The magnitude and feedbacks of future methane release from the Arctic region are unknown. Despite limited documentation of potential future releases associated with thawing permafrost and degassing methane hydrates, the large potential for future methane releases calls for improved understanding of the interaction of a changing climate with processes in the Arctic and chemical feedbacks in the atmosphere. Here we apply a “state of the art” atmospheric chemistry transport model to show that large emissions of CH4 would likely have an unexpectedly large impact on the chemical composition of the atmosphere and on radiative forcing (RF). The indirect contribution to RF of additional methane emission is particularly important. It is shown that if global methane emissions were to increase by factors of 2.5 and 5.2 above current emissions, the indirect contributions to RF would be about 250% and 400%, respectively, of the RF that can be attributed to directly emitted methane alone. Assuming several hypothetical scenarios of CH4 release associated with permafrost thaw, shallow marine hydrate degassing, and submarine landslides, we find a strong positive feedback on RF through atmospheric chemistry. In particular, the impact of CH4 is enhanced through increase of its lifetime, and of atmospheric abundances of ozone, stratospheric water vapor, and CO2 as a result of atmospheric chemical processes. Despite uncertainties in emission scenarios, our results provide a better understanding of the feedbacks in the atmospheric chemistry that would amplify climate warming.
 Methane (CH4) is an important greenhouse gas with a radiative forcing (RF) of 0.48 Wm−2, due to anthropogenic activity since preindustrial time [Forster et al., 2007], being second only to CO2 among the anthropogenic greenhouse gases. Its distribution and growth are well documented [Forster et al., 2007; Ramaswamy et al., 2001] showing a significant increase in atmospheric concentrations since preindustrial times. Analyses of ice core data for the last 650,000 years show that atmospheric CH4 concentrations varied from approximately 400 ppb during glacial periods to approximately 700 ppb during interglacial periods. The tropospheric average concentration is currently about 1,800 ppb, representing an approximate 2.5 increase since preindustrial time. The atmospheric concentrations in 2005 correspond to an atmospheric burden of 4,900 Tg CH4 (1 Tg = 1012 g). Observations since 1984, for which there are continuous measurements, show an increase in atmospheric abundances of CH4 by about 10%. Growth rates have decreased significantly since the early 1990s, but with pronounced interannual variations [Rigby et al., 2008].
 CH4 is a chemically active greenhouse gas, and its long-term distribution is affected by changes in both emissions and the atmospheric oxidation rate. Estimates of current global CH4 emissions are in the range 500 to 600 Tg CH4 yr−1 [Denman et al., 2007].
 Atmospheric CH4 is removed through oxidation by the hydroxyl radical (OH), mainly in the troposphere:
 A small fraction is also removed by surface deposition. In the stratosphere, where water vapor is in the range of only a few ppm, CH4 oxidation contributes to water vapor buildup. Since reaction (R1) also represents a significant loss path for OH, additional CH4 emission will suppress OH and thereby increase the CH4 lifetime, implying further increases in atmospheric CH4 concentrations [Isaksen and Hov, 1987; Prather et al., 2001]. This represents a positive chemical feedback, with a feedback factor estimated to be about 1.4 (uncertainty range 1.3 to 1.7) for current atmospheric conditions [Prather et al., 2001]. The nonlinearity in the chemical system could result in a significantly enhanced feedback factor for large CH4 emissions causing large perturbations [Isaksen, 1988].
 The growth in atmospheric CH4 concentrations, a result of an imbalance between sources and sinks, observed over several decades, is reduced markedly during recent years [Dlugokencky et al., 2003]. For instance, estimates reported in IPCC AR4 give close to a balance (about 1 Tg CH4 yr−1 difference between emissions and loss) for the time period 2000 to 2005 [Denman et al., 2007], compared to 33 Tg CH4 yr−1 during the 1980s. Although there are no clear indications whether the changes in trends are due to stabilization in emissions or enhanced atmospheric loss through increased OH levels [Prather et al., 2001], Dlugokencky et al.  point to a possible stabilization of emissions in recent decades. This could be a result of changes in emissions from anthropogenic sources and from wetlands [Bousquet et al., 2006]. However, analyses of atmospheric CH4 changes in 2007 indicate a renewed growth of CH4 [Rigby et al., 2008].
