1. Department of Earth Sciences, University of Oxford, Parks Road, Oxford OX1 3PR , UK
2. Danish Centre for Earth System Science, University of Copenhagen, Juliane Maries Vej 30, DK-2100, Denmark
3. Fossil Fuels and Environmental Geochemistry, Drummond Building, University of Newcastle, Newcastle-Upon-Tyne NE1 7RU, UK

Correspondence to: Stephen P. Hesselbo¹ Correspondence and requests for materials should be addressed to S.P.H (e-mail: Email: stephen.hesselbo@earth.ox.ac.uk).

The better-known positive carbon-isotope excursion of the Early Toarcian is well illustrated by European organic-poor marine carbonates ( Figs 1 and 2), whereas the negative excursion is recorded primarily in marine organic-matter and associated carbonate (the latter previously considered to be largely of diagenetic origin²). The negative excursion is also present, albeit very rarely, in Tethyan limestones lacking organic enrichment⁸. Viewed together, these ¹³C curves depict a period of gradual rise in isotopic values which is interrupted by first a negative excursion, and then continued immediately afterwards as a positive excursion (Fig. 2).

Figure 1: Palaeogeographic maps for Toarcian sections in northwestern Europe^{19, 28}.

Land areas are shown in medium grey; organic-rich shale^{1, 2} in dark grey. 1, Wales, Mochras Farm borehole; 2, England, Hawsker Bottoms; 3, Denmark, Bornholm; 4, Portugal, Porto de Mós; 5, Spain, Fuente de la Vidriera.

High resolution image and legend (50K)

Figure 2: Carbonate carbon-isotope data through selected European Toarcian sections.

Wales, Mochras Farm borehole (ref. 2 and new data); Portugal, Porto de Mós²⁹, Spain, Fuente de la Vidriera⁸. See Fig. 1 for locations. We note that ¹³C_carb values from belemnites can reach up to +6 in the upper falciferum zone at the location shown in Fig. 3 (compare ref. 2). Tethyan zonal divisions are correlated with those of northern Europe using the first occurrence of the ammonite Hildaites¹⁴ as a time-synchronous marker. VPDB, Vienna Pee-Dee belemnite.

High resolution image and legend (72K)

Details of the negative excursion in marine organic matter are provided by a section at Hawsker Bottoms, Yorkshire, England (Fig. 3 ). There, the relationship of the excursion to black-shale deposition and high-resolution ammonite biostratigraphy is simply determined. Fossil wood is also present, preserved as coal (some pyritized) and 'jet', a form of coal that will take a high polish. Carbon-isotope values for wood (¹³C_wood) and marine organic matter ( ¹³C_org) show a close parallelism through the sampled upper tenuicostatum and lower falciferum zones (Fig. 3 ). The negative ¹³C excursion and high total organic carbon (TOC) characterize the exaratum subzone of the falciferum zone.

Figure 3: Carbon-isotope data through the lower Whitby Mudstone, Hawsker Bottoms, Yorkshire.

Graphic log from ref. 30. Biostratigraphy from ref. 14 and references therein. Negative excursion begins at level a, shows a large abrupt decrease in ¹³C values at level b, and terminates at level c. No palaeontological or sedimentary evidence exists for a hiatus at level b. Slight divergence between wood and bulk organic-isotope ratios may be due to changing relative abundance of organic components, but this cannot have been more than a minor factor in the genesis of the excursion. Approximate timescale is based on reported laminae thickness.

High resolution image and legend (80K)

Jet has previously been rejected as a monitor of atmospheric carbon-isotope composition because it is known that marine organic carbon has been incorporated during diagenesis⁵. To assess the isotopic effects of this process, six samples were subjected to solvent extraction and pyrolysis-gas chromatography-mass spectrometry (Py-GC-MS) of the residues (a method similar to ref. 9). Extractable organic-matter content ranges from 6.4% to 18.8% and is closely correlated with TOC (Fig. 3). The phenolic and napthalenic composition of the residue confirms terrestrial (wood) dominance. Extracted bitumen could not account for the isotopic excursion in raw samples because impregnation would have to have been more than 50% to move pure wood values away from a -25 baseline to the observed extent. Confirmation comes from carbon-isotope analyses of the residues, which show only minor differences from unextracted samples. We conclude that a primary stratigraphic signal is retained in the fossil wood, despite bitumen contamination from the enclosing shale.

