Transfer of organic carbon through marine water columns to sediments – insights from stable and radiocarbon isotopes of lipid biomarkers

Compound-specific C and C compositions of diverse lipid biomarkers (fatty acids, alkenones, hydrocarbons, sterols and fatty alcohols) were measured in sinking particulate matter collected in sediment traps and from underlying surface sediments in the Black Sea, the Arabian Sea and the Ross Sea. The goal was to develop a multiparameter approach to constrain relative inputs of organic carbon (OC) from marine biomass, terrigenous vascular-plant and relict-kerogen sources. Using an isotope mass balance, we calculate that marine biomass in sediment trap material from the Black Sea and Arabian Sea accounted for 66–100 % of OC, with lower terrigenous (3–8 %) and relict (4–16 %) contributions. Marine biomass in sediments constituted lower proportions of OC (66–90 %), with consequentially higher proportions of terrigenous and relict carbon (3–17 and 7– 13 %, respectively). Ross Sea data were insufficient to allow similar mass balance calculations. These results suggest that, whereas particulate organic carbon is overwhelmingly marine in origin, pre-aged allochthonous terrigenous and relict OC become proportionally more important in sediments, consistent with pre-aged OC being better preserved during vertical transport to and burial at the seafloor than the upper-ocean-derived marine OC.


Introduction
The dynamics of sources, sinks and processes that control the burial of organic carbon (OC) in marine sediments have important implications for the global carbon cycle, paleoceanographic reconstructions and understanding climate variability (Berner, 1982;Hedges and Keil, 1995;Burdige, 2007;Zonneveld et al., 2010).Assigning the provenance of sedimentary OC remains a difficult task.Most burial occurs on continental margins where terrigenous material constitutes a significant proportion of the burial flux (Hedges et al., 1997;Burdige, 2005).But even at open-ocean locations remote from the continents where marine OC dominates the water column flux via the biological pump, a terrigenous component delivered by long-range aeolian transport (Zafiriou et al., 1995;Gagosian and Peltzer, 1986;Eglinton et al., 2002;Kawamura et al., 2003) is still recognizable in sediments (Prahl et al., 1989;Wakeham et al., 2002;Zonneveld et al., 2010).Myriad biogeochemical and sedimentological processes during transport through the marine water column and at the sediment-water interface affect the quantity and nature of sedimentary OC.The structural characterization of most marine organic matter is incomplete (Hedges et al., 2000;Lee et al., 2004), but it is generally thought that marine OC is more reactive than terrigenous plant and relict sediment OC (Cowie and Hedges, 1984;Wakeham et al., 1997).Even a small fraction of the most recalcitrant relict carbon is amenable to bacterial assimilation (Petsch et al., 2001(Petsch et al., , 2003;;Pearson et al., 2005;Wakeham et al., 2006).Selective enrichment of terrigenous OC in marine sediments over what is observed in the water column may result from differences in the intrinsic reactivity of the organic molecules themselves, protection by degradation-resistant macroorganic matrices and mineral surfaces, or environmental conditions (oxygen availability, oscillating redox, microbial consortia present) (Hedges and Keil, 1995;Hedges et al., 2001;Aller, 1994;Wakeham and Canuel, 2006;Burdige, 2007).
A suite of geochemical tools are applied to characterize the source(s) and fate of OC in the marine water column and sediments.Elemental compositions and ratios (e.g., OC/N) are often combined with carbon isotope analyses at the bulk level.Natural abundance stable carbon isotopes (δ 13 C) give insight into carbon source, carbon assimilation pathways and carbon flow in marine ecosystems and food webs (Hayes, 1993;Fry and Sherr, 1994;Freeman, 2001;Pearson, 2010).Natural-abundance radiocarbon analyses ( 14 C OC or fraction modern f m ) add the dimension of "age" to the character of organic matter and help define the residence time and redistribution of OC (Blair et al., 2003;Ingalls and Pearson, 2005;Griffith et al., 2010).Molecular analyses of biomarkers can distinguish between marine (e.g., sterols, alkenones), terrigenous (plant waxes and lignin phenols) and relict (alkanes with a carbon preference index of ∼ 1 and often an unresolved complex mixture) materials in the heterogeneous mixture that is sedimentary OC.However, biomarker compounds are often present at low concentrations, and extrapolations to bulk OC are formidable.Single-compound (compound-specific) isotope analyses help in this respect because they combine the source specificity of biomarkers with δ 13 C biomarker -derived information on carbon flow (Freeman et al., 1990;Hayes, 2001;Freeman, 2001) and 14 C biomarkerderived ages that indicate the mixing of old with modern OC (Eglinton et al., 1997;McNichol and Aluwihare, 2007;Ingalls and Pearson, 2005).Early investigations in the marine water column showed that, whereas sinking particulate OC comprised of fresh planktonic detritus has a young radiocarbon age, surface sedimentary OC has older ages of a "pre-aged" and/or relict terrigenous component (Druffel et al., 1996;Wang et al., 1998;Hwang et al., 2010).Compound-specific radiocarbon measurements confirm preaged biomarkers in sedimentary OC (review by Ingalls and Pearson, 2005), and compound-class radiocarbon analyses have subsequently shown pre-aged OC in water column particulate matter (Wang et al., 2001;review by McNichol and Aluwihare, 2007).However, to date, combined biomarkerstable-carbon-radiocarbon studies of marine particulate matter are few (an exception being Ingalls et al., 2006), largely due to sample size limitations.
The present investigation attempts to fill this gap.The combination of lipid biomarker composition with molecular stable carbon and radiocarbon isotopes offers a threedimensional approach for investigating OC sources and transport and alteration processes in the ocean.In this study, compound-specific stable carbon and radiocarbon isotopes were measured on multiple lipid biomarkers in sinking particulate matter collected in sediment traps and from underlying surface sediments to evaluate the provenance of sedimentary organic matter.Three sets of paired sediment-trapsurface-sediment samples (POM -particulate organic matter; SOM -sedimentary organic matter) from the Black Sea, Arabian Sea and Ross Sea were investigated as representatives of oceanic regions characterized by widely disparate OC sources and depositional environments.Organic carbon content (%OC), atomic C / N ratios (C / N (a) ), δ 13 C and 14 C of bulk materials and individual biomarkers (fatty acids, hydrocarbons, alkenones, alcohols and sterols) were measured to identify major molecular and isotopic compositions, and thereby shifts in relative amounts of marine, terrigenous (preaged vascular-plant-derived with continental residence times of decades to centuries) and relict (derived from eroded ancient sedimentary rocks and petrogenic material) carbon, in POM and underlying SOM.Throughout this discussion it is important to remember that the sediment trap samples are short snapshots in time (up to 6 months) whereas the surface sediments may represent centuries.

