Global Biogeochemical Cycles

Stable carbon isotopic evidence for methane oxidation in plumes above Hydrate Ridge, Cascadia Oregon Margin

Authors


Abstract

[1] The transport and consumption of methane in the water column in the vicinity of the cold seeps of Hydrate Ridge on the Cascadia Oregon Margin were characterized using measurements of the stable carbon isotope composition of methane. The δ13C-CH4 values measured in the water column ranged from approximately −65 to −16‰, PDB. The combination of measured methane concentration data and the stable carbon isotope values from the same depths support the hypothesis of biogenically produced methane which enters the water column from dissolving bubbles released from cold seepages, likely as a consequence of destabilized methane hydrate (δ13C-CH4 = −65‰, PDB). Kinetic fractionation factors, α, associated with aerobic bacterial methane oxidation in the water column were calculated using a Rayleigh distillation equation applied to a subset of the data. Fractionation factors ranged from 1.002 to 1.013 (mean = 1.008) and were in the lower end of the range of those reported in the literature, a result likely due to the influence of temperature and mixing in plume waters. The fraction of methane remaining after oxidation calculated using the same Rayleigh model approach suggests that the aerobic oxidation of methane in the water column over Hydrate Ridge is nearly quantitative.

1. Introduction

[2] Methane is an important, radiatively active trace gas whose concentration in the troposphere has more than doubled since the beginning of the industrial era [Cicerone and Oremland, 1988]. The release of methane to the contemporary atmosphere results in a radiative forcing that is approximately 23 times that of carbon dioxide on a mole per mole basis [Lelieveld et al., 1998], thus considerable efforts have been made to identify and constrain the magnitudes of individual methane sources and sinks. Though estimates of individual source magnitudes in the global methane budget are constrained to no better than an order of magnitude, most analyses of the contemporary methane cycle are in agreement that wetlands, both natural and artificial, constitute the largest annual sources of methane to the atmosphere [Cicerone and Oremland, 1988; Fung et al., 1991]. Other anthropogenically-related methane sources such as releases from natural gas exploration and transmission, domestic ruminants, and biomass burning make up the majority of the remainder of the annual global methane source to the atmosphere.

[3] The magnitude assigned the annual global ocean methane source is generally thought to be small in comparison to the methane sources listed above. In two recent studies the annual global ocean source was estimated to represent up to 2% of the total annual source of methane to the atmosphere [Bange et al., 1994; Bates et al., 1996]. However, studies of the sea-air flux of methane belie the turnover of much larger quantities of methane by methane-oxidizing bacteria in select ocean environments. Methanotrophs are known to play an important role in regulating methane flux to the troposphere from surface ocean environments such as major coastal upwelling zones [Cynar and Yayanos, 1992; Iverson, 1996], seasonally anoxic inlets and seas [Ward et al., 1987, 1989; Scranton et al., 1993; Lammers et al., 1995], and shelf waters receiving significant input from estuarine river basins [Scranton et al., 1995], i.e., marine environments where methane concentrations can be many orders of magnitude larger than open ocean surface waters equilibrated with the contemporary atmosphere (approximately 40–80 nL/L [e.g., Bates et al., 1996]).

[4] Methane hydrates and hydrocarbon seepages found at the world's convergent margins have drawn considerable attention because of the biota that live symbiotically with methane oxidizing bacteria [Kulm, 1986; Ritger et al., 1987; Suess and Whiticar, 1989; Carson et al., 1990; Linke et al., 1994; Suess et al., 1998, 1999]. More recently, attention has been focused on the question of the potential impact of these “cold seeps” on tropospheric methane levels on annual, decadal and millennial timescales [Paull et al., 1991; Dickens et al., 1995]. There is general consensus that methane hydrates at the world's continental margins represent a significant methane reservoir, orders of magnitude larger than the contemporary atmospheric methane burden [Kvenvolden, 1993; Gornitz and Fung, 1994]. Gornitz and Fung [1994] estimated that the global reservoir of methane trapped in hydrate is between 26 and 139, or between 18,570 and about 99,360 Pg CH4 (1 Pg = 1 × 1015 g) under STP conditions. The possibility that methane from destabilized hydrate could represent a much larger methane source to current or past atmospheres than has yet been considered has motivated the study of the factors controlling the release, transport and fate of methane from destabilized hydrates found in the in the sediments of accretionary prisms [Suess et al., 1998, 1999].

[5] In two recent studies, the potential for methane releases from cold seeps to affect the carbon cycle of the deep ocean in the vicinity of cold seeps on accretionary margins has been investigated [Suess et al., 1998, 1999]. The focus of this study is a section of the Cascadia Accretionary Complex located off the coast of Oregon known as “Hydrate Ridge” (see Figure 1).

Figure 1.

Bathymetry map of the North Summit of the Second Accretionary Ridge (a.k.a. Hydrate Ridge), Oregon Margin, showing CTD locations. Map shows location of Hydrate Ridge study area with respect to the Oregon coastline.

[6] In this study, we use stable carbon isotopic measurements of methane extracted from samples taken from the water column to identify the source of the methane, the mechanisms by which methane is transported into the water column, and the fate of the methane released from the seafloor.

