warming and biosphere feedbacks, which will somewhat reduce the expected net increase in land ecosystem carbon sink. But this is very uncertain and fast-moving research area, so projections may well change!


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s