Sites 994, 995, and 997 were drilled to 700-750 mbsf on the Blake Ridge within the same stratigraphic units. These three sites form a 10-km-long transect that extends from the ridge flank, where there is no BSR, to the ridge crest where there is a very strong BSR (Fig. 1, Fig. 2). The holes extended through the base of gas hydrate stability (~450 mbsf) and 250-300 m into the underlying sediments.
Most of the sediments recovered during Leg 164 accumulated during the Pliocene and Miocene (Okada, Chap. 33, this volume) at very rapid rates (up to 350 m/m.y.). The sediments were deposited by the south-flowing Western Boundary Undercurrent that sweeps southward along the Atlantic Margin (Shipboard Scientific Party, 1972; Gradstein and Sheridan, 1983). The stratigraphic sequence is composed of lithologically rather homogeneous, nannofossil-rich clays and claystones and variable amounts of opalline silica. Intersite and downhole variation in sediment physical properties and lithology was minimal. The lithologic similarity allows the distribution of gas hydrate and the origins of the BSR (~450 mbsf) to be studied without the complication of lithologic factors. The sediments typically contain between 0.5% and 1.5% organic carbon, enough to generate significant quantities of microbial methane.
The cored sediments were very gassy and experienced vigorous expansion during recovery. Sediments frequently extruded from the liners as the cores arrived on deck. Sediment recovery between ~190 and 480 mbsf tended to be low (averaging ~50%). The gas is predominately methane (~99%) with secondary amounts of carbon dioxide, and only trace amounts of ethane and other hydrocarbon gases.
Recovered gas hydrate occurred as nodules and veins. The largest sample was a >30-cm-long piece of massive gas hydrate from Site 997 on the Blake Ridge (Fig. 3). Decomposing gas hydrates yielded gas that was ~99% methane and ~1% carbon dioxide. Volumetric ratios of the evolved gas and water for these hydrate samples ranged from 130 to 160, which is a stoichiometric ratio that suggests at least 70% cage filling of Structure I methane gas hydrates (Sloan, 1990). The most commonly observed hydrate occurrences were associated with thin (~1 mm) veins of white hydrate, which could be seen on the surfaces of quickly opened cores from Site 996 on the Blake Ridge Diapir.
Measurements of interstitial water chloride concentrations from high-resolution sampling provided quantitative estimates of the minimum amounts of in situ gas hydrate in individual samples. During gas hydrate formation, water and methane removal leave residual pore waters increasingly saline. Over time, locally elevated chloride concentrations associated with gas hydrate formation diffuse away. When sediment-hosted gas hydrates decompose during drilling and core recovery, they release water and gas back into the pore space, freshening the pore waters. Pore-water profiles from Sites 994, 995, and 997 trend to fresher values to depths of ~200 mbsf. From ~200 mbsf to the depth of the BSR (~450 mbsf), pore-water chloride concentrations are highly variable and characterized by local, anomalously fresh values (Fig. 4). Beneath ~450 mbsf, the chloride values are nearly constant. Departures from base-line chloride values can be used to calculate the minimum amount of gas hydrate that occurs in the samples. Shipboard calculations of the amount of gas hydrate that decomposed in individual samples were quite variable, but most samples had 1%-2% gas hydrate filling the total sediment volume and some ranged up to 14% within the zone from 200 to 450 mbsf. Nearly as much gas hydrate was inferred to occur at Site 994, which was not associated with a BSR, as at Sites 995 and 997, where a well-developed BSR is identified.
An extensive suite of logs was run at each deep site. Well logs show distinct zones of higher velocity and electrical resistivity that are coincident with zones where chloride anomalies indicate the presence of gas hydrates (Fig. 5). Shipboard physical properties data (e.g., porosity, grain density, water content, etc.) do not explain these variations. Thus, it is inferred that the log parameters are related to properties that are lost during core recovery (e.g., free gas and gas hydrate). Calculations of the amount of gas hydrate needed to produce these offsets are consistent with predictions from the chloride data. Below 450 mbsf, sediment velocities are equal to or less than seawater (1500 m/s), indicating the presence of free gas.
