Impact of methane seeps on the local carbon-isotope record: a case study from a Late Jurassic hemipelagic section

Authors


Helmut Weissert, Geological Institute, ETH Zurich, 8092 Zurich, Switzerland. Tel.: +41 44 632 3715; fax: +41 44 632 1075; e-mail: helmut.weissert@erdw.ethz.ch

Abstract

An Oxfordian (Late Jurassic) hemipelagic succession from Beauvoisin (SE France) contains a pronounced, short-lived negative excursion in the bulk-carbonate carbon-isotope record, with an amplitude of 4‰. It was shown previously that the Beauvoisin paleoenvironment was impacted by hydrocarbon seepage. New isotopic data corroborate that methane was a significant constituent of these hydrocarbons. The negative excursion was caused by transient enhanced precipitation of 13C-depleted carbonate, mediated by anaerobic oxidation of methane. Despite its local diagenetic origin, the Beauvoisin excursion is similar in shape and duration to globally recognized negative C-isotope excursions that have been related to catastrophic, massive dissociation of methane hydrate. Shape and duration of negative excursions therefore cannot be used as an argument when determining their origin if they have not been shown to represent a global perturbation of the carbon cycle.

Introduction

Short-lived negative excursions are common in Phanerozoic carbon-isotope records. They can have multiple origins. Excursions recognized in several records from different paleolatitudes and paleoenvironments and in both inorganic- and organic-carbon records represent global perturbations of the carbon cycle. One such perturbation is sudden and massive dissociation of methane hydrates, used to explain e.g. the Paleocene-Eocene Thermal Maximum (PETM), Aptian (Early Cretaceous), Middle Oxfordian (Late Jurassic), and Toarcian (Early Jurassic) excursions (e.g. Dickens et al., 1995; Hesselbo et al., 2000; Jahren et al., 2001; Padden et al., 2001; Thomas et al., 2002). Negative excursions that are local and/or only present in carbonate and not in organic matter, have commonly been attributed to meteoric diagenesis or to oxidation of organic matter and subsequent precipitation of 13C-depleted carbonate cement (e.g. Humphrey et al., 1986; Hollis et al., 2005). Negative excursions also occur in modern methane-bearing sedimentary successions. These are caused when anaerobic oxidation of methane (AOM) induces authigenic carbonate precipitation (e.g. Rodriguez et al., 2000).

Here we investigate if negative C-isotope excursions preserved in fossil sedimentary successions can be related to AOM. Our natural laboratory is located near Beauvoisin in SE France, one of the few Mesozoic sites where methane seepage has been well documented based on sedimentological, isotope geochemical and biomarker evidence (Gaillard et al., 1985; Rolin et al., 1990; Hofmann and Bernasconi, 1998; Peckmann et al., 1999; Peckmann and Thiel, 2004). We will demonstrate that AOM altered bulk carbonate and produced a negative C-isotope excursion similar to excursions formed by methane hydrate dissociation.

Investigated section

The investigated section near Beauvoisin (SE France) is part of the hemipelagic ‘Terres Noires’ Formation of the subalpine basin (Fig. 1a), a deep epicontinental basin affected by synsedimentary tectonics related to the opening of the Alpine Tethys. The section is dated as cordatum ammonite zone of the Early Oxfordian (Bourseau, 1977), hence it is clearly older than the Middle Oxfordian transversarium zone bearing the negative C-isotope excursion interpreted by Padden et al. (2001) as a ‘methane hydrate dissociation spike’. The succession consists of dark marlstones with carbonate contents around 30 wt% (Fig.1b) and organic carbon contents near 0.5 wt% (Tribovillard et al., 1986). Carbonate consists of micrite, coccoliths, foraminifera and other small fossils; organic matter is a mixture between marine and terrestrial components (Tribovillard et al., 1986). The marlstones are bedded at decimetre to metre scale and show an alternation of lighter and darker colour (Fig. 1c) related to variations in the sediment composition, possibly caused by climatic cycles (Tribovillard et al., 1986). The deep basinal depositional environment of Beauvoisin is characterized by varying moderate (on average about 25 m Myr−1; Gaillard et al., 1992) sedimentation rates.

Figure 1.

 (a) Location map, paleogeography and stratigraphy of the studied Beauvoisin section. Location of the section is marked with a white star. Paleogeographical map modified and redrawn after Ziegler (1988). (b) Carbonate content and isotope record of the studied section showing a pronounced negative excursion in the marlstone bulk carbonate record. δ13C nodules measured in the cores of the nodules. These values are significantly lower in the lower ∼9 m of the section. (c) Photograph of the section showing large seep carbonate bodies, and with section line indicated. (d) A transect through one of the nodules shows increasing δ13C from −30‰ at the centre to −26‰ at the rim.

