Previous   |   Next

doi:10.2204/iodp.proc.311.213.2010

Gas hydrate formation—from in situ methane production or deeper methane sources?

A general model for deep-sea gas hydrate formation by removal of methane from upwardly expelled fluids was developed by Hyndman and Davis (1992). In this model, mostly microbial methane produced over a thick sediment section pervasively migrates upward to form gas hydrate as it enters the stability zone. The gas hydrate concentration is therein predicted to be greatest just above the BSR associated with the BGHSZ.

With the five sites established during Expedition 311 and combinations of observations made during Leg 204, there are now several data sets available to help modify and further develop the pore fluid expulsion model. These data sets include

• δ¹³C isotopic composition of methane and carbon dioxide gases from sediment core samples and gas from void spaces in the recovered cores (Pohlman et al., 2009), gas analyses from pressure core samples, and gas hydrate samples;

• δ¹³C isotope composition of the dissolved inorganic carbon (DIC) (Torres and Kastner);

• Pore fluid chlorinity and related fluid advection modeling (see the "Expedition 311 summary" chapter; Malinverno et al., 2008; Wortmann et al. 2008);

• Strontium (Sr) and lithium (Li) components of the pore water following earlier results from Legs 146 and 204 (Teichert et al., 2005; Kastner et al., 1995b); and

• Iodine, bromine, and ammonium pore fluid constituents (Lu et al., 2008).

Inorganic pore water constituents and upward fluid migration constraints

Downhole trends in inorganic pore water constituents (e.g., chlorinity) suggest a strong component of pore water migration from below with distinctly different source compositions (Fig. F6). This inferred upward pore water migration is a critical component of the dewatering process in the Cascadia accretionary prism, where the incoming ascending sediment section is deformed and squeezed by the tectonic processes that form the accretionary wedge. Modeling conducted by Malinverno et al. (2008) assumed relatively low advection rates (a maximum of 0.017 cm/y was inferred), which is about an order of magnitude lower than rates used in previous studies (Bekins and Dreiss, 1992; Wang et al., 1993; Hyndman and Davis, 1992). Fluid advection modeling is also strongly dependent on accurate sedimentation rates. Sedimentation rates at the Expedition 311 transect sites were estimated from diatom biostratigraphy (see the "Expedition 311 summary" chapter; Akiba et al.) and decrease markedly from Sites U1325 and U1326 (>400 m/m.y.) to Site U1329 (<100 m/m.y.). Sedimentation rates in the Cascadia Basin seaward of the deformation front are even higher, reaching ~1000 m/m.y. at Site 888, located ~60 km south of the Expedition 311 transect (Westbrook, Carson, Musgrave, et al., 1994). However, note that Site U1329 is dominated by a marked unconformity at ~136 mbsf, where ~5 m.y. of sediments were eroded (a jump from 1.6 to 6.4 Ma was observed across the unconformity; Akiba et al.).

The presence of an advecting pore fluid from greater depth is also required to explain other observed inorganic pore water constituents, including iodine, bromine, and ammonia (Lu et al., 2008). The advection rates calculated by Lu et al. (2008) have very similar values to those modeled by Malinverno et al. (2008) and range from 0.015 cm/y at Site U1325 to 0.06 cm/y at Site U1326. All other sites have values of ~0.03 cm/y. Additional modeling to achieve a better fit between the predicted and observed depth profiles of pore water halogen and ammonia constituents was also conducted by Lu et al. (2008) by incorporating lateral advection through fractures/faults.

Sites U1325 and U1326 show increasing pore water chlorinity/salinity with depth, whereas Sites U1327, U1328, and U1329 show (to various degrees) pore water freshening with depth. The observed increase in pore water chlorinity/salinity is attributed to diagenetic processes (e.g., through low-temperature reactions where volcanic ash is transformed to zeolite, releasing salt). Given that no information exists about the exact type zeolite involved in the actual reaction at depth and the possibilities of having generated zeolites from very low temperatures (as low as 5°C in the case of phillipsite), it is difficult to determine an exact depth range for this reaction. For example, diagenetic alteration of ash to zeolite was reported to occur from 200 m to 11 km below seafloor by Iijima and Utada (1966), but it likely occurs as soon as sediment is deposited (i.e., diagenetic alteration and related salt production happens throughout the entire sediment column).

