Deep Gulf of Mexico seeps are not a significant source of methane to the atmosphere

Abstract

Hydrocarbon seeps release fossil methane into the marine environment, but emission of this seep-derived methane to the atmosphere is challenging to constrain. Here, we measure the concentration and radiocarbon content of dissolved methane in seawater above seeps in the northern Gulf of Mexico that have previously been linked to considerable atmospheric emissions. In bottom waters above the seeps, methane radiocarbon content is close to zero, confirming the release of fossil seep methane. However, radiocarbon signatures approach modern values at shallower depths, indicating that only ~21% of the methane in surface waters is sourced from seeps. We observe a mid-depth methane concentration maximum and radiocarbon minimum at ~200 m below the surface, but this likely reflects lateral advection of fossil methane from shallower seep fields. Our findings are consistent with previous radiocarbon fingerprinting in coastal regions, and suggest that seeps deeper than ~400 m are not a major contributor to atmospheric methane emissions.

Introduction

Methane (CH₄) is an important greenhouse gas with an atmospheric warming potential 80 times higher than carbon dioxide (CO₂) over a 20-year time frame¹. Over the last two centuries, CH₄ emissions have nearly doubled, primarily resulting from human activities, and the atmospheric CH₄ inventory is persistently rising at an average global rate of 19 ± 8 Tg CH₄/yr¹. It is anticipated that CH₄ emissions will further increase in response to agricultural practices and increased natural gas utilisation as a fossil fuel, as well as perturbation of natural CH₄ sources by climate and land-use change².

Oceans make a considerable contribution to natural CH₄ emissions to the atmosphere³, with sources that can be classified broadly as either modern or fossil. Modern oceanic CH₄ is formed via anaerobic or aerobic methanogenesis from relatively modern carbon precursors (e.g., refs. ^4,5,6,7,8), while fossil oceanic CH₄ is released at the seafloor via hydrocarbon seeps, decomposing clathrates, or degraded subsea permafrost (e.g., refs. ^{4,5,6,9,10,11,12,13}). Fossil oceanic CH₄ sources are of particular interest because of (i) the potential for CH₄ clathrates to dissociate and subsea permafrost to degrade due to ocean warming, driving a positive climate feedback¹⁴ and (ii) the importance of constraining the natural sources of fossil CH₄ to the atmosphere in the context of increasing anthropogenic emissions of fossilised CH₄¹⁵.

In oceanic systems, CH₄ clathrate hydrate dissociation has the potential to release fossil CH₄ into the water column and atmosphere. These ice-like structures, composed of water and CH₄ molecules, store approximately 1800 Gt C (10¹⁵ g) of CH₄ globally and are sensitive to intermediate depth warming^14,16. In shallow (<200 m) arctic regions¹⁷, decomposing clathrates can contribute CH₄ directly to the atmosphere via ebullition, however, some emissions have been suggested to be episodic rather than consistent¹⁸. In addition to ebullitive emissions, diffusive sea-to-air fluxes of dissolved seep CH₄ can be substantial, as observed at Coal Oil Point in the Santa Barbara Channel^19,20.

Outside of these shallow regions, fossil CH₄ released via hydrocarbon seeps or decomposing clathrates into deeper water has been shown to drive negligible atmospheric emissions (e.g., refs. ^20,21,22). Measurements of the natural radiocarbon (¹⁴C) content of CH₄ above active seafloor seeps along US margins found fossil CH₄ dissolved in deep waters, but only modern CH₄ in surface waters when the total water column depth was >500 m⁹. These results highlight the efficiency of bubble dissolution, aerobic oxidation of CH₄, and dilution⁹ at limiting the transfer of CH₄ to the atmosphere from seeps at depths where dissociating hydrates are found in mid-latitude regions (500–600 m). However, contrasting findings suggest that deepwater (>500 m) hydrocarbon seeps can drive substantial diffusive sea-to-air CH₄ fluxes, particularly in locations where gas bubbles coexist with oil, protecting CH₄ bubbles and slowing dissolution rates (e.g., Gulf of Mexico²³).

The transport of CH₄ bubbles from the seafloor to surface waters is influenced by myriad factors, including: (i) the bubble size distribution and release depth²⁴; (ii) water temperature, salinity, and dissolved [CH₄]; and (iii) the presence of surfactant, hydrate, or oil coatings on the bubble surface²⁵. Simulations using bubble plume models²⁶ suggest that bubbles without an oil coating released at depths >230 m are unlikely to release CH₄ to the atmosphere. Nonetheless, recent work modeling the dynamics of oil-coated bubbles in deep ocean conditions²⁷ found that 5 mm diameter gas bubbles released at a water depth of approximately 470 m could transport fossil CH₄ to the surface ocean, though only <0.5% of the initial CH₄ remained in the bubble. Once dissolved in seawater, CH₄ can be diluted through mixing, laterally transported outside the seep field, and oxidized via microbial activities (e.g., refs. ^28,29,30), further minimizing atmospheric release. Therefore, due to the complex interplay of physical, chemical, and biological factors influencing the transfer of seep-derived CH₄, the contribution of fossil oceanic CH₄ to the atmospheric budget remains uncertain.

