The Methane Time Bomb

Atmospheric methane peaks from sampling stations at Barrow northern Alaska and Alert northern Canada (off N Greenland) are centred on January of each annual cycle (Figures 1 and 4). This makes it highly unlikely that the annual cycle in methane concentrations is caused by the freeze thaw cycle of Arctic tundra.

Rather, the rise of methane concentrations 1983 to 1990 and subsequent flattening of the curve, combined with annual cycles with mid-winter peaks (Figure 2), matches Russian (or N American) natural gas production.

Figure 1 Atmospheric methane concentration data from Alert station, Canada. The annual peaks are centred on January each year when the Alert station is surrounded by frozen ground and frozen seas for more than 500 miles on all sides.

Data source is NOAA: Alert and Barrow.

In a comment last week, commenter Todd De Ryck linked to this blog article by Dr Jason Box. Amongst other things the article said:

Methane emission from the Arctic shelf has a maximum in September -October. [when sea ice minimum occurs]

I was already conditioned to believe that the annual cycle present in Arctic atmospheric methane data was due to the natural freeze thaw cycle of the circum Arctic tundra. And I was therefore already conditioned to believe that the cycle maxima present in the NOAA data presented by Dr Box were centred on late summer. My intention was to write a quick and cheerful post placing the NOAA data into appropriate context by converting to ppm (parts per million)  from ppb (parts per billion) and using a more appropriate y-axis scale (Figure 3). I then checked the data on an expanded x-axis (time axis) scale and was quite astonished to find that the methane peaks were centred on January (Figures 1 and 4).

There are two Arctic stations with extensive data series. One at Alert, Canada, that sits on the edge of the Arctic Ocean off the N coast of Greenland and the other at Barrow on the N coast of Alaska. The Barrow data are more noisy (Figures 2 and 4)  but show the same trend as the Alert data with peaks centred on January (compare Figures 1 and 4).

Figure 2 The Barrow Alaska sampling station has the longest record of Arctic atmospheric methane data starting in 1983. The annual cycle defined by the majority of data points is plain to see but there are a large number of points that scatter above and below the main trend. Those below are clearly data errors. Those above were described by Dr Box as Dragon’s breath alluding to degassing of the tundra or submarine clathrates. That story will have to wait to another day. The annual peaks of the Barrow methane data are also centred on January (Figure 4). 

Figure 3 The Barrow Alaska data, 1983 to 2013, converted to ppm and plotted on a 0 to 100 ppm y-axis scale. This is the same data as plotted in Figure 2. There is hardly any methane in the atmosphere and plotted at this scale it is impossible to discern any annual or decadal variation. 10,000 ppm = 1%, thus 100 ppm = 0.001%. Methane in the arctic atmosphere has a concentration of 2 ppm = 0.0002% of the atmosphere. The rise in methane since 1983 is about 0.2 ppm or 0.00002% of the atmosphere. NOAA assert that increasing methane from pre-indutrial levels accounts for 20% of anthropogenic climate forcing. Source of chart image is IGBP.

Figure 4 The Barrow Alaska data from 2004 to present also have methane peaks centred on January each year.

Figure 5 If the Arctic methane concentration data are not controlled by the annual freeze thaw cycle of the tundra then another explanation is required. The rise in methane 1983 to 1990 matches the rise in Soviet Union gas production but not the subsequent fall nor the dip associated with the 2008 financial crash.

In mid winter it is very difficult to explain the methane peaks by a natural cyclic thawing process since both sampling sites are surrounded by hundreds of miles of frozen ground, snow and sea ice at this time. So what is the origin of the annual cycle in the data? I have no axe to grind here and can but observe that Russian natural gas production rose until 1990 and thereafter plateaued and that Russian (indeed N hemisphere) natural gas production and consumption goes up in winter normally reaching a peak in January and February (Figure 5). If anyone else has a better idea then please offer it in the comments.

I have Googled around a bit looking for articles on Arctic methane and there are of course an abundance of them describing methane bubbles, plumes, melting permafrost, tipping points, ice breakers crunching their way through the Arctic sea ice trying to avert global melt down. But none of those I read in my quick search include reference to the beautiful NOAA data sets. I wonder why not?

Note added 10 September:

Commenters Peter Shaw and Greg Goodman suggested that the annual cycle with summer troughs may be due to optical destruction of methane. In other words sunlight accelerates the oxidation of methane to CO2. Roger Andrews posted the chart below that shows the methane cycle is reversed in the southern hemisphere. Optical destruction of methane in summer is a leading contender to explain the annual cycles.

