Widespread natural methane and oil leakage from sub-marine Arctic reservoirs

The study addresses the challenge of distinguishing anthropogenic from natural geological fossil methane emissions in the atmosphere, given overlapping isotopic signatures (δ13C, δD) among sources. Due to a mismatch between top-down and bottom-up methane budgets and limited source-specific markers, mapping and quantifying geological methane inventories is needed. Marine methane sources are often attributed to microbial production in shallow sediments, and thermogenic seepage is underrepresented in global emission maps, especially in the Arctic. The Barents Sea shelf, with significant hydrocarbon reserves and a history of substantial uplift and glacial erosion that can unseal reservoirs and enhance fluid discharge, provides a natural laboratory to test whether eroded, formerly glaciated structures leak hydrocarbons. The study aims to document and analyze methane and oil leakage in the northern Norwegian Barents Sea, evaluate controls on seep distribution, and assess implications for methane transfer to the ocean and potentially the atmosphere.

Prior work highlights rapid atmospheric methane growth and uncertainties in the global methane budget, with discrepancies between top-down inversions and bottom-up inventories. Isotopic similarities hinder partitioning of fossil methane sources (e.g., hydrocarbon extraction, coal mining, natural marine and terrestrial seepage). Global marine methane emissions have been largely attributed to microbial processes in shallow sediments, and gridded Arctic emission maps show only a few seep areas dominated by microbial gas. However, the Arctic shelves host extensive hydrocarbon reserves and have experienced multiple erosive Quaternary glaciations that can promote fluid discharge by unsealing reservoirs, altering fault permeability, tilting reservoirs, and inducing pressure-driven seal failure. Oil-coated gas bubbles have been hypothesized to reduce dissolution and enhance methane transport to the surface. Evidence from the Barents Sea indicates compromised sealing potential due to uplift and erosion; traces of hydrocarbons in dry wells and sediments suggest widespread paleo-leakage, and modeling indicates significant overburden removal, motivating reassessment of thermogenic seepage contributions in glaciated basins.

The study integrates: (1) multibeam echosounder water-column backscatter to map gas flares; (2) SAR satellite imagery (Sentinel-1) to detect oil slicks; (3) 2D seismic reflection data and interpretations to relate seepage to subsurface structures and stratigraphy; (4) subbottom profiler (Chirp) to image the upper ~0–30 m subseafloor; and (5) seawater and sediment geochemical measurements. Data sources: four CAGE cruises (CAGE 18-1: May 2018; CAGE 19-2: July 2019; CAGE 20-2: July 2020; CAGE 21-4: August 2021) and an areal MAREANO dataset. - Multibeam acquisition/processing: Kongsberg EM302 (CAGE) with 120° opening angle, 432 soundings, swath up to ~3× water depth; ping rate auto-adjusted to water depth; vessel speed 6–8 kn; CTD (SBE 911plus) for sound velocity. MAREANO data used Kongsberg EM710. Water-column visualization performed in FMMidwater (QPS). Only the observable water volume (approximately 50% of swath, due to side-lobe artifacts) was used for flare picking. Manual picking of flare roots in fan-view was performed once per overlap. - Flare classification: Flares were categorized as weak, medium, or strong by apparent scattering strength and water-column height. Verification via raw amplitude thresholds (e.g., strong > ~65; weak < ~85) and random inspections; constant color ranges and single interpreter ensured consistency. Density maps produced via quartic kernel interpolation (search radius 500 m). - SAR oil slicks: Seven Sentinel-1 images (April–October 2020) were selected under suitable weather conditions. Oil slick outlines identified manually in ArcGIS from low-backscatter patches; emission points derived from repeated outlines in three persistent slick areas. - Seismic: One GI air gun source; 100 m streamer with 32 channels at 3.125 m spacing (four 25 m P-Cable sections). Processing in Radex Pro: CDP binning (3.125×3.125 m), bandpass (10-25-300-400 Hz), spherical divergence correction, bubble removal, NMO to water velocity, zero-offset demultiple, post-stack Kirchhoff migration. Interpretation in Petrel. - Subbottom profiler (X-STAR Full Spectrum Sonar): 1.5–9 kHz linear sweep at 0.3 Hz to image 0–30 m subseafloor; allowed distinction of soft marine sediments, glacigenic deposits (~3–10 m), and lithified strata. - Water sampling and analysis: CTD rosette samples into 120 mL bottles; 1 mL 1 M NaOH added; 5 mL N2 headspace added prior to GC-FID (ThermoScientific Trace 1310) for dissolved CH4 quantification. Sediment gas composition from gravity cores via headspace technique (5 mL sediment + 5 mL 1 M NaOH) analyzed on Trace 1310 with TG-BOND alumina (Na2SO4) column. - Sea–air methane flux: Bulk flux F = k(Cw − Co), with k = 0.251 u10^0.5 Sc^−0.5; u10 converted from measured wind at 22.4 m: u10 = Umeas (Zmeas/10)^−0.11; ~20% uncertainty in gas exchange parameterization. Equilibrium concentrations computed via Bunsen solubility at in situ T–S and atmospheric pressure. - Study areas: Sentralbanken High (detailed mapping), and comparison sites at Storbanken High and anticlines on Kong Karls Platform. Structural and stratigraphic context compiled from regional geology and NPD interpretations.