Study sites and samples
Paired sediment trap and surface sediment (∼ 0-2 cm) samples from three settings (Fig. 1) were studied: the Black Sea (anoxic water column; high biogenic and high terrigenous OC); the Arabian Sea (oxygen minimum zone; high biogenic OC, low terrigenous OC); the Ross Sea, Antarctica (high biogenic flux, relict continental OC).The choice of these locations was predicated in part on the availability of large amounts of archived sinking particulate matter to facilitate compound-specific radiocarbon analyses.

Black Sea
The Black Sea site was in the southwestern Black Sea (42 • N, 32 • E), at station BS of the joint US-Turkish-German sediment trap program 50 km north of Asmara and 15 km from the base of the continental slope (Hay and Honjo, 1989;Hay et al., 1990).Anoxic conditions prevail below ∼ 120-150 m water depth (Sorokin, 1983).Underlying sediments contain up to ∼ 6 % OC (Ross and Degens, 1974;Premuzic et al., 1982;Calvert et al., 1991).Biogenic and lithogenic particle fluxes and sediment accumulation have been reported by Hay (1987) and Hay et al. (1990).The site is offshore of the rugged Pontic Mountains on the north Anatolian coastline, and the continental margin is particularly steep, typically 6 • between 100 and 2000 m depths (Ross et al., 1974).Lithogenic material delivered by the Sarkarya Nehri, Filyos and Kocacay Rivers, usually during winter/spring, is deposited on the shelf, is frequently resuspended by storms and is subsequently transported offshore as either surface, mid-water (∼ 150 m depth) or bottom-water nepheloid layers and turbidites; a surface plume extending out to the study site is sometimes visible on LANDSAT imagery (Hay, 1987).Sedimentation of lithogenic material out of the water column is enhanced by biogenic material derived from the annual succession of blooms of the coccolithophorid Emiliania huxleyi (spring) and pennate diatoms Rhizosolenia sp.(summer), producing the characteristic light-dark laminae (varve couplets) of Black Sea sediments (Hay, 1987;Hay et al., 1991).White laminae are comprised almost exclusively of coccolithophores of E. huxleyi, whereas dark laminae are predominately terrigenous clay minerals.
Sediment trap material (18.5 grams dry weight, gdw) from the moored BS trap at ∼ 250 m depth under strongly anoxic conditions is a composite of a 7-month-long time series collection (October 1985-April 1986).Trap material was preserved with buffered formalin and stored refrigerated at 4 • C. Surface sediment (0-2 cm; 109.7 gdw) was composited from multicores collected at 2200 m water depth during the 1988 Black Sea expedition (Murray and Izdar, 1989;Hay and Honjo, 1989) at approximately the same location as the trap mooring.Sediments were stored frozen.Radiocarbon dating of sediments near the study site gave a sedimentation rate of 26 cm kyr −1 (Arthur and Dean, 1998).Thus the 0-2 cm sediment interval represents approximately 1 century.

Arabian Sea
Arabian Sea samples were collected at mooring site MS-1 in the northwestern Arabian Sea (Oman Margin) during the US Joint Global Ocean Flux Study (JGOFS) Arabian Sea Process Study (ASPS) in 1994-1995 (Smith et al., 1998).MS-1 was approximately 160 km from the Oman coast off Ra's Sharbatat (17 • 41 N, 58 • 51 E) at a water depth of 1445 m.High seasonal productivity during monsoon-driven upwelling enhances export into deep waters where remineralization of sinking organic matter depletes dissolved oxygen down to levels of ∼ 5 µM, producing the world's largest open-ocean oxygen minimum zone (OMZ) (Smith et al., 1998).More than 50 % of the annual particle flux in the central Arabian Sea occurs during the boreal-summer southwest monsoon.Biogenic material, primarily diatom-derived, dominates over lithogenic material (Haake et al., 1996;Honjo et al., 1999), but some terrigenous material is delivered to the northwestern Arabian Sea as dust from the Horn of Africa (Somalia and Ethiopia) and from the Arabian Peninsula by strong summer monsoon winds (Ramage et al., 1972;Sirocko and Sarnthein, 1989;Dahl et al., 2005).
Sediment trap material (35.5 gdw) was a composite of material collected in time series traps deployed in the OMZ at ∼ 500 and ∼ 900 m between May 1995 and January 1996, covering the southwest monsoon period (Wakeham et al., 2002).Mercuric chloride was used as a biocide.Upon recovery, trap samples were sealed and stored refrigerated at 4 • C. Surface sediments (0-2 cm) were composited (142.7 gdw) from multicores collected in 1995 at the same location as the trap deployment and stored frozen.Passier et al. (1997) have estimated the sedimentation rate for this part of the Oman Margin at 5 cm kyr −1 ; the sediment sample thus represents about 400 years.

Ross Sea
The Ross Sea site was in the southwestern Ross Sea where diatoms, primarily Nitzschia sp., and Phaeocystis antarctica dominate the phytoplankton community and vertical flux during the austral summer bloom (Arrigo et al., 2002;Dunbar et al., 2003).Sediments are largely biogenic oozes (biogenic silica 10-30 %) with low OC (0.1-3 %, averaging 1.5 %) and negligible biogenic carbonate (Dunbar et al., 1985(Dunbar et al., , 1989)).Terrigenous material, primary lithogenics with low OC content, is delivered by glaciers that drain the polar plateau and by aeolian transport from the ice-free Dry Valleys and accounts for 2-25 % of the vertical flux through the water column.In the western Ross Sea, ice-rafted debris constitutes ∼ 10 % of sediments (Anderson et al., 1984) and aeolian sedimentation (either through sea ice or directly onto the sea surface) could be up to 50 % of sediments in nearshore areas with limited glacial ice cover (Bentley, 1979, cited in Dunbar et al., 1989).Nonetheless, sedimentary OC is dominated by biogenic water column sources.Sediments are resuspended, mixed and redistributed within a pervasive nepheloid layer.(Dunbar et al., 2003).Trap material was preserved with 3 % formalin and stored at 4 • C. Surface sediments (515.1 gdw) were obtained from box cores (Ohkouchi et al., 2003) and stored frozen.
Sedimentation rate determinations for the Antarctic margin often use acid-insoluble organic carbon (AIOC) due to a lack of calcareous formaninifera (Domack et al. 1989;Licht et al., 1996;Harris et al. 1996;Andrew et al., 1999).DeMaster et al. (1996) reported AIOC-derived sedimentation rates in this region of the Ross Sea of ∼ 4.5 cm kyr −1 .AIOC-based chronology, however, is complicated by "contamination" by unknown amounts of relict OC (Sackett et al., 1974).To overcome this problem, Ohkouchi et al. (2003) applied a compound-specific radiocarbon analysis of sedimentary fatty acids for cores from the Gentoo and Adelie sites and found a 1200-2000 year offset between the ages of fatty acids and AIOC.The fatty acid-derived sedimentation rate was 7.5 cm kyr −1 vs. an AIOC-derived rate of 15 cm kyr −1 .The sediment sample therefore represents 130-250 years of deposition.

