Impact of interannual and multidecadal trends on methane-climate feedbacks and sensitivity

Methane is a significant anthropogenic greenhouse gas, and its atmospheric concentration has increased sharply since 2007, reaching record levels in 2020 and 2021 despite declining fossil fuel emissions. This increase, coupled with a decrease in the δ¹³CH₄ isotope ratio, points towards biogenic emissions as a major driver, potentially linked to surging emissions from wetlands and permafrost, alongside rising fossil fuel emissions and reduced biomass burning. A further contributor is the weakening of atmospheric and soil methane sinks. The uncertainty surrounding methane lifetime feedback necessitates improved understanding of past CH₄ and δ¹³CH₄ variations. Warming generally promotes positive methane-climate feedbacks through enhanced methanogenesis, while increased atmospheric hydroxyl radicals ('OH) leads to increased oxidation and a negative feedback. However, exceptions exist, such as secondary positive feedback from increased atmospheric CO limiting 'OH, or from increased biogenic volatile organic compounds (BVOCs). Soil sinks also exhibit contradictory behavior, sometimes decreasing due to increased precipitation. Precipitation (Pr) significantly influences emissions, with high Pr leading to increased wetland emissions and low Pr resulting in increased wildfires. Sea surface temperature (SST) influences the sink through 'OH and chlorine, and indirectly via LSAT, Pr, and terrestrial 'OH. Recent findings highlight the significant yet uncertain role of oceanic methane production by cyanobacteria and phytoplankton. This study uses a causal analytical method to quantify the contributions of temperature and precipitation to changes in CH₄ and δ¹³CH₄, differentiating feedback signs to identify underlying processes.

The literature extensively documents the recent acceleration in atmospheric methane concentration, highlighting the roles of various sources (e.g., wetlands, permafrost, fossil fuels, agriculture) and sinks ('OH, soil). Studies have investigated the isotopic signatures of methane sources, showing differences between biogenic and fossil fuel sources. The impact of climate change on these sources and sinks has been a focus, including the effects of temperature, precipitation, and sea surface temperatures on methane emissions and oxidation rates. While the general understanding points towards a complex interplay of positive and negative feedbacks, the quantification of their relative contributions and their temporal variations remain uncertain, particularly regarding the role of the atmospheric and soil sinks, and the contribution of oceanic sources. This research builds upon previous studies by employing a more sophisticated causal analysis to disentangle these complex interactions.

The study employs an empirically verified causal analytical method to quantify the causal contributions of temperature (T) and precipitation (Pr) to changes in atmospheric methane concentration (CH₄) and its isotope ratio (δ¹³CH₄). A material balance equation is used to separate climate-driven contributions from non-climate-driven contributions, differentiating positive and negative feedback processes. The causal contributions from T and Pr to CH₄ and δ¹³CH₄ are quantified using the normalized information flow (nIF), adjusted by their covariance to differentiate positive and negative feedbacks. Two approaches are used to assess the contributions of SST: (i) exclusive means, assuming negligible direct oceanic influences; and (ii) area-means, assuming negligible indirect influence between SST and terrestrial contributions. Non-climate contributions are estimated by subtracting the area-means from observed trends. Five periods are classified for CH₄ and two for δ¹³CH₄ based on variations in climate-driven contributions. The normalized information flow (nIF) quantifies the causal sensitivity, accounting for uncertainties and differentiating positive and negative feedbacks. The method utilizes monthly zonal mean CH₄ and δ¹³CH₄ data from NOAA and WDCGG, respectively, along with gridded LSAT, SST, and precipitation data from NOAA. Data processing involved two-step regressions to reconstruct station data into monthly zonal mean matrices, removing seasonal trends to improve interannual variability estimates. The analysis uses a 49-month moving window to capture interannual variations. The global mean methane-climate feedback sensitivity is calculated as a function of global mean surface temperature (GMST) anomalies.

The study's analysis reveals that both positive and negative methane-climate feedbacks contribute to the observed increase in atmospheric methane concentration. Interannually, positive feedbacks (e.g., increased wetland emissions and wildfires due to higher land surface air temperature) are often followed by increased methane due to weakened sinks (lower sea surface temperatures). Over decadal timescales, the study identifies alternating rate-limiting factors for methane oxidation: positive feedback dominates when CH₄ is limiting, and negative feedback when ·OH is limiting. The estimated historical methane-climate feedback sensitivity (≈0.08 W m⁻² °C⁻¹) is significantly higher than the IPCC AR6 estimate, largely due to the inclusion of positive contributions from negative feedbacks. The analysis successfully captures the observed interannual variations in CH₄ and δ¹³CH₄, revealing interannual oscillations in dominant feedback mechanisms. The study identifies a contemporaneous pattern between SST/LSAT contributions with positive correlations and Pr contributions with negative correlations, and vice versa, suggesting interferences between oceanic and terrestrial contributions. The tropics show a clear interannual alternation of warming-drying and cooling-wetting trends, influenced by ENSO. At higher latitudes, the alternation is less pronounced, with positive LSAT and SST feedbacks dominating. The multidecadal analysis reveals two stages of decreasing methane growth rate, attributed to growing negative feedbacks in the late 1980s-1990s and after the late 1990s. However, post-2012, positive feedbacks strengthen, potentially due to weakened sinks and strengthened biogenic sources. The strong positive SST feedback observed, particularly around 30°N East Pacific in 2013-2015, is attributed to cyanobacterial bloom direct methane production. The analysis suggests a potential causality from SST to terrestrial contributions, influencing LSAT, Pr, and terrestrial 'OH concentrations. Projections suggest increased multidecadal variability of climate contributions at higher GMST, potentially amplifying decadal climate variability.

The findings challenge the IPCC AR6's assessment of methane-climate feedback sensitivity, highlighting the significance of considering both positive and negative feedback mechanisms and their dynamic interplay. The higher sensitivity estimated in this study is attributed to the inclusion of lagged positive contributions arising from negative feedbacks, which are often overlooked in simpler models. The observed interannual and multidecadal oscillations emphasize the importance of understanding the nonlinear dynamics of methane sources and sinks, influenced by climate variability and processes like ENSO. The identification of cyanobacterial blooms as a potential driver of oceanic methane emissions adds complexity to the existing feedback frameworks. The study's methodology demonstrates the value of causal analysis in understanding complex climate systems, and its findings emphasize the need for improved models that account for these intricate interactions to better project future climate change.

This study demonstrates that the IPCC AR6's estimate of net methane-climate feedback sensitivity is likely underestimated due to nonlinearly lagged responses from oscillating positive-negative feedbacks. The interannual and multidecadal variability of methane-climate feedback is likely to be amplified at higher temperatures, potentially amplifying climate variability. Future research should focus on refining models to better capture the complexities of these feedback loops, exploring mitigation strategies, and incorporating more detailed data on oceanic methane production and the influence of ocean eutrophication. The study's methodology offers a valuable approach for disentangling causal relationships in complex climate systems.

The study's primary limitation lies in the potential underestimation of negative methane-climate feedback contributions due to the influence of anthropogenic emissions. The calibration factor used in the analysis is an approximation, and the assumption of mutual exclusivity between LSAT and Pr contributions might lead to some oversimplification. The limited availability of δ¹³CH₄ data introduces uncertainty in the reconstructed trends, particularly regarding the accurate separation of biogenic and fossil fuel emissions. Certain region-specific results, especially in the southern hemisphere, require further investigation and physical explanations.