High Risk of Permafrost Thaw Northern Soils Will Release Huge Amounts of Carbon in a Warmer World, Say Edward A

.... .

Abrupt thaw, as seen here in Alaska's Noatak National Preserve, causes the land to collapse, accelerating permafrost degradation and carbon release. High risk of permafrost thaw Northern soils will release huge amounts of carbon in a warmer world, say Edward A. G. Schuur, Benjamin Abbott and the Permafrost Carbon Network.

rctic temperatures are rising fast, largely because of the realization that organic greenhouse gas with about 25 times more

and permafrost is thawing. Carbon carbon is stored much deeper in frozen soils warming potential than C02 over a 100-year A released into the atmosphere from than was thought. Inventories typically meas period. However, waterlogged environments permafrost soils will accelerate climate ure carbon in the top metre of soil. But the also tend to retain more carbon within the change, but the magnitude of this effect physical mixing during freeze-thaw cycles, in soil. It is not yet understood how these fac remains highly uncertain. Our collective combination with sediment deposition over tors will act together to affect future climate. estimate is that carbon will be released more hundreds and thousands of years, has buried The ability to project how much carbon will quickly than models suggest, and at levels permafrost carbon many metres deep. be released is hampered both by the fact that that are cause for serious concern. The answers to three key questions will models do not account for some important We calculate that permafrost thaw will determine the extent to which the emission processes, and by a lack of data to inform the release the same order of magnitude of of this carbon will affect climate change: How models. For example, most large-scale mod carbon as deforestation if current rates of much is vulnerable to release into the atmos els project that permafrost warming depends deforestation continue. But because these phere? In what form it will be released? And on how much the air is warming above them. emissions include significant quantities of how fast will it be released? These questions This warming then boosts microbial activity methane, the overall effect on climate could are easily framed, but challenging to answer. and carbon release. But this is a simplification. be 2.5 times larger. Abrupt thaw processes can cause ice wedges Recent years have brought reports from KNOWN UNKNOWNS to melt and the ground surface to collapse, 1 5 the far north of tundra fires , the release of As soils defrost, microbes decompose the accelerating the thaw of frozen ground • Evi 2 3 ancient carbon , CH, bubbling out oflakes ancient carbon and release CH4 and carbon dence of rapid thaw is widespread: you can 1 and gigantic stores of frozen soil carbon . dioxide. Not all carbon is equally vulner see it in the 'drunken' trees that tip danger The latest estimate is that some 18.8 mil able to release: some soil carbon is easily ously as a result of ground subsidence, and lion square kilometres of northern soils hold metabolized and transformed to gas, but in collapsed hill slopes marked by scars from about 1,700 billion tonnes of organic car more complex molecules are harder to break landslides. These are just some of the complex bon'- the remains of plants and animals down. The bulk of permafrost carbon will processes that models don't include. that have been accumulating in the soil over be released slowly over decades after thaw, At the same time, few data are available to thousands of years. That is about four times but a smaller fraction could remain within support these models because of the difficul more than all the carbon emitted by human the soil for centuries or longer. The type of ties of gathering data in extreme environ activity in modern times and twice as much gas released also affects the heat-trapping ments. Only a handful of remote field stations as is present in the atmosphere now. potential of the emissions. Waterlogged, around the world are collecting data to sup This soil carbon amount is more than low-oxygen environments are likely to con port this research, even though the perma

three times higher than previous estimates, tain microbes that produce CH1 - a potent frost zone covers about almost one-quarter

