13CCH4 of –65
to –77) are consistent with an in situ microbial source for the methane (Fig. F4C). The gas hydrate at these sites is, therefore, referred to as reaction dominated (Claypool et al., this volume).
Estimates of average gas hydrate content (Milkov et al., 2003) mask the great heterogeneity in gas hydrate distribution at SHR (Tréhu et al., 2004b). This heterogeneity results from a variety of processes. Foremost is that two fundamentally different regimes operate to generate the gas hydrate deposits. Both regimes may operate simultaneously in any given locality. Throughout the region, some of the gas hydrate forms from methane generated in the upper few 100 m of sediment by microbial remobilization of organic carbon. Superimposed on this reaction-dominated regime is a transport-dominated regime in which methane is focused from a large volume of deeply buried sediment into a conduit that feeds methane into the GHSZ to form rich gas hydrated deposits near the structural summit. These two regimes correspond to the more general distributed low-flux and focused high-flux regimes discussed by Tréhu et al. (in press).
Modulating the gas hydrate distribution, especially within the reaction-dominated regime, is the effect of lithology. Gas hydrate forms preferentially in relatively coarse grained sediments (Weinberger et al., 2005; Su et al., this volume; Grŕcia et al., this volume), leading to large vertical variations in gas hydrate on the scale of centimeters. Borehole logging data have been used to further characterize azimuthal variability, which can range from 0% to 90% of the pore space at a given depth (Janik et al., 2003).
The methane in the cored sediments recovered from northern flank of the summit (Sites 1244–1247) and from the slope basin (Sites 1251 and 1252), was generated by microbial methanogenesis (Claypool et al., this volume). At these sites, the relative rates of alkalinity production and removal are such that the onset of methanogenesis is clearly observed as a change in slope in the alkalinity profiles (Tréhu, Bohrmann, Rack, Torres, et al., 2003).
Measured alkalinity at any depth beneath the zone of sulfate reduction results from total dissolved inorganic carbon (DIC) addition by organic matter oxidation minus DIC removal as methane and authigenic carbonate. Alkalinity production due to organic matter oxidation can be calculated and extrapolated from the observed sulfate gradient (Borowski, this volume;
Claypool et al., this volume). The difference between the projected and observed alkalinity is a measure of the amount of DIC removed to form methane and authigenic carbonate (Fig. F4A). Methane production rates estimated with this approach are on the order of 10 mmol/m3/yr in sediments just beneath the sulfate reduction zone and rapidly decrease to rates of <0.1 mmol/m3/yr at depths greater than 100 mbsf (Fig. F4B). These rates are comparable to those found in other high-productivity continental margin settings and lead to methane supersaturation at of 30–50 mbsf, which corresponds to the observed onset of gas hydrate in the sediments at most Leg 204 sites (Claypool et al., this volume). Within the gas hydrate stability zone, the hydrocarbon composition (C1/C2 > 10,000) and isotopic data (13CCH4 of –65 to –77) are consistent with an in situ microbial source for the methane (Fig. F4C). The gas hydrate at these sites is, therefore, referred to as reaction dominated (Claypool et al., this volume).
The carbon isotope and hydrocarbon observations (Fig. F4) are consistent with an observed increase in dissolved iodide and bromide concentrations (Fehn et al., this volume), as these halogens are released into the pore fluids during marine organic matter diagenesis. The general observations for halogen concentrations are in good agreement with those found for other gas hydrate locations (e.g., Peru margin, Martin et al., 1993; Blake Ridge, Egeberg and Dickens, 1999; Nankai Trough, Fehn et al., 2003). However, age estimates based on 129I measurements suggest that the iodine and methane in the pore waters sampled at Hydrate Ridge is older than 15 Ma, indicating fluid migration from early Eocene sources (Fehn et al., this volume). Although sediments of that age do occur ~40 km to the east, these old ages are not consistent with other geochemical data that indicate in situ methane generation. The reason for this discrepancy remains unresolved.
The transport-dominated regime is associated with active venting of gas at the SHR summit (Sites 1248–1250), where gas seepage supports rapid growth of gas hydrates near the seafloor and vigorous anaerobic methane oxidation (Torres et al., 2002; Boetius et al., 2000; Boetius and Suess, 2004). Geophysical, geochemical, and lithologic data all indicate that Horizon A acts as a conduit delivering gas that has migrated from greater depth in the accretionary complex to the summit vents. Horizon A can be mapped as continuous surface that laps onto the accretionary complex to the east and intersects the BSR on the west. Updip from Site 1245, its amplitude is anomalously large (Fig. F5A). LWD density logs and shipboard porosity and grain density measurements can be used to calculate the free gas content of the pore space (Tréhu et al., 2004a) and indicate gas saturation high enough (50%–90%) to support a connected gas column that extends from at least the depth of Horizon A at Site 1245 to the summit. Assuming that the gas pressure at Site 1245 is equal to the hydrostatic pressure, gas pressure at the BSR beneath the summit may exceed lithostatic pressure, a condition that is compatible with migration of gas through fracturing of the overlying sediments (Fig. F5B, F5C). Coring at Sites 1245, 1247, 1248, and 1250 clearly indicated the anomalous nature of this 2- to 4-m-thick, volcanic ash–rich horizon, as illustrated by the in situ bulk density and grain size distribution at Site 1245 (Fig. F5D). We speculate that the overpressures evolved as gas hydrate formed where Horizon A entered the GHSZ, thus sealing this boundary and focusing gas toward the summit. Sampling of Horizon A within the GHSZ (which was not done during Leg 204 because of time constraints) is needed to test this hypothesis.
