Blog: Source-driven Doubling in Methane


By James Levine. 18th January 2012.

As a potent greenhouse gas, the amount of methane in the Earth’s atmosphere affects our climate. It also has bearing on the quality of air we breathe, as it influences the ability of the atmosphere to rid itself of pollutants (including other greenhouse gases) that could reach levels harmful to human health if allowed to accumulate. We therefore need to understand the causes of variations in the concentration of this key constituent, and an ability to account for past variations is a prerequisite for meaningful future predictions.

Air bubbles trapped in Antarctic ice (pictured) reveal large variations in the concentration of methane, [CH4], over the last 800,000 years that appear to broadly track changes in climate. The question at the heart of a recent study by Levine et al. (2011) is, why did [CH4] almost double from 360 parts per billion by volume (ppbv) at the so-called Last Glacial Maximum (LGM; around 21,000 years ago) to 700 ppbv in the relatively warm pre-industrial era (PI; about 200 years ago)? Did the amount of methane emitted from natural sources such as wetlands increase, or did the rate at which methane was removed from the atmosphere decrease, allowing those emissions to accumulate to higher concentrations? The relative contributions made by changes in methane sources and changes in methane ‘sinks’ have long been debated.

Early ‘bottom-up’ model estimates of the changes in methane sources could only explain around half the change in [CH4], appealing to a reduction in the main methane sink—oxidation by the hydroxyl radical (OH)—as the climate warmed; the warming (and increasingly wet) climate would have seen an increase in vegetation and an increase in the amount of volatile organic compounds emitted from vegetation (e.g. isoprene) that compete with methane for reaction with OH. However, the warming climate would have also been accompanied by an increase in humidity, fuelling greater OH production, and an increase in the reactivity that OH shows towards methane. So there were opposing influences on the rate of methane removal by OH.

Using a computer model of the Earth’s atmosphere, we find that the changes in emissions from vegetation would have had a significant influence on [CH4] of the sort previously proposed, but this would have been almost entirely negated by the accompanying changes in humidity and chemical reaction rates, the implication being the LGM-PI change in [CH4] was essentially entirely source driven. Meanwhile, estimates of the change in methane emissions from wetlands during this period have increased, and now include almost precisely that needed to explain the change in [CH4] without recourse to changes in the rate of methane removal. We therefore conclude that it is plausible the LGM-PI change in [CH4] was entirely source-driven, and the changes in methane sources and sinks between the LGM and the PI could be reconciled thus.

Levine, J. G.1, E. W. Wolff1, A. E. Jones1, L. C. Sime1, P. J. Valdes2, A. T. Archibald3,4, G. D. Carver3,4, N. J. Warwick3,4, and J. A. Pyle3,4 (2011), Reconciling the changes in atmospheric methane sources and sinks between the Last Glacial Maximum and the pre-industrial era, Geophys. Res. Lett., 38, L23804, doi:10.1029/2011GL049545.

1British Antarctic Survey, High Cross, Madingley Road, Cambridge, UK.
2School of Geographical Sciences, University of Bristol, Bristol, UK.
3Centre for Atmospheric Science, Department of Chemistry, University of Cambridge, Cambridge, UK.
4National Centre for Atmospheric Science, University of Cambridge, Cambridge, UK.

Blog: Of Mammoth Significance?


By Gail Riekie. 9th May 2011.

Methane researchers will all be accustomed to the old jokes about flatulent cows, but fewer will be familiar with the debate about the possible role of methane emissions from other large bodied herbivores in an earlier era.

Prior to the arrival of humans, around 12,500 years ago, a vast population of now extinct megafauna (mammoths, camelids, giant ground sloths, to name but a few) roamed the Americas.  The megafauna’s dramatic demise is considered to be the earliest catastrophic event attributable to our own species.

A Nature Geoscience article published last year by Smith et al. (1) suggested that the drop in atmospheric methane concentrations observed in the ice core record at the onset of the Younger Dryas, ~12,800 years ago, occurred at an unusually rapid rate, and could be explained, at least in part, by the loss of methane emissions from megafauna.  Using estimates of the amount of methane emitted by the 114 species of herbivores known to have become extinct by the end of the Pleistocene epoch, Smith and co-authors calculated that the loss of megafauna could account for between 12.5 and 100% of the observed methane decrease. They also suggested that the onset of the ‘Anthropocene’ should be recalibrated to 13,400 years before present, when humans first started migrating on a large scale into the Americas.

