Low δ13C in tests of live epibenthic and endobenthic foraminifera at a site of active methane seepage



[1] To investigate the use of benthic foraminifera as a means to document ancient methane release, we determined the stable isotopic composition of tests of live (Rose Bengal stained) and dead specimens of epibenthic Fontbotia wuellerstorfi, preferentially used in paleoceanographic reconstructions, and of endobenthic high-latitude Cassidulina neoteretis and Cassidulina reniforme from a cold methane-venting seep off northern Norway. We collected foraminiferal tests from three push cores and nine multiple cores obtained with a remotely operated vehicle and a video-guided multiple corer, respectively. All sampled sites except one control site are situated at the Håkon Mosby mud volcano (HMMV) on the Barents Sea continental slope in 1250 m water depth. At the HMMV in areas densely populated by pogonophoran tube worms, δ13C values of cytoplasm-containing epibenthic F. wuellerstorfi are by up to 4.4‰ lower than at control site, thus representing the lowest values hitherto reported for this species. Live C. neoteretis and C. reniforme reach δ13C values of −7.5 and −5.5‰ Vienna Pee Dee Belemnite (VPDB), respectively, whereas δ13C values of their empty tests are higher by 4‰ and 3‰. However, δ13C values of empty tests are never lower than those of stained specimens, although they are still lower than empty tests from the control site. This indicates that authigenic calcite precipitates at or below the sediment surface are not significantly influencing the stable isotopic composition of foraminiferal shells. The comparatively high δ13C results rather from upward convection of pore water and fluid mud during active methane venting phases at these sites. These processes mingle tests just recently calcified with older ones secreted at intermittent times of less or no methane discharge. Since cytoplasm-containing specimens of suspension feeder F. wuellerstorfi are almost exclusively found attached to pogonophores, which protrude up to 3 cm above the sediment, and δ13C values of bottom-water-dissolved inorganic carbon (DIC) are not significantly depleted, we conclude that low test δ13C values of F. wuellerstorfi are the result of incorporation of heavily 13C-depleted methanotrophic biomass that these specimens feed on rather than because of low bottom water δ13CDIC. Alternatively, the pogonophores, which are rooted at depth in the upper sediment column, may serve as a conduit for depleted δ13CDIC that ultimately influences the calcification process of F. wuellerstorfi attached to the pogonophoran tube well above the sediment/water interface. The lowest δ13C of live specimens of the endobenthic C. neoteretis and C. reniforme are within the range of pore water δ13CDIC values, which exceed those that could be due to organic matter decomposition, and thus, in fact, document active methane release in the sediment.

1. Introduction

[2] Massive releases of 13C-depleted methane from marine sediments along continental margins have been invoked to explain large carbon isotopic shifts and rapid climate changes in the geologic record. In present-day environments mud volcanoes and cold seeps are considered to be a significant source of fossil methane in the atmosphere and the ocean [e.g., Dimitrov, 2002; Milkov et al., 2003, 2004]. It is a matter of debate whether the current flux of methane from mud volcanoes and cold seeps has changed over time, and if so, how past changes in seep activity are documented in the geological record. It is suggested that characteristic species or a specific faunal composition of benthic foraminifera may indicate methane release at the seafloor [e.g., Bernhard, 2003], or that the δ13C of benthic foraminiferal tests are characteristically depleted, since they calcified in pore water or near-seafloor water, with a locally depleted δ13CDIC because of oxidation of methane [e.g., Hill et al., 2003; Rathburn et al., 2003].

[3] Here we present benthic foraminiferal isotope data from a site of active methane venting at the northern Norwegian continental margin. The overall aim of this study is to further investigate whether or not in a Recent methane-rich environment, the benthic foraminiferal δ13C is significantly influenced by near-seafloor release of methane. This is important for the interpretation of low fossil benthic δ13C as indicator of past methane release in the geologic record.

1.1. Recent Release of Methane and Benthic Foraminiferal δ13C

[4] Benthic foraminifera and their stable isotopic composition have been studied at several sites of active methane discharge. These include the Gulf of Mexico [Sen Gupta et al., 1997] and off northern Norway [Mackensen et al., 2004] in the Atlantic, and off the coasts of California [Rathburn et al., 2000; Bernhard et al., 2001; Hill et al., 2003; Rathburn et al., 2003], Mexico [Herguera et al., 2004], Oregon [Torres et al., 2003; Hill et al., 2004], and Japan [Akimoto et al., 1994] in the Pacific, as well as the Adriatic Sea in the Mediterranean [Panieri, 2003]. However, it is not yet clear whether or not methane venting at the seafloor is recorded by the δ13C of benthic foraminiferal tests.

