Update- See Correction of an error at bottom
Here at Skeptical Science, there is an ongoing effort to combat disinformation from those who maintain that climate change is a non-issue or non-reality. From time to time, however, individuals or groups overhype the impacts of climate change beyond the realm of plausibility. Some of this is well-intentioned but misguided. For those who advocate climate literacy or for scientists who engage with the public, it is necessary to call out this stuff in the same manner as one would call out a scientist who doesn’t think that the modern CO2 rise is due to human activities.
Many overblown scenarios or catastrophes seem to involve methane in the Arctic in some way. There are even groups out there declaring a planet-wide emergency because of catastrophic, runaway feedbacks, involving the interplay between high latitude methane sources and sea ice.
About a week ago, a Nature article by Gail Whiteman, Chris Hope, and Peter Wadhams came out analyzing the "Vast Costs of Arctic Change." The Whiteman article is an honest and thoughtful commentary about the economic impacts of a changing Arctic climate. I will not comment on their economic modeling here, but rather on a key scenario assumption that they use which calls for vast increases in Arctic-sourced methane to the atmosphere. In this case, they have in mind a very rapid pulse of 50 Gigatons of methane emanating from the East Siberian Shelf (see image, including Laptev and East Siberian sea). Note: 1 GtCH4= 1 Gigaton of methane = 1 billion tons of methane. Whiteman et al. essentially assume that this "extra methane" will be put in the atmosphere on timescales of years or a couple decades. This article has been widely publicized because it calls for an average of 60 trillion dollars on top of all other climate change costs. Since this was discussed in a prediction context rather than as a thought experiment, it demands analysis of evidence.
In this article, I will argue that there is no compelling evidence for any looming methane spike. Other scientists have spoken out against this scenario as well, and I will encompass some of their arguments into this piece. In summary, the reason a huge feedback is unlikely is because of the long timescale required for global warming to reach some of the largest methane hydrate reservoirs (defined later), and because no evidence exists for such an extreme methane concentration sensitivity to climate in the past record. Permafrost feedbacks are of concern, but there is no basis for assuming a dramatic "tipping point" in the atmospheric methane concentration.
Methane (CH4) is a greenhouse gas. It absorbs thermal energy that the Earth is trying to shed into outer space, and can thus warm the surface of the planet. Its concentration in the modern atmosphere is a little bit shy of 2 parts per million by volume (ppm), compared to roughly 0.72 ppm in 1750 or 0.38 ppm in typical glacial conditions. Like CO2, methane has not risen to modern day concentrations during the entirety of the now ~800,000 year long ice core record.
So what about Whiteman's scenario?
For perspective on how big 50 GtCH4 is, I've used data from David Archer's online methane model to see how atmospheric methane concentrations would change in response to such a big carbon injection. You can do this as a back-of-envelope calculation by noting that 1 ppm is about 2.8 GtCH4 if it all stays as methane and isn't removed, but this model lets you see the decay timescale too. For methane, the decay back to original concentrations occurs within decades, whereas for CO2 it takes millennia (CH4 is rapidly oxidized by the hydroxyl radical in the atmosphere). Therefore, CO2 dominates the long-term climate change picture but the methane spike can induce very large transitory effects.
I've run two scenarios in which the 50 GtCH4 injection takes 1 year and 10 years to complete (red and blue lines, respectively). The model starts with pre-industrial CH4 concentrations in years -10 through zero. The modern concentration of methane is shown as a horizontal orange line.
Everything having to do methane in the ice core record resides below the orange line in Figure 1 (at least within the resolution of the cores). So we're potentially talking about a very big change, which the Whiteman article contends is likely to be emitted fairly soon and should have implications for Arctic policy.
For many, the primary concern about “big” abrupt changes in atmospheric CH4 stems from the large quantity of CH4 stored as methane hydrate or in permafrost in the Arctic region. These terms are defined below. It should be noted that globally, wetlands are the largest single methane source to the modern atmosphere. Most of that contribution is from the tropics and not from high latitudes (even if the Arctic was to start pumping harder). The Denman et al., 2007 carbon cycle chapter in the last IPCC report is a useful reference.
Nonetheless, the Arctic is a region that is quite dynamic and is changing rapidly. The high latitudes are currently a CO2 sink and CH4 source in the modern atmosphere, and it’s not implausible that the effectiveness of the sink could diminish (or reverse) or that the methane source could enhance in the future, since we expect a transition to a warmer, wetter climate with an extended thawing season. This makes the carbon budget in the Arctic a “hot” place for research.