The pore water freshening with depth was expected to be a margin-wide phenomenon and simply a function of distance from the deformation front based on results from Leg 204 (Kastner et al., 1995a; Torres et al., 2004). The source of freshwater was attributed to low-temperature clay dehydration processes (e.g., smectite-to-illite transformation) that typically occur over a temperature range of 60°–160°C (e.g., Bekins et al., 1994). With a temperature gradient of ~60°C/km at Site U1327, the source of this freshwater pool is at a depth of ~1000–2750 mbsf.

The clear separation between salt and freshwater sources along the margin remains a puzzle because the two assumed diagenetic reactions could occur over the same depth range. Furthermore, the distribution of volcanic ash and smectite should not be drastically different across the accretionary prism. Bartier et al. indicate only modest variation in the clay mineralogy among all of the sites, and ash layers have been reported at all sites, including Site 888 (Westbrook, Carson, Musgrave, et al., 1994; see the "Expedition 311 summary" chapter).

However, the apparent complex segregation of the margin from more saline to fresher formation waters can be explained by a relatively simple model that incorporates the change in fluid expulsion rates along the margin, as modeled by Hyndman and Davis (1992) (Fig. F8). The expulsion rate is expected to be at maximum ~15 km landward of the deformation front, and little fluid is expelled along the transect until a distance of ~10 km. The pore water freshening reported in Figure F8 for Leg 204 and Expedition 311 sites was determined at a constant depth of 200 mbsf and is all relative to seawater. Pore water freshening is seen only at sites ~15 km east of the deformation front, and the freshening quickly increases landward. The slowdown of the modeled expulsion rate is reflected in the decrease in freshening between Sites U1327 (Site 889) and U1329; however, the erosion and associated drastic difference in sedimentology at Site U1329 may also be a factor.

Furthermore, at Sites U1325 and U1326, little water has been expelled from depth, and, because the source of any freshwater must come from greater depth (~1000 mbsf), there is little mixing of the pore waters whose salt contents increased from the ash-to-zeolite transformation. As soon as freshwater is expelled in large quantities, freshening overprints salt formation and the pore waters show an overall reduction in chlorinity. Additionally, the accretionary prism is relatively thin near the deformation front, which reduces the possible region for freshwater generation (which requires higher temperatures—i.e., a thicker sediment column). Thus, there is a relatively lesser amount of source material beneath Sites U1325 and U1326 compared to the more mature, thicker prism farther east beneath Site U1327 (Site 889) and others.

Additional constraints on the deep fluid sources come from lithium and strontium data (e.g., Teichert et al., 2005; Kastner et al., 1995b). Here we use the measured concentration-depth profiles of lithium and strontium as well as ⁸⁷Sr/⁸⁶Sr ratios (Fig. F9) to discern types of fluid sources that may have contributed to the pore water sampled along the Expedition 311 transect (M. Kastner and M. Torres, unpubl. data).

The mobility of lithium is temperature dependent. At lower temperatures, lithium is partitioned into clay minerals, whereas it is leached into the pore fluid at temperatures >70°C. At the thermal gradients measured during Expedition 311 (~60°C/km), inferred in situ lithium leaching mainly occurs at depths >1 km (see also Kastner et al., 1995b). Thus, higher-than-seawater lithium concentrations found at shallower depths are indicative of an upward-migrating deep fluid source. Pore water lithium concentrations increase with depth at all sites (especially Sites U1327 and U1329) and with distance away from the deformation front along the transect (Fig. F9A). A shift to elevated lithium concentrations even at shallow depths of <50 mbsf was observed at Sites U1327 and U1329, which is in very good agreement with the previously stated evolution of fluid expulsion rates along the drilling transect. However, the cold vent Site U1328 surprisingly does not show much evidence for equivalent deeper fluid advection, although it is only 3.5 km south and about the same distance from the deformation front as Site U1327.

