DISCUSSION

TOC concentrations of sediments from the 13 Leg 175 drill sites are greater than the average of 0.3 wt% given by McIver (1975) based on Deep Sea Drilling Project (DSDP) Legs 1–33, a value that can be considered representative of typical deep-sea sediments. The elevated TOC concentrations are probably a consequence of the combination of elevated paleoproductivities, which would have delivered abundant organic matter to the sediments, and the high accumulation rates of sediments, which would have improved preservation of the organic matter. Associated variations in CaCO₃ concentrations reflect varying combinations of changes in delivery of calcareous material, dilution by noncalcareous components, and carbonate dissolution fueled by oxidation of organic matter.

Sedimentary gases are elevated at all 13 drill sites on the southwest African margin. Both CO₂ and methane are abundant in the cored sediments. The most likely source of the CO₂ is from oxidation of some of the marine organic matter that is present in the sediments. The evolution of dissolved interstitial CO₂ from oxidation of organic matter involves a sequence of terminal electron acceptors—first interstitial oxygen, followed by nitrate, manganese (IV) oxides, iron (III) oxides, and finally sulfate—in the transition from oxic to suboxic to anoxic sediment conditions (Froelich et al., 1979; Schulz et al., 1994). Dissolved oxygen and nitrate tend to disappear rapidly with depth in marine sediments; sulfate is then used as a source of oxygen in generating CO₂. Oxidation continues deeper into the sediments until interstitial sulfate is exhausted. It should be recognized that only a small fraction of sedimentary organic matter needed to be oxidized to markedly increase dissolved interstitial CO₂ levels.

Two sources of the abundant methane are possible. First, hydrocarbon gases from deeper, thermally mature sediments may have migrated into the gas-rich zone. Evidence for migration of methane into porous sediments from deeper sources has been found at Sites 762 and 763 on the Exmouth Plateau, where a thermogenic source exists in underlying Jurassic rocks (Meyers and Snowdon, 1993). Mesozoic sources of thermogenic gas are indeed known on the African margin between the Niger delta and the Walvis Ridge. However, high C₁/C₂ ratios, which are representative of microbial gases (Claypool and Kvenvolden, 1983; Whiticar, 1996), and the absence of major contributions of higher molecular weight hydrocarbon gases (see individual site chapters, this volume) indicate that little, if any, of the gas in the sites drilled during Leg 175 is derived from the thermal degradation of organic matter.

A second possible source of methane, the dominant hydrocarbon gas in these sediments, is in situ formation by methanogenic microorganisms. Methanogenesis becomes important in sediments that lack interstitial sulfate but contain metabolizable organic matter. This process has two stages in which a consortium of microorganisms first converts the various forms of organic matter into short-chain alcohols and acids, CO₂, and H₂. In the second stage, some of these products are converted to methane by a metabolically limited group of obligate anaerobic microbes (Mechalas, 1981; Whiticar, 1996). The general reactions that represent the production of methane in the second stages are these:

CH₃COOH → CH₄ + CO₂ (acetic acid fermentation) and

CO₂ + 4H₂ → CH₄ + 2H₂O (CO₂ reduction).

The reduction of CO₂ to methane is the dominant process in marine sediments (Wiese and Kvenvolden, 1993: Whiticar, 1996).

At all 13 Leg 175 drill sites, a correspondence between the disappearance of interstitial sulfate and the increase in methane concentrations was found (see individual site chapters, this volume). A microbial origin of methane is supported by this observation, inasmuch as Claypool and Kvenvolden (1983) observe that the presence of interstitial sulfate inhibits microbial methanogenesis in marine sediments. Similar in situ microbial production of methane has been inferred for high gas concentrations in Pliocene–Pleistocene sediments from DSDP Site 532 nearby on the Walvis Ridge (Meyers and Brassell, 1985), Sites 897 and 898 on the Iberian Abyssal Plain (Meyers and Shaw, 1996), and also in middle Miocene sediments from Site 767 in the Celebes Sea (Shipboard Scientific Party, 1990). Moreover, Cragg et al. (1992) report the existence of viable microbes to depths of ~500 mbsf in the sediments of the Japan Sea. The abundance of microbial gases in sediments from Sites 1084 and 1085 suggests the similar presence of viable microbial communities to sub-bottom depths as great as 600 mbsf along the southwest African margin.