2. Study Area: Cascadia Accretionary Complex

[7] The primary surface manifestations of the accretionary prism located at the Oregon Margin are a series of north-south trending ridges separated by a series of sedimentary basins. The Second Accretionary Ridge represents the primary focus of this study. The ridge is approximately 2–3 km wide and about 20 km long. Reflection seismology has revealed the presence of landward-dipping reflectors under the ridge; the folding and faulting patterns of these reflectors together with the current location of their surface expressions (outcrops) suggest that the ridge is undergoing shortening and uplift [Suess and Bohrmann, 1997; Trehu et al., 1995].

[8] The Second Accretionary Ridge itself may be considered as two separate summits. The North Summit (44°40′N, 125°06′W) or “Hydrate Ridge” as it is now known, is associated with extensive carbonate build-up from bacterial conversion of methane to carbonate [Kulm, 1986; Ritger et al., 1987; Suess and Whiticar, 1989], high fluid venting rates [Carson et al., 1990; Linke et al., 1994] and a prominent BSR [Suess et al., 1999; Trehu et al., 1995]. The surface expression of the thrust fault underlying Hydrate Ridge is a rock outcrop near which extensive carbonate pavements, chimneys and clusters of chemosynthetic clams have been observed.

3. Field Sampling and Methods

3.1. CTD Hydrocasts

[9] Water samples were collected with combined Conductivity-Temperature-Depth (CTD)-rosette hydrocasts made on three separate research cruises aboard the R/V Sonne to Hydrate Ridge during the summer of 1996. SO 109/1 occupied eight CTD stations over Hydrate Ridge from May 30 to June 4; SO 109/2 occupied seven CTD stations between June 22–25, and SO 110-1a occupied five CTD stations between July 9–15, 1996. The locations relative to Hydrate Ridge of CTD stations where samples were analyzed for both methane concentration and stable isotope composition are shown in Figure 1. All water samples were taken on the up-cast and sample locations in the water column were based on the presence of features in the CTD data, particularly the presence of transmissometry minima (associated with particulate maxima) observed on the downcast.

3.2. Remotely Operated Platform for Ocean Sciences (ROPOS)

[10] ROPOS is a remotely operated submersible capable of collecting water samples with a suction-sampling unit fitted to the seven-function manipulator arm and linked to a series of sealed 2-L Lexan canisters. Water samples were collected on ROPOS Dive 339 and 340 over bacterial mats, actively bubbling cold seeps and clam fields at the summit of Hydrate Ridge. Methane extracted from these water samples was analyzed onshore for stable carbon isotope composition. We expect that the degassing of methane occurred during transfer of the water to the glass sample bottles, and that this affected the eventual concentration measured in these samples. However, the kinetic fractionation factor for dissolved methane in disequilibrium with the atmosphere is small (αk = 0.9992; [Knox et al., 1992]) and we are therefore confident that the isotope values measured in waters obtained from ROPOS sampling are close to those that would have been measured had the samples been obtained via Niskin sampling. A full description of the features and functionality of ROPOS can be found at http://ropos.com.

4. Analytical Methods

[11] Methane was extracted from water samples on board the research vessel after each hydrocast using the vacuum ultrasonic technique described by Lammers and Suess [1994]. Briefly, 1.15-L sample bottles filled with seawater from the CTD-Rosette sampler were fitted to an apparatus that applied a vacuum to first create a headspace and then repeated pulses of ultrasonic energy to drive dissolved gases into the headspace. The apparatus allows for the subsequent compression of the extracted gas under ambient atmospheric pressure and a determination of the volume of gas extracted. A sample port fitted with a septum allowed for aliquots of extracted gas to be syringed and analyzed on board for methane concentration using standard flame ionization detector gas chromatography (FID-GC). Selected results of these analyses were first reported by Suess et al. [1999].

[12] For each water sample, a 10- or 15-mL glass sample vial crimp-sealed with a gray butyl septum was fitted to the vacuum extraction apparatus via a needle, in such a way that the turn of a valve could expose the vial to a vacuum or allow gas driven from the water sample to enter. A short time (30 s) was allowed for the headspace and vial system to come to isotopic equilibrium before the vial was removed from the needle. The extracted gas in the vial was stored on-shore for later δ13C-CH4 analysis.

[13] Gas samples were analyzed for δ13C-CH4 using continuous-flow isotope ratio mass-spectrometry (CF-IRMS). Briefly, aliquots of gas samples were processed using a sample purification and cryo-distillation apparatus coupled directly to the gas chromatograph, combustion interface, water removal system and isotope ratio mass spectrometer (Finnigan MAT 252, Bremen, Germany) [Merritt et al., 1995]. Analyses were completed at the Biogeochemistry Facility at the University of Victoria, Canada [Grant, 2000]. Isotope results are expressed relative to the PeeDee Belemnite standard, (in ‰ PDB), using the usual δ notation, where

equation image

The precision of the CF-IRMS analyses was ±0.4‰ PDB.

5. Results

5.1. The δ13C-CH4 and [CH4] Profiles

[14] Depth profiles of [CH4] and δ13C-CH4 in the water column are shown in Figures 2a–2b (SO109/1), 2c–2d (SO109/2), and 2e and 2f (SO110-1a). Hand-contoured two-dimensional sections of [CH4] and δ13C-CH4 based on stations SO 109/1 34-1, 45-1, 37-1 and 35-1 are shown in Figures 3a and 3b.

