
==== Front
Proc Natl Acad Sci U S A
Proc Natl Acad Sci U S A
PNAS
Proceedings of the National Academy of Sciences of the United States of America
0027-8424
1091-6490
National Academy of Sciences

38466851
202316878
10.1073/pnas.2316878121
datasetDatasetvideoVideoresearch-articleResearch Articleearth-sciEarth, Atmospheric, and Planetary Sciences413
Physical Sciences
Earth, Atmospheric, and Planetary Sciences
Cold seep formation from salt diapir–controlled deep biosphere oases
Chowdhury Anirban a
Ventura Gregory T. todd.ventura@smu.ca
a 1
Owino Yaisa a
Lalk Ellen J. b
MacAdam Natasha c
Dooma John M. a
Ono Shuhei b https://orcid.org/0000-0002-1348-9584

Fowler Martin d
MacDonald Adam c
Bennett Robbie e
MacRae R. Andrew a https://orcid.org/0009-0004-4636-814X

Hubert Casey R. J. f https://orcid.org/0000-0002-8691-8116

Bentley Jeremy N. a
Kerr Mitchell J. a https://orcid.org/0009-0003-0042-7178

aDepartment of Geology, Saint Mary’s University, Halifax, NS B3H 3C3, Canada
bDepartment of Earth, Atmospheric, and Planetary Sciences, Massachusetts Institute of Technology, Cambridge, MA 02139
cNova Scotia Department of Natural Resources and Renewables, Halifax, NS B3J 3J9, Canada
dApplied Petroleum Technology (Canada) Ltd., Calgary, AB T3A 2M3, Canada
eNatural Resources Canada, Geological Survey of Canada-Atlantic, Dartmouth, NS B2Y 4A2, Canada
fGeomicrobiology Group, Department of Biological Sciences, University of Calgary, Calgary, AB T2N 1N4, Canada
1To whom correspondence may be addressed. Email: todd.ventura@smu.ca.
Edited by Michael Manga, University of California, Berkeley, München, Germany; received September 29, 2023; accepted January 24, 2024

11 3 2024
19 3 2024
11 9 2024
121 12 e231687812129 9 2023
24 1 2024
Copyright © 2024 the Author(s). Published by PNAS.
2024
https://creativecommons.org/licenses/by-nc-nd/4.0/ This article is distributed under Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND).

Significance

We provide compelling evidence that highly biologically diverse deep marine cold seeps can arise, and be further maintained, by methane supplied from methanogenic archaeal deep biosphere oases. These oases are nested directly above salt diapirs more than 1 km below the ocean seafloor and mark recurring features in the diapiric province of Canada’s Scotian Margin. They likely arise from nutrient and dissolved inorganic carbon loading by the convective flow of formation water that is driven by the high heat exchange of the underlying salt diapir. These deep biosphere—diapir—cold seep interlinked ecosystems may occur elsewhere around the globe as the geophysical conditions leading to their formation are also found in other passive margin salt provinces.

Deep sea cold seeps are sites where hydrogen sulfide, methane, and other hydrocarbon-rich fluids vent from the ocean floor. They are an important component of Earth’s carbon cycle in which subsurface hydrocarbons form the energy source for highly diverse benthic micro- and macro-fauna in what is otherwise vast and spartan sea scape. Passive continental margin cold seeps are typically attributed to the migration of hydrocarbons generated from deeply buried source rocks. Many of these seeps occur over salt tectonic provinces, where the movement of salt generates complex fault systems that can enable fluid migration or create seals and traps associated with reservoir formation. The elevated advective heat transport of the salt also produces a chimney effect directly over these structures. Here, we provide geophysical and geochemical evidence that the salt chimney effect in conjunction with diapiric faulting drives a subsurface groundwater circulation system that brings dissolved inorganic carbon, nutrient-rich deep basinal fluids, and potentially overlying seawater onto the crests of deeply buried salt diapirs. The mobilized fluids fuel methanogenic archaea locally enhancing the deep biosphere. The resulting elevated biogenic methane production, alongside the upward heat-driven fluid transport, represents a previously unrecognized mechanism of cold seep formation and regulation.

cold seep
deep biosphere
diapir
methanogenesis
methane
Canadian Government | Natural Sciences and Engineering Research Council of Canada (NSERC) 501100000038 RGPIN-2018-06147 Gregory Todd Ventura Research Nova Scotia (RNS) 501100020170 2142 Gregory Todd Ventura
==== Body
pmcCold seeps are among the most biologically diverse ecosystems of deep ocean environments (1, 2). Their rich biota is sustained by a chemosynthetic fauna that primarily derives energy from the oxidation of venting hydrocarbons and reduced sulfur expelled from deep within the sedimentary basin. Typically, the most abundant of these fluids is methane, which especially in the presence of sulfate, activates the anaerobic oxidation of methane (AOM).[1] CH4+SO42-→HCO3-+HS-+H2O.

AOM consumes a significant proportion of the carbon sourced within a basin making this metabolic process the main hydrocarbon sink in marine sediments (3, 4). It is therefore not surprising that cold seeps are globally common features of petroleum-producing basins (5) where they act as guides for offshore oil and gas exploration (6).

Over 29 petroleum-producing basins, including major regions such as the Gulf of Mexico, Persian Gulf, and Campos Basin, are hosted in extensive salt tectonic provinces (7). The less dense and ductile behavior of salt results in the deformation of localized rock strata upon sediment loading. This leads to conditions favorable for thermogenic hydrocarbon migration and the formation of stratigraphic and structural salt-tectonic-related traps, which ultimately results in the petroleum reservoir creation (7) that makes up an estimated 50% of the world’s known total petroleum reserve (8). Salt tectonism further deforms the surrounding strata producing anticlinal bedforms and abundant faults along the flanks and above the crests of the salt structures. These more shallow basin features channelize leaking fluids (9) to feed the cold seep ecosystems on the ocean floor (10).

Not included in the petrogenic cold seep model is the impact of biologically produced methane from the deep biosphere. Methanogenic activity is globally ubiquitous in marine sedimentary systems resulting in the formation of gas hydrates that contain an estimated ∼500 to 2,500 Gt of methane carbon (11). Its production is closely associated with the basal region of the hydrate stability zone (12), which is easily mapped across the ocean basins by the identification of the bottom simulating reflection anomaly in seismic surveys. Gas hydrate instability can result in periodic releases of methane where it often produces pockmarks on the ocean floor. These tend to be ephemeral events and mainly impact the local shallow subsurface microbiome (13, 14), largely precluding the establishment of diverse ocean floor cold seep ecosystems.

The Scotian Slope is home to a large salt tectonic system that is subdivided into diapir- and canopy-dominated provinces (15, 16). Here, cold seeps capable of hosting a rich diversity of benthic life (17) are associated with salt diapirs along the continental slope of the Shelburne Subbasin, offshore Nova Scotia, Canada (18). Recent studies indicate there is a diverse microbial community that inhabits the near-surface and deeply buried sediments within and around the Scotian Slope cold seeps (19, 20). Near-surface hydrocarbon seep sediments have been shown to host microbial communities that include members capable of oxidizing C2+ alkanes (19). Furthermore, certain thermophilic bacterial endospores are more prevalent in the vicinity of seeping thermogenic hydrocarbons and are likely derived from petroleum reservoirs deep within the basin that connect to the seabed via active cold seeps (21).

In this study, we provide evidence that, contrary to expectation, biologically diverse ocean floor cold seeps can also be an extension of methanogenic archaeal deep biosphere oases that are supported by the interplay between salt-generated basin deformation and heat-driven groundwater circulation. In this regard, the manner in which passive margin cold seeps are formed and maintained may only be partially understood.

