The definitive CO2/CH4 comparison post

There is a new push to reduce CH4 emissions as a possible quick ‘win-win’ for climate and air quality. To be clear this is an eminently sensible idea – as it has been for decades (remember the ‘Methane-to-markets’ initiative from the early 2000s?), but it inevitably brings forth a mish-mash of half-remembered, inappropriate or out-of-date comparisons between the impacts of carbon dioxide and methane. So this is an attempt to put all of that in context and provide a hopefully comprehensive guide to how, when, and why to properly compare the two greenhouse gases.

Historical comparisons

First of all, let’s be clear about the relative magnitude of the gas concentrations. In 2020, CO2 was at ~410 parts per million, while CH4 was around 1870 parts per billion (or 1.87 ppm, a factor of more than 200 smaller). However the relative rise since the pre-industrial is three times larger for CH4, around 150%, compared to the 50% increase in CO2.

The radiative forcing from these changes in concentrations can be easily calculated using standard formulas (from Etminan et al, 2016 which supersede the slightly simpler ones from IPCC TAR), as about 2 W/m2 for the CO2 change and 0.65 W/m2 for CH4.

But methane’s role in atmospheric chemistry and as a source of stratospheric water vapour means that it has a bigger effect on climate than just the direct effect of its concentration. Methane emissions have a feedback on its own lifetime, serve as an ozone precursor, and reduce the production of sulphate and nitrate aerosols (and consequently indirect cloud-aerosol effects), all of which amplify its net warming effect to about 1.2 W/m2 (to about 60% of the CO2 effect since 1750). There is also a very small impact of the CH4 oxidation to CO2 itself for any fossil-fuel derived methane.

(a) Historical radiative forcing by emissions (W/m2) (IPCC fig TS.15a).(b) Historical concentrations of CO2, CH4 and N2O (IPCC AR6 fig TS.9b)

This implies that if you convert the impacts of each set of emissions into temperatures, as was done in the IPCC AR6 report, you get about 0.75ºC from the changes in CO2 and 0.5ºC for CH4 (from the late 19th C, see figure below) or 1ºC and 0.6ºC, respectively, from 1750. Thus despite the smaller concentrations and changes in methane compared to carbon dioxide, the impacts are comparable.

Historical contributions to global warming by emissions (IPCC AR6 SPM)

Stocks and flows

Before we go any further though, we need to understand that the effective perturbation time for CO2 and CH4 in the atmosphere are very different. CO2 emissions embed themselves in the atmosphere/biosphere/upper-ocean carbon cycle and have very long-term impacts (under natural conditions, some 15% of the CO2 perturbation will still be in the atmosphere thousands of years from now). In contrast, methane has a perturbation time-scale of about 12 years. This implies that the impact of CO2 on temperature is cumulative (a function of the total emitted CO2 or stock), while the impact of CH4 is a function of current (~decadal) emissions (the flows). Stabilizing temperature effects from CO2 means getting down to net-zero anthropogenic emissions, while stabilizing temperature effects from CH4 means simply stabilizing emissions.

The impacts of emissions of CH4 compared to CO2 then will have a time-varying component. Over a short time, the enhanced effectiveness of methane will be important but on very long time scales the effects of CO2 will be dominant. This is the source of the difference between the “Global Warming Potential” (GWP) numbers calculated at 20 years or 100 years which have been used for decades. You might recall that GWP is defined as the ratio on per-kg basis of the temperature impact of other greenhouse gases compared to CO2 over a specific time period. But as is clearly stated in AR6, the suitability of comparative emission metrics depends on your end goal or values.

For instance, if you use GWP-100 to trade off emissions on the way to a temperature stabilization scenario, it simply doesn’t work (since you can’t balance any net CO2 emissions with a particular level of CH4 emissions – you would need to have constantly decreasing CH4). Hence, newer concepts like GWP* have been developed that take that into account. Nonetheless, the UNFCCC (and the EPA) use the GWPs from IPCC AR4 for calculating CO2eq emissions and have not updated them as the science has progressed.

Forward-facing comparisons

People tend to be most interested in comparisons related to future choices, and it’s worth bearing in mind that while there are many ways to do this, most don’t relate to real choices that people have, nor do they clearly match up with a consistent set of values. I’ll return to that issue below. So let’s go:

  • Molecule-to-molecule concentrations: On a per-ppm basis, methane is 25 times more effective as a direct greenhouse gas. Including the indirect effects, increases that to 45 times as effective.
  • kg-to-kg: On a mass basis, methane is 70 times more effective as a greenhouse gas. This takes into account of the different molecular weights of the molecules. That would mean 126 times as effective including indirect effects.
  • kgC-to-kgC: an equal amount of kgC as CH4 or CO2 gives rise to the same ppm change, so kgC-to-kgC, methane is again 45 times more effective as a greenhouse gas.
  • kg to kg emitted: This is where it starts to get hairy because of the different timescales. Current (AR6) estimates for fossil-sourced methane are ~83 for GWP-20 and ~30 for GWP-100 (AR6 Table 7.15). (It’s slightly smaller than this for biogenic (non-fossil) methane since the oxidation product of CO2 in that case is carbon neutral). The assessed uncertainties in these values (largely related to direct and indirect aerosol effects) are ±25 and ±11. The AR4 value for methane GWP-100 was 25.
  • kgC emitted to kgC emitted: For some applications, for instance judging the impact of flaring natural gas vs. releasing it directly into the atmosphere, the kg-to-kg comparisons are not relevant, since the same amount of carbon is being emitted, rather than the same total mass. For that, the GWP-like value over 100 years, choosing to release methane directly would be 30*16/44 = 11 times worse than flaring [Corrected 9/20/21].
  • Emissions for temperature stabilization: Each additional GtC of carbon dioxide contributes to about 0.00165ºC of eventual warming (the TCRE), while a sustained TgCH4/yr of methane emissions (0.00075 GtC/yr), leads to ~3 ppb increase of methane concentrations (AR6 Table 5.2), about 0.0024 W/m2 in total radiative forcing, and, assuming a median climate sensitivity of 3ºC for 2xCO2, roughly 0.002ºC of equilibrium global warming. That implies you need a sustained reduction of 0.8 TgCH4/yr (0.0006 GtC/yr of methane) to compensate for a one-off GtC pulse of CO2.

Whatever way you slice this it implies that CH4 reductions have an outsize effect on climate, as well as an undeniably positive impact on air pollution, crop yields and public health (mainly through ozone reductions). It is therefore not a complicated decision to pursue methane reductions, taking care not to assume that doing so gets you off the hook for reducing CO2, whatever the EPA says.

I’d like this page to be useful and current, so if you think I should add an additional comparison, or use case, or if you think I’ve got something wrong, please let me know in the comments.


  1. M. Etminan, G. Myhre, E.J. Highwood, and K.P. Shine, “Radiative forcing of carbon dioxide, methane, and nitrous oxide: A significant revision of the methane radiative forcing”, Geophysical Research Letters, vol. 43, 2016.

The post The definitive CO2/CH4 comparison post first appeared on RealClimate.