Elemental analysis
Freeze-dried and acidified (Hedges and Stern, 1984) trap material (bulk POM) and sediments (bulk SOM) were analyzed for organic carbon (%OC) and total nitrogen (TN) with a Fisons (Model EA 1108) elemental analyzer.

Lipid analysis
The extraction, cleanup and isolation of fatty acids, hydrocarbons, alcohols and sterols by preparative capillary gas chromatography (PCGC; Eglinton et al., 1996) are outlined in Fig. 2. All laboratory glassware and SiO 2 were precombusted at 500 • C for 8 h before use.Freeze-dried POM and SOM were Soxhlet-extracted with methylene-chloride : methanol (DCM : MeOH, 2 : 1 v/v) for 72 h.Extracts were washed with 5 % NaCl solution, and solvent lipid extracts (SLEs) were partitioned into DCM, after which the DCM fraction was dried over Na 2 SO 4 .SLEs were saponified using 0.5 N KOH at 100 • C for 2 h, and nonsaponifiable lipids were extracted out of the alkaline mixture with hexane after which the pH was adjusted to < 2 with 6N HCl, and acids were extracted with hexane.Nonsaponifiable lipids were fractionated on 5 % deactivated silica gel into a hydrocarbon fraction eluted with hexane, an alkenone fraction eluted with 10 % ethylacetate in hexane and an alcohol-sterol fraction eluted with 25 % ethylacetate in hexane.Straight-chained hydrocarbons were separated from branched and cyclic hydrocarbons by urea adduction.Alkenones were isolated by sequential silica gel, AgNO 3 -silica gel chromatography and urea adduction after Ohkouchi et al. (2005).Alcohols and sterols were acetylated with pyridine and acetic anhydride.Acids were methylated with BF 3 : MeOH, and the fatty acid methyl esters (FAMEs) were purified on columns of activated SiO 2 .

Stable and radiocarbon analysis
The isolation of purified individual hydrocarbons, sterol acetates, alcohol acetates and FAMEs was by preparative capillary gas chromatography (PCGC; Eglinton et al., 1996;Wakeham et al., 2006).An Agilent 5980II gas chromatograph equipped with an HP 7673 autoinjector, a Gerstel CIS-3 cooled injection system and a Gerstel preparative fraction collector (PFC) was fitted with an RTX-1 megabore (60 m × 0.53 mm ID × 0.5 µm film) capillary column.The GC temperature program was 60 • C (1 min), 20 • C min −1 to 160 • C, 4 • C min −1 to 300 • C and isothermal at 300 • C for 20 min.An effluent splitter directed 1 % of the column effluent to the flame ionization detector, and the remaining 99 % was sent to the zero-dead-volume splitter of the PFC.The PFC was operated at 320 • C, and U-tube traps were held at room temperature.Purified fractions were checked for purity and quantified by gas chromatographymass spectrometry (Agilent 6890 gas chromatograph, Agilent 5793 mass spectrometer, 30 m × 0.25 mm ID J&W DB-5 capillary column).Individual, composited compounds or operational classes were transferred to glass ampules and flamesealed for isotope analysis.
Stable and radiocarbon isotope measurements were made at the National Ocean Sciences Accelerator Mass Spectrometry (NOSAMS) Facility at the Woods Hole Oceanographic Institution.δ 13 C values are reported relative to the Vienna Pee Dee belemnite (precision ± 0.2 ‰), and 14 C values are reported according to Stuiver and Polach (1977), using the year of sample collection for age correction.Acidified POM and SOM were transferred to precombusted Vycor tubes containing CuO and Ag powder.Sample extracts and isolated biomarkers were transferred with solvent to precombusted Vycor tubes, and after evaporating the solvent, 100 mg precombusted CuO was added to the tube.Samples were combusted to CO 2 at 850 • C for 5 hours.After purification and quantification, a split of the CO 2 was analyzed for δ 13 C on a VG Micromass Optima isotope ratio mass spectrometer.The remaining CO 2 was reduced to filamentous graphite over either Fe or Co powder.Radiocarbon analyses of both large and small samples were performed using standard NOSAMS procedures (McNichol et al., 1994;von Reden et al., 1998;Pearson et al., 1998).Processing and combustion blanks of a hydrocarbon fraction isolated from a south Louisiana crude oil had replicate 14 C values between −980 and −998 ‰ (f m 0.03 and 0.001, respectively).Contributions of added methyl carbon derived from methanol in FAMEs and acetyl carbons from acetic anhydride in alcohol and sterol esters were removed by isotopic mass balance (Pearson, 2000;Wakeham et al., 2006).The δ 13 C and 14 C values of carbon in the BF 3 : MeOH and acetic anhydride reagents were calculated by measuring the δ 13 C and 14 C values of palmitic acid and cholesterol standards and of methyl palmitate and cholesteryl acetate prepared using the same lots of BF 3 -MeOH and acetic anhydride, respectively, and isolated by PCGC.
Several FAs (fatty acids) had unusually high 14 C values and were not used in subsequent calculations of ranges, means and standard deviations of radiocarbon data (shown in bold italics in Tables 1-3).The reasons for these enrichments are unknown (see Wakeham et al. (2006) for additional discussion).Radiotracers had never been used in the Skid- away Institute laboratory, which was thoroughly checked for any radiocarbon contamination, and crude-oil process blanks were free of modern carbon.The most enriched FA (most often but not always 16 : 0, 18 : 1 and 18 : 0) were compounds isolated in the highest concentrations and thus ones for which larger amounts of carbon were analyzed by accelerator mass spectrometry (AMS) rather than being the less abundant compounds.Such enrichments were not systematic, and in several cases replicate AMS analyses from splits of the same isolates were made, but with similar results indicating that any contamination must have occurred prior to or during PCGC workup.However, Levin and Kromer (1997) suggested that the average 14 C of atmospheric CO 2 between 1980 and the present may have been ∼ 200 ‰, whereas analysis of post-bomb sediment from the Santa Monica Basin by Pearson and Eglinton (2000) indicates that an average 14 C of atmospheric CO 2 around 1960 may have been ∼ 400 ‰.Thus while it is not possible to completely rule out the incorporation of higher amounts of post-bomb 14 C into these biomarker 14 C, other marine biomarkers and bulk OC are not consistent with such a scenario.