[5] One set of papers suggests that methane is not recorded by the tests of benthic foraminifera. Stott et al. [2002] measured pore water dissolved inorganic carbon (DIC) concentrations and δ13CDIC values, as well as the δ13C of tests of live benthic foraminifera from uppermost sediments in the Santa Barbara Basin off California. They suggest that today methane-derived CO2 has no discernable influence on pore water δ13CDIC, and that benthic foraminiferal δ13C values, as low as those recorded and interpreted by Kennett et al. [2003] to reflect past methane release, can easily be explained by enhanced photosynthate carbon rain and carbon oxidation in the basin. At sites with anaerobic methane oxidation offshore Oregon, Torres et al. [2003] showed that tests of live (Rose Bengal stained) benthic foraminifera have δ13C not significantly lower than those observed in nonventing sediments in spite of extremely low pore water δ13CDIC, but that dead foraminifera (unstained empty tests) have δ13C significantly lower than those of live ones. Supported by additional geochemical evidence, they concluded that authigenic carbonate precipitation could explain extremely low δ13C values in fossil foraminifera. This recently was corroborated by Herguera et al. [2004], who reported on live specimens from active methane venting sites offshore of Mexico. They showed that tests did not have distinctly lower δ13C than expected from bottom water and pore water δ13CDIC. Moreover, the range of δ13C was small, and the low values can be fully explained in terms of oxidation of organic carbon.

[6] Another set of papers, however, indicates that methane can be recorded by benthic foraminiferal tests. Rathburn et al. [2000], showed that δ13C of fossil foraminifera from cores off northern California were consistently lower than those of stained specimens, and suggested that this reflects stronger seep activity in the past. Hill et al. [2003] examined the isotopic composition of single specimens of unstained foraminifera of late Glacial age from modern methane seeps in Santa Barbara Channel, California, and revealed an extreme carbon isotopic variability and very low δ13C. On the basis of scanning electron microscopy, they doubt that overgrowth has affected the isotopic composition. The δ13C values of live benthic foraminifera from a cold methane seep in Monterey Bay, California were found to be lower than those from nonseep sites [Rathburn et al., 2003], but not as much as indicated by the fossil values reported by Hill et al. [2003]. Rathburn et al. [2003] concluded that it is the high variability in the stable carbon isotopic composition within live conspecific specimens rather than low absolute values that records past methane seepage. Most recently, Hill et al. [2004] showed low δ13C of live (Rose Bengal stained) foraminifera from Hydrate Ridge off Oregon, in contrast to foraminifera examined by Torres et al. [2003] from the same seep area. Obviously, additional studies from different environments are needed to (1) document the capability of live benthic foraminifera to record the presence of methane enriched pore water and bottom water in their test δ13C and (2) better understand how methane influences the δ13C of empty tests in the sediment.

[7] In this study from the Håkon Mosby mud volcano (HMMV) we aim to determine whether live benthic foraminifera record the presence of methane via significantly lower δ13C values of their tests. We further ask whether there are differences between species occupying different microhabitats, and whether negative δ13C excursions in benthic foraminiferal tests are recorded during primary calcification or as the result of authigenic calcite precipitation and diagenetic overgrowth. To tackle these questions, we determined the stable isotopic composition of the tests of live and dead specimens of three species from the HMMV and a control site without any influence of methane. Fontbotia wuellerstorfi [Schwager, 1866] is most often used in paleoceanographic reconstructions and regarded as most reliable recorder of bottom water DIC. Therefore this species was chosen as representative of an almost strictly epibenthic lifestyle. Cassidulina neoteretis [Seidenkrantz, 1995] and Cassidulina reniforme [Sejrup and Guilbault, 1980] were investigated as they represent shallow infaunal, low-oxygen tolerating species [Wollenburg and Mackensen, 1998b, 1998a; Husum and Hald, 2002, 2004].

1.2. Håkon Mosby Mud Volcano

[8] The continental margin between northern Norway and Svalbard is characterized by major submarine slides, large-scale mass wasting, and seafloor features such as pockmarks, mud volcanoes and diapirs produced by mobilized gas, fluids and sediments [Vogt et al., 1999]. The HMMV is a site of active methane venting located at about 72°N and 14°E within a slide scar on the glacial submarine Bjørnøya Fan on the SW Barents Sea continental slope. The fan covers the entire continental slope and reaches a thickness of approximately 3.1 km beneath the mud volcano [Hjelstuen et al., 1999]. The HMMV is a circular structure, approximately 1 km in diameter that rises 8 to 10 m above the neighboring seafloor (Figure 1) at roughly 1250 m water depth [Hjelstuen et al., 1999].

Figure 1.

Location of sites at Håkon Mosby mud volcano (HMMV) occupied by R/V Atalante and PR/V Polarstern in 2001 (black asterisks) and 2004 (white asterisks), respectively. White lines give ROV tracks in 2001. Read 12 to 25d as ATL12 to ATL25d, and 002 to 015 as PS66002 to PS66015. Microbathymetric map is based on remotely operated vehicle (ROV)-based echo sounding in 2003 (© IFREMER). Full range of color represents water depths between 1249 and 1289 m.

[9] Three concentric zones can be delineated across the HMMV [Vogt et al., 1999]. A central zone of about 200 m diameter has uniform sediments that have been interpreted as a mud breccia [Shilov et al., 1999]. These sediments are highly reduced containing high hydrogen sulfide concentrations and dissolved methane that reaches 12.5 mM L−1 with a δ13C value of −60.8‰ VPDB at 19–21 cm subbottom depth [Lein et al., 1999]. In this inner central zone, methane with a δ13CCH4 value of −61.6‰ is also released into the water column [Damm and Budeus, 2003]. Adjacent to the inner zone are reduced sediments containing gas hydrates. Large parts of these sediments are covered by widespread, irregularly shaped patches of bacterial mats [Lein et al., 1999]. In an outer hummocky peripheral zone, surface sediments are oxidized down to approximately 10 cm subbottom depth, and densely populated by two species of pogonophoran tube worms [Pimenov et al., 1999; Smirnov, 2000].