From strontium concentration and isotopic ratio data (Fig. F9B–F9C) it is apparent that the values at Site U1329 below the unconformity are the only anomalous values found along the Expedition 311 transect. It is also evident that the entire set of values from all sites is overprinted by diagenesis from carbonate precipitation (Fig. F9D), especially for Site U1327—a finding that was already noted by Kastner et al., (1995b) for nearby Site 889. Similar observations were made by Teichert et al. (2005) for data from Leg 204. Teichert et al. (2005) also proposed a mixing line (dotted black line in Fig. F9D) for the strontium isotope and concentration data for the southern Cascadia margin using seawater and bottom fluids collected from ODP Sites 1027 and 1026, similar to the previous mixing line proposed by Kastner et al. (1995b).

The strontium concentrations and isotopic ratios at all sites from all previous drilling (including Expedition 311, with the exception of Site U1329), can easily be explained by the mixing of seawater and the deep fluid source measured at Sites 1026 and 1027. The apparent end-member inferred for Site U1329 could be different (but basaltic in nature) from the incoming plate (M. Torres, pers. comm., 2008). On the other hand, the observation based on Leg 146 that the sites drilled from Northern Hydrate Ridge to offshore Vancouver Island communicate with the same deep fluid is striking (Kastner et al., 1995b), and it would make sense if Site U1329 also communicates with the same deep fluid. Thus, the observed local deviations at Site U1329 may indicate some mixing with pore fluids influenced by in situ reactions with certain more radiogenic detrital phases, though this is not necessarily the case for regional variations. This is further emphasized by the fact that the anomalous values were obtained only for samples from depths below the unconformity at Site U1329, where sediments older than 6.0 Ma were recovered, in contrast to all other sediments of mainly younger than 1.0 Ma age (Akiba et al.).

Methane production from CO₂ reduction and dissolved inorganic carbon isotopic composition

The δ¹³C values of methane range from a minimum of –82.2‰ near the deformation front at Site U1326 to a maximum of –39.5‰ at the most landward location Site U1329 (Fig. F10A). The δ¹³C enrichment of methane has classically been interpreted as evidence for the transition from microbial to thermogenic methane. However, the associated CO₂ sampled during Expedition 311 exhibits a similar trend of δ¹³C enrichment, with values ranging from –22.5‰ to +25.7‰ (Fig. F10A). The magnitude of the carbon isotope separation between methane and CO₂ is consistent with the kinetic isotope effect (KIE) that occurs during microbially mediated carbonate reduction. Furthermore, the gaseous hydrocarbon content is composed of >99.8% (by volume) methane and the methane has uniform δD_CH4 values (–172‰ ± 8‰) that are also consistent with carbonate reduction. Little evidence was found of any thermogenic gas source along the transect, which is distinct from the thermogenic methane signatures seen at Barkley Canyon located ~60 km southeast of the Expedition 311 transect sites (Pohlman et al., 2005). These combined results suggest that microbial CO₂ reduction is the predominant source of the methane. The increase in δ¹³C with sediment depth at each site is a closed-system KIE from preferential consumption of ¹²CO₂ during methanogenesis, which drives the residual CO₂ and accumulated methane toward more enriched ¹³C values. Similar observations were made at Blake Ridge (Leg 164; Paull et al., 2000) and SHR (Leg 204; Claypool et al., 2006). The trend of ¹³C enrichment in DIC (Torres and Kastner) is similar to that of CO₂, with the difference being related to equilibrium isotope effects that occur during the degassing of CO₂ from the dissolved phase (Emrich et al., 1970). The δ¹³C-DIC values are near –5‰ at Cascadia Basin Site 888 and near +32‰ for Site U1329 (Fig. F10B).

An additional explanation for the downcore increase in δ¹³C composition is the possibility of acetoclastic methanogenesis at Sites U1327 and U1329 (Heuer et al., 2007, 2009). Most often, acetate concentrations increase in the sediments from <5 µM at the sediment/water interface to a maximum of ~100 µM at 250 mbsf. However, Site U1328 deviates from this relationship, where maximum values reach almost 800 µM at 250 mbsf. The δ¹³C values of acetate range from –48‰ to –8‰. However, the largest enrichment of δ¹³C acetate (about –10‰) occurs over the depth range from 130 to 220 mbsf at Site U1327 (Heuer et al., 2007, 2009).