- Extensive seepage at Sentralbanken High: 4,137 acoustic flares (bubble emission sites) identified over ~660 km² of multibeam coverage. Flares cluster into distinct 1–>7 km² patches with sharp boundaries; solitary flares are only 2.7% of the total. Maximum flare densities exceed 100 seeps km⁻² in central apex areas. - Flare strength distribution: 620 strong, 1,384 medium, and 2,133 weak flares. Stronger and taller flares are concentrated within central clusters; rarefied clusters at flanks show ≤60 flares km⁻² with >76% weak flares. - Vertical reach: 89 flares reached to within <50 m of the sea surface before terminating or leaving the echosounder footprint, implying more may reach the upper water column than observed. Elsewhere it is noted that ~80 flares reached the surface mixed layer, but none were observed to reach the sea surface. - Structural controls: Highest flare densities do not consistently correlate with mapped faults or bright spots beneath the seabed in the apex where reservoir units subcrop and are thickest. On flanks where overburden remains and reservoir thickness tapers, lower-density seep clusters (<50 seeps km⁻²) correlate with faults down to the top Kobbe Formation, indicating fault-controlled seepage where seals persist. No correlation found between pockmarks and active seepage. - Other sites: Kong Karls Platform anticlines (Snadd Fm. subcropping) correlate with 597 flares (discrete line survey). Storbanken High shows 2,646 flares within 2,810 km² (MAREANO dataset), with distinct flare strings correlating to subcropping Snadd, Tubåen, and Stø Formations. - Regional total and comparison: Across the three Barents Sea sites, 7,380 seeps were identified within 3,730 km² surveyed, indicating a seep density far exceeding ~1,000 seeps over 30,000 km² on the Western Svalbard margin and 570 seeps over 94,000 km² on the northern US Atlantic margin. Given limited coverage and observational constraints, the actual number is likely higher, making this a global hotspot of submarine methane release. - Methane in seawater: Throughout Sentralbanken, the water column was supersaturated with methane relative to atmospheric equilibrium at all stations and depths, indicating a pervasive plume. Sea–air methane flux from the surface mixed layer ranged from 1.2 to 3 µmol m⁻² d⁻¹, lower than reported at some shallow Svalbard seeps (2–20 µmol m⁻² d⁻¹), and subject to variability from stratification, currents, and wind speed. - Oil leakage: Persistent oil slicks were detected by SAR; oil droplets were observed reaching the sea surface during sampling, and some seabed seeps emit oil-coated bubbles. However, methane concentrations measured 5 m below slicks were not higher than outside slicks, and no surface ebullition or flares reaching the surface were observed, suggesting bubble collapse and methane dissolution in the upper water column despite oil coatings. - Atmospheric signal: Prior atmospheric measurements over Sentralbanken showed methane mixing ratios exceeding 2000 ppb in autumn–winter (background ~1950 ppb), previously attributed to distant land sources; these findings motivate re-evaluation given local marine emissions. - Geological context and persistence: The Barents Sea’s uplift and km-scale glacial erosion have exposed and partially uncapped Triassic–Jurassic reservoirs (e.g., Kobbe, Snadd, Tubåen, Stø) beneath limited sealing units (e.g., Hekkingen/Fuglen). Degassing through erosion likely enabled continuous leakage since the last major ice-sheet collapse ~15 ka BP. Similar exhumed, eroded, hydrocarbon-bearing basins across polar shelves may host widespread natural thermogenic leakage.