Results

Two wide-azimuth 3D seismic surveys (named Shelburne and Tangier), totaling an area of 18,400 km2 spanning 500-m to >4,000 m water depth within the Scotian Slope, were mapped to a sediment basin depth of 3,000 mbsf (meters below seafloor) (SI Appendix, Fig. S1). Here, the margin is home to a complex allochthonous salt province extending across much of the slope. Vertical diapirs dominate, but other salt structures, such as salt canopies, are present in the northeastern region of the Tangier survey. For the mapped region, 99 diapirs were identified and basic parameters were measured (SI Appendix, Fig. S2). These diapirs rise from the Late Triassic–Early Jurassic Argo Formation, initially deposited near the bottom of the basin (SI Appendix, Fig. S3). Their emplacement has disturbed the overlying and laterally adjacent rock strata, creating complex polygonal, radial, and crestal faults systems as well as subsided minibasins adjacent to the salt structures (Fig. 1 and SI Appendix, Figs. S1 and S10–S18). Diapirs located further down the continental slope (≥2,500-m water depth) reach closer to the seafloor (~0 to 400 mbsf) than those nearer to the shelf break that are thickly covered (~1,500 mbsf) (SI Appendix, Fig. S3).

Fig. 1. (A) Map of the Scotian Slope with yellow polygons indicating the locations of the Shelburne and Tangier 3D seismic surveys. (B) Combined RMS amplitude and coherence maps of the Shelburne 3D data block for three depth intervals below sea floor (red and black font for RMS amplitude and coherence, respectively) (see SI Appendix, Fig. S1 for separate maps that include the Tangier survey). Brighter map colors indicate seismic amplitude anomalies with DHIs and other types of amplitude anomalies illuminated in red. White line at the right of the map marks an ocean bottom sediment heat flow transect line from Negulic and Louden (2017) with a recorded DHI directly over a diapir (pink polygon in vertical 2D seismic crossline). Letters B and M indicate biogenic and mixed biogenic/thermogenic gas findings from ROV push core, gravity, and piston coring surveys (51, 18, 49, 17, 23, 24). Locations of seep sites 2A-1 and 2B-1 are marked with white arrows (see Fig. 3). (C) Coherence maps showing the structure of the subsurface geology at 500 mbsf showing geological structures such as radial and crestal faults, effects of contourites, and intersected salt bodies. Ellipses indicate the boundary of radial faults (yellow), the boundary of intersected salt diapir (blue), and the region covered by the overlying DHI located at ~300 mbsf (red).

Seismic attribute root mean square (RMS) amplitude maps were generated for multiple depth intervals (300 mbsf ± 100 m; 1,100 mbsf ± 400 m; and 1,250 mbsf ± 250 m) resulting in a near-continuous survey down to 1,500 mbsf (Fig. 1 and SI Appendix, Fig. S1) for both the Shelburne and Tangier 3D data blocks. Coherence maps were also generated to enhance discontinuous structural geologic features such as faults, paleochannels, salt diapirs, and intervening minibasins. Detected strong amplitude anomalies (“bright spots” in RMS amplitude maps) were interpreted to be direct hydrocarbon indicators (DHIs) if they are not likely lithologically controlled (such as basal fill of incised paleocanyons). These two seismic attribute maps were overlain and interpreted using associated vertical sections to identify potential fluid-flow migration pathways from 1,500 mbsf to the seafloor (Fig. 1). Across the geographic area, 37 DHIs were identified (SI Appendix, Table S1). All of these are located directly above salt structures (Fig. 1 and SI Appendix, Figs. S10–S18).

The location of the gas hydrate stability zone (GHSZ), marking the depth methane becomes trapped within clathrates, was identified in the seismic data by a pronounced bottom simulating reflector located at 0 to 100 mbsf. Below the GHSZ, methane gas has the potential to move more freely to shallower basin zones. All of the detected DHIs in the surveys are below the GHSZ at 200 to 500 mbsf (22), where they commonly occupy fault-bounded crestal half-grabens draped by a seal of relatively impermeable contourite muds and mass transport deposits. Most of these DHIs (n = 22) are intersected by diapir-generated faults, some of which reach unconsolidated sediments near the ocean floor, thereby providing potential migration pathways for deeply sourced fluids (Fig. 3 and SI Appendix, Figs. S10–S18). The fluids forming these amplitude anomalies appear to be vertically ascending from below the DHI, based on their position at local structural culminations; potentially following the structural conduits formed by the underlying diapirs along flanking anticlinal bedding planes of adjacent minibasins and the diapir-related crestal faults as well as amplitude dimming characteristic of gas chimneys residing below various DHIs (SI Appendix, Fig. S19). Two more shallow seepage anomalies interpreted to be gas chimneys were found directly over salt diapirs in the Shelburne survey (referred to as sites 48 and 49) but do not have an associated DHI. These reside along a portion of the slope interpreted to have an inadequate seal to form a trap (SI Appendix, Fig. S2). Surprisingly, in the Shelburne and Tangier surveys, virtually no verified DHIs, including major resolvable gas chimneys, exist away from salt bodies even when potential structural features for fluid transport and fluid trapping are present off axis of the diapir crest (Fig. 1 and SI Appendix, Figs. S10–S18 and Table S6).

Ocean Floor Cold Seeps.

In 2020, an automated underwater vehicle (AUV) cruise surveyed 10 prospective seep sites in the Shelburne and Tangier area based on the presence of DHI mapping, geophysical evidence of fluid escape structures, and geochemical data generated from three prior piston and gravity coring surveys (23, 24). The acquired multibeam benthic maps were then used to direct a 2021 remotely operated vehicle (ROV) push coring survey that targeted seven prospective seep locations (SI Appendix, Fig. S4–S6). Only two of these sites, 2A-1 (The Hole) and 2B-1 (Clamshell); marking DHI #s 34 and 32 (Fig. 1A and SI Appendix, Fig. S2) located at 2,691.9- and 2,740-m water depth, respectively, had carbonate mounds with observable gas bubbling up from the seafloor. Gas seepage rates were estimated at 1.48 and 0.01 Mg y−1 for 2A-1 and 2B-1, respectively (SI Appendix, Table S2). These seep sites were vastly more biologically diverse than the surrounding largely barren ocean floor mud-scape (SI Appendix, Text and Figs. S5 and S6).

Gas Geochemistry Speciation and Source Origin of Hydrocarbons.

The hydrocarbon gases methane (C1), ethane (C2), propane (C3), and carbon dioxide (CO2) captured at these sites were measured by three different sampling techniques: i) bubbling seep gas directly captured escaping from the ocean floor, ii) void gas pockets collected from the push core sediment by drilling through the core liner, and iii) headspace gas trapped within either crimp sealed glass serum vials or Isojars containing push core sediments. The bubbled seep gas collected at the ocean floor of both sites using a non-isobaric gas sampler is methane dominated (92.6% for 2A-1, The Hole and 90.7% for 2B-1, Clamshell, respectively) with the sum of all other hydrocarbon gases making up less than 0.1% of the bulk volume. The C1/C2+3 ratios at both sites are >1000 (SI Appendix, Table S3), indicating a biogenic origin for the methane gas (25, 26). Where known, CO2 averaged 2.2% of the gas volume. Similar gas speciation patterns were also observed for the other sampling techniques (SI Appendix, Table S3).

Carbon and Hydrogen Isotope Geochemistry.

Methane carbon (δ13C) and hydrogen (δD) isotopic composition of the non-isobaric gas collected bubbled seep gas from The Hole and Clamshell are nearly identical (δ13C narrowly ranging from −70.6 ± 0.2 to −71.0 ± 0.3‰ and δD ranging from −193.5 to −199.5‰) indicating a biogenic origin (Fig. 2A and SI Appendix, Table S4). The δ13C values closely align between seep, void, and sediment headspace gas (Fig. 2B and SI Appendix, Fig. S7). However, the void space and headspace gas isotope values of each site have larger ranges that likely indicate a significant kinetic isotope effect (KIE). Site 2A-1 where the gas discharge rate is highest has a max/avg 13C fractionation range (εseep–sediment sampled gas) of 13.4 and 6.3‰, respectively. For 2H, it is −66.8 and −58.2‰, respectively (SI Appendix, Table S5). For site 2B-1 where the gas discharge rate is low, the max/avg range of 13C fractionation (εseep–sediment sampled gas) is 13 and 5.5‰, respectively, and for 2H is −35.5 and −20.7‰. The contrast in the range of isotope values is likely due to the different seepage rates at each site along with potentially variable sources of methane production and oxidation by the shallow sediment chemosynthetic microbiome.

