RESULTS AND DISCUSSION

Figure F2 shows the chromatograms of two total lipid fractions (Samples 175-1077A-9H-2, 75-77 cm, and 175-1084A-12H-6, 140-143 cm), with compounds listed in Tables T2 and T3. A wide variety of biomarkers from both terrestrial plants and marine sources such as algae, (cyano) bacteria, and archaea is present in the samples. These biomarkers will be discussed according to compound class.

Terrigenous Lipids

n-Alkanes

n-Alkanes in the studied sediments range in carbon number from 18 to 35, with C₂₅-C₃₃ n-alkanes the most dominant homologues (Fig. F3). A strong odd-over-even carbon number predominance (carbon preference index [CPI]) (CPI_25-33 = 3-10) (Table T4) was observed in all samples, indicating a predominantly terrigenous origin of the n-alkanes. These distributions resemble those of n-alkanes from leaf waxes of higher plants (Kolattukudy, 1976; Eglinton and Hamilton, 1967) and in eolian dust samples (Gagosian et al., 1981, 1987; Simoneit et al., 1977), supporting a terrigenous origin.

Concentrations of n-alkanes in sediments from Sites 1076 and 1077 are much higher than in sediments from the southern sites (see Table T4). These differences can probably be related to the different oceanographic and climatologic settings for both regions. The northern sites (1076, 1077, and to a minor extent also 1079) receive river input of terrigenous organic matter, whereas the terrigenous compounds in the southern sites are mainly transported via the wind and diluted by a high production of marine lipids resulting from upwelling activity. The variations in concentrations of leaf wax components at the southern sites possibly result from changes in the strength and direction of the trade winds, influencing the strength of marine upwelling and, therefore, the relative dilution of terrigenous by marine lipids.

At the northern sites (Sites 1076, 1077, and 1079), the n-alkane distributions maximize at n-nonacosane (n-C₂₉, structure I) (see Fig. F4 for structures). In samples from Sites 1082, 1084, and 1085, n-untriacontane (n-C₃₁) is most abundant. The chain length distributions of the n-alkanes could thus reflect different sources of the terrigenous compounds in both regions. The average chain length (ACL) of the C₂₅-C₃₃ n-alkanes changes from ~29.6 in the northern sediments to ~30.4 at the southern sites (Table T4). It has been suggested that plants produce longer-chain compounds in warmer climates (e.g., Poynter et al., 1989). The observed shift in our limited data set contradicts this interpretation. At present, vegetation in southern Africa consists predominantly of savanna-type grasslands, whereas in the more northern, tropical parts, lowland rainforest and Afromontane forest dominate. Therefore, an influence of vegetation type on chainlength of terrigenous leaf lipids seems more likely. Leaf lipids derived from grasslands may on average have longer chain lengths than do leaf lipids from plants in rainforests (Cranwell et al., 1973). Obviously, the limited sample number of this study prevents any further conclusions on this subject, which will be addressed in future research.

Fatty Acids and n-Alcohols

Fatty acids in the sediments range in carbon number from C₁₆ to C₃₀with only small amounts of C₁₆-C₁₈ fatty acids. n-Alcohols occur in the carbon number range of C₂₀-C₃₀. The highest concentrations of fatty acids and n-alcohols occur in the sediments close to the Congo River, where also the largest variation could be detected (Table T4). The lowest concentrations of both compound classes are present in the sediments from the Walvis Basin (Site 1082). Even carbon numbered long-chain compounds dominate the distributions of both compound groups. The dominance of the long-chain compounds, together with the observed strong even-over-odd carbon number predominance (for both classes, CPI_22-28 = 5-11), indicates a terrigenous origin for fatty acids and n-alcohols. Series of long-chain fatty acids and n-alcohols with a strong even carbon number predominance are characteristic constituents of surface waxes of higher plants (Kolattukudy, 1976). Therefore, the biological source of these compounds is closely related to that of the leaf wax-derived n-alkanes. Similar to the n-alkanes, the high concentrations of fatty acids and wax alcohols in the Congo plume sediments suggest a predominant fluvial supply, whereas the smaller amounts at other sites are more readily explained to be derived from dust (Gagosian et al., 1981, 1987; Simoneit et al., 1977). An algal origin of the long-chain fatty acids, however, cannot be completely excluded (Volkman et al., 1998). Marine microalgae are not a major source of long-chain n-alcohols but may contain minor amounts of these compounds (for a review see Volkman et al., 1998). However, the strong covariance of the concentrations of fatty acids and n-alcohols with the concentrations of n-alkanes, together with their high CPI values, indicate that autochthonous sources of fatty acids and n-alcohols are of minor importance. Higher plant leaf waxes are the main source.

