Hydrate Ridge is a 25-km-long and 15-km-wide ridge in the Cascadia accretionary complex, formed as the Juan de Fuca plate subducts obliquely beneath North America at a rate of ~4.5 cm/yr (Fig. F1A). Sediment on the subducting plate contains large volumes of sandy and silty turbidites. At present, most of this sediment appears to be accreted to the continental margin, either by offscraping at the deformation front or by being underplated beneath the accretionary complex some tens of kilometers east of the deformation front (MacKay et al., 1992; MacKay, 1995) (Fig. F2).
Hydrate Ridge has been the site of many geological and geophysical cruises since cold seeps were first discovered on this part of the margin nearly 20 yr ago (Kulm et al., 1986). It is characterized by a northern peak having a minimum depth of ~600 m and a southern peak with a depth of ~800 m (Fig. F1B). Hydrate Ridge appears to be capped by hydrate, as indicated by a nearly ubiquitous and strong bottom-simulating reflector (BSR) (Trehu et al., 1999).
A regional two-dimensional (2-D) multichannel seismic survey was acquired in 1989 as a site survey for ODP Leg 146, which was designed primarily to study dynamics of fluid flow in accretionary complexes. The location where an upward deflection of the BSR is cut by a fault on the northern summit of Hydrate Ridge was selected to drill Site 892 (ODP Leg 194 Scientific Party, 1993). At this site, massive H2S-rich hydrates were recovered from 2 to 19 meters below seafloor (mbsf) (Kastner et al., 1995). No hydrate was recovered from near the BSR, but geochemical pore water and temperature anomalies suggested the presence of disseminated hydrate in the pore space to 68 mbsf (Kastner et al., 1995; Hovland et al., 1995). Vertical seismic profiles (VSPs) indicated the presence of free gas for at least 20 m beneath the gas hydrate stability zone (MacKay et al., 1994). Trehu and Flueh (2001) argue, based on seismic velocities and attenuation, that free gas is present for 500–600 m beneath the BSR at Site 892. Methane at Site 892 and in seafloor gas hydrates elsewhere is primarily of biogenic origin (Kvenvolden, 1995), but higher-order hydrocarbons of thermogenic origin are also present (Hovland et al., 1995; Schluter et al., 1998).
Since 1996, there have been several cruises per year, which have generated an extensive database of swath bathymetry, deep-towed side-scan, and seafloor observations and samples collected via submersible and remotely operated vehicle (Suess and Bohrmann, 1997; Torres et al., 1998, 1999; Bohrmann et al., 2002; Linke and Suess, 2001). In addition, a high-resolution three-dimensional (3-D) seismic survey was recently conducted in the immediate region of planned drilling (Trehu and Bangs, 2001).
Side-scan data, seafloor camera tows, and diving with manned and remotely operated submersibles demonstrated the presence of extensive massive carbonate pavement on the northern summit of Hydrate Ridge (Carson et al., 1994; Clague et al., 2001; Sample and Kopf, 1995; Bohrmann et al., 1988; Greinert et al., 2001). Until recently, massive authigenic carbonate pavement was thought to be absent on the southern summit of Hydrate Ridge. During Alvin dives in 1999, however, a 50-m-high carbonate "pinnacle" was discovered 250 m southwest of the summit (Torres et al., 1999). Deep-towed side-scan data indicate that the pinnacle is located in the center of a buried carbonate apron with a diameter of ~250 m (Johnson and Goldfinger, unpubl. data). Authigenic carbonates on the Cascadia margin form from methane oxidation in the sediments and discharge of isotopically light dissolved inorganic carbon at the seafloor and into the ocean. The relative absence of carbonate on the southern summit of Hydrate Ridge is thought to indicate that this hydrate/gas system is younger than that on the northern summit, providing a spatial proxy for temporal evolution of this hydrate-bearing accretionary ridge (Trehu et al., 1999). This interpretation is supported by U-Th dating of recovered carbonates (Teichert et al., 2001), which indicates that the pinnacle is <10,000 yr old, whereas the carbonate carapace on northern Hydrate Ridge is >100,000 yr old.
