DISCUSSION

Precipitation of Carbonates

Mixing of sulfate- and methane-bearing fluids is occurring at Site 996 (Paull, Matsumoto, Wallace, et al., 1996) and other vent locations (Suess and Whiticar, 1989; Borowski et al., 1996; Wallmann et al., 1997). Sulfate gradients with sulfate concentrations approaching 0 mM near the seafloor (0.10-6.95 mbsf; Paull, Matsumoto, Wallace, et al., 1996) suggest that sulfate depletion is driven by methane flux from below (Borowski et al., 1996) and that anaerobic methane oxidation is the principal sulfate-consuming process at Site 996. The upward methane flux, which is required to sustain these reactions above the Blake Ridge Diapir, is most likely occurring along fluid conduits that were observed in seismic reflection data and in the form of vertical, gas hydrate-cemented veins in the sedimentary section.

The authigenic carbonate cements at Site 996 are primarily composed of aragonite (Table 1; Fig. 5), which is consistent with authigenic aragonite formation at other vent sites (e.g., Hovland et al., 1987; Matsumoto, 1990; Jørgensen, 1992; Savard et al., 1996). Temperature, the degree of pore-fluid supersaturation, and the presence of oxic or anoxic conditions seem to be the major controlling factors for the precipitation of different carbonate minerals (Burton and Walter, 1987; Hovland et al., 1987; Burton, 1993). The relative importance of those mechanisms, however, is still poorly understood. Elevated temperatures generally seem to favor aragonite precipitation, whereas changes in saturation state seem to have little effect. There have been no definitive means (petrographic, isotopic, or geochemical) for distinguishing between carbonate cements formed under oxic or anoxic conditions. Although aragonite formation has generally been associated with oxic environments at or near the seafloor (Longman, 1980; Hovland et al., 1987, Matsumoto, 1990), the co-occurrence of aragonite and authigenic pyrite in almost all Site 996 samples (Fig. 3F) is difficult to reconcile with oxic conditions, because pyrite can only form in anoxic, sulfate-reducing environments. Thus, the aragonite cements at Site 996 must have formed very near the sediment/water interface in an oxygen-depleted environment where the generation of HCO₃^- by anaerobic methane oxidation resulted in oversaturation with respect to aragonite.

The darker rims on many of the intraclasts may be dissolution rinds that contain less CaCO₃. The interaction of sulfide-rich pore fluids with oxygenated seawater (HS^- + 2O₂ SO₄^-2 + H⁺) produces locally acidic water (Paull and Neumann, 1987), generating dissolution in carbonates at the sediment/water interface. The rough and partially dissolved surfaces of aragonitic bivalve shells also indicate exposure (either syn- or postdepositionally) to a locally corrosive environment at the seafloor (Fig. 4C). The complex juxtaposition of dissolution surfaces with cemented sediment suggests that the position of the oxic-anoxic interface may fluctuate over time, perhaps in response to variations in the flux of methane from below.

Controls on the Geochemical and Isotopic Composition of the Authigenic Carbonates

The depletion in ¹³C of the authigenic carbonates at Site 996 (¹³C as low as -48) indicates that CO₂ derived from methane oxidation was the primary carbon source for these carbonates. Based on methane ¹³C values of -62 to -72 at Site 996 (Paull et al., Chap. 7, this volume), we estimate that up to 75% of the carbon incorporated into authigenic carbonates at this site was derived from biogenic methane.

Measurements of the oxygen isotopic composition of pore water of near surface sediments above the Blake Ridge Diapir revealed ¹⁸O values of 0.1 to 0.9 with a mean of 0.3 (Borowski et al., 1997). Using the ¹⁸O_aragonite/T relationship established by Hudson and Anderson (1989), the ¹⁸O values of the aragonite cements indicate a precipitation temperature of about 4ºC. This is in good agreement with the observed bottom-water temperature of 3.5ºC at Site 996 (Paull, Matsumoto, Wallace, et al., 1996). It is therefore reasonable to infer that these cements precipitated in, or near, isotopic equilibrium with the regional bottom water.

