Methane madness

by Stephen Shafer on August 10, 2019


Rising unacknowledged  emissions of methane from the natural gas supply chain  are dangerously under-estimated as  a driver  of global heating and must be ended.

Stephen Q. Shafer  MD MA MPH



                                                                                                  Aerial view of fracking pad in Pennsylvania  photo source Smithsonian magazine

Summary:  The  natural gas industry  is now  the largest source  of human-influenced methane releases to the air, in the form of “fugitive emissions”  leaked  from the industry’s supply chain.  Though recent measurements are lacking, annual emissions are surely  rising  worldwide as natural gas production booms  via  ” fracking.”    This paper presents a range of  possible impacts from  these fugitive emissions on the atmosphere’s growing  capacity  to trap heat radiated from earth over the next two decades.  The range uses  estimates of leakage proportions from the literature, to which a GWP20 is applied instead of the customary   GWP100.   

If  methane emissions keep going up  2019-2023,  this short-lived super-potent greenhouse gas will wreck all  hopes of keeping global  temperature rise under + 1.5 C   as of 2030.   Those hopes may already be vain.  Getting  total annual human-influenced methane  emissions to stop rising  will improve the slim chance.  Natural gas production cannot be allowed to increase year-on-year.    A worldwide ban on new fracking  is the best action.

           It’s no secret now to dedicated climate activists and fracktivists that the potent  greenhouse gas (GHG)  methane leaks into the atmosphere from the  natural gas supply chain.  Even those  who’ve  heard about these  fugitive emissions, however, usually  know them as a proportion of total  natural gas production, e.g. “2%”  or “7%.”   It’s better to see them as  a quantity  that can be compared on a mass  and temporo-spatial distribution basis  to other streams of atmosphere-heating GHGs such as methane from livestock or  CO2 emitted  in  air travel.  This short paper provides a GHG analysis of  natural gas  through an assessment that though  not full life cycle  factors  in supply chain losses, not just combustion. 

            To make this  briefing  useful to non-scientists,  scientific terms and abbreviations are minimized.  An unavoidable few are listed below.  It supposes basic knowledge, e.g. that combusting fossil fuels releases carbon dioxide, or that natural gas is mostly methane.    It skips over the physics of  how greenhouse gases heat the atmosphere, which are discussed   a  shorter blog  titled  “Why We Must Control Methane Emissions Now.”


cf          cubic foot   

CH4    methane

cm        cubic meter in this context cubic meter, not centimeter 

CO2      carbon dioxide

CO2-e          CO2 equivalent, which   measures  the efficiency of a specified mass of a GHG at trapping heat in the atmosphere   over a specified time horizon   relative to that of an equal mass of CO2 released  at the same time.  CO2-e has units of mass, e.g. metric tons.  To find CO2-e multiply the  mass of  a  GHG emission  by that gas’s  global warming potential  (GWP).

fugitive  refers to gaseous emissions to air that are neither intended   nor  readily detectable or measurable

GHG     greenhouse gas

Gt          gigaton = billion metric tons = 1 petagram (Pg)

GWP      global warming potential expresses the average efficiency of  a  GHG  at trapping  heat in the atmosphere over a specified time span relative to  that  of  CO2  at the same time.  GWP  has  no mass units, but as noted, does have a time dimension,  often called the “horizon.”  A 100 year horizon is most common. For short-lived atmospheric pollutants like methane a shorter horizon such as 10 or 20 years   is warranted,  though  not adopted  by the Intergovernmental Panel on Climate Change.

mt          metric ton

mmt     million metric tons = 1 Tg = 1 teragram = 1 megaton    Tg used for methane and mmt  for CO2 and CO2-e

NG       natural gas  

tcf         trillion cubic feet

Tg         teragram = 1 mmt


Part 1.  Why “natural gas” may not be better for the atmosphere than coal.   

