RESULTS AND DISCUSSION

Interstitial Waters

Site 994

The ¹⁸O values of the interstitial waters from Site 994 show a fairly smooth variation with depth with values ranging from 0.30 to -0.37 (Table 1, Fig. 2A). The ¹⁸O profiles show consistent characteristics within three depth zones: an upper (top 250 mbsf), a middle (250-500 mbsf), and a lower unit (500-700 mbsf). The middle unit roughly corresponds to the gas hydrate zone (Paull, Matsumoto, Wallace, et al., 1996). In the upper unit, ¹⁸O values decrease with depth from ~0.3 to -0.4, whereas values of the lower unit increase slightly from -0.2 to reach between 0 and 0.2. In the middle unit, ¹⁸O values increase with depth, showing a sharp drop at 456.25-485.93 mbsf, which coincides with the bottom-simulating reflector (BSR). The ¹⁸O_H2O values from samples within the gas hydrate zone represent mixing of water derived from dissociated gas hydrate and from pristine in situ pore water. To evaluate the meaning of the ¹⁸O_H2O data, one must be able to separate these signals, so we have constructed a regression curve representing the ¹⁸O_H2O of the in situ pore water only (dotted curve, r² = 0.9905, Figure 2A). This curve is based on best-fit polynomial function using only data above (2 data points) and below (6 data points) the gas hydrate zone. Apparent positive shifts in ¹⁸O_H2O from the baseline may be related to the existence of gas hydrate in these sediments, as gas hydrate should yield ¹⁸O-enriched water to the pore waters.

Site 997

The ¹⁸O of the interstitial water samples from Site 997 range from 0.25 (5.8 mbsf) to -0.54 (347.95 mbsf) (Fig. 2B). The overall trend is similar to that of Site 994, decreasing in the top 200 m and increasing in the lower unit below ~500 mbsf. The regression for the ¹⁸O values of the pristine pore water was constructed based on 7 data points above and 3 data points below the gas hydrate zone (r² = 0.9890, Fig. 2B). The curve indicates that the pristine pore waters of gas hydrate zone are significantly depleted in ¹⁸O. Values from samples of the gas hydrate zone are highly variable, and the positive excursions are much larger relative to those of Site 994. The magnitude of positive shifts in gas hydrate zone is 0.3-0.6, almost twice as large as those of Site 994.

Oxygen Isotopic Anomalies vs. Chloride Anomalies

The dilution of pore waters by freshwater derived from dissociating gas hydrate decreases the chloride concentration in recovered interstitial waters. The difference between measured chloride concentration in recovered samples and that of calculated, in situ values is the chloride anomaly of gas hydrate-zone samples (Paull, Matsumoto, Wallace, et al., 1996). The ¹⁸O anomalies are defined as the difference between ¹⁸O of recovered interstitial waters and ¹⁸O of pristine in situ pore waters, which is given by the polynomial fits of the respective sites (Fig. 2A, B). If the gas hydrate dissociation is also responsible for the deviation from the baseline ¹⁸O values by adding ¹⁸O to pore waters, ¹⁸O_H2O anomalies should be correlative with chloride anomalies. To test this possible correlation, both anomalies are shown in Figure 3. The occurrence and magnitude of chloride and ¹⁸O anomalies are generally correlative. For example, the largest measured ¹⁸O anomaly (451.64 mbsf) is directly correlative with a large chloride anomaly. The covariation between chloride and ¹⁸O anomalies is also shown in Figure 4, where a least-squares linear regression gives a correlation coefficient (r²) of 0.76.

Gas Hydrates

Results of ¹⁸O measurements of water derived from gas hydrate samples are given in Table 2 along with gas and water volumes, and chloride concentrations. These ¹⁸O values were corrected for mixing with interstitial water, probably coming from sediment adhering to the gas hydrate samples. Gas hydrates are assumed to contain only freshwater (e.g., Hesse and Harrison, 1981), whereas interstitial waters are assumed to have a baseline value of chloride (Paull, Matsumoto, Wallace, et al., 1996).

Chloride concentrations of the interstitial waters are given in Figure 5. Chloride concentrations within the interval 200-450 mbsf (gas hydrate zone) are highly irregular, with sharp spikes toward lower concentration values. These chloride fluctuations are interpreted to represent differing amounts of gas hydrate contained within the sediment samples, which contribute freshwater to sediment pore waters as they dissociate during core recovery and processing. Paull, Matsumoto, Wallace, et al. (1996) calculated the in situ chloride concentration through the gas hydrate zone using a polynomial best fit to chloride concentration data above and below the gas hydrate-containing zone (dashed line in Figure 5). Baseline chloride concentration values are 505 mM and 509 mM for Site 994 and 997, respectively.

Because gas hydrate should be composed of freshwater, chloride values above 0 mM represent introduction of pore water of varying amounts into the gas hydrate water. Chloride concentrations of five water samples derived from gas hydrate are 5 to 62 mM (Table 2). Using the in situ chloride concentration values listed above, the mole fraction of gas hydrate water in water samples is 0.990-0.878. Correcting for pore water contamination, ¹⁸O values of gas hydrate water is 2.67 for Section 164-994C-31X-7 and 2.82-3.51 (mean = 3.2) for Section 164-997A-42X-3. Note that the isotopic value for Site 997 samples differs by about 20%.