More recently in Nature Geoscience, doubt has been cast on some of these claims (2). Brook and Severinghaus assert they are inconsistent with constraints imposed by the ice core record, noting that the drop in concentration of atmospheric methane discussed by Smith and co-authors and more usually attributed to decreased wetland emissions, is not in fact uniquely rapid in rate, and that furthermore it is likely that the megafauna extinctions occurred too early to be the cause of the observed methane decrease.

The debate about changes in atmospheric methane concentration (causes, timings and rates) in the atmosphere over the past 125,000 years is doubtless not over yet.  The ‘Early Anthropocene’ hypothesis, originally proposed by William Ruddiman (3) in the context of the onset of human-related activities such forest clearance (8000 years ago) and rice agriculture (5000 years ago) is also still a live topic, as discussed elsewhere on this website in the news item ‘Early Anthropocene doubt’. Watch this space.


(1) Smith, F.A., Elliot, S.M. and Lyons, S.K. (2010) Methane emissions from extinct megafauna. Nature Geoscience, 3, 374-375.

(2) Brook , E.J. and Severinghaus, J.P. (2011) Methane and megafauna. Nature Geoscience, 4, 271-272.

(3) Ruddiman, W. (2003) The anthropogenic greenhouse era began thousands of years ago. Climate Change. 61, 261-293.

Methane in the Earth System

Nisbet1 (2)

16th June 2014.

Methane in the earth system – what can we measure, understand and do about recent trends?

By Michelle Cain.

The topic of the latest Cambridge Centre for Climate Science (CCfCS) symposium was “Methane in the Earth System”, held on 5 June 2014, and supported by MethaneNet. The aim of the symposium was to bring a diverse range of speakers together to generate discussion on this broad theme. Not only did we hear from scientists about this topic, but also an engineer from the government Department for Energy and Climate Change (DECC), who was able to enlighten us on how DECC thinks about methane in terms of the UK policy environment.

Professor Euan Nisbet kicked off the afternoon with a whistle-stop tour of methane through the ages and across the globe: from methanogenesis 4 Gya (4×109 years ago) through to the present day. Entitled “Is methane the canary in the mine?”, the talk referred to the idea that methane may be like a lever that kick-starts periods of global warming. Is an increase in atmospheric methane a sign that the temperature is going to follow a similar trend? Or perhaps it’s the other way around, and increased temperature is responsible for the increasing atmospheric methane concentrations?

Observing and understanding the recent trends in atmospheric methane was the main focus of Professor Nisbet’s talk. The trends stabilised around the turn of this century, but have been rising since about 2007. The puzzle of why this has happened is yet to be solved. Measurements of isotopes of carbon in methane show that the global average isotopic ratio is getting “lighter”, meaning the sources are becoming more characteristic of wetland, and less like combustion. However anthropogenic emissions inventories show increases in recent years. If both of these trends are true, presumably the wetland source must be increasing at a greater rate than the anthropogenic source. This could be due to weather events like floods and the position of the inter-tropical convergence zone. The big question that follows on from this is whether this is a short- or a long-term trend. Professor Nisbet showed many examples of measurements, both short- and long-term, spanning the tropics to the poles, which will go some way to resolving this question in the fullness of time.

Dr Nic Gedney followed with a talk about one of the key aspects of understanding the methane trends – the modelling of methane emissions from wetlands. Both temperature and precipitation are critical for modelling wetland location and therefore methane emissions and potential feedbacks. For example, the Wetland and Wetland CH4 Inter-comparison of Models Project (Melton et al 2013) found that methane emissions increased when there was increased precipitation, and that global wetland area decreased when temperature increased. The model results revealed disagreements in both spatial and temporal extent of wetland areas and therefore methane emissions.

Dr Gedney also showed recent developments to the Joint UK Land Environment Simulator (JULES), a “process-based model of carbon, energy and water exchange between atmosphere and land surface.” The model uses drainage and topography to calculate inundation. Updating the model with a representation of organic soil and with new topography has improved the model compared to observations.

To make further improvements to the complex task of modelling methane emissions from wetlands, it will be essential to improve on observations of wetland extent. There are currently striking differences between inundation products derived from satellite observations and wetland models. However, wetlands are not always detectable until you “step into one and end up knee deep in water”, and so it’s clear that detection from space will have its limitations. Nonetheless, improvements to current measurement techniques will be a valuable step towards better understanding wetland areas.

This theme was continued in Professor John Burrows’ talk about measuring the anthropocene (in particular methane) from space, which he described as “very much work in progress”. The anthropocene is a new epoch, in which humans are changing the earth system dramatically. Examples of this include: burning fossil fuels, releasing other pollutants, causing the ozone hole and changing land use. Without measurements, it’s impossible to understand or manage how we are altering the earth system and climate.