Fig. 2. (A) Bulk hydrogen and carbon isotope cross-plots from seep sites 2A-1 and 2B-1 for directly sampled seep gas (blue circles), push core void space (orange squares), and sediment head space gas (gray triangles, this study; brown triangles, Fowler, 2022). Two kinetic isotope effects (KIEs) are identified with FS and FP indicating fractionation from substrate and product, respectively (gray circle 90% confidence ellipse). (B) Bulk methane carbon and hydrogen isotope cross-plot with microbial carbonate reduction (CR), fermentation (F), early mature thermogenic gas (EMT), oil-associated thermogenic gas (OA), and late mature thermogenic gas (LMT) source fields (25, 26). (C) Methane clumped isotope vs. εMethane cross-plot showing 13CH3D=12CH4 chemical isotopic reversibility (Φ ranging 0 to 1), derived methane formation temperatures in relationship to isotopic equilibrium (27), radiocarbon ages (yellow boxes) and pMC (red stars) of the methane gas.

Clumped Methane Isotope Geochemistry.

Nine samples from seep sites 2A-1 (The Hole) and 2B-1 (Clamshell) were collected for methane-clumped isotope analysis (SI Appendix, Table S4). These represent two seep gas samples with replicates (n = 4), three void gas samples, and two sediment gas samples. The Δ13CH3D values range from 4.6 and 8.2‰ indicating a microbial methane origin (SI Appendix, Fig. S8). Collectively, the gas speciation, carbon, and clumped isotopic data indicate the methane emitted from these two seeps was formed by microbial carbonate reduction (i.e., hydrogenotrophic methanogenesis) (Fig. 2B):[2] CO2+4H2→CH4+2H2O.

All other ROV-visited seep sites produced elevated sediment head space gas of biogenic (n = 7) and mixed biogenic/thermogenic (n = 1) origin or no elevated gas content (n = 3).

Clumped isotopes of methane can be further used to determine the formation temperature of the hydrocarbon gas when under equilibrium conditions (28, 29). Only the bubbling seep gas collected from both sites (Fig. 2C) yielded near-equilibrium clumped methane isotope values (27). The other sampling methods (void and sediment headspace gas collection) produce a range of increasingly fractionated isotope values that progressively depart from equilibrium formation temperature estimates. These results indicate methane oxidation of gas microseepage in the shallow sediments either overprints the original methane formation temperature or represents methane entirely sourced from the environment that undergoes its own non-equilibrium fractionation kinetics (30, 31). These results also suggest the accurate determination of deep basin methane formation temperatures is most reliably achieved through direct sampling of ebullient gas.

The ebullient gas collected from seep sites 2A-1 and 2B-1 yielded Δ13CH3D values of 4.6 and 4.9‰ (Fig. 2 B and C and SI Appendix, Table S4 and Fig. S8) that record apparent geothermal temperatures of 64 +10-10 and 55 +10-13 °C, respectively. These temperatures are at or near the 60 to 80 °C upper limit of microbial methanogenesis for marine sedimentary environments (32, 33). Such higher methane formation temperatures are increasingly observed in marine environments and may require faster CO2 supply rates to sustain the archaeal community (32, 34). Mapping these temperatures to the locally derived geophysical and modeled geothermal depth changes (35) at the seep sites results in a methane formation depth of ~1,000 m for both seeps. The formation depths extend far below the underlying DHI onto the crest of the more deeply buried salt diapirs (Fig. 3). Collectively, for this region of the Scotian Slope, when the gas geochemical results of this cruise are merged with three prior piston and gravity coring surveys (2015, 2016, and 2018) on the Scotian Slope (23, 24), all but two gas producing sites yielded biogenic or a mixed biogenic/thermogenic gas (corresponding to diapirs hosting DHI #8 of the Shelburne survey area and DHI #37 of the Tangier survey; Fig. 1 and SI Appendix, Table S1 and Figs. S2 and S18).

Fig. 3. 2D seismic inline cross-sections of seep sites 2A-1 and 2B-1 (red upside-down triangles). Yellow ellipses mark DHI locations. Black lines denote faults. The colored background (right-hand panels) shows modeled geothermal temperature changes (35). Red polygons outline the possible area of methane formation based on derived methane isotopologue formation temperatures, modeled geothermal basin temperature gradients showing heat chimney effects, and structural relationships in the basin that could allow for fluid transport.

Radiocarbon Age Dating of Methane.

Lastly, nine methane-isolated gas samples were further selected for radiocarbon age dating. Of these, only four samples were pure enough to enable reportable age measurements. The Clamshell seep gas recorded an age of >44,600 y (pMC = 0.03). The void gas sample from The Hole at seep site 2A-1 had an age of 44,800 y (± 7,400 y; pMC = 0.38). Two other 2B-1 void gas samples had ages ranging from 36,100 (± 2,500 y; pMC = 1.11) and 37,100 (± 2,900 y, pMC = 0.99; Fig. 2C and SI Appendix, Table S4). The obtained radiocarbon ages of seep methane provide evidence that at least some component of the carbon in the methane was initially sourced from water column DIC [CO32−, HCO3−, and CO2 (aq)].

Discussion

Methane Assent within the Basin.

The original isotopic composition of methane might potentially change during its assent from the deep subsurface. Equilibrated free methane that becomes trapped at the base of the GHSZ is unlikely to re-equilibrate. The stored gas is then later subject to being released by sublimation on the top horizon of the GHSZ (36). Isotope fractionation associated with this process is poorly constrained but is expected to only affect δD (and not δ13C) values by a few per mil (37). However, at site 2A-1 where a high methane discharge rate emanates from a small DHI that is itself intersected by a pronounced top of the diapir to seafloor crestal fault likely requires fluid flow to largely bypass the GHSZ (Fig. 3 and SI Appendix, Table S2). Isotope fractionation of methane by biodegradation during transit is also unlikely based on the acquired δ13CCH4, δD CH4, and Δ13CH3D seep gas values (SI Appendix, Figs. S7 and S8). The methane from the deep subsurface is therefore unlikely to have significantly changed the recorded formation temperature.

Gas Production by Deep Biosphere Oases.

Recent studies show a diverse microbial community inhabits the near surface and deeply buried sediments within the Scotian Slope (19, 20). Thermophilic endospores sourced from deep within the basin and brought to the surface through cold seep fluid flow were detected in shallow surface and core sediments in the surrounding area (21, 38). Their presence is consistent with a deep biosphere that likely includes biodegradation of hydrocarbons from a yet known petroleum-hosting reservoir. Our finding of a dominantly biogenic gas–producing region in what has historically been viewed as a largely thermogenic petroleum play area is noteworthy. The closest analog to our finding is the microbial activity associated with Gulf of Mexico salt tectonism that arises by the deep infiltration of marine and meteoric water to produce dissolution of anhydrite. The elevated supply of Ca+ and sulfate enables oxidation of petrogenic hydrocarbons by AOM in the oil and gas reservoirs trapped on the flanks of the salt diapirs (39, 40). The sustained process marks a mechanism of cap carbonate formation and thus represents a previously unrecognized carbon sink (40, 41).