As noted by Pohlman et al. (2009), although the CO₂ and methane profiles are influenced by closed-system KIE, additional sources and sinks also affect the mass balance of CO₂ and methane. A completely closed system would result in δ¹³C-CO₂ values that continually increase with depth, whereas the δ¹³C-CO₂ profiles at each site approach constant values with increasing depth. Constant δ¹³C values with increasing depth for CO₂ and methane may be explained by mixing with a deeper isotopically uniform gas reservoir (Paull et al., 2000) or contributions from organic matter fermentation that balance losses to carbonate reduction (Claypool et al., 1985).

As mentioned before, the carbon isotopic composition of the residual DIC shows a general progressive enrichment with distance from the deformation front as well as downcore for all sites along the Expedition 311 transect, which can be explained by a KIE from preferential consumption of the lighter ¹²CO₂ during methanogenesis. However, Site U1326 shows a distinctively different depth profile compared to the other sites in that there are two intervals where δ¹³C values are ¹³C depleted (Fig. F11). The upper anomalous interval (minimum δ¹³C is around –4‰ Peedee belemnite [PDB]) extends from ~40 to ~140 mbsf, with the top of the gas hydrate reported at ~47 mbsf. The maximum δ¹³C value of +11‰ PDB is reached at ~155 mbsf before values decrease again to around +4‰ PDB. This distinct pattern toward lighter carbon isotopic composition is a possible indicator of the influx of nondepleted pore water from depth.

Regional in situ methane production and vent-associated gas migration

The δ¹³C composition of the methane recovered from gas hydrate samples from 44 and 53 mbsf at Site U1326 is identical to the composition of core void and sediment methane gas, which suggests that methane in the gas hydrate samples was formed in situ near the site of gas hydrate nucleation and was not derived from a deeper source that migrated into the GHSZ (Pohlman et al., 2006). This contrasts with observations made at the cold vent at Site U1328. The cold vent is characterized by a cap of gas hydrate–rich sediments that extends to ~40 mbsf (see the "Site U1328" chapter). The depth profiles of δ¹³C composition of core void and sediment methane gas at Site U1328 are similar to those seen at the other sites along the transect. Values gradually decrease from around –62‰ PDB at 300 mbsf to about –72‰ PDB at the seafloor. However, within the uppermost 50 mbsf there is a secondary trend of increasing δ¹³C composition to values similar to those observed at the bottom of the hole. Gas hydrate samples recovered within the depth interval where the core gas methane was ¹³C enriched have similar values, suggesting that the gas hydrate–bound gas and dissolved gas in that interval are derived from a deep-migrated source (Pohlman et al., 2006).

The main source of methane in gas hydrates sampled during Expedition 311 is CO₂ reduction, which raises a key question: What are the production rates? Several microbiological experiments were conducted to determine the rate of methane production from Expedition 311 sediments (Yoshioka et al., 2010) and Leg 204 (Colwell et al., 2008). However, accurate in situ microbial methane production rates are difficult to determine using incubation experiments, which often result in unrealistically high production rates (Colwell et al., 2008). Nevertheless, methane production rates at Sites U1327 and U1329 were determined to be on average ~1 pmol/cm³/day, with highly variable rates with depth (Yoshioka et al., 2010). The value of ~1 pmol/cm³/day is comparable to values obtained by Colwell et al. (2008), who reported a rate of 0.017 fmol/cell/day using an average cell count of ~1000 cells/g and a density of ~1.7 g/cm³. Another independent estimate of methane production can be obtained from the progressive enrichment of the δ¹³C composition of residual DIC (Torres et al., 2007) which yields results similar to estimates made by Colwell et al. (2008).

Yoshioka et al. (2010) determined that methane production at Site U1327 is higher within the depth interval where gas hydrate is inferred to occur than in the near-surface, gas hydrate–free sediments or the sediments below the BSR. They also show that at Site U1327 methanogens are more abundant in the gas hydrate–bearing sediments than in sediments at other depths. In contrast, at Site U1329 there is no evidence for a substantial amount of gas hydrate in the sediments. Methane production rates at this site are also relatively high in the sediments between 70 and 140 mbsf (comparable to those at Site U1327). This suggests that the enhancement of methane production rates in the hydrate-bearing interval at Site U1327 is likely unrelated to the actual occurrence and distribution of gas hydrate but instead may be related to the location of the site within the accretionary prism or sediment composition (high methane production rates were only seen in accreted sediments below 100 mbsf).

Top of page   |   Previous   |   Next