Given that ground-based measurements will always be sparse, we need space-based techniques to get good global coverage. Ideally, a combination of low earth orbit (which has good global coverage) and geostationary orbit (high resolution, but less wide coverage) would provide the best information. Unfortunately, the current processes for getting a new satellite built are very slow, and it’s possible that the space-based measurements of methane will take a step backwards if new instruments are not developed quickly enough. CarbonSat is the next big project being considered for deployment by the European Space Agency. This would measure methane at a 2km by 2km resolution, however if successful, it won’t be launched until 2022.

Professor Burrows showed many different applications of satellite data in understanding recent methane trends. Data from the SCIAMACHY satellite have added to the quantification and identification of the recent increasing methane trend, which is mainly found in the tropics and northern mid-latitudes. These data have then been put into inverse methods to work out emissions. An inversion is a technique for calculating a best estimate of emissions by reconciling observations and a model using a cost function. Bergamaschi et al (2013) suggest that anthropogenic sources are causing the trend, and that wetlands and biomass burning dominate interannual variability. Another inversion study by Houweling et al (2014) was unable to discriminate between Asian anthropogenic and wetland sources. There are many other recent studies contributing to this debate, and this issue is far from resolved.

We heard more about inverse methods in Dr Matt Rigby’s talk, in which he explained why we should take some inverse modelling – including some of his own – with “a pinch of salt”. The key factor is how uncertainties are estimated. There are uncertainties in both the prior information and in the transport model used. Bayes’ theorem requires that these should be independent, but this is not always the case in inversions that use this assumption. Defining the size of an uncertainty may be simply an estimate, and may be “tuned” in concert with other uncertainties. Dr Rigby demonstrated an alternative approach to eliminate this particular problem: the use of a hierarchical Bayesian estimation, which estimates the uncertainties in the uncertainties. An example of this from Ganesan et al (2014), showed that a hierarchical approach tended to have a larger uncertainty, but one that is arguably closer to the true uncertainty.

Another approach for tackling the issue of uncertainty is to test a model’s sensitivity to its parameters. This can reveal large and sometimes non-linear sources of error, which are often simply untested and therefore ignored. Dr Rigby suggested that statistical emulators might be a good way to test for such systematic errors.

The concentration of atmospheric OH (the main sink for methane) is another uncertainty that affects our understanding of methane. Dr Rigby showed a time series of derived global OH concentrations based on methyl chloroform observations. This showed a low variability (<5%) in the past decade, although this is poorly constrained.

Although there is room for improvement in terms of reducing uncertainties, inverse methods are a key component to understanding methane emissions. The UK uses such techniques to verify bottom-up national emissions inventories, and is a world-leader in this kind of approach. The UK measurement network is also relatively dense, with recent projects like GAUGE adding to this.

We had a shift in perspective for the final talk of the day when we heard from Dr Philip Sargent, an engineer from DECC. Dr Sargent gave us a flavour of the concerns DECC has when it comes to methane. One aspect of DECC’s role is to weigh up which policy options emit the least greenhouse gases. For example, how does UK fracked gas compare to imports of liquid natural gas or via a pipeline? The UK must reduce carbon emissions if it is to meet the Committee on Climate Change’s 4th carbon budget in 2025.

The heart of the matter is that DECC must balance policies and spending related to methane with those related to carbon dioxide. Therefore metrics that allow comparisons to be made are essential. Dr Sargent posed an open question regarding what time horizon should be used for global warming potential (GWP). The Intergovernmental Panel on Climate Change (IPCC) says that using 100 years is an arbitrary choice. Alternatively, perhaps it is better to specify a cut-off date, and use that as the time horizon. Evidence, including uncertainties, to support the use of a particular metric and time horizon is one particular point that DECC is interested in, as it can be used for economists to base their models on. However, Dr Sargent also noted that it is preferable to continue using GWP (likely with more useful time horizons), as this term is well understood among policy makers – vital for international negotiations.

Formal proceedings ended with a discussion about the kinds of evidence the science community can usefully provide to government, which continued in more depth over the poster reception. A take home point from this final session was the importance of finding opportunities for policy makers to communicate with scientists. Sometimes talking to someone is the most effective route to providing them with the information that they require, but having that chance to talk can often be the limiting factor.


Bergamaschi, P., et al. (2013), Atmospheric CH4 in the first decade of the 21st century: Inverse modeling analysis using SCIAMACHY satellite retrievals and NOAA surface measurements, J. Geophys. Res. Atmos., 118, 7350–7369, doi:10.1002/jgrd.50480.