 Increased CH4 emissions affect climate in several ways: Directly through increased CH4 concentrations and indirectly through the chemical feedback on CH4 levels and through production of O3 and stratospheric H2O. Furthermore, CO2 will increase since it is the end product of atmospheric CH4 oxidation. In the current atmosphere the indirect RF is approximately the same as the direct methane RF, taking into account the effect on its own lifetime, on ozone, and on stratospheric water vapor [Forster et al., 2007].
 In this study we use a global Chemical Transport Model (CTM), the Oslo CTM2 [Isaksen et al., 2005; Søvde et al., 2008], to estimate the impact of additional CH4 emissions on the atmospheric concentrations of the climate gases CH4, O3, stratospheric H2O, and CO2, and on RF from these forcing agents. The study covers a wide range of hypothetical methane emission scenarios, up to about 5 times the current emission rate. Although there is no evidence supporting the higher emission in this range, we include them in order to demonstrate the particularly strong positive feedback in the chemical system from large methane releases and the general impact on atmospheric composition and on climate forcing.
 We further consider two potentially important natural sources of atmospheric CH4 in relation to the adopted scenarios: (1) Conversion of organic carbon to CH4, and release when permafrost thaws; and (2) release of CH4 hydrates in marine sediments. Earlier studies have demonstrated that large releases of CH4 from natural sources during warming events can have significant impacts on atmospheric CH4 levels and may have potential synergistic effects leading to increased and/or sustained global warming. Observed surface temperatures in recent years show significant warming, indicating Arctic warming of more than a factor 2 greater than the global mean value [Hansen et al., 2007]. Permafrost thawing could be more extensive than previously predicted [Camill, 2005; Osterkamp, 2005], with large potential for methane emission.
 Atmospheric CH4 has a global average atmospheric lifetime of approximately 8 to 10 years [Denman et al., 2007]. Thus, atmospheric composition changes initiated by methane emissions from Arctic warming depend critically on the time horizon of the CH4 releases from thawing permafrost and degassing from marine hydrates (centennial to millennium timescales).
 Our calculations are based on a range of hypothetical future emission scenarios for CH4. There are currently no studies that quantify large CH4 emission from permafrost thawing and destabilization of marine gas hydrates in the Arctic in a future warmer climate.
2. CH4 Emissions From the Arctic Region
 We consider two major sources of CH4 emissions from the warming Arctic: (1) Methane produced from microbial degradation of labile organic carbon that becomes bioavailable as permafrost thaws; and (2) methane released from gas hydrate deposits as they dissociate in response to climate warming. Thawing permafrost may also promote emissions from other methane sources in the Arctic, but the amount of methane that could potentially be produced by microbial processes in thawed soils or release of methane from gas hydrates far exceeds that associated with other Arctic sources. There is evidence that continuous permafrost is actively thawing in many circum-Arctic regions, both onshore and in the shallow offshore continental shelves [Rachold et al., 2007].
2.1. Methane and Thawing Permafrost
 Thawing of permafrost at a rate of 0.04–0.10 m yr−1 has been observed in some terrestrial upland regions [Osterkamp, 2005], and it is shown that temperatures have increased at depths as great as 25 to 30 m below the surface at some locations in the Arctic during the last two decades [Isaksen et al., 2007; Osterkamp and Jorgenson, 2006]. Camill  finds that if the current rate of temperature increase (which is at least twice that of the global mean temperature increase) continues at his sites in boreal peatlands in Arctic Canada, the permafrost will have disappeared by the end of this century. The observations at this measurement site are consistent with a warming of at least 7–8°C in a high-CO2-emission scenario.
 Northern soils sequester an estimated 1,672 Pg (1 Pg = 1e+15 g) of organic C, 88% of which is stored in perennially frozen ground [Tarnocai et al., 2009]. The yedoma ice complex within the continuous permafrost in Northern Siberia is of particular interest for understanding the interplay among thawing permafrost, organic carbon, and methane emissions. The yedoma complex consists of ice-rich frozen loess and loess-related sediments that cover approximately 1 million km2. These sediments have organic carbon content of approximately 27 kg C/m3 [Walter et al., 2006] and represent a large reservoir (∼450 Pg C) of labile organic carbon [Zimov et al., 1997]. An additional 240 Pg C are sequestered in continuous permafrost in thick fluvial deltaic deposits. Thawing of this permafrost and particularly of the yedoma ice complex could produce large CH4 emissions via anaerobic degradation of the labile carbon, as demonstrated in laboratory incubations by Zimov et al. . In order to have sustained anaerobic conditions in thermokarsting soils, meltwater needs to be retained in the yedoma complex.
