1. Introduction
The Kimmeridgian is the second stage of the Upper Jurassic and is dated from ~ 155.6 ± 4 Ma to 150.8 ± 4 Ma (Gradstein, Ogg & Smith, Reference Gradstein, Ogg and Smith2004; Ogg, Ogg & Gradstein, Reference Ogg, Ogg and Gradstein2008). International standardization of the geological time scale implies that the boundary between two stages should be represented by a physical outcrop, called a Global Boundary Stratotype Section and Point (GSSP). Thereafter, this outcrop could be used as a widely recognized reference in stratigraphy and therefore should be carefully studied. To be accepted as a GSSP, such an outcrop must exhibit a large set of biostratigraphic (appearance and disappearance of taxa that characterize the stage boundary), chemostratigraphic and magnetostratigraphic data. The base of the Kimmeridgian stage is not yet defined by an official reference succession. Difficulties mainly come from the provincialism of the fauna that characterized Tethyan and Boreal palaeobiogeographic realms during Late Oxfordian and Early Kimmeridgian times (Sykes & Callomon, Reference Sykes and Callomon1979). Indeed, each biogeographic province has its own specific biostratigraphic scale and it is therefore difficult to correlate the Oxfordian/Kimmeridgian boundary, notably between the Sub-Boreal and the Sub-Mediterranean provinces (e.g. Cariou & Hantzpergue, Reference Cariou and Hantzpergue1997; Zeiss, Reference Zeiss2003; Wierzbowski, Reference Wierzbowski2010). Those parallel scales are based on ammonite faunas and, for the time being, correlations between distinct palaeobiogeographic areas remain very uncertain. For these reasons, it is difficult to find a good GSSP candidate for the Oxfordian/Kimmeridgian boundary. The biostratigraphic issues are connected with a global palaeoclimatic (Cecca et al. Reference Cecca, Martin Garin, Marchand, Lathuiliere and Bartolini2005; Sellwood & Valdes, Reference Sellwood and Valdes2006) and palaeoenvironmental (Bartolini et al. Reference Bartolini, Pittet, Mattioli and Hunziker2003) context that engendered the scarcity of stratigraphic markers, in particular the lack of ammonites in the Tethyan carbonate platform. The biostratigraphic trouble is therefore connected to the specific eustatic (Haq, Hardenbol & Vail, Reference Haq, Hardenbol and Vail1987; Jacquin et al. Reference Jacquin, Dardeau, Durlet, Graciansky and Hantzpergue1998) and/or tectonic (Golonka, Reference Golonka2004; Louis-Schmid et al. Reference Louis-Schmid, Rais, Schaeffer, Bernasconi and Weissert2007) situation during the Upper Oxfordian/Lower Kimmeridgian interval.
Historically, the Kimmeridgian sedimentary series located near the village of Kimmeridge (Dorset, South of England, Fig. 1) is often used as a non-official reference to represent the base of the Kimmeridgian stage. Indeed, it is the place where Salfeld (Reference Salfeld1913) defined the Oxfordian/Kimmeridgian boundary, as indicated by the replacement of the pictoniid ammonite Ringsteadia by its descendant Pictonia. Unfortunately, this outcrop does not fulfil many criteria for being accepted as a GSSP (Wright, Reference Wright2003; Wierzbowski et al. Reference Wierzbowski, Coe, Hounslow, Matyja, Ogg, Page, Wierzbowski and Wright2006). Recently, a coastal exposure of the Flodigarry Shale Member of the Staffin Shale Formation (Staffin Bay, Isle of Skye, Scotland) has been proposed as a potential stratotype for the Oxfordian/Kimmeridgian boundary (Wierzbowski et al. Reference Wierzbowski, Coe, Hounslow, Matyja, Ogg, Page, Wierzbowski and Wright2006). This member is represented by a succession of bituminous and silty shales as well as limestone lenses and beds deposited through the Oxfordian/Kimmeridgian boundary. Furthermore, this outcrop seems to fulfil the criteria, as defined by Remane et al. (Reference Remane, Bassett, Cowie, Gohrbrandt, Lane, Michelsen and Naiwen1996), to be accepted as the GSSP for the base of the Kimmeridgian stage (Wierzbowski et al. Reference Wierzbowski, Coe, Hounslow, Matyja, Ogg, Page, Wierzbowski and Wright2006).