One especially interesting feature of southern Hydrate Ridge is the abundance of massive hydrate at the seafloor near its summit. This was first discovered in 1996, when >50 kg of massive hydrate was recovered in a television-guided grab sample (Bohrmann et al., 1998). The samples show dense interfingering of gas hydrate with soft sediment (Fig. F3A). In most cases, pure white hydrate occurs in layers several millimeters to several centimeters thick. Host sediment is often present as small clasts within the pure gas hydrate matrix. On a macroscopic scale, the fabric varies from highly porous (with pores of up to 5 cm in diameter) (Fig. F4B) to massive (Suess et al., 2001). Thin sections show a structure in which gas bubbles have been filled with hydrate (Fig. F3C). Wet bulk densities of 80 hydrate samples measured on board the Sonne range from 0.35 to 7.5 g/cm3. Pore space was estimated from the change in sample volume before and after compression to ~160 bar (Bohrmann et al., 2000). The samples show high variability in pore volumes ranging from 10–70 vol%, and the values are negatively correlated with sample density. From this correlation, the end-member density at zero porosity was estimated to be ~0.81 g/cm3. This value is lower than the theoretical density of pure methane hydrate (0.91 g/cm3). Field-emission scanning electron microscopy indicates that this is due to submicron porosity of the massive hydrate (Techmer et al., 2001).
The low bulk density of the natural methane hydrates from Hydrate Ridge results in a strong positive buoyancy force, implying that the hydrate remains on the seafloor only because of the shear strength of the host sediment. Unusual seafloor topography observed on southern Hydrate Ridge during Alvin and ROPOS surveys, which is characterized by mounds and depressions with a wavelength of a few meters, may result from spontaneous breaking off of hydrate from the seafloor. This may be an important mechanism for transporting methane from the seafloor to the atmosphere.
Vigorous streams of methane bubbles have been observed emanating from vents on the seafloor on the northern and southern peaks of Hydrate Ridge (Suess and Borhmann, 1997; Suess et al., 1999; Torres et al., 1998, 1999) as well as from a similar, but smaller, reflective high in the accretionary complex known as SE Knoll (Figs. F1B, F3D). Because the seafloor at all three sites is well within the hydrate stability zone (Fig. F3E), the presence of methane bubbles beneath and at the seafloor suggests rapid transport of methane to the seafloor from sediments beneath the hydrate stability zone. Because seawater is undersaturated in methane, their persistence in the water column suggests that they are protected by a thin coating of hydrate (Suess et al., 2001; Rehder et al., unpubl. data). Disappearance of the acoustic "bubble" plumes at 450–500 m below the sea surface (near the top of the hydrate stability zone) suggests that the hydrate shell dissociates and that most of the methane in the bubble plumes dissolves in the ocean rather than reaching the atmosphere.
Two 2-D multichannel seismic profiles across southern Hydrate Ridge acquired in 1989 suggested a complicated subsurface plumbing system related to the presence of hydrate and free gas. Prior to a 3-D high-resolution seismic survey in 2000 (Trehu and Bangs, 2001), the relationship between subsurface reflections and the summit vents was not known because no profiles crossed the summit. The 3-D survey comprised 81 profiles spaced 50 m apart spanning the region between the two southern lines from the 1989 survey (Fig. F1C). Shots from two generator-injector (GI) guns fired simultaneously were recorded on the Lamont-Doherty Earth Observatory portable 600-m-long, 48-channel towed streamer and on an array of 21 UTIG four-component ocean bottom seismometers (OBSs). The locations of the ship and the streamer were determined via differential Global Positioning System (GPS) and four compasses, respectively, and 3-D fold was monitored during the cruise to identify locations where additional data were needed. Excellent data quality was obtained in spite of strong winds and high seas. The data contain frequencies up to ~250 Hz, providing considerable stratigraphic and structural resolution.
Figure F5 shows east-west line 230 from the data volume. The data have been 3-D prestack time migrated and then converted to depth using velocities from a 3-D P-wave velocity model derived from tomographic inversion of first arrivals recorded on the OBSs (Arsenault et al., 2001). The profile is coincident with line 2 from the 1989 site survey (Fig. F2A ). Locations of several of the planned drill sites are also shown as are boundaries between an upper facies of folded and uplifted sediments that unconformably overlies a stratigraphic sequence in which seismic layering is less pronounced. This facies in turn overlies a low-frequency, incoherent zone interpreted to be highly deformed accretionary complex material.