 Thermokarst lakes (lakes that have an underlying thaw bulb and that are widespread in some parts of the Arctic) are among the best places to study the potential release of CH4 during permafrost degradation. Plug and West  consider thermokarst expansion rates of up to 8 m in a single year. A strong amplification of CH4 emissions from Arctic lakes due to thawing of permafrost has already been observed in response to ongoing climate warming. For instance, Walter et al.  described significant increases (∼60%) in CH4 emissions from North Siberian thermokarst lakes between 1974 and 2000, associated with a more moderate (15%) increase in lake area during the same period. Calculations by Walter et al.  suggest that ∼50 Pg CH4 will be released from Siberian thermokarst (thaw) lakes as yedoma permafrost thaws over a timescale of centuries to millennia. CH4 emission from lakes in other nonyedoma permafrost-dominated regions in the Arctic is also expected as thermokarst lakes form and expand in conjunction with warming and thawing of permafrost.
 The remainder of nonyedoma permafrost in the Arctic contains ∼1,000 Pg C [Tarnocai et al., 2009]. Assuming that 10% of this permafrost will thaw beneath lakes and 17% of the C in the thawed permafrost will be converted to CH4, similar to the efficiency of CH4 production in yedoma, we estimate that 23 Pg CH4 can be released from nonyedoma thermokarst lakes. Together, thawing of yedoma and nonyedoma permafrost could thus release up to 73 Pg CH4 to the atmosphere.
 Several studies indicate that high-latitude wetlands have CH4 emissions that are sensitive to increased temperatures. For example, northern peatlands that experience permafrost thaw [Christensen et al., 2004; Wickland et al., 2006] have enhanced CH4 emissions. Due to uncertainties associated with future surface moisture and redox regimes, enhanced wetland emissions are not explicitly included as a methane source here.
 The CH4 emission rate and the magnitude of its accumulation in the atmosphere strongly depend on the rate of permafrost thaw. Under current assumptions of warming in the Arctic, permafrost thaw and CH4 release are likely to occur over timescales of centuries to millennia. In scenarios of more enhanced warming of the Arctic than considered in current climate models [IPCC, 2007] the timescale for CH4 release due to permafrost thawing could be shorter. Although the current global contribution from North Siberian thermokarst lake emissions of CH4 are small, Walter et al.  indicate that these emissions have increased significantly (∼60%) between 1974 and 2000.
2.2. Methane Hydrates
 Gas hydrates (clathrates) are a solid, ice like form of mostly methane, which occur beneath and possibly within [Dallimore and Collett, 1995] onshore permafrost and also in subsea permafrost that persists in some high-latitude regions to water depths as great as ∼90 m. Both permafrost-associated gas hydrates and the shallowest part of the deepwater marine gas hydrate system are susceptible to dissociation (breakdown to methane and water) under conditions of a warming Arctic climate.
 The most recent review of the numerous published estimates of the amount of methane sequestered in global gas hydrate deposits converges on a range of 3 to 40 × 1015 m3 of methane [Boswell and Collett, 2011], which converts to a range of ∼1,600 to 21,000 Pg C. This consensus range brackets some older estimates (3000 Pg C in the work of Buffett and Archer ) and a recent estimate of 1,000 to 10,000 Pg C by Krey et al. . Based on the estimates by Soloviev et al. , Shakhova et al. [2010a] conclude that one quarter of the Arctic ocean shelf contains 540 Pg CH4 in gas hydrates. This yields an estimated ∼1,600 Pg C within gas hydrates associated with subsea permafrost on the Arctic Ocean continental shelves. It is important to note that the formerly terrestrial sediments on these very shallow shelves contain significant additional carbon in nonhydrate form. Like the carbon trapped in terrestrial permafrost, this additional carbon is subject to microbial degradation and CO2 and CH4 production as the subsea permafrost thaws.
 1. Subsea Permafrost: There is substantial evidence that subsea permafrost is undergoing rapid degradation at high northern latitudes [e.g., Rachold et al., 2007; Shakhova et al., 2005]. The current rate of subsea permafrost degradation is unknown, and acceleration in this degradation with recent changes in sea ice cover and thus ocean temperatures is expected, but not yet fully documented. Still, dissociation of methane hydrate that is currently capped by or contained within subsea permafrost is very likely occurring now [e.g., Shakhova et al., 2010b] and should increase as warming affects the ocean-atmosphere system. Methane released from these hydrates would be emitted into shallow seas where relatively little is likely to be oxidized before reaching the atmosphere. Shakhova et al.  speculate that 50 Pg CH4 could be released abruptly at any time from gas hydrates associated with subsea permafrost. Although there is no basis for estimating the rate of such a release, this value is used as a worst case scenario for the numerical model studies.