Tower or lense-shaped limestone bodies (Fig. 1c) with horizontal and vertical extents of up to several metres contain a fauna characteristic of hydrocarbon seep environments (Rolin et al., 1990; Gaillard et al., 1992; Peckmann et al., 1999). Biomarkers found in these seep limestones give evidence for the presence of archaea-performing AOM (Peckmann and Thiel, 2004). In analogy to modern hydrocarbon seeps (Ritger et al., 1987; Campbell, 2006), the marlstones contain numerous carbonate nodules between 5 mm and 10 cm in diameter, arranged in layers parallel to the bedding of the marlstones. These layers occur in intervals of a few centimetres up to several tens of decimetres.

We sampled 35 m of this section at 50 cm resolution for isotope analyses, a few metres away from one of the largest seep limestone bodies (Fig. 1c). When possible we collected marlstone and nodules from the same level. We used SEM analysis for investigating micritic carbonate. For detailed description of methods see Appendix.

Results

Carbonate contents in the marlstones fluctuate between 22.9 and 33.8 wt% (Fig. 1b). Nodules frequently show concentric layering (Fig. 1d), probably related to different pyrite contents. XRD analyses revealed that carbonate in the marlstones consist of calcite and small amounts of ankerite. Other mineral phases in the marlstones include quartz, chlorite and pyrite. All marlstone samples throughout the section have identical mineral compositions. Nodules are dominated by calcite and high-Mg-calcite (∼6.4 mol% Mg), quartz and pyrite. Dolomite was not detected by XRD either in marlstones or in nodules.

The marlstone bulk-carbonate δ13 C-record mostly fluctuates between 1 and 2‰, but is marked in its lower part by a pronounced negative spike. This excursion starts at 2‰ and decreases linearly to −2‰ within 4 m of sediment (Fig. 1b), corresponding to an estimated duration of several 104 years based on ammonite stratigraphy. The values rapidly return to 1‰ within 1 m, then gradually increase to 2‰ within another 15 m of sediment, and remain stable for the remainder of the section. This pattern does not correlate with carbonate content and any of the other data measured: oxygen-isotope values in marlstones are fairly constant at δ18O = −1.1 ± 0.24‰ (1 σ), except for two samples with δ18O = −2‰ and −3‰ respectively. Organic matter in marlstones shows stable δ13C values between −25‰ and −24‰ throughout the section (Fig. 1b). C/N ratios of the organic matter fluctuate between 15 and 20.

The δ13C values of the nodules measured in their cores vary between −37‰ and −17‰, whereby in the lower part of the section they average approximately −34‰, and above 9 ms they average approximately −25‰ (Fig. 1b). A transect through one of the nodules shows increasing δ13C from −30‰ at the centre to −26‰ at the rim. Oxygen-isotope values in the nodules average 0.17 ± 0.25‰ (1 σ). The δ13C values of the nodules in the lower part of the section are significantly lower than those reported from the seep deposit limestones (Peckmann et al., 1999).

SEM analyses document many well-preserved coccoliths, occasionally in formation as coccospheres (Fig. 2a). In all the samples, but especially in nodules, coccoliths are overgrown, but not replaced, by cement (Fig. 2b). Carbonate cement occurs as idiomorphic crystals sized between 2 and 15 μm (Fig. 2c), and as aggregates of rod-shaped calcite micro-crystals (Fig. 3). These rods are up to 3 μm long and occur as criss-cross aggregates of individual rods with open space in between. The aggregates seem to fill pore space (Fig. 4a). Aggregates commonly show moulds with the shape of euhedral carbonate crystals, or contain such crystals (Fig. 4b).

Figure 2.

 SEM micrographs from marlstone sample 6 (a) and from nodule sample 3.5 (b, c). (a) Entirely preserved coccosphere indicating early cementation of the sediment, as this structure possibly would have been destroyed in non-lithified sediments during burial. (b) Overgrowth of well preserved coccolith by possibly 13C-depleted calcite cement. (c) Single euhedral crystal, possibly calcite or dolomite. Note that dolomite does not occur in amounts detectable by XRD.

Figure 3.