The gas geochemistry also supports the mechanical scenario indicated by geophysical and lithologic data. Horizon A shows the signature of migrated gas at all sites (i.e., relatively low C1/C2 and high 13C) at which it was sampled (Fig.
F5E,
F5F), as do the shallow gas hydrates at the summit (Fig.
F5F). Using the isotopic and chemical composition of gas samples,
Claypool et al. (this volume) argue that the migrated gas component includes both previously buried microbial methane (~65% of the total gas) as well as thermogenic hydrocarbons (10%–15%). The isotopic composition of the thermogenic gas suggests an origin at temperatures in the range of 125°–135°C, or depths of 2.0–2.3 km (Claypool et al., this volume;
Tréhu, this volume). This temperature is attained in underplated, duplexed, and accreted sediments (Fig. F3B). Seismic velocities obtained from vertical seismic profiles (Tréhu et al., this volume) and sonic logs (Guerin et al., this volume) provide further support for a gas-charged conduit (Fig.
F5G). Low in situ P-wave velocity indicates abundant gas in the pore space. Low S-wave velocity suggests gas pressure high enough to result in a decrease in sediment shear strength and/or increase in pore space.
Although there is a consensus that Horizon A is the primary conduit feeding the summit vents and gas hydrate deposits, the mechanism whereby the gas migrates from Horizon A to and through the GHSZ remains controversial. Based on a one-dimensional, nonsteady-state transport reaction model, Torres et al. (2004b) argue that abundant methane in the free gas phase must be supplied from below in order to reproduce the dissolved chloride and gas hydrate distributions observed at the ridge summit (Fig. F5H). They reproduce the observed Cl– concentration and gas hydrate saturation by assuming vertical variation in the kinetic constants controlling methane dissolution and gas hydrate nucleation. Comparison of the internal pressures of gas hydrate crystallite and gas bubbles to the effective stress indicates that the depth at which gas hydrate can form by pushing away sediment grains corresponds approximately to the depth at the base of the shallow gas hydrate deposit (Fig. F5H). They conclude that capillary effects inhibit hydrate formation below ~30 mbsf, allowing methane gas to coexist with gas hydrate within the thermodynamically defined GHSZ (Clennell et al., 1999) and that gas migrates toward the ridge summit through a temporally variable network of pathways (Fig. F5I) controlled by catenary transport and/or pressure-dependent flow within hydrofractures (Tréhu et al., 2004a; Weinberger and Brown, 2006).
In contrast, based on the observations of high Cl– concentration at shallow depth (Fig. F5H), which suggest that free gas may locally be in thermodynamic equilibrium (Milkov et al., 2004b), Milkov et al. (2005) and Liu and Flemings (2006) have argued that formation of gas hydrate where Horizon A enters the GHSZ led to formation of high-salinity pore fluid near the base of the GHSZ, which shifted the hydrate stability field sufficiently to preclude further gas hydrate formation. In this steady-state model, free gas migrates vertically beneath the carbonate pinnacle that overlies the intersection of Horizon A and the BSR and then is deflected laterally at ~30 mbsf to emerge at the seafloor in the summit bubble vents (Fig. F5I). Torres et al. (2005) discuss the evidence for and against these two models for moving gas through the GHSZ and conclude that both models are plausible and that neither model can be excluded at the present time.
Transport of methane dissolved in pore water appears to be a secondary process at the summit. Whereas lithium concentration and isotopic data suggest some migration of water along Horizon A (Tréhu, Bohrmann, Rack, Torres, et al., 2003; L.-H. Chan and M.E. Torres, pers. comm., 2006), boron and strontium isotope data (Teichert et al., 2005b) are inconclusive. In contrast, strontium, boron, and lithium concentration and isotope data suggest that the deep fluid is transported along Horizon B, sampled at Site 1244 (Teichert et al., 2005b) Drilling confirmed that Horizon B (Fig. F2) results from a 10-m-thick zone of multiple turbidites that contain some minor free gas below the GHSZ (Site 1244) (Tréhu et al., this volume; Guerin et al., this volume) and gas hydrate within the GHSZ (Site 1246). Unlike Horizon A, Horizons B and B´ do not appear to contain enough free gas at present to develop high gas pressure and drive free gas migration through the GHSZ and may represent a fluid regime intermediate between the focused and diffuse flow end-members. Where Horizon B and B´ were sampled within the GHSZ (at Site 1246), they contained more hydrate than adjacent fine-grained sediments, illustrating the modulating effect of lithology on gas hydrate distribution (see "Lithologic Effects," below).