Pentacyclic Triterpenoids

Two pentacyclic triterpenoid alcohols were detected in samples from Sites 1076, 1077, and 1079 (see Fig. F2). They were identified as taraxerol (taraxer-14-en-3-ol; Structure II in Fig. F4) and -amyrine (olean-12-en-3-ol; Structure III in Fig. F4) by comparison with mass spectra published by Killops and Frewin (1994) and relative retention times. The largest quantities were found in sediments located below the Congo River plume (Sites 1076 and 1077) and a smaller amount in sediments from the Angola margin (Site 1079). In samples from the Congo, the triterpenoid concentrations varied strongly, and they were below the detection limit in samples from the southern Sites 1082, 1084, and 1085. Another triterpenoid compound present in trace amounts in samples from Sites 1076, 1077, and 1079 is friedelan-3-one (Structure IV in Fig. F4), a triterpenoid ketone.

All three pentacyclic triterpenoids (Structures II-IV in Fig. F4) are widespread in higher plants (Beaton et al., 1955) and therefore represent a terrigenous source. Their occurrence has been reported in several marine sediments (e.g., Brassell and Eglinton, 1983; Volkman et al., 1987). -Amyrine has been isolated from several species of mangroves, including some species of Rhizophora (Ghosh et al., 1985). In Rhizophora mangle, it was found to be a major component of the epicuticular wax of the leaves, whereas taraxerol was largely bound in the cutin fraction of the leaves (Frewin et al., 1993; Killops and Frewin, 1994). Another mangrove species, Avicennia germinans, contains smaller amounts of -amyrine and taraxerol (Killops and Frewin, 1994). At present, the dominant mangrove species in the Congo area is Rhizophora racemosa (Moguedet, 1980). We suggest therefore, that this species, or its predecessor, is the major source of -amyrine and taraxerol in the sediments. The strong variation in triterpenoid content between the samples could thus give an indication of the proximity and the extent of mangrove vegetation in the Congo area.

Interestingly, none of the pentacyclic triterpenoid compounds could be detected in sediments from sites south of the ABF. This is in agreement with the present restriction of mangrove swamps to humid, tropical environments north of the ABF. Furthermore, it strongly favors the hypothesis of fluvial rather than atmospheric transport of mangrove lipids. The northward-flowing Benguela Coastal Current could limit southward transport of lipids at the ABF.

Marine Lipids

Long-Chain Alkenones

All sediment samples contain C₃₇ and C₃₈, and to a lesser extent C₃₉ and di- and triunsaturated methyl and ethyl ketones (Structure V in Fig. F4) (de Leeuw et al., 1980). The highest concentrations of these compounds were found in samples from Site 1084 (up to 1400 µg/g TOC), but the highest variability was detected in samples from Site 1076, where the maximum concentration was slightly smaller. Concentrations of alkenones in sediments from other sites are much smaller in range and absolute amount. These lipids are exclusively biosynthesized by haptophyte algae like Emiliania huxleyi and Gephyrocapsa spp. (Volkman et al., 1980, 1995; Marlowe et al., 1984) and were detected in numerous marine sediments (e.g., Brassell et al., 1986a), including the South Atlantic. Changes in coccolithophorid assemblages seem not to have affected alkenone production (Müller et al., 1997). Sedimentary alkenone concentrations have also been used as an indicator for marine primary production (Villanueva et al., 1998; Schubert et al., 1998; Hinrichs et al., 1999). The U^k₃₇´ index, the ratio of the diunsaturated to the sum of the di- and triunsaturated C₃₇ alkenones, is extensively used in paleoceanography as a temperature proxy. It has been shown that the U^k₃₇´ index is strongly correlated to sea-surface temperature (e.g., Brassell, 1993). The calculated sea-surface temperature estimates using the temperature calibration for the South Atlantic by Müller et al. (1998) are given in Table T4. The lowest sea-surface temperature is reconstructed for samples from below the upwelling activity. This suggests upwelling of cold South Atlantic Central Water as the main cause. Highest temperatures and largest temperature variability were detected for samples from Site 1082 in the Walvis Basin. Because of our limited data set, we regard any suggestion on the cause of this large range of temperatures as highly speculative. Alkenone concentrations are generally elevated during times of lower sea-surface temperatures (Table T4). This observation indicates that haptophyte algae are strongly dependent on upwelling events, which decrease sea-surface temperature and increase nutrient concentrations.