 Rod-shaped microcrystals of calcite that are common in both marlstone and nodules throughout the Beauvoisin section. (a) Rods commonly occur in criss-cross aggregates with open pore space in between. (b) Their cross sections are sub-circular to polygonal. Length is a few μm, diameter below ∼200 nm.

Figure 4.

 (a) In places, aggregates of rod-shaped microcrystals seem to fill pore space. (b) Aggregates of rods commonly show moulds with the shape of euhedral carbonate crystals (white arrows), or contain such crystals (black arrow).

Discussion

All measured carbonates show uniform δ18O values throughout the section (Fig. 1b). Using a seawater oxygen-isotope value of −1‰ SMOW and average sample O-isotope values, mean temperatures for the ambient water during calcite precipitation of 12.2 °C (marlstones) and 6.4 °C (nodules) can be calculated (Kim and O'Neil, 1997). This is in agreement with an oceanic O-isotope signal only affected by early marine diagenesis. Together with the results of SEM microscopy (Figs 2–4) and XRD analyses, we consider that neither meteoric nor late high-temperature diagenesis altered the isotopic composition.

The low δ13C-values of nodules from the lower part of the Beauvoisin section (δ13C < −30‰, Fig. 1b,d) indicate that methane-derived carbon contributed significantly to nodule formation. The nodules in the upper part of the section have considerably higher δ13C-values (δ13C > −28‰, Fig. 1b). In this case, it cannot be determined if methane or sedimentary organic matter was the main carbon source.

Coleman (1993) suggested that the formation of nodules similar to those from Beauvoisin is linked to microbial activity. More recently, it has been shown in modern methane-bearing successions that nodular microbial aggregates and the formation of carbonates preferentially occur within the zone of AOM (Rodriguez et al., 2000; Reitner et al., 2005; Treude et al., 2005; Meister et al., 2006). AOM mediated by microbes is represented by the net reaction CH4 + SOinline image → HCOinline image + HS + H2O (Ritger et al., 1987). This reaction increases carbonate alkalinity and thus may cause the precipitation of 13C-depleted carbonate minerals (Ritger et al., 1987; Peckmann and Thiel, 2004).

It has been suggested that at times of elevated sedimentation rates, the zone of AOM will continuously move up in the sediment column adapting to the changing diffusion gradient of sulphate supplied from the ocean water (Rodriguez et al., 2000). The zone of AOM may be stabilized at a particular level, if sedimentation rates are low or if there is a break in sedimentation (Raiswell, 1988; Rodriguez et al., 2000), provided that other factors controlling depth of the zone of AOM [like the upward flux of methane (Dickens, 2001)] are constant. During such intervals, precipitation of 13C-depleted carbonates caused by AOM will be concentrated at a particular level in the sediment column, creating a low bulk-carbonate δ13C-value (Rodriguez et al., 2000). For Beauvoisin, rhythmic changes in sedimentation rate have been proposed (Tribovillard et al., 1986). Accordingly, the Beauvoisin excursion may reflect an interval of low sedimentation rates and stabilization of AOM.

A local diagenetic origin of the Beauvoisin excursion is corroborated by the fact that it is not recorded in bulk organic matter (Fig. 1b). If the excursion was caused by a mechanism affecting the total exchangeable carbon reservoir in oceans and atmosphere, like catastrophic methane hydrate dissociation, it should be recorded in all the materials produced from this reservoir, carbonates as well as organic matter. Furthermore, other sections do not record a negative excursion within the cordatum zone (Lavastre, 2002; Buschaert et al., 2004; Wierzbowski, 2004; Tremolada et al., 2006).

Despite their local origin related to AOM, the patterns of the negative excursions recorded at Beauvoisin and in the modern succession on Blake Ridge offshore south-eastern North America (Rodriguez et al., 2000) are strikingly similar to those recorded during the PETM and Toarcian events explained with a global perturbation of the carbon cycle (Fig. 5). In all records, the decrease in δ13C is linear in shape and occurs within a few 104 years; or in the case of the PETM event, within a few 103 years (Thomas et al., 2002). Recovery to pre-excursion values occurs over a much longer period. Apparently, local diagenetic processes may produce within a geologically short time an isotopic pattern which mimics negative C-isotope anomalies caused by the perturbation of the global carbon cycle. This intriguing similarity seems to be coincidental. However, AOM can occur at similar depths over extensive areas (e.g. Borowski et al., 1999). Thus, one might think of a situation where AOM causes negative excursions in C-isotope records at similar stratigraphic horizons across large regions, triggered by climatic or oceanographic changes. Such excursions would not be recorded in δ13Corg if most of the organic matter stems from the water column, although organic remains from archaea involved in AOM may be significantly depleted in 13C (Peckmann et al., 1999).