The data show considerable stratigraphic and structural complexity both above and below the BSR. Certain reflective horizons are anomalously bright, and these amplitude anomalies are consistent for hundreds of meters. In particular, we point out the event labeled "A" on Figure F5. This reflection has an amplitude that is ~10 times greater than that of adjacent stratigraphic events and two times greater than that of the BSR. Relative true-amplitude seismic sections illustrating characteristics of reflection A and of overlying actively venting features near the southern summit of Hydrate Ridge are shown in Figure F4A, F4B, F4C, and F4D. Locations of sections are shown on the side-scan map to illustrate the relationship between reflection A and seafloor venting. A north-south slice through the data volume indicates that this bright, negatively polarized stratigraphic horizon is continuous with the bright "spot" immediately underlying the BSR beneath the summit (Fig. F4D). We speculate that this surface transports methane-rich fluids toward the summit of southern Hydrate Ridge and predict that variations in stratigraphic permeability favored fluid flow along this horizon, which may be an unconformity. We further speculate that diagenetic reactions have resulted in a feedback effect, enhancing flow along this surface. These speculations will be tested during Leg 204 by drilling at proposed Sites HR3a, HR4a, and HR4c.
Chaotic bright reflectivity is observed just beneath the seafloor at the summit (line 300; Fig. F4B). This reflectivity pattern is observed only at the summit and is almost exactly coincident with the "tongue" of intermediate-strength seafloor reflectivity northeast of the "pinnacle" observed in deep-towed side-scan data. This pattern also underlies the acoustic "bubble" plume that was observed each time the southern summit was crossed during the seismic data acquisition cruise. We speculate that this pattern indicates the depth extent of massive hydrate, and we will test this speculation by drilling at proposed Site HR4b. Whereas it appears that reflection A is a primary source of fluids for the summit vents, the mechanism whereby methane migrates to the seafloor is not imaged in the seismic data. We speculate that the region between reflection A and the seafloor is broken by small faults that are not well resolved in the seismic data but that permit methane-rich fluids to rise vertically from reflection A to the seafloor.
Complicated reflectivity patterns are also observed east of the southern Hydrate Ridge axis and are associated with an active secondary anticline (anticline A in Fig. F5). The "double BSR" originally identified on line 2 from the 1989 site survey (labeled C in Fig. F5) shallows to the south and merges with the BSR along 3-D line 274. It is continuous with a package of bright, regionally extensive reflections that cut across the BSR (labeled B and B' in Figs. F4 and F5). Although these reflections are very strong, they are "ringy," and polarity cannot be unambiguously determined, unlike for the BSR and for reflection A. They are also pervasively faulted, with offsets consistent with tensional cracking in response to uplift and folding. Their amplitude does not change abruptly as they cross the BSR, although there is a slight increase in amplitude as depth decreases. The lack of change in amplitude of reflection B as it crosses the BSR suggests that the high reflectivity is not a result of free gas beneath the BSR, which should form hydrate on entering the hydrate stability zone, thus changing the reflectivity. We speculate that reflection C is an unconformity at the base of an uplifted and deformed slope basin, within which the BSR is relatively weak. We further speculate that permeability of the slope basin sediments is generally low, that sedimentary horizons marked by reflections B and B' have higher permeability than adjacent strata, and that fluids rising through the accretionary complex migrate along the unconformity until they reach the B and B' horizons. The high amplitude of these reflections may result from carbonate formation along this horizon as a result of fluid flow. Proposed Sites HR1a and HR1b target the bright reflection pair B below and above the BSR. Proposed Site HR1c targets reflection C.