Carbon Isotopic Composition of Pore Fluids and Authigenic Carbonates

The carbon isotopic composition of CO₂ in near-surface sediments at Site 996 varies between -28 and -45 (Paull et al., Chap. 7, this volume). Such negative ¹³C values are typical for concurrent microbial-mediated oxidation of sedimentary organic matter and methane originating from depth (Suess and Whiticar, 1989). As indicated by ¹³C values of -62 to -72, methane that is consumed at Site 996 is predominately biogenic in origin (Paull et al., Chap. 7, this volume). As authigenic carbonates derive their carbon from the pore-water CO₂ pool (Suess and Whiticar, 1989), the similarity between the ¹³C values of the carbonates (-30.5 to -48.4) throughout the sedimentary section and pore-water CO₂ (-28 to -45) at shallow depth indicates that the carbonates have formed from this shallow CO₂ pool. Furthermore, carbon isotope values of CO₂ at Site 996 are most negative near the sediment/water interface and become more positive with depth (Fig. 8). The ¹³C values of recovered authigenic carbonates, however, do not show such a trend (Fig. 9). This dissimilarity again suggests that carbonate nodules recovered from greater depth probably formed near the seafloor, from a ¹³C-depleted carbon pool similar to the one that currently exists.

There are also indications that the carbon isotopic composition of the CO₂ pool at any one spot changes over time. An isotopic profile of ¹³C values across an aragonite-cemented vein within one of the carbonate nodules (Fig. 10) shows a symmetric trend towards more ¹³C-depleted carbon isotope values from the rim to the center of the vein. This trend can be explained as the result of a shift in the methane pool from less ¹³C-depleted to strongly ¹³C-depleted methane, possibly in response to changes in flow rate (Suess and Whiticar, 1989) or recycling of light carbon in the shallow methanogenic zone (Borowski et al., 1997). Alternatively, this may reflect local changes in the relative importance of the different carbon sources to the pore fluids.

Oxygen Isotopic Variations

As a first approximation, ¹⁸O values of the carbonate nodules indicate that they precipitated in or near isotopic equilibrium with the regional bottom water. Closer inspection, however, shows that the oxygen isotopic composition of carbonates recovered from shallow depth (0-3 mbsf) differs from those recovered from deeper in the sediment (Fig. 6, Fig. 9). This difference in the oxygen isotopic composition is roughly 1 (Fig. 6). Although currently highly speculative, this difference in ¹⁸O might be related to changes in bottom-water conditions, corresponding to glacial and interglacial time periods. However, the relatively poor depth and age control on the carbonate precipitates does not allow a detailed comparison to established Pleistocene ¹⁸O records.

Strontium Isotope Ratios

Secular variations in the strontium isotopic composition of seawater are well known (e.g., Elderfield, 1986; Farrell et al., 1995). The strontium isotopic composition of carbonate minerals will reflect the strontium isotopic composition of the water in which they formed (Hess et al., 1986). Thus, strontium isotopes can be used to constrain the source of fluids for authigenic carbonate precipitation in marine sediments (Sample and Reid, 1998). To evaluate the source of fluids at this site, we compared the ⁸⁷Sr/⁸⁶Sr values of authigenic carbonates to the ⁸⁷Sr/⁸⁶Sr values of pore fluids in nearby sedimentary sections not influenced by venting fluids (Sites 994, 995, and 997).

Because all of the samples that were recovered from Site 996 are Pleistocene in age (Paull, Matsumoto, Wallace, et al., 1996), only a nominal variation in the Sr isotopic composition of the authigenic carbonates is to be expected if the carbonates have formed at or near the seafloor in contact with Pleistocene bottom waters. If, however, there has been addition of strontium from depth, one would expect a different Sr isotopic composition because pore waters from greater depth are known to vary in this area. At Sites 994, 995, and 997 on the Blake Ridge, the Sr isotopic composition at and below the base of gas hydrate stability ranges from 0.709006 to 0.709043 (Fig. 7). If the source of the Sr captured in the authigenic carbonates was being carried in fluids moving upward along a fault above the Blake Ridge Diapir, the ⁸⁷Sr/⁸⁶Sr values should reflect this deeper pool. At Site 996, however, pore-water Sr values are clearly indistinguishable from measurements of modern (0.709175) and contemporaneous seawater (Farrell et al., 1995). A comparison with Farrell's seawater curve (Fig. 7) also shows that the ⁸⁷Sr/⁸⁶Sr values of carbonate precipitates recovered throughout the sedimentary section are approximately consistent with the expected ⁸⁷Sr/⁸⁶Sr composition based on the age of their host sediment. Thus, our data do not indicate an addition of Sr from deep-seated fluids at Site 996.