 2. Methane hydrates in terrestrial permafrost near the top of the gas hydrate stability zone are unlikely to contribute significantly to the net methane flux from these deposits. A critical exception may be beneath deep thermokarst thaw bulbs, where the superposition of climate-induced surface warming and possible thaw bulb growth may so perturb temperatures near the top of the gas hydrate stability zone that methane hydrate at depths of only a few hundred meters might be susceptible to dissociation.
 3. Deepwater marine hydrates: Most deepwater gas hydrates are unlikely to be a major source of atmospheric methane in climate scenarios that consider less than several millennia. Dissolution and oxidation are likely to strongly reduce the amount of methane that reaches the atmosphere once it is emitted at the seafloor. One exception is gas hydrate in continental margin sediments at pressure-temperature conditions close to the gas hydrate phase boundary. For example, contemporary climate warming may be triggering gas hydrate dissociation and methane emissions along the West Spitsbergen margin [Westbrook et al., 2009]. In that case, much of the methane is emitted rapidly and in bubble streams, meaning that significant amounts could reach the atmosphere. In other cases, oxidation and dissolution might be expected to prevent much methane from crossing the ocean-atmosphere interface. The numerical model studies presented here do not include a potential contribution from this population of potentially degrading gas hydrates.
 Submarine landslides associated with gas hydrates and/or underlying free gas are possible for CH4 release from marine sediments (e.g., discussion and references in the work of Archer ). It is assumed that single landslides can release up to about 5 Pg C as CH4, based on estimates of past major landslides. This can be regarded as an upper estimate from single landslides, since a recent estimate indicates lower release [Paull et al., 2007]. Our models do not include this as a potential carbon source.
2.3. Model Inputs
 Clearly, large uncertainties exist in the amount of carbon stored in Arctic permafrost as methane hydrates, the fraction emitted as CH4 from thawing permafrost and the timing and magnitude of potential emissions from thawing permafrost and gas hydrates. Combining these with our limited knowledge of the rate and degree of long-term global warming it is clear that our estimates of the impact of future emissions of CH4 on atmospheric composition and on climate are also associated with large uncertainties. In order to demonstrate the potential contribution of high-latitude release of CH4 to global warming, several scenarios are considered here. The scenarios discussed in section 5 are based on the potential future emissions of CH4 estimated for the Arctic permafrost region and for CH4 stored in hydrates. The estimates include (1) sustained emissions over a specified time horizon from permafrost thawing necessary to reach a defined level of atmospheric CH4 enhancement and (2) a hypothetical release of 50 Pg of CH4 from the hydrate pool as a pulse over a limited period of time (1 year) or as sustained emissions over several decades. Although the high-emission scenarios are unlikely to occur, they are compatible with the current knowledge of the cumulative magnitude of CH4 that might be emitted from permafrost thawing and from CH4 hydrate destabilization.
3. Atmospheric CH4 Oxidation
 This section provides a brief description of the atmospheric chemistry leading to the formation of greenhouse gases from CH4 emissions.
 CH4 oxidation leads to enhanced formation of ozone in the troposphere and lower stratosphere through a sequence of reactions involving NOx compounds. The CH3 resulting from reaction (R1) is oxidized and the reaction products are photolyzed in the presence of sunlight:
 M is an air molecule (usually N2), participating in a three-body reaction, and “hν” represents the solar photon flux. Through this cycle ozone is efficiently formed in the presence of NOx, CO and CH4. The end product of the OH and HO2 formation from CH4 is water vapor. As two H2O molecules are formed from each CH4 molecule, water vapor enhancements due to additional CH4 releases can be important in relative terms in the dry stratosphere. Reaction (R9) yields CO2, constituting another important product from CH4 oxidation. The result of the CH4 oxidation chain is thus the formation of the three greenhouse gases O3, H2O, and CO2, which comes in addition to the enhancement of CH4 concentrations due to direct emissions.