Alkyl Diols and Alkyl Keto-ols

Long-chain saturated C₂₈-C₃₂ alkyl diols (Structure VI in Fig. F4) were detected in all samples. The highest concentrations were found in samples from Sites 1076 and 1077, off the mouth of the Congo River (Table T4). Samples from Sites 1082 and 1085 contain only small amounts of these compounds (cf. Fig. F2), whereas samples from Sites 1079 and 1084 have intermediate quantities. The highest variability in alkyl diol content was also detected in samples from Site 1076. The major sources are probably microalgae of the class Eustigmatophyceae (Volkman et al., 1992; Volkman et al., 1999). The high concentrations of alkyl diols in sediments underlying the Congo River plume may point to a contribution by freshwater eustigmatophyte algae in the low-salinity waters (Volkman et al., 1999), although marine eustigmatophytes of the genus Nannochloropsis are also known to contain alkyl diols (Volkman et al., 1992). Alkyl diols have been reported to occur in sediments from various areas (de Leeuw et al., 1981; see review in Versteegh et al., 1997). Remarkably, samples from the Congo area predominantly contain the C₃₀ homologue, whereas the C₂₈ compound is the dominant alkyl diol in sediments from the southern upwelling region. A similar change was detected for the isomeric composition of the diols. 1,15-Isomers dominate diol distributions from sites north of the ABF, whereas sediments south of the ABF contain smaller amounts of the 1,15-diols but more of 1,14-diols. Versteegh et al. (2000) reported the same change in isomer composition in surface sediments and in sediments from the oxygen isotope Stage 5-6 transition from this area. The differences are most likely to be explained by different algal communities living under contrasting temperature and nutrient conditions.

C₃₀ and C₃₂ alkyl keto-ols (Structure VII in Fig. F4) are present in all samples. Their amounts do not exceed trace levels, except in sediments from Site 1079 on the Angola margin. In these sediments, the C₃₀ and C₃₂ alkyl keto-ols are major components. By comparison of chain length and isomer distributions of diols and keto-ols, Versteegh et al. (1997) concluded that both compound classes are diagenetically independent. Therefore, the high amounts of keto-ols in Angola margin sediments seem to reflect a specific planktonic source of an as-yet-unidentified nature.

Midchain Hydroxy Fatty Acids

In sediments from Site 1084, C₂₆ and C₂₈ midchain hydroxy fatty acids are present in relatively large amounts. At Site 1082, they occur in trace amounts, whereas they could not be detected at other sites. Both compounds are dominated by the 12-hydroxy-isomer. The C₂₈ 12-hydroxy fatty acid (Structure VIII in Fig. F4) has already been detected in organic matter-rich sediments, like Mediterranean sapropels (ten Haven et al., 1987), sediments from the Oman margin (ten Haven and Rullkötter, 1991), those from the Arabian Sea (Versteegh et al., 1997; Prahl et al., 2000), and from K/T boundary sediments (Yamamoto et al., 1996). Their phytoplanktonic origin was indicated by occurrence in sediment trap samples from high productive Arabian Sea waters (Prahl et al., 2000). Midchain hydroxy fatty acids have been reported from the genus Nannochloropsis (Gelin et al., 1997) but only homologues C₃₀. Therefore, organisms different from the Nannochloropsis investigated must have been the source of the detected compounds in the investigated South Atlantic sediments. Because these compounds could only be identified in sediments from Sites 1082 and 1084, which are potentially influenced by upwelling, an origin from planktonic organisms living at high nutrient availability seems likely.

Steroids

A range of C₂₆-C₃₀ sterols was detected. An overview of compounds identified in the TLC-5 fraction of Sample 175-1084A-12H-6, 140-143 cm (Fig. F5), is given in Table T3. The sterol distribution is dominated by only few components: cholesta-5-en-3-ol (cholesterol; Structure IXa in Fig. F4), 24-methylcholesta-5,22-dien-3-ol (Structure IXb in Fig. F4), 24-ethylcholest-5-en-3-ol (Structure IXc in Fig. F4), and 4,23,24-trimethylcholest-22-en-3-ol (dinosterol; Structure IXd in Fig. F4). All other detected sterols contribute only minor to trace amounts. The sterol content is highest in sediments from drill sites under upwelling cells (Sites 1084 and 1085) where phytoplankton primary production is high. Sediments from the Congo (Sites 1076 and 1077) and the Walvis Basin (Site 1082) contain minor quantities of sterols, and only trace amounts were detected in samples from the Angola margin (Site 1079). A conclusive interpretation of sterol distributions is complicated by the multiple sources of most sterols (Volkman, 1986).