Bulk elemental compositions
Organic carbon contents (%OC) of POM (Fig. 3a) were 8.7 % in the Black Sea (BS), 6.4 % in the Arabian Sea (AS) and 5.5 % in the Ross Sea (RS).Total nitrogen (%TN) contents were 1.1, 0.86 and 0.92 %, respectively, for the BS, AS and RS trap material.Thus, atomic C / N ratios were 7.7, 7.4 and 5.9 (Fig. 3b).Sediment %OC and %TN were lower: 1.8, 2.3 and 0.54 % OC for the BS, AS and RS, respectively; 0.20, 0.29 and 0.08 % TN for the BS, AS and RS, respectively.C / N (a) ratios for sediments were somewhat higher (9.4,7.7, 6.5 for BS, AS and RS, respectively) than for POM.Except for the BS sediment, C / N (a) ratios of the other trap and sediment samples were sufficiently Redfield-like to indicate the predominance of marine OC.The higher C / N (a) ratio of the BS sediment suggests a somewhat higher component of terrigenous OC; alternately, the higher C / N (a) ratio could result from preferential loss of nitrogen during degradation of OC,

Bulk stable carbon isotopes
BS and AS POM δ 13 C OC values were typical for marinedominated OC: −22.9 and −22.4 ‰ for the Black Sea and Arabian Sea, respectively (Fig. 4a).δ 13 C OC values for the corresponding SOM were both slightly more negative (BS: −25.3 ‰) and more positive (AS: −20.8 ‰) compared to the corresponding POM.The 13 C-depletion in the BS sediment could reflect a greater long-term/time-averaged terrigenous C 3 -plant OC component (Collister et al., 1994;Conte and Weber, 2002;Chikaraishi et al., 2004) Peninsula (Parker et al., 2004;Dahl et al., 2005), but again sampling timescales for trap and sediment are different.The Ross Sea is a very different environment.Both POM and SOM had δ 13 C OC values of −27.9 ‰, significantly isotopically depleted compared to BS and AS samples.Algal biomass at high latitudes is typically depleted in 13 C relative to algal OC at lower latitudes (Rau et al., 1991a, b;DeHairs et al., 1997;Freeman, 2001).Several factors may be involved in this differential photosynthetic isotope fractionation, including high algal growth rates, higher dissolved CO 2 concentrations associated with low seawater temperatures and carbon assimilation mechanisms (Rau et al., 1991a, b;DeHairs et al., 1997;Freeman, 2001).In the open southwest Ross Sea, diatom and Phaeocystis biomass is −28 to −27 ‰ (Villinski et al., 2000).Sea-ice algae would add 13 C-enriched OC (Gleitz et al., 1996;Gibson et al., 1999;Villinski et al., 2000), but there is not a significant source of algal OC at the study site.Antarctic kerogen and coal have δ 13 C OC values of −22 and −24 ‰, respectively (Sackett, 1986;Burkins et al., 2000), and soils from the McMurdo Dry Valley region have a wide range of δ 13 C OC values: ∼ −30 to ∼ −18 ‰ (and C / N (a) of 11 ± 4) (Burkins et al., 2000).
Solvent lipid extracts (SLE) would help bridge the gap between bulk OC and individual biomarkers.In all three sample pairs, SLEs were depleted by up to 4 ‰ in 13 C compared with the respective bulk OCs (Fig. 4b).δ 13 C SLE values for both traps and sediments followed the same trend among samples as δ 13 C OC : δ 13 C AS−SLE > δ 13 C BS−SLE >δ 13 C RS−SLE .Lipids are a relatively abundant fraction of OC in phytoplankton and zooplankton but are usually only a few percent of OC in particulate matter and sediments (Wakeham et al., 1997;Wang and Druffel, 2001).They would thus not be major contributors to particulate and sedimentary δ 13 C OC compared to more abundant proteins and carbohydrates that are ∼ 4-6 ‰ enriched in 13 C relative to lipids.Intraclass comparisons of isotopic compositions in marine systems are few.A study of δ 13 C of particulate and sedimentary OC in the northeast Pacific and Southern Oceans found that total hydrolyzable amino acids (THAA) and total carbohydrates (TCHO) in phytoplankton, zooplankton and sediments were enriched in 13 C by about ∼ 2 and ∼ 3 ‰, respectively, relative to OC, whereas lipids were depleted by ∼ 4 ‰ relative to OC (Wang et al., 1998;Wang and Druffel, 2001).

Bulk radiocarbon isotopes
Radiocarbon isotopic values were 21 and 14 ‰ for BS and AS trap OC ( 14 C OC ) and −199 and −138 ‰ for BS and AS sediments, respectively (Fig. 4c and Tables 1-3, which also list radiocarbon ages and fraction modern, f m ).RS POM and SOM were depleted in 14 C (−208 and −355 ‰, respectively) relative to the two other sites. 14C DIC values for prebomb and post-bomb mixed-layer DIC (dissolved inorganic carbon) in the Black Sea are ∼ −70 and 100 ‰, respectively (Jones and Gagnon, 1994) and in the AS they are ∼ −60 and ∼ 70 ‰ (Stuiver and Östlund, 1983;Southon et al., 2002); these values determine the 14 C of autochthonous marine biomass.Pre-bomb and post-bomb 14 C DIC values in the Ross Sea are lower than elsewhere: −130 and −100 ‰, respectively (Berkman and Forman 1996;Gordon and Harkness 1992;Hall et al., 2010).In the AS and RS, upwelling of 14 C-depleted deep waters and the short residence time of surface waters lead to the high surface-water reservoir ages (∼ 700 years and ∼ 1100 years for AS and RS, respectively, vs. ∼ 400 years for open-ocean locations, including the BS; Stuiver and Braziunas, 1993;Siani et al., 2000;Southon et al., 2002;Key, 2004;Hall et al., 2010).Thus trap OCs displayed post-bomb signatures for their respective regions but additionally contained older carbon, especially in the Ross Sea.Sediments contained significant contributions of old carbon, and corrected 14 C OC ages (Tables 1-3) are significantly older than estimated geological ages (100-400 years, admittedly based on 14 C OC -derived sedimentation rates).
Trap and sediment SLEs for the BS and RS had higher 14 C SLE values than corresponding 14 C OC values (Fig. 4d): 14 C SLE values were −96 and −150 ‰ for BS trap and sediment, respectively, and −154 and −211 ‰ for RS trap and sediment.That the BS and RS trap lipid fractions had lower 14 C SLE values than 14 C OC values indicates that some old carbon was extractable (e.g., plant waxes and petroleum hydrocarbons).But the higher sediment 14 C SLE values than sediment 14 C OC suggested that residual OC remaining after solvent extraction must be, by extension, something still older, such as nonextractable kerogen, especially in the RS sediment.On the other hand, 14 C SLE of the AS trap material was enriched (66 ‰) compared to the 14 C OC value, but the sediment was depleted (−173 ‰) relative to its 14 C OC value.Thus solvent extraction of the AS trap mate- rial released a greater proportion of fresh, young algal lipid into the SLE, but extraction of the sediment recovered an SLE with a greater proportion of older lipid.In the only other reports, to our knowledge, that measured 14 C SLE (Wang et al., 1998;Wang and Druffel, 2001), plankton and sedimentary lipids in the northeastern Pacific and Southern oceans were, respectively, similar to or lower than 14 C of total OC.Lipids, THAA and TCHO all had similar 14 C values in plankton, but in sediments lipids usually had lower 14 C signatures than total hydrolyzable amino acids (THAA) and total carbohydrates (THCO).

Biomarker isotopic compositions
PCGC isolation for compound-specific isotope analyses targeted the most abundant biomarkers of marine, terrigenous plant and relict origins.Short-chain n-fatty acids (C 14 −C 24 , abbreviated as 14 : 0 FA, etc.), C 37 +C 38 -alkenones, a shortchain n-fatty alcohol [16ROH (n-hexadecanol)] and sterols [cholest-5-en-3β-ol (cholesterol), abbreviated as 27 5 ; 24-methylcholesta-5,22-dien-3β-ol, 28 5,22 ; and 4,23,24trimethylcholest-22-en-3β-ol (dinosterol), 30 22 ] were designated as marine biomarkers.Long-chain even-carbonnumber n-acids (C 26 −C 30 ) and n-alcohols (C 24 −C 28 ) and long-chain odd-carbon-numbered n-alkanes (C 27 and C 29 ) were tagged as terrigenous, vascular-plant markers; in some cases the long-chain fatty acids could be isolated individually but in other cases composites (e.g., C 24 +C 26 +C 28 fatty acids) were required.Long-chain even-carbon-numbered n−alkanes (composited C 24 +C 26 +C 28 ) and, in the AS sediment, short-chain C 15 +C 16 +C 17 +C 18 alkanes (including some unresolved complex mixture that could not be removed) were used as relict markers since these alkanes are not abundant in marine or terrestrial biomass.As will be shown below, using long-chain n-alkanes as either terrigenous plant or relict markers is problematic because, in the present samples, there is overlap between odd-vs.even-chain lengths and plant vs. relict sources.Thus it is possible that long-chain n-alcohols, if present in sufficient abundances, may represent the best "terrigenous" biomarkers since they would have neither modern marine (major alcohols in wax esters are C 16 and C 18 ; Wakeham, 1982) nor relict sources.
Highly branched C 25 alkenes (HBI) of diatom origin (Belt et al., 2000) were the most abundant hydrocarbons in AS trap material, less so in AS sediments, but they were underlain by a UCM that could not be removed, thus yielding unrealistically 14 C-depleted results.In the RS POM and SOM, only fatty acids, 16 ROH and sterols (cholest-5-en-3β-ol and 24-methylcholesta-5,22-dien-3β-ol) could be isolated.Stable carbon and radiocarbon isotope results were weighted according to their abundance within each sample to give means for each of the three sources (marine, terrigenous and relict) that are designated below as δ 13 C M and 14 C M , δ 13 C T and 14 C T , and δ 13 C R and 14 C R .

Black Sea
For the Black Sea POM, short-chain fatty acids, long-chain alkenones and sterols had δ 13 C values between −26.1 to −23.2 ‰ (Fig. 8a and Table 1), generally 1-4 ‰ more 13 Cdepleted than bulk POM (−22.9 ‰).Interestingly, alkenones were the most 13 C-depleted of this group.Overall these "marine" biomarkers had an abundance-weighted average δ 13 C M of −25.3 ± 1.1 ‰ (Table 4), typical for marine lipids.Radiocarbon values for the marine biomarkers (excluding the highly 14 C-enriched 18 : 1 fatty acid outlier at 288 ‰) ranged from 65 to 146 ‰, giving 14 C M of 78 ± 9 ‰, considerably enriched in 14 C relative to bulk POM ( 14 C OC −199 ‰) (Fig. 8b).Long-chain even-carbon-number acids and alcohols and odd-carbon-number alkanes had δ 13 C values ranging from −30.1 to −27.0 ‰, yielding a mean δ 13 C T of −28.7 ± 1.6 ‰, roughly 3‰ depleted in δ 13 C compared to marine lipids.Terrigenous markers had a much greater range of 14 C values (1 ‰ for the fatty acids, −44 ‰ for the alcohols and −181 ‰ for the alkanes) but a mean 14 C T of −75 ± 94 ‰.It is possible that some fraction of the long-chain fatty acids are zooplankton (wax ester)-derived young carbon, whereas the long-chain alkanes may contain a relict component.Thus, the long-chain alcohols may represent the best "terrigenous" biomarkers since they would have neither modern marine (major alcohols in wax esters are C 16 and C 18 ; Wakeham, 1982) nor relict sources with a 14 C of −1000 ‰.The single sample of "relict" biomarkers isolated by PCGC was a composite of C 24 +C 26 +C 28 nalkanes that had a δ 13 C R of −29.1 ± 0.2 ‰ and a 14 C R of −677 ± 10 ‰; this 14 C value probably also reflects a mix of moderately pre-aged terrigenous ( 14 C of −44 ‰ might be reasonable if the alcohols are a good representative of terrigenous OC) and radiocarbon-dead ( 14 C R of −1000) relict carbon.
Marine biomarkers in the BS sediment had δ 13 C values ranging from −31.9 to −26.0 ‰ (mean δ 13 C M −28.8 ± 1.8‰), about 2.5 ‰ depleted in 13 C relative to both BS bulk SOM (−25.3 ‰) and the marine group of BS POM (also −25.3 ‰).Radiocarbon contents of marine biomarkers ranged from −46 to 75 ‰ (mean 14 C M 30 ± 10 ‰, excluding 16 : 0 at 214 ‰ and 18 : 0 at 374 ‰), considerably enriched in 14 C relative to bulk POM ( 14 C OC −199 ‰) but depleted in 14 C compared to marine biomarkers in BS POM.Terrigenous biomarkers displayed δ 13 C values between −31.9 to −29.7 ‰ (mean δ 13 C T −30.5 ± 0.7 ‰).Thus, although there was a considerable overlap in δ 13 C values for marine and terrigenous groups, concentration weighting yielded an offset of ∼ 2 ‰, as would be expected.Plantwax alkanes (C 27 and C 29 ) and alcohols (C 24 and C 26 ) were strongly depleted in 14 C relative to the marine lipids (range −231 to −100 ‰), with a mean 14 C T of −171 ± 58 ‰.Long-chain even-carbon-numbered [C 24 +C 26 +C 28 ] alkanes had a δ 13 C value of −29.2 ‰ and a 14 C R of −609 ± 8 ‰, again suggesting they are pre-aged but not exclusively relict.A similar spread in δ 13 C and 14 C values for marine, vascular-plant and relict-hydrocarbon biomarkers has been previously reported for Black Sea and Arabian Sea sediments (Eglinton et al., 1997).

Arabian Sea
The marine biomarkers in the AS trap POM had δ 13 C values between −27.1 and −16.7 ‰ (mean δ 13 C M −23.5 ± 2.5 ‰). ) and 14 C contents ranging from −91 to 111 ‰ (Fig. 9a and  b and Table 2) (mean 14 C M 64 ± 20 ‰) , excluding C 14:0 at 178 ‰ and HBI alkenes at −514 ‰).The negative 14 C value for the HBI alkenes is due to the inclusion of a UCM which could not be removed.The only vascular-plant compounds in the AS trap in sufficient quantity for AMS analysis were analyzed as a composite of [C 27 +C 29 ] n-alkanes, giving a δ 13 C T of −28.4 ‰ and a 14 C T of −320 ± 24 ‰.The single sample of [C 24 +C 26 +C 28 ] n-alkanes gave a δ 13 C R of −28.5 ± 0.2 ‰ and a 14 C R of −731 ± 14 ‰.

Ross Sea
Ross Sea POM and SOM contained only fatty acids and alcohols/sterols in sufficient abundance for compound-specific isotopic analysis.In keeping with a 13 C isotope depletion in high-latitude regions, noted above, fatty acids had δ 13 C values ranging from −35.3 to −31.7 ‰ and alcohol/sterol values ranging from −35.4 to −31.8 ‰ (Fig. 10a and Table 3), together giving a mean δ 13 C M of −33.2 ± 1.8 ‰.Radiocarbon contents ranged from −216 to −100 ‰, with the alcohols/sterols slightly depleted relative to the fatty acids, for a 14 C M of −155 ± 47 ‰ (Fig. 10b).Fatty acids and sterols/alcohols in RS sediments had δ 13 C values ranging from −36.6 to −31.7 and −34.3 to −33.0 ‰, respectively, for a mean δ 13 C M of −33.2 ± 1.9 ‰.Radiocarbon contents of fatty acids from RS sediments were more variable than for RS POM, with values ranging from −302 to −83 ‰.Sterols from RS sediments were very unchanged from sterols in POM with 14 C values between −202 and −178 ‰ for sterols.Overall, this gave a mean 14 C M of −105 ± 31 ‰, also essentially the same as for 14 C M of the trap material.The two vascular plant FAs (24 : 0 and 26 : 0) gave a δ 13 C T value of 31.8 ± 0.28 ‰ and a 14 C T value of 255 ± 66 ‰.