Natural gas is seen widely as a  “bridge fuel” from  coal  to  renewables, particularly  for  electricity generation , which  uses about  35%   of  annual production.  Industrial (e.g. pulp and paper, stone, glass ) uses 34%,  residential  17%  and commercial 12%.  Its designers  build and try to extend   this bridge saying  it  causes  less global warming  and air pollution than would  the coal  (or oil) consumed  in electricity  generation .  “Better  than  coal”  fits  over  a  50-100 year  future in which  the proportion of natural gas withdrawn that is lost  to the air as methane is low.  The opposite applies to  a shorter term (ten-twenty years),  especially with   higher  loss proportion.  Tanaka et al (2019)  and Xiaochun Zhang et al (2016) offer good discussions of the debate with many nuances and caveats.  Among the imponderables to be considered are these

  • time horizon and, related to that ,
  • pace at which  renewable energy sources penetrate the electricity generation sector
  • proportion of NG lost from supply chain as methane

            I believe the time horizon must be twenty years at most.  Within that time, renewable energy sources (wind and solar particularly) must be supplying most electricity.  To paraphrase Fridtjof  Nansen, “The impossible just takes a little bit longer.” Encouraging news comes often.

            Compared to the two papers above, this discussion is primitive but illustrative.  It starts by comparing  the GHG emissions in metric tons  CO2  due to combusting a quantity of   natural (roughly 95%  methane)  to the emissions  as CO2-e associated with  losses of methane from moving that same quantity through  the natural  gas supply chain,  from in-ground  to end  user pre- combustion.

            Combustion  of  41,900 metric tons of NG  (~ = 2.16 trillion btu ~= 2.12 bcf  ~ = 60 million cm  )  releases 115,000 metric tons of CO2 [ @ 0.053 metric tons CO2 per million btu].    

Important aside on measurement units:  The quantity 41,900 mt  used here  has no special significance .  It has roughly the same  heat energy value that is contained  in  100,000 cubic meters of liquefied natural gas carried by a large LNG tanker.  Natural gas is not usually measured by mass.  It  is not seen as a greenhouse gas   but as a commercial material  of  very low density that is theoretically 100% contained until combustion.  Thus,  production of it  is counted  in volume (e.g.  cubic feet or meters)   or heat value (e.g.  btu or tonnes of oil equivalent),  not in mass.   GHGs like methane, emphatically not contained,   are measured by mass.  This paper is thus  unconventional, but not wrong, to  measure NG by mass.  As cubic meters  in the pipe,  natural gas  is “cleaner than coal, ” a “bridge fuel.”  As metric tons outside  the pipes it  is methane, the bane of vegans and fracktivists alike,   a GHG  86 to 104 times more efficient at trapping heat  in its ten years of life in the atmosphere than is CO2.

            A  conservative estimate for the proportion of natural gas  lost as unacknowledged, or fugitive,  emissions  of methane is 2.3%.  Raimi and  Aldana  tabulated  29 reports, all after 2011.  The point estimate in nine of these  was 4% or higher, in four, it was   >2% and <4%.  In the other sixteen, it was  less than 2%.

            Applying  five  hypothetical  supply-chain loss proportions (2%, 2.5%, 3%, 3.2%  and 4%) to the 41,900 metric tons (~= 2.16 trillion btu) of NG  in the example gives supply-chain methane losses in metric tons as 840, 1040, 1260, 1341 and 1670,  respectively.  All five proportions  are  in the lower 55% of the range given by Raimi and Aldana, 

            The  Global Warming Potential  (GWP20)  of methane for a 20 year time horizon is 86,  higher than the GWP100,  cited in most recent places as 25  to  34.   The strong  case for GWP20  for methane is presented elsewhere.   Applying GWP20 to a  loss proportion of  3.2%,  then, fugitive emissions of methane  related to marketing  41,900 mt of NG amount to  115,000 mt CO2–e, equal to  the release from combustion alone

 total emiss graph

Graph 1.  GHG emissions in thousand metric tons CO2-e  for life cycle of 2.16 trillion btu  (~=41,900 mt )natural gas, from below-ground  through  combustion,  by  proportion of natural gas  lost as methane in supply chain and type of gas emitted,  CO2 vs. CH4     GWP20 of 86 used for CH4. Total emissions = sum of the blue and red bars. For 3.2% losses, that is 230,000 mt CO2-e.