[10] The areas with pogonophores at the peripheral zone of the HMMV display high microbial activity and high rates of methane oxidation and sulfate reduction in the upper 20 cm of the sediment [Pimenov et al., 1999]. Certainly, above these areas the bottom water methane concentration is higher than background level. However, the δ13CCH4 signature is not significantly enriched versus source values, as would be expected for sedimentary methane oxidation [Damm and Budeus, 2003]. This suggests that methane in the water column at these sites is not released from the sediments below, but is transported away from the central crater faster than it is oxidized. Significant release of methane directly into the bottom water occurs only in the central zone [Damm and Budeus, 2003].

2. Material and Methods

[11] On R/V L'Atalante's 2001 expedition in the Norwegian and Greenland Sea, sediment samples were taken at the HMMV by multiple corer and push corer (Figure 1), the latter operated during dives with a remotely operated vehicle (ROV). Bottom water samples were obtained by horizontal bottom water sampler and rosette sampler, or after recovery, from push cores and the tubes of the multiple corer (Table 1). Sampling was performed at the three characteristic environments, which had been selected from initial video surveys with the ROV (Figures 1 and 2). In addition, a control site was sampled 12 nautical miles south of the seepage area at 1265 m water depth. This site has similar water depth, overlying water mass, and sea surface productivity.

Figure 2.

Seafloor images (©IFREMER) of the HMMV: (a) active center of the mud volcano, (b) bacterial mats south of the center of the crater (ROV track 19 in Figure 1), and (c) the hummocky outer rim of the crater densely populated by pogonophoran tube worms. Diameter of push corer as handled by the manipulator is 6 cm (ROV track 25 in Figure 1).

Table 1. Stable Isotopic Composition of Water Samples Taken by Rosette Sampler Mounted on a Conductivity-Temperature-Depth Probe (CTD), Horizontal Bottom Water Sampler (HBS), Multiple Corer (MUC), and Push Cores Operated by a Remotely Operated Vehiclea
GearSiteWater Depth, mδ13CDIC, ‰ VPDBδ18O, ‰ VSMOW
  • a

    Water depth is approximate and just gives length of wire.

CTDATL81250 0.27
HBSATL13, N11261 0.28
HBSATL13, N212611.070.28
HBSATL13, N312611.070.28
HBSATL13, N412611.050.29
HBSATL13, N512611.080.23
Push coreATL1912601.000.22
Push coreATL25a12620.57 
Push coreATL25a1262 0.23
Push coreATL25c1262−2.580.26
Push coreATL25c12621.03 
Push coreATL25d1260−0.48−0.51

[12] During onshore investigation of the 2001 material it became obvious that because of low standing stocks of benthic foraminifera in the seepage area, the study would benefit from additional sample material. Consequently, to increase the number of samples and statistical reliability of measurements we resampled bottom water and surface sediments at the HMMV in the summer of 2004 on Polarstern cruise ANT-XX/1 with the aid of a video-guided multiple corer (Figure 1).

[13] At each site (Table 2), immediately after recovery, surface sediment cores were cut in 1-cm thick slices from the surface to a maximum of 30 cm depth, preserved in Rose Bengal stained alcohol, and kept cool until sieving over a 63-μm mesh on shore. For determination of the stable carbon isotope ratio of DIC and the stable oxygen isotope ratio of water, water samples were filled into 50 mL glass vials, sealed with wax under 4°C air temperature, and kept cool until further treatment on shore. The DIC samples additionally were poisoned with a saturated solution of HgCl2.

Table 2. Stable Isotopic Composition of Benthic Foraminiferal Testsa
  • a

    The σ gives standard deviation of n samples. Number of specimens per sample depends on species, i.e., F. wuellerstorfi (1–4), C. neoteretis (10–25), and C. reniforme (40). The δ13C and δ18O values are normalized to bottom water δ13CDIC of 1.11 ‰ VPDB and equilibrium of 180, respectively.