4. Experimental Setup
4.1. Model Description
 We apply the global-scale Oslo CTM2 model (Oslo Chemical Transport Model). The model runs with meteorology from the European Centre for Medium Range Weather Forecasts (ECMWF) and uses the Second-Order Moments Scheme [Prather, 1986] for advective transport. Surface emissions are based on the EDGARv3.2 database [Olivier and Berdowski, 2001] for anthropogenic emissions, and on the work of Müller  for natural emissions. CH4 levels are prescribed in the troposphere with latitudinal distributions based on observations from the ESRL network [Dlugokencky et al., 2010; Masarie et al., 1991]. The model has a resolution, which is 2.8 × 2.8 degrees (T42) horizontally and 60 vertical layers in the troposphere and stratosphere, and has extensive tropospheric and stratospheric chemistry which has been validated in model-model and model-measurement inter comparisons of relevance to CH4 perturbations [Gauss et al., 2006; Isaksen et al., 2005; Shindell et al., 2006]. The model is further documented by Berglen et al.  and Søvde et al. .
4.2. Simulations and Definitions
 In the Oslo CTM2 model simulations atmospheric CH4 levels are fixed. The estimates of corresponding emissions are made with a forward explicit box model and described in more detail in section 5. Five simulations are made with 2003 meteorology. A base case simulation with present-day CH4 levels, and 4 perturbation cases with tropospheric CH4 levels multiplied by 2.5 (“2.5 × CH4”), four (“4 × CH4”), seven (“7 × CH4”), and thirteen (“13 × CH4”). For chlorine and bromine levels we use values suggested by WMO  for 2100, as we are considering long-term future changes. In each of the simulations the global lifetime of CH4 is calculated from a mass-weighted average of loss through reaction with OH. The base case simulation was spun up for several model years until a stable chemical composition was reached in both the troposphere and the stratosphere. Starting from the resulting chemical fields, both the base case and the perturbation cases were run for an additional 4 years to reach a new equilibrium corresponding to the respective CH4 levels.
 In the calculations we have used current atmospheric water vapor content. Since water vapor is expected to increase in a future warmer climate the calculations were repeated for a 40% increase in tropospheric water vapor (but no other changes). Such perturbations in tropospheric water vapor are expected to be representative for strongly perturbed future climate conditions and key compounds in the interactive climate-chemistry system like OH and O3 will be affected. We found that the calculated tracer and lifetime perturbations were only slightly affected by this increase (less than 10% impact).
 We define an enhancement factor η to quantify the effect of CH4 emissions on its own concentration and lifetime. η is given as the ratio of relative change in CH4 burden over relative change in CH4 emissions, i.e., η = (M2 − M1)/M1/(e2 − e1)/e1, where M and e represent the global CH4 burden and global CH4 emission at steady state, respectively. Subscript 1 is for the base case, and subscript 2 is for the perturbation cases. The feedback in atmospheric concentration from emissions is positive if η > 1. We define further the ratio of burden α = M2/M1 and the ratio of CH4 lifetimes as β = τ2/τ1.
5.1. Atmospheric Enhancement of CH4
 Estimates of current atmospheric CH4 lifetime are in the range 8 to 10 years [Shindell et al., 2006]. In the Oslo CTM2 model the lifetime of CH4 amounts to 9.1 years in the base case. Table 1 summarizes the main results for the four perturbation cases. For example, in the 4 × CH4 case we choose a fourfold increase in atmospheric CH4, i.e., α = 4 (or ΔM/M1 = 3) in the Oslo CTM2 model calculation and obtain a new CH4 lifetime of τ2 = 14.7 years, giving a relative increase in lifetime of β = 1.61. The 7 × CH4 case with α = 7 almost doubles the CH4 lifetime, β = 1.98.
Table 1. Changes in CH4 Burden Chosen for the 2.5 × CH4, 4 × CH4, 7 × CH4, and 13 × CH4 Cases Addressed by Oslo CTM2 Model Runs, Along With the Lifetimes τ2 as Modeled by Oslo CTM2 and the Increase With Respect to the Base Case Lifetime τ1a
Name of Case
Lifetime of CH4 (years), τ2
Δe in Pg CH4 yr−1
Also shown are the emission enhancements required to bring about the chosen enhancements in atmospheric CH4. M1 = 4.9 Pg CH4, e1 = 540 Tg CH4 yr−1, and τ1 = 9.1 years refer to CH4 burden, emission and lifetime in the base case. ΔM and Δe denote the absolute increases from the base to the perturbation cases. The enhancement factor η (see text) is a measure of the positive feedback in the chemistry of CH4 and is calculated from the changes in CH4 lifetime and burden at the new equilibrium. The three last rows show additional cases, which have not been simulated by the Oslo CTM2 model, but are based on emission estimates assuming thawing of permafrost, and a sudden release of CH4 from marine sediments.