To further elucidate the dynamics of fossil CH₄ released from seeps and resolve discrepancies in the release of fossil oceanic CH₄ to the atmosphere^23,31, we analyzed the concentration and ¹⁴C content (Δ¹⁴C-CH₄) of CH₄ dissolved in seawater across the northern Gulf of Mexico. We sampled above previously identified seafloor CH₄ seeps and in locations without known fossil CH₄ influences (Fig. 1 and Supplementary Fig. 1). Our study aims to partition the contribution of fossil CH₄ in the waters of the Gulf of Mexico, specifically at two seep sites in Green Canyon block 185 (GC-185); Bush Hill (BHS) and a shallow seep (SS) at a water depths of approximately 550 m and 130 m, respectively. BHS is a well-studied seep site associated with oil emissions, resulting in some bubbles being coated with oil²³. The CH₄ flux from this seep is highly variable and modulated by hydrate formation and dissociation^32,33. Despite this variability, the seep is long-lived and supports a hydrocarbon-based ecosystem, as evidenced by authigenic carbonate precipitation^34,35,36. In contrast, the SS sites located on the continental shelf of Louisiana do not exhibit oil leakage from the seafloor. To provide a broader environmental context, we also included a site with no observed seafloor seeps as a background site, and collected several surface-only samples, including those from the Mississippi River (MR) plume and an offshore location.

Fig. 1: Sample locations for radiocarbon (¹⁴C) and concentration of CH₄ in the northern Gulf of Mexico.

Yellow polygons (with white lines) represent the seep anomalies identified by the Bureau of Ocean Energy Management (BOEM). Vertical ¹⁴C-CH₄ distributions were investigated at the BHS, SS, and Background sites, and surface-only samples were collected from the ATGOM-U9, -U10, -U12, and MR plume sites.

Results and discussion

Concentration and ¹⁴C content of dissolved CH₄ in the water column

Measured CH₄ concentrations at the BHS site were around 60 nM in bottom waters and reached their lowest levels (<1 nM) at a depth of ~250 m (Figs. 2a and 3 and Supplementary Tables 2 and 3). The concentration exhibited an intermediate peak at ~100 m (17.7 nM), then decreased to 4.3 nM at the surface. Although the magnitudes differed, the vertical structure of the [CH₄] profile we observed was consistent with findings from a previous study²³ conducted at the same site. At the SS site, water column CH₄ concentrations were considerably higher than those at BHS, except for the bottom water, where [CH₄] was 35.6 nM. Interestingly, the maximum CH₄ concentration of 79.3 nM was observed at a depth of 80 m, approximately 50 m above the seafloor. From there, the concentration gradually decreased towards the surface, where it was measured to be 7.1 nM. For the background site, the bottom water concentration started at 2.1 nM and gradually increased to a maximum of 5.6 nM at a depth of 65 m, before decreasing gradually to 2.6 nM at the surface (Fig. 2a). Atmospheric [CH₄] was in the ranges of 1.9 ppm to 2.4 ppm along the entire our study track in Gulf of Mexico (Supplementary Fig. 2a).

Fig. 2: Methane concentration and radiocarbon content at sampling sites.

Vertical profiles of: a dissolved concentrations of measured (colored symbols) and seep-derived CH₄ estimated using F_s × [CH₄] (open symbols), b Δ¹⁴C, and c Seep-sourced fraction of CH₄ (F_s). F_s is calculated using two possible ¹⁴C-CH₄ end-member values for CH₄ not sourced from seeps: 132 percent modern carbon (pMC) and 100 pMC, representing atmospheric equilibrium and CH₄ formed from contemporary organic matter, respectively. For surface F_s calculation, 132 pMC was used for CH₄ not sourced from seeps due to air-sea equilibrium at this station. Surface ¹⁴C contents and concentrations of CH₄ are also shown in Fig. 3.

Fig. 3: Surface distributions of Δ¹⁴C and concentration of CH₄.

Colored diamonds represent ∆¹⁴C-CH₄ (pMC). White dots represent sample sites for the CH₄ concentration measurements.