The origin of the methane probably involves a reduction mainly of the abundant interstitial CO₂ in the sediments and minor in situ microbial fermentation of marine organic matter. Because these two processes are carried out by distinct microbial populations, the abundance of methane ultimately depends on satisfying the metabolic requirements of both pathways. Comparison of the carbon-isotopic contents of CO₂ and methane in ODP Site 533 sediments has indicated that most of the methane originates from the reduction of CO₂ (Galimov and Kvenvolden, 1983). Inasmuch as the obligate anaerobes involved with CO₂ reduction are particularly metabolically limited (Mechalas, 1981; Whiticar, 1996), this pathway can become blocked, despite an abundance of interstitial CO₂.

Considerable differences in the concentrations of CaCO₃ and TOC exist in the sediments from the four depositional regimes, yet the gas profiles are similar. For example, CaCO₃ concentrations differ by about a factor of 10 in sediments from Sites 1076 and 1085, whereas the residual CO₂ and methane concentrations are virtually the same (Fig. 2, Fig. 5). Similarly, TOC concentrations in Site 1076 sediments are about twice those at Site 1985, yet this difference does not appear in the gas concentrations. The sediments having the highest concentrations of organic carbon, those at Site 1084, also have the highest concentrations of CO₂; this is understandable given that oxidation of organic matter leads to CO₂ production. Surprisingly, these same sediments have the lowest methane concentrations throughout much of the sequence (Fig. 4), which implies that the metabolic requirements of the CO₂-reducing microbes were poorly satisfied at this location. It may be significant that much of the sedimentary sequence from Site 1084 smelled strongly of organo-sulfur compounds, as opposed to simply H₂S. This observation may reflect fundamental biogeochemical differences that limited microbial CO₂ reduction at Site 1084.

Moreover, the general lack of correspondence between concentrations of carbonate carbon and organic carbon and concentrations of CO₂ and methane in sediments from the four depositional regimes implies that only a small percentage of organic matter is converted to gas. For example, even at saturation, methane represents only ~0.1% of the organic matter that is present in unconsolidated sediments (Whiticar, 1996). The limiting factor to microbial gas production is generally not the availability of organic matter, but the availability of terminal electron receptors.

Sedimentary sequences from the 13 Leg 175 drill sites typically did not contain enough gas to develop gas-expansion pockets until below the sub-bottom depth where interstitial sulfate was depleted. Concentrations of headspace methane often reached a maximum somewhat deeper than the sulfate depletion depth, but concentrations of headspace CO₂ commonly continued to increase with depth in the sequences (Fig. 2, Fig.3, Fig. 4). In these patterns, the concentrations of headspace CO₂ and methane were approximately the same. Compositions of gases from expansion pockets were usually similar to those of the headspace gases from the same cores. A significant difference, however, between the compositions of headspace and expansion-pocket gases was found in the sedimentary sequences from the Angola-Namibia upwelling regime. Methane typically was three to four times more abundant than CO₂ in the gas-pocket compositions of these sediments, whereas methane was at equal or lesser concentrations than CO₂ in the headspace samples.