Numbers apply to peak concentrations reached after 1 year.
Figure 1 shows β as a function of CH4 increase. The circles show the increases of CH4 lifetime as represented by β = τ2/τ1 for the four different cases of enhancement modeled by Oslo CTM2. To obtain intermediate values of β we apply cubic interpolation, which is assumed to approximate the relation between β and α ( = M2/M1) well within the considered range.
5.2. Calculated Atmospheric Changes in H2O, OH, and Ozone
Figure 2 shows model calculated changes in zonal-mean atmospheric water vapor, OH, and ozone for the 4 × CH4, 7 × CH4, and 13 × CH4 cases. Related to the negative height dependence of CH4 mixing ratio, stratospheric water vapor increases strongly with height. More CH4 leads to more stratospheric water vapor, and this increase is also significant below 30 km where stratospheric water vapor is radiatively efficient. OH decreases strongly in most of the troposphere due to increased CH4 oxidation. The strongest decrease in OH is found in the lower troposphere with reductions by more than 50% in the three cases. In the stratosphere, OH increases as a result of enhanced production from water vapor. The ozone increase is particularly large in the upper tropical troposphere where the ozone radiative forcing is strong [Ramaswamy et al., 2001], while in the lower stratosphere the ozone increase is modest, with regions where ozone even decreases slightly when CH4 is perturbed. In the upper stratosphere ozone is reduced due to the more efficient removal by HOx (odd hydrogen) reactions.
5.3. Relating the Chosen CH4 Enhancements to Emissions and Available Storage
 A forward explicit box model is applied to calculate the evolution of CH4 as function of time assuming different (sustained) CH4 emission rates. The calculation starts at present CH4 levels, and at each time step (0.01 years) it uses the CH4 lifetime that corresponds to the CH4 concentration calculated by the box model for the current time step. The relation between CH4 lifetime and concentration calculated in the Oslo CTM2 simulations (Figure 1) is taken as input to the box model. The resulting evolution of CH4 is shown in Figure 3. 96%, 92% and 87% of the chosen increases are reached after 50 years in the 2.5 × CH4, 4 × CH4 and 7 × CH4 cases, respectively, reflecting the increase in lifetime with increasing emission.
Table 1 also lists the additional CH4 emissions (Δe) needed to reach the adopted equilibrium enhancement values in the different perturbation experiments. These additional emissions are switched on in the forward explicit box model from the beginning. The base case level is consistent with an emission of 540 Tg CH4 yr−1. Total equilibrium emissions can thus be obtained by adding 540 Tg CH4 yr−1 to the values given in Table 1. However, an important question in relation to the equilibrium value is the timescale for the additional emission.
 Our highest estimate for thermokarst lakes is a 100 to 200 Tg CH4 yr−1 emission, which can be sustained for a time period of up to 500 years. Such enhanced emissions would increase the CH4 lifetime by about 7 to 15%, and the CH4 burden by factors of 1.3 to 1.6 (Table 1).
 To illustrate the atmospheric impact on CH4 of large pulse type releases, we assume a hypothetical case of 50 Pg CH4 emission, corresponding to CH4 hydrate releases from the Siberian Arctic shelf (as defined in section 2). The atmospheric impact for a pulse emission (over 1 year) and for sustained emissions over several decades, are considered. Our calculations (Figure 3) show that a 1 year pulse emission of 50 Pg CH4 would lead to abundances peaking at about 11 times its current level. Table 1 gives the yearly emissions needed to give equilibrium abundances in the atmosphere 2.5 times, 4 times and 7 times current abundances. A total of 50 Pg CH4 can sustain the additional emissions in the three cases for 111, 62.5, and 36.5 years, respectively, giving enhancements that are approaching the equilibrium values.
 For the sustained 100 Tg CH4 yr−1, 200 Tg CH4 yr−1 and the 1 year 50 Pg CH4 emission cases, the CH4 lifetime dependence derived from the four explicit Oslo CTM2 simulations (shown in Figure 1) is taken into account in the calculation of CH4 evolution.