Ganesan, A. L., Rigby, M., Zammit-Mangion, A., Manning, A. J., Prinn, R. G., Fraser, P. J., Harth, C. M., Kim, K.-R., Krummel, P. B., Li, S., Mühle, J., O’Doherty, S. J., Park, S., Salameh, P. K., Steele, L. P., and Weiss, R. F.: Characterization of uncertainties in atmospheric trace gas inversions using hierarchical Bayesian methods, Atmos. Chem. Phys., 14, 3855-3864, doi:10.5194/acp-14-3855-2014, 2014.

Houweling, S., Krol, M., Bergamaschi, P., Frankenberg, C., Dlugokencky, E. J., Morino, I., Notholt, J., Sherlock, V., Wunch, D., Beck, V., Gerbig, C., Chen, H., Kort, E. A., Röckmann, T., and Aben, I.: A multi-year methane inversion using SCIAMACHY, accounting for systematic errors using TCCON measurements, Atmos. Chem. Phys., 14, 3991-4012, doi:10.5194/acp-14-3991-2014, 2014.

Melton, J. R., Wania, R., Hodson, E. L., Poulter, B., Ringeval, B., Spahni, R., Bohn, T., Avis, C. A., Beerling, D. J., Chen, G., Eliseev, A. V., Denisov, S. N., Hopcroft, P. O., Lettenmaier, D. P., Riley, W. J., Singarayer, J. S., Subin, Z. M., Tian, H., Zürcher, S., Brovkin, V., van Bodegom, P. M., Kleinen, T., Yu, Z. C., and Kaplan, J. O.: Present state of global wetland extent and wetland methane modelling: conclusions from a model inter-comparison project (WETCHIMP), Biogeosciences, 10, 753-788, doi:10.5194/bg-10-753-2013, 2013.

The Continued Rise of Atmospheric Methane Concentrations


12th February 2014.

A new article in Science has discussed the recent changes in atmospheric levels of methane, as well as examining possible drivers for these changes. As highlighted on MethaneNet last October, global methane concentrations increased by 12 ppb per year in the 1980s. This rise slowed in the 1990s, and then stabilised entirely from 1999 to 2006, before concentrations began to increase again at a rate of 6 ppb per year.

As would be expected, methane growth rate varies across different regions of the globe. For instance, growth has been above global levels in the southern tropics since 2007 and the authors suggest that this may have been due to wet summers stimulating the expansion of wetland area. Additionally, they point out the large increase in Arctic methane that occurred solely during 2007, but also suggest that catastrophic emission scenarios of methane from hydrates are unlikely. Modelling of methane concentrations thus shows that tropical wetlands were responsible for driving atmospheric concentration growth in 2007, and that since then the tropics and northern mid-latitudes have been important contributors. There are also anthropogenic sources to consider. Coal mining activities have expanded across some areas of the globe, particularly China, and fracking has increased in popularity in the US.

Additional data are needed to reconcile top-down and bottom-up estimates of atmospheric methane, and more isotope studies that would allow emission sources to be identified. The authors conclude by warning that funding for the monitoring of greenhouse gases is shrinking at a time when they are sorely needed.


Nisbet, E.G., Dlugokencky, E.J., Bousquet, P. 2014. Science, 343, 493-495.

New Estimates of Decadal Global Methane Dynamics. MethaneNet.

Image: Ed Dlugokencky, NOAA CMDL. This plot shows methane observations from 1993 to 2007 showing the growth of methane, the seasonal variations and the difference between northern and southern hemispheres

New Estimates of Decadal Global Methane Dynamics


16th October 2013.

A review of global methane sources and sinks has recently been published in Nature Geoscience by an international group of scientists.  They attempted to explain decadal changes in atmospheric methane concentrations using a combination of bottom-up and top-down estimates.  Methane concentrations increased by 12 ppb per year in the 1980s, but this rate dropped to 6 ppb per year in the 1990s, before stabilising between 1999-2006.  Since then, concentrations have once again started to rise.

Global methane sources are grouped into three categories: biogenic (methanogens), thermogenic (released from natural seeps and fossil fuel usage) and pyrogenic (released from soil and biomass during wildfires, as well as from fossil fuels and biofuels).  The main methane sink (90%) is oxidation by hydroxyl radicals in the troposphere.  Other small sinks include soil methanotrophy, reactions with stratospheric chlorine and oxygen radicals, and reactions with tropospheric chlorine radicals from sea salt.