Atl12−    2C. reniforme
Atl12−1.65 0.84      C. neoteretis
Atl12−0.15 0.01      F. wuellerstorfi
Atl18    −5.51 0.69  C. reniforme
Atl18−2.49 0.360.17    2C. reniforme
Atl18    −5.90    C. neoteretis
Atl18−1.950.350.89     2C. neoteretis
Atl18    − wuellerstorfi
Atl19−3.34 0.21      C. reniforme
Atl19−1.960.510.59     2C. neoteretis
Atl22    −4.770.610.680.164C. neoteretis
Atl22−2.820.430.85     2C. neoteretis
Atl22    −1.070.440.170.123F. wuellerstorfi
Atl22−1.89 0.33      F. wuellerstorfi
Atl22    −4.16 0.67  C. reniforme
Atl22−2.21 −0.08      C. reniforme
Atl22    −7.710.360.630.032M. zaandami
ATL25a−2.51 −0.05      C. reniforme
ATL25a−2.15 0.93      C. neoteretis
Atl25d    −7.460.100.860.102C. neoteretis
Atl25d−2.200.510.840.21    2C. neoteretis
Atl25d    −3.291.640.150.062F. wuellerstorfi
Atl25d−2.38        F. wuellerstorfi
Atl25d    −5.44 0.42  C. reniforme
Atl25d−2.490.570.170.19    2C. reniforme
Atl28    −1.36 0.78  C. neoteretis
Atl28−1.15 0.86      C. neoteretis
Atl280.02 0.39      F. wuellerstorfi
Atl28    − wuellerstorfi
Atl28    −1.69 0.24  C. reniforme
Atl28−1.52 0.19      C. reniforme
PS66002-3    − wuellerstorfi
PS66002-3−2.270.650.400.08    6F. wuellerstorfi
PS66002-3    −3.85 0.74  C. neoteretis
PS66003-2−2.64 1.11      C. neoteretis
PS66011-2−2.28 0.91      C. neoteretis
PS66013-2−    2F. wuellerstorfi
PS66013-2−1.88 0.93      C. neoteretis
PS66015-2    −0.770.640.300.085F. wuellerstorfi
PS66015-2−0.600.450.310.03    4F. wuellerstorfi
PS66015-2    −2.84 0.90  C. neoteretis
PS66015-2−2.38 0.94      C. neoteretis

[14] In the laboratory Rose Bengal stained, hyaline, and pristine tests of benthic foraminifera were counted as live specimens and separated from empty tests. About 15 stained specimens of Fontbotia wuellerstorfi and Cassidulina reniforme each, as well as 35 Cassidulina neoteretis were examined with scanning electron microscopy to check for authigenic calcite overgrowth or signs of dissolution. We acknowledge that using Rose Bengal stain to distinguish live specimens and empty tests may overestimate the number of living specimens at the time of sampling. Dead foraminiferal cytoplasm can be stained from weeks to months after an individual's death [e.g., Jorissen et al., 1995; Bernhard et al., 2001]. All pogonophora tubes were examined for attached benthic foraminifera.

[15] The stable isotopic composition of benthic foraminiferal tests was determined with a Finnigan MAT 251 isotope ratio gas mass spectrometer directly coupled to an automated carbonate preparation device (Kiel II) and calibrated via NIST 19 international standard to the Vienna Pee Dee Belemnite (VPDB) scale. We used 1–4, 10–25, and about 40 specimens per single measurement of Fontbotia wuellerstorfi, Cassidulina neoteretis, and Cassidulina reniforme, respectively. Stable isotope values are given in δ notation versus VPDB (Table 2). The precision of the measurements at 1σ based on repeated analyses of an internal laboratory standard (Solnhofen limestone) over a 1-year period was better than ±0.08 and ±0.06‰ for oxygen and carbon isotopes, respectively.

[16] DIC was extracted from seawater with phosphoric acid in an automatic preparation line (Finnegan Gasbench I) coupled online with a Finnigan MAT 252 mass spectrometer to determine its 13C/12C ratio. All samples were run at least in duplicate. Results are reported in δ notation relative to the VPDB scale with an external reproducibility of ±0.1‰ at 2σ. For the oxygen isotope determination of water, 7 mL of water were equilibrated in 13 mL headspace with CO2 gas using an automated Finnigan equilibration device, online connected with a Finnigan MAT Delta-S mass spectrometer. At least two replicates (including preparation and measurement) were run for each oxygen isotope determination. Results are reported in δ notation relative to the VSMOW scale with an external reproducibility of ±0.03‰ at 1σ.

[17] To describe differences between means of δ13C of samples at HMMV and at the control site, and between means of intraspecific δ13C of live and dead specimens we used a two-tailed t test [Davis, 1973]:

display math

with M1 and M2 = means of δ13C, and n1 and n2 = number of measurements at HMMV and at the control site or of stained and empty tests, and σΔ = pooled estimate of the population standard deviation.

3. Results

[18] Bottom water directly overlying the sediment from multiple corer tubes and push cores at Site ATL25a, Site PS66002, and ATL25c revealed δ13CDIC values of 0.57‰, −2.12‰, and −2.58‰, respectively (Table 1 and Figures 1 and 3) . This is significantly less than bottom water δ13CDIC of 1.11 ± 0.04‰, which was determined from bottom water sampled with the horizontal bottom water sampler and the rosette sampler from 0.4 to 5 m above the seafloor respectively.

Figure 3.

Stable isotopic composition of (a) water samples taken from push cores or multiplecorer tubes (squares) and rosette water sampler (circles) and (b) water sampler values only (open circles) with calculated bottom water mean of δ18O and δ13CDIC of 0.29 ± 0.02‰ Vienna Standard Mean Ocean Water and 1.11 ± 0.04‰ Vienna Pee Dee Belemnite (VPDB), respectively.