Figure 2.

(a, b) SO 109/1 CTD profiles of δ13C-CH4 and [CH4] with depth; (c, d) SO 109/2 CTD profiles of δ13C-CH4 and [CH4] with depth; (d, e) SO 110-1a CTD profiles of [CH4] and δ13C-CH4 with depth. Arrows indicate presence of data points that exceed the maximum on the abcissa in Figure 2e: CTD 16-1: [CH4] = 19,259 nL/L at 565 m, and CTD 12-1: [CH4] = 73,586 nL/L at 490 m.

Figure 3.

Hand-contoured east-west sections of (a) [CH4] and (b) δ13C-CH4 based on hydrocast data from stations SO109/2 34-1, 45-1, 37-1, and 35-1.

5.2. Summary

[15] When the methane concentration and carbon isotope data from all CTD-hydrocast stations occupied around Hydrate Ridge are plotted together with the ROPOS data as in Figure 4, several features of the data are apparent. The methane concentration-isotope data can be grouped into distinct zones: (1) a near-bottom zone above the summit of Hydrate Ridge (represented by the ROPOS samples) with δ13C-CH4 = approximately −63‰ and [CH4] = approximately 100,000 nL/L; (2) a near-bottom zone over the sedimentary basin to the east of Hydrate Ridge (depths of 720 to 900 m at station 35-1), with δ13C-CH4 = −40 to −50‰ and [CH4] between approximately 60 and 120 nL/L); (3) a plume zone (350 to 700 m), with δ13C-CH4 = −65 to −16‰ and [CH4] between approximately 40 and 74,000 nL/L; (4) an above-plume zone (100 to 350 m) where δ13C-CH4 = −25 to −35‰ and [CH4] = approximately 40 to 100 nL/L; and (5) a surface zone (0 to 100 m) where δ13C-CH4 = −44 to −47‰ and [CH4] = approximately 70 to 100 nL/L.

Figure 4.

Depth profiles of (a) δ13C-CH4 and (b) [CH4] from all CTD hydrocasts. Also shown are major zonations of data: deep water over the continental slope, plume depth, and mixing zone between oxidized plume methane and surface waters with atmospheric equilibration signal. Note range of δ13C-CH4 hydrate endmember in relation to most 13C-depleted (-65‰) and most 13C-enriched (−16‰) plume δ13C-CH4 values. Note logarithmic scale on [CH4] axis. Arrows indicate the presence of four data points that exceed the maximum methane concentration on the abcissa in Figure 4b: two samples from SO110/1a previously described in Figure 2e, and two ROPOS samples from Dive 340, [CH4] = 331,666 nL/L sampled over a bacterial mat, and [CH4] = 59,040 nL/L.

[16] The processes that may control the methane distribution in the water column include mixing and bacterial oxidation and these are indicated in Figure 4 together with the stable carbon isotope values of the known endmembers: methane from hydrate recovered from the seafloor, and the atmosphere. The most important results for the understanding of the transport and fate of the methane released from the seafloor (see discussion) are the presence of high methane concentrations with δ13C near −65‰ in the observed plumes, and the presence of low concentrations of highly 13C-enriched methane (δ13C = −16‰) in the upper plume horizon.

6. Discussion

6.1. The δ13C-CH4 and [CH4] Profiles

6.1.1. Identity of Source

[17] The stable isotope measurements reported above are the first for the water column above the Cascadia Oregon Accretionary Complex. Previously, the only other stable carbon isotope measurements of methane from this region are from methane recovered from hydrate and sediment gas at ODP Leg 146, Site 892 at the summit of Hydrate Ridge. Stable carbon isotope values of −65.4‰ to −69.0‰ were measured in the first aliquots obtained from sediment samples containing dissociating methane hydrate [Hovland et al., 1995]. The most 13C depleted stable carbon isotope values measured in this study ranged from −60 to −65‰ and are consistent with methane produced from bacterial methanogenesis [Claypool and Kaplan, 1974; Whiticar et al., 1986; Whiticar, 1999]. The methane is most likely produced from methanogenesis by carbonate reduction occurring in the sediment below the sulfate reduction zone (SRZ). Within the SRZ, methanogenesis by carbonate reduction or by methyl-type fermentation does not proceed to any great extent because methanogens are unable to compete for acetate or for hydrogen, which is utilized more efficiently by sulfate reducing bacteria and acetogens [Daniels et al., 1980; Alperin et al., 1992]. Methanogenesis in the SRZ could proceed via one of the non-competitive substrate pathways (e.g., mono-, di- or tri-methylamine, dimethylsulfide) but any methane produced would likely be small and quickly recycled as substrate by the sulfate reducers [Whiticar and Faber, 1986; Alperin et al., 1988]. Though methanogenesis is likely ongoing throughout the accretionary prism, at depths greater than approximately 750 m, thermogenic methane generation can also occur as a result of the geothermal gradient of 51°C/km [Hunt, 1979; Whiticar et al., 1995].

