Site 994 is part of a transect of holes on the southern flank of the Blake Ridge that extends from an area where a BSR is not detectable to an area where an extremely well-developed and distinct BSR exists (Fig. 5). The transect is situated where variations in the development of the BSR and seismic blanking are especially distinct. However, the geology and topography along this transect are relatively simple (Fig. 6), which provides an opportunity to assess basic properties of hydrate bearing sediments and to understand lateral hydrate variations caused by local lithologic, chemical, and hydrologic factors. Site 994 is situated at the end of the transect where the BSR is not detectable and, thus, serves as a background or reference site.
At Site 994, we recovered a 700-m-thick sequence that is dominantly composed of clay and calcareous nannofossils (Fig. 6). Three major lithologic units were identified, based on downward decreasing contents of calcareous nannoplankton. Unit I (0-14 mbsf) has the highest nannofossil contents (up to 45%) correlating with the lightest colors of any sediment found in the hole. Unit II (14-160 mbsf) consists of decimeter- to meter-thick layers of dark greenish gray nannofossil-rich clay and more carbonate-rich greenish gray nannofossil clay. Nannofossil contents average 25%, with values of 40% and higher in the lighter colored beds. Unit III (160-703.5 mbsf), which extends to the bottom of the hole and comprises the remainder of the Pliocene to uppermost Miocene, consists of a monotonous succession of dark greenish gray diatom-bearing nannofossil rich clay, with average CaCO3 contents of 20%. Within most of the section (from 180 mbsf to the bottom of the hole), grain densities are homogeneous. Within the same interval, wet-bulk density and strength increase linearly, whereas water contents gradually decrease with increasing depth. The lithologic homogeneity of sediments at Site 994 makes it an ideal reference section for comparative gas hydrate studies.
The section obtained in Hole 994C contains a continuous record spanning from the Holocene to latest Miocene (~6 Ma) in age (Fig. 7). All nannofossil zones and subzones were present, and no obvious hiatuses were identified. Measurement of discrete paleomagnetic samples allowed determination of approximate positions for the Brunhes/Matuyama (40 mbsf) and Gauss/Matuyama (150 mbsf) magnetochron boundaries and for the Olduvai subchron (90-115 mbsf). Estimated sedimentation rates increased consistently downsection, reaching a maximum of 400 m/m.y.
Sediments in many cores were very gassy. Most cores underwent extensive gas-driven self extrusion, sometimes splitting the liners near the shoe, and one core burst its liner within the core barrel on the drill floor. Disruption of the cores due to gas expansion began occurring at ~60 mbsf. Poor recovery below 190 m is largely a result of vigorous degassing. The gas is mostly methane with secondary amounts of CO2. Slight increases in ethane contents with depth were observed; however, the sediments did not reach the zone of thermogenic hydrocarbon production.
The chloride profiles from Holes 994A, B, and C (Fig. 8) contain anomalously low-chloride spikes superimposed on a trend of generally decreasing chloride concentration from typical seawater values at the top of the section to ~90% of seawater at 300-400 mbsf. A slight increase in chloride toward the bottom of the hole is present. The lowest measured chloride concentration is 438 mM. In the interval from 100 to 450 mbsf, chloride anomalies were found in 50% of the interstitial water samples, and they are particularly abundant and well developed at 380-440 mbsf. Detailed studies on cores from Hole 994D revealed that the anomalous low-chloride spikes extend over less than 1.5 m. The spikes are interpreted as evidence for the presence of gas hydrate, which melts and dilutes the interstitial water during coring and processing.
Gas hydrate was sampled from two cores. Section 164-994C-31X-7 (258 mbsf) contained several white nodules of hydrate that ranged in volume from ~4 to 25 cm3. Gas volume, gas composition, and water volume of a solid piece of hydrate were determined. The gas composition of the hydrate was 98.78% methane, 1.22% CO2, 86 ppm ethane, and 2 ppm propane. Volumetric calculations show that the cage occupancy (percentage of potential gas molecule sites in the hydrate crystal lattice that are actually filled) of the hydrate sample is at least 80%. In Section 164-994D-4X-1 (261 mbsf), a piece of hydrate that was less than 1 cm3 was found. In the region where solid hydrate was recovered from the cores, the physical properties data do not reveal any significant changes in any sediment properties. Many cores recovered from 240 to 430 mbsf contained anomalously cold zones (measured using temperature probes on the catwalk) indicating that gas hydrate had decomposed within these cores, even though hydrate was not visually observed.