Cholesta-5-en-3-ol (cholesterol) in marine sediments is derived from a variety of planktonic organisms, including haptophyte algae (Volkman, 1986). Zooplankton was found to be a major source of cholesterol in the Peru upwelling system through dietary alteration of phytosterols (Volkman et al., 1987). Therefore, the intermediate to high abundance of cholesterol in the samples from Site 1084 probably indicates extensive grazing of primary marine biomass by zooplankton.

The occurrence of the C₂₈ sterol 24-methylcholesta-5,22-dien-3-ol in sediments is often ascribed to a diatom lipid contribution. However, it was also detected in cultures of dinoflagellates (Teshima et al., 1980; Goad and Withers, 1982) and haptophytes (Volkman et al., 1998) and detected in the water column after massive blooms of Emiliania huxleyi (Conte et al., 1995). Its occurrence in sediment samples from Sites 1082, 1084, and 1085, all characterized by a substantial long-chain alkenone content (i.e., 80-1400 µg/g TOC) (Table T4), thus cannot be used to specifically indicate remnants of diatom biomass. More likely, a combined contribution of haptophytes and diatoms was responsible for the relatively high content of 24-methylcholesta-5,22-dien-3-ol.

Another C₂₈ sterol proposed as a marker compound for diatoms is 24-methylcholesta-5,24(28)-dien-3-ol because of its elevated concentrations in some centric diatom species (Volkman et al., 1998), although it also occurs in marine dinoflagellates (Nichols et al., 1984). It was only detected in minor amounts in samples from Site 1084 and was below the detection limit in all other samples. Specific diatom occurrence from sterol distributions is, therefore, only indicated at the site of the most substantial upwelling intensity.

Another major sterol is 24-ethylcholest-5-en-3-ol, together with its stanol analog (24-ethyl-5-cholestan-3-ol) present in smaller amounts. Both are commonly associated with terrestrial higher plants, in which they are the dominant sterols (Huang and Meinschein, 1976). In environmental settings, where terrigenous material is scarce, the main contributors likely are microalgae like diatoms and xantophyceae (Volkman et al., 1998). The source of these compounds in samples containing high concentrations of long-chain n-alkanes may have been terrestrial higher plant detritus as in the samples from the Congo area. In samples from upwelling sites, which contain minor or neglectable amounts of other terrigenous lipids, marine phytoplankton sources seem more likely.

The occurrence of 4,23,24-trimethylcholest-22-en-3-ol (dinosterol), accompanied by a range of other 4-methyl sterols, is ascribed to dinoflagellates. Dinosterol is almost uniquely produced by dinoflagellates (Boon et al., 1979b; Robinson et al., 1984), with minor quantities occurring in diatoms (Volkman et al., 1993). High concentrations of dinosterol were found in samples from Sites 1076, 1079, and 1084, with the highest variability in the Congo Fan sediments. Dinoflagellates, therefore, are considered to have significantly contributed to the marine extractable lipids in the Congo plume and in the upwelling cells, although the detection as a major lipid does not necessarily imply that dinoflagellates were dominant constituents of the phytoplankton community in these environments. Other less abundant sterols are less specific and may be derived from a variety of algal sources (cf. Volkman, 1986).

In the apolar fractions of samples from the upwelling Sites 1084 and 1085, and to a lesser extent also in samples from Sites 1079 and 1082, series of sterenes, steradienes, and steratrienes are present. The most abundant compounds are C₂₇ and C₂₈ homologues. They are diagenetic transformation products of sterols. Their formation proceeds via microbiologically mediated reduction and/or dehydration of sterols (Gagosian and Farrington, 1978; review in de Leeuw and Baas, 1986).