6.1.2. Coincident [CH4] Maxima/δ13C-CH4 Minima

[18] Methane produced within the accretionary prism can enter the water column dissolved in vent fluid, and via dissolution from bubbles. Bubbles form as methane is released from destabilized hydrate into free-flowing fluid with [CH4] at the saturation point (>50 mM in the first 100 m of sediment, assuming a bottom depth of 500 m and bottom water temperature of 5°C [Duan et al., 1992]). The bubbles and surrounding fluid migrate upward along faults in the accreted sediment until they are vented into the water column with release rates that appear to be controlled by tidally-induced changes in the local pressure field [Torres et al., 2002; Tryon et al., 2002]. The methane distribution in the water column is then controlled by bubble dissolution and rise rates, advection by currents, diffusion, mixing with “background” seawater and by bacterial oxidation. The observed methane concentration and isotope profile results from the particular combination of these processes at work at a given location and the time period over which they act.

[19] In some of the SO109/1 profiles, the presence of coincident methane concentration maxima and isotope minima observed suspended above the seafloor likely resulted from a release of methane on the summit of Hydrate Ridge, followed by its subsequent advection along isopycnal surfaces over the slope basin to the east and north. This is supported by the profiles in Figures 2a and 2b, and the sections in Figures 3a and 3b, in which the methane concentration maxima and carbon isotope minima occur at similar depths even as the seafloor depth increases from station 45-1 to 35-1, and northward to station 44-1. The source appears to be the summit of Hydrate Ridge, an area of known active venting [Torres et al., 2002; Tryon et al., 2002; Tryon and Brown, 2001; Suess et al., 1999]. In a similar study of the methane dynamics in the water column above the vents of the Eel River Basin, the coincident concentration maxima and isotope minima over a downslope station was consistent with the advection of dissolved methane within the depth zone of hydrate stability from an upslope, benthic source [Valentine et al., 2001].

[20] Methane maxima and isotope minima suspended above the seafloor could also result from the venting of methane bubbles from the seafloor into stably stratified system with multiple, layered mixing regimes [Radlinski and Leyk, 1995]. While the hydrography of the entire water column over Hydrate Ridge has not yet been the focus of a long term study, the mixing regime down to 300 m depth over the continental slope off the Oregon coast is known to subject to the influences of both the California Current and Poleward Undercurrent [Pierce et al., 2000; Huyer et al., 1998].

6.1.3. Evidence for Bubble Transport

[21] In some cases (e.g., SO 109/2, CTD 118-1), the δ13C-CH4 values measured in the water column show the endmember stable isotope value of the methane extracted from hydrate samples (approximately −65‰) [Hovland et al., 1995]. The similarity of the observed δ13C-CH4 to that measured in methane recovered from hydrate-bearing sediment suggests that the methane in these parts of the plume is delivered to the water column without significant loss at the sediment-water interface. The transport of methane in bubbles would effectively bypass oxidation by sulfate-reducing bacteria living both within the sediment and at the benthic interface (Figure 5). Similar cases in which ebullitive releases allow methane to escape microbial oxidation have been shown for freshwater tundra lakes and wetlands using stable isotope techniques [Chanton et al., 1992].

Figure 5.

Schematic of methane dynamics at the Second Accretionary Ridge cold seeps. Ebullition from cold seep vents allows methane to enter the water column and escape the intense oxidation activity at the vent openings. The occurrence and location of the Sulfate Reduction Zone (SRZ; 0–20 meters below the seafloor (mbsf) and methanogenic zone (>20 mbsf) are discussed extensively in relation to the results from ODP Leg 146, Hole 892 [Whiticar et al., 1995]. Thermogenic methane generation is confined to depths greater than approximately 750 m owing to the geothermal gradient at Hydrate Ridge of 51 °C/km [Hunt, 1979; Whiticar et al., 1995]. The BSR = Bottom Simulating Reflector, was observed at approximately 74 mbsf at Hole 892 [Hovland et al., 1995]. Distortions of the BSR towards the surface are intended to represent the destabilization of hydrate by warm vent fluids and its subsequent reformation in shallower sediments due to the presence of favorable conditions for the formation of H2S-CH4 hydrate [Suess et al., 1999].

[22] In contrast to the hot fluids released by hydrothermal vent systems, the fluids released from the Hydrate Ridge vent systems do not show any temperature anomaly and show only a slight freshening relative to the ambient bottom water [Suess et al., 1999]; they are therefore not buoyant to any appreciable degree. Flow-through chamber experiments conducted during these same cruises have suggested that once released to the bottom water in proximity to the vents, fluids are subject to the intense methane oxidizing activity of bacteria, either present at the benthic interface (on and in the tissues of chemosynthetic clams and in bacterial mats) or expelled together with the fluid from the vents [Suess et al., 1999]. Such intense methane oxidation would likely leave any residual methane derived exclusively from vent fluid with a 13C-enriched signature, but this signature is not associated with any of the methane concentration maxima observed in this study.

