106 cells mL-1.Previous analysis of bacterial distributions in deep sediment samples from six ODP legs in the Pacific (Parkes et al., 1994) and the Atlantic Oceans (Cragg et al., 1997) has demonstrated an exponential relationship between bacterial populations and sediment depth. The consistent bacterial/depth relationship found at other sites with widely differing oceanographic settings suggests that the deep bacterial biosphere is ubiquitous in marine sediments. This relationship also allows comparison with and interpretation of bacterial population data in more complex environments, for example, the sediment at Cascadia Margin (ODP Leg 146; Cragg et al., 1995), where tectonic activity results in the expulsion of large quantities of fluids and volatiles throughout the sediments, and the formation of gas hydrates in some areas.
The total bacterial populations in Leg 164 sediments are consistent with all previous ODP data (Fig. 5). Below ~10 mbsf, however, numbers are comparatively high, and often exceed the upper 95% prediction limit. Bacterial populations are particularly stimulated around the BSR (Fig. 5).
Organic carbon concentrations increased with depth in Leg 164 sediments. Consequently, there is no correlation between organic carbon and the total bacterial population at any of the three sites. Elevated organic carbon at depth may, however, contribute to higher bacterial numbers in the deeper layers of Blake Ridge sediments compared to other ODP sites previously analyzed (Fig. 5).
The deepest marine
sediment samples previously examined for bacterial populations were from ODP Leg
161 (Alboran Sea), at a depth of 647 mbsf (Cragg et al., in press). This study
extends the depth of known biosphere in marine sediments by a further ~100 mbsf;
the deepest sample collected during Leg 164, from 749 mbsf at Site 997,
represents the deepest marine sediment sample collected and examined as part of
a continuous depth sequence for bacterial populations to date. Even at this
depth, the bacterial population at Blake Ridge is substantial, at 1.8
106 cells mL-1.
With increasing amounts of hydrate at the three sites (Table 1) and the associated increases in free gas (Dickens et al., 1997), the relative number of cells involved in division increases. The frequency of dividing and divided cells increases from 9.5% at Site 994, through 10.8% at Site 995 to 13.1% at Site 997. This suggests that the presence of the hydrate/free-gas accumulations stimulates bacterial growth and is confirmed by increases in depth-integrated bacterial productivity (thymidine incorporation), which increases from 33 pmol m-2 per day at Site 994 to 198 pmol m-2 per day at Site 995 within the inferred hydrate zone (a complete data set for thymidine incorporation rates is unavailable for Site 997). Although thymidine incorporation is a reliable index of bacterial productivity in sediments (Wellsbury et al., 1996), conversion into absolute growth rates (Moriarty, 1990) is problematical (Robarts and Zohary, 1993; Wellsbury et al., 1993), and hence growth rates have not been calculated.
The total bacterial population increase, which occurs immediately below the inferred hydrate zone (Fig. 6), is associated with the zone of free gas trapped beneath the "frozen" hydrate (Dickens et al., 1997). Below the inferred hydrate zone, the population increases are such that total bacterial numbers are significantly higher than those immediately above the BSR (P < 0.005 and P < 0.002 at Sites 995 and 997). Thus, the population increases suggest an increased biological availability of methane within the free-gas zone.
Total bacterial populations in the samples of solid hydrate recovered from Site 997 (331 mbsf) were much lower than those predicted from a general regression of total populations and depth at Leg 164 sites, as it contained only 2.1% of the expected bacterial population. Thus, solid gas hydrate might physically exclude bacteria, or be colonized by limited numbers of specialized bacteria. Hence, increases and variable distributions in bacterial populations within the inferred hydrate zone (Site 995, 428 mbsf; Fig. 6) probably reflect the patchy distribution of solid hydrate and localized higher free gas concentrations that bacteria can utilize, similar to the situation just below the BSR.
In near-surface sediments at Blake Ridge, carbon flow is dominated by bacterial sulfate reduction (Fig. 2B, Fig. 3B), which occurs at rates up to four orders of magnitude higher than methanogenesis. Increasing sulfide concentrations (TRIS) and the rapidly decreasing interstitial water sulfate concentrations are both independent confirmations of sulfate reduction. In addition, depth distributions of viable populations of sulfate-reducing bacteria in near-surface Leg 164 sediments reflect this trend (Fig. 2A, Fig. 3A). Below ~20 mbsf, pore-water sulfate concentrations are depleted to values of <0.2 mM. Associated with this decrease is an increase in headspace methane, exceeding ~20,000 ppmv by depths of ~30 mbsf. As methane gas increases, so the rates of bacterial methane oxidation increase (to ~160 nmol mL-1 per day). Rates of methanogenesis also increase deeper in the sediment (Fig. 2B, Fig. 3B)
Potential rates of bacterial activity increase sharply around the base of the inferred hydrate zone and the free-gas zone beneath (e.g., Site 995, Fig. 7). Anaerobic methane oxidation, methanogenesis from both acetate and H2:CO2, acetate oxidation, sulfate reduction, and thymidine incorporation are all stimulated. These data clearly demonstrate that the sediments near and below the BSR form a biogeochemically dynamic zone, with carbon cycling occurring through methane, acetate, and carbon dioxide.