The provenance of POM and SOM in the Black Sea and Arabian Sea
Concentration-weighted isotope values (Table 4) can be used to constrain the ranges of isotopic compositions of marine, terrigenous and relict biomarkers and, by extension, organic carbon, of Black Sea and Arabian Sea POM and SOM.Few biomarkers could be isolated from the Ross Sea in sufficient quantities for AMS analyses.Overall, δ 13 C M > δ 13 C T ∼ δ 13 C R and 14 C M > 14 C T 14 C R (Fig. 11).Furthermore, SOM biomarkers were generally 13 C-depleted and 14 C-depleted relative to their corresponding POM samples, indicative of higher proportions of 13 C-depleted but older, pre-aged OC in sediments.The range of δ 13 C R was relatively small compared to δ 13 C M and δ 13 C T of POM and SOM, but the span of 14    mix of old but not radiocarbon-"dead" terrigenous vascularplant ( 14 C values similar to average 14 C T values) and truly relict ( 14 C = 1000 ‰) OC.
The relative contributions of marine, pre-aged terrigenous and relict OC (f M , f T , andf R ) to sediment trap material and sediments can be estimated in the Black, Arabian and Ross Seas using an isotopic mass balance.Plots of δ 13 C M, T, R vs.
14 C M, T, R in both the sediment traps and sediment (Fig. 11) show significantly different values for the radiocarbon content of the different carbon pools but relatively small differ- ences in the stable isotopic content in the Black and Arabian seas and in the Ross Sea sediments.The small differences in δ 13 C coupled with significant error make it difficult to use the stable isotopic composition as a discriminating factor in determining the sources of the bulk material.However, the large differences observed in the 14 C values make it possible to use these data to constrain the relative amounts of relict, terrestrial and marine components.
For each sample, there is a 14 C value assigned to the bulk SLE, fresh marine material, fresh terrestrial material and a relict-terrestrial mix.Bulk material is assumed to be a mixture of fresh marine and other "added" material, where the added material refers to the fresh terrestrial and relictterrestrial mix.Given this, we can derive the relative amounts of marine and "added" material.
solving this equation for f added as follows: The value of 14 C added can range from that for relict material (−1000 ‰) to that of the bulk material.Using this information, we can calculate the relative amounts of marine and "added" material as a function of the 14 C of the "added" material.Further, we can combine the radiocarbon value measured on the terrestrial portion with the value of truly relict material to calculate how much of the "added" material might come from fresh terrestrial material using the equations below.
Truly relict material has a 14 C value of −1000 ‰, and fresh terrigenous material has the values determined in this study ( 14 C T , Table 5).As stated earlier, the measured value of 14 C R appears to be a mixture of truly relict material and terrigenous material with a suite of ages.If we assume that the value we measured for "relict" material (Table 5) is a good representation of 14 C added , then we have a unique solution to the mass balance.This solution is indicated by the vertical lines in Fig. 12 and the values listed in Table 5.Using this model, we force ourselves to an extreme situation where all the "added" material is either vascular or totally dead.Thus, the relative fractions of terrestrial and relict material must be considered maximum and minimum values, respectively.
For the calculations, concentration-weighted δ 13 C M , δ 13 C T , δ 13 C R, δ 13 C B , 14 C M , 14 C T and 14 C B values listed in Table 4 were used.The relative fractions of marine, terrestrial and relict material (f M , f T , and f R ) as a function of 14 C added are shown in Fig. 12.It is most likely that the marine component is dominant in both the trap and sediment samples, although the possible values range from > 80 to 0 %.Fresh vascular material can account for 0 up to almost 100 %.The amount of relict material is constant at a low proportion, ranging from 0 to < 20 % in both the trap and sediments.In the AS, the marine component is dominant in both the trap and sediments; the trap data indicate that there can be virtually no relict or vascular material input to this sample, while the sediment data suggest that up to 30 % of the material may come from either relict or vascular sources.In the RS sediment, marine material dominates in almost all instances, with possible values ranging from 90 to 30 %. Relict material can account for > 10 to 0 % and vascular from 0 to < 70 %.
In the Black Sea and Arabian Sea cases here, it is likely that the marine component contributes the most material because a marine biogenic flux dominates at both sites.No comparable combined biomarker, 13 C and 14 C studies of POM at other sites exist to our knowledge.But the high amounts of marine OC in the Black Sea and Arabian Sea sediments contrasts with other environments where similar biomarker-isotopic mass balances have been made, although using a narrower range of biomarkers.In nearshore sediments on the northern California Margin off the high-energy Eel River that erodes ancient sedimentary rocks in its watershed (Blair et al., 2003), fractional contributions from marine, terrestrial and relict components were estimated at 0-10, 50-75 and 30-50 % of total OC (Drenzek et al., 2009).On the Washington Margin off the lower-energy Columbia River, marine, terrestrial and relict OC contribute 2, 89-95 and 3-9 % of bulk sedimentary OC (Feng et al., 2013).Sediments in the Beaufort Sea had f M , f T and f R values of 13-27, 36-42 and 34-37 %, respectively, with the elevated f R values consistent with the well-defined petrogenic signature of n-alkanes and polycyclic aromatic hydrocarbons in Mackenzie River and Beaufort Sea sediments (Yunker et al., 1993(Yunker et al., , 2002)).For surface sediments in the southwest Black Sea near our study site, Kusch et al. (2010) estimate an f R of ∼ 18 % of OC.In the Santa Monica Basin of the California Borderlands, 80-87 % of the n-alkanes were of terrigenous plant origin, whereas up to 20 % of the alkanes were derived from petroleum or shales (Pearson and Eglinton, 2000), but since alkanes are not abundant in most marine biota, no estimation of the marine component was made.
For the Black Sea, the graphs for the relative contributions look similar for both the trap and sediment samples.A strict interpretation of these results suggests that there is a greater relative amount of nonmarine material in the sediment trap than in the sediments themselves.Given the limitations of our data set, we believe that it is more likely that the relative amounts of material are actually very similar.This is somewhat surprising for two reasons.First, under a differential degradation/preservation case (Wakeham and Canuel, 2006), a more labile component of marine OC would be selectively degraded as POM moved between the trap depths and the surface sediments, leaving behind increased proportions of selectively preserved, more refractory terrigenous and relict components in sediments.Alternately, lateral advection of terrigenous and relict OC below the trap depths either by mid-depth or bottom currents might simply allow this pre-aged material to bypass the traps to be deposited directly on the sediments.advective transport of continental material containing terrigenous and relict carbon to the study site is likely via surface, mid-water (∼ 150 m depth) or bottom-water nepheloid layers and turbidites.Surface and mid-water plumes moving offshore could carry some continental material to the shallow (∼ 250 m) sediment trap.But a greater amount of terrigenous and relict carbon might, after temporary storage on the narrow continental shelf, be remobilized to move laterally down the steep continental slope under the trap.This type of advective remobilization of pre-aged material is well documented at other locations for both marine-derived alkenones (e.g., Ohkouchi et al., 2002;Mollenhauer et al., 2003) and terrigenous/relict OC (e.g., Aller et al., 2004;Mead and Goñi, 2006;Mollenhauer and Eglinton, 2007;Kusch et al., 2010;Hwang et al., 2010).Aeolian transport of leaf wax OC and petrogenic OC to the sediments of the central Black Sea have been reported (Wakeham 1996;Eglinton et al., 1997), but the importance of aeolian delivery to the southwestern Black Sea is unknown.Nonetheless, a small leaf wax/petrogenic signal was detected in the trap POM.Because the trap was already deployed within the anoxic zone (which starts at 120-150 m), OC degradation in the water column and surface sediments should be depressed, so that degradation might not be a significant cause for any reduction in f M in the sediments.
In the Arabian Sea, the sediment trap material indicates that the material is all marine while the sediments show a small influence of nonmarine material.The Arabian Sea site was in an area of intense upwelling dominated by high export of diatomaceous material (Wakeham et al., 2002), remote from fluvial inputs but potentially affected by aeolian transport of pre-aged OC off the Arabian Peninsula and Horn of Africa during the windy monsoon periods (Dahl et al., 2005).The AS trap was deployed within the OMZ where organic matter degradation has reduced dissolved oxygen concentrations to ∼ 5 µM (Smith et al., 1998), but AS sediments were collected at 1400 m water depth where bottom waters are oxygenated.Indeed within and below the OMZ, OC fluxes decreased 5-10 fold between the trap and surface sediments, even as %OC did not change as dramatically (Lee et al., 2000 for OC;Wakeham et al., 2002 for AS lipid fluxes).Comparative studies have shown conclusively that, among other things (e.g., intrinsic reactivity of organic molecules and protection by macromolecular organic matrices and mineral surfaces), oxygen availability is a key control on OC and lipid degradation/preservation in water columns and sediments (e.g., Hedges and Keil, 1995;Gong and Hollander, 1997;Wakeham and Canuel, 2006;Burdige, 2007;Mollenhauer and Eglinton, 2007).Enhanced degradation (diminished preservation) in the oxygenated AS should be greater than in the anoxic Black Sea, and if marine OC is more labile than terrigenous/relict OC, f M in the AS POM and SOM would, as observed, be higher than in the BS.The extensive production of petroleum reserves on the Arabian Peninsula and intense tanker traffic in the Arabian Sea is an additional source of refractory and radiocarbon-dead petrogenic OC to AS particulate matter and sediments.
The RS sediment results are consistent with previous studies suggesting that marine material is the most important source of organic matter in the sediments (Arrigo et al., 2002;Dunbar et al., 2003).The solutions that suggest a relatively large input of vascular-plant material seem unrealistic given these prior studies.The Ross Sea data set is very limited and in fact does not include unambiguous lipids of either terrestrial or relict origins with which to constrain the isotopic compositions of these endmembers.