Fine-scale distribution of gas hydrate varies strongly with depth over distances of centimeters, with gas hydrate present in lenses and nodules of submillimeter to centimeter thickness. These lenses and nodules are heterogeneously distributed, occurring in clusters that are several meters thick and having orientations that range from horizontal to vertical (Tréhu et al., 2004b; Janik et al., 2003; Abegg et al., this volume, submitted [N1]). A statistical correlation between grain size and hydrate content as inferred from infrared temperature anomalies indicates that turbidite deposits containing more silt and sand than the intervening hemipelagic deposits are preferential sites for gas hydrate formation (Weinberger et al., 2005; Su et al., this volume). Similar lithologic overprinting of the gas hydrate distribution predicted by controls on gas hydrate distribution have been documented at Blake Ridge (Kraemer et al., 2000; Ginsberg et al., 2000), although the effect is not as strong because of the generally more homogeneous character of the Blake Ridge sediments. Comparison of results from Leg 204 to those from Expedition 311 (Expedition 311 Scientists, 2006), where numerous sand-rich horizons associated with gas hydrate were studied in detail, will further refine our understanding of the factors that control microscopic properties of gas hydrate growth.
One such cluster of gas hydrate lenses is associated with Horizon B (Fig. F2B) at Site 1246 (Fig. F6). Two coarse-grained turbidites characterized by high density and high magnetic susceptibility were found at 54–56.5 and 63.5–67 mbsf. These two horizons result in a strong, "ringy" reflection. The base of each turbidite was associated with cold infrared (IR) temperature anomalies; the lower one was also sampled for pore water analysis and yielded low pore water chloride indicative of 22% gas hydrate in the pore space, similar to the estimate of 28% gas hydrate determined from the IR data. At present, Horizon B is broken into many discontinuous segments because of normal faulting that is characteristic of the eastern flank of Hydrate Ridge. Because of evidence for free gas in Horizon B at Site 1244 and because of geochemical evidence for migration of fluids from greater depth, we speculate that Horizon B may have acted as an important conduit for free gas migration and may have fed a former seafloor vent system prior to being tectonically disrupted. Alternatively, the vertical faults in this region may facilitate upward fluid flow of fluids supersaturated with methane, which release this methane to form hydrate when they intersect Horizon B because the grain size and pore structure facilitate gas hydrate nucleation.
Strong lateral heterogeneity in gas hydrate distribution near the base of the GHSZ is indicated by a comparison of results in several different holes, spaced 30–40 m apart, at Site 1251. At this site, the reflection strength of the BSR and of the underlying reflectivity, which is attributed to the presence of lithologically controlled free gas, vary over short distances (Fig. F7A). Resistivity and nuclear magnetic resonance (NMR) data from Hole 1251A were interpreted to indicate free gas below the BSR but were not interpreted to show gas hydrate just above the BSR (Collett et al., this volume), although the resistivity data indicate that some gas hydrate is possible (Fig. F7B). In contrast, a 12-m-thick zone of relatively high gas hydrate content (~15% of pore space) was observed immediately above the BSR in Hole 1251D based on IR and chloride data (Fig. F7C). In Hole 1251H, the sonic log indicates gas hydrate enrichment of a few percent immediately above the BSR (Guerin et al., this volume).
This basal hydrate-rich layer may be the result of recycling of methane released in response to upward migration of the base of the gas hydrate stability zone triggered by rapid deposition of the overlying slope basin sediments. Sediments presently above the BSR were deposited at a rate of 60–160 cm/k.y. (Tréhu, Bohrmann, Rack, Torres, et al., 2003). A sharp decrease at the BSR in the C1/C2 ratio of gas exsolved from pore water is consistent with this model (Claypool et al., this volume). Although no significant variation in grain size in sediments associated with gas hydrate at this site has been documented, a more detailed look at variations in grain size and other lithologic parameters is warranted because the seismic data suggest that the distribution of free gas beneath the BSR is lithologically controlled. A similar concentration of gas hydrate just above the BSR may also be present at Site 1247. Such patchy lenses of anomalously high gas hydrate concentration at the base of the GHSZ have the potential to lead to deeply buried zones of high pore pressure and slope instability if these hydrates are destabilized as a result of tectonic uplift or a drop in sea level (Xu, 2004).