            Graph 1 displays CO2-e emission from  the life cycle of 2.16 trillion btu  (~41,900 metric tons) of  natural gas.   The size relationships between combustion emissions of CO2 and CO2 equivalents of fugitive emissions of methane will be the same for each loss proportion regardless of the quantity of NG.  With a  supply-chain loss proportion of  just  3.2%,  for example,  fugitive emissions of methane from natural gas are 1341 mt.   By  GWP20, this converts  to  115,000 mt CO2-e,  equal to  the mass of CO2 released from combustion of 41,900 mt NG.  Thus, marketing and burning any quantity  of natural gas doubles the mass of CO2-e when the loss proportion is as low as   3.2%.

            Parties promoting the expansion of natural gas use declare  it’s cleaner than coal.  In regard to GHG, the assertion is based only on CO2 released from combustion.   Combusting 2.16  trillion  btu of  bituminous coal at 206 lb CO2 / million btu does  emit 202,000  mt CO2,  much more than the 115,000 mt CO2 from combusting 2.16  trillion  btu  of NG.   Nonetheless, with  a supply-chain methane loss proportion of only 2.5 % added to the combustion emissions, natural gas (115,000 mt CO2 from combustion plus 89,000 mt CO2-e from fugitive emissions = 204 mt CO2-e)  looks worse than coal.   That comparison, however,  is also biased.

            The 202,000  mt CO2-e figure for coal does not account for other GHG releases in the life cycle of coal from extraction to end-use,  such as  mining operations, washing, transportation and methane releases from underground mining.  Nor does it incorporate black carbon,  a short-lived aerosol enormously more efficient than CO2 at trapping heat in the atmosphere.    I can’t do the math on black carbon, can only say that comparing bituminous coal to NG on a life cycle basis might show that  natural gas even with  (say) 4%  supply-chain loss would  look better for atmospheric heating  than coal,  btu for  btu.  I don’t have the needed figures.

            Yet the question of which is “better” is irrelevant.  Humankind cannot  afford  to  not  rapidly bring to nil   CO2 emissions from  all sources  and  to  not start cutting methane emissions from all sources beginning now.  Tanaka et al (2019) observe   “Several studies caution about potential side-effects [of NG as a bridge fuel] —that an expansion of natural gas may delay the deployment of less carbon-intensive technologies such as renewables, representing carbon lock-in from fossil fuel infrastructure, and thereby postponing the transition to a decarbonized society.”  “May” here is the wrong word.  It should be “will.”  Laissez faire on natural gas production  subverts  renewables.   It’s as if in January 1942 the War Production Board   had not halted all production of civilian motor vehicles in favor of war machinery but had reduced it by just 50%, proclaiming “We can do both.”   We couldn’t have then, and we can’t now.  

            The best way to reduce  methane emissions from all human-influenced sources is to go after the biggest single stream: methane losses in the natural gas supply chain.  This will not be done through promises  to  reduce the loss proportion by upgrading equipment and procedures.  It  must be done by ending the growth of  natural gas production before 2021,  then starting  a steady trend downward in production

Part 2  Global and domestic trends in natural gas marketing,  and  how they relate to trends in methane emissions        

            Figures for annual emissions of methane worldwide are probably not very accurate.  There  is much disagreement among different sources. A respected source is the multinational Emissions Database for Global Atmospheric Research (EDGAR).  I could not  find  EDGAR  figures beyond  2012 in the public domain.  For a perspective on world emissions to that time, Graph 2 below  shows stable, even slightly lessening emissions from 1990 through 2002 then a 16% rise over the next decade.


  Graph 2.  Annual human-influenced methane emissions worldwide by year 1990-2012 in Tg (million metric tons)  Data source EDGAR       graph by Shafer  On the x axis,  92 is 1992,  102 is 2002 and so forth.