Sterol Ethers

In samples from Sites 1084 and 1085, relatively large amounts of sterol ethers are present; they occur in minor amounts at Site 1079. They commonly consist of C₂₇ and C₂₈, ⁵ and ^5,22 steroid moieties, ether bound to a nonyl or decyl moiety (Structure X shows the C₂₇⁵decyl ether in Fig. F4). These compounds have been identified by comparison with mass spectra reported by Schouten et al. (2000a). Their distributions show distinct differences from the sterol distributions. Steroid moieties larger than C₂₈, or stanol analogs, could not be detected. A diagenetic origin of sterol ethers from alkylation of the sterols therefore seems unlikely. The occurrence of sterol ethers has earlier been reported for Walvis Bay diatomaceous ooze (Boon and de Leeuw, 1979a), and from sediments from the Miocene Monterey formation and the Arabian Sea (Schouten et al., 2000a). All these settings, including the samples under consideration, are known to receive a large amount of organic matter from diatoms, from which these compounds could be derived. It is assumed that the sterol ethers either represent primary biosynthetic products or are derived from zooplankton feeding on specific source organisms. Copepods are known to convert algal sterols to steryl chlorin esters (Talbot et al., 1999). In that case, algal sterols were converted with equal efficiency to steryl esters, resulting in the same ratio for sterols and steryl esters. We suggest that the formation of sterol ethers could take place by a similar mechanism during herbivory, although no sterol ether-producing organism has yet been identified.

Loliolide and iso-Loliolide

An omnipresent compound pair in all investigated samples is loliolide (Structure XI in Fig. F4) and its isomer, iso-loliolide. They are the dominant lipids in the TLC-8 fraction of Sample 175-1084A-12H-6, 140-143 cm. Highest concentrations in samples from Site 1084 are followed by slightly lower contents in samples from Sites 1076 and 1079. At Site 1076 the highest variability was detected. Loliolide and iso-loliolide are known as end products of anaerobic degradation of fucoxanthin (Klok et al., 1984; Repeta, 1989). Fucoxanthin is the major carotenoid in diatoms and haptophytes. Loliolide and iso-loliolide were earlier reported from sediments receiving abundant marine phytoplankton biomass, like the Mediterranean sapropels (ten Haven et al., 1987), the Namibian margin (Klok et al., 1984), and the Peruvian shelf (Repeta, 1989). The elevated concentrations of loliolide and iso-loliolide in the sediments from Sites 1076, 1079, and 1084 therefore indicate a large carotenoid contribution from diatoms and/or haptophyte algae at these sites. Furthermore, it is concluded that anaerobic degradation of carotenoids to loliolides has taken place in the sediments.

Biphytane Diols

Two C₄₀ cyclic biphytanediols were detected in the TLC-5 fraction of Sample 175-1084A-12H-6, 140-143 cm. By comparison with published mass spectra, they were identified as the C₄₀ dicyclic (Structure XIIa in Fig. F4) (Schouten et al., 1998) and C₄₀ tricyclic biphytanediol (Structure XIIb in Fig. F4) (Schouten et al., 2000b). Investigations of all samples showed that those lipids are present in highest concentrations and with the largest variation in samples from Site 1084. Although these compounds have not yet been identified in organisms, they were encountered in various sediments from, for example, the Indian Ocean, Arabian Sea, Angola Basin, Walvis Bay, and Madeira Abyssal Plain (Schouten et al., 1998, 2000a). From the structural similarities with sedimentary C₄₀ ether-bound biphytanes, it was proposed that biphytane diols are biosynthesized by archaea (Hoefs et al., 1997; Schouten et al., 1998). The presence of intact archaeal tetraether lipids was proven for Sample 175-1084A-25X-3, 20-23 cm, by high performance liquid chromatography following the methodology described by Hopmans et al. (2000). Their high abundance in water-column samples of the Black Sea and the Cariaco Basin and in surface sediments from the Indian Ocean (Hoefs et al., 1997) suggests that they are already produced in the water column. The carbon skeleton structure of the sedimentary C_40:3 biphytane diol (and ether-bound biphytane) differs from that of thermophilic archaea (Schouten et al., 1998). Therefore, it is assumed that these compounds derive from nonthermophilic archaea (Hoefs et al., 1997). Possibly, the pelagic archaea are associated with low-oxygen conditions (Hoefs et al., 1997).

Hopanoids

Small amounts of hop-17(21)-ene and neo-hop-13(18)-ene were detected in all samples. These compounds are present in slightly elevated concentrations in samples from Sites 1082, 1084, and 1085, influenced by upwelling. For neo-hopenes, a direct biological origin was suggested (J.S. Sinninghe Damsté et al., unpubl. data), but hop-17(21)-ene is likely to be derived from diploptene by isomerization during early diagenesis.