The Δ¹⁴C-CH₄ values observed at our sampling sites spanned a range from 1.5 to 131.9 percent modern carbon (pMC) (Figs. 2b and 3). In our deepest measurement at BHS (470 m; 80 m above the seafloor) Δ¹⁴C-CH₄ was 1.5 pMC, but ¹⁴C content increased to 36.1 pMC at a depth of 200 m, where the CH₄ concentration decreased considerably to 1.7 nM. The Δ¹⁴C-CH₄ value then decreased to 14.8 pMC at 100 m, corresponding to the subsurface concentration maximum (Fig. 2b), before increasing rapidly to 103.9 pMC in the surface water. At the SS site, the Δ¹⁴C-CH₄ value was 3.4 pMC in bottom waters, and increased to 10.4 pMC and 82.6 pMC at 30 m and the surface, respectively. At the background site where no seafloor seeps have previously been reported, the Δ¹⁴C-CH₄ values in the bottom and surface waters were 103.1 pMC and 131.9 pMC, respectively. It should be noted that the deepest samples for Δ¹⁴C-CH₄ measurement were collected approximately 80 m above the seafloor for BHS and 10–20 m for the SS and background sites, to avoid contact between our sampling system and seafloor (see “Methods” section). Therefore, it is likely that the bottom water close to the seafloor may have Δ¹⁴C-CH₄ values even closer to 0 pMC in the two seep sites (Fig. 2b).

Methane sources in the water column

The Δ¹⁴C-CH₄ value of 103.1 pMC in bottom waters at the background site (Fig. 2), is similar to that expected for CH₄ produced from contemporary organic matter (~100 pMC)^5,37. This shallow background site (approximately 170 m of water depth), therefore, likely contains CH₄ produced recently from anaerobic respiration of organic matter in the sediment. Recent studies have shown that the ¹⁴C content of sedimentary organic matter in the northern GOM ranges from approximately 80–90 pMC, with higher (younger) values closer to the land^38,39,40. CH₄ produced biologically in sediments at the background site would therefore likely contain ~90 pMC, since CH₄ produced from recently fixed organic matter usually matches the ¹⁴C content of the organic precursor^4,5,6,7. The slightly elevated Δ¹⁴C-CH₄ value (103.1 pMC) we observed in the bottom water of the background site could result from mixing with surface waters, where Δ¹⁴C-CH₄ is at atmospheric equilibrium (131.9 pMC). Therefore, measured Δ¹⁴C-CH₄ at the background site is most consistent with a modern organic matter source of CH₄, with negligible influence of seep-sourced fossil CH₄.

The Δ¹⁴C-CH₄ measured in near-bottom waters at both seep sites was found to be strongly depleted (<4.0 pMC) compared to the contemporary value observed in the ocean surface and the background site (Figs. 2 and 3). These depleted Δ¹⁴C-CH₄ values cannot be explained by the ¹⁴C content typically found in GOM organic matter (~90 pMC), ruling out a more modern biological sedimentary source for this CH₄. It is widely reported that CH₄ derived from gas-hydrate dissociation and natural hydrocarbon seeps is devoid of ¹⁴C⁹. Our Δ¹⁴C-CH₄ measurements at the BHS and SS sites therefore confirm that the elevated CH₄ in bottom waters was indeed derived from natural seeps fed by gas-hydrate dissociation and other sub-seafloor hydrocarbon reservoirs. We note that CH₄ leakage from the oil and gas infrastructure near this study site in the northern Gulf of Mexico may also represent an additional fossil CH₄ source to the water column (Supplementary Fig. 1). At both seep sites, we observed a general decline in CH₄ concentration and a transition towards modern Δ¹⁴C-CH₄ values moving from the seafloor toward surface (Fig. 2). Previous studies in the northern GOM have reported similar vertical concentration profiles^23,41,42 and suggested that the removal was attributed to oxidation^42,43. However, if oxidation were the only process influencing the dissolved CH₄ concentration, the Δ¹⁴C-CH₄ profile would remain relatively close to the bottom water endmember values (e.g., 1.5 pMC at BHS) since the unit of pMC removes isotopic fractionation effects (e.g., microbial oxidation) through ¹³C normalization⁴⁴. The increase in Δ¹⁴C-CH₄ with altitude above the seafloor indicates the introduction of contemporary CH₄, suggesting that dilution of seep-plume waters with background waters (with low [CH₄] and modern Δ¹⁴C-CH₄) also influences the CH₄ distribution in these seep sites.

To calculate the fraction of CH₄ dissolved in the water column that originates from seeps (F_S), a radiocarbon isotopic mass balance was conducted (Eq. 1).

$${}^{14}C_{m}={}^{14}C_{s}\cdot {F}_{{{{\rm{s}}}}}+{}^{14}C_{b}\cdot (1-{F}_{{{{\rm{s}}}}})$$

(1)