In sediments from Cascadia Margin, ODP Leg 146, increasing amounts of hydrate within the sediments resulted in increasing rates of anaerobic methane oxidation (Cragg et al., 1996; Cragg et al., 1995). At Blake Ridge, the largest quantities of hydrate and associated free gas were present at Site 997. Anaerobic methane oxidation rates in the "hydrate zone" at Site 997 reached a maximum of 109 nmol mL-1 per day, compared to a maximum of 19 nmol mL-1 per day at Site 995. Methane oxidation rates are clearly stimulated in the free-gas zone below the BSR (Fig. 7); thus, increasing amounts of free gas associated with increasing amounts of hydrate impact on methane oxidation rates. These data are in good agreement with the increases in bacterial populations directly within the free-gas zone (Fig. 6) and further substantiate the hypothesis that it is the presence of the free gas associated with the hydrate deposits, rather than the hydrates themselves, that impacts on microbial activity.
Sulfate-reducing bacteria have been implicated in anaerobic methane oxidation (Iversen and Jørgensen, 1985; Schulz, et al., 1994). Both sulfate reduction and methane oxidation were stimulated below the BSR (Fig. 7), but these processes are not closely correlated. The presence of deep interstitial water sulfate may be an indicator of seawater contamination during drilling (Paull, Matsumoto, Wallace, et al., 1996). However, anoxic sulfide oxidation could also produce sulfate (Bottrell et al., unpubl. data; Fossing and Jørgensen, 1990; Schulz et al., 1994). The electron acceptor(s) for this oxidation and methane oxidation might be supplied by fluid flow into the hydrate zone (P.K. Egeberg, pers. comm., 1998). In addition, other electron acceptors such as ferric iron and humic acids may also be important (Lovley et al., 1996).
The extraordinary depth profile of pore-water acetate at Site 995 (Fig. 3) demonstrates the importance of acetate in biogeochemical cycling in deep marine sediments (Fig. 7). Concentrations of acetate were surprisingly high, reaching ~15 mM at 691 mbsf, ~1000 times higher than "typical" near-surface concentrations. These values were determined by the use of two different, independent analytical techniques. The specific enzymatic HPLC technique was used (ion-exclusion chromatography) and independently confirmed using isotachophoresis on interstitial water samples from Site 997 (Wellsbury et al., 1997).
Rates of potential acetate metabolism in Site 995 sediment are extremely high, reflecting the very high pool size in the interstitial water. Actual turnover rates of labeled [1-(2)14C] acetate, however, were comparatively small (<0.4%/hr below 28 mbsf), so when multiplied by the large pool size this would result in an amplification of any 14C counting errors, which might result in some overestimates of turnover rates. Despite this possibility, acetate turnover is so large that it must be a major energy source for deep sediment bacteria, and consistent with this, the rate of labeled acetate turnover to methane, which is independent of the acetate pool size, increases in the deeper sections (e.g., rates of turnover of labeled acetate to methane increase by 85× between 500 and 600 mbsf). However, as the acetate concentration increases with depth, this means that acetate production must exceed consumption, and thus there must be a significant source of acetate within the deeper zones of both Site 995 and Site 997. Some of this acetate may be produced in situ, as demonstrated by Wellsbury et al. (1997), because temperature increases during burial will increase the bioavailability of organic matter. The increasing acetate concentrations are accompanied by increasing concentrations of other volatile fatty acids (e.g., formate, lactate and propionate [data not shown]) and are consistent with increasing concentrations of dissolved organic carbon at depth in Blake Ridge sediments (Egeberg and Barth, 1998). However, these rates of acetate metabolism are extremely high, and could not be sustained without influx of organic carbon into the sediment; hence in situ rates are likely to be lower than these potential rate measurements. However, Egeberg and Barth (1998) suggest that high dissolved organic carbon is supplied at Site 997 by upward migration of pore water, contributing significantly to the pool of metabolizable carbon. These increases in volatile organic acids may be specific to gas hydrate sites, explaining the high quantities of biogenic methane in the sediments.
Rates of acetate methanogenesis below the BSR are two to three orders of magnitude higher than H2:CO2 methanogenesis. Methane oxidation rates at the base of the hydrate zone at Site 995 are 10 times greater than H2:CO2 methanogenesis, similar to previous results from Leg 146, where methane oxidation also exceeded methanogenesis (from H2:CO2). In contrast, acetate methanogenesis at Site 995 exceeds methane oxidation through and below the BSR, which is consistent with high accumulations of methane below the BSR (Dickens et al., 1997).