Conclusions
This study examined the compound-specific 13 C and 14 C compositions of diverse biomarker indicators of marine biomass, terrigenous vascular-plant and relict sources of organic carbon in sediment traps and underlying surface sediments in the Black Sea the Arabian Sea and the Ross Sea.Using an isotopic mass balance approach, it was possible to constrain relative inputs from these three sources, with marine biomass accounting for 66-100 % of extractable lipids and organic carbon in Black Sea and Arabian Sea sediment trap material.The remaining 3-8 % derive from terrigenous and 4-16 % from relict sources.Sediments contained lower proportions of marine biomarkers (66-90 %) and consequentially higher proportions of terrigenous and relict carbon (3-17 and 7-13 %, respectively).These results suggest that although particulate organic carbon is overwhelmingly marine in origin, there are significant proportions of pre-aged terrigenous and relict OC present.Because these latter fractions become proportionally more important in sediments, it is likely that they are better preserved than the marine component and/or that they reach the sediments by lateral advection rather than only by the vertical sinking that affects the upper ocean-derived marine POC.This approach demonstrates the strengths, and limitations, of such a multiparameter approach for studying marine OC cycling and budgeting.

Figure 1 .
Figure 1.Sampling locations in the Black Sea, Arabian Sea and Ross Sea.

Figure 2 .
Figure 2. Scheme of extraction, isolation and analysis of biomarkers.

Figure 4 .
Figure 4. (a) The δ 13 C and (b) 14 C values of bulk organic carbon (OC) and (c) δ 13 C and (d) 14 C values solvent extractable lipids (SLE) for trap POM and sediments.

Figure 5 .
Figure 5. Histograms of relative abundances of (a) fatty acids, (b) hydrocarbons and (c) sterols/alcohols in POM and SOM from the Black Sea.Carbon numbers are given for fatty acids, alkanes and alcohols; sterol abbreviations are 27D5,22 = 27 5,22 , etc.

Figure 6 .
Figure 6.Histograms of relative abundances of (a) fatty acids, (b) hydrocarbons and (c) sterols/alcohols in POM and SOM from the Arabian Sea.Carbon numbers are given for fatty acids, alkanes and alcohols; sterol abbreviations are 27D5,22 = 27 5,22 , etc.

Figure 7 .
Figure 7. Histograms of relative abundances of (a) fatty acids, (b) hydrocarbons and (c) sterols/alcohols in POM and SOM from the Ross Sea.Carbon numbers are given for fatty acids, alkanes and alcohols; sterol abbreviations are 27D5,22 = 27 5,22 , etc.

Figure 8 .
Figure 8.The (a) δ 13 C and (b ) 14 C values of bulk POC and SOC and individual biomarkers for trap and sediments from the Black Sea (see also Table 1).Filled symbols are trap biomarkers; open symbols are sediment biomarkers.All values have been corrected for procedural blanks and any derivative carbon, as needed.Dashed lines are pre-bomb (lower) and post-bomb (upper) mixedlayer 14 C DIC .

Figure 9 .
Figure 9.The (a) δ 13 C and (b) 14 C values of bulk POC and SOC and individual biomarkers for trap and sediments from the Arabian Sea (see also Table 2).Filled symbols are trap biomarkers; open symbols are sediment biomarkers.All values have been corrected for procedural blanks and any derivative carbon, as needed.Dashed lines are pre-bomb (lower) and post-bomb (upper) mixedlayer 14 C DIC .

Figure 10 .
Figure 10.The (a) δ 13 C and (b) 14 C values of bulk POC and SOC and individual biomarkers for trap and sediments from the Ross Sea (see also Table 3).Filled symbols are trap biomarkers; open symbols are sediment biomarkers.All values have been corrected for procedural blanks and any derivative carbon, as needed.Dashed lines are pre-bomb (lower) and post-bomb (upper) mixedlayer 14 C DIC .

Figure 11 .
Figure 11.Concentration-weighted average δ 13 C and 14 C values of marine (M), terrigenous (T) and relict (R) lipids in (a) Black Sea, (b) Arabian Sea and (c) Ross Sea traps and sediments.

Figure 12 .
Figure 12.Relative amounts of marine (M, blue solid line), terrestrial (T, red dashed line) and relict (R, black dashed line) carbon as a function of 14 C added in Black Sea and Arabian Sea trap material (a and c, respectively) and in Black Sea, Arabian Sea and Ross Sea sediments (b, d, and e, respectively).Only real solutions are depicted.The solid black line indicates the solution when14 C added is equal to the value measured on compounds defined as relict.There are not enough data to construct a graph for the Ross Sea sediment trap.

Table 1 .
Stable carbon and radiocarbon isotope data for Black Sea POM and SOM.Bold italicized values were not used in calculations (see text).
M stands for marine; T for terrigenous; R for relict FA for fatty acid; ROH for alcohol; HC for alkane

Table 2 .
Stable carbon and radiocarbon isotope data for Arabian Sea POM and SOM.Bold italicized values were not used in calculations (see text).
from the heavily wooded Anatolian coast than was present in the short- term trap sample.The relative enrichment of the AS sediment compared to the trap material could be the result of the addition of isotopically enriched aeolian-transported OC derived from C 4 -grasses in the arid Horn of Africa and Arabian www.biogeosciences.net/11/6895/2014/Biogeosciences, 11, 6895-6914, 2014

Table 3 .
Stable carbon and radiocarbon isotope data for Ross Sea POM and SOM.Bold italicized values were not used in calculations (see text).
FA stands for fatty acid; ROH stands for alcohol

Table 4 .
Abundance-weighted mean table carbon and radiocarbon isotope values for composited marine, terrigenous and relict biomarkers.

Table 5 .
Isotopic values used in mass balance calculations.The values of f M , f T and f R are those values calculated assuming that the actual value of 14 C added is that measured on the compounds defined as relict.