Here, ¹⁴C_m is the measured value of Δ¹⁴C-CH₄ at a specific location and depth, ¹⁴C_s is the end-member Δ¹⁴C-CH₄ value for seep-sourced CH₄ (0 pMC), and ¹⁴C_b is the Δ¹⁴C-CH₄ value for a “background” end-member. For samples in surface water, where CH₄ exchanges rapidly with the atmosphere, we used the modern atmospheric end-member of 131.9 pMC measured at the background station for ¹⁴C_b. In deeper waters, we set ¹⁴C_b to either the atmospheric value or the modern organic matter value of 100 pMC to derive two different estimates of F_s. This balance reveals that CH₄ in bottom waters at both seep sites is almost entirely (>99%) seep-derived, and the seep contribution gradually decreases with altitude above the seafloor (Fig. 2c and Supplementary Table 3), reaching 21% and 37% at the surface of BHS and SS, respectively. At BHS, this vertical trend is interrupted by a subsurface maximum of ~85% seep-sourced CH₄ at ~100 m (Fig. 2c), which we discuss in the following sections. We note that F_s was estimated under the assumption of a simple two-endmember mixing between atmospheric and seep-derived components. Therefore, this represents the maximum potential seep contribution, since CH₄ in the surface layer can also be produced from modern dissolved organic carbon (DOC) through aerobic methanogenesis (e.g., ref. ⁸). In the Gulf of Mexico, the radiocarbon (¹⁴C) content of surface DOC is approximately 90 pMC³⁸, suggesting that the true seep contribution to surface waters could be considerably lower than our estimated F_s if there is also a CH₄ endmember derived from this DOM.

The vertical structure of F_s we calculate at the BHS site indicates that seep-sourced CH₄ is largely confined at depth and makes a limited contribution to surface waters, even in this highly active field of >500 m deep seeps. A similar observation was reported in the US-Atlantic margins, where seep CH₄ does not reach the surface or atmosphere in water columns deeper than 500 m⁹. In addition to its limited vertical propagation, our results also indicate that seep-sourced CH₄ is localized in its lateral extent. Our background station showed no contribution of seep-sourced CH₄ in bottom waters, even though it is situated close to the margin (within ~100 km) of the dense seep fields (Fig. 1), and is downstream with respect to the large-scale GOM loop current⁴⁵. Surface current velocities were approximately 0.1 ms⁻¹ during this study (Fig. 4), thus, it would take approximately 10 days for water masses to travel from the seep field to the background site. Assuming a CH₄ oxidation rate constant of 0.2 per day, as measured previously in this region⁴⁶, fossil seep CH₄ inputs would likely not be observed at the surface of the background site, which agrees with the results of this study. This suggests seep CH₄ is not transported far from the confines of the seepage region, similar to a previous conclusion in the Santa Barbara Basin²⁸. Furthermore, the fact that the seep-sourced Δ¹⁴C-CH₄ signal at the BHS site is replaced by a contemporary background signal within a few hundred meters of the seafloor suggests that even within the dense seep field, much of the [CH₄] profile is not seep-sourced. It therefore seems that seep CH₄ is confined only to the waters surrounding individual seep mounds/anomalies (km scale), as opposed to the more regional-scale dispersal and oxidation which occurred during a large anthropogenic CH₄ and oil hydrocarbon release⁴⁷ and in enclosed basins⁴.

Fig. 4: Surface currents in the northern Gulf of Mexico.

Daily-mean geostrophic surface currents during the sample collection on: a 7/7, b 7/8, c 7/9, and d 7/10/2022. Data for the surface currents were obtained from NOAA (https://cwcgom.aoml.noaa.gov/erddap/griddap/miamicurrents). Colored dots represent ¹⁴C-CH₄ dissolved in surface waters in units of pMC.

Surface water radiocarbon in methane

The Δ¹⁴C-CH₄ signatures measured in surface waters at our background site and in the center of the Gulf of Mexico (ATGOM-9) were 131.9 pMC and 131.8, respectively (Fig. 3), which is comparable to the current atmospheric value of 130–133 pMC⁴⁸. In pristine marine environments that are unaffected by river inflow and local anthropogenic activities like nuclear power (releasing gaseous ¹⁴C as CH₄), surface water Δ¹⁴C-CH₄ tends to reach equilibrium with atmospheric values^5,13. This confirms that at our background site (and throughout the central GOM) surface water CH₄ is controlled largely by atmospheric equilibrium and shows no influence of natural seepage or local anthropogenic activities.

In contrast to the background site, the surface water Δ¹⁴C-CH₄ values at both seep sites were depleted relative to the atmosphere and the background site, indicating that the surface CH₄ at these locations is influenced by additional sources, although more so at the SS site. The SS site, located approximately 120 km from the shore, is likely influenced by freshwater sources such as the Atchafalaya and Mississippi Rivers (AR and MR, respectively), as evidenced by the surface water salinity of 29.6. However, the Δ¹⁴C-CH₄ value of MR plume waters (collected ~16 km from one of the main MR outflows at a salinity of 15.0) was 90.3 pMC, and only ~31% of the surface water at the SS site was sourced from the MR plume according to salinity balance. The Δ¹⁴C-CH₄ value of 82.6 pMC we measured in surface water at the SS site, therefore, cannot be explained as a mixture between the MR plume water and surface water equilibrated with the atmosphere, which would yield ~121.0 pMC.