[23] Intermittent streams of methane bubbles were observed rising in streams from vent conduits and carbonate chimneys on Hydrate Ridge during SO110/1a [Suess et al., 1999]. Methane bubbles in the lower 100m of the water column at Hydrate Ridge have smaller dissolution rates because they are still in the hydrate stability field; thus they can survive to reach greater heights above the seafloor [Rehder et al., 2002b, 2001, 2000]. The presence of anomalous, high concentrations of methane in CTD 12-1 and CTD 16-1 (see Figure 2e), an order of magnitude larger than the methane concentrations determined in water sampled only 15–20 m above or below, could be explained by the capture of methane bubbles during the tripping of the Niskin bottles and their subsequent dissolution prior to recovery [Collier, 1997]. Though the physics controlling the distribution and dissolution of bubbles formed in the sediment are beyond the scope of this study, they have been and continue to be the current focus of both laboratory, computer modeling [Leifer et al., 2000] and field (e.g., TECFLUX 2000) investigations.

6.1.4. Evidence for Methane Oxidation

[24] Stable carbon isotope values as enriched as −16.3‰ were measured in the water column, together with the lowest methane concentrations (approximately 40 nL/L) (see Figures 2a and 2b, SO110-1a, CTD 37-1). These isotope values are the best evidence that methane oxidation is ongoing in the water column.

[25] At least three different processes could result in a 13C-enrichment from the hydrate-associated methane stable carbon isotope value (−65‰) in the water column. (1) There could be contributions of a 13C-enriched methane source to the water column (such as thermogenically generated methane) in a water mass located above the bacterial methane observed in proximity to the seeps, either from the seeps themselves or in waters advected from the continental shelf. (2) The 13C-depleted methane dissolving from bubbles or in vent fluid released from the vent field could mix with 13C-enriched “background” methane in the water column. (3) Populations of methanotrophic bacteria living at the sediment-water interface, or within the water column, could oxidize the methane released into the water column, preferentially utilizing 12CH4, and leaving the remaining pool 13CH4 enriched. While the data do not rule out the influence of any of the three processes listed above, only the third (i.e., oxidation of bacterial or thermogenic methane) can account for the observations of δ13C-CH4 values > −20‰ in the water column.

[26] The existence of thermogenic hydrocarbons in sediments on the Second Accretionary Ridge has been inferred from the stable carbon isotopic analysis of gas samples taken from expansion void gases in a sediment core at Ocean Drilling Program Site 892 [Whiticar et al., 1995]. Observations of admixtures of bacterial and thermogenic gas from ODP site 892 sediment cores suggest that it is highly unlikely that the methane transported to the sediment-water interface at Hydrate Ridge would reflect only the thermogenic endmember. However, the possibility that there exist other seep sites nearby that access thermogenic methane reservoirs from deeper within the accretionary prism cannot be ruled out by these data.

[27] Another possible thermogenic methane source in these off-shelf waters are hydrocarbon seeps located on the continental shelf to the east of Hydrate Ridge [Rehder et al., 2002a, 1999a]. Water column sampling along onshore to offshore transects has been recently conducted to determine the possible influence of the advection of methane from these on-shelf sources on the carbon isotope profile in the waters above Hydrate Ridge. The observed range of δ13C-CH4 values for thermogenically generated methane is generally accepted to be from −50 to −20‰ [Whiticar, 1999], and δ13C-CH4 values as 13C-enriched as −16‰ are observed in the water column above Hydrate Ridge. However, advected thermogenic methane from shelf seeps that is then oxidized would also result in such enriched δ13C-CH4 values. The coexistence of deep and shallow methane plumes reflecting separate biogenic and thermogenic endmember sources has been postulated previously to explain the observed carbon isotope distribution of methane in the Suruga Trough, Japan [Tsunogai et al., 1998].

[28] The degree to which the observed methane distribution results from mixing with ambient seawater is difficult to determine because the “background” methane concentration and isotope composition are unknown. We select the depth profiles of [CH4] and δ13C-CH4 from CTD 29-1 as an approximate background profile, as this station was located over the First Accretionary Ridge where no active vents were observed. The most notable feature of this profile (not shown) is the presence of a single, small methane concentration maximum (most likely resulting from the influence of terrigenous slope sediments) of approximately 60 nL/L at 2300 m depth, with δ13C-CH4 = −54‰. Above the depth of the upper [CH4] maximum (1600 m), the δ13C-CH4 values lie consistently in the −46 to −47‰ range, despite the fact that the methane concentration in these samples drops to near 20 nL/L. Stable carbon isotope values as 13C-enriched as −16‰ cannot be produced from the linear mixing of methane of −65‰ to the range of background water column δ13C-CH4 values discussed above.

6.1.5. Locus of Methane Oxidation

[29] The oxidation of methane at the benthic interface as the seep fluid enters the water column can be inferred from the presence of extensive carbonate pavements [Ritger et al., 1987; Suess and Whiticar, 1989; Kulm and Suess, 1990; Paull et al., 1992] as well as from time series of methane concentrations, dissolved oxygen and sulfate in closed chambers placed over vent sites [Linke et al., 1994; Suess et al., 1999].

[30] The consumption of free methane gas released as bubbles in vent fluid would not occur at the sediment-water interface as does the oxidation of methane dissolved in vent fluid [Suess et al., 1999] as methanotrophs present in the vent fluid or at the benthic interface would not have access to methane in free gas form. The bacterial population in the water column could respond, however, to increasing concentrations of methane contributed from either bubbles or vent fluid.