A similar AOM-dominated deep biosphere for this portion of Scotian Slope is unlikely. There is a near-complete absence of thermogenic hydrocarbons at the two seep sites (SI Appendix, Table S3), and the seismic reflection data are unclear with respect to the presence of pronounced cap carbonate deposits above diapirs for this part of the diapiric province (Fig. 3). There are no deeper DHIs below the salt diapir edifices, and the thickness of rock strata at and above the diapirs is marked by thermally immature, relatively organic lean intervals with poor petroleum generative potential (35). The seep methane at sites 2A-1 and 2B-1 is also in near equilibrium with δ13CCH4 and δDCH4 values most consistent with non-biodegradation processes (SI Appendix, Fig. S7). Therefore, it appears in the absence of elevated bacterial sulfate reduction and AOM, hydrogenotrophic methanogenesis, utilizing the oxidized form of a carbon substrate (CO2 or acetate common to freshwater environments), dominates the metabolic strategy of the deep biosphere. The most likely sources of DIC and metabolites are from adjacent formation waters and by input from the overlying ocean water column, which collectively can also result in partial dissolution of evaporitic minerals at the crests of the salt diapirs (40). Taken together, these results suggest a two-end member, substrate-dependent flexibility to salt diapir-deep biosphere ecosystems. However, the preferential enrichment of such biologic growth factors requires additional mechanistic influences.

Salt Tectonism and the Local Petroleum System.

Almost all identified DHIs across this portion of the Scotian Slope are found directly over top salt diapirs. They extend laterally outward to no further than the radius of the underlying salt structure (Fig. 4A). Such a confined fluid-hosted setting is difficult to explain for this petroleum-producing basin. Prominent potential structural traps are abundant off-axis of most DHI-hosted salt structures due to the extensive network of crestal faulting, the presence of anticline-generating bedforms in contourite deposits at a similar stratigraphic level, and widespread distribution of a variety of other potential stratigraphic seals at lateral pinch-outs at sites of onlap or truncation beneath erosional unconformities (Fig. 3 and SI Appendix, Figs. S10–S18). Furthermore, larger fault systems including growth faults are common adjacent to, but also sometimes independent of, the salt diapirs themselves. The many stratigraphic pinch-outs on the flanks of minibasins or growth-fault-related anticlinal traps would similarly also produce seismic amplitude anomalies if they were appropriately charged. Yet these off-diapir structures do not have clear DHIs in the investigated 1,500 m thickness of the RMS amplitude maps or seismic sections. Furthermore, only 27% of the diapirs have recognizable DHIs or gas chimneys. Most DHIs appear in the northern, more shallow-water sections of the Shelburne survey, where the diapirs are more deeply buried, suggesting the thickness of cover also has a relationship to DHI development (Fig. 1 and SI Appendix, Figs. S2 and S3).

Fig. 4. (A) Coherence map of DHI 14 (green polygon) showing the location and diameters of its underlying salt diapir (yellow line) and the horizontal extent of the resulting radial fault system (blue dotted line). The red-lined graph indicates measured ocean floor sediment heat anomalies resulting from the salt diapir (42). (B) Cross-plots of numerical modeled Rayleigh (Ra) and Péclet numbers (Pe) for different hypothesized basin circulation depths considering isotropic sediment (Left) and fault impacted (Right) permeabilities (green fields indicate conditions favorable for convective flow) (B, T, and M mark diapir sites with biogenic, thermogenic, and mixed source methane) (SI Appendix, Table S6). Conditions for ground-water circulation are reached when Ra and Pe values are ≥40 and 1, respectively. (C) Cross-plot of the Shelburne and Tangier salt diapir depths and the ratio of a diapir’s diameter to its overlying radial fault diameter (referred to here as its η value), indicating diapirs are typically found at ≥1,000 mbsf with a η ≥ 1.2. (D) Model of DHI and cold seep formation. When diapirs that produce radial fault systems have high enough η values, cold, nutrient-rich seawater is able to convectively flow to the diapir crest where it enhances deep biosphere methane production. When diapirs reach close to the ocean floor or have narrow radial faults their η values converge to 1, thereby removing the potential of convective flow from the overlying water column.

Heat Chimney Effect and Convective Flow Model.

Salt is an effective conductor of heat. Its thermal conductivity is 6.1 W m–3 K–1 at 0 °C, decreasing to 4.1 W m−3 K−1 at 100 °C (43). The thermal conductivity is much higher than typical clastic and carbonate sediments values that range from 1.5 to 2.5 W m−3 K−1 (44). This well-known feature within salt provinces is credited for elevating the thermal maturity of source rocks above salt structures and the cooling of rocks that lie below or adjacent to salt bodies (45). The differential heat flow focused atop diapirs is known as a salt chimney effect (43, 46, 47). For the Shelburne Subbasin, a regional seafloor surface sediment heat flow of 35 mW m−2 has been established (42). The heat flow dramatically increases precisely overtop deeply buried diapirs (for example reaching 45 W m−2 over top of diapir #14; Figs. 1 and 4A) (42). The dramatic and predictable heat flow changes provide direct evidence of salt chimney effects within the Shelburne diapiric province. While the elevated heat flow represents a mechanism for deep formation water to reach the cold seep, it is unclear why biogenic methane production within such a diapiric petroleum system would also be elevated. Here, we posit the contribution of convective flow from adjacent formation waters or from the overlying water column linked by radial faulting increases bicarbonate and nutrient supplies in a subsurface circulation cell. To evaluate this hypothesis, we adopted a 2D isotropic numerical groundwater model for the prediction of subsurface convection based on classical Rayleigh–Taylor instability theory for a homogeneous sedimentary system (Fig. 4B and SI Appendix, Table S6). For this model, Rayleigh numbers (Re) ≥ 40 indicate unstable systems where convective heat transfer dominates the thermal behavior. Péclet numbers (Pe) ≥ 1 indicate groundwater flow will begin to alter the subsurface temperature field (48). A groundwater system undergoes convective flow when these two conditions are collectively met (43). The geothermal temperature estimates and depth profiles of DHI sites for three hypothetical seafloor linked convective cells were tested in the isotropic sediment-groundwater flow model: i) down to the depth of the DHI; ii) to the top of the diapir; and iii) to the base of the diapir. From these parameters, only the deepest convective cell iii is predicted to form in a partial number of DHI locations (Fig. 4B and SI Appendix, Table S6). However, changing the sediment permeability (see SI Appendix, Table S6 for details) to incorporate expected fluid movement along the radial fault splays results in predicted convective flow for the hypothesized ocean floor-top of the salt diapir circulation cell ii. The shallowest circulation cell reaching down to the DHI i was not found to have favorable conditions for fluid circulation (Fig. 4B and SI Appendix, Table S6).

Although numerical modeling suggests conditions could enable convective flow, other factors must also affect seep formation as not all diapirs form DHIs or produce gas chimneys. We therefore investigated the geometric relationship between a diapir and the overlying sediment strata as a further control on a cold seep–diapir–deep biosphere complex. A ratio of the diameter of the salt’s heat chimney (assumed here to be the diameter of the salt diapir) to the diameter of the radial fault system at its closest point to the seafloor (herein referred to as its η value) is used as a means of determining the capacity for convective flow from the overlying ocean water column. When compared to the diapir’s depth from the ocean floor (Fig. 4C), DHI sites are found to have η values ≥1.2. Except for the only detected thermogenic gas site (SI Appendix, Fig. S19), DHI-hosting diapirs are also deeply buried (≥1,000 mbsf). For our study region, diapirs that do not host DHIs have a broader range of diapir depths, but generally much less extensive radial fault systems (lower η values). These results suggest convective influxes of DIC- and nutrient-rich seawater can contribute to the elevated methanogenic activity of the deep biosphere if the radial fault system sufficiently extends beyond the perimeter of the diapir’s heat chimney.