Fourteen in situ temperature measurements were attempted between 0 and 445 mbsf in Hole 994C (Fig. 9). Based on data from eight successful deployments of the Adara temperature tool and water sampler temperature probe (WSTP) tools, the average geothermal gradient in the uppermost 320 m is estimated at 38.6°C/km. Average heat flow is 35 mW/m2 over the entire depth range of the measurements, but 45 mW/m2 within the upper 100 m of the hole. Anomalously low temperature (6-8°C) measurements were made at four depths between 300 and 425 mbsf. Although there is no simple explanation for the shape of the equilibration paths associated with the anomalous records, the low in situ temperatures are consistent with those measured close to solid hydrate in the recovered cores. Such anomalously low temperatures might be produced by the endothermic decomposition of solid hydrate.
Rock-magnetic studies indicate two prominent features superimposed on a continuous sequence of magnetite to pyrite reduction. At 260 mbsf, immediately below the hydrate recovery in Core 164 994C-31X, and at 365 mbsf, coincident with an anomalously low interstitial-water chloride content, rock-magnetic signatures are similar to those found on samples from Leg 146 at the base of hydrate concentrations and "fossil hydrate zones."
A suite of wireline logs was run, including neutron density, neutron porosity, resistivity, P-wave sonic, shear sonic, and geochemical logs. In addition, several measurements with the geochemical tool in the inelastic mode were made in an attempt, for the first time in ODP history, to determine carbon/oxygen ratios. Initial analysis of the neutron density, resistivity, and sonic logs shows distinct changes in P-wave velocity and resistivity in two important regions (Fig. 10 and 11). The first corresponds to the zone below 220 mbsf from which gas hydrate was recovered. In this zone, velocity and resistivity have pronounced positive spikes. The second region at 420 mbsf, where velocity and resistivity decrease sharply, is near the predicted base of the hydrate stability field. In contrast, the density log does not show changes at corresponding depths. The formation factor (observed resistivity/seawater resistivity) increases with depth from 220 to 420 mbsf and then abruptly decreases to lower values. The relatively constant density throughout this interval suggests that the rise in resistivity is not entirely due to changes in porosity and may reflect increasing amounts of gas hydrate with depth in this zone.
Vertical seismic profiles were conducted at depths of 110-650 mbsf during five lowerings of the three-component Woods Hole Oceanographic Institution (WHOI) borehole seismometer in Hole 994D. Difficulties with the clamping arm restricted the acquisition to two walkaway VSPs at 650 and 482 mbsf and eight zero-offset clamps over the same depth range. Zero-offset air gun shots were fired with the tool suspended in the hole at 20-m intervals from 570 to 110 mbsf. A stacked record section shows clear first arrivals from 250 to 650 mbsf. A preliminary P-wave velocity model (Fig. 12) shows elevated velocities (with respect to background levels) in a zone from 320 to 420 mbsf and a pronounced low-velocity zone from 550 to 650 mbsf. The low-velocity zone may be due to the presence of free gas bubbles dispersed in the sediments at this depth.
Estimates of gas hydrate amounts in the sediments before recovery at Site 994 are made by assuming that diffusive equilibration prohibits significant and nonsystematic interstitial chloride concentrations from occurring between closely spaced samples (Fig. 8). Thus, chloride spikes are only a result of gas hydrate decomposition during sample recovery. To produce an estimate of the interstitial-water salinity through the zone between 200 and 450 mbsf, where erratic chloride values were measured, a polynomial was fit to the relatively smooth chloride data above 200 m and below 450 m. All but one of the measured chloride concentrations in this zone have lower values than the calculated in situ values, with some significantly lower. The differences between the calculated in situ chloride concentrations and the measured chloride concentrations were used to establish the relative chloride anomaly that is associated with each sample. The calculated chloride anomalies enable the amount of gas hydrate that was present in these samples to be estimated. Corrections for the porosity of the samples were made using the shipboard physical properties data. The estimated percent volume of the samples that was occupied by gas hydrate had a skewed distribution, ranging to as much as 7% (at 391 mbsf), with a mean value of 1.3% ± 1.8% and a median value of 1% for all the interstitial-water samples that were collected between 200 and 450 mbsf.
Calculations of the percentage of gas hydrate that is required to explain the observed change in the well-logging resistivity trend(Fig. 11) indicate that the general trend through this interval (between 212.0 and 428.8 mbsf) can be explained by the pervasive addition of up to 2.9% gas hydrate. The same calculation indicates that the horizon with the highest concentration (~239 mbsf) contains as much as 9.5% gas hydrate. It is remarkable that independent estimation of the amounts of gas hydrate from different data sets (chloride anomalies and logging data) yield similar values of a few percent gas hydrate disseminated in the sediments.
In summary, although very little gas hydrate was recovered from Site 994, the interstitial-water chloride anomalies, temperature anomalies in recovered cores, and patterns in the well-log data all indicate that gas hydrate occupies an average of 1% or more of the sedimentary section from 220 to 430 mbsf. The hydrate is inferred to occur as finely dispersed crystals within homogenous sediments. All of the inferred hydrate occurs well above the predicted base of gas hydrate stability at this site.