[19] At four out of 11 sites (ATL18, −22, −25d, and PS66002) within the HMMV and at Site PS66015 just 600 m outside the northern rim of the mud volcano we found Rose Bengal stained benthic foraminifera, which we considered live at the time of sampling, in sufficient numbers for stable isotopic determination (Table 2 and Figures 1 and 47). Although the 11 sites cover different environments at HMMV, live Fontbotia wuellerstorfi, Cassidulina neoteretis and Cassidulina reniforme are only found in samples from areas colonized by pogonophores. In addition to the sites with live foraminifera, empty tests were analyzed from Sites ATL12, −19, −25a, and PS66003, and −013, which were sampled from areas covered by bacterial mats (Table 2 and Figures 1 and 47). Sites ATL25c (from the center of active methane discharge into the water column) and PS66011 (from bacterial mats) did not yield any pristine benthic foraminiferal tests. Examination of stained benthic foraminiferal tests with scanning electron microscopy did not reveal authigenic calcite overgrowth or signs of dissolution.

Figure 4.

Stable isotopic composition of live (solid symbols) and dead (open symbols) F. wuellerstorfi from sites within the HMMV (Sites PS66002, −003, −013, ATL12, −18, −22, −25a, and −25d), Site PS66015 from the outer margin, and reference Site ATL28 (R) not influenced by methane seepage. Error bars give standard deviation of replicates at single sites. Note different scale of axes. The δ13C and δ18O values are normalized to bottom water δ13CDIC of 1.11‰ VPDB and equilibrium δ18O, respectively.

Figure 5.

Stable isotopic composition of live (solid symbols) and dead (open symbols) C. neoteretis from sites within the HMMV (Sites PS66002, −003, −011, −013, ATL12, −18, −19, −22, −25a, and −25d), Site PS66015 from the outer margin, and reference Site ATL28 (R) not influenced by methane seepage. Error bars give standard deviation of replicates at single sites. Note different scale of axes. The δ13C and δ18O values are normalized to bottom water δ13CDIC of 1.11‰ VPDB and equilibrium δ18O, respectively.

Figure 6.

Stable isotopic composition of live (solid symbols) and dead (open symbols) C. reniforme from sites within the HMMV (Sites ATL12, −18, −19, −22, −25a, and −25d) and reference Site ATL28 (R) not influenced by methane seepage. Error bars give standard deviation of replicates at single sites. Note different scale of axes. The δ13C and δ18O values are normalized to bottom water δ13CDIC of 1.11‰ VPDB and equilibrium δ18O, respectively.

Figure 7.

Stable isotopic composition of live (solid symbols) and dead (open symbols) F. wuellerstorfi (circles), C. neoteretis (diamonds), and C. reniforme (squares) from all sites within the HMMV including PS66015 at the outer margin, and the reference site (R) not influenced by methane seepage. Error bars give standard deviation of replicates at single sites. The δ13C and δ18O values are normalized to bottom water δ13CDIC of 1.11‰ VPDB and equilibrium δ18O, respectively.

[20] Stable carbon isotope values of live and dead Fontbotia wuellerstorfi from reference Site ATL28 perfectly reflect the δ13CDIC of bottom water, and the measured δ18O values fit into the range of known disequilibrium as well. All δ13C values of live F. wuellerstorfi from inside HMMV are significantly lower than bottom water δ13CDIC (Tables 2 and 3a and Figure 4). Even the average δ13C of stained F. wuellerstorfi at PS66015 outside the crater are less than that of bottom water. All live specimens of F. wuellerstorfi from within HMMV are from pogonophores fields, picked from 49 stained specimens of the tube worm Sclerolinum contortum [Smirnov, 2000]. Specimens of the tube worm Oligobrachia haakonmosbiensis [Smirnov, 2000] and dead specimens of S. contortum did not carry any attached benthic foraminiferal tests. Dead specimens of F. wuellerstorfi with low δ13C were found at pogonophores Sites ATL22, −25d, and PS66002. At Sites ATL12 and PS66013, which were covered by bacterial mats, δ13C values of dead F. wuellerstorfi do not significantly deviate from bottom water δ13CDIC. The δ18O values of live specimens of this species vary within a narrow range of about 0.2‰ (Figure 4).

Table 3a. Student's t Characteristics of δ13C Distribution Within Håkon Mosby Mud Volcano Versus Control Site
SpeciesCalculated t ValueProbability, %Critical t ValueaH0 Rejected
  • a

    Critical values of t for appropriate degrees of freedom and selected levels of significance are given.

F. wuellerstorfi2.18952.02yes
C. neoteretis2.7297.52.57yes
C. reniforme2.77952.35yes

[21] The δ13C values of live and dead Cassidulina neoteretis at the reference site are 1.26 ± 0.15‰ lower than bottom water δ13CDIC of 1.11 ± 0.04‰ (Tables 2 and 3a and Figure 5), whereas inside the volcano live specimens are between 3.85 and 7.53‰ lower, all of which are from areas populated by pogonophores. Most interestingly, δ13C values of empty tests are consistently less depleted, no matter whether they are from pogonophores fields or bacterial mats, they all plot between −3.12 and −1.65‰, thus are still lower than the reference value, but considerably higher than the isotopic composition of live specimens. The δ18O values do not differ between stained and empty tests and vary around 0.85‰.