The BSR is anomalously shallow in the saddle between axis of Hydrate Ridge and anticline A (Fig. F5). If one assumes that seafloor temperature, sediment velocity between the seafloor and the BSR, and fluid and gas composition are known, the apparent thermal gradient can be calculated from observed BSR depth (e.g., Zwart et al., 1996). Beneath Hydrate Ridge, the depth of the BSR is generally consistent with a temperature gradient of 0.06°C/m, assuming an average velocity between the seafloor and BSR of 1.6 km/s, as determined from OBS data, and a seafloor temperature of 4°C at 800 m and 3°C at 1200 m, as indicated by hydrographic data. A BSR uplift of ~20 m, implying a temperature gradient of ~0.07°C/m, is suggested in the saddle between anticline A and the crest of Hydrate Ridge with the same assumptions. However, if reflections B and B' are caused by carbonates in the hydrate stability zone, then average velocity may be higher and the inferred temperature gradient may be lower. Assuming an average velocity of 1.80 km/s above the BSR almost eliminates the apparent thermal gradient anomaly. A slightly lower temperature gradient is suggested for the slope basin to the east, suggesting fluid flow toward Hydrate Ridge, although this may also be, in part, due to lateral variations in sediment velocity. Leg 204 will provide critical constraints for decreasing the uncertainty in deriving constraints on fluid flow from observations of BSR depth.
Communities of tube worms, bacterial mats, clams, and other fauna are associated with seafloor hydrates and methane vents on Hydrate Ridge and elsewhere (e.g., Kulm et al., 1986; MacDonald et al., 1989; Suess et al., 1999, 2001; Sassen et al., 2001). Microorganisms are at the base of the food chain in these communities. Recent work suggests the complex complementary relationships between sulfate reducing, methanogenic, and methanotrophic microorganisms in hydrate-bearing sediments (e.g., Parkes et al., 2000; Boetius et al., 2000). These microorganisms must be playing an important role in methane formation and oxidation and are therefore a critical component of the hydrate system. Identification of these organisms and determination of their abundances, spatial variability, and rates of activities is just beginning.
Particularly interesting are recently discovered organisms that play a critical role in anaerobic methane oxidation (AMO), which is the process forming the carbonates that remain in the geologic record and record of the history of past fluid flow and hydrate formation and dissociation (e.g., Sample and Kopf, 1995; Bohrmann et al., 1998; Greinert et al., 2001). Very high rates of AMO have been measured in sediment overlying massive gas hydrates on southern Hydrate Ridge (Boetius et al., 2000) and attributed to structured aggregates consisting of a central cluster of methanotropic archaea coated by sulfate-reducing bacteria. That microbes oxidize methane by utilizing sulfate in the absence of oxygen was long suspected by geochemists, based on interstitial sulfate and methane gradients and Borowski et al. (1996), who showed that steep sulfate gradients and shallow depths to sulfate-methane interface are a consequence of the increased influence of AMO, but Boetius et al. (2000) were the first to observe the microorganisms that presumably catalyze anaerobic methane oxidation. These bacterial aggregates appear to be abundant in sediments of Hydrate Ridge and mediate AMO when enough sulfate is available.
Analysis of sulfide minerals provides a possible opportunity to reconstruct past biological activity because most of the reduced sulfide produced during bacterial sulfate reduction in nonhydrate-bearing sediments is ultimately sequestered in various iron phases, which usually involve multiple steps terminating in the formation of sedimentary pyrite. In the Cascadia margin, the sequestering of sulfide into the clathrate structure (e.g., Kastner et al., 1995; Bohrmann et al., 1998) essentially removes it from further reaction with ferrous iron complexation. There is a wealth of information on the significance of iron sulfide interactions in marine sediments (e.g., Berner, 1970; and many others). The burial of this mineral phase contributes significantly to the oxygen level of the atmosphere, the sulfate concentration in seawater, and the pH of the oceans over geologic timescales (e.g., Garrels and Perry, 1974; Holland, 1978; Boudreau and Canfield, 1993). Another significant effect of H2S sequestering by hydrates is the development of anomalous intervals of high greigite content at the intervals in which gas hydrates were recovered or were inferred to exist (Housen and Musgrave, 1996). Based on the rock magnetic properties at Site 889, Housen and Musgrave (1996) identified the presence of a "fossil gas hydrate zone" that extended downward to 295 mbsf during the last glaciation.