            Worldwide  production of  natural  gas  has increased much more than have methane emissions recorded by EDGAR. over the same epoch,   by about 70%  1990-2012, as seen in Figure 1 below. 

 worldNG since1990

Figure 1.   World production natural gas 1990-2018 by region and year in billion cubic meters

            Figure  2  below,  from the Global Gas Report     graphs later years 2010-2016   in the above series.  It shows that all the considerable step-up  in NG  “production” since 2010 (and undoubtedly for  a couple of years before that )  is “unconventional.”  

 natgasprod world

Figure 1. World natural gas/fossil gas  production by year 2010-2016 and by method of production in billion cubic meters.  Link to source

            Extending the series in Figure 1 with   a 2017 (3768 bcm)  datum from another source, world  production of  NG  rose by 507 bcm/yr between 2010 and 2017, an average of  +  72 bcm/yr .   That is an increase of about 48.1 million metric tons (Tg)  natural gas    per year during the current boom in   “unconventional”   production,  better known as  fracking.   If we apply to the annual world production figures through 2017 three loss proportions within the range of estimates given above  by Raimi and Aldana ,  a steady rise in fugitive methane emissions from the natural gas supply chain is predicted..  Graph 3 below presents this.

 Tg methane emitted

 Graph 3.  Imputed methane losses in Tg  from natural gas supply chain worldwide by estimate of  %  total production presumed  lost in supply chain and  year    Data source for NG  production: p. 14 of

            Before speculating further about how much  methane escapes annually from supply-chain losses, a salient question regarding the worldwide trends above  must be addressed.  “If methane losses from the natural gas supply system are an appreciable fraction  of  production,  why have methane emissions risen  by only about 20% in the same epoch during  which natural  gas production went up 70% ? “ A related,  harder,  question applies to  the  USA.   “Why are recorded annual methane emissions flat or falling in this country (Graph 4 below)  while  natural  gas production has climbed rapidly in the same epoch (Graph 5)?”

            There is not a solid answer to  either question,  yet the divergence between natural gas production growth and  official CH4 emission figures does not  lessen the significance of supply-chain losses as a major source of methane.   The emission figures for USA and  the  world are very likely under-estimated.   We know that atmospheric methane levels are rising apace since about 2006 (see Appendix),  and no one denies  that at least some methane is lost from the supply chain.  It is  a safe assumption that if natural gas production is trending up steeply over time, then methane emissions from that stream are as well. 

 Ch4 by yr USA

 Graph 4  Methane emissions Tg/yr USA by year 2002-2017.  Emissions in CO2-e from  source were divided by GWP100 figure of 25 to convert to Tg  methane.


 natgas usa by yr

Graph 5.  Natural gas marketed production  by year 2006-2018 USA in trillion cubic feet  data source      

Part 3.  Projecting the amount of methane fugitive emissions and their contribution to atmospheric heating-up        

            Returning to graph 3  above,   the 4% figure for 2017 is 99.2 Tg/yr.  4%  is slightly above the median value for loss % among the twenty-nine  reports tabulated by Raimi and Aldana  mentioned above.  99 Tg/yr is less than the 157 Tg/yr   for a recent year attributed to the oil and gas sector  by Howarth in a 2019 lecture,  but  close to  the 96 Tg ascribed by Saunois et al to “fossil fuels” for 2012.  99.2 Tg  CH4 adjusted by GWP20 means 8.5 billion metric tons CO2-e  of  methane worldwide for one calendar year from NG alone.   Whatever one sets as an acceptable carbon budget  for  up  to  whatever future year one chooses, 8.5 billion metric tons  CO2-e  in  one calendar year from just one of several sources of the second-most important GHG  is intolerable.

            Non-CO2  GHGs  (methane, nitrous oxide and F-gases) and aerosols are not counted in most carbon budgets, in which  aggressive mitigation is [prayerfully] assumed.  If they were, the budgets for molecular CO2 itself  would have to be smaller than those proffered,   and very much smaller when   GWP20 of 86   is used for methane,  (as it should be  for any epoch  less than twenty years).        