[22] Stained and empty Cassidulina reniforme tests at reference Site ATL 28 are by 1.61 ± 0.12‰ lower in δ13C (Tables 2 and 3a and Figure 6), whereas live specimens from pogonophoran tube worm fields inside the mud volcano are between 4.16 and 5.51‰ lower than bottom water δ13CDIC of 1.11 ± 0.04‰. Similar to C. neoteretis tests, most surprisingly empty test values plot much heavier at pogonophores as well as bacterial mat sites, namely, around −2.5‰. In contrast to the other species investigated, C. reniforme δ18O values of live specimens from sites inside HMMV are offset from their corresponding empty tests.

[23] Student's t statistics were used to evaluate whether samples of live foraminifera δ13C values from seep sites are from the same population as conspecifics from the control site, which is formulated as null hypothesis (Table 3a). As expected from Figures 4 to 7 already, it is confirmed that differences in δ13C values of all species from inside HMMV and the control site are significant to the 95% level at the minimum. On the contrary, comparison of δ13C values of PS66015 from the outer margin with δ13C values from inside HMMV clearly indicates that they are from statistical parent populations with equal variances (Figure 1). Furthermore, analogous student t testing for checking the mean δ13C of stained and empty tests, confirm that live and dead specimens of C. reniforme and C. neoteretis are not from the same statistical population, whereas δ13C values of F. wuellerstorfi are (Table 3b).

Table 3b. Student's t Characteristics of δ13C Distribution of Live Versus Dead Conspecific Specimens Within Volcano
SpeciesCalculated t ValueProbability, %Critical t ValueaH0 Rejected
  • a

    Critical values of t for appropriate degrees of freedom and selected levels of significance are given.

F. wuellerstorfi0.46901.39no
C. neoteretis2.3697.52.18yes
C. reniforme4.8099.53.71yes

4. Discussion

[24] Three main characteristic environments across the HMMV (central zone, bacterial mats and pogonophore colonies) were analyzed for their content of benthic foraminifera. Empty tests of benthic foraminifera were found in all samples except from the central area. However, only areas colonized by pogonophores, yielded calcareous cytoplasm-containing benthic foraminiferal tests in densities sufficient for stable isotope analyses. It was shown previously that foraminiferal abundance is lower at Guayamas Basin sites influenced by methane than at control sites and lowest in areas covered by bacterial mats [Perez et al., 2004]. Consistent with these findings, within HMMV the central area of active methane discharge is barren of living foraminifera. At sites with pogonophores, where methane is efficiently oxidized within the upper sediment by microorganisms [Milkov et al., 2004], we found high numbers of live specimens of both epifaunal as well as infaunal species. Standing stocks of between 20 and 100 specimens per cm3 may be favored by the fact that both oxygen as well as ample food is available in the uppermost sediment, the latter because the process of microbial methane oxidation is accompanied by synthesis of additional methanotrophic biomass. Indeed, total organic carbon contents of 0.7 to 1.3% in these sediments at HMMV are relatively high [Pimenov et al., 1999; Milkov et al., 2004]. However, these values are still within the range of total sedimentary carbon contents from similar environments and water depth without methane seepage at the Norwegian continental margin [Mackensen et al., 1985; Mackensen and Hald, 1988].

4.1. Low δ13C: Habitat-Dependent DIC Signal or Food-Related Signal?

[25] Both C. neoteretis and C. reniforme occupy a shallow infaunal microhabitat [Wollenburg and Mackensen, 1998b; Husum and Hald, 2004] and like many other endobenthic foraminiferal species they are probably able to tolerate low-oxygen conditions [cf. Bernhard and Bowser, 1999; Bernhard and Sen Gupta, 1999]. Although reproduction of C. neoteretis is related to phytodetritus deposition, the species does not invade the phytodetritus on top of the sediment [Mackensen and Hald, 1988; Seidenkrantz, 1995]. Therefore it was suggested that C. neoteretis feeds on bacteria associated with phytodetritus decomposition rather than on photosynthate organic matter itself [Gooday and Lambshead, 1989]. So C. neoteretis seems to be well preadapted to an environment enriched in methanotrophic biomass as found at HMMV in the uppermost sediment in pogonophores colonies.

[26] Generally, δ13C values of tests of infaunal species qualitatively reflect low δ13CDIC values of the pore water where the species calcified [McCorkle et al., 1990, 1995; Mackensen et al., 2000; Mackensen and Licari, 2004]. Lein et al. [1999] reported pore water δ13CDIC values from surface sediments inside the HMMV of −18.7‰ in the top 3 cm. From methane seeps in Monterey Bay, pore water δ13CDIC values as low as −45 or −59‰ in sediments below the upper 2 to 4 cm are measured [Rathburn et al., 2003]. Taking −18.7‰ as average pore water δ13C in the upper 3 cm of sediment at face value and 1.11‰ as bottom water δ13C, minimum δ13C values of −7.5‰ of C. neoteretis and −5.5‰ of C. reniforme tests, suggest that live specimens in this study exclusively calcified in the uppermost surface sediment reflecting a very steep pore water δ13C gradient within the top 3 cm of surface sediments. After correction for vital effects of C. neoteretis and C. reniforme of −1.3 and −1.6‰, respectively, pore water δ13CDIC values between about −3.9 and −6.2‰ can be estimated for the uppermost sediment close to the sediment/water interface. Moreover, significantly enriched δ18O values of live C. reniforme tests compared to empty ones and those from the control site suggest calcification in 18O-enriched pore water because of gas hydrate dissociation below the oxic surface sediment layer, a process occurring in peripheral areas of the HMMV [Hesse and Harrison, 1981; Ginsburg et al., 1999]. Since C. reniforme at HMMV occupies a microhabitat positioned deeper in the sediment than C. neoteretis (J. Wollenburg and A. Mackensen, unpublished data, 2005) it may calcify closer to the zone of clathrate stability where pore water is more strongly influenced by 18O-enriched water from dissociated clathrates, and thus is the only species investigated here that exhibits an increased δ18O value.

