Carbon Feedback From Thawing Permafrost Will Likely Add 0.4°F – 1.5°F To Total Global Warming By 2100

Posted on  

"Carbon Feedback From Thawing Permafrost Will Likely Add 0.4°F – 1.5°F To Total Global Warming By 2100"

  • Thawing permafrost will release carbon to the atmosphere that will have an appreciable additional effect on climate change, adding at least one quarter of a degree Celsius by the end of the century and perhaps nearly as much as one degree (about 1.5°F).
  • The permafrost feedback response to our historic emissions, even in the absence of future human emissions, is likely to be self-sustaining and will cancel out future natural carbon sinks in the oceans and biosphere over the next two centuries.
  • Unfortunately, there are several good reasons to consider the outlook in this study as rosy — as the authors themselves make clear. However, as bad and inevitable as they are, feedbacks from the permafrost are just the (de-)frosting on the fossil fuel cake that we are busy baking. It is still up to us to influence how severe climate change is going to be.

by Andy Skuce, via Skeptical Science

Many papers have looked at the expected contribution of thawing permafrost to climate change. For example, Schaeffer et al. (2011) and Schuur and Abbott (2011) have both published estimates of the effect that the thawing and decomposition of organic matter in Arctic soils will have on future climates. Aspects that these studies neglected were the feedback that the permafrost carbon release would have on causing further permafrost degradation and the varying response that the carbon release would have on the climate in different emission scenarios and for a range of climate sensitivities.

To explore this matter further, a recent paper in Nature Geoscience (paywalled) by Andrew MacDougall, Christopher Avis and Andrew Weaver couples together climate and carbon-cycle models. Using the University of Victoria Earth System Climate Model adapted to include a permafrost response module, the researchers calculated the contribution to climate warming of thawing permafrost over a range of varying parameters.

Figure 1. Taken from MacDougall et al. (2012) showing the additional warming induced by permafrost thawing for four diagnosed emissions pathways (DEP, see text below for explanation). The coloured areas are the ranges of likely additional temperature ranges and the black lines show the median responses. The uncertainty within each DEP run results from uncertainties in the density of carbon in the permafrost and the climate sensitivity (the temperature effect of a given rise in carbon dioxide concentration in the air). Figure with original caption here.

Emissions pathways

The four scenarios or diagnosed emissions pathways (DEP) were derived from the Representative Concentration Pathways (RCP) used for the upcoming Fifth Assessment Report of the IPCC. RCPs are not emissions scenarios but rather curves of atmospheric CO2 concentration with time. The numbers 2.6, 4.5, 6.0 and 8.5, correspond to the radiative forcing in W/m2 in 2100. For use in climate models the emissions scenarios have to be backed out of the RCPs to give the DEPs.

Figure 2. The DEP’s used in the MacDougall et al study. The figure corresponds to Figure S2 in their Supplementary Material. The top graph shows the annual human CO2 emissions pathways and the lower graph the cumulative emissions. The jitter in the top curves is due to variations in solar cycles. Note that the most optimistic curve DEP 2.6 has emissions that drop to zero from 2020 to 2070 and, following that, net human sequestration of atmospheric carbon. The most pessimistic pathway has emissions that continue on their recent trajectory until 2070, when they level off. Roughly speaking, the pathways can be categorized as Heroic (2.5); Prudent (4.5); Procrastinating (6.0); and Oblivious (8.5).

Other variables of the model

a) Size of the carbon pool. MacDougall et al vary the carbon concentration of the permafrost from 15.8 to 26.3 kg/cubic metre, which leads to a range of 837-1206 PgC (billion tonnes of carbon) for the upper 3.5 metres of permafrost. This figure is larger than that assumed in previous studies. Consequently, the upper-end range of carbon release by 2100 calculated in this paper, 508 PgC, is much larger than the previous studies that did not use coupled carbon-climate models. For example, Schaeffer et al calculated 138 PgC as their upper limit. The entire permafrost soil carbon pool is estimated to contain 1700 PgC, about twice the mass of the carbon in the current atmosphere.

b) Climate sensitivity. The other important variable used in the model is the climate sensitivity, which varies between 2.0 and 4.5°C for a doubling of atmospheric CO2, with a central value of 3.0°C.

Counter-intuitive results

The results shown in Figure 1 are very consistent for the 2100 median forecasts, ranging between 0.23 and 0.27°C of extra warming due to permafrost feedbacks. By 2300, the range of median has increased to 0.37 to 0.73°C. The results are a little counter-intuitive, with the biggest feedbacks in temperature terms not coming with the worst emissions pathways. On one hand, the amount of permafrost carbon released on the most pessimistic pathways is larger (39, 58, 67 and 101 ppm CO2 respectively for the four DEP’s), and occurs sooner. On the other, in the lower DEP cases, the permafrost carbon released to the atmosphere has a stronger effect due to less saturation of the infra-red absorption bands.