5.4. Atmospheric Enhancement of CO2
 CH4 is oxidized to CO2 in the atmosphere, and CO2 decays according to different loss processes. Our calculation of the evolution of CO2 (shown in Figure 4) from oxidation of the CH4 released from thawing permafrost or hydrate storage is based on the formula of IPCC-AR4 [Forster et al., 2007], according to which the decay of a pulse of CO2 with time t is given by
Where a0 = 0.217, a1 = 0.259, a2 = 0.338, a3 = 0.186, τ1 = 172.9 years, τ2 = 18.51 years, and τ3 = 1.186 years. The pulses at each time step are set equal to the amount that is oxidized from the additional atmospheric CH4 from permafrost emissions. After 50 years, the CO2 values are 4.4, 7.3, and 11.3 ppm (ppm) for the 2.5 × CH4, 4 × CH4, and 7 × CH4 cases, respectively. For comparison with Figure 3, Figure 4 also includes the CH4 hydrate case for which the CO2 enhancement peaks at around 10 ppm after 40 to 50 years. It has to be noted that the CO2 shown in Figure 4 does not include directly emitted CO2 from permafrost thaw. The inclusion of CO2 in these calculations is consistent with the inclusion of CH4 oxidation to CO2 in the GWP for CH4 [Boucher et al., 2009].
5.5. Radiative Forcing
 Detailed off-line atmospheric radiative transfer schemes are used in the calculations of radiative forcing (RF), from changes in ozone and stratospheric water vapor (sH2O) [Myhre et al., 2000; Myhre et al., 2007]. For CO2 and CH4, expressions given by Ramaswamy et al.  are used. The radiative forcing is not related to a particular year but to the time when the CH4 increase is reached. The 13 × CH4 case refers to a shorter time horizon (30 years) while the other cases are based on a 50 year time horizon.
 Results are shown in Figure 5 for the 4 × CH4, 7 × CH4 and 13 × CH4 cases. Radiative forcings due to CH4 are listed in Table 2. In the 4 × CH4 case the RF caused by direct emissions of CH4 and the RF due to the lifetime increase are comparable and add up to 1.3 Wm−2. The RF of 0.5 Wm−2 from ozone, calculated for the same case, is probably stronger than the forcing since preindustrial time [Gauss et al., 2006]. For sH2O we calculate nearly quadrupling (0.3 Wm−2) of the current sH2O forcing (0.08 Wm−2) [Myhre et al., 2007]. In the 7 × CH4 and 13 × CH4 cases the indirect CH4 RF clearly dominates the direct one, due to more pronounced OH changes. The total RF is 2.2, 3.6, and 5.4 Wm−2 for the 4 × CH4, 7 × CH4 and 13 × CH4 cases, respectively.
Table 2. Radiative Forcing Due to Directly Emitted CH4 (CH4 Direct Forcing), the Lifetime Change of CH4 (CH4 Indirect Forcing), and the Chemical Enhancements Due to the Feedbacks in the 4 × CH4, 7 × CH4, and 13 × CH4 Casesa
Name of Case
CH4 Direct Forcing
CH4 Indirect Forcing
Indirect Forcing FromO3, strat. H2O and CO2
For the “100 Tg” and “200 Tg” cases (in the table listed in italics) no Oslo CTM2 calculations have been made. Thus the split into direct and indirect components of radiative forcing is based on extrapolated values for β. Contributions to indirect radiative forcing from perturbations of O3, stratospheric H2O and CO2 are not included in the table.
4 × CH4
7 × CH4
13 × CH4
 The importance of CH4 emissions through atmospheric chemistry perturbation is illustrated by a 3.5- to fivefold increase in the total RF compared to the contribution from direct emissions, increasing in importance with the increase in CH4 emission. The impact of chemical perturbations by CH4 emission increases is substantially higher than what is included in the GWP value for CH4 used in IPCC [Forster et al., 2007], where a twofold increase in RF was adopted to account for indirect effects at current conditions. The CH4 lifetime feedback represents the most important contribution among the additional radiative forcings due to changes in atmospheric composition, followed by ozone (in particular tropospheric ozone) and sH2O.
 According to our simulations, sustained CH4 emissions of 100 Tg yr−1 to 200 Tg yr−1 due to permafrost thawing will correspond to a RF of 0.3 to 0.6 Wm−2 in a new equilibrium (reached after a few decades). Although the amount of CH4 released to the atmosphere is large, the short lifetime (8 to 10 years) compared to the time horizon for emission (up to 500 years) means that equilibrium in atmospheric concentration is reached at an early stage in the time period. As an illustration of the potential of permafrost CH4 releases, the increase of the CH4 concentration by a factor 2.5 compared to current CH4 concentrations would correspond to an extra emission of 450 Tg CH4 yr−1 from permafrost thawing and give a radiative forcing of about 1.5 Wm−2. However, the emission would be over a shorter time horizon than in the 100 Tg CH4 yr−1 to 200 Tg CH4 yr−1 emission cases.