While direct evidence of advective flow comes from the presence of the seeps, and more specifically, the seepage rate of gas ebullition at sites 2A-1 and 2B-1 (SI Appendix, Table S2); evidence of convective seawater flow is expectedly more difficult to obtain as the net downward flux will be non-linearly divided across the total number of radial fault limbs (i.e., >20 faults for site 2A-1). The radiocarbon ages may provide further insight as these data could be interpreted as marking a ≤0.05 m y−1 seawater draw down where the radial fault system extends out ~500 m beyond the perimeter of the heat chimney formed by the diapir that is buried 1.1 km below the 2A-1 cold seep. Such a slow convective flow rate would be difficult to directly measure from near-surface sediment porewater diffusion gradients. However, the radiocarbon age data of the methane more likely indicate a portion of the methane seepage is instead sourced from fermentative processes in the shallow subsurface. Under this condition, based on a regionally calibrated 0.3 to 0.6 mm y−1 sedimentation rate (49), the 14C portion of the methane entering into the ebullient stream via the biodegradation of buried sedimentary organic matter, will have a maximum depth of ~13.4 to 26.9 mbsf in line with other seep-related surveys (50). Future seep studies should aim to link clumped isotope and radiocarbon analytical methods to resolve mixed biogenic gas signatures for the further identification of cold seep methane sources.

Conclusion

This study provides evidence that the elevated geothermal gradients overlying deeply buried salt diapirs drives the formation of a deep biosphere—diapir—cold seep interlinked ecosystem. These ecosystems are established by the heat-driven convective movement of DIC- and nutrient-rich formation waters and downward transported seawater through the crestal fault network residing above the diapir. The movement of waters near the edifice of the salt diapir feeds methanogenic archaea to give rise to a deep biosphere oasis. These oases elevate biogenic methane production, which then migrates to the seafloor where it fuels cold seeps along the Scotian Slope of Atlantic Canada. In this context, the deep biosphere is shown to be responsible for creating and sustaining the existence of what can be among the most biologically diverse deep ocean floor ecosystems. As the biogenic gas umbilical is dependent on the hydrographic controls of the larger diapir-fault architecture, we provide a method to track these interlinked ecosystems based on their geophysical character. These results may therefore lead to better models for where cold seeps can form, the biogeochemical processes that sustain them, and the type of hydrocarbons they host.

Methods

Geophysical Surveys.

The upper 5,500 m of the Shelburne 3D Seismic Survey (conducted by Shell Canada Ltd. in 2013), a large, wide azimuth survey with a coverage of 10,400 km2 (RPS, 2013; Canada-Nova Scotia Offshore Petroleum Board (CNSOPB) program number NS24-S6-3E) was investigated alongside the Tangier 3D Seismic Survey (conducted by BP Canada Exploration Ltd. in 2014, CNSOPB program number NS24-B71-1E) spanning 8,000 km2. Common depth point (CDP) bin spacing/output grid for Shelburne 3D is 12.5 m by 50 m (time sampling rate: 2 ms). The Shelburne survey used both full 3D anisotropic Kirchhoff prestack time migration (PSTM) and full-volume anisotropic Kirchhoff prestack depth migration (PSDM) with vertical transverse isotropy. BP Exploration Ltd. acquired four exploration licenses (Parcels 5 to 8; EL 2431-2434) from the Nova Scotia Call for Bids NS12-1 in the deep-water Scotian Slope. CDP bin spacing/output grid for Tangier 3D is 25 m and 25 m (time sampling rate: 2 ms). Both seismic surveys are in SEG-Y format and are post-stack, depth-migrated. Together, these surveys range from ~500- to >4,000-m water depths along the Scotian Slope.

These data blocks were mapped for the presence of DHIs and other types of seepage evidence and basin structural features that could promote migration of hydrocarbons using v2019 Schlumberger’s Petrel © seismic software. Seismic horizons were picked based on the difference in reflection geometry, continuity, and configuration, known as seismic facies. Horizons were commonly chosen for mapping based on strong reflections with good continuity and differences in seismic facies. For this study, we picked and regionally mapped horizons across the shallow subsurface for both surveys. Seismic attributes analysis included RMS amplitude and coherence attribute extraction. For the Shelburne data block, the RMS amplitude extraction search windows was ± 20 m from the chosen horizon. RMS amplitude maps were extracted for 0, 200, 500, 750, 1,000, 1,500, and 2,500 mbsf. For the Tangier data block, RMS amplitude extraction search window was ±25 m from the chosen horizon. RMS amplitude maps were extracted are: 3,250 m (P1), 3,000 m (N1), 2,750 m (N2), 2,500 m (N3), 2,250 m (N4), 2,000 m (N5), and 1,750 m (N6) from sea level. For both seismic surveys, the search window for coherence extraction was 0 m (i.e., at the horizon itself). DHIs are evaluated based on the nature of the amplitude anomalies (e.g., bright spots, dim spots, polarity reversal, etc.) which are unusual for their stratigraphic context (e.g., isolated amplitude anomalies). An anomaly was assigned as a fluid-derived “bright spot” DHI if a strong lateral amplitude variation is seen in seismic section without characters typical of a stratigraphic feature (e.g., paleocanyon with expected distinct lithology, or an unconformity surface with compaction or cementation expected beneath it). If stratigraphic explanations are not tenable, and especially if there is a plausible structural control (e.g., truncation against a fault at a structural culmination), the amplitude anomalies are interpreted to be created by fluid migration. Seismic sections with prospective DHIs are annotated and highlighted to document the abundance of fluid-related amplitude anomalies. Isocore maps of seafloor to diapir crests for the Shelburne data block were also used to determine diapir depths (SI Appendix, Table S1 and Fig. S3).

Seep Surveys.

A 2020 autonomous underwater vehicle (AUV) cruise conducted by the Bedford Institute of Oceanography’s CCGS Hudson, surveyed 10 prospective seep sites in the Shelburne and Tangier area based on the presence of DHI mapping and other geophysical data relating to fluid escape features and/or geochemical data generated from three prior piston and gravity coring surveys (SI Appendix, Fig. S4) (18, 49, 51). In 2020, the HUGIN 6000 AUV (Kongsberg Maritime) was deployed from the offshore supply ship Pacific Constructor. The AUV collected high-resolution geophysical data over a 2.5 km by 2.5 km area. It was steered ~40 m above the seafloor for multibeam bathymetric, side scan sonar, and sub-bottom profiling. Data collection was done by a Kongsberg EM 2040 multibeam echosounder and an EdgeTech 2205 sonar system. The EM 2040 multibeam bathymetric data were acquired at a 400 kHz frequency, with a continuous waveform pulse and synchronized with Doppler velocity log. CARIS and EIVA suite was used for multibeam bathymetric data processing performed on the. Side scan sonar data were acquired at 230-kHz frequency with post data processing being completed with SonarWiz. The sub-bottom profiler operated over a 1 to 9 kHz frequency range with a 20-ms pulse. High-resolution seismic data were integrated and analyzed using I Kingdom SuitIIHS Markit Ltd.). Acoustic travel times for high-resolution sub-bottom profiler lines were converted into depths using an average seismic velocity of 1,500 m s−1.

The resulting multibeam benthic maps were then used to direct a 2021 Triton XLX remote-operated vehicle (ROV) operated by Helix Robotics Solutions Ltd., push core survey of the seep sites using the Atlantic Condor, operated by Atlantic Towing Ltd. Push core devices and core transport units (quivers) were designed by Todd Ventura and Jeremy Bentley at Saint Mary’s University and built by Velocity Machining and Welding Ltd. in Dartmouth NS. Push core sediments were collected at “hot spot” locations deemed to be of scientific interest for all seep sites (SI Appendix, Table S1 and Figs. S5 and S6) (17). High-definition video recordings were made using a Simrad OE 1366 SD color zoom camera (Kongsberg, Norway).

Sample Collection and Geochemical Analysis.

Seepage rate analysis.

A seepage rate was measured by the collection of gas using a graduated funnel and measurement of the area of gas ebullition on the seafloor (SI Appendix, Table S2). An annual rate of methane emission assuming all the hydrocarbon gas reached the ocean surfaces was calculated as[3] CH4molarmass×CH4moleemissionrate×525,600s1000000,

where the mole rate CH4 emission is[4] Pocean bottom×Seepage rateRTocean bottom,

and the Pocean bottom was estimated by[5] p×g×h,

with ρ being the density of seawater at 1,030 kg m−3, g the acceleration due to gravity 9.81 m s−2, and h the height of the overlying water column.