Also counter intuitively, the uncertainty of the permafrost feedback is reduced in the higher emission scenarios. In these scenarios, there is less uncertainty that we will unleash all of the permafrost, and quickly. Not only that, but following the pessimistic pathways means that we will have already have inflicted so much violence on the atmospheric chemistry that the extra damage caused by the permafrost feedback will be felt relatively less strongly.

The industrial shutdown experiment

The results in Figure 1 imply that we are in for at least some backlash from the permafrost whatever emissions pathway we choose to follow. Because the real world will resemble a coupled climate-carbon cycle model, the releases of permafrost carbon will result in more warming, which will cause still more carbon to be released. In other words, at some level the cycle will become self-sustaining, even without any further human encouragement. MacDougall et al investigated this by imagining a complete shutdown of human emissions in 2013 and in 2050.

Figure 3. Showing the atmospheric concentration of CO2 following a shutdown of human emissions in 2013(left) and, after following DEP 8.5 for 39 years, a shutdown in 2050 (right). The dotted blue line shows the results at a climate sensitivity of 3.0°C and the upper and lower lines 4.5° and 2.0° respectively. Selected and modified from Figure S8 in the Supplementary Information.

What Figure 3 demonstrates is that even in the case where human emissions cease altogether, the atmospheric CO2 concentration will remain more-or-less constant in the most likely climate sensitivity case. Essentially, for a climate sensitivity of 3°C, the self-sustaining permafrost-related emissions match the uptake of CO2 from the oceans and the terrestrial biosphere. Note that a self-sustaining feedback is not the same thing as a runaway greenhouse effect.

Permafrost feedbacks and the trillionth tonne

Allen et al (2009) argued that what matters most in determining safe levels of emissions is not so much the rate but the cumulative amount. Framed in this way, total emissions of a trillion tonnes of carbon will lead to a most likely warming of 2°C, a somewhat arbitrary, but widely accepted limit on the amount of warming that the world can endure without a high risk of catastrophic consequences. According to the trillionthtonne.org website, humans have already emitted 561 billion tonnes of carbon from fossil fuels, cement production and land use changes. We therefore have about 440 billion tonnes left to emit before we cross the trillion tonne limit. On present emission trends, this will occur in 2043.

MacDougall et al predict median permafrost emissions of 174 billion tonnes of carbon by 2100, which are more-or-less independent of the future pathway of our other emissions. Such feedbacks were not factored in to the trillionth tonne analysis, or were assumed to kick in after the “safe” limit of 2°C had been breached. Accordingly, the 440 billion tonnes we can still emit over the next few years and stay under the trillion-tonne limit needs to be reduced from 440 to 270 billion tonnes of fossil fuel, cement and land use emissions, a reduction of some 40%. Avoiding dangerous climate change becomes a lot harder once we face up to the permafrost feedback.

Why even this bleak prospect may be optimistic

Alert readers may have already noticed that this article has not yet used the word “methane”. When organic matter in the permafrost is thawed and decomposes it produces mostly CO2 but also small amounts of methane, particularly so in the wetlands that are prevalent in areas of thawing permafrost. Schuur and Abbott (2011) polled 41 experts on permafrost decay who estimated that about 3% of the carbon released from the permafrost will be in the form of methane. Methane has a restricted lifetime in the atmosphere, measured in decades, but while present in the air it has a greenhouse effect some 25 times that of CO2 over a 100-year period and higher values over shorter periods. According to Schuur and Abbott, the small amount of methane is responsible for approximately half of the warming effect from the permafrost emissions.

The UVic model does not simulate methanogenesis. That is to say that it does not model the generation of methane—all of the permafrost carbon that goes into the atmosphere in the model is in the form of CO2. This is a significantly conservative simplification over the time period studied.

Also, their model assumes only purely thermal degradation of the permafrost. Physical erosion, for example at coastlines, is not considered. Their model accounts only for permafrost down to a depth of 3.5 metres and there is plenty of carbon stored below those depths that was excluded from their modelling.

Finally, this study does not consider any contribution of methane from methane hydrates, either from under permafrost or under ice sheets, nor from fossil methane currently trapped under an impermeable seal of continuous permafrost.

Summing-up

  • Thawing permafrost will release carbon to the atmosphere that will have an appreciable additional effect on climate change, adding at least one quarter of a degree Celsius by the end of the century and perhaps as much as one degree. (In comparison, Swart and Weaver (2012) calculated that combustion of the in-place resources of the Alberta oil sands would increase temperatures by 0.24-0.50°C.)
  • The temperature effect of the coming permafrost feedback is not sensitive to the emission pathway that we choose to follow.
  • The permafrost feedback response to our historic emissions, even in the absence of future human emissions, is likely to be self-sustaining and will cancel out future natural carbon sinks in the oceans and biosphere over the next two centuries.
  • Unfortunately, there are several good reasons to consider the outlook in MacDougall et al. as rosy; as the authors themselves make clear. However, as bad and inevitable as they are, feedbacks from the permafrost are just the (de-)icing on the fossil fuel cake that we are busy baking. It is still up to us to influence how severe climate change is going to be.