A pressure core sampler (PCS) was used with unprecedented success (Dickens et al., 1997). The device takes a small sediment core (42 mm diameter, up to 0.86 m long) from the bottom of the bore hole and seals the core into a pressure housing so that recovery occurs under in situ pressure. Thus, PCS cores contain all their original gases until they are opened to the atmosphere. PCS cores recovered from Sites 995 and 997 produced between 0.8 and 46 L of methane per liter of pore space (Dickens et al., Chap. 11 and Chap. 43, both this volume). Gas volumes captured by the PCS often grossly exceeded (~10 times) the concentration needed for gas saturation under surface conditions (Fig. 6). Such observations make it easy to understand why the other cores expanded so dramatically during recovery. Above the base of gas hydrate stability, the excess gas was presumably supplied from dissociation of gas hydrates that would have occupied up to 8% of the sediment volume. The excess gas that was captured at and beneath the base of gas hydrate stability demonstrates that free gas is associated with the BSR and occurs intermittently throughout the sedimentary section below. In fact, the volume of free gas trapped in sediments beneath the base of gas hydrate stability rivals the amounts stored as gas hydrate. Dickens et al. (1977) used the measured amount of gas and the known regional distribution of the BSR (Dillon and Paull, 1983) to estimate that ~35 Gt of methane carbon is stored in the Blake Ridge (15 Gt are in gas hydrate, 5 Gt dissolved in pore water, and 15 in free gas).
The velocity structure of the sediments and the precise depth of the BSR were determined with vertical seismic reflection profiles (Holbrook et al., 1996; Fig. 7). The velocities of sediments overlying the BSR only reached ~1850 m/s, consistent with the presence of a few percent gas hydrate. However, velocities as low as ~1400 m/s were measured beneath the BSR at Site 997, indicating the presence of free gas.
More than nine in situ temperature measurements establish the temperature gradient to depths of up to 415 mbsf at each site (Ruppel, 1997). Extrapolation of these thermal gradients to depths that correspond with the BSR allows the experimentally predicted gas hydrate phase boundaries (assuming hydrostatic pressure) to be determined. The predicted temperatures are 0.5°C to 2.9°C colder than experimental data predict the base of gas hydrate stability would be.
Many cores came up surprisingly cold and some cores were partially frozen. The cool temperatures measured in the zone where gas hydrates (200-450 mbsf) were common are interpreted to be an artifact of the recent decomposition of gas hydrates and the cooling associated with the expansion of escaping gas. Gas hydrate dissociation is endothermic and thus cools the surrounding sediments. However, surprisingly cold temperatures were also measured on freshly recovered cores from Site 997 below the gas hydrate-bearing zone. These low temperatures are attributed to cooling associated with the expansion of escaping gas bubbles.
Short holes (50-67 mbsf) were drilled on the flanks and crest of the Cape Fear Diapir (Sites 991, 992, and 993) and the Blake Ridge Diapir (Site 996). Increases in salinity observed at these sites suggest that salt cores exist in the diapirs at depth.
Sites 991, 992, and 993 all lay within the scar of the Cape Fear Slide. A thin veneer (2.05 m) of Holocene material covered the slide scar at Site 991, but older materials were exposed at the seafloor at Sites 992 and 993. The sediments from Sites 991 and 992 were strongly deformed.
Site 996 is located where gas-rich plumes occur in the water column above a pockmarked seafloor associated with active chemosynthetic biological communities (Paull et al., 1995). Beneath these sea-floor features, a small fault extends downward toward the BSR, suggesting that fluid migration is associated with gas hydrate-bearing sediment below. Gas hydrates were common in cores from the uppermost sediment column.
The shallow cores (>20 mbsf) evolved enough hydrogen sulfide to saturate hand-held sensors (>100 ppm). Because hydrogen sulfide concentrations decrease with depth and are not elevated at other sites in the region, the hydrogen sulfide is believed to be produced locally from sulfate via anaerobic methane oxidation (Whiticar and Faber, 1986) near the seafloor. Thus, the methane supplied from below and sulfate derived from seawater react to produce hydrogen sulfide within the surface sediments. Pervasively cemented carbonate horizons and fossil seep communities were found in the subsurface. Thus, seepage appears to have occurred at least intermittently at this site throughout the Pleistocene.