Gas sampling and analysis.

Hydrocarbon gases (methane, ethane, and propane) were analyzed from i) ocean floor bubbling seep gas, ii) void gas pockets inside of push corer liner, and from iii) push core sediment headspace gas (SI Appendix, Tables S3 and S4). Seep gas was directly collected from bubbles escaping the seafloor at sites 2A-1 and 2B-2 using a custom-designed non-isobaric seep gas sampler (United States Provisional Application No. 63/392,224). Once onboard the vessel, the collected seep gas was transferred by a hand pump to Isobags. Subsamples of push cores were collected for geochemical headspace analyses in which 5 mL of sediment from the base of the push corer was subsampled into a 20-mL glass serum vial with 5 mL 1M NaOH as a biocide. The serum vials were sealed using pre-treated (1 M KOH bath and rinsed overnight) Bellco blue butyl stoppers and stored in the refrigerator.

A 1-mL aliquot of the head space gas was removed using a Hamilton airtight syringe and injected into an Agilent 7890 series gas chromatograph with a flame ionization detector (GC-FID) at Saint Mary’s Department of Geology. The GC-FID was adapted with appropriate valve and injection systems by Wasson-ECE Instrumentation. Analysis of trace combustible hydrocarbon phases used an alumina-PLOT capillary analytical column (50-m × 0.53-mm × 10-μm) with a flow rate of 11.9 ± 0.5 mL min−1 of ultrapure He carrier gas. The initial oven temperature was set to 35 °C for 6 min and then ramped 10 °C min−1 to 200 °C for 2.5 min. Analysis of non-combustible volatile species (e.g., CO2) was made using an alumina-PLOT analytical column (30-m × 0.53-mm × 10-μm) with an 8.65 ± 0.5 mL min−1 flow rate of ultrapure He as the carrier gas. The oven temperature was initially set to 35 °C for 6 min and then ramped 10 °C min−1 to 110 °C for 7 min followed by a second ramp of then 20 °C min-1 to 200 °C for 10 min. Identification and standardization of gas species was performed using standard gas mixtures supplied by Matheson Tri-Gas and SCOTTY® Specialty Gases. Lower limits of quantification are approximately 10−12 mol L−1 varying slightly based on compound and further based on a minimum integration area of 0.5 units (i.e., in Agilent ChemStation software). Standards for quantification of gas and identification of gas peak retention time used Scott 1, Scott 2, and customized diluted standards. GC-FID runs were replicated at intervals to check for error and sensitivity of the process.

Void and sediment gas sampling and analysis.

Void space samples were extracted from the push core liners using a hand drill and syringe. They were then stored in pre-evacuated 50-mL glass serum bottles. Whole-round core sediment samples (1.5 cm) were sectioned, and vacuum sealed into gas-barrier bags, and frozen at −80 °C until extraction. Extraction of gas from frozen sediment samples was accomplished using a custom-designed gas extractor that thaws sediment samples under vacuum to minimize air contamination. The stainless-steel sample chamber in the gas extractor was kept at −20 °C using an isopropanol/LN bath while being loaded with frozen sediment. The cylinder was then placed under vacuum for 15 min before the cold bath was removed and then warmed overnight to room temperature. A GC-FID was then used to measure the C1 to C3 gas concentrations of extracted gas. The GC-FID was equipped with a 10´column packed with HayeSep-Q (VICI), and operated at a temperature of 90 °C, with helium as the carrier gas, and calibrated using SCOTTY® gas mixes. Methane was extracted and purified from the total gas with a series of cryogen traps (dry ice, LN2) using an automated preparative GC system (27).

Bulk stable isotope analyses.

Stable isotope ratios of carbon (δ13C-CH4) and hydrogen (δD-CH4) are reported (SI Appendix, Table S4) using standard delta notation with Vienna Pee Dee Belemnite (VPDB) and Vienna Standard Mean Ocean Water (VSMOW) standards for the ratios 13C/12C and D/H:[6] δ 13C= 13C 12Csample 13C 12CVPDB-1.

[7] δD=DHsampleDHVPDB-1.

Measured isotope values are reported in per mil (‰). The isotope scale was calibrated by the measurements of IAEA reference standards NGS-1 and NGS-3 (27). The δ13C-CH4 and δD-CH4 values in this study are based on isotopologue absorption that are proportionally comparable to 13CH4/12CH4 and D/H values (36).

Clumped isotope analyses.

The abundances of methane isotopologues were quantified using a Tunable Infrared Laser Direct Absorption Spectrometer (TILDAS) at the Isotope Geobiology Lab at MIT (52, 53). Cryogenically purified methane was extracted from each gas sample. The instruments’ two bellows were loaded with equal volumes of commercial methane reference gas (AL1) alongside a natural sample gas for 8 to 10 measurement cycles (27). Approximately 7 to 10 mL of methane (at room temperature and pressure) was used for analysis of seep gas collected by the nonisobaric gas sampler. For void and sediment gas samples, 1 mL of methane (at room temperature and pressure) was used for analysis. The TILDAS is a non-destructive analytical method. All measured sample-standard gas sample pairs were recovered for future use. Methane isotopologue abundances of the sample with respect to the stochastic distribution were determined relative to the AL1 calibration reference gas (27). Δ13CH3D values are reported in per mil (‰) notation and represent the deviation of multiply substituted isotopologue 13CH3D abundance from that of the stochastic distribution in which:[8] Δ 13CH3D=ln 13CH3D 12CH4 13CH4 12CH3D.

A methane formation temperature calculated from the Δ13CH3D sample values assumes the measurement is reflecting equilibrium conditions (54). These measurements are therefore referred to as an “apparent temperature.” The temperature-dependent equilibrium constant (K) from the isotopologue exchange reaction:[9] CH4+ 12CH3D↔ 13CH3D+ 12CH4,

is primarily a function of temperature, and the apparent temperature of equilibrium in Kelvin can be derived from Δ13CH3D values as (54).[10] Δ13CH3DT=-0.11011000T3+1.04151000T2-0.52231000T,

The temperature dependence for the value of Δ13CH3D (Eq. 10) can yield slightly different results from recent experimental calibrations (27, 55–57) affecting both a reported Δ13CH3D value and the apparent temperature. However, the calibration uncertainty for Δ13CH3D values is less than the 95% CI of our measurements (<0.1‰ vs. ca. 0.2‰) and does not significantly alter the results with both approaches yielding consistent apparent temperatures within 1.5 to 4.0 °C with the calibration using Eq. 10 yielding slightly higher apparent temperatures (36).

Radiocarbon analyses.

Radiocarbon age dating was conducted on purified methane samples at NOSAMS AMS laboratory at Woods Hole Oceanographic Institution, MA, USA (see SI Appendix, Table S4 for more details).

Heat Flow Modeling.

The 2D modeled capacity for subsurface convection was based on parameterized conditioned as presented by Canova et al. (43). Calculation of the critical permeability threshold is based on the classical Rayleigh–Taylor instability theory to demarcate convective flow in a homogeneous system. Rayleigh numbers (Ra) were calculated based on the formula:[11] Ra=kzβ∇TgV∼Kθ,

in which k is permeability, z is thickness of the reservoir, β water expansivity, ∇T is temperature chance over domain, g is gravity, Kθ is thermal diffusivity of saturated matrix, and v˜ is water viscosity (see SI Appendix, Table S6 for further details). Ra values ≥ 40 indicate conditions of unstable systems where convective heat transfer will dominate the thermal behavior. Calculation of Péclet numbers was based on the formula:[12] Pe=AdvectivetranferrateDiffusivetransferrate=0.056kzgV∼ΓKθγ,

where Γ and ϒ are the ratio of volume heat capacity of saturated matrix to fluid and the coefficient of fluid density dependence on salinity, respectively. Pe ≥ 1 indicate groundwater flow will begin to alter the subsurface temperature field. Data for the modeled parameters are obtained from this study’s geophysical surveys (SI Appendix, Table S1) as well as other published sources detailed in SI Appendix, Table S6. We also introduce a fault convective transport capacity measure (η):[13] η=DfDd,

in which Df and Dd are the diameters of the upper overlying radial fault system and the diameter of underlying salt diapir. The upward fault propagation angle (θ) was held to a constant 35° with the vertex residing within the diapir. The angle was determined based on a comparative review of diapir containing 2D seismic sections. The radial fault diameters were predicted based on the following function:[14] b2=2⌈h2-200+0.5b1tan180-θ⌉tanθ,

where b1 is the measured diapir diameter and h2 is based on isocore-determined depths (SI Appendix, Table S6 and Fig. S3) and b2 is the calculated diameters of the upper extent of diapiric radial fault propagation. Several DHI cross-sections were mapped to test their agreement to the model based on a η value measured from the 2D inline cross-sections (n = 5). These tests resulted in ≥10% variance from the predicted diapir fault propagation estimate. The modeled diameters assume circular salt structures in aerial view. However, not all DHI-hosted diapirs conform to this design. Some are more elliptical or canopy shaped.

Supplementary Material

Appendix 01 (PDF)

Dataset S01 (XLSX)

Movie S1. 2A-1 Seep Site

Movie S2. 2B-1 Seep Site

We thank the Nova Scotia Provincial Government for support of this project and spear-heading the various research cruises and surveys that made the integrative data collection possible. We especially thank Carey Ryan of Net Zero Atlantic, project manager for two consecutive Genome Canada funded Genomics Applied Partnership Program (GAPP) grants who was responsible for the larger group research initiatives. In addition, we especially thank Kim Doane, Executive Director of the Department of Natural Resources and Renewables for her support of the Organic Geochemistry Research Group at Saint Mary’s University. Natural Resources Canada provided one of the AUV multibeam bathymetric maps used for the ROV seep survey. Greg Slater of McMaster University suggested the use of radiocarbon age dating of the methane and Cliff Walters for editorial feedback. We thank the captain and crew of the Atlantic Condor, and the Helix Robotics Solutions Ltd., team, for a successful pushcore expedition and inaugural deployment of the Modular Ocean Research Infrastructure (MORI) vision from Dalhousie University. Jeff Seewald of Woods Hole Oceanographic Institution, Peter Girguis of Harvard University, and Andres Teske of University of North Carolina helped with push corer designs. Thanks to Britta Fiander of Genome Atlantic for supervision and guidance over the GAPP application and funding period. Thanks to Andrew Stevens of Velocity Machining & Welding Inc., Dartmouth, Nova Scotia, for help with the manufacture of the push core system and non-isobaric gas sampler. Porewater chemical analysis was aided by undergraduate students Macy Giles, Elish Redshaw, and Andrew Drane. Clifford Voss of the Water Resources Research Center, University of Hawaii at Mānoa, provided guidance for the numerical groundwater model. Lastly, we thank the reviewers and editors of PNAS for their help in the review and editing of this manuscript. Funding for this work was provided from Genome Atlantic and Genome Canada (GAPP); Research Nova Scotia (grant no. 2142); Mitacs (grant no. IT12481 and IT29547), together with Net Zero Atlantic and the Government of Nova Scotia. This research has been supported by the Natural Sciences and Engineering Research Council of Canada (NSERC; grant no. RGPIN-2018-06147), the Canada Research Chairs (CRC) program, and the Canada Foundation for Innovation (CFI; JELF–CRC, John R. Evans Leaders Fund).

Author contributions

A.C., G.T.V., N.M., S.O., A.M., R.B., and J.N.B. designed research; A.C., G.T.V., E.J.L., N.M., J.M.D., M.F., and J.N.B. performed research; A.C., J.N.B., and M.J.K. contributed new reagents/analytic tools; A.C., G.T.V., Y.O., E.J.L., N.M., J.M.D., S.O., and M.F. analyzed data; and A.C., G.T.V., E.J.L., N.M., J.M.D., S.O., M.F., A.M., R.B., R.A.M., and C.R.J.H. wrote the paper.

Competing interests

The authors declare no competing interest.

Data, Materials, and Software Availability

All study data are included in the article and/or supporting information.