Furthermore, given the ¹⁴C content of DOC measured in the GOM (~90 pMC, 37), aerobic methanogenesis using a DOC precursor cannot explain the Δ¹⁴C-CH₄ observed at the SS site. Instead, the low Δ¹⁴C-CH₄ in surface waters likely results from a considerable input of CH₄ originating from the seafloor seep (F_s = 37%, Supplementary Table 3). Joung et al.^9,28 observed depleted Δ¹⁴C-CH₄ values in surface waters off the US Atlantic and Pacific margins and concluded that natural seeps on the seafloor can deliver CH₄ to the ocean surface at sites with depths shallower than 350 m. Our measurements at the SS site (130 m water depth), therefore, align with previous studies, confirming that SSs can serve as sources of fossil CH₄ to the atmosphere.

At the deep seep (BHS) site, the surface Δ¹⁴C-CH₄ value of 103.9 pMC was found to be slightly depleted compared to atmospheric levels, and slightly higher than contemporary DOC³⁸. The CH₄ concentration of 4.2 nM measured in these surface waters was approximately double the air-sea equilibrium concentration of 2.1 nM at the conditions of 2.4 ppm atmospheric pCH₄ (the highest atmospheric concentration observed near BHS), salinity of 31, and temperature of 31 °C (Supplementary Figs. 2a and 3), and cannot be explained by bubble-induced supersaturation which is at most 5% under the wind speed conditions we observed <10 km hr⁻¹ (Supplementary Fig. 2b)⁴⁹. This indicates additional sources of CH₄ to the mixed layer, beyond atmospheric equilibration. In oligotrophic oceans with limited phosphate availability, CH₄ can be produced through bacterial breakdown of methylphosphonate⁸. However, even attributing all of the CH₄ supersaturation in BHS surface waters (2.3 nM, 55% of total CH₄) to biological production would result in a Δ¹⁴C-CH₄ value of 108.9 pMC, based on mixing calculations using ¹⁴C values of DOC measured in the GOM (~90 pMC, 38). This estimate demonstrates that aerobic CH₄ production alone cannot account for the observed Δ¹⁴C-CH₄ of 103.9 pMC at the BHS site. Furthermore, the surface salinity at BHS of 34.0 indicates minimal influence from the MR plume. Therefore, it is likely that surface CH₄ for the BHS site includes a contribution from seep sources (F_s = 21% according to our radiocarbon mass balance, Supplementary Table 3). At a neighboring site within GC-185 with a bottom depth of 910 m (ATGOM-U10, ~60 km from BHS), we measured Δ¹⁴C-CH₄ of 112.0 pMC in surface waters, corresponding to a seep CH₄ contribution of 15% (Figs. 2 and 4, and Supplementary Table 3). These findings support previous work that attributed elevated [CH₄] in the surface of deep water columns (bottom depth > 500 m) at GC-185 to seafloor seeps^23,31. However, the concentrations we observed were relatively low, and we estimate that seeps contribute a minority of surface CH₄. Additionally, it is unclear whether the seep-derived CH₄ in the surface originates from the deep seafloor seeps themselves (e.g., BHS) or is transported laterally from other seep-impacted sites.

Lateral advection of CH₄ from SS sites

The presence of depleted Δ¹⁴C-CH₄ values in the surface water of the BHS site in GC-185 suggests the existence of fossil (or seep-sourced) CH₄ in the mixed layer. There are two potential mechanisms for the delivery of fossil CH₄ to surface waters at the BHS: vertical migration from the known seafloor seep at this site, or lateral advection of water impacted by shallower seeps. It seems unlikely that vertical diffusion or upwelling are the only transport mechanisms of fossil CH₄ between the seafloor and the sea surface, since the concentration of dissolved CH₄ reaches a minimum at intermediate depths, both in our observations (Fig. 2a) and previous studies (e.g., refs. ^23,41,43,47).

Bubble transport from the seafloor to the surface has been hypothesized as a vector for seep-derived CH₄ to the surface at this site. Solomon et al.²³ tracked bubble plumes from the seafloor to the surface using a submersible and analyzed CH₄ concentrations in samples collected during ascent. They reported surface CH₄ concentrations as high as 1609 nM (608 nM in surface water of BHS). However, the surface CH₄ concentration we measured (4.3 nM) was considerably lower than those reported by Solomon et al.²³, but similar to those reported by Hu et al.³¹ (ranging from 1.72 to 4.48 nM). To date, no studies in Gulf of Mexico surface water above deepwater seeps have observed similar high concentrations of dissolved CH₄ as were reported in Solomon et al.²³.