This piece was originally published at Skeptical Science and was reprinted with permission.

Related Posts:

« »

17 Responses to Carbon Feedback From Thawing Permafrost Will Likely Add 0.4°F – 1.5°F To Total Global Warming By 2100

  1. Alex Smith says:

    To dig deeper into the permafrost story, you can hear an interview with Professor Antoni Lewkowicz. Plus a speech by Dr. Charles Koven from the American Academy for the Advancement of Science special presentation “What If the Permafrost Thaws” February in Vancouver, Canada. Here is the one hour Radio Ecoshock special on permafrost as an mp3.
    http://www.ecoshock.net/eshock12/ES_120606_Show_LoFi.mp3

    More details and links in the blog for that program here.
    http://www.ecoshock.info/2012/05/what-if-permafrost-thaws.html

  2. Tami Kennedy says:

    I found the Industrial Shutdown experiment to be very troubling. Using the 2013 shutdown with the optimistic 2 degree sensitivity the permafrost on its own reduces the rate of CO2 decline such that Bill McKibben’s 350 ppm number is ~ the year 2300. If the current level of CO2 is resulting in taking 100 year weather events and turning them into yearly affairs we have some adjusting to do. And nobody is talking about shutting off population growth. Especially if the UN guesstimate is 10 billion stable by 2100.

  3. Mike Roddy says:

    As noted above, this doesn’t even factor methane releases from the Arctic Shelf and Siberian lakes.

    There is another situation that we need to be aware of, even though the science is incomplete. North American conifer forests are some of the largest carbon sinks on the planet. They evolved under cold winter and cloudy conditions. We have already seen bark beetle outbreaks, but massive tree deaths remain a distinct possibility, something not covered in the literature since temps must rise further before things spiral out of control- similar to in the Amazon.

    Tree death, including fires, will release carbon and heat, and alter precipitation patterns, creating further feedbacks. Places like the Sierras and lower Rockies are especially vulnerable, but this will hit BC and the Northwest in some form. All of these forests are weakened from industrial logging, further complicating matters.

    • Jack Burton says:

      I thought we are already seeing massive tree deaths in Alaska, Canada and in parts of the Western USA.
      I live in the upper Midwest boreal forest bordering Canada. Indeed this forest survives well in cold and cloudy and rainy conditions. That has been the traditional weather pattern up here and the forests have been very healthy.
      But in the last decade or so this weather has become rare. We now see hot dry summers and much longer warmer falls. In the past 5-6 years we have seen the beginning of severe fires, drought death and violent thunderstorm blow downs of huge areas of trees.
      The forest is under great stress.

      • Mike Roddy says:

        You’ve got a point, trees are dying already, but I guess it depends on the definition of “massive”.

  4. KELLI2L says:

    ? Would the Permafrost be melting if there was no drilling, etc., going on ?

    • David B. Benson says:

      Yes.

      • Rob Benjamin says:

        Reply to KELL12L: “? Would the Permafrost be melting if there was no drilling, etc., going on ?”

        Yes, if we continue doing the same nothing we are doing now to stop it. No, if we made a global effort to capture and use atmospheric methane instead of drilling for it.

  5. When complex systems fail, they fail suddenly and catastrophically. When a human’s kidneys fail, the damage isn’t limited to that single crucial subsystem but extends to the entire body. Other subsystems that depend on a detoxified blood supply also fail, and those failures further cascade to other subsystems.

    The ice cap is one of Earth’s vital organs. Its removal is likely to cause cascading effects that even good work like the MacDougall paper in this post can’t anticipate. We simply haven’t seen anything like it in detail at the planetary scale.

    I’m sometimes reduced to hoping that there’s a negative feedback lurking in our ecosystem that hasn’t been identified, one that will counter the obvious and distressing positive feedbacks.

    It seems to me that saving the ice cap is a potent, tangible goal. It’s much more intuitive than 350 ppm (though there’s a definite link). It seems that scientists could explain the consequences well enough that even James Inhofe would have to say, well, I hate to admit it, but I can see that’s right. Or others would see it, reject any further dithering, and move on to prevention.

    It’s like getting a diagnosis of kidney disease. We can’t see our kidneys, or often feel them until we’re really sick, but we understand that we can’t function normally without them.

    I’m deeply skeptical that the planetary equivalent of dialysis–geoengineering–is the answer. But no sane person would elect to continue behavior that destroyed his kidneys because he counted on dialysis. Prevention is only sane response.