The Dolomite Nodule

A single dolomite nodule, with a fabric similar to our "type 3" authigenic carbonate (Fig. 3E), was recovered from 51.6 mbsf in Hole 996E. Dolomite (composed of 45 mol% Mg) occurs as microcrystalline cement. The isotopic characteristics of this nodule suggest an origin distinct from the rest of the authigenic carbonates in this section. The ¹³C values of the dolomite nodule (-13.1 to -19.2) are significantly more enriched in ¹³C than those of authigenic aragonite recovered elsewhere in the sedimentary section. Furthermore, the ¹⁸O values (4.8-5.4) indicate a temperature of formation significantly higher (18º-24ºC, using the ¹⁸O/temperature relation of Northrop and Clayton [1966]), than those of the other authigenic carbonates at this site. Although we could assume that the dolomite nodule formed from ¹⁸O-depleted water (estimated to be -4.0 to -4.5 standard mean ocean water [SMOW], based on Northrop and Clayton, [1966]), oxygen isotope measurements of pore water at Site 996 (Egeberg, Chap. 22, this volume) do not indicate such light ¹⁸O values.

Although authigenic dolomites have been described from various vent locations (e.g., Matsumoto, 1990; Jørgensen, 1992; Kopf et al., 1995), we do not know at this point if the dolomite nodule at Site 996 is directly related to the venting process. The isotopic similarity of the nodule to dolomite found elsewhere on the Blake Ridge at Sites 994, 995, and 997 (Rodriguez et al., Chap. 30, this volume), and the fact that it is the only sample of its kind at Site 996, suggests an origin independent of fluid venting.

CaCO₃ Precipitation at Site 996

Our data indicate that precipitation of authigenic carbonates at this site took place at (or near) the seafloor and that continued carbonate formation with increasing sediment depth has not occurred. This inference is supported by the observation that precipitates do not show any significant petrographic or mineralogical changes with depth and that aragonite is the primary authigenic carbonate at Site 996. In addition, both microscopy and electron microprobe analysis do not provide evidence for multistage cementation or changes in the mineral chemistry with depth (mineralogical changes would be expected if nodules grew, or continued to grow, at several depth intervals). Furthermore, ¹³C values of the authigenic carbonates (-30.5 to -48.4) are similar to the carbon isotopic composition of pore water CO₂ at shallow depth (-28 to -45). At greater depths, ¹³C values of CO₂ become progressively ¹³C enriched (+10.5 at 43 mbsf), demonstrating that the authigenic carbonates could not have formed in equilibrium with bicarbonate at ambient depth. If these observations can be extrapolated to other vent sites, precipitation of authigenic carbonate at vent sites takes place only at or near the sediment/water interface, and the fluid conduit itself will not be distinctly marked. This means that only the seafloor expression of vent sites will be preserved (as carbonate crusts) and that fossil fluid conduits will be difficult to find in the sedimentary record.

The occurrence of authigenic carbonates at several depth horizons, therefore, can be used to estimate the duration, and lateral shifting, of fluid venting above the Blake Ridge Diapir. Given that the deepest aragonite nodule was recovered from about 30 mbsf, and assuming a constant sedimentation rate of ~48 m/m.y. (Paull, Matsumoto, Wallace, et al., 1996), we can infer that seepage has been going on in this area for at least 600,000 yr. Variations in the oxygen isotopic composition (which might be related to glacial and interglacial periods) support this inference.