Solomon et al.²³ also observed a mid-depth [CH₄] minimum, and suggested that the extremely high surface concentrations they observed could result from bubbles larger than 4.5 mm in size that dissolved in the surface layer, based on a simplified bubble propagation model. In that model, large oil-coated bubbles fully dissolved at the bottom of the mixed layer or shallower, where they encountered sharp density gradients between plume fluids and ambient seawater, causing CH₄ to accumulate to 30–954 times atmospheric equilibrium. However, a recent modeling study by Gros et al.²⁷ demonstrated that while large oil-coated bubbles can reach the ocean surface, most of the CH₄ they carry will dissolve into the seawater during their ascent. Using a dynamic bubble model (Texas A&M Oil Spill/Outfall Calculator—TAMOC), they simulated 5 mm diameter CH₄ bubbles with oil coatings and found no meaningful fraction (<0.1%) of the original CH₄ was delivered to surface waters when the release depth was greater than 200 m. This suggests that bubbles formed at the BHS site, with a water depth of ~550 m, would not carry a substantial CH₄ content by the time they reach the surface. Moreover, recent models (e.g., refs. ^26,27) suggest that rising bubbles rapidly exchange gases with surrounding waters during their whole ascent, rather than dissolving specifically in surface waters. Consequently, the BHS CH₄ concentration profiles observed here and in previous work^23,31, which decrease rapidly with altitude before increasing again near the surface, likely cannot be explained by the dynamics of vertical bubble transport alone. Instead, the CH₄ depletion at mid depths suggests that CH₄ released from bubbles is rapidly removed by oxidation and dilution⁴³, and confined within a few hundred meters of the seafloor, while there is an additional influence from lateral transport in the upper water column.

Our measurements reveal a subsurface CH₄ concentration peak centered at ~100 m at the BHS site, where [CH₄] is ~4-fold higher than at the surface. This further undercuts the hypothesis that elevated surface [CH₄] is driven by bubbles “breaking” at the base of the mixed layer²³, since the mixed layer was 20–30 m during our sampling (Supplementary Fig. 1), which is typical for summertime at this site⁴⁴. Subsurface [CH₄] peaks have been observed in previous studies in the Gulf of Mexico^23,50,51, and are often attributed to CH₄ production by microbial processes^41,50. However, at the BHS site, the [CH₄] peak is accompanied by a subsurface minimum in Δ¹⁴C-CH₄ of 14.8 pMC (Fig. 2b), which is far too low to be explained by microbial production from a modern organic matter precursor. As for the source, our observations suggest that the elevated subsurface [CH₄] originates from a seep source (F_s reaches a peak of ~85% at 100 m, Supplementary Table 3), albeit not the deep seep at this specific location.

One previous study attributed a mid-depth CH₄ peak observed in the northern Gulf of Mexico to lateral advection from the edge of the continental shelf across the shelf break⁵¹. We therefore explored whether advection of seep-derived CH₄ from SSs on the continental shelf could explain the elevated [CH₄] with depleted Δ¹⁴C-CH₄ (relative to atmosphere) that we observed in the top ~100 m at the BHS site. During our sampling period, surface currents in the region surrounding BHS were directed towards the south or southwest (Fig. 4), creating favorable conditions for advection of shelf water across the shelf break. In fact, for most of the sampling period, the BHS site was directly downstream from our SS site (36 km northeast of BHS), allowing our SS samples to be interpreted as an upstream endmember.

According to a simple mixing and equilibration calculation (e.g., ref. ²⁰), the surface CH₄ concentration observed at BHS could be explained as a mixture of ~44% surface water from the SS site and ~56% atmosphere-equilibrated water, which could either represent equilibration of the SS precursor during its advection, or dilution with equilibrated background water. In turn, this would predict a surface Δ¹⁴C-CH₄ value of 95.5 pMC at BHS, which is similar to the observed value of 103.9 pMC (although we acknowledge this calculation does not account for losses due to oxidation or additional inputs like microbial production). Therefore, it is highly plausible that lateral advection of surface water masses carrying CH₄ from SSs, or from any leaks in the oil/gas infrastructure, was responsible for the presence of fossil CH₄ in surface water at the BHS site (Fig. 5). The subsurface (~100 m) CH₄ concentration peak and Δ¹⁴C-CH₄ minimum at BHS can then be interpreted as advected near-bottom waters from the SS site or other similar seeps on the continental shelf, which contain almost exclusively seep-sourced CH₄ (Figs. 2c and 5). The depth of this peak would therefore depend on the depth of shelf waters being advected offshore, and the penetration depth of cross-shelf currents, which may vary both in space and time.

Fig. 5: Conceptual model of fossil CH₄ cycling in the northern Gulf of Mexico.

A schematic of our main conclusions, in which ¹⁴C-depleted CH₄ released from deep seeps is removed due to oxidation and mixing before penetrating the surface, whereas CH₄ released by shallow seeps is advected offshore and emitted to the atmosphere.