 A release of 50 Pg (CH4) from gas hydrates over a 30 year time period could give a RF in excess of 4 Wm−2. The larger calculated RF, results from the fact that this case represents a more rapid emission than our 7 × CH4 case.
 The release of up to 5 Pg C estimated for single submarine landslides (see discussion in section 2) can enhance atmospheric CH4 by nearly 2.5 ppm, and give a RF slightly larger than 1 Wm−2. This is substantially higher than the 0.2 Wm−2 (only the direct RF was taken into account) from a similar CH4 release discussed by Archer . It has to be noted, however, that more recent estimates of the emission from the Storegga Landslide are lower than 5 Pg C [e.g., Paull et al., 2007], and that there is no evidence that the CH4 released escaped to the atmosphere after the event.
6. Discussion and Conclusion
 Relating the magnitude and timing of CH4 release in a warming Arctic to the adopted scenarios applied in the model studies is challenging due to uncertainties associated with the rate of permafrost warming and thawing, the size and distribution of the large organic C pool [Tarnocai et al., 2009] and its fate in microbial decomposition under aerobic versus anaerobic conditions [Schuur et al., 2008; Walter et al., 2007], and the size and stability of the hydrate reservoir. However, given the magnitude and vulnerability of the permafrost organic carbon pool and the large permafrost-associated CH4 hydrate reservoir, quantification of the feedbacks associated with the release of CH4 from these systems is important for climate modeling.
 There is a potential for the release of large amounts of carbon stored in the Arctic to the atmosphere as CH4 following the thawing of permafrost and warming of oceans [Buffett and Archer, 2004; Schuur et al., 2008; Walter et al., 2007, 2006]. Considering the positive feedback (increase in lifetime) and the non linearity in atmospheric chemistry through amplification of atmospheric CH4 (represented by the enhancement factor η), we have applied a wide range of CH4 emission scenarios to explore the impact of such interactions. The impact on methane lifetime does not only apply to the additionally emitted CH4 but also to background levels of CH4.
 Although the high-CH4 scenarios applied in this study are unlikely, they demonstrate the strong CH4 feedbacks in the climate system, with large amplification of atmospheric composition changes and RF compared to the direct RF of CH4 emissions.
 The Oslo CTM2 has been thoroughly tested with respect to OH and CH4 lifetime through model comparisons and through perturbation studies [Hoor et al., 2009; Shindell et al., 2006]. These studies show that there are noticeable differences in model estimates. Another model uncertainty in the Oslo CTM2 calculations may arise from the use of today's meteorology, uncertainties in emissions of other gases than CH4 (e.g., emissions of NOx and CO), and the formulation of dry deposition (O3) and wet deposition (NOx) [Isaksen et al., 2005; Solberg et al., 2008]. Yet, the uncertainties in the atmospheric chemistry calculations are likely to be small compared to the uncertainties in the amount and time horizon for CH4 emission from thawing permafrost or degassing hydrates.
 Since current atmospheric CH4 lifetime ranges between 8 and 10 years, the RF of CH4 emissions from permafrost thawing and CH4 hydrate degassing depends critically on the rate of emissions. Additional studies linking CH4 emissions to the possibilities for large future warming in the Arctic are needed.
 Although CH4 emissions from permafrost thawing and hydrate degassing are likely to have a very different spatiotemporal distribution than anthropogenic emissions, the impact on climate will be global in the same way as the impact from CH4 emitted in other regions, when we take into account the multiyear lifetime of CH4.
 Fossil fuel CO2 emissions have increased substantially over the last decade and is now 40% higher than in 1990 [Le Quéré et al., 2009; Myhre et al., 2009]. The continued increase in greenhouse gas emissions toward the end of this century has the potential to produce significant warming at high northern latitudes well beyond what has been observed during the last decades [Hansen et al., 2007; IPCC, 2007]. There is a possibility that the Arctic temperature increases could be followed by extensive permafrost thawing, with enhanced CH4 emission from thermokarst lakes [Walter et al., 2006], with later release of CH4 from gas hydrates that would eventually be affected by warming temperatures. Considering the large, nonlinear atmospheric chemistry feedbacks discussed here, future CH4 emissions from permafrost deposits could be a larger concern for climate warming than previously thought.