Figure 5.

 Comparison of the patterns of four negative C-isotope excursions. The two globally recognized excursions on the left side, Paleocene-Eocene thermal maximum (PETM) and Toarcian, have been explained with the catastrophic dissociation of methane hydrate (see text for references). The modern Blake Ridge and the fossil Beauvoisin cordatum zone excursions (this study) on the right-hand side are interpreted to have been caused by transient enhanced precipitation of 13C-depleted carbonate mediated by AOM (see text for references). The striking similarity in the pattern especially between the Toarcian, Blake Ridge and Beauvoisin excursions indicates that the duration and shapes of negative excursions are not necessarily diagnostic for their origin. C-isotope curves are not to scale in vertical direction.

The rod-shaped calcite crystals occurring in both marlstones and limestone nodules are common in Beauvoisin, but absent or extremely rare in time-equivalent and other Oxfordian sections lacking evidence for methane (Louis, 2006). Similarly shaped aragonite crystals have been reported from ooids (Fabricius and Klingele, 1970; Folk and Lynch, 2001), where they grow within an organic substrate (Loreau and Purser, 1973). The composition and concentration of an organic substrate determine the shape of crystals growing within it (e.g. Braissant et al., 2003). To decipher whether rods of Beauvoisin are related to organic substances produced by the consortia of microbes involved in AOM would require a method of micrometre-scale sampling for C-isotope analyses.

Conclusions

Beauvoisin is the first described Mesozoic example of a methane seep locality featuring several key characteristics of its modern counterparts, including seep limestone bodies, carbonate nodules and a negative excursion in the carbonate C-isotope record. This excursion was caused by transient enhanced precipitation of 13C-depleted carbonate mediated by AOM. The pattern of the excursion is strikingly similar to that observed in cases where massive dissociation of methane hydrate has been inferred to have caused the excursion. Hence, the duration and shapes of negative excursions in carbonate C-isotope records are not necessarily diagnostic for their origin, even if they are recognized in several sections. A diagnostic feature for global-scale excursions seems to be a covariance in δ13Ccarb and δ13Corg patterns. The occurrence of rod-shaped micro-crystals of calcite in Beauvoisin marlstones and nodules may be related to AOM.

Acknowledgements

The manuscript benefited from suggestions by P. Meister and T. Bontognali. We gratefully acknowledge the constructive reviews by J. Peckmann, G. Dickens, J. Whiteside and J. O'Neil. The Elektronenmikroskopisches Zentrallabor of the University of Zürich and A. G. Bittermann are acknowledged for providing lab-time and expertise with SEM-EDX analyses. We thank P. Pellenard, P.-Y. Collin, D. Fortwengler, M. Coray, M. Sanchez-Roman and H. Lan for assistance in the field and laboratory. Financial support was provided by the Swiss Science Foundation.

Appendix: Methods

Samples for carbonate C-isotope analyses were drilled and analysed following standard procedures with a VG Prism mass spectrometer equipped with an automated carbonate preparation module, where samples are reacted in 100% phosphoric acid at 90 °C to obtain CO2, or with a Thermo Delta V plus mass spectrometer equipped with a KIEL IV carbonate preparation module. Standard deviation of replicate analyses of a laboratory-internal standard (Carrara marble) calibrated to NBS 19 was better than ±0.1‰, standard deviation of replicate sample analyses was better than ±0.2‰. Samples for organic carbon C-isotope analyses were powdered and decarbonated with diluted HCl, and measured with a Carlo Erba elemental analyser coupled in continuous flow to a VG Optima mass spectrometer. Precision of replicate analyses of the laboratory-internal standard atropina calibrated to NBS 22 was better than ±0.3‰. All isotope values are reported in delta notation relative to Vienna PeeDee Belemnite (VPDB).

Samples for SEM microscopy were broken and fresh surfaces coated with gold before analysis either with a Philips XL30 or with a Zeiss Supra 50VP equipped with EDX. Inorganic carbon contents were determined with a UIC CM 5012 CO2 Coulometer. Data for mineral phase analysis were collected in the angular range of 5–80°2θ with 0.017 °step size and 40 s counting time, using a Philips X'Pert PRO MPD θ-θ powder diffractometer equipped with an X'Celerator linear detector. Phase search was performed in Pdf-2 crystal structure database using X'Pert High Score Plus software.

Ancillary