[27] Fontbotia wuellerstorfi, in contrast to C. neoteretis and C. reniforme, is unable to tolerate periods of low or no oxygen. Its microhabitat is epibenthic and, wherever possible, well elevated above the seafloor [Lutze and Thiel, 1989]. Indeed at HMMV, we found live F. wuellerstorfi almost exclusively attached to the pogonophoran tube worm S. contortum that is most abundant on oxidized surface sediments [Pimenov et al., 1999; Smirnov, 2000], and usually protrude about 3 cm into the water column (Figure 8). On the contrary, the second most abundant tube worm Oligobrachia haakonmosbiensis [Smirnov, 2000] occupies reduced surface sediments, and is thus less favorable for foraminifera and consequently occurs without attached F. wuellerstorfi.

Figure 8.

Scanning electron microscope image of (a) live (Rose Bengal stained) benthic foraminifer F. wuellerstorfi attached to live (Rose Bengal stained) pogonophoran tube worm S. contortum. (b) Higher magnification of same specimen as Figure 8a.

[28] The δ13CDIC values, 3.2‰ lower than average bottom water δ13CDIC of 1.11 ± 0.04‰ measured in directly overlying water from multiple-corer tubes at Site PS66002, may indicate the influence of isotopically light CO2 from methane oxidation. Water samples from just 40 cm above seafloor, however, do not show deviations from average bottom water mass δ13CDIC. Since in bottom water directly above the HMMV no methane is oxidized [Damm and Budeus, 2003], and stratification of the lowermost water column in this kind of environment that would allow for such high differences in deepwater δ13CDIC values is very unlikely, the depleted δ13CDIC values measured in multiple-corer water are probably due to pore water, which by gassing out and decompression during recovery of the multiple corer was admixed to the overlying bottom water in the tube [Mackensen et al., 1993]. Consequently, these depleted values are not regarded reflecting the actual bottom water DIC isotopic composition.

[29] Alternatively, extremely low δ13C values of foraminiferal tests may be the result of 13C-depleted ingested methanotrophic biomass on which the foraminifera prey as suggested by Hill et al. [2004]. This explanation relies on the hypothesis that benthic foraminiferal tests partly reflect the isotopic composition of ingested organic matter via incorporation of isotopically light metabolic CO2 into their carbonate shells. The magnitude of this 13C depletion is proportional to the amount of metabolic CO2 within the foraminifera's internal CO2 pool and is species specific [Erez, 1978; McConnaughey, 1989; McConnaughey et al., 1997]. Although methanotrophic biomass δ13C can reach around −60‰ [e.g., Hinrichs, 2001], at HMMV the δ13C of bacterial mats overlying reduced methane-rich sediments vary between −17.6 and −53.0‰ [Milkov et al., 2004], and sedimentary total organic carbon δ13C values of −30.7‰ are measured [Pimenov et al., 1999], much lower than −24.0‰, typical of photosynthate organic carbon in this area [Lein et al., 1999]. The δ13C of pogonophoran tube worm tissue varies from −34.9 to −56.1‰ [Pimenov et al., 1999; Milkov et al., 2004]. Demersal fishes that consume these tube worms reach values as low as −51.9‰ [Milkov et al., 2004].

[30] Fontbotia wuellerstorfi is a passive suspension feeder [e.g., Goldstein, 1999]. At HMMV F. wuellerstorfi lives attached to the tube worm species S. contortum, which is abundant only on oxidized surface sediments [Pimenov et al., 1999; Smirnov, 2000] and by this substantially protruding above the actual sediment surface (Figure 8). Elevated, F. wuellerstorfi scavenges with its pseudopodia particulate organic matter out of the near-bottom turbulent layers [Lutze and Thiel, 1989], which is in this specific environment isotopically extremely depleted. This food, via metabolic CO2, may deplete the internal CO2 pool of the foraminiferal cell in 13C, which finally may result in a decrease of its calcitic test δ13C value. Culture experiments with planktonic foraminifera, however, show only small shifts in shell δ13C despite of large variations in food δ13C [Spero and Lea, 1996], but virtually nothing is known about the isotopic composition of benthic foraminiferal food and its influence on test calcite in a deep-sea environment. So the severely 13C-depleted particulate organic matter at HMMV in the end may cause a negative δ13C signal of the foraminiferal tests, substantially deviating from bottom water DIC. Seasonal differences in the amount of phytodetritus mixed to the diet of the foraminifera and other reasons for varying amounts of extremely depleted versus less depleted organic material may account for the variability between samples and conspecific specimens [cf. Mackensen et al., 1993; Rathburn et al., 2003].

[31] Although not much is known about the ecology of pogonophores at HMMV, it is clear that these organisms do not have a gut but live on metabolic waste products of methylotrophic symbionts, which gain energy by oxidizing methane that they receive via the worm from below the sediment/water interface [Smirnov, 2000]. The resulting CO2 directly, or as dissociated to bicarbonate, may pass through the outer organic wall of the tube worm to which F. wuellerstorfi is attached (Figure 8), and in the end the foraminifer might use it during the calcification process. In this case the pogonophoran tube would just serve as a conduit for the depleted carbonate ions resulting from the methane oxidation taking place at their anchoring depth below the sediment/water interface.

4.2. Low δ13C: Primary Calcification Signal or Diagenetic Overgrowth?

[32] In this study we document a consistent offset between δ13C values of Rose Bengal stained and empty tests of endobenthic species C. neoteretis and C. reniforme, such that empty tests are much less depleted in 13C and closer to control values than stained ones (Figure 7). This is important, because hitherto it was documented that fossil or dead benthic foraminiferal tests at modern methane seeps are much more depleted than Recent or stained ones. To explain these findings, it was demonstrated, that empty and fossil shells were overgrown by precipitates from either highly 13C-depleted pore water or by bacterially mediated methane oxidation and associated carbonate precipitation [Rathburn et al., 2000; Hill et al., 2003; Rathburn et al., 2003; Torres et al., 2003]. Alternatively, it was suggested that stained specimens, just slightly depleted compared to empty tests, have been calcified during times of less or no methane seepage at a specific site [Rathburn et al., 2000].

[33] However, there is evidence that in dysoxic environments Rose Bengal staining does not properly differentiate between live individuals and those that have been dead for several weeks or even months, because a specimen brightly stained with Rose Bengal merely indicates the presence of cytoplasm, but nothing about the viability of that cytoplasm [e.g., Bernhard et al., 2001]. So, although it is not likely that stained individuals are recrystallized or internally overgrown, parts of the outer shell, where the organic lining might already have been decomposed or eroded, might be overgrown by carbonate precipitates. If, however, authigenic precipitates on stained but dead individuals occurred, why then do empty tests from the same sample show significantly higher δ13C values than stained ones, indicating less or no overgrowth? We rather suggest our results demonstrate that disturbance of sediments due to upward convection of pore water and fluid mud during active methane venting phases at these sites mingles tests just recently calcified with older ones secreted at intermittent times of less or no methane discharge. The degree of mixing in a given sample hence depends on the vicinity of the site to active or former venting and the average test size of the species, the latter simply because smaller tests are more prone to physical sediment transport by fluids and gases. Consequently, the smallest species occupying the deepest infaunal microhabitat, namely, C. reniforme, exhibits the highest and statistically most significant discrepancies between live and dead values.

[34] The above interpretation is supported by a study inferring temporal variation in seep activity from very low δ13C of benthic foraminifera in surface sediments with active methane release, underlain by sediments with isotopically nondepleted tests, and farther down core, by sediments with low values again [Panieri, 2003]. Moreover, faunal elements of Paleogene age that we encountered in surface sediment samples investigated in this study, and late Pliocene benthic foraminiferal species encountered in about 3-m-long gravity cores by Shilov et al. [1999], suggest that at the HMMV, mud volcano deposits have been transported up from below the Pliocene/Pleistocene boundary in a depth of 3 km.

5. Conclusions

[35] We show that at HMMV stained tests of the benthic foraminiferal species Fontbotia wuellerstorfi, Cassidulina neoteretis and Cassidulina reniforme are strongly depleted in 13C, more than hitherto reported. Although endobenthic C. neoteretis and C. reniforme may record a pore water signal, which partly might be explained by enhanced organic carbon decomposition, epibenthic F. wuellerstorfi δ13C values, more than 4‰ depleted versus bottom water dissolved inorganic carbon, are very unlikely to be due to this process, but rather the result of incorporation of 13C-depleted methanotrophic biomass upon which the species feeds. Possibly but less likely, depleted F. wuellerstorfi calcite δ13C values may partly be caused by low δ13CDIC that was transported within the pogonophoran tube from depth in the sediment up to F. wuellerstorfi living attached to the tube well above the sediment/water interface.

[36] We found empty tests of endobenthic species much less 13C depleted than stained specimens from the same site. Consequently, if stained specimens we analyzed for their stable isotopic composition were alive at the time of sampling, our observations are the best proof to date that methane release at cold seeps can be directly recorded in benthic foraminiferal shells during primary calcification of the test.


[37] We are grateful to shipboard scientific parties and masters and crews of R/V Atalante and PR/V Polarstern for help and professional guidance during sampling, as well as Beate Hollmann, Günter Meyer, and Susanne Wiebe for technical help during sample preparation and stable isotope measurements. Useful and constructive comments of Joan Bernhard, Gerald Dickens, and Juan Carlos Herguera helped in clarifying the final version. We acknowledge funding of the Deutsche Forschungsgemeinschaft under grant MA1942/4.