Extended hopanoids identified in the sediments are 17,21(H)-homo-hopane, 17,21(H)-homo-hopan-31-ol, 17,21(H)-dihomo-hopan-32-ol (Structure XIII in Fig. F4), a C₃₃ hopane-diol, and a C₃₅ hopane-keto-diol. The latter two compounds were tentatively identified by their mass spectra (Fig. F6). They are present in all investigated samples but occur in highest amounts in samples from the Congo Fan (Sites 1076 and 1077). Common precursors of hopanoid compounds are bacteriohopanepolyols (Rohmer et al., 1989). Hopanoids are exclusively biosynthesized by prokaryotes as membrane rigidifiers and can be considered the bacterial counterparts of steroids in cell membranes of algae (Rohmer et al., 1989). Recently, it has been inferred that extended hopanoids in sediments are predominantly derived from cyanobacteria and much less from heterotrophic bacteria or bacterial reworking (Schoell et al., 1994; Summons et al., 1999). Sediments located under the Congo plume are therefore considered to receive a larger cyanobacterial contribution than the other sediments investigated.

Wax Esters

Wax esters identified in the TLC-1 fraction of Sample 175-1084A-12H-6, 140-143 cm, range from C₃₂to C₄₀ and have exclusively saturated fatty acid and alcohol moieties. The fatty acid moieties range from C₁₄ to C₂₄ and the alcohol moieties from C₁₆ to C₂₄. For both counterparts, the C₁₆, C₁₈, and C₂₀ homologues are quantitatively the most important ones. Wax esters are minor components of the total extractable organic matter and they could only be detected in the TLC fraction. Besides their occurrence in leaf waxes of higher plants (Kolattukudy, 1976), wax esters were reported to occur in pelagic marine animals, especially copepods. Copepods contain relatively large amounts of wax esters as energy reserves (Sargent et al., 1976). The storage of wax esters in copepods is connected to intermittent food supply with periods of starvation. The occurrence of wax esters in sediments from Site 1084 thus could be potentially indicative of zooplankton lipids at upwelling-influenced sites.

Perylene

In the TLC-2 fraction of Sample 175-1084A-12H-6, 140-143 cm, perylene (Structure XIV in Fig. F4) was identified as a major compound. It is the only detected polycyclic aromatic hydrocarbon in all samples. It was earlier identified in Namibian shelf sediments by Wakeham et al. (1979). Perylene was reported to be most abundant in anoxic diatomaceous sediments (Louda and Baker, 1984). Although the actual precursor compound is still unknown, in situ diagenetic formation of perylene under anoxic conditions is commonly assumed (Silliman et al., 1999). Venkatesan's (1988) conclusion that diatoms are the most likely source of a precursor compound fits well with our finding of perylene in sediments below the Lüderitz upwelling cell. Besides suggesting an origin from diatom lipids, the presence of perylene is related to a large abundance of labile organic matter, leading to a strong oxygen consumption in the sediments.

Organic Sulfur Compounds

Two tricyclic triterpenoid thianes were detected in minor amounts in the apolar fractions of samples from Site 1084. By comparison with mass spectra reported by Werne et al. (2000), they were identified as monounsaturated malabaricane thianes (Structure XV in Fig. F4).

A sulfur compound present in trace amounts in sediments from Sites 1084 and 1082 was identified as a C₂₀ isoprenoid thiophene, 3-methyl-2-(3,7,11-trimethyldodecyl)-thiophene (Structure XVI in Fig. F4). Its occurrence in sediments from Walvis Bay was reported earlier (Brassell et al., 1986b; ten Haven et al., 1990). Sulfur incorporation into phytol-derived phytadienes leads to the formation of isoprenoid thiophenes (Brassell et al., 1986b). Source organisms of phytol are chlorophyll-producing photoautotrophic algae.

Several isomers of highly branched isoprenoid (HBI) thiolanes were detected in sediments from Site 1084 and, in much lower amounts, from Site 1082. They were identified as C₂₅ monounsaturated and saturated HBI thiolanes (Structure XVII in Fig. F4) by comparison with published mass spectra (Kohnen et al., 1990). These compounds derive from the reaction of inorganic sulfur species with specific double bonds of the precursor HBI polyenes during early diagenesis (Kohnen et al., 1990). Precursor compounds of the HBI thiolanes are HBI polyenes biosynthesized by specific diatom species like Rhizosolenia setigera and Haslea sp. (Volkman et al., 1994; Belt et al., 1996; Sinninghe Damsté et al., 1999). Desulfurization of the polar fraction of Sample 175-1084A-12H-6, 140-143 cm, released C₂₅ (Structure XVIII in Fig. F4) and C₂₆ HBI (Structure XIX in Fig. F4) compounds (Fig. F7) in quantities similar to those of the long-chain alkenones. The C₂₆ HBI compound most probably originates from the same diatom species as the C₂₅ HBI (Rospondek et al., 1997). The presence of these compounds in samples from Site 1084 therefore indicates an important contribution of diatom lipid material to sediments deposited in upwelling-influenced environments. Rapid sulfur incorporation into alkenes may account for the absence of unsulfurized HBI compounds in samples from Site 1084 (cf. Werne et al., 2000; Kohnen et al., 1990).

This rapid formation of organosulfur compounds is in line with the high rate of sulfate reduction in the uppermost meters of sediment at Sites 1082 and 1084 (Wefer, Berger, Richter, et al., 1998), suggesting a high abundance of labile organic matter and low-oxygen conditions at these locations. In other sediments, organic sulfur compounds were not detected. The formation of organic sulfur compounds at other locations could possibly be inhibited by a higher availability of reactive iron.

Site Summaries

The majority of biomarker lipids in sediment samples drilled during Leg 175 in the eastern South Atlantic are of marine, autochthonous origin. Terrigenous lipids, detected in minor but varying amounts, were transported into the marine environment via dust or fresh water. Their contribution to the sedimentary lipid material could be used as a measure of the relative importance of terrigenous vs. marine organic matter deposition in each setting. Because of the limited number of samples in this study and their different ages, paleoceanographic interpretations are restricted. The study is solely intended to provide an inventory and underline the potential of detailed biomarker analyses for paleoceanographic purposes.

Figure F8 shows the concentrations of selected biomarkers or compound classes to illustrate the lipid composition at each site, as summarized below.

Site 1076

Site 1076 is the shallow-water drill site in the lower Congo Basin. The sediments contain the highest amounts of terrigenous lipids of all investigated samples (Table T4; Fig. F8). The concentrations of n-alkanes, derived from leaf waxes of higher plants, and taraxerol, most probably derived from mangroves, show strong variations between the samples (Fig. F8). This suggests a dependence of the concentrations of terrigenous lipids on fluctuations in river discharge, similar to the late Quaternary (Jansen et al., 1984; van der Gaast and Jansen, 1984; Jansen, 1990; Schneider et al., 1997). Long-chain alkenone concentrations in samples from Site 1076 display a strong variability with elevated concentrations when sea-surface temperatures were low (Table T4). This is in agreement with the results from the late Quaternary of stronger coastal upwelling during cold times with low river discharge (Jansen and van Iperen, 1991). On average, samples from Site 1076 contain the highest concentrations of alkyl diols and intermediate amounts of dinosterol and loliolide. These compounds also exhibit strong variability in concentrations, suggesting that the production of eustigmatophytes, dinoflagellates, and diatoms covaried with river discharge of nutrients or coastal upwelling. The presence of large amounts of eustigmatophyte-derived lipids may indicate a significant contribution by eustigmatophyte algae from brackish waters. Cholesterol is present in trace amounts and biphytanediols are absent. Zooplankton herbivory and pelagic archaea are therefore suggested to have been of minor importance.

Site 1077

Site 1077 is the intermediate-depth drill site on the Congo Fan. Samples from this location have the lowest organic carbon content of all investigated sediments. The content of terrigenous lipids, n-alkanes, and taraxerol is lower than in the sediments from Site 1076 but still elevated compared to the more southern sites (Table T4). On average, concentrations of alkenones, loliolide, alkyl diols, and dinosterol are lower than at Site 1076 (Table T4). Slightly higher concentrations were detected for biphytanediols and cholesterol (Table T4). This suggests that haptophyte algae, diatoms, eustigmatophytes, and dinoflagellates on average were less important as primary producers. In contrast, archaea generally contributed slightly higher amounts to the lipid material (Fig. F8). This change agrees, despite the limited sample numbers, with a zonation of plankton groups depending on the proximity to the river mouth (van Iperen et al., 1987; Ufkes et al., 1998). The influence of fluviatile nutrients (on diatoms) and river-induced upwelling (on haptophytes) thus would be much smaller at Site 1077 than at Site 1076.

Site 1079

Samples from Site 1079 on the Angola margin contain intermediate concentrations of alkenones, loliolide, dinosterol, cholesterol, and alkyl diols. Terrigenous lipids (e.g., n-alkanes and taraxerol) are present in low but significant amounts in both samples (Table T4). A direct supply of river-transported terrigenous lipid material seems unlikely because of the lack of rivers draining into this area. However, alongshore transport from the Kunene River might carry terrigenous lipids to this site. The composition of marine lipids, with alkenones, loliolide, and dinosterol as the major components, is, except for the high content of alkyl keto-ols, not significantly different from the lipid composition of sediments in the Congo area. The elevated sedimentary organic carbon contents of these samples therefore probably result from a similar plankton community, with haptophyte algae, diatoms, dinoflagellates, and unidentified, keto-ol-producing microalgae as main lipid producers. The supply of nutrients from subsurface waters, likely involving the Angola Dome nearby, may sustain high productivity.

Site 1082

Sediments at Site 1082 in the Walvis Basin were reported to receive abundant biological material from displaced upwelling filaments (Wefer, Berger, Richter, et al., 1998). The elevated organic carbon content of the studied samples (Table T1) may therefore reflect a strong phytoplankton signal. Most commonly, upwelling induces high diatom productivity. Only low amounts of loliolide are present in the sediments. However, HBI thiolanes clearly indicate a significant supply of diatom-derived lipid material. The average concentration of alkenones in the sediments is relatively high (Fig. F8). Biphytanediols are present in significantly elevated amounts, in contrast to the more northern sites. Other biomarkers only contribute small to minor quantities. Especially the terrigenous lipid contribution is very low: the concentrations of n-alkanes are low, and taraxerol is absent (Table T3). The n-alkanes at this location are most probably predominantly wind derived, whereas mangrove lipids are probably absent in the samples because of the lack of coastal mangrove vegetation in this area and because of the ABF in the north inhibiting southward transport.

Site 1084

Investigated sediments from Site 1084 off Lüderitz Bay contain the highest amount of organic carbon of all studied samples (up to 13%) (Table T1). Sediments from this site should contain the most pronounced coastal upwelling signal because of their close vicinity to the Lüderitz upwelling cell (Wefer, Berger, Richter, et al., 1998). This is well expressed in the lipid geochemistry of these samples. Sediment from Site 1084 contains the highest concentrations of biphytanediols, long-chain alkenones, loliolide, and cholesterol. Alkyl diols and dinosterol are present in elevated amounts (Table T4; Fig. F8). The presence of other plankton-derived compounds, like midchain hydroxy fatty acids, sterols, sterol ethers, hopanoids, isoprenoid thiophenes, and sulfurized HBIs, also indicates the importance of plankton biomass contribution. The investigated sediments from Site 1084 contain the highest amounts of phytoplankton, pelagic archaea, and zooplankton markers of all investigated samples. Land-derived lipids are present only in trace amounts. The n-alkanes exhibit only small concentrations. Taraxerol, like at Site 1082, is absent. Of all studied samples, the sediments from Site 1084 contain the most diverse upwelling-derived lipids. The organic geochemical phytoplankton signals are much more pronounced than at Site 1082.

Site 1085

Site 1085 is located in the southern Cape Basin, situated under a seasonal upwelling regime. Its sediments show less pronounced upwelling signals than those at Site 1084. The contribution of marine-derived compounds, like alkenones, loliolide, alkyl diols, and dinosterol, is on average smaller than at Site 1084 and in the same range as at Site 1082. The presence of elevated amounts of sterols and sterol ethers, however, suggests phytoplankton production with subsequent zooplankton grazing. The sediments from Site 1085 contain only trace quantities of n-alkanes, indicating negligible transfer of material from the nearby Orange River. Taraxerol could not be detected in this setting, which agrees with the present absence of mangrove vegetation in southern Africa. The lipid compositions of sediments from Site 1085 resemble those of sediments from Site 1082. Both settings receive their biological signal from seasonal upwelling or mixing of upwelled, nutrient-rich waters with low-productivity surface waters, in contrast to the persistent year-round upwelling at Site 1084.