In these discussions, it is important to clarify what sort of methane source we're talking about.
Methane hydrate is a solid substance that forms at low temperatures / high pressures in the presence of sufficient methane. It is an ice-like substance of frozen carbon, occurring in deep permafrost soils, marine continental margins, and also in deeper ocean bottom sediments. It's also very concentrated (a cubic foot of methane hydrate contains well over 100 times the same volume of methane gas).
On the decade-to-century timescale, the liberation of methane from the marine hydrate reservoir (or the deep hydrates on land) should be well insulated from anthropogenic climate change. Deep ocean responses by methane are a very slow response (many centuries to millennia, Archer et al., 2009). Methane released in deep water also needs to evacuate the water column and get to the atmosphere in order to have a climate impact, although much of it should get eaten up by micro-organisms before it gets the chance. These issues are discussed in a review paper by O’Connor et al., 2010.
There’s also carbon in near-surface permafrost, which is the more vulnerable carbon pool during this century. Permafrost is frozen soil (perennial sub-0°C ground), and can also encompass the sub-sea permafrost on the shelves of the Arctic Ocean. This includes the eastern Siberian shelf, a very shallow shelf region (only ~10-20 m deep, and very broad, extending a distance of 400– 800 km from the shoreline). This is a bit of a special case. These subsea deposits formed during glacial times, when sea levels were lower and the modern-day seafloor was instead exposed to the cold atmosphere. The ground then became submerged as sea levels rose (going into the warmer Holocene). The rising seas have been warming the deposits for thousands of years. Because of their exposure during the Last Glacial Maximum, the shelves may be almost entirely underlain by permafrost from the coastline all the way down to a water depth of tens or even a hundred meters (e.g., Rachold et al., 2007 and this USGS page).
There's actually no good evidence of shallow hydrate on the Siberian shelves, even though there are substantial quantities of subsea permafrost. Hydrate may exist deeper down however, more than 50 meters below the seafloor. The stability of these hydrates is sustained by the existence of permafrost, and it's not quite clear to what extent hydrate can also be stored within the permafrost layer.
The estimates of the amount of methane in these various Arctic reservoirs are very uncertain. Ballpark numbers are a couple thousand gigatons of carbon (GtC) stored in hydrates in global marine sediments (e.g., Archer et al., 2009) of which a couple hundred gigatons of carbon are in the Arctic Ocean basin, and between 1000-2000 GtC in permafrost soil carbon stocks (e.g., Tarnocai et al., 2009) after you include the deeper deposits. For comparison, there is a bit over 800 GtC in the atmosphere, of which about 5 Gt is in the form of methane, and estimated ~5000 GtC in the remaining fossil fuel reserve. These numbers seem big compared to the atmosphere, but for methane direct comparison isn't too relevant unless you put it in rapidly, since it has such a short lifetime in the atmosphere. Large amounts of CO2, in contrast, last much longer.
A couple years ago, Shakhova et al. (2010a) reported extensive methane venting in the eastern Siberian shelf and suggested that the subsea permafrost could become unstable in a future warmer Arctic. Shakhova et al (2010b) cite ~1400 Gt in the East Siberian Arctic Shelf, which comprises ~25% of the Arctic continental shelf and most of the subsea permafrost. Shakhova et al (2010c) ran through a few different pathways in which they argued for 50 GtCH4 release to the atmosphere either in a 1-5 year belch or over a 50-yr smooth emission growth, which they suggest, “significantly increases the probability of a climate catastrophe.” This assessment was the foundation for the concern in the recent Whiteman Nature article, linked at the top.
The physical mechanism outlined by some of these authors is related to the rapid reduction in Arctic summer sea ice observed over the last few decades, which allows for greater amounts of solar radiation to penetrate the waters around the Arctic shelf. Warming water propagates down in the well-mixed layers tens of meters to the seabed, and might melt frozen sediments underneath. Because the shelf in this region is shallow (compared to other regions), one doesn't need to wait a long time for the seafloor to feel the atmosphere-surface forcing, and methane leakage might have an easier escape path to the atmosphere. Allegedly, this has been leading to an acceleration of methane flux.