            Some idea of the role  of  methane in a hypothetical all-GHG accounting (as opposed to  one for  CO2 only)  can come from  USA data; comparable figures  for the world are hard to find.  From a recent  EPA inventory through 2017,  Tg of methane  emitted from all human-influenced sources for each chosen year were calculated by  dividing the  CO2-e  listed in mmt  by 25, the  GWP100  for methane often used by EPA.   Multiplying that imputed  CH4  annual release in Tg (mmt)  by GWP20  of 86 returned a much higher value for methane emissions in CO2-e  than is shown  in the EPA table.    The results appear in graph 6 .   For each of the four years included 2002-2017,  there are two columns.  The  CO2-e ascribed to  each gas in the original EPA table is shown in column a,  while column b shows the CO2-e when methane has a GWP20 of 86.  On the average, the grand total in column b is about 24% higher than that in column a and the share of total  due to methane almost three times higher.


 Graph 6.  GHG emissions USA in four separate years in million metric tons CO2-e, by gas (CO2 vs CH4 vs  “all other”  (N2O + F-gases)).  For each year the a column uses GWP100 for methane and the b column GWP20

            Graph 6  is most disturbing.   It shows that when methane is given a GWP20,   total US GHG emissions as CO2-e have been, are being or will be  under-counted by about 24% , approximately 1.6  Gt  CO2-e (1.6 billion metric tons CO2-e).  1.6 Gt CO2-e  is  more  than the emissions (1.4 Gt CO2-e)  of the entire industrial sector of the USA  in 2015. Fugitive emissions from fossil fuels  (largely natural gas) make up, conservatively, 40% of that 1.6 Gt CO2-e or 640  mmt  CO2-e .  This is  more than the 582 mmt CO2-e attributed (using GWP100  of 25 )  to agriculture for all GHGs in 2017.

            No matter what GWP is used ,  the fact remains  that,  because of methane’s  short lifetime in the atmosphere,   emissions rising at any rate add to atmospheric heating-up.  Level annual emissions contribute to atmospheric heating without adding to it, while falling total emissions cause relative cooling.  If humankind wants a fighting chance of keeping global average surface temperature less than 1.5 degrees centigrade higher after 2030  than it was in the pre-industrial baseline, natural gas withdrawals must stop rising year-on-year  in 2020.   This turnaround is  the only  avenue for significant course correction on  GHG management over  the next ten years that can have a discernible effect when those years have elapsed. 

          As Figure 1 showed, all the annual increment in natural gas production  worldwide in the last decade  is due to “unconventional methods” i.e. fracking.  A worldwide ban on new fracking  must be enacted in 2019.  The US should lead.

              Thanks to Ken Dolsky for  his  meticulous help with  line-editing   All remaining  errors of orthography or clarity or fact are  on my head.

             Permission is hereby granted  to  reproduce this blog  elsewhere in whole or in part as long as the permalink is cited .


Appendix  on trends in atmospheric methane levels

Global atmospheric methane levels have risen enormously in the last 150 years.  Click on the link to see a graph  too faint to reproduce here.   Since about 2006  there has been a decided upsurge after a brief plateau

global ch4 levels

Additional readings.

Millar et al  (Nature Geoscience Sept 18 2017:10:741-747)  850 Gt CO2-e

Glen Peters  Beyond Carbon Budgets Nature Geoscience 11 June 2019 378-383 withdrawals withdrawals and production

Katsumasa Tanaka,  Otávio Cavalett, William J. Collins &   Francesco Cherubini  Asserting the climate benefits of the coal-to-gas shift across temporal and spatial scales Nature Climate Change volume 9, pages 389–396 (2019)

Xiaochun Zhang,  NP Myhrvold,  Z Hausfather  K  Caldeira  Climate benefits of natural gas as a bridge fuel and potential delay of near-zero energy systems  Applied Energy (2016) 167: 317-322




Previous post:

Next post: