Contrasting lipid biomarker composition of terrestrial organic matter exported from across the Eurasian Arctic by the five great Russian Arctic rivers

Authors

  • Bart E. van Dongen,

    1. Department of Applied Environmental Science (ITM), Stockholm University, Stockholm, Sweden
    2. Now at School of Earth, Atmospheric and Environmental Sciences, University of Manchester, Manchester, UK.
    Search for more papers by this author
  • Igor Semiletov,

    1. International Arctic Research Center, University of Alaska, Fairbanks, Alaska, USA
    2. Pacific Oceanological Institute, Russian Academy of Sciences, Vladivostok, Russia
    Search for more papers by this author
  • Johan W. H. Weijers,

    1. Department of Marine Biogeochemistry, Royal Netherlands Institute for Sea Research, Texel, Netherlands
    2. Now at Organic Geochemistry Unit, Bristol Biogeochemistry Research Centre, School of Chemistry, Bristol University, Bristol, UK.
    Search for more papers by this author
  • Örjan Gustafsson

    1. Department of Applied Environmental Science (ITM), Stockholm University, Stockholm, Sweden
    Search for more papers by this author

Abstract

[1] Surface sediments outside the great Russian Arctic rivers (GRARs; Ob, Yenisey, Lena, Indigirka, and Kolyma) were investigated for their lipid biomarker composition to elucidate compositional distinctions of the exported organic matter (OM) across this continent-scale climosequence of the Siberian Arctic. The lipid biomarker composition is dominantly terrestrial (high molecular weight (HMW) n-alkanols, n-alkanoic acids, n-alkanes; branched glycerol dialkyl glycerol tetraethers (GDGTs), β-sitosterol, other sterols, and triterpenoids) with only minor marine contributions (e.g., the ratio of terrigenous-to-aquatic n-alkanes was 17–80, the TOC/TN ratio was 10–16, and the branched and isoprenoid tetraether (BIT) index >0.88). There is a large contribution of C23 and C25 homologues relative to other long-chained n-alkanes, suggesting substantial contribution of probably Sphagnum-derived OM. The C23 and C25 contribution decreases eastward, signaling either a decrease in the potential contribution of Sphagnum or a shift within the n-alkane distribution of Sphagnum species or increased aeolian input, due to more arid conditions in the east. Other distinctions in molecular OM composition across the climosequence include increased concentrations of both HMW n-alkanoic acids and β-sitosterol relative to HMW n-alkanes in the eastern sediments. This suggests that the OM exported by the eastern GRARs is, despite their higher bulk radiocarbon ages, less degraded, which is consistent with increasing permafrost and a shorter annual thaw period in eastern Siberia. Taken together, this benchmark study of the current composition of terrestrially exported OM suggests distinguishing continent-scale trends in molecular composition of the OM across the west-east set of GRARs, which reflects both differences in vegetation and climate. If the climate in the eastern Russian Arctic region becomes more like the current state in the western part, these results would predict a greater degree of decomposition of the old terrestrial OM released by the eastern GRARs and thus greater remineralization and release as CO2 and methane.

1. Introduction

[2] Estuarine surface sediments of major rivers provide an integrated archive of terrestrial organic matter (OM) exported from the heterogeneous mosaic of the upstream drainage basin. River discharge is particularly important for the Arctic Ocean since, although it contains only 1% of the world ocean volume [Menard and Smith, 1966], it receives 11% of the global input of freshwater and terrestrial dissolved organic matter (DOM) [Aagaard and Carmack, 1989; Gordeev et al., 1996; Anderson et al., 1998; Macdonald et al., 1998]. The drainage basins of the Arctic rivers store large amounts of OM in their taiga and tundra soils, which account for 25–33% of the entire global soil reservoir of OM [Waelbroeck et al., 1997; Oechel et al., 2000].

[3] Numerical climate models forecast an amplified warming in the continental Arctic region [e.g., Zwiers, 2002]. This makes it reasonable to expect that the Arctic is one of the first areas where climate warming effects on large-scale biogeochemical cycles will be observed [Macdonald, 1996; Dickson et al., 2000; Serreze et al., 2000] most pronounced in the east Siberian region [ACIA, 2004]. For instance, the huge carbon inventory in Siberian soils is predicted to be affected by global warming through permafrost destabilization [Stendel and Christensen, 2002] with wide-reaching consequences such as increasing river discharge [e.g., Savelieva et al., 2000; Peterson et al., 2002; 2006; Yang et al., 2002], and changing plant cover. This will in turn change the input fluxes and composition of terrigenous carbon to coastal seas [Dittmar and Kattner, 2003; Benner et al., 2004; Guo et al., 2004]. The provenance and molecular composition of OM are key characteristics that affect the large-scale biogeochemical cycling of carbon. However, it is still largely unclear how system responses to climate change will impact the delivery of terrestrial OM to the Arctic Ocean. While the OM entering the Mackenzie–Beaufort Sea system has been well characterized at a molecular compositional level [e.g., Yunker et al., 1995, 2002; Goñi et al., 2000, 2005; Belicka et al., 2004; Drenzek et al., 2007], such important baseline data is much less detailed for the Eurasian Arctic and then only for single rivers [e.g., Peulve et al., 1996; Fahl and Stein, 1997; Zegouagh et al., 1998; Fernandes and Sicre, 2000; Fahl et al., 2003; Stein and Macdonald, 2004]. The major objective of the present study is to provide a detailed and coherent benchmark of the lipid biomarker composition of the terrestrial OM exported by the great Russian Arctic rivers (GRARs; the Ob, Yenisey, Lena, Indigirka, and Kolyma), spanning a continent-scale (140° longitude) variation in climate, permafrost, hydrology and vegetation coverage. A simultaneous investigation of all five GRARs offers the possibility to evaluate links between specific features of this climosequence and the molecular-level OM composition, which may prove beneficial in predicting the effects on OM composition and thus large-scale carbon cycling under a changing climate.

2. Materials and Methods

2.1. Study Area and Sampling

[4] The five GRARs have different drainage basin areas, hydrology and annual discharge (Table 1) and the locations of their river mouths span over 4000 km along the Eurasian rim of the Arctic Ocean (Figure 1). The watersheds of the two west Siberian rivers (Ob and Yenisey) are mainly located in the region of nonpermafrost and/or permafrost islands and flow into the Kara Sea. The east Siberian rivers, on the other hand, are located in the permafrost region that covers approximately half of the territory of Russia and flow into the Laptev Sea (Lena) and the East Siberian Sea (Indigirka and Kolyma; Figure 1a). In addition, both the Kolyma and Indigirka rivers have their headwater almost entirely within the Arctic. In contrast, the other three rivers begin their flow far south of the Arctic region [Woo, 1992], with the Arctic defined as the area north of the 10°C July isotherm [AMAP, 1998].

Figure 1.

(a) Map of the Siberian Arctic region, based on a map obtained from the National Geophysical Data Center, National Oceanic and Atmospheric Administration, U.S. Department of Commerce [Jakobsson et al., 2000], with color-coded bathymetry and topography, flow paths of the great Russian Arctic rivers (GRARs) as thick black lines, sampling locations as open circles, and sample ID as mentioned in Table 1. (b) Land cover map of Russia based on the SPOT4-vegetation satellite data (http://terranorte.iki.rssi.ru) [Bartalev et al., 2003; Bartholomé and Belward, 2005; Groisman and Bartalev, 2006] with flow paths of the GRARs as black lines.

Table 1. Sample Location, Hydrological Data, and Bulk Geochemical Data of the Surface Sediments
 ObYeniseyLenaIndigirkaKolyma
Sample ID12N-8IK43-46aIK-71
Geological and physiographic regionsbwest Siberian lowlandswest Siberian lowlandscentral Siberian plateaueast Siberian highlandseast Siberian highlands
Latitude; longitude72°65′N;73°44′E72°61′N;79°86′E71°96′N;129°54′E72°06′N;150°46′E–71°02′N;152°60′E70°00′N;163°70′E
Basin area, 106 km2c2.54–2.992.44–2.592.43–2.490.360.65–0.66
Water runoff, km3/ac402–429580–620528–53054–61103–132
TSM, 106t/ac,d13.4–16.513.2–14.511.7–17.612.916.1
TOC, mg/ge9.2 ± 0.219.4 ± 0.34.8 ± 0.214.6 ± 0.217.3 ± 1.1
δ13CTOC, ‰−27.4 ± 0.1−26.5 ± 0.1−25.0 ± 0.1−26.6 ± 0.1−26.7 ± 0.1
TNf, mg/g0.92 ± 0.011.85 ± 0.030.39 ± 0.030.99 ± 0.021.09 ± 0.04
TOC/TN mass ratio10.0 ± 0.110.5 ± 0.112.3 ± 0.814.7 ± 0.215.9 ± 1.2
Δ14C, ‰g−314 ± 3−175 ± 3−608 ± 3−527 ± 3−503 ± 2

[5] The climate of northeastern Eurasia (central Siberian plateau and the east Siberian highland), containing the drainage basins of the three eastern GRARs (Figure 1), is severe, with average summer temperatures between +7°C and +9°C, winter temperatures below −40°C and semiarid–arid conditions. The most arid conditions can be found in the basins of the Kolyma and Indigirka rivers [FAO, 2001; Naumov, 2004]. In contrast, the west Siberian lowland, the drainage basin of the two western GRARs Ob and Yenisey, holds a transitional place between this extreme continental climate and the less extreme Atlantic climate of Europe, with average summer temperatures comparable to northeastern Eurasia but much higher winter temperatures (around −20°C) [FAO, 2001; Kremenetski et al., 2003; Karavaeva, 2004]. Consequently, the drainage areas of these two western rivers, and to a lower degree the Lena River, have a higher annual precipitation and higher precipitation to evaporation ratios compared to the drainage areas of the Kolyma and Indigirka rivers.

[6] The GRARs have a combined freshwater discharge around 1700 km3/a, by far the greatest portion of the total freshwater discharge to the Arctic Ocean [Opsahl et al., 1999]. The majority of this annual discharge (>80%) is concentrated in the time period from April to June. The hydraulic discharge of Indigirka and Kolyma are combined just about 10% of the GRAR total (Table 1). In contrast, the total suspended matter (TSM) discharge is comparable for all five GRARs (Table 1). This relatively lower hydraulic discharge but higher TSM yield of the eastern rivers is comparable to the North American Mackenzie River [Gordeev et al., 1996], although the amount of solids that the Mackenzie River delivers is far greater than the Siberian rivers [Yunker et al., 1995; Macdonald et al., 1998; Stein, 2000; Yunker et al., 2002]. This likely reflects the greater relief and thus a stronger weathering of the east Siberian highland drainage basins of Indigirka and Kolyma. It should be noted that besides river transport also coastal erosion forms a significant flux of TSM to the Arctic, especially in near shore zones [Semiletov, 1999; Semiletov et al., 2005].

[7] The vegetation cover for all drainage basins varies from north to south (Figure 1b). The northern areas are characterized by different tundra and wetland vegetation. The west Siberian lowland is the world's largest high-latitude wetland, and possesses over 900,000 km2 of peatlands [Kremenetski et al., 2003]. The dominant vegetation types are mosses (mainly Sphagnum), sedges, shrubs and trees. Southward, particularly in the drainage areas of the Ob, Yenisey, and Lena rivers, the vegetation becomes more dominated by different types of forests (Figure 1b) [Bordovskiy, 1965; FAO, 2001; Kremenetski et al., 2003; Bartholomé and Belward, 2005; Groisman and Bartalev, 2006].

[8] Surface sediment samples of the Lena, Indigirka and Kolyma estuaries were collected using a van Veen grab sampler (dimensions 20 × 30 cm), a light weight sampler designed to take large samples in soft bottoms [Riddle, 1989; Somerfield and Clarke, 1997], during the second Russia-United States cruise 2004 in the east Siberian and Laptev seas. Typically, between 2 and 4 kg of sediments were collected to have enough material for subsequent compound-specific radiocarbon analysis. For this reason sediment from 4 locations (IK43-IK46) were combined for the Indigirka estuarine sample to provide a sufficiently large sample. In addition, surface sediment samples of the Ob and Yenisey estuaries were obtained during the third Russia-United States cruise 2005 in the Kara Sea, using the same van Veen grab sampler, used by the same operator and using the same protocol. Detailed sampling locations and other parameters are listed in Table 1. In all cases sediment integrity was first visually inspected before approximately 0–2 cm was measured manually and subsampled from the van Veen sampler; the rest was discarded. The samples were kept frozen at −20°C until processed in the laboratory.

2.2. Extraction and Fractionation

[9] The bulk sediments were thoroughly mixed prior to sub sampling and small amounts of material were used for elemental analyses (performed by the UC Davis Stable Isotope Facility; http://stableisotopefacility.ucdavis.edu) and radiocarbon analyses (at the National Ocean Sciences Accelerator Mass Spectrometry Facility at Woods Hole Oceanographic Institution). For biomarker analysis, sub samples, typically between 10 and 20 g, were freeze dried, a mixture of three standards (tetracosane-2H50, tricontane-2H62 and hexadecan-2-ol) was added, and the sediments were extracted using a Soxhlet apparatus with dichloromethane/methanol (DCM/MeOH, 2:1, v/v) for 24 h. The total lipid extracts (TLEs) were concentrated using rotary evaporation and treated with activated (2N HCl) copper curls to remove elemental sulfur. Subsequently, aliquots of the TLEs were separated into three fractions using Bond-Elut® column chromatography (Strata NH2; 5 μm, 70 A; Varian, Middelburg, Netherlands) [e.g., Kim and Salem, 1990] and elution with DCM/isopropanol (2:1 v/v; 12 ml; “neutral lipid fraction”), an acetic acid solution (a 2% solution in diethyl ether; 12 ml; “acid fraction”) and MeOH (12 ml; “phospholipid fraction”). For the present study, further analyses were only performed on the neutral lipid and acid fractions. The neutral lipid fraction was further separated into two fractions using a column packed with 100% activated alumina by elution with hexane/DCM (9:1 v/v; 3 ml; “hydrocarbon fraction”) and DCM/MeOH (1:1 v/v; 3 ml; “polar fraction”). The hydrocarbon, polar and acid fractions were analyzed using gas chromatography/mass spectrometry (GC/MS). Blanks were run to ensure that no contamination was introduced during the extraction and separation procedure and were found to be <1:100 for all target analytes. The acid fractions were, after addition of an eicosanoic acid-2H39 standard, derivatized with BF3 in MeOH to convert acids into their corresponding methyl esters. Subsequently, very polar compounds were removed by column chromatography over silica gel with DCM as eluent. Prior to analyses the polar fractions were dissolved in bis(trimethylsilyl)trifluroacetamide (BSTFA) and heated (70°C; 60 min) to convert alcohols into their corresponding trimethylsilyl ethers. Subsequently, hexane was added and the hexane layers were washed three times with small amounts of acetonitrile.

[10] Analyses of glycerol dialkyl glycerol tetraether (GDGT) membrane lipids of the east Siberian River estuary sediments were carried out on the polar fractions using the method reported by Hopmans et al. [2004]. In short, the polar fractions were dissolved in hexane/isopropanol (99:1 v/v) and filtered using a 0.45 μm cutoff, 4 mm diameter polytetrafluoroethylene filter. The filtered fractions were analyzed using high-performance liquid chromatography atmospheric pressure chemical ionization mass spectrometry (HPLC-APCI-MS).

2.3. Instrumental Analyses

[11] Identification and quantification was performed using a Thermoquest Finnigan GC8000, equipped with an on-column injector and helium as the carrier gas, interfaced to a Fison MD800 MS. A Supelco Equity™-5 fused silica capillary column (30 m x 0.25 mm) coated with 5% diphenyl/95% dimethylsiloxane (film thickness 0.25 μm) was used. The hydrocarbon and polar fractions were injected at 50°C and the oven was programmed to 130°C at 20°C/min and then at 4°C/min to 300°C where it was held isothermal for 20 min. The acid fractions were injected at 70°C and the oven was programmed to 130°C at 20°C/min and then at 10°C/min to 300°C where it was held isothermal for 10 min. The mass spectrometer was operated with electron ionization at 70 eV, scanning a mass range of m/z 50 to 650 using a cycle time of 2 scans s−1. The interface was set to 250°C with the ion source at 200°C. Compounds were identified by comparison of mass spectra and retention time of standards, when possible, or with those reported in literature. The n-alkanes, n-alkanoic acids, n-alkanols, and sterols were quantified from the m/z 57 + 71, m/z 74 + 87 + 143 and m/z 75 mass chromatograms and total ion current (TIC) chromatograms, respectively.

[12] Identification and quantification of the GDGTs was performed using an HP1100 series LC-MS equipped with an autoinjector (Palo Alto, California, United States). Column, flow rates and gradient conditions of the eluent were similar to those described by Hopmans et al. [2004]. Detection was achieved using atmospheric pressure positive ion chemical ionization mass spectrometry (APCI-MS) of the eluent using the same conditions as described by Hopmans et al. [2004]. GDGTs were detected by single ion monitoring of their [M + H]+ ions, quantified by integration of the peak areas and comparing these to an external calibration curve prepared with known amounts of pure crenarchaeol (IV; Figure 2d). The recovery of the internal standards tetracosane-2H50, tricontane-2H62 and hexadecan-2-ol of all sediments were determined, averaging 82.0 ± 6.2%, 80.0 ± 7.9% and 81.0 ± 6.7% (±1 s.d.; n = 10), respectively.

Figure 2.

Total ion current gas chromatography (GC) chromatograms of the (a) hydrocarbon fraction of the Indigirka, (b) acid fraction of the Yenisey, and (c) polar fraction of the Indigirka River estuary surface sediment. The n-alkanes are represented by solid circles; internal standard (IS); hopanes are represented by plus symbols; hopenes are represented by asterisks; n-alkanoic acids are represented by stars; n-alkanols are represented by open inverted triangles; triterpenoids are represented by solid inverted triangles. Letters in Figure 2c indicate cholesta-5,22E-dien-3β-ol (a); cholest-5-en-3β-ol (b); 5α-cholestan-3β-ol (c); 24-methylcholest-5,22E-dien-3β-ol (d); 24-methylcholest-5-en-3β-ol (e); 24-methyl-5α-cholestan-3β-ol (f); 24-ethylcholest-5,22E-dien-3β-ol (g); 24-ethylcholest-5-en-3β-ol (h); and 24-ethyl-5α-cholestan-3β-ol (i). Numbers indicate carbon chain length, and ββ indicates stereochemistry of hopanes at C-17 and C-21 positions. (d) High-performance liquid chromatography mass spectrometry (HPLC/MS) base peak chromatogram of tetraether lipids in the surface sediment of the Kolyma River estuary.

3. Results

3.1. Bulk Analyses

[13] The total organic carbon (TOC) contents of the estuarine samples range from 4.8 to 19.4 mg/g. The TOC/TN ratio varies from 10.0 to 15.9 and the δ13CTOC range from −25.0 to −27.4‰ (Table 1). The Δ14C values vary from −608 to −314‰ (Table 1).

3.2. Extractable Organic Matter

3.2.1. Hydrocarbon Fraction

[14] Hydrocarbon fractions are all dominated by a homologous series of C16–C34n-alkanes with a predominance of the odd-numbered homologues and the C27n-alkane being the most abundant homologue in all sediments (Figure 2a and Table 2). The average chain length (ACL; defined in Table 3) is comparable for all samples, with values between 28.0 and 28.6 (Table 3). The n-alkane carbon preference index (CPI; defined in Table 3) varies, with values ranging from 3.2 to 7.3 (Table 3). In addition, the ratio of C23n-alkane to the sum of C23 and C29n-alkanes ranged from 0.35 to 0.47. Further, the ratio of terrigenous-to-aquatic n-alkanes (TARn-alkanes; as defined in Table 3) ranged from 17 to 80 (Table 3). Hopanes and hopenes are also present in all samples (Figure 2a), with 17β(H),21β(H)-homohopane (C31ββ hopane) and C30 hop-17(21)-ene being predominant, but in substantially lower amounts than the n-alkanes. However, with a signal to noise ratio >4 we were still able to detect them clearly. We estimate that the total amounts of hopanes and hopenes in the Lena, Indigirka and Kolyma samples are below 0.02, 0.09 and 0.06 mg/g OC, respectively.

Table 2. Concentrations of Long-Chained n-Alkanes, n-Alkanoic Acids, and n-Alkanols in Surface Sediments of Russian Arctic River Estuaries
 ObYeniseyLenaIndigirkaKolyma
  • a

    Units are μg/g Sed.

  • b

    High molecular weight (HMW), sum of the concentrations of C20–C34n-alkanes.

  • c

    Sum of the concentrations of C20–C30n-alkanoic acids.

  • d

    Sum of the concentrations of C20–C32n-alkanols.

  • e

    Units are mg/g OC.

Concentrationsa
HMW n-alkanesb17182.62520
Σ(C27 + C29 + C31) n-alkanes7.09.41.41311
HMW n-alkanoic acidsc2.44.21.58.017
Σ(C24 + C26 + C28) n-alkanoic acids1.42.70.84.59.0
HMW n-alkanolsd20433.13235
Σ(C26 + C28 + C30) n-alkanol13292.01921
 
Concentrationse
HMW n-alkanesb1.80.910.551.71.2
Σ(C27 + C29 + C31) n-alkanes0.760.480.280.910.65
HMW n-alkanoic acidsc0.260.220.310.550.96
Σ(C24 + C26 + C28) n-alkanoic acids0.150.140.160.310.52
HMW n-alkanolsd2.22.20.642.22.0
Σ(C26 + C28 + C30) n-alkanol1.41.50.431.31.2
Table 3. Source-Diagnostic Distributions of n-Alkanes, n-Alkanoic Acids and n-Alkanols in Surface Sediments of Russian Arctic River Estuaries
 ObYeniseyLenaIndigirkaKolyma
  • a

    CPIi−n, carbon preference index, = equation imageΣ(Xi + Xi+2 + … + Xn)/Σ(Xi−1 + Xi+1 + … + Xn−1) + equation imageΣ(Xi + Xi+2 + … + Xn)/Σ(Xi+1 + Xi+3 + … + Xn+1), where X is concentration.

  • b

    ACLi−n, average chain length, = Σ(iXi + … + nXn)/ΣXi + … + Xn), where X is concentration.

  • c

    Ratio of the C23n-alkane to the sum of the C23 and C29n-alkanes.

  • d

    TARn-alkane, ratio of terrigenous-to-aquatic n-alkanes, = sum of the C27, C29 and C31n-alkanes to the sum of the C17 and C19n-alkanes.

  • e

    TARn-alkanol, ratio of terrigenous-to-aquatic n-alkanols, = sum of the C26 and C28n-alkanols to the sum of the C16 and C18n-alkanols.

n-Alkanes
CPI25–33a3.26.55.55.57.3
ACL25–33b28.028.528.628.528.2
C23/(C23 + C29)c0.470.380.370.330.35
TARn-alkaned8049175046
 
n-Alkanoic Acids
CPI20–30a4.24.54.14.44.2
ACL20–30b25.326.024.825.224.3
 
n-Alkanols
CPI20–32a5.66.613.98.28.9
ACL20–32b26.626.626.625.926.2
TARn-alkanole7914548106115

3.2.2. Acid Fraction

[15] All acid fractions are dominated by a homologous series of C14–C30n-alkanoic acids, with the C24 member being the most abundant, showing an even-over-odd carbon number predominance (Figure 2b and Table 2). The n-alkanoic acid CPI (defined in Table 3) is similar for all samples, with values between 4.1 and 4.5 (Table 3). In contrast, the n-alkanoic acid ACL (defined in Table 3) varies from 24.3 to 26.0 (Table 3).

3.2.3. Polar Fraction

[16] All polar fractions contain a homologous series of C16–C32n-alkanols, dominated by the C28 homologue (in the Kolyma estuary sediment the C26 homologue) and a predominance of the even numbered homologues (Figure 2c and Table 2). The n-alkanol ACL, CPI and TARn-alkanol ratios (all defined in Table 3) span ranges of 25.9–26.6, 5.6–13.9 and 48–145, respectively (Table 3). Substantial amounts of steroids, in particular β-sitosterol (24-ethylcholest-5-en-3β-ol; 3 to 15 times more abundant than the other sterols), are also present in all samples (Figure 2c and Table 4). β-sitosterol is in fact the most abundant compound present in the polar fraction of the Lena, Indigirka and Kolyma estuary sediments (Figure 2c and Table 4). Other steroids that are also present in substantial amounts include cholest-5-en-3β-ol, 24-methylcholest-5-en-3β-ol (campersterol), 24-ethylcholest-5,22E-dien-3β-ol (stigmasterol), 24-ethyl-5α-cholestan-3β-ol (Figure 2c and Table 2). In contrast, the marine-derived 4,23,24-trimethyl-5a-cholest-22-en-3b-ol (dinosterol) [Boon et al., 1979; de Leeuw et al., 1983] could not be detected in any of the sediments. In contrast to the sterols, triterpenoids, represented by friedelan-3-one, α-amyrin (urs-12-en-3β-ol) and β-amyrin (olean-12-en-3β-ol) are only present in minor amounts (Figure 2c).

Table 4. Concentrations of Sterols in Surface Sediments of Russian Arctic River Estuaries
 ObYeniseyLenaIndigirkaKolyma
  • a

    Units are μg/g Sed.

  • b

    Cholesterol.

  • c

    Brassicasterol.

  • d

    Campesterol.

  • e

    Stigmasterol.

  • f

    β-sitosterol.

  • g

    Units are mg/g OC.

Concentrationsa
Cholesta-5-en-3β-olb1.00.970.100.700.91
24-Methylcholest-5,22E-dien-3β-olc0.691.050.070.540.59
24-Methylcholest-5-en-3β-old0.841.20.121.32.05
24-Ethylcholest-5,22E-dien-3β-ole1.22.10.351.32.2
24-Ethylcholest-5-en-3β-olf3.57.01.92539
 
Concentrationsg
Cholesta-5-en-3β-olb0.110.050.020.040.06
24-Methylcholest-5,22E-dien-3β-olc0.080.050.010.030.04
24-Methylcholest-5-en-3β-old0.090.060.030.080.14
24-Ethylcholest-5,22E-dien-3β-ole0.130.110.070.080.15
24-ethylcholest-5-en-3β-olf0.380.360.401.42.7

3.2.4. GDGTs

[17] The east Siberia River estuary sediments contain substantial amounts of GDGTs, including both nonisoprenoidal branched GDGTs and isoprenoidal GDGTs (Figure 2d). All sediments contain much higher relative abundances of branched GDGTs compared to the isoprenoidal GDGTs (Table 5). This difference is clearly reflected in the BIT index [Hopmans et al., 2004] (defined in Table 5), showing values between 0.88 and 0.92 (Table 5).

Table 5. Concentrations of GDGTs and BIT Indices in Surface Sediments of East Siberian Arctic River Estuaries
 LenaIndigirkaKolyma
  • a

    Units are μg/g Sed.

  • b

    Sum of glycerol dialkyl glycerol tetraether (GDGT) I and IV (Figure 2d).

  • c

    Sum of GDGT V, VI, and VII (Figure 2d).

  • d

    Units are mg/g OC.

  • e

    Branched and isoprenoid tetraether (BIT) index = index. calculated following the equation of Hopmans et al. [2004]; (BIT index = Σ(V + VI + VII)/Σ(V + VI + VII + IV)] (Figure 2d)).

Concentrationsa
Isprenoidal GDGTsb0.42.82.8
Branched GDGTsc103749
 
Concentrationsd
Isprenoidal GDGTsb0.090.190.16
Branched GDGTsc2.12.52.8
BIT indexe0.920.880.91

4. Discussion

4.1. Terrestrial Versus Marine Origin of Organic Matter in GRAR Estuary Sediments

[18] Bulk geochemical analysis of all GRAR estuarine sediments indicates terrestrially dominated OM sources. The TOC/TN ratios (10 to 15.9; Table 1) fall between those of C/N of marine OM (around 6) and terrestrially derived OM (>15) [Bordovskiy, 1965; Hedges et al., 1986]. However, TN is here used in this ratio and not total organic nitrogen (TON), which, as seen in other Arctic sediments, could cause a shift in the C/N ratio to lower values due to inclusion of inorganic N such as NH4+ in clay interstitial layers [Schubert and Calvert, 2001; Fahl et al., 2003; Drenzek et al., 2007]. The depleted δ13CTOC values of −25.0 to −27.4‰ (Table 1) are typical for land plants using the C3 synthesis pathway [Meyers, 1997] and comparable to earlier observations in these regions [e.g., Guo et al., 2004; Semiletov et al., 2005], supporting the predominantly terrestrial origin of these sediments. Besides material of terrestrial biosynthetic origin, recent research showed that 1–10% of the OC in the GRAR estuaries is present as combustion-derived black carbon [Guo et al., 2004; M. Elmquist et al., Source apportionment of black carbon in surface sediments from the seven largest Pan-Arctic rivers using natural abundance radiocarbon and molecular markers, submitted to Global Biogeochem. Cycles, 2007; hereinafter referred to as Elmquist et al., submitted manuscript, 2007].

[19] The abundant presence of biomarkers such as C16–C34n-alkanes (with an odd-over-even predominance), C14–C30n-alkanoic acids and C16–C32n-alkanols (both with an even-over-odd predominance; Table 2), sterols such as β-sitosterol (Table 4 and Figure 3) and smaller amounts of higher plant triterpenoids [Bray and Evans, 1961; Eglinton and Hamilton, 1963, 1967; Pant and Rastogi, 1979], further supports a predominantly terrestrial origin. The n-alkane and n-alkanoic acid concentrations are comparable to those observed in the estuary of the Mackenzie River (e.g., station 5 in the work of Drenzek et al. [2007]), with estimations of 0.50 mg/g OC and 0.45 mg/g OC, respectively (note that the Mackenzie River concentrations are measured after saponification). The CPIs of the n-alkanes range from 3.2 to 7.3 (Table 3), with the exception of the Ob sample, comparable to those typically observed for extant plants (>5) [Rieley et al., 1991; Hedges and Prahl, 1993] and those reported earlier for Russian Arctic sediments [Fahl and Stein, 1997; Fernandes and Sicre, 2000; Fahl et al., 2003]. Significantly lower n-alkane CPIs are reported for the Mackenzie River estuary sediments (1.8 to 2.7) [Yunker et al., 1993, 1995, 2002; Drenzek et al., 2007], most likely reflecting a large petrogenic input to this North American river. The GRAR CPIs of the n-alkanoic acids (4.1 to 4.5) and the n-alkanols (5.5 to 13.3; Table 3) are similarly in agreement with a predominantly higher plant origin.

Figure 3.

Abundances of n-alkanes, n-alkanoic acids, n-alkanols, and β-sitosterol along the 140° longitude Eurasian-Arctic climosequence. Ob (O); Yenisey (Y); Lena (L); Indigirka (I); and Kolyma (K).

[20] The slightly lower n-alkane CPI value for this Ob sample (CPI = 3.2) is comparable to n-alkane CPIs of 0.8–3.5 for other Ob estuary sediments [Fahl et al., 2003]. The absence of both a detectable unresolved complex mixture (UCM), typical of heavily biodegraded OM [Gough and Rowland, 1990, 1991; Gough et al., 1992; van Dongen et al., 2003] and abiogenic but thermally more stable 17α(H),21β-hopanes [Mackenzie, 1984; Brassell, 1985] makes an input of OM with a thermally mature origin, such as natural oils seeps, weathering of petroleum source rocks or even petroleum contamination, less likely. Furthermore, recent PAH analyses of this sediment sample indicates a methylphenanthrenes to phenanthrene ratio that also supports a very low petrogenic input in the Ob estuary sample (Elmquist et al., submitted manuscript, 2007). In addition, values as low as the Ob CPI are commonly observed in sediments and this value is still consistent with a predominantly higher plant origin [Brassell et al., 1978]. The large variability in the CPI value for the terrestrial end-member prevents CPI-based quantitative assessment of the fractional contribution of land sources to the total OM.

[21] The concentrations of the low molecular weight (LMW) n-alkanes (C17 and C19) and n-alkanols (C16 and C18), originating from aquatic phytoplankton, are much lower than their high molecular weight (HMW) terrestrial counterparts in all five GRAR estuary samples. Thus the associated TARn-alkanes and TARn-alkanol, spanning 17–80 and 48–123 (Table 3), respectively, supports an overwhelmingly terrigenous source of the OM.

[22] A predominantly nonterrestrial source for the hydrocarbons observed in the sediments of the Lena estuary has been suggested [Zegouagh et al., 1996, 1998]. That proposition was based on an absence of HMW n-alkanoic acids and hydroxy n-alkanoic acids. In contrast to Zegouagh et al. [1996] substantial amounts of HMW n-alkanoic acids were observed in all estuary sediments of the present study, including the Lena estuary (Table 2 and Figures 2b and 4) . Comparable to Zegouagh et al. [1998], no hydroxy n-alkanoic acids have been observed, but this does not necessarily indicate an absence of terrestrially derived material. For instance, North Sea sediments, although dominated by terrestrially derived OM, were also characterized by the absence of detectable amounts of hydroxyl n-alkanoic acids [van Dongen et al., 2000]. The overwhelming presence of many other terrestrial biomarkers and complete absence or only minor amounts of marine biomarkers, as also shown in several other, both Russian and Canadian/American, Arctic sediment studies [e.g., Yunker et al., 1995, 2002; Peulve et al., 1996; Fahl et al., 2003; Stein and Macdonald, 2004], suggest a predominantly terrestrial source for the hydrocarbons, including those observed in the Lena River estuary.

Figure 4.

Ratios of the (a) C23n-alkane to the sum of the C23 and C29n-alkanes, (b) high molecular weight (HMW) n-alkanoic acids to the HMW n-alkanes, (c) β-sitosterol to HMW n-alkanes, and (d) the n-alkanoic acid carbon preference index (CPI) to n-alkane CPI along the 140° longitude Eurasian-Arctic climosequence. Ob (O); Yenisey (Y); Lena (L); Indigirka (I); Kolyma (K).

[23] Biomarker analyses also indicate, comparable to earlier Arctic sediment studies [Yunker et al., 1995; Fahl et al., 2003] the presence of substantial amounts of specific sterols synthesized by higher plants, such as campesterol and foremost β-sitosterol (Table 4). However, these sterols are not exclusively synthesized by higher plants and additional input from phytoplankton and bacteria, although a minor process at best [Volkman, 2005], is possible [Gagosian et al., 1983; Volkman, 1986; Volkman et al., 1998]. Brassicasterol, generally synthesized by both marine and freshwater phytoplankton, is also abundantly present (Figure 2c). Fahl et al. [2003] observed a decreasing trend in brassicasterol concentrations in surface sediments from Ob and Yenisey estuaries northward to the southern Kara Sea, making the freshwater origin the most likely source in these systems. In addition, no other specific marine steroids, including dinosterol, were observed, making the higher plant/freshwater phytoplankton origin the most likely source for the majority of the steroids observed in these sediments. The absence of dinosterol in surface sediments of Arctic River estuaries has been noted before and could, for instance, not be positively identified in Mackenzie River sediment samples [Yunker et al., 1995].

[24] Branched glycerol dialkyl glycerol tetratetraethers (GDGTs; V, VI, VII in Figure 2d) are ostensibly synthesized by anaerobic soil bacteria and thus most likely of pure terrestrial origin [Weijers et al., 2006b], whereas the isoprenoid GDGTs in the Arctic estuarine sediments likely have a mixed terrestrial and marine origin [Powers et al., 2004; Weijers et al., 2004, 2006a, 2006b; Herfort et al., 2006b; Kim et al., 2006]. Methanogenic archaea in peats produce predominantly GDGT I and to a lesser extent GDGTs II and III [e.g., Kates et al., 1993; Pancost et al., 2000] and marine Crenarchaeota predominantly synthesize GDGT I and crenarchaeol (GDGT IV in Figure 2d) [Sinninghe Damsté et al., 2002]. Although crenarchaeol has also been detected in peats and soils [Weijers et al., 2004, 2006a], its amounts are considerably lower than the branched GDGTs resulting in BIT values of soils typically >0.90 [Weijers et al., 2006a]. The BIT index values of 0.88 to 0.92 obtained for the eastern GRARs are typical for sediments receiving high amounts of fluvial terrestrial organic matter [Hopmans et al., 2004; Herfort et al., 2006a; van Dongen et al., 2006; Weijers et al., 2006a], and again in support of a predominantly terrestrial origin of OM in these sediments.

[25] To summarize, both bulk geochemical and a broad suite of molecular biomarker analyses of this first comprehensive study of the five GRAR estuaries indicate that the OM across the continent-scale climosequence of the coastal Siberian Arctic are overwhelmingly dominated by terrestrial sources. The majority of this land-derived OM is transported to the estuaries by fluvial processes, by riverine run-off and coastal erosion, as aeolian transport provides only a very minor input [e.g., Stein and Macdonald, 2004].

4.2. Distinguishing Among Different Terrestrial Sources of the Exported Organic Matter

[26] The terrestrial OM transported to all GRAR estuarine sediments, as illustrated above, are originating from diverse sources. For instance, n-alkanes, n-alkanoic acids, n-alkanols, β-sitosterol, campesterol and triterpenoids are all indicative of a higher plant origin, brassicasterol for freshwater phytoplankton origin, while branched GDGTs traces an anaerobic soil bacteria origin.

[27] The C23 to (C23 and C29) n-alkane with values around 0.4 (Table 3) further suggests an influx of OM rich in midchain (besides C23 also predominantly C25) n-alkanes, following Nichols et al. [2006] probably Sphagnum-derived OM, to the estuarine sediments. Indeed in the Russian Arctic, large areas are dominated by tundra, bogs and marshes [Bordovskiy, 1965; FAO, 2001; Kremenetski et al., 2003; Kimble, 2004; Bartholomé and Belward, 2005; Groisman and Bartalev, 2006]. All these vegetation types are rich in Sphagnum, indicating that Sphagnum-rich peats are likely to be a major source of OM to the GRARs. Unfortunately, the n-alkane distributions in Sphagnum species have not been specifically determined for Russian Arctic ecosystems. We are therefore using published analyses of Sphagnum species in non-Arctic peat bogs showing large amounts of C23 and C25n-alkanes [Corrigan et al., 1973; Baas et al., 2000; Nott et al., 2000; Pancost et al., 2002; Nichols et al., 2006] as an available reference point.

[28] The C23 to (C23 and C29) n-alkane ratio shows a clear trend from west (0.47) to east (0.33; Table 3 and Figure 4a). The west Siberian lowland, covering the drainage areas of both the Ob and Yenisey, is indeed predominantly covered with Sphagnum-rich peatland (>900,000 km2) [Kremenetski et al., 2003]. In contrast, and also in agreement with the observed trend, the extent of peatland and other Sphagnum rich land cover is much lower in the eastern areas (Figure 1b). The basins of the Kolyma and Indigirka rivers are characterized by relatively arid conditions, with summer precipitation below 100 to 150 mm, a droughty period that lasts for 30 to 100 d and a precipitation-to-evaporation ratio between 0.3 and 0.5 [Naumov, 2004]. In contrast, the summer precipitation in the basin of the more western Siberian rivers (Ob, Yenisey and Lena) is much higher (up to 400 mm) with a precipitation-to-evaporation ratio above 1.3 to 1.5, indicating that the habitats in these areas are relative more wet [FAO, 2001; Karavaeva, 2004; Naumov, 2004]. These arid conditions can also influence the observed trend. Besides an increased Aeolian plant wax transport and potential for preferential transport of n-alkanes at the lower end of the vascular plant range these arid conditions may possibly also cause a shift in the n-alkane distribution within the Sphagnum species. Recent research indicates that Sphagnum species growing under drier conditions, in contrast to species growing in relatively wet habitats, such as S. magellanicum and S. capillifolium [Andrus, 1980; Hájková and Hájek, 2004a, 2004b], are dominated by the higher n-alkanes (C27 to C33n-alkanes) [Nott et al., 2000]. This suggests that a shift in the n-alkanes distribution may reflect not only a variation in relative amounts of Sphagnum-derived OM transported to the estuaries but also a shift in drainage basin aridity.

[29] Differences in vegetation and/or environmental conditions could also be the cause of other molecular-compositional shifts observed along the Eurasian-Arctic climosequence. For instance, the ratios of both HMW n-alkanoic acids and β-sitosterol to HMW n-alkanes show clearly increasing trends from west to east of 0.14–0.83 and 0.21–2.3, respectively (Figures 4b and 4c). These trends may signal increased preservation due a lesser extent of microbial remineralization, as will be discussed below. However, the relative amounts and proportions of all biomarkers are dependent on the plant species [e.g., Gunstone et al., 1994; Christie, 2003, and references therein] indicating that shifts in biomarker ratios may also indicate shifts in vegetation cover. As espoused above, the detailed molecular composition may also vary for similar plant species growing at different sites or under different environmental conditions [Maffei, 1996; Sachse et al., 2006]. This is mainly due to the different functions these compounds have within the plants. For instance, sterols act as rigidifiers in cell membranes while HMW n-alkanes and n-alkanoic acids act as a protective coating in leaf waxes. This indicates that, although the observed ratios are largely influenced by degradation processes, a response to different environmental conditions, such as the more arid conditions in the basins of the Kolyma and Indigirka rivers, cannot be excluded.

4.3. Trends in Degradation of Exported Terrestrial Organic Matter

[30] Decay of OM occurs initially in the water column and then in the microbially active upper zones of the sediment, and involves the loss of diagnostic characteristics (functional groups) of the original biological compounds [Mackenzie et al., 1982; Maxwell and Wardroper, 1982; Brassell et al., 1984; de Leeuw et al., 1989]. Thus the presence of substantial amounts of compounds such as n-alkanoic acids, n-alkanols and sterols relative to n-alkanes (Table 2 and 4) suggest a low proportion of OM decay. Furthermore, the n-alkanol distributions are characterized by a maxima at n-C28 (or n-C26), comparable to particulate matter samples of the Mackenzie River [Yunker et al., 1995]. This indicates the presence of “free” n-alkanols rather than more refractory “bound” n-alkanols [Cranwell, 1981; Yunker et al., 1995], suggesting the presence of fresh material in the river sediments. Bacterial biomarkers, such as hopanes and hopenes, which are typically produced by bacteria during the early decay of OM, are also only present in relatively small amounts in the GRAR sediments. The relatively small contribution of these bacterial biomarkers thus also suggests a low degree of OM decay. In addition, recent research indicates substantial amounts of bacteriohopanepolyols (BHP), the precursors of these hopenes/hopanes, in the estuary sediments of the east Siberian rivers (B. E. van Dongen et al., Bacteriohopanepolyols in surface sediments of the east Siberian River estuaries, manuscript in preparation, 2007), further supporting the low degree of OM decay. It should be noted that although degradation of OM seems to be at a low degree, recent research indicates that degradation of OM transported by both coastal erosion and riverine run-off causes strong carbon dioxide and methane out gassing in the Arctic Siberian coastal zones [Semiletov, 1999; Shakhova et al., 2005; Semiletov et al., 2007; Shakhova and Semiletov, 2007].