As a response to the first paper from Shakhova on enhanced methane fluxes, Petrenko et al (2010) criticized the authors for misunderstanding several of their references and primarily for the logical implications of their conclusions. For example,
“A newly discovered CH4 source is not necessarily a changing source, much less a source that is changing in response to Arctic warming. Shakhova et al. do acknowledge these distinctions, but in these times of enhanced scrutiny of climate change science, it is important to communicate all evidence to the scientific community and the public clearly and accurately”
Another paper, Dmitrenko et al (2011) reinforced this statement and came to the conclusion that there is currently no evidence that Arctic shelf hydrate emissions have increased due to global warming. This is also discussed in the review article by O'Connor et al (2010, linked above).
The work done by the Dmitrenko paper shows that although the changing Arctic atmosphere has led to warmer temperatures throughout the water column (over the eastern Siberian shelf coastal zone), it takes a very long time for the permafrost feedback at the bed to respond to this signal. They noted that the deepening of the permafrost table should only have been on the order of 1 meter over the last several decades, which does not permit a rapid destabilization of methane hydrate.
It is important to emphasize that simple point source emission estimates are not often suitable for determining changed sources and sinks over the last few decades, and thus don't tell you how that translates into atmospheric concentration. This should be kept in mind when seeing dramatic videos of methane venting from a shelf or exploding lake, which might not actually have much to do with global warming.
In 2008, there was a comprehensive report on Abrupt Climate Change from the U.S. Climate Change Science Program, which is a bit dated but nonetheless makes a statement reflecting most of current scientific thinking. Quoting Ch. 5 Brook et al (2008):
"Destabilization of hydrates in permafrost by global warming is unlikely over the next few centuries (Harvey and Huang, 1995). No mechanisms have been proposed for the abrupt release of significant quantities of methane from terrestrial hydrates (Archer, 2007). Slow and perhaps sustained release from permafrost regions may occur over decades to centuries from mining extraction of methane from terrestrial hydrates in the Arctic (Boswell, 2007), over decades to centuries from continued erosion of coastal permafrost in Eurasia (Shakova [sic] et al., 2005), and over centuries to millennia from the propagation of any warming 100 to 1,000 meters down into permafrost hydrates (Harvey and Huang, 1995)"
One of the primary reasons we don't think there's as much methane sensitivity to warming as has been proposed by Shakhova, and argued for in the Whiteman Nature article, is because there's no evidence for it in the paleoclimate record. This has been a point made by Gavin Schmidt on Twitter (a compilation of his many tweets on the topic here) but the objections to the Nature assumptions have been further echoed in recent days by other scientists working on the Arctic methane issue (e.g., here, here).
One can argue from a process-based and observations-based approach that we don't understand everything about Arctic methane feedback dynamics, which is fair. Nonetheless, the methane changes on the scale being argued by Whiteman et al. should have been seen in the early Holocene (when Summer Northern Hemispheric solar radiation was about 40 W/m2 higher than today at 60 degrees North, 7000-9000 years ago). Even larger anomalies occurred during the Last Interglacial period between 130,000 to 120,000 years ago, though with complicated regional evolution (Bakker et al., 2013).
Both of these times were marked by warmer Arctic regions in summer without a methane spike. It's also known pretty well (see here) that summertime Arctic sea ice was probably reduced in extent or seasonally free compared to the modern during the early Holocene, offering a suitable test case for the hypothesis of rapid, looming methane release.
It should be noted that Peter Wadhams did offer a response recently to the criticisms of the Whitehead Nature piece (Wadham is a co-author) but did not address why this idea has not been borne out paleoclimatically.
Yesterday, an objection to the paleoclimate comparison cropped up in the Guardian suggesting that the early Holocene or Last Interglacial analogs are not suitable pieces of evidence against rapid methane release. They aren't perfect analogs, but the argument does not seem compelling. The Northeast Siberian shelf regions have been exposed many times to the atmosphere during the Pleistocene when sea levels were lower (and not covered by an ice sheet since at least the Late Saalian, before 130,000 years ago, e.g., here). As mentioned before, when areas such as the Laptev shelf and adjacent lowlands were exposed, ice-rich permafrost sediments were deposited. The deposits become degraded after they are submerged (when sea levels increase again), resulting in local flooding and seabed temperature changes an order of magnitude greater than what is currently happening. Moreover, the permafrost responses have a lag time and are still responding to early Holocene forcing (some overviews in e.g., Romanovskii and Hubberten, 2001; Romanovskii et al., 2004; Nicolsky et al., 2012). A book chapter by Overduin et al., 2007 overviews the history of this region since the Last Glacial Maximum. These texts also suggest that large amounts of submarine permafrost may have existed going back at least 400,000 years. It therefore does not seem likely that the seafloor deposits will be exposed to anything in the coming decades that they haven't seen before.