[31] The presence of methane-oxidizing bacteria in the deep-sea has been inferred from studies of methane and oxygen concentrations in waters of increasing age in the Deep-Water Conveyor [Scranton and Brewer, 1978] as well as from studies of the changing ratio of methane to CFC-11 in newly formed deep water [Rehder et al., 1999b]. In the Bering Sea, 14CH4 incubation experiments revealed a bacterial population in deep, off-shelf waters that converted nearly 90% of the utilized methane carbon into carbon dioxide rather than cell carbon [Griffiths et al., 1982].

[32] Microbial oxidation activity has been determined in the water column within the methane plume above the Second Accretionary Ridge by M. de Angelis using 14CH4 incubations (M. de Angelis, personal communication, 2000). Results from incubation studies conducted during the summer of 1999 on samples obtained from CTD hydrocasts over the Hydrate Ridge support the hypothesis that most of the methane released to the water column is consumed in the bulk water column below the thermocline, rather than in near-bottom waters in close proximity to the seep sites. The low methane oxidation rates observed in near-bottom waters are both consistent with the rapid drop in methane concentrations observed between the seep fluid and near-bottom waters, the introduction of methane to the water column by bubble dissolution, and with the continual advection of water column methane oxidizers away from the vent site (M. deAngelis, personal communication, 2000).

6.2. Methane Oxidation: Rayleigh Distillation

[33] [CH4] and δ13C-CH4 from each hydrocast are plotted against each other for each cruise in Figure 6. The general trend in the plot is that larger methane concentrations are associated with the most 13C-depleted methane, and smaller concentrations of methane with the more 13C-enriched methane. A Rayleigh distillation model of the type discussed by Coleman et al. [1981], i.e.,

equation image

can be used to determine the fraction f of the methane remaining (1-f is therefore the methane plume consumed by oxidation), assuming the kinetic isotope fractionation factor, α, and the starting stable isotopic composition of the plume, (δ13C-CH4)o, are known. Alternatively, the starting isotopic composition and methane concentration of the plume water may be assumed and kinetic fractionation factors calculated for aerobic methane oxidation using stations receiving water advected from the starting location. The closed system Rayleigh model in equation (1) assumes that the methane reservoir has one sink (bacterial oxidation), no inputs and that the observed isotopic composition is not affected by mixing. Though this is an oversimplification of the case of a methane plume in the water column above the Hydrate Ridge, the version of the Rayleigh model represented by equation (1) is a first attempt to understand the methane dynamics in the water column. In addition, a methane plume suspended above the seafloor could approximate a closed isotopic system if methane contributions from the seafloor (from bubbles or vent fluid) were interrupted by tidal forcing or redirection by bottom currents.

Figure 6.

Scatterplot showing [CH4] versus δ13C-CH4 for CTD data from all cruises.

[34] Below, we apply a Rayleigh modeling approach to the CTD hydrocast data from SO 109/1 located in the methane plume (data between 350 and 680 m in depth) in order to establish lower bounds on and the kinetic fractionation factor associated with aerobic methane oxidation and the fraction of methane oxidized in the water column. We chose this subset of the data because of the presence of CTD stations located directly over (CTD 45-1), and receiving advected methane from, the summit source vents (CTD's 34-1, 35-1, 37-1 and 44-1). For both calculations, a value for the initial stable carbon isotope composition of −65‰ was assumed, based on the measurements of methane extracted from hydrate-bearing sediment at ODP site 892 by Hovland et al. [1995].

6.2.1. Fractionation Factors

[35] Using a similar approach to that of Sansone et al. [1999], we compared the methane concentration and isotopic composition at the methane maximum over station 44-1 to all other SO 109/1 hydrocast data within the plume zone (as defined above), and solved for α using equation (1). We obtained kinetic isotope fractionation factors of 1.002 to 1.013 (mean = 1.008). These values fall in the lower end of the ranges calculated for the aerobic oxidation of methane from laboratory and field studies of methanotrophs (α = 1.003 to 1.035; see Table 1) but are closer to the kinetic fractionation factors determined in other marine and estuarine systems [Tsunogai et al., 2000; Sansone et al., 1999; Wen et al., 1996].

Table 1. Kinetic Fractionation Factors for the Aerobic Oxidation of Methane Derived in Field and Laboratory Settings
αSettingSource
1.005 to 1.031laboratory cultureBarker and Fritz [1981]
1.013 to 1.025laboratory cultureColeman et al. [1981]
1.022 to 1.025boreal forest soilReeburgh et al. [1997]
1.0121 to 1.0183temperate grassland/forestSnover and Quay [2000]
1.003 to 1.021freshwater swampHappell et al. [1994]
1.0042 to 1.012estuarineSansone et al. [1999]
1.006 to 1.035marine: subtropical gyreSansone et al. [2001]
1.004 to 1.006marine: hydrothermal plumeTsunogai et al. [2000]
1.008marine: hydrothermal plumeWen et al. [1996]
1.002 to 1.013marine: cold seep plumethis study

[36] The fractionation factors in this cold marine environment could be smaller than the highest fractionation factors observed in laboratory cultures of methanotrophs (1.025 at T = 26°C [Coleman et al., 1981]) and forest soils due to the inverse relationship between temperature and kinetic fractionation factor [Tsunogai et al., 2000; Whiticar, 1999]. However, mixing likely plays a large role in the evolving plume on the timescale of methane oxidation at this site and causes fractionation factors calculated using a Rayleigh model to be depressed from the values that would be obtained from a true, closed isotopic system.

