Results of our analyses are presented in Table 1. Our most important observation is that there is a distinct change in lithology across the upper hydrate zone between 185 and 260 mbsf. Calcium carbonate content of sediment averages about 25% above and below the upper hydrate zone, and about 8% within the upper hydrate zone (Fig. 3). Our values are similar to carbonate concentrations determined on ship (Paull, Matsumoto, Wallace, et al., 1996).
Nearly all Sc in marine sediment exists in the aluminosilicate fraction (Schmitz et al., 1991; Dickens and Owen, 1996). Therefore, scandium can be used as a first-order proxy for the amount of aluminosilicates (in this case, clay), provided the amount of Sc in the aluminosilicate fraction is known (Schmitz et al., 1991; Dickens and Owen, 1994). A value of 18 ppm Sc is used as an average of total concentration of Sc in a mixture of deep-sea clay and terrigenous shale (Gromet et al., 1984; Faure, 1991). There is a broad increase of about 10% in the clay content across the upper hydrate zone (Fig. 4), with a few peaks in clay content. This is consistent with logging results (Paull, Matsumoto, Wallace, et al., 1996).
There is also an increase in percent amorphous silica; the analytical results indicate that siliceous microfossils (as percent opal) are generally more abundant in the hydrate zone and can make up to 8% of the sediment (Fig. 5). Shipboard observations (Paull, Matsumoto, Wallace, et al., 1996) as well as our own (Pl. 1A, B) show this silica to be comprised mostly of diatom tests.
We do not know the abundance of quartz and other lesser components at Site 994 beyond qualitative estimates from smear slides. However, the sum of carbonate, clay (as determined from Sc), and biogenic silica is close to 90% for our samples. This is consistent with observations that there are only four major (>5%) components in the sediment (carbonate, clay, biogenic silica, and quartz), and that the quartz content averages about 10% (Paull, Matsumoto, Wallace, et al., 1996).
There is a highly
significant (
< 0.001) positive correlation
between bulk porosity (from Paull, Matsumoto, Wallace et al., 1996) and percent
biogenic silica (Fig. 6, Fig.
7). This indicates that an increase in percent opal results in an
increase in the porosity of the sediment (Tribble et al., 1992), and that the
sediment in the upper hydrate zone is significantly more porous than the
surrounding sediment.
There are two ways that diatoms increase the porosity of sediment. Diatom tests folded on themselves or crushed against each other (Pl. 1B) create large (on the order of 5 mm × 50 mm) pore spaces. The microporosity of the sediment is also a factor. Pores in the diatoms are regularly about 2 mm in diameter (Pl. 1A, B). These openings are much larger than the pore space of surrounding sediment (Pl. 1A, B).
The porosity difference between diatoms and their surrounding matrix (carbonate, clays, and quartz) may have a significant impact on the formation and distribution of gas hydrate in sediment. Capillary forces between grains are inversely proportional to the radii of the menisci of interfacial water (Clennell et al., 1995; Ruppel, 1997). The menisci are between the solid methane hydrate and the unfrozen pore water, and the contact angle is effectively 180°. In fine-grained sediments, the menisci have extremely small radii (Klausner, 1991). This results in strong capillary forces and decreased hydrate formation in sediment dominated by carbonate and clay. However, sediment with abundant diatoms has a high number of large and round pore spaces. The presence of these openings provide nucleation sites for gas hydrate that are uninhibited by capillary forces between grains. Diatom tests should promote hydrate formation relative to the surrounding matrix.
The upper hydrate zone in sediment on the Blake Ridge can be explained by a depositional mechanism involving fluid flow and lithological change. The presence of abundant free gas bubbles below the bottom-simulating reflector (BSR) on the Blake Ridge (Dickens et al., 1997) indicates that water is supersaturated with methane at depth. Pore-water chloride profiles on the Blake Ridge are best explained by invoking significant upward advection of fluids (Egeberg and Dickens, unpubl. data). We suggest that methane-supersaturated fluids flow upward through the sediment column at Site 994. Gas hydrate is not easily precipitated in sediment dominated by clay and carbonate because hydrate cannot nucleate in the small pore spaces. However, when the supersaturated fluids encounter sediment with abundant diatoms, gas hydrate precipitates in the large and round openings between and within the diatoms. The depth interval between 260 and 380 mbsf (i.e., between the upper and lower gas hydrate zones) lacks abundant gas hydrate because there are relatively few diatom nucleation sites (Fig. 5).
The proposed depositional mechanism for gas hydrate in upper sediment on the Blake Ridge has analogs in other environments. For example, Tardy and Novikoff (1988) and Putnis et al. (1995) discussed sediment systems where precipitation of minerals from supersaturated solutions is inhibited in fine-grained sediments, but occurs in adjacent coarse-grained lithologies.
Another recent explanation for the existence of an upper hydrate zone is the possibility that chemically distinct gas hydrates may be stable in different burial depths (Kastner et al., 1998). However, evidence indicates that Hole 994C hydrate is chemically homogenous and contains only Structure I hydrate (with only trace amounts of propane). Consequently, the hydrate at Site 994 should not be vertically heterogeneous because of chemical variations.
The Blake Ridge is a contourite deposit created by sediment deposition from the thermohaline-induced Western Boundary Undercurrent (Shipboard Scientific Party, 1972). Primary deposition is by hemipelagic accumulation of terrigenous and biogenic components as well as by biogenic tests sinking from surface water. This deposition is then modified through reworking and redeposition of sediment by contour currents of varying intensity (Pinet et al., 1981; Paull, Matsumoto, Wallace, et al., 1996). Lithologic variations at Site 994 may be caused by fluctuations in bottom-current intensity, terrigenous output, or carbonate and biogenic silica productivity.
The diatom-rich interval at Site 994 was deposited during the late Pliocene (between about 3.5 and 2.7 Ma) (Paull, Matsumoto, Wallace, et al., 1996). This corresponds to a time of global warming and an acceleration of thermohaline circulation, especially in the North Atlantic. Several locations from around the world show an increase in diatom abundance at this time interval, presumably because the change in oceanic circulation led to elevated nutrient fluxes in multiple regions (e.g., Dickens and Owen, 1996). We speculate that the increased abundance of diatoms at Site 994 is causally linked to the change in oceanographic conditions and surficial productivity rather than some local effect on sedimentation. The diatom-rich interval observed at Site 994 may, therefore, extend over a large area. Gas hydrate amounts may be higher than expected wherever this unit exists on the Blake Ridge; that is, the upper hydrate zone at Site 994 may be a regional phenomenon (although it would occur at different depths depending on sedimentation rate).