Reconciling methane dynamics in GOM surface waters

Previous studies have reported contrasting CH₄ dynamics in surface waters of the Gulf of Mexico GC-185 region. Solomon et al.²³ documented surface CH₄ concentrations ranging from 650-1605 nM, representing considerable supersaturation (30-954 times) above atmospheric equilibrium, whereas Hu et al.³¹ observed only modest supersaturation (4.48 nM, ~2 times equilibrium) at the same location. Hu et al.³¹ deliberated on the elevated CH₄ levels observed by Solomon et al., attributing them to localized hotspots of CH₄ emanating from seafloor seeps, which were not pinpointed during resampling. They proposed that the elevated CH₄ could originate either from seeps within the sampled sites, or be advected from shallower seep locations to the sampling site. Our new ¹⁴C-CH₄ measurements confirm that seep-sourced CH₄ is present in the surface and mid-depths at BHS, but suggest that it only makes a minor contribution (<21% or <1 nM, Fig. 2 and Supplementary Table 3) to the dissolved mixed-layer CH₄ inventory. In addition, our Δ¹⁴C-CH₄ and CH₄ concentration measurements suggest that much of the near-surface seep-derived CH₄ was laterally transported to the BHS site, rather than transmitted vertically from the seep at ~550 m depth as posited by Solomon et al.²³. During the sampling period for that prior study, (August 14–23, 2003), surface currents were also directed southwards towards the center of the Gulf of Mexico, though considerable variations were observed over the span of hours and days (Supplementary Fig. 4). This flow pattern would be conducive to water mass transport from shallower seep regions to the deeper sites, as postulated for our sampling period, and it may plausibly explain the high CH₄ concentrations reported by Solomon et al. Even if the elevated CH₄ was transmitted vertically from the 550-deep seep, the considerably lower CH₄ concentrations observed in follow-up studies at the same site suggest that CH₄ hotspots in surface waters at BHS are extremely limited in space and time, similar to other seep fields²⁸. Alternatively, seepage at BHS could be ephemeral in bubble size distribution, bubble coatings, or seafloor gas release rates, allowing for short episodes that are favorable for bubble propagation through the water column. However, previous studies have shown that under normal conditions, very little CH₄ should reach surface waters from 550 m depth in this environment⁵², which would further diminish the annual atmospheric emissions of oceanic seep-derived CH₄.

Overall, our results indicate that fossil seep-derived CH₄ makes a limited contribution to surface waters in the GC-185 sector, and suggest that the impact of deep seep-derived CH₄ on atmospheric emissions is localized and minor despite the presence of numerous seafloor seeps in the northern Gulf of Mexico. Furthermore, the advection of water masses from SS sites, rather than vertical transmission from deep seeps, appears to be an important factor controlling the water column distribution of seep-derived CH₄ in the northern Gulf of Mexico (Fig. 5). This supports and builds upon previous findings from the US Atlantic and Pacific margins⁹, which showed that fossil CH₄ is not transmitted to the surface from seeps >500 m deep. Here, we showed that this applies even in regions of dense and active seepage, where oil may provide a protective bubble coating. This strengthens the building consensus that dissociating hydrates in mid-latitudes, which occur at depths >500 m, are unlikely to drive substantial emissions of CH₄ to the atmosphere, especially as compared to direct anthropogenic emissions of fossil CH₄ to the atmosphere⁵³.

Methods and materials

Site description

The northern Gulf of Mexico harbors a vast number of seep anomalies, with over 36,000 identified (BOEM, https://www.boem.gov/oil-gas-energy/mapping-and-data/map-gallery/seismic-water-bottom-anomalies-map-gallery) (Fig. 1). Among these, the BHS site located in Green Canyon 185 (GC-185) is well-known and situated at a depth of approximately 550 m. The GC-185 gas seeps originate from gas reservoirs present in deep sediments, as well as gas-hydrate dissociation^{23,34,54,55,56,57,58}. Geographically, the major fault system underlying the GC-185 block connects to the Jolliet hydrocarbon and oil reservoirs, intersecting the seafloor at the active site of hydrocarbon discharge (refer to Fig. 2 in 34). Consequently, the gas bubbles in the GC-185 seep fields are occasionally coated with oil, and ongoing gas venting results in visible bubble plumes in the water column as recorded by echo sounders. Additionally, the presence of oil sheen on the water surface can be observed by satellites^23,55,58. While sonar was not used to confirm active bubble release from BHS during our sampling, we did observe oil-coated bubbles surfacing at the site. Methane concentrations in this area are strongly elevated, reaching up to 13,660 nM at the bottom and up to 1600 nM in the surface layer²³.

A SS site was selected for this study based on previous observations conducted by the BOEM (Fig. 1). This site is located in the continental shelf with a water depth of approximately 130 m and is situated around 130 km from the Louisiana coast. While there is no specific information regarding the geographic features of this site, it is likely to share similarities with the GC block, as it is only about 35 km northeast of GC-185. However, according to the BOEM’s seismic seep field anomaly investigations, this location is not categorized as an oil-related anomaly (neither confirmed-oil nor possible-oil in both positive and negative anomalies). Additionally, due to the shallow water depth (~130 m), it is not expected to harbor gas hydrates on the seafloor. Consequently, the bubbles observed in this system would be classified as “not oil coated,” indicating their origin from natural hydrocarbon seeps.