This entry was posted in Climate change and tagged , , , . Bookmark the permalink.

22 Responses to The Methane Time Bomb

  1. J I Pitman says:

    I just had a look at the CH4 data for Summit Greenland

    and these observations also show a clear seasonal cycle which peaks in Jan-Feb.

    The methane C13/C12 however shows a summer peak for this location

    The paper by J D Miller (2002) compares the seasonal cycles of CH4 in both the Northern Hemisphere and SH, and suggests that the C13/C12 isotopic ratio may be useful in elucidating the differing origins of CH4, as biological delta 13 is -60 0/00, coal and gas -40, and biomass burning -25.

    • Euan Mearns says:

      John, thanks for links, key chart from the paper below. Had a quick scan and believe they conclude the d13C data is inconclusive. There is only half a per mil variation. What would be interesting to see is the d13C variance across the annual cycle. PS comments are set so that once you have one approved you should be able to post freely. Maybe you need to use the same name.

  2. Eoin Licken says:

    Hi Euan. I’m no expert on methane, but studies of snow crystals on the ground (what the French call nivologie as part of avalanche risk measurement) show that air permeates snow on the ground. So it’s not a given that methane is not coming from the ground.

    – Eoin LICKEN (Grenoble, France)

  3. Andrew says:

    Well, I’m no physical chemist, but I gather that the Bunsen coefficient of solubility of methane in saline water is inversely temperature-dependent, so I presume that there should be a temperature-driven cycle of methane entering and leaving solution in the sea and, conversely, the atmosphere.

    • Euan Mearns says:

      Andrew, this sounds as good an idea as any. But it is the exact opposite of warming leading to melting of clathrates having anything to do with rising levels of Arctic methane.

    • Greg Goodman says:

      I don’t follow that argument. Warmer temps; less solubility; more outgassing. This is the way it works with most ( all ? ) gases: the temperature dependency of Henry’s “constant”.

      That would give the expected summer peak not what Euan has pointed out.

  4. Greg Goodman says:

    Decomposition ? Cumulative integral thereof .

    Spring summer is gowth period, autumn is die off and rotting phase. Off the top of my head I would have expected Dec peak, not Jan but I’m not at all au fait with biology of the region.

    IIRC UV is a key factor in CH4 breaking down , expect it to drop quickly in the summer and linger in the winter.

  5. Peter Shaw says:

    Euan –
    I can offer a fairly parsimonious hypothesis which I’ve not seen elsewhere, and that someone with your background might miss.
    Were I given the task of destroying trace methane in air, I’d very probably shine strong light through it (optionally with a selected solid aerosol as catalyst). The rate of this sort of slow photo-oxidation typically varies as the concentration of reactant and average light intensity.
    If the only substantial change in the system is seasonal variation of insolation, you’ll get a regular cycle with its peak at the solstice – as you show in the post.
    Any temperature-related effects will reduce and skew the cycle (for an example of skew from lags in the system, see the familiar monthly CO2 record from Mauna Loa).
    The charts above are consistent with a fairly steady large-scale emission rate, lacking indications of seasonal or anthropogenic sources.

    If records for Antarctica are mirror-image, methane sources are global. If not, the data indicate a NH source, probably continental (for which I have supporting argument).

    I suggest the data “outliers” may be significant. There could generally be good mixing, with occasional atypical air masses – eg those passing over rich point-sources, and subsiding high-altitude air remote from sources.

    The amplitude of the seasonal variation offers a fairly independent estimate of methane residence-time. At steady-state, the higher the (constant) emission-rate, the larger the seasonal swing. For a continental source, I thumbsuck less than a decade.

    • Greg Goodman says:

      “I suggest the data “outliers” may be significant. ”

      Indeed, it does not seem objective or consistent to call one side “dragons breathe” but the other side erroneous data.

      ” a fairly parsimonious hypothesis which I’ve not seen elsewhere”

      Except for about two lines above that comment. 😉

      The catalisy may be ozone , which is also UV dependent, heavily so in lower stratosphere, ozone holes etc.

    • Euan Mearns says:

      Were I given the task of destroying trace methane in air, I’d very probably shine strong light through it (optionally with a selected solid aerosol as catalyst). The rate of this sort of slow photo-oxidation typically varies as the concentration of reactant and average light intensity.