32 NATURE VO L 4 80 1 DE C EMBER 2 01 1 ,c; 2011 Macmillan Publ1shers Limited. All rights rese rved of the Northern Hemisphere's land area. carbon loss has happened before: a 7-14% soils, emissions from these soils are unlikely The field studies that do exist confirm that decrease has been measured in soil carbon to overshadow those from the burning of fos permafrost thaw is tightly linked to ground inventories across thousands of sites in the sil fuels, which will continue to be the main subsidence and soil moisture as well as tem temperate-zone United Kingdom as a result source of climate forcing. Permafrost carbon 9 perature. So modelling carbon emissions of climate change • Also, data scaled up release will still be an important amplifier from permafrost thaw is much more complex from a single permafrost field site point to of climate change, however, and is in some than a simple response to temperature alone. a potentialS% loss over a century as a result ways more problematic: it occurs in remote Models have flaws, but experts intimately of widespread permafrost thaw. These field places, far from human influence, and is dis familiar with these landscapes and processes results generally agree with the collective persed across the landscape. Trapping carbon have accumulated knowledge about what carbon-loss projection made by this survey, emissions at the source - as one might do at they expect to happen, based on quantitative so it should indeed be plausible. power plants- is not an option. And once data and qualitative understanding of these Across all the warming scenarios, we pro the soils thaw, emissions are likely to continue systems. We have attempted to quantify this ject that most of the released carbon will be for decades, or even centuries. expertise through a survey developed over in the form of C02, with only about 2. 7% in The scientific community needs to collect several years. the form ofCH1 • However, because CH4 has more data and develop more-sophisticated a higher global-warming potential, almost models to test the hypotheses presented by SURVEY SAYS half the effect of future permafrost -zone car this survey. Fortunately, awareness of the Our survey asks what percentage of the bon emissions on climate forcing is likely to problem is increasing and these are starting surface permafrost is likely to thaw, how be from CH1• That is roughly consistent with to happen. The US Department of Energy, much carbon will be released, and how much the tens ofbillions of tonnes of CH1 thought for example, has initiated a project called of that carbon will be CH1, for three time to have come from oxygen-limited environ Next-Generation Ecosystem Experiments periods and under four warming scenarios ments in northern ecosystems after the end - Arctic, which aims to improve the rep 10 that will be part of the Intergovernmental of the last glacial period • resentation of these processes in large-scale Panel on Climate Change Fifth Assessment All this points towards significant carbon models. NASA is pursuing an Arctic-Boreal Report. The lowest warming scenario pro releases from permafrost-zone soils over Vulnerability Experiment, which aims to jects 1.5 oc Arctic warming over the 198S- policy-relevant timescales. It also highlights improve satellite observations of this region. 2004 average by the year 2040, ramping up to important lags whereby permafrost degra The Vulnerability of Permafrost Carbon 2 oc by 2100; the highest warming scenario dation and carbon emissions are expected Research Coordination Network funded considers 2.5 oc by2040, and 7.S°C by2100. to continue for decades or centuries after by the US National Science Foundation, of In all cases, we posited that the temperature global temperatures stabilize at new, higher which we are part, is bringing together peo would remain steady from 2100 to 2300 so levels. Of course, temperatures might not ple and observations to synthesize results that we could assess opinions about the time reach such high levels. Our group's estimate and validate models. These are just some of lag in the response of permafrost carbon to for carbon release under the lowest warm the many international initiatives aimed at temperature change. ing scenario, although still quite sizeable, is filling these research gaps. The survey was filled out this year by about one-third of that predicted under the In the meantime, our survey outlines the 41 international scientists, listed as authors strongest warming scenario. additional risk to society caused by thawing here, who publish on various aspects of per Knowing how much carbon will be of the frozen north, and underscores the mafrost. The results are striking. Collectively, released from the permafrost zone in this urgent need to reduce atmospheric emis we hypothesize that the high warming sce century and beyond is crucial for determin sions from fossil-fuel use and deforestation. nario will degrade 9-1S% of the top 3 metres ing the appropriate response. But despite the This will help to keep permafrost carbon of permafrost by 2040, increasing to 4 7-61% massive amount of carbon in permafrost frozen in the ground. • by 2100 and 67-79% by 2300 (these ranges are the 9S% confidence intervals around the Edward Schuur is in the Department group's mean estimate). The estimated car ofBiology at the University ofFlorida, bon release from this degradation is 30 billion Gainesville, Florida 32611, USA. Benjamin to 63 billion tonnes of carbon by 2040, reach Abbott is in the Institute ofArctic Biology ing 232 billion to 380 billion tonnes by 2100 at the University ofAlaska, Fairbanks, and S49 billion to 86S billion tonnes by 2300. Alaska 99775, USA. The other experts in the

These values, expressed in C02 equivalents, Permafrost Carbon Research Network combine the effect of carbon released as both are listed at go.nature.com/kkidom.

C02 and as CH1• e-mail: [email protected] Our estimate for the amount of carbon released by 2100 is l.7-S.2 times larger than 1. Mack, M. C. eta/. Nature 475, 489-492 (20 11). those reported in several recent model 2. Schuur, E. A. G. eta/. Nature 459, 556-559 6 8 (2009). ling studies - , all of which used a similar 3. Wa lter, K. M., Zimov, S. A., Chanton, J.P., Verbyla, warming scenario. This reflects, in part, our D. & Chapin, F. S. lll Nature 443,71-75 (2006). perceived importance of the abrupt thaw 4. Tarnocai, C. eta/. Global Biogeochem. Cycles 23, processes, as well as our heightened aware GB2023 (2009). 5. Jorgenson, M. T., Shur, Y. L. & Pullman, E. R. ness of deep carbon pools. Active research Geophys. Res. Lett. 33, L02503 (2006). is aimed at incorporating these main issues, 6. Schaefer, K., Zhang, T., Bruhwiler, L. & Barrett, A. along with others, into models. P. Tel/us B 63, 165-180 (2011). 7. Koven, C. D. eta/. Proc. Nat/ Acad. Sci. USA 108, Are our projected rapid changes to the 14769- 14774 (2011). permafrost soil carbon pool plausible? The 8. Schneider von Deim ling, T. eta/. Biogeosciences survey predicts a 7-11% drop in the size of Discuss. 8, 4727-4761 (2011). the permafrost carbon pool by 2100 under 9. Bellamy, P. H. Loveland, P. J., Brad ley, R. 1., Lark, R. M. & Kirk, G. J. Nature 437, 245-248 (2005). the high-warming scenario. That scale of 'Drunken' trees reveal areas of subsidence. 10.Fischer, H. eta/. Nature 452, 864-867 (2008).