The results confirm that eroded, formerly glaciated structural highs in the northern Barents Sea act as major natural sources of thermogenic methane and oil to the ocean, addressing the knowledge gap about geological methane emissions in the Arctic. The observed clustering, high densities, and correlation with subcropping reservoir units indicate that regional exhumation and partial removal of seals by uplift and glacial erosion precondition reservoirs for persistent leakage independent of active faulting in apex areas, while faults facilitate seepage on flanks where overburden remains. Methane supersaturation throughout the water column and flares reaching the upper layers demonstrate efficient bubble transport, potentially augmented by oil coatings that reduce dissolution; nevertheless, bubbles collapse before reaching the surface, and methane largely dissolves into the surface mixed layer with modest but nonzero air–sea flux. The seep densities and persistence imply that natural geological emissions from such settings are underestimated in current Arctic and global methane budgets. The findings suggest that analogous basins across the Arctic and North Atlantic margins could substantially contribute to natural fossil methane emissions, complicating partitioning between anthropogenic and natural sources and necessitating improved mapping and quantification for budget reconciliation.

This study documents one of the largest known Arctic cold seep regions, revealing widespread natural methane and oil leakage from exhumed hydrocarbon reservoirs on the northern Norwegian Barents Sea shelf. Using integrated echosounder, seismic, SAR, and geochemical data, the authors show that seepage is concentrated over partially uncapped Triassic–Jurassic reservoirs, producing thousands of persistent seeps with exceptionally high local densities and methane supersaturation throughout the water column. Although bubbles do not reach the sea surface, methane likely contributes to atmospheric flux via dissolution and sea–air exchange. These results indicate that natural thermogenic emissions from formerly glaciated, erosional basins are underrepresented in current budgets and may materially affect the regional and global carbon cycle. Future work should prioritize systematic, spatially extensive mapping of seepage across analogous Arctic and North Atlantic basins, improved quantification of bubble-mediated methane transfer (with oil effects), year-round flux measurements under varying oceanographic and meteorological conditions, and integration of geological seep inventories into top-down atmospheric inversions to better partition natural and anthropogenic fossil methane.

- Observational constraints: Multibeam water-column data are limited by side-lobe artifacts, reducing the observable volume to ~50% of the swath and hindering detection of the true termination height of tall flares except beneath nadir; overlap was only partially mitigating. Discrete line coverage at Kong Karls Platform and nonuniform coverage overall imply undercounting of seeps. - Seismic resolution: Conventional seismic could not resolve the thin Quaternary veneer (~3–10 m) that may influence near-seafloor migration pathways; subbottom profiler provided only local coverage. Bright spots did not consistently correlate with seabed seepage, limiting predictive power from reflection attributes alone. - Attribution: While geological context strongly indicates thermogenic sources, direct source apportionment (e.g., isotopic signatures of emitted gas at scale) is limited in the presented text; oil presence suggests a thermogenic component but mixed sources cannot be fully excluded at all sites. - Flux quantification: Sea–air methane flux estimates are based on discrete surface water sampling and bulk parameterizations with ~20% uncertainty, and are sensitive to transient oceanographic and wind conditions; no direct eddy covariance or continuous flux measurements were reported. - Seasonality and persistence: Multi-year but episodic surveys constrain persistence indirectly; lack of full seasonal coverage may miss temporal variability in seep strength and atmospheric coupling. - Spatial extrapolation: Inferences to other Arctic basins are based on geological analogy and limited targeted surveys; comprehensive basin-wide mapping is needed to quantify regional contributions.