The results of the review showed that the main natural emitter of methane during the 2000s was natural wetlands (a bottom-up approach estimated 217 Tg per year), with lakes and rivers (40 Tg per year), and geological sources (54 Tg per year) also being important contributors.  Considering anthropogenic sources, agriculture and waste (200 Tg per year) released more methane than fossil fuels (96 Tg per year).  Tropospheric hydroxyl reactions, the primary sink, were estimated to consume 528 Tg per year, whilst soil methanotrophy was calculated as 28 Tg per year.

It seems that the yearly variations in methane emissions have been driven by wetlands, but the drivers of decadal variations are more enigmatic.  One suggestion for the stabilisation of methane emissions during the early 2000s is that changes occurred due to decreasing and stabilising methane emissions from the fossil fuel industry, coupled to the same trend in emissions associated with rice growing.  A second possibility is a combination of stabilising fossil fuel emissions and decreasing microbial emissions.  The explanation for the renewed increase in methane concentrations after 2006 is that wetland methane emissions were stimulated by high temperatures in northern latitudes and increased rainfall over the tropics.  Additionally, fossil fuel emissions increased due to shale gas extraction in the US, and coal exploitation in China and India.

The paper concludes by discussing uncertainties in their calculations.  They point to the need to improve wetland mapping, as well as a general shortage of datasets for wetland methane fluxes that are needed to validate models.  They also note that decadal trends for both natural and anthropogenic methane emissions are extremely uncertain, making it difficult to correctly attribute the source for atmospheric changes.

Reference: Kirschke, S., et al. 2013. Three decades of global methane sources and sinks. Nature Geoscience, 6, 813-823.

Image:  AIRS 2011 annual mean upper troposphere(359Hpa) methane mixing ratio.

Recent Methane Puzzle

Grey World@flickr

16th August 2011.

Methane has featured prominently on the pages of the journal Nature in recent weeks.

Two different explanations for unexpected changes in methane concentration in the atmosphere in the last twenty years were proposed in the same volume (Aydin et al., and Kai et al., 2011). A commentary on these two papers (Heimann, 2011) and an earlier review paper ‘Non-CO2 greenhouse gases and climate change’ (Montzka et al., 2011) provided useful context to the debate.

Aydin and colleagues took the novel approach of measuring the atmospheric variability of ethane in the firn (perennial snowpack) in Greenland, and used this to reconstruct estimates of methane from fossil fuel sources. They concluded that the amount of methane emitted from fossil fuels during the twentieth century is “strikingly different from bottom-up estimates”. Their inferred pre-1980 methane emissions from fossil fuel sources are double those from standard databases based on fossil fuel production, and a subsequent sharp post-1980 decline of 30% explains, according to their model, the observed slow down in the rate of increase in atmospheric methane.

By contrast, Kai and colleagues suggest that the late twentieth century pattern is best explained by reduced methane emissions from microbial source, in particular from rice paddy fields. They constrained atmospheric methane models with carbon isotope data, using the fact that microbially sourced methane is relatively depleted in 13C whereas that from fossil fuels is relatively enriched. They also used 2H/1H isotope ratios in the atmospheric methane to control for possible changes in the atmospheric sink. According to their models, the observed changes in methane concentration are incompatible with a fossil fuel source.

The review paper by Montzka (a co-author on the Aydin paper) and colleagues gives some insight into why explaining changes in global atmospheric methane concentrations is such a vexed issue. Methane has a relatively short lifetime of methane the in the atmosphere (~9 yr). There is a delicate balance between methane sources and sinks. The major sources are highly sensitive to both climate change and human activity, and both positive and negative feedback mechanisms operate. The magnitude of the hydroxyl radical sink is influenced by complex atmospheric processes. All these factors combined result in high short-term variability in the amounts of methane in the atmosphere, which will need to be better understood if future attempts to reduce the amount of methane in the atmosphere are to be effective.


Aydin, M., Verhulst, K.R., Saltzman, E.S., Battles, M.O., Montzka, S.A., Blake, D.R., Tang, Q. and Prather, M.J. (2011). Recent decreases in fossil-fuel emissions of ethane and methane derived from firn air. Nature, 476, 198-201.

Kai, F.M., Tyler, S.C., Randerson, J.T. and Blake, D.R. (2011). Reduced methane growth rate explained by decreased Northern Hemisphere microbial sources. Nature, 476, 194-197.

Heimann, M. (2011). Enigma of the recent methane budget. Nature, 476, 157-158.

Montzka, S.A., Dlugokencky, E.J. and Butler, J.H. (2011). Non-CO2 greenhouse gases and climate change. Nature, 476, 43-49.