      Peter / Greg I like this idea. Never occurred to me.

      “I suggest the data “outliers” may be significant. ”

      Indeed, it does not seem objective or consistent to call one side “dragons breathe” but the other side erroneous data.

      I agree. If you look at the Box article I link to you’ll see that the low outliers are absent. The data error label is mine. There may also be some data errors in the high scattered points. The way this work is done is once or twice a week they collect 3 to 7 flask samples and these are analysed as replicates. If you look at Figure 2, the red oval surrounds 6 replicates from one day – so they are all high, unlikely to be instrument error. So yes, it seems likely that there are transient puffs of higher concentration air passing over from time to time. The prevailing wind is off the frozen sea. But Prudoe Bay is not far to the East. And so I wonder if anomalous puffs may be linked to that. The data from Alert have far fewer high outliers.

  6. Greg Goodman says:

    Euan, something very odd going on with your time scale in fig 2. 🙁

  7. When all else fails, engage the data. 😉

    Here’s a plot of the CO2 seasonal cycles at Barrow in the NH and Baring Head in the SH. The cycle is far larger at Barrow and the two cycles are out of phase. This is clearly a result of seasonal sequestration and release of CO2 from vegetation/soils.

    Now here’s the same plot for methane. The amplitude of the seasonal cycles is similar in both places and the cycles are also close to being in phase:

    Preliminary conclusion: The seasonal methane cycle doesn’t change with location and is unrelated to vegetation/soils.

    • Euan Mearns says:

      Great data Roger. But how to Hell do you interpret that? You should be obliged to provide an explanation 😉 Can you post links to data sources and explain a little about how you normalised the plots. The CO2 Barrow data looks suspicious – far too regular.

      • Euan: The interpretation here is that the guy who put the second chart together screwed the dates up 🙁 Here’s what it should look like:

        Sorry about that.

        The fact that the amplitude of the methane cycle doesn’t change much while the amplitude of the CO2 cycle does is still intriguing, though.

        I calculate the seasonal cycle by subtracting a running twelve-month average of the monthly means from the raw monthly mean. The Barrow CO2 cycle is indeed as regular as shown.

        • Euan Mearns says:

          CO2 concentrations are highest at high N latitude, so that tallies with the slightly higher amplitude on the Barrow data.

          Sunlight UV destruction of CH4 is back in the lead with the No Thanks coming a close second.

          • Temperature is actually in the lead. Methane peaks at Barrow and Baring tend to occur about a month after the winter solstice and methane troughs about a month after the summer solstice; i.e slightly out of phase with sunlight but in (inverse) phase with temperature.

            Average CO2 is about 4ppm higher at Barrow than at Baring, but the large differences in the amplitude of the seasonal cycle occur because there’s a lot more carbon-chomping vegetation at high northern latitudes.

  8. Euan Mearns says:

    The great tragedy of Science: the slaying of a beautiful hypothesis by an ugly fact


  9. Peter Shaw says:

    Roger –
    Thanks for the SH data.
    A reading of your charts might be:
    > A seasonal NH consumption of CO2, which cycle is progressively attenuated by transmission over water (nb high solubility);
    > Destruction of methane globally, greater where this is concentrated close to its NH (continental?) source; this cycle not attenuated over water (low solubility).

  10. Dominik Utz says:

    What if the maximum in methane concentrations is caused by a combination of the polar vortex systems (huge wind systems practically decoupling the (higher?) atmosphere from the surrounding). Causing e.g. the ozone depletion by increasing the amount of reactive fluor chlor components -> destroying ozone? Don´t know if this would work also for the lower atmosphere and methane. As far as I know methane is a very powerful but instable green house gas and therefore easily destroyed in the atmosphere. Photo-dissociation might be a major cause? In the winter time (especially in December) the photon flux is of course zero.. polar night -> no destruction of methane but due to sublimation and/or other processes still created and increasing due to the effect of the vortex wind system which encapsulates the part of the atmosphere. When the light comes back the vortex wind system breaks ups (Exchange with lower latitudes -> decreasing amount of methane) and of course als the photo dissociation sets in again.
    The lower errors might be caused by solar storm/flare events and higher solar activity in general. Higher solar activity normally means higher UV fluxes causing higher photo dissociation rates.

    But all of these is just speculation.
    some links:
    vortext flows:
    Methane and photo dissociation:



Comments are closed.