In Yorkshire, the up-section trend towards more negative ¹³C_org begins at the same level as that where TOC values start to rise, but the isotopic values show an abrupt decrease about 2.4 m above this level (Fig. 3). Time represented in this section has been estimated previously based on an average 50-m thick lamina-couplet (assumed to be annual)^{10, 11}. The overall shift to lower ¹³C_org values occurs through a thickness of strata representing a minimum duration of about 100 kyr (some laminae may have been lost to non-deposition or erosion) and most of this change takes place between levels b and c in Fig. 3, representing 70 kyr. Importantly, the 2 jump across less than 22 cm of strata at level b (Fig. 3) may represent a much shorter interval, presumably less than about 5 kyr if there is no major hiatus. The decay pattern of the negative anomaly from its peak at level c to the top of the section analysed closely approximates an exponential curve, with relative stability attained after about a further 80 kyr. Thus, close similarities exist with patterns of timing described for the Late Palaeocene thermal maximum (LPTM)¹².

We also observe similar isotopic patterns in Toarcian wood from a near-shore setting, within the Bagå Formation, Bornholm, eastern Danish Basin¹³ (Figs 1 and 4). Samples analysed comprise two sets: (1) macroscopic wood, recognizable to the naked eye, up to branch-sized; and (2) microscopic wood (0.25 to 1 mm average dimension). Where possible, coal (hard, black, unstructured, higher density) and charcoal (soft, black, structured, lower density) were identified and analysed separately. Apart from wood and leaf cuticle there is little organic matter present in the section—the sediments comprise sand, silt and mud—and thus there is no possibility of contamination by marine carbon.

Figure 4: Stratigraphic and carbon-isotope data through the upper Bagå Formation at Korsodde, southwestern Bornholm.

The up-section transition to marine deposits is gradual and precedes the initiation of the negative excursion, but the Early Toarcian strata are truncated by a sharp erosion surface at 44 m. Graphic log, palynology, microfossil zonation, inferred environments and sea-level changes are summarized from ref. 13 with slight modification. Grey shading in biostratigraphic columns indicates uncertainty resulting from absence of index taxa. Stage assignment is based solely on microfossil data. Equivalent ammonite biostratigraphy is based on correlation of ¹³C excursion to the UK reference sections (Figs 2 and 3). Sp.-Le. indicates the Spheripollenites–Leptolepidites zone. c, s and b are the clay, sand and boulder grain-size categories.

High resolution image and legend (88K)

A striking feature of the ¹³C_wood data from Bornholm is the major negative excursion (Fig. 4). The ¹³C_wood values are not influenced by the nature of the preservation (that is, coal versus charcoal). Data from macrofossils and microfossils plot along identical stratigraphic trends, except between 44 and 46 m, where palynological evidence indicates an Aalenian age. Only at this higher level is it likely that the microscopic wood is reworked. Identification of the negative excursion allows precise correlation with the ammonite biostratigraphy for northwestern Europe¹⁴ illustrated in Fig. 3. Palynology indicates an abrupt decrease in salinity concurrent with initiation of the excursion¹³.

Our results prompt a re-evaluation of the nature and origin of the Early Toarcian oceanic anoxic event (OAE). It could be argued that isotopically light carbon was derived from a huge deep-ocean reservoir which was receiving organic carbon from surface environments but not returning it to the surface as CO₂ until a time of sudden overturn, much as suggested for the negative carbon-isotope excursion at the Permian–Triassic boundary (for example, ref. 15). However, necessary consequences of this mechanism are deposition of widespread anoxic or strongly dysaerobic facies before the earliest falciferum zone, and coeval development of a positive carbon-isotope excursion in the shallow-oceanic and atmospheric reservoirs. Neither of these phenomena is known from the stratigraphic record¹.

Recent age determinations indicate coincidence of the Early Toarcian OAE with large igneous-province formation, as is also the case for other OAEs and the LPTM (for example, ref. 16). A U-Pb age of 181.4 1.2 Myr BP from an ammonite-calibrated ash layer in western Canada corresponds to the variabilis zone of northwestern Europe¹⁷, and Ar-Ar dates from the Karoo–Ferrar continental flood basalts give ages of 183 1 Myr BP (ref. 18). The Early Toarcian OAE is two ammonite zones below variabilis, and is therefore roughly 2 Myr older. Volcanogenic CO ₂ from Karoo–Ferrar sources (¹³C -7) is unlikely to have effused rapidly enough to cause such a large excursion directly (compare with ref. 7). However, massive volcanism, together with intensified rift-related tectonic activity¹⁹, could conceivably have led to environmental changes sufficient to cause methane-hydrate dissociation⁷ in continental margin sediments, and generation of the negative ¹³C excursion (¹³C _methane -60). If this explanation is correct, Early Toarcian methane release occurred during eustatic sea-level rise (estimated to be about 30–90 m over approximately 1.5 Myr (refs 10, 20)), and therefore increasing hydrostatic pressure and hydrate stability. Thus, the proximal trigger for hydrate dissociation must have been raised temperature at the ocean floor and consequent realignment of the temperature gradient in the fault-disrupted sediment pile.