A background site situated at a water depth of approximately 170 m was also chosen for comparison. This site is located at the head of De Soto Canyon, approximately 60 km off Pensacola Beach, Florida, USA. No seep activity was observed at this site, as reported by the BOEM (Fig. 1).

Sample collection

For the ¹⁴C-CH₄ analysis, water samples were collected from a total of six sites, including three sites for vertical profile investigations and three surface-only sites. The vertical investigation sites consisted of two seep sites, namely BHS in GC-185 and an SS, along with one background site. The purpose of these vertical profile sites was to examine the CH₄ dynamics in the water column. In addition, surface-only samples were collected from three different locations. The first surface-only sample was taken from the MR plume, which represents a distinct water mass influenced by riverine inputs. The second surface-only sample was obtained from a seep field near GC-185, specifically referred to as ATGOM-U10. Lastly, a sample was collected from a far-offshore location with a water column depth of approximately 3000 meters, referred to as ATGOM-U9. These sampling sites were chosen to provide a comprehensive understanding of the CH₄ distribution and dynamics, encompassing both vertical profiles and surface conditions in various locations within the study area.

The details of the ¹⁴C-CH₄ sample collection and measurements can be found in the works of Sparrow and Kessler^6,59, as well as Joung et al.^9,13,28,60. The equipment specifications used in this study are provided in Sparrow and Kessler⁵⁹. The process of water sample collection and preparation involved two main steps: field gas sample collection and laboratory gas sample purification. In the field, CH₄ was extracted from the water and compressed, then stored in a small gas cylinder. Briefly, a discharge pump was used to draw water from the desired depths and introduce it to the gas extraction system on the deck. The pumped water underwent filtration through sequentially decreasing pore sizes of 100 µm, 50 µm, and 10 (or 5) µm to remove particles that could potentially damage the membrane filters. The filtered water was then passed through the membrane filters, which consisted of a water and a vacuum side. Vacuum pressure was applied to the vacuum side of the membrane while the water flowed through the water side, facilitating the extraction of dissolved gases from the water. The extracted gases were compressed and stored in a gas cylinder, which was transported back to the laboratory for further processing of the samples.

In the laboratory, the gas samples in the cylinder were analyzed for their concentrations, including other carbon-containing gases such as CO₂ and CO, to assess the recovery of carbon from CH₄ during the final stage of analysis. The gas cylinder was connected to a vacuum line to purify the gas samples. The extracted gas sample was then passed through a molecular sieve and the first liquid nitrogen (LN) trap to remove CO₂. The remaining gases, including CO and CH₄, were directed through the first oven set at 600 °C to convert CO to CO₂. The resulting CO₂ was subsequently removed by the second LN trap. Finally, the gases were introduced to the second oven, which was set at 1000 °C to convert CH₄ to CO₂. The CO₂, now trapped in the third LN trap, was transferred to a pre-combusted glass tube and flame-sealed. This CO₂-containing glass tube was sent to the Keck Carbon Cycle Accelerator Mass Spectrometer (AMS) facility at UC Irvine for determination of its ¹⁴C values.

To assess the performance of the sample purification system, two standards containing different carbon species (CO₂, CO, and CH₄) and concentrations were processed alongside every two to three days, following the same procedures as the samples. Analysis of the standards revealed ¹⁴C-CH₄ values of approximately 0 pMC (Supplementary Table 1). Additionally, the standards were monitored continuously to ensure the complete removal of CO₂ and CO, as well as the full conversion of CH₄ to CO₂ and subsequent trapping.

To evaluate background carbon in the purification system, ultra-zero-grade air was passed through the system for 1–2 h. This test was conducted daily, just prior to running the samples. The background carbon concentration was found to be less than 0.005 µg/L, which amounted to less than 0.4% of the sample carbon collection (250 µg-C when processing 180 L of sample). Since the gas sample volume used for analysis was typically less than 100 L, the percentile contribution of the background carbon was even lower than 0.3%, similar to the instrumental background level of 0.1% (or 2–3 ‰,⁶¹). Nevertheless, this background carbon is not expected to have any impact on the ¹⁴C-CH₄ values obtained from the samples.

Our measurement technique for ¹⁴C in CH₄ requires approximately 250 µg of carbon in total. The water volume for CH₄ extraction can vary depending on the CH₄ concentrations in the water samples. For instance, surface waters with a concentration of 2 nM would typically require over 35,000 L of water, taking into account the gas extraction efficiency of the membrane filters and system cleaning⁹. Furthermore, discrete bottle samples were also collected at the same locations as the ¹⁴C-CH₄ sample collection sites to determine the concentrations of CH₄. The CH₄ concentrations were analyzed following the method described by Weinstein et al.⁶².