Locality map of the Isle of Skye in Scotland and location of the Flodigarry section (star); (map adapted from Ordnance Survey).
Before the proposal of Wierzbowski et al. (Reference Wierzbowski, Coe, Hounslow, Matyja, Ogg, Page, Wierzbowski and Wright2006) is accepted by the Kimmeridgian Working Group of the Subcommission on Jurassic Stratigraphy, it is very important to have an extensive and multidisciplinary study of this outcrop. To date, several studies have been conducted. They covered the fields of lithology, biostratigraphy (using both micro- and macrofossils), sedimentology, isotopic geochemistry and magnetostratigraphy (Turner, Reference Turner1966; Hudson & Morton, Reference Hudson and Morton1969; Wright, Reference Wright1989; Stancliffe, Reference Stancliffe1990; Morton & Hudson, Reference Morton, Hudson and Taylor1995; Riding & Thomas, Reference Riding and Thomas1997; Wright & Cox, Reference Wright and Cox2001; Hesketh & Underhill, Reference Hesketh and Underhill2002; Matyja, Wierzbowski & Wright, Reference Matyja, Wierzbowski and Wright2004, Reference Matyja, Wierzbowski and Wright2006; Wierzbowski, Reference Wierzbowski2004; Wierzbowski et al. Reference Wierzbowski, Coe, Hounslow, Matyja, Ogg, Page, Wierzbowski and Wright2006; Nunn et al. Reference Nunn, Price, Hart, Page and Leng2009; Przybylski et al. Reference Przybylski, Ogg, Wierzbowski, Coe, Hounslow, Wright, Atrops and Settles2010). But paradoxically, while these shales are particularly rich in organic carbon, available data about their organic molecular composition are very scarce and limited to the preliminary study of Foster et al. (Reference Foster, Grice, Riding and Greenwood2003), which precedes the proposition of Wierzbowski et al. (Reference Wierzbowski, Coe, Hounslow, Matyja, Ogg, Page, Wierzbowski and Wright2006). However, relevant information can be expected from the molecular composition of organic matter such as evolutionary (first appearance datum and last appearance datum of fossil molecules), palaeoenvironmental (palaeobiology, palaeoclimate, water palaeochemistry, etc.) or diagenetic signals. Indeed, solvent-soluble organic matter contains a wide diversity of molecular biomarkers that derive from biomolecules synthesized by organisms that lived within or in the vicinity of the depositional environment. Their study enables the determination of the nature of these organisms, the environment in which they grew and the diagenetic transformation that affected the initial biomolecules (Killops & Killops, Reference Killops and Killops2005; Peters, Walters & Moldowan, Reference Peters, Walters and Moldowan2005).
The Upper Oxfordian–Lower Kimmeridgian interval of Flodigarry is associated with particular evolutionary, palaeoenvironmental or diagenetic changes. The aim of this study is to supply additional information on biomarkers, palaeoenvironmental conditions and diagenetic history of this section of high stratigraphic interest using organic molecular geochemistry.
2. Geological setting
The potential stratotype outcrop for the Oxfordian/Kimmeridgian boundary proposed by Wierzbowski et al. (Reference Wierzbowski, Coe, Hounslow, Matyja, Ogg, Page, Wierzbowski and Wright2006) is located on the eastern coast of the Trotternish Peninsula of the Isle of Skye (northwest of Scotland), near Flodigarry village (Fig. 1). The beach faces the Poldorais sea strait, to the north of Staffin Bay. The Flodigarry outcrop lies on the tidal zone and is accessible only at low tide. The action of meteoric elements, sea and seaweed quickly changes the look of the foreshore. Jurassic shales and limestone beds are partially covered by Paleocene basaltic lava flows and boulders.
The studied deposits lie in the Hebrides basin. During Late Jurassic times, they were located near a palaeolatitude of 40° N, at the confluence of the Arctic boreal ocean and the large epicontinental sea covering the whole of Western Europe (Thierry, Reference Thierry, Dercourt, Gaetani, Vrielynck, Barrier, Biju-Duval, Brunet, Cadet, Crasquin and Sandulescu2000). The Hebrides basin was formed during extensional phases of the evolution of the Central and North Atlantic oceans (Fig. 2). This basin is part of a series of Mesozoic basins extending from Morocco to Greenland (Ziegler, Reference Ziegler1990; Morton & Hudson, Reference Morton, Hudson and Taylