Rapid shift in methane carbon isotopes suggests microbial emissions drove record high atmospheric methane growth in 2020–2022

Abstract

The growth rate of the atmospheric abundance of methane (CH₄) reached a record high of 15.4 ppb yr⁻¹ between 2020 and 2022, but the mechanisms driving the accelerated CH₄ growth have so far been unclear. In this work, we use measurements of the ¹³C:¹²C ratio of CH₄ (expressed as δ¹³C_CH4) from NOAA’s Global Greenhouse Gas Reference Network and a box model to investigate potential drivers for the rapid CH₄ growth. These measurements show that the record-high CH₄ growth in 2020–2022 was accompanied by a sharp decline in δ¹³C_CH4, indicating that the increase in CH₄ abundance was mainly driven by increased emissions from microbial sources such as wetlands, waste, and agriculture. We use our box model to reject increasing fossil fuel emissions or decreasing hydroxyl radical sink as the dominant driver for increasing global methane abundance.

Methane (CH₄) is the second-most abundant anthropogenic greenhouse gas and has global warming potential (GWP) of 28 over 100 y (1); as a result, CH₄ has consequential near-term radiative effects and is a prominent target for mitigation (2). Following a short pause in growth from 1999 to 2006, both the abundance and growth rate of atmospheric methane have been increasing (3). During 2020–2022, the observed CH₄ growth rate reached a record high since NOAA measurements began in 1983, averaging 15.4 ± 0.6 ppb yr⁻¹ (4). Understanding the mechanisms driving this accelerated growth is essential for predicting its future climate impact and providing scientific support for climate mitigation strategies (2).

The carbon isotopic composition of atmospheric CH₄ (δ¹³C_CH4) is a powerful tool for tracking the sources and sinks of atmospheric CH₄. Different CH₄ sources have distinctive δ ¹³C_CH4 values: Microbial CH₄ emissions (wetlands, livestock, landfills, etc.) have lower δ¹³C_CH4 values (global mean of –62‰) than pyrogenic (biomass and biofuel burning, global mean of –24‰) and fossil fuel CH₄ emissions (global mean of –45‰) (5). Various sinks of atmospheric CH₄ also have distinctive isotopic effects. Therefore, combined observations of atmospheric CH₄ mole fraction and δ¹³C_CH4 can provide unique constraints on the changes of global CH₄ sources and sinks during the post-2006 rapid CH₄ growth.

The National Oceanic and Atmospheric Administration’s Global Monitoring Laboratory (NOAA/GML) has been carefully monitoring the global CH₄ burden through the Global Greenhouse Gas Reference Network (GGGRN) for over four decades. The collaboration between NOAA/GML and the Institute of Arctic and Alpine Research (INSTAAR) at the University of Colorado Boulder has enabled δ¹³C_CH4 measurements from the GGGRN since 1998, currently measuring weekly or biweekly from 22 globally distributed background sites (6). The dataset has been widely used for studying the evolution of global CH₄ sources and sinks (7–9). Here, we report our most recent observations of atmospheric CH₄ mole fractions and δ¹³C_CH4 values through the end of 2022 and then use a box model to examine and quantify the contributions of potential drivers of the record-high CH₄ growth rate.

Results and Discussion

The global average methane growth rates in 2020, 2021, and 2022 reached record levels of 15.2 ± 0.45, 17.9 ± 0.45, and 13.1 ± 0.8 ppb yr⁻¹, significantly higher than the average growth rates of 9.2 ppb yr⁻¹ in 2014–2020, and 5.3 ppb yr⁻¹ in 2008–2014 (Fig. 1A). Meanwhile, we observed the lowest global average δ¹³C_CH4 in the observational record: –47.67 ± 0.01‰ in 2022. The global δ¹³C_CH4 growth rate from 2020–2022 was –0.09 ± 0.01‰ yr⁻¹, a much faster decrease than –0.04 ± 0.02‰ yr⁻¹ in 2014–2020 and –0.03 ± 0.02‰ yr⁻¹ in 2008–2014 (Fig. 1A).

Fig. 1.

(A) Trend of globally averaged CH₄ abundance (in gray) and δ¹³C_CH4 (purple) from the NOAA/GML GGGRN. Mean growth rates of CH₄ mole fraction and δ¹³C_CH4 are shown for the following time periods: 1983–1998, 1999–2006, 2008–2014, 2014–2020, and 2020–2022. (B) Colocated δ¹³C_CH4 measurements at Alert (Canada), Svalbard (Norway), and Antarctica by INSTAAR, NIWA, TU/NIPR, and MPI. Each dataset is fitted with a trend in the same color.

The rapid decrease in δ¹³C_CH4 in 2020–2022 is observed by multiple long-term monitoring programs: Max Planck Institute (MPI), National Institute of Water and Atmospheric Research (NIWA), and Tohoku University and National Institute of Polar Research (TU/NIPR, Fig. 1B), which have independent sampling schemes, analytical techniques, and data processing and quality protocols. These observations exhibit similar trends confirming the accelerated decreasing trend in atmospheric δ¹³C_CH4 in 2020–2022 (SI Appendix).

To investigate potential drivers for the rapid CH₄ growth, we used a box model (10) to reconstruct the time series of global average CH₄ mole fraction and δ¹³C_CH4. Initial model emissions and sinks prior to 1999 are based on optimized values from a global 3-D inverse model (8) to allow the model to reach steady state with respect to CH₄ mole fractions and δ¹³C_CH4 during 1999 to 2006. We treated the time series as four segments (1999–2006, 2008–2014, 2014–2020, and 2020–2022), each with distinct CH₄ and δ¹³C_CH4 growth rates (Fig. 1A). We conducted different simulations to test the isotopic response to possible CH₄ growth drivers (Fig. 2): 1) decreased OH in the troposphere (OH); 2) increased fossil-fuel emissions (FF); 3) increased microbial emission (MICR). In each simulation, we adjusted the flux of each source/sink category in each time segment to match the observed CH₄ growth rate and then compared the resulting simulated atmospheric δ¹³C_CH4 values to our observations.

Fig. 2.

(A) Modeled response of CH₄ mole fraction and δ¹³C_CH4 due to different CH₄ growth drivers. (B) Emissions and CH₄ lifetime relative to OH for each scenario.

Our model shows that only the MICR simulation displays a decrease in δ¹³C_CH4. However, increasing only microbial emissions resulted in lower δ¹³C_CH4 than the observations, so we also adjusted fossil fuel emissions to best fit both the observed CH₄ mole fraction and δ¹³C_CH4 (Fig. 2). Our best-fit result of the MICR simulation (SI Appendix) required an increase of microbial emissions over the steady state mean by 14 Tg yr⁻¹ in 2008 with a concurrent increase in fossil emissions of 10 Tg yr⁻¹; then in 2014, the microbial emissions increased by an additional 22 Tg yr⁻¹, and fossil emissions increased by 3 Tg yr⁻¹. These results are consistent with previous inverse modeling studies (8, 11, 12) that suggested approximately 85% of CH₄ growth during 2007–2020 was due to increased microbial emissions. To capture the rapid growth in CH₄ mole fraction and the decline of δ¹³C_CH4 in 2020–2022, our model suggests an increase in microbial emissions of 32 Tg yr⁻¹ in 2020 with no increase in fossil CH₄ emissions required to match observations.

Decreases in biomass burning emissions between 10 to 30% over the past 2 decades (13, 14) could also explain some of the observed changes in δ¹³C_CH4. Such decreases allow for more fossil emissions due to high δ¹³C_CH4 from biomass burning. However, even considering the decreased biomass burning emissions, our model still suggests the post-2020 CH₄ growth is almost entirely driven by increased microbial emissions (SI Appendix). Likewise, we modeled 1) a small increasing trend in OH number density (15), 2) an alternate OH fraction factor, and 3) a more negative δ¹³C_CH4 value of fossil fuel emissions. In all scenarios, emission increases dominated by microbial sources are required to track both atmospheric CH₄ and δ¹³C_CH4 (SI Appendix). In this underconstrained problem, there are many ways to adjust model parameters to fit the model to the atmospheric data; however, all of the reasonable solutions require very large increases in microbial emissions. (An example of an unrealistic scenario would be an extreme case where biomass burning emissions decline to zero by 2020; only then do fossil fuel emission increases become comparable to those from microbial sources.)

Atmospheric δ¹³C_CH4 does not allow us to differentiate between anthropogenic microbial sources (livestock, landfills) and natural ones (wetlands), so further study is necessary to investigate the potential climate feedback hypothesis (16). However, our box model suggests that microbial emissions played an even more significant role during 2020–2022 than in the years since 2008, which is in general agreement with studies that emphasize the key role of wetland emission increases to the recent global CH₄ budget (11, 12, 17, 18).

Materials and Methods

Atmospheric δ¹³C_CH4 of background air samples collected from the GGGRN are measured using an Isotope Ratio Mass Spectrometer equipped with a custom-built extraction system which traps methane from whole air, focuses the sample, separates it from other carbon-containing compounds, combusts it to CO₂, and measures it relative to a standard (6). Data extension and integration techniques were used to convert global measurements of CH₄ and δ¹³C_CH4 from the GGGRN into global averages and growth rates.

We used a two-box model with time steps of 0.2 y to investigate changes in sources and sinks that could match our observations of CH₄ and δ¹³C_CH4. The box model specifies CH₄ emissions from microbial, fossil, and pyrogenic sources with prescribed δ¹³C values of –61.7‰, –44.8‰, and –24.3‰, respectively (5). Sinks include uptake by soil microbes, and oxidation by OH, Cl, and O(¹D), all of which have associated kinetic isotope fractionation factors. The model was tuned to match observations from 1999–1996 and then adjusted to test the isotopic effects of different source/sink scenarios. More details are available in SI Appendix.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

Data have been deposited in NOAA Global Monitoring Laboratory Data Repository (https://doi.org/10.15138/JQEV-PF31) (19).