An effective means to increase bottom-water temperatures would have been a change in the global thermohaline circulation such that bottom-water formation switched from high to low latitude. This general mechanism has been proposed to explain the release of methane during the LPTM. Physical palaeoceanographic modelling suggests that bottom-water temperatures in Tethys could have risen by up to 5 °C in this manner²¹. Water flow through the northwestern European seaway (Fig. 1) is likely to have been related to global thermohaline circulation (as in ref. 22): with southward flow driven by low-latitude deep-water formation, modelled mean salinity in the seaway is reduced by 3–5 compared to northward flow²¹. The data for northwestern Europe indicate that salinity reduction, initiation of black-shale deposition and the beginning of the negative carbon-isotope excursion are all synchronous. Although the OAE is undeniably a global phenomenon, the introduction of cool, nutrient-rich bottom waters of northerly origin could help to explain the extraordinary nature of the organic enrichment in the northwestern European seaway.

The negative excursion in organic matter is -4 to -7, depending on location and chosen baseline (Figs 3 and 4). Some of this variability may be attributed to taxonomic or taphonomic factors. The size of the negative excursion in marine carbonates is also variable (-2 to -5) with some considerable uncertainty stemming from diagenetic effects. Most geological data suggest a two-fold amplification of ¹³C_org variations relative to stratigraphically equivalent ¹³C_carb (ref. 16), thought to be a consequence of changing atmospheric pCO₂ occurring in parallel with reservoir-scale shifts in isotopic values. Although higher atmospheric CO₂ concentrations are a likely result of methane release²³ they are as yet unquantified by geological data. As a basis for discussion, we take the magnitude of the excursion in the total exchangeable carbon reservoir to be between -2 and -3.5.

The mass of methane-hydrate carbon necessary to cause the negative excursion over this short timescale can be estimated using simple mass-balance equations⁷. Taking present-day mass and ¹³C estimates, we calculate that 1.5 10¹⁸ to 2.7 10 ¹⁸g of carbon is required for excursions of -2 or -3.5 respectively. These figures are 14–24% of the estimated present-day gas-hydrate reservoir (compare 14–19% for the LPTM using estimates of reservoir mass and isotopic composition derived from refs 7 and 24). If the synchronous burial of light organic carbon is taken into account, the mass of methane-derived carbon necessary to produce the excursion is very much larger. Release and oxidation of methane in such quantities would have reduced oceanic O₂ levels and thus helped promote organic-carbon burial, irrespective of any productivity changes driven by redistribution of nutrients. It is also likely that the bicarbonate and carbonate balance of the oceans would have been strongly disturbed and carbonate solubility increased⁷. In deep-water Tethyan settings, the common absence of primary carbonates carrying the negative carbon-isotope signal^{1, 2} is thus a predicted consequence of gas-hydrate dissociation.

Our estimates indicate that the Early Toarcian OAE could have involved the release of twice as much methane carbon than was the case for the LPTM. A negative ¹³C excursion is also associated with initiation in the Cretaceous of the Aptian OAE, or Selli event (about 120 Myr ago): the isotopic shift in this case is -2.5 to -3 in carbonates²⁵ and as great as -7 in wood¹⁶. The Selli event, which shows many overall similarities to the Early Toarcian OAE, was also characterized by enhanced organic-carbon burial, and had an estimated duration of less than 400 kyr (ref. 26). Thus, the Early Toarcian OAE and Selli event²⁷ together appear to be the two largest proposed hydrate-dissociation events out of at least three similar occurrences over the past 200 Myr of Earth history.

References

Acknowledgements

We thank F. Surlyk and G. Pedersen for help with this study. Useful critical comments were provided by J. Hudson. Palaeoceanographic modelling was carried out by C.J.B. in part while at the Geological Institute, University of Copenhagen. Organic carbon-isotope analyses were carried out at the University of Oxford Radiocarbon Unit thanks to T. O'Connell and M. Humm. D.M. Jones and P. Donohoe helped with Py-GC-MS analysis. Scientific discussion with L. Nielsen, G. Henderson and C. Jones is gratefully acknowledged. D.R.G. is funded by ESSO, C.J.B. by the Danish Natural Science Research Council, and H.S.M.B. by a NERC–industrial consortium (Rapid Global Geological Events project).