    Save the ice cap.

    • David Goldstein says:

      I have created an 18 minute presentation (PowerPoint and script) interweaving my personal story of liver disease and transplant with the story of climate change. I have begun offering at high schools and colleges. Much of what you mention in your metaphor of kidney disease is part of my presentation. My personal health story has been pretty dramatic and I am hoping my storytelling method may engage folks where ‘facts and figures’ fail to do so. I will gladly send you a copy if you are interested. Email me at dagold56@hotmail.com

    • “It seems to me that saving the ice cap is a potent, tangible goal. It’s much more intuitive than 350 ppm (though there’s a definite link).”

      Good point from a climate change communications perspective. Also, aside from permafrost and so on, the loss of Arctic albedo is a calamity in and of itself. Whadham has calculated that the increased heat gained from the recently exposed Arctic Ocean is equal to the forcings of 24 years of human carbon output.

      Wadham wants to geoengineer the Arctic which is 1) not going happen because of international politics driven by all the resources lying beneath the ice and, 2) is fraught with unintended consequences.

      But just getting the idea -SAVE THE ARCTIC ICE- out there — I’m getting a bumper sticker — could be an effective meme.

    • Brooks Bridges says:

      Far too late to save the Arctic icecap. A recent post on this site: “An Illustrated Guide To 2012 Record Arctic Sea Ice Melt” contains a link to another site, (search for: “Wipneus”).

      The second graph on this site shows exponential fits to minimum Arctic Sea Ice Volume for various months of the year. Obviously there are error bands but the curves themselves show that by 2021 we have a strong possibility of 7 months of ice free Arctic.

      • You’re probably right, but the outcome isn’t 100% certain yet. The ice could stabilize at a new diminished level for a while, and, like the polar bears, serve as a rallying point for people.

        More likely the ice is gone, so the increased heat gain from the lost Albedo will roughly double in a decade, hastening the permafrost breakdown and so on.

        But we’ve got to try to draw a line somewhere. Were already past 350 ppm, so what’s next?

  6. john atcheson says:

    Clathrates contain as much or more carbon (as methane) as the permafrost does, and if the permafrost melts the clathrates are likely to also. So you can effectively double these numbers to .8 to 3 degrees F.

    We really need to do a better job of modeling systems, not isolated components of the entire system. Our failure to do so is one reason we keep having headlines like Mann’s about being surprised at how ice melting, sea level rise, warming etc is proceeding faster …

    Of course it is. The world is a system. Take Krill … they sequester a lot of carbon because of their diurnal habits .. as their population plummets, their contribution to sequestering carbon does too. There are some 18 of these kind of major feedbacks. We need to know what waring looks like when ALL of them are operating.

    We’d eliminated a lot of these head-exploding “surprises” about the speed of warming if we did.

  7. fj says:

    As described many times by scientists that this is just one experiment that is better not done and — on top of what’s happened and continues to accelerate — the likelihood of something going terribly wrong bares the specter of inevitable and humanity must smarten up real fast.

  8. Giving this article a more thorough second read, I believe it is way too optimistic.

    In the author’s own words:

    “Their model accounts only for permafrost down to a depth of 3.5 metres and there is plenty of carbon stored below those depths that was excluded from their modelling. Finally, this study does not consider any contribution of methane from methane hydrates, either from under permafrost or under ice sheets, nor from fossil methane currently trapped under an impermeable seal of continuous permafrost.”

    Some of the permafrost is a mile or more thick. I don’t know how deep the authors expect the melting to go by 2100, but 3.5 meters seems like a conservative minimum. How much are they accounting for Arctic amplification?

    Second, at this point in the game, I don’t think the relevant figure for methane is its GG potency over a 100 year horizon, but rather over a 20 year horizon during which it is 75-100 times more potent than CO2. I say 20 years because the deadlines that we’re working against are in the range of 20 years or less. If we don’t get our act together within 20 years at the outside, those who remain on the planet in 2100 will be living in caves.

    Then there is the great gain in warming coming from the loss of the Arctic ice albedo. Wadham has calculated a gain of about 24 years worth of human CO2 emissions already, and that will only increase as the ice shrinks. Although that warming will have global impacts, much of it will be expressed locally — that is, above the Arctic circle, right in the permafrost regions.

    There is also the matter of tundra fires as the region dries out and warms up. Will these fires become more extensive? Will they melt some of permafrost? (Will their soot land on the region’s ice and hasten its decline?).

    Mathematically challenged as I am, it seems to me that if we look at the whole system, rather than just anthropogenic CO2 as the driver of permafrost melt and climate change, that the overall “permafrost carbonization effect” will be at least twice as much (by 2100) and come down about three times as fast as the authors predict.

    With luck, everybody will be able to laugh at my predictions in 20 years.