Supporting Information

This article is a PNAS Direct Submission.
==== Refs
1 E. Suess, Marine cold seeps and their manifestations: Geological control, biogeochemical criteria and environmental conditions. Int. J. Earth Sci. 103 , 1889–1916 (2014).
2 P. Pop Ristova, F. Wenzhöfer, A. Ramette, J. Felden, A. Boetius, Spatial scales of bacterial community diversity at cold seeps (Eastern Mediterranean Sea). ISME J. 9 , 1306–1318 (2015).25500510
3 A. Boetius , A marine microbial consortium apparently mediating anaerobic oxidation of methane. Nature 407 , 623–626 (2000).11034209
4 V. J. Orphan, C. H. House, K.-U. Hinrichs, K. D. McKeegan, E. F. DeLong, Methane-consuming Archaea revealed by directly coupled isotopic and phylogenetic analysis. Science 293 , 484–487 (2001).11463914
5 S. B. Joye, M. W. Bowles, V. A. Samarkin, K. S. Hunter, H. Niemann, Biogeochemical signatures and microbial activity of different cold-seep habitats along the Gulf of Mexico deep slope. Deep Sea Res. Part II Top. Stud. Oceanogr. 57 , 1990–2001 (2010).
6 A. G. Judd, M. Hovland, Seabed Fluid Flow: The Impact of Geology, Biology and the Marine Environment (Cambridge University Press, 2007).
7 M. R. Hudec, M. P. A. Jackson, Terra infirma: Understanding salt tectonics. Earth Sci. Rev. 82 , 1–28 (2007).
8 H. R. Grunau, A worldwide look at the cap-rock problem. J. Pet. Geol. 10 , 245–265 (1987).
9 M. Römer , Seafloor methane seepage related to salt diapirism in the northwestern part of the German North Sea. Front. Earth Sci. 9 , 556329 (2021).
10 E. Suess, “Marine cold seeps: Background and recent advances” in Hydrocarbons, Oils and Lipids: Diversity, Origin, Chemistry and Fate, H. Wilkes, Ed. (Springer International Publishing, 2018), pp. 1–21, 10.1007/978-3-319-54529-5_27-1.
11 A. V. Milkov, Global estimates of hydrate-bound gas in marine sediments: How much is really out there? Earth Sci. Rev. 66 , 183–197 (2004).
12 J. C. Fry, R. J. Parkes, B. A. Cragg, A. J. Weightman, G. Webster, Prokaryotic biodiversity and activity in the deep subseafloor biosphere. FEMS Microbiol. Ecol. 66 , 181–196 (2008).18752622
13 D. L. Valentine, Emerging topics in marine methane biogeochemistry. Annu. Rev. Mar. Sci. 3 , 147–171 (2011).
14 D. Giovannelli , Diversity and distribution of prokaryotes within a shallow-water pockmark field. Front. Microbiol. 7 , 941 (2016).27379070
15 M. Albertz, C. Beaumont, J. W. Shimeld, S. J. Ings, S. Gradmann, An investigation of salt tectonic structural styles in the Scotian Basin, offshore Atlantic Canada: 1. Comparison of observations with geometrically simple numerical models. Tectonics 29 , 1–29 (2010).
16 M. E. Deptuck, K. L. Kendell, “Chapter 13–A review of Mesozoic-Cenozoic salt tectonics along the Scotian Margin, eastern Canada” in Permo-Triassic Salt Provinces of Europe, North Africa and the Atlantic Margins, J. I. Soto, J. F. Flinch, G. Tari, Eds. (Elsevier, 2017), pp. 287–312, 10.1016/B978-0-12-809417-4.00014-8.
17 R. Bennett, P.-A. Desiage, Expedition report 21 CONDOR: Scotian Slope, August 14-29, 2021 (Open File 8889, Geological Survey of Canada, 2022), 53 p. 10.4095/329977. Accessed 23 May 2022.
18 D. C. Campbell, CCGS Hudson Expedition 2016-011, phase 2. Cold seep investigations on the Scotian Slope, offshore Nova Scotia, June 15-July 6, 2016. 88. https://geoscan.nrcan.gc.ca/starweb/geoscan/servlet.starweb?path=geoscan/fulle.web&search1=R=313603 (2019). Accessed 5 March 2021.
19 X. Dong , Thermogenic hydrocarbon biodegradation by diverse depth-stratified microbial populations at a Scotian Basin cold seep. Nat. Commun. 11 , 5825 (2020).33203858
20 C. Li , Bacterial anomalies associated with deep sea hydrocarbon seepage along the Scotian Slope. Deep Sea Res. Part Oceanogr. Res. Pap. 193 , 103955 (2023).
21 D. A. Gittins , Geological processes mediate a microbial dispersal loop in the deep biosphere. Sci. Adv. 8 , eabn3485 (2022).36026445
22 D. C. Mosher, A margin-wide BSR gas hydrate assessment: Canada’s Atlantic margin. Mar. Pet. Geol. 28 , 1540–1553 (2011).
23 M. Fowler, J. Webb, H. Olsen, F. Ashraf, S. Guldbrandsen, Geochemistry data report for 2016 Scotian Slope piston coring program. 310 (2017).
24 M. Fowler, J. Webb, S. Gulbrandsen, L.-K. Austnes, F. Ashraf, Geochemistry data report for 2018 Scotian Slope coring program (2019).
25 M. J. Whiticar, Carbon and hydrogen isotope systematics of bacterial formation and oxidation of methane. Chem. Geol. 161 , 291–314 (1999).
26 A. V. Milkov, G. Etiope, Revised genetic diagrams for natural gases based on a global dataset of >20,000 samples. Org. Geochem. 125 , 109–120 (2018).
27 D. T. Wang , Nonequilibrium clumped isotope signals in microbial methane. Science 348 , 428–431 (2015).25745067
28 D. A. Stolper , Formation temperatures of thermogenic and biogenic methane. Science 344 , 1500–1503 (2014).24970083
29 D. A. Stolper , Distinguishing and understanding thermogenic and biogenic sources of methane using multiply substituted isotopologues. Geochim. Cosmochim. Acta 161 , 219–247 (2015).
30 T. Giunta , Extreme methane clumped isotopologue bio-signatures of aerobic and anaerobic methanotrophy: Insights from the Lake Pavin and the Black Sea sediments. Geochim. Cosmochim. Acta 338 , 34–53 (2022).
31 J. Gropp, Q. Jin, I. Halevy, Controls on the isotopic composition of microbial methane. Sci. Adv. 8 , eabm5713 (2022).35385305
32 A. Wilhelms , Biodegradation of oil in uplifted basins prevented by deep-burial sterilization. Nature 411 , 1034–1037 (2001).11429600
33 F. Inagaki , Exploring deep microbial life in coal-bearing sediment down to ~2.5 km below the ocean floor. Science 349 , 420–424 (2015).26206933
34 I. M. Head, D. M. Jones, S. R. Larter, Biological activity in the deep subsurface and the origin of heavy oil. Nature 426 , 344–352 (2003).14628064
35 OETRA, Play Fairway Analysis Atlas – Offshore Nova Scotia. Nova Scotia Department of Energy Report. 347 (2011). https://oera.ca/research/play-fairway-analysis-atlas.
36 E. Lalk , Clumped methane isotopologue-based temperature estimates for sources of methane in marine gas hydrates and associated vent gases. Geochim. Cosmochim. Acta 327 , 276–297 (2022).
37 A. Hachikubo , Isotopic fractionation of methane and ethane hydrates between gas and hydrate phases. Geophys. Res. Lett. 34 , 1–5 (2007).
38 J. E. Rattray , Elevated bacterial endospores associated with thermogenic hydrocarbon seeps in deep sea sediments. Org. Geochem. 177 , 104568 (2023).
39 H. Drake , Isotopic evidence for microbial production and consumption of methane in the upper continental crust throughout the Phanerozoic eon. Earth Planet. Sci. Lett. 470 , 108–118 (2017).
40 K. H. Caesar, J. R. Kyle, T. W. Lyons, A. Tripati, S. J. Loyd, Carbonate formation in salt dome cap rocks by microbial anaerobic oxidation of methane. Nat. Commun. 10 , 808 (2019).30778057
41 S. Kotelnikova, Microbial production and oxidation of methane in deep subsurface. Earth Sci. Rev. 58 , 367–395 (2002).
42 E. Negulic, K. E. Louden, The thermal structure of the central Nova Scotia Slope (eastern Canada): Seafloor heat flow and thermal maturation models. Can. J. Earth Sci. 54 , 146–162 (2016).
43 D. P. Canova, M. P. Fischer, R. S. Jayne, R. M. Pollyea, Advective heat transport and the Salt Chimney Effect: A numerical analysis. Geofluids 2018 , e2378710 (2018).
44 A. F. Birch, H. Clark, The thermal conductivity of rocks and its dependence upon temperature and composition. Am. J. Sci. 238 , 529–558 (1940).
45 U. T. Mello, G. D. Karner, R. N. Anderson, Role of salt in restraining the maturation of subsalt source rocks. Mar. Pet. Geol. 12 , 697–716 (1995).
46 A. Wilson, C. Ruppel, Salt tectonics and shallow subseafloor fluid convection: Models of coupled fluid-heat-salt transport. Geofluids 7 , 377–386 (2007).
47 Q. G. Zhuo , The salt chimney effect: Delay of thermal evolution of deep hydrocarbon source rocks due to high thermal conductivity of evaporites. Geofluids 16 , 440–451 (2016).
48 M. Huysmans, A. Dassargues, Review of the use of Péclet numbers to determine the relative importance of advection and diffusion in low permeability environments. Hydrogeol. J. 13 , 895–904 (2005).
49 D. C. Campbell, A. Normandeau, CCGS Hudson Expedition 2018-041: High-resolution investigation of deep-water seabed seeps and landslides along the Scotian Slope, offshore Nova Scotia, May 26–June 15, 2018. 66. https://geoscan.nrcan.gc.ca/starweb/geoscan/servlet.starweb?path=geoscan/fulle.web&search1=R=314695 (2019).
50 J. W. Pohlman , Methane sources in gas hydrate-bearing cold seeps: Evidence from radiocarbon and stable isotopes. Mar. Chem. 115 , 102–109 (2009).
51 D. C. Campbell, A. MacDonald, CCGS Hudson Expedition 2015-018, geological investigation of potential seabed seeps along the Scotian Slope, June 25-July 9, 2015. 75 (2016).
52 S. Ono , Measurement of a doubly substituted methane isotopologue, 13CH3D, by tunable infrared laser direct absorption spectroscopy. Anal. Chem. 86 , 6487–6494 (2014).24895840
53 Y. Gonzalez , Precise measurements of 12CH2D2 by tunable infrared laser direct absorption spectroscopy. Anal. Chem. 91 , 14967–14974 (2019).31663335
54 J. Bigeleisen, M. G. Mayer, Calculation of equilibrium constants for isotopic exchange reactions. J. Chem. Phys. 15 , 261–267 (2004).
55 M. A. Webb, T. F. I. Miller, Position-specific and clumped stable isotope studies: Comparison of the Urey and path-integral approaches for carbon dioxide, nitrous oxide, methane, and propane. J. Phys. Chem. A 118 , 467–474 (2014).24372450
56 Q. Liu, Y. Liu, Clumped-isotope signatures at equilibrium of CH4, NH3, H2O, H2S and SO2. Geochim. Cosmochim. Acta 175 , 252–270 (2016).
57 D. L. Eldridge , Comparison of experimental vs theoretical abundances of 13CH3D and 12CH2D2 for isotopically equilibrated systems from 1 to 500 °C. ACS Earth Space Chem. 3 , 2747–2764 (2019).