Several possibilities may contribute to this difference between gas compositions. First, most of the sites in the Angola-Namibia upwelling regime were drilled to 600 m, which is deeper than those in the other depositional regimes. The solubility of methane increases rapidly with increasing pressure (Culberson and McKetta, 1951), which increases with sub-bottom depth. The sediments from deeper in the upwelling regime may have actually had higher in situ concentrations of methane. Second, interstitial alkalinity reaches very high levels in the sediments from Sites 1082 and 1084 (see individual site chapters, this volume). Much of this alkalinity increase involves the addition of CO₂ from organic matter degradation. This gas becomes incorporated into the bicarbonate-dominated inorganic carbon equilibria. When sediment cores are brought onto the JOIDES Resolution, in situ pressures are released. If abundant methane is dissolved in sediment pore waters, it bubbles out of solution. Elevated pressures also increase the solubility of CO₂, but incorporation of this gas into the carbonate equilibria delays its exsolution when pressures are released. The period of heating that is part of the headspace procedure compensates for the slower release of CO₂ from solution. The greater proportion of methane in the expansion-pocket gases than in the headspace gases may consequently reflect the difference in the mobilities and solubilities of CO₂ and methane, rather than their true in situ concentrations.

A third possibility exists. The two factors that make the mobilities of sedimentary CO₂ and methane different may systematically bias the headspace procedure results toward lower methane concentrations. Because the plug of sediment that is used in the headspace technique is not collected until the core is cut into 1.5-m sections some 10–15 min after arrival on deck, a significant fraction of the interstitial methane may exsolve, either escaping to the atmosphere in the space between the core and core liner or collecting in the gas pockets, where it is lost when they are punctured to minimize core expansion. A further consequence of this potential biasing of headspace analyses toward lower methane values is that gas-pocket analyses may be simultaneously biased toward higher methane concentrations.

Gas hydrate formation can greatly increase the concentration of gases present in sediments above the limit of their solubility in interstitial water. In gas hydrate, water crystallizes into a cubic lattice rather than the hexagonal lattice of pure water. Large volumes of low molecular weight gases—methane, ethane, CO₂, H₂S, and nitrogen —fit into the cubic lattice. Methane is easily accommodated into the three-dimensional framework of water molecules. As much as 170 volumes of methane (at standard temperature and pressure) can theoretically be accommodated by one volume of hydrate (Kvenvolden and McMenamin, 1980). Actual accommodation values commonly are less, as exemplified by hydrate recovered from the Peruvian margin at ODP Sites 685 and 688, in which ~100 volumes of methane evolved from one volume of solid hydrate (Kvenvolden and Kastner, 1990). Gas hydrate forms under specific pressure and temperature combinations that are satisfied at water depths as shallow as 500 m at temperatures of 4° to 6°C (Claypool and Kvenvolden, 1983). The presence of gas hydrate is often inferred from acoustic BSRs in the seabed, and BSRs were detected near some of the planned Leg 175 drill sites during precruise seismic surveys.

The presence of substantial amounts of gas in the Lower Congo Basin and Angola margin depositional regimes was implied by features in the seismic traces made during predrilling site surveys. Drill sites were ultimately selected to avoid what appeared to be gas chimneys, but BSRs were so extensive that they were unavoidable. After two sites (six holes) had been drilled in which little direct evidence of gas hydrate had been found, a special test was conducted to search for the existence of hydrate layers. Headspace gases were sampled at 3-m intervals between 100 and 130 mbsf at Hole 1077B to detect the presence of methane hydrate that might be responsible for a conspicuous seismic reflector seen at this level. No extraordinary amounts of methane were found; headspace concentrations ranged between 7,000 and 41,000 ppmv. Furthermore, none of the frothing of sediments, the disruption of sediment structures, the frosting of the outside of the core liner, or the dilution of interstitial chloride that accompanies the endothermic decomposition of hydrate and the release of its contents (Kvenvolden and Barnard, 1983) were observed. It was concluded that despite the common occurrence of BSRs, gas hydrate is not strongly developed along this part of the southwest African margin.

A possible alternate explanation for the conspicuous seismic reflectors on this margin is that they represent lenses of dolostone. These rocky layers were encountered at several sites north of the Walvis Ridge, and they presented obstacles to drilling. For example, the two holes at Site 1080 were terminated prematurely at 37 and 52 mbsf, respectively, because of drilling refusal from these layers. The dolostones could be formed by precipitation of diagenetic dolomite as a result of alkalinity increases that follow in situ oxidation of organic matter, with consequent oversaturation of interstitial bicarbonate.