[31] Relatively larger amounts of HMW n-alkanoic acids or sterols, in particular β-sitosterol, compared to HMW n-alkanes (Figures 4b and 4c), are observed in the estuary sediments of the eastern Siberian rivers compared to the western estuaries, indicating a lower degree of decay toward the east. Following Fahl et al. [2003], this is further illustrated by the summed concentrations of campesterol and β-sitosterol as a function of the HMW n-alkanes (Figure 5). Similar to Fahl et al. [2003], a linear correlation for the western GRAR estuaries is observed. However, both samples from the Kolyma and Indigirka estuaries fall outside this correlation, with both showing an excess of β-sitosterol and campesterol, indicating a lower degree of OM decay in the eastern GRARs. The ratio of n-alkanoic acid CPI to n-alkane CPI also show a trend from west (>1) to east (<0.6; Figure 4d) with the n-alkanoic acids CPIs staying almost constant (between 4.1 and 4.5) and the n-alkane CPIs increasing from west (3.2) to east (7.3; Table 3). Although this trend eastward could be attributed to mixing of various plant inputs, the lower degradation of functionalized lipids such as HMW n-alkanoic acids to n-alkanes to the east can also cause higher n-alkane CPIs.

Figure 5.

Correlation plot of the summed concentrations of campesterol and β-sitosterol versus the summed concentrations of the three most predominant high molecular weight n-alkanes.

[32] A previous study using pyrolysis-GC/MS of similar surface sediments from across this entire climosequence showed a shift in the pyrolysate contribution of furfurals, a representative of polysaccharides, from the west (∼32 % of most dominant pyrolysate products) to the east (∼42%) [Guo et al., 2004]. This indicates the presence of more labile material in surface sediments of the eastern river estuaries. Furthermore recent analyses indicate a strong eastward increase in dissolved methane concentration, indicating an increase of the OC liability [Shakhova et al., 2008]. Hence both studies support the lesser degree of decomposition of the OM toward the east. In addition, earlier observations indicate that the watersheds of the eastern rivers are characterized by more permafrost and a shorter vegetative season, which apparently tends to preserve this OM in a less degraded state.

[33] Radiocarbon analyses indicate that the Δ14C values are generally lower in the east, −503 and −527‰ for the Kolyma and Indigirka river estuaries, compared to −175 and −314‰ for the western Ob and Yenisey estuaries (Table 1). These Δ14C values correspond to apparent 14C ages ranging from 1490–2970 years BP in the west to 5560–5960 years BP in the eastern part. The larger abundance of permafrost in the east, which delays the transport time, is probably the cause for these differences. Hence despite that fluvially released OM in the eastern GRARs is older, they have a molecular biomarker composition that, taken together, strongly indicates fresher and less degraded material. Compound specific radiocarbon analyses may shed additional light on this apparent decoupling between age and degradation [e.g., Drenzek et al., 2007].

5. Conclusions and Potential Implications

[34] This study has demonstrated the utility of estuarine sediments of major rivers as integrating archives of terrestrial organic matter released and exported from vast and heterogeneous drainage basins. Both specific molecular biomarkers (n-alkanes, n-alkanoic acids, n-alkanols, triterpenoids, branched GDGTs and steroids) and geochemical bulk parameters (δ13CTOC and TOC/TN ratios) indicate a predominance of terrestrially derived OM in estuary sediments of all five great Russian Arctic rivers with only minor contributions from marine productivity.

[35] The detailed n-alkane distributions indicate a possible contribution of Sphagnum-derived OM, particularly in the west Siberian rivers. A suite of molecular-level proxies suggest a trend along the west-east climosequence of decreasing degree of decomposition of the river-exported organic matter toward east Siberia. Eastward increasing relative abundance of terrestrial compounds with functional groups such as sterols and n-alkanoic acids as well as very low abundance of bacterial biomarkers such as hopanes and hopenes combine to suggest that the OM, despite older reservoir age, is structurally “fresh”. Although differences in changing vegetations or environmental conditions also could have played a role, these differences in reservoir age and degradation state of the material are most likely both caused by the greater extent of permafrost in the eastern area.

[36] The current study is providing a coherent view of the large-scale variations in organic matter dynamics on the entire Siberian Arctic shelf. The GRARs integrates vegetation and climate differences between their large drainage basins and sediments in a given estuary are more homogeneous in composition than the composition of the drained soil. Many indices combined in this study to suggest a complete dominance of terrestrial sources, and thus very little complication of marine in-mixing, in the organic matter found in the estuarine samples. There is geochemical consistency in the geochemical parameters illustrated in Figures 4 and 5 that are consistent with the known variations in climate and vegetation between the drainage basins. This benchmarking study of the organic geochemical composition of the five GRAR estuaries suggests that the organic matter in the west is consistent with a significant contribution from Sphagnum. The composition also suggests that the exported OM becomes less degraded toward the east. Several of the conclusions of this study are consistent with those drawn by Guo et al. [2004] in the same area, albeit that earlier study used less sophisticated and informative data.

[37] The predicted amplified warming in the Eurasian Arctic continental region, particularly in east Siberia, could have huge consequences such as reduction in the volume of the permafrost and an increase in summertime active-layer soil thickness with consequences for the release and degree of decomposition of the OM. This suggests that the ongoing environmental changes in the northern ecosystems could have major effects on the composition of the OM delivered to the Arctic estuaries, particularly in the eastern Siberian area, plausibly shifting the molecular composition to more “west-Siberian-like”. The corollary is that a warming of the eastern Siberia area would be predicted to yield greater remineralization and thus release of, on average 6000-year-old, fluvial and eroded organic matter as carbon dioxide and methane.

Acknowledgments

[38] This project was supported financially by the Swedish Research Council (VR contract 621-2004-4039), a senior research fellowship to Ö.G. (VR contract 629-2002-2309), the Swedish Foundation for Strategic Environmental Research (Mistra contract 2002-057), the Far-Eastern Branch of Russian Academy of Sciences (Project Environmental Changes in the East-Siberian Region) and the International Arctic Research Center of the University Alaska Fairbanks (by the Cooperative Institute for Arctic Research through NOAA Cooperative Agreement NA17RJ1224), and the Russian Foundation for Basic Research (04-05-64819). The authors thank Dr. O. Dudarev and A. Charkin for assistance during sampling and Z. Kukulska, M. Elmquist, Dr. Z. Zencak, Dr. Y. Zebühr, Dr. E. Hopmans, and H. Gustavsson for technical assistance.

Ancillary