[37] As an illustration of the potential contribution of mixing to the observed methane distribution, we approximate the isotope mass balance of methane in a water parcel with background methane concentration with a linear mixing equation as follows:

equation image

where Cox, Cbkgd, and Cmeas are the concentrations of oxidized methane originally contributed from the seafloor, background methane and that measured in the water parcel, respectively, δox, δbkgd, and δmeas are the carbon isotope compositions of the oxidized methane, background methane and that measured for the overall parcel, respectively and x is some real number that takes on values between 0 and 1. Assume the background methane concentration and isotope composition of methane in the water column are 70 nL/L and −47‰, respectively (i.e., roughly that from atmospheric equilibration). The Rayleigh distillation model can then be used to calculate the carbon isotope value that would be paired with a methane concentration of 70 nL/L if the methane distribution were 100% controlled by oxidation. Departures from the isotope value predicted from the Rayleigh model can then be considered as the result of the mixing of the water parcel with background water; an estimate of contribution of mixing can be obtained by solving for the fraction of background water (1-x in equation (2) above) that must be mixed with the oxidized seafloor-contributed methane to produce a water parcel with the observed concentration and isotope composition. Since Cox = Cbkgd = Cmeas = 70 nL/L in such a case, equation 2 can be re-written to solve for x as

equation image

[38] Data points where the methane concentrations are approximately equal (i.e. within ±2 nL/L) to the assumed background (70 nL/L) are presented in Table 2 together with their associated δ13C-CH4 values.

Table 2. Estimate of Contribution of Mixing of Background Water ([CH4] = 70 nL/L, δbkgd = −47‰) to Water Parcels With Near Background Methane Concentrations
CTDRegionDepth, m[CH4],nL/Lδ13C-CH4,‰ABC
δox,a(1-x)bδox, ‰(1-x)δox, ‰(1-x)
  • a

    Expected carbon isotope value from oxidation alone, calculated using a Rayleigh distillation equation with α = 1.02, (δ13C-CH4)o = −65‰, initial methane concentration in parcel from seafloor source ([CH4]o) as described in A and B below. See equation (2) and section 6.2.1.

  • b

    Contribution of mixing with background water calculated using equation (3). A: with [CH4]o = 1,200 nL/L; B: with [CH4]o = 10,000 nL/L, all other parameters as in A. C: δbkgd = −42‰, all other parameters as in A.

35-1deep88572−42.89−9.840.8931.70.95−9.841.0
45-1above-plume32167−29.3−8.420.5433.10.78−8.420.62
6-1plume47572−18.7−9.840.2431.70.64−9.840.28
10-1plume39969−38.1−9.000.7632.60.89−9.000.88

[39] Carbon isotope values computed from the Rayleigh model (with α = 1.02, (δ13C-CH4)o = −65‰ and initial [CH4] equal to 1,200 nL/L (scenario A) or 10,000 nL/L (scenario B) are presented together with values for 1-x calculated using equation 3 in the adjacent columns. Calculated this way, mixing with background water accounts for between 24% (at plume depth, 475 m, CTD 6-1) to 90% (885 m depth; CTD 35-1) of the methane observed in these parcels. These numbers rise to 64 to 95% when a tenfold higher starting methane concentration is assumed (scenario B). If a background carbon isotope composition of −42‰ is assumed, with all other parameters as in scenario A, mixing is calculated to account for between 28% and 100% of the methane composition of a water parcel with background methane concentration (scenario C).

[40] While these calculations serve as a useful illustration of the effect of mixing on the isotopic composition of water parcels, they oversimplify the actual processes taking place in the water column because mixing and oxidation occur simultaneously, rather than as the discrete, sequential events represented in the mixing model. In reality, the observed methane distribution is affected by contributions from background methane that has been both mixed into the water parcel and oxidized. A rigorous calculation of the extent to which the observed methane distribution above Hydrate Ridge is the result of mixing would require normalization methane concentrations to a conservative tracer, such as 3He.

6.2.2. Fraction of Methane Oxidized

[41] Using CTD hydrocast data from SO109/1 and equation (1) to solve for the fraction of methane oxidized, 1-f. Assuming α = 1.005, 1-f ranges from 0.58 to near 1.0; with α = 1.025, 1-f varies from 0.16 to 0.87. The sensitivity of this calculation to the α value assumed is shown in Figure 7, where δ13C-CH4 values from SO109/1 hydrocasts are plotted against the residual methane, f, calculated from equation (1) using three different α values. This result suggests that bacteria are highly effective in removing methane from the water column.

Figure 7.

Fraction of methane reservoir remaining calculated using a Rayleigh distillation model and SO109/1 δ13C-CH4 data showing sensitivity of calculation to α value assumed. Plots are for α values of 1.005 (circles), 1.010 (squares) and 1.025 (triangles). Fraction of methane oxidized is 1-f.