Oxygen Isotopic Fractionation Between Gas Hydrate and Ambient Water

The ¹⁸O of interstitial waters and gas hydrates at Sites 994 and 997 are collectively shown in Figure 6. The difference in ¹⁸O values between gas hydrate and the ambient water (¹⁸O_GH-IW) is 3.1 at Site 994 and 3.3-3.8 (mean = 3.6) at Site 997. Given that recovered massive gas hydrates were in isotopic equilibrium with ambient interstitial waters (T = 12º-16ºC and P = 31 MPa), the equilibrium isotopic fractionation factor (_GH-IW) at each site is 1.0034 (Site 994) and 1.0037-1.0040 (Site 997). For comparison, the oxygen-isotopic fractionation of ice-water is 1.0027-1.0035 (O' Neil, 1968; Craig and Hom, 1968; Jakli and Staschewski, 1977) and that of the THF hydrate-water association is 1.00268 ± 0.00003 at 0°-4ºC (Davidson et al., 1983).

The gas hydrate occurrence depth does not necessarily indicate that gas hydrate formed at that depth. The hydrate may have been formed at a shallower depth and subsequently buried deeper within the sediment column. Thus, recovered gas hydrates are not necessarily in isotopic equilibrium with the ambient pore waters. A wide range in ¹⁸O values for Site 997 gas hydrate seems to suggest that the gas hydrate formed over a long time and incorporated water from temporally and spatially changing pore waters. Alternatively, the massive gas hydrate at Site 997 may have formed in fractures associated with upward-migrating water enriched in ¹⁸O.

Gas Hydrate Amounts

As mentioned above, ¹⁸O anomalies of interstitial waters of the gas hydrate zone are explained as the results of mixing of pristine pore water with isotopically heavy water released by the dissociation of gas hydrate. Assuming that the Blake Ridge gas hydrate has a stoichiometric composition (CH₄ · 5.75 H₂O; M = 16.0 + 103.5 = 119.5 g) and the density of methane hydrate is 0.91 g/cm³ (MacDonald, 1990; Makogon, 1997), the volume ratio of gas hydrate water to gas hydrate is 103.5 : 119.5/0.91 = 0.79 : 1. Using the gas hydrate-water fractionation (¹⁸O_GH-IW) value of 3.1 at Site 994 and 3.6 at Site 997, which were derived above, the relationship between the measured ¹⁸O_H2O values (¹⁸O_M) and the pristine pore water (¹⁸O_P) at Site 994 is expressed as:

(1 - 0.21X) ¹⁸O_M = 0.79 X (¹⁸O_P + 3.1) + (1 - X) ₁₈O_P, (1)

where X is the pore saturation of gas hydrate, that is, the fraction of pore space occupied by gas hydrate, and (1 - 0.21X), 0.79X, and (1 - X) are mole fraction of mixed water, gas hydrate water, and pristine pore water, respectively. Solving Equation 1 for X gives:

X = (¹⁸O_M - ¹⁸O_P)/[2.4 + 0.21 (¹⁸O_M - ¹⁸O_P)]. (2)

For Site 997, Equation (2) should be,

X = (¹⁸O_M - ¹⁸O_P)/[2.8 + 0.21 (¹⁸O_M - ¹⁸O_P)]. (3)

Gas hydrate amounts (pore saturation%) in sediments are calculated and given in Figure 7 along with gas hydrate estimates from chloride anomalies (Paull, Matsumoto, Wallace, et al., 1996).

At Site 994, five samples in the gas hydrate zone yield a gas hydrate saturation of 6.2%, which is almost twice as much as the estimates based on chloride anomalies (Paull, Matsumoto, Wallace, et al., 1996). At Site 997, the gas hydrate amount shows a rapid increase to 16% in the interval 230-340 mbsf in the gas hydrate zone, and a sharp drop at 348 mbsf, before reaching a maximum value of 25% at 451 mbsf just above the BSR depth. The average amount of gas hydrate is around 12%, which is also about two times greater than the amount based on chloride anomalies.

The observed discrepancy between the two estimates is not great but not insignificant. The possible causes of the discrepancy are (1) uncertainty in the determination of baselines (in situ trends) of both ¹⁸O and chloride concentration, (2) differential chemical effects between ¹⁸O and Cl^- during mechanical squeezing (Cave et al., 1998), and (3) sample deterioration, which may have occurred during storage of water samples. The baseline factor is thought to be most critical, because its calculation may involve large uncertainties. Also, the sampling density may partially explain the apparent discrepancy. Samples for chloride analyses were more densely spaced than those for isotope analyses, plus they were taken in gas hydrate-free sediments, whereas sparsely spaced ¹⁸O samples were obtained mainly from hydrate occurrences and therefore did not unambiguously detect hydrate-free horizons within the gas hydrate zone.

Determining the amount of methane gas hydrate stored in marine sediments is of paramount importance. Assessing the environmental impact of marine gas hydrates is wholly dependent on knowing gas hydrate amounts and occurrence. Moreover, any assessment of the resource potential of gas hydrates is impossible without knowing the amount of gas hydrates in the target areas. Several alternative approaches toward gas hydrate estimates include geophysical surveys, well logging, and sediment temperature measurements, but the geochemical approach is likely to give the most reliable quantitative estimates presently. These geochemical methods are dependent on the chemical composition of pore waters unaffected by sampling and recovery; therefore, it is crucial to develop new technologies to recover in situ pore waters from gas hydrate-bearing sediments. In this way, local estimates of gas hydrate amounts may be refined, and the global inventory of marine gas hydrate can be incrementally improved.