What about other times in the past? Fairly fast methane changes did occur during the abrupt climate change events embedded within the last deglaciation (e.g., Younger Dryas), just before the Holocene when the climate was still fluctuating around a state colder than today. These CH4 changes were slower than the abrupt climate changes themselves, and have been largely attributed to tropical and boreal wetland responses rather than high latitude hydrate anomalies. Marine hydrate destabilization as a major driver of glacial-interglacial CH4 variations has also been ruled out through the inter-hemispheric gradient in methane and hydrogen isotopes (e.g., Sowers, 2006)
To be fair, we don't have good atmospheric methane estimates during warmer climates that prevailed beyond the ice core record, going back tens of millions of years. Methane is brought up a lot in the context of the Paleocene-Eocene Thermal Maximum (PETM, 55 million years ago). During this time, proxy records show global warming at the PETM (similar to what modern models would give for a quadrupling of CO2), extending to the deep ocean and lasting for thousands of years. In addition, there were substantial amounts of carbon released. It may very well be that isotopically light carbon came from a release of some 3,000 GtC of land-based organic carbon, rather than a destabilization of methane hydrates, although this is a topic of debate and ongoing research (see e.g., Zeebe et al., 2009; Dickens et al., 2011).
It's also important to emphasize that any destabilization of oceanic methane hydrates at the PETM, or any other time period, would imply that the carbon release is a feedback to some ocean warming that occurred first- perhaps on the order of 1000 years beforehand. Furthermore, once methane was in the atmosphere, it would oxidize to CO2 on timescales significantly shorter than the PETM itself (decades.) Unfortunately, there is no bullet-proof answer right now for what caused the PETM, but rather several hypotheses that are consistent with proxy interpretation. However, methane cannot be the only story.
To be clear, CH4 is important as we go forward, and is already a key climate forcing agent behind CO2 (coming in at ~0.5 W/m2 radiative forcing since pre-industrial times). Additionally, methane is quite reactive in the atmosphere, and the effect of other things like tropospheric ozone, aerosols, or stratospheric water vapor are partly slaved to whatever is happening to methane (Shindell et al., 2009). This means methane emitted has a bigger collective impact on climate than if you just do the radiative forcing calculation by comparing methane concentration changes to what it was in 1750.
Permafrost thawing is also going to be important in the coming century (this is a good paper), and the uncertainties pretty much go one way on this. There's not much wiggle room to argue that permafrost will reduce CH4/CO2 concentrations in the future. This is also likely to be a sustained release rather than one big catastrophic event. For example, permafrost was not included in Lenton (2008) as a "tipping point" for precisely the reason that there's no evidence for any "switch" of rapid behavior change. Much of the carbon is also likely to be in the form of CO2 to the atmosphere, and even implausible thought experiments of catastrophic methane release (see David Archer's post at RealClimate) give you comparable results in the short-term as to what CO2 is going to do for a long time.
The observed methane venting from the East Siberian shelf sea-floor to the atmosphere is probably not a new component of the Arctic methane budget. Furthermore, warming of the Arctic waters and sea ice decline will likely impact subsea permafrost on longer timescales, rather than the short term.
Methane feedbacks in the Arctic are going to be important for future climate change, just like the direct emissions from humans. This includes substantial regions of shallow permafrost in the Arctic, which is already going appreciable change. Much larger changes involving hydrate may be important longer-term. Nonetheless, these feedbacks need to be kept in context and should be thought of as one of the many other carbon cycle feedbacks, and dynamic responses, that supplement the increasing anthropogenic CO2 burden to the atmosphere. There is no evidence that methane will run out of control and initiate any sudden, catastrophic effects. There's certainly no runaway greenhouse. Instead, chronic methane releases will supplement the primary role of CO2. Eventually some of this methane oxidizes into CO2, so if the injection is large enough, it can add extra CO2 forcing onto the very long term evolution of global climate, over hundreds to thousands of years.
Errata Update: Gavin Schmidt let me know that in the first version of this post, I used gigatons of carbon instead of gigatons of methane. I mistakingly read the Shakhova paper as an injection of carbon. Since the molecular weight of carbon is 12 g/mol, and CH4 is 16 g/mol, then 1 GtC=1.33 GtCH4. The figure in the post has been revised accordingly and doesn't impact the argument here.
Posted by Chris Colose on Thursday, 1 August, 2013
 |
The Skeptical Science website by Skeptical Science is licensed under a Creative Commons Attribution 3.0 Unported License. |