[42] Other studies have also shown that aerobic oxidation in the water column represents a significant sink for seepages of methane. In a study of a methane plume in the Barents Sea, methane concentrations from CTD profiles and the diffusive fluxes they implied were used to determine that approximately 98% of the methane released from a crater field or “pockmark” at 300 m depth appeared to be oxidized before reaching surface waters [Lammers and Suess, 1995]. A recent study of methane oxidation in the water column over a region of actively dissociating methane hydrates used a similar closed-system Rayleigh model approach to show that about 45% of the methane was consumed in the water column [Valentine et al., 2001]. However, as discussed in relation to the data from this study, the value obtained from closed system model calculations likely underestimates the true fraction of methane oxidized in the water column due to the contribution of mixing to the most 13C-enriched values observed. Additionally, a kinetic fractionation factor of 1.025 was used in the calculation, a value at the high end of the range of kinetic fractionation factors reported for the aerobic oxidation of methane in marine systems (section 6.2.1 and Table 1). Using a fractionation factor more reflective of this range, such as 1.015, yields an oxidized fraction of 64% for the same isotope data, before mixing is taken into account.

[43] The question of whether the methane released into the water column from Hydrate Ridge can be brought to the surface to exchange with the atmosphere while still oversaturated with respect to equilibrated surface waters is obviously of prime concern. The issue of wind-driven upwelling of methane from shelf and slope sources off the Oregon coast has been the subject of a recent study [Rehder et al., 2002a, 1999a]. Though summer upwelling appears to drive the appearance of regions of the surface ocean supersaturated in methane above shelf seepages, the upwelling waters appear to be drawn from 100–300 m depth [Barber and Smith, 1981]. As the Hydrate Ridge methane plumes have so far been observed between 400 and 500 m depth (this study and work by Suess et al. [1999]), it remains to be demonstrated that methane released into the water column from the Hydrate Ridge seeps is upwelled to the surface.

7. Conclusions

[44] The analysis of the stable carbon isotope composition of methane has added a significant dimension to the study of methane biogeochemistry at the Cascadia Accretionary Complex. In particular, the analysis above has been able to address questions of the identity of the methane source, the mechanism of methane release, the activity of methane-oxidizing bacteria in the water column, and the capacity of methane oxidizers to consume methane.

[45] The methane observed in the water column above the Hydrate Ridge is bacterial in origin, with measured δ13C-CH4 values as 13C-depleted as −65.4‰. The presence of methane in the water column with stable carbon isotope values so close to that of methane extracted from hydrate samples is strong evidence that the observed methane plume is derived from bubbles formed from a common source of biogenic methane produced from carbonate reduction in accreted sediments.

[46] The trend in the CTD-hydrocast data towards more 13C-enriched δ13C-CH4 values with low methane concentrations, and the observation of δ13C-CH4 values as 13C-enriched as −16.3‰ are evidence for methane oxidation within the water column itself. Rayleigh models fitted to the [CH4] and δ13C-CH4 data from SO109/1 yield a kinetic fractionation factors for the bacterial oxidation of methane of 1.002 to 1.013 (mean 1.008). These values fall at the low end of the range of previously reported fractionation factors from laboratory and field experiments (α = 1.004 to 1.035), a result consistent with the significant contribution of mixing to the methane distribution observed. Despite this fact, a closed system Rayleigh model approach has shown that methane oxidation in the water column is highly effective at removing methane from the water column. Methane introduced to the water column appears to be consumed nearly quantitatively, a result that is consistent with the findings of 14CH4 incubation work. Though significant evidence exists to suggest that the upwelling of methane from bottom waters over the shelf and upper slope off the Oregon coast occurs, it remains an open question whether wind-driven upwelling processes can draw methane in waters from 400–500 m depth over Hydrate Ridge to exchange with the atmosphere.

[47] Estimates of the fraction of methane released into the water column that could survive bacterial oxidation to exchange with the atmosphere therefore depend on a knowledge of the fraction of methane entering the water column as bubbles. Closed sample chamber techniques cannot discern methane released in bubbles from that dissolved in vent fluid. Direct collection of bubble-gas, longer-term video monitoring of vents, and more extensive hydrocast profiling of the methane plumes must be conducted in order to estimate the volumes of methane gas released. Improved estimates of the flux of methane in bubbles would represent valuable input for computer model scenarios of seafloor methane releases and their potential influence on, and response to past and present climates.

Acknowledgments

[48] The authors would like to thank Stephan Lammers, Evelyn Zuegler, and Nicola Jones for their collection of the samples, and the captain and the crew of the R/V Sonne for their continuing dedication to marine geochemical work. This manuscript benefited greatly from helpful discussions with Bob Collier, Gregor Rehder, Marie deAngelis, Magnus Eek, and Robie MacDonald, and from reviews by David Valentine and one anonymous reviewer. Paul Eby and Magnus Eek at the Biogeochemistry Facility, University of Victoria, provided invaluable assistance and advice with the CF-IRMS methodology. The map was drawn using Online Mapping Tools for GMT (http://gmt.soest.hawaii.edu) at http://www.aquarius.geomar.de/omc. These investigations were partially supported by the Federal Ministry of Education, Research and Technology, Bonn, Germany (grants 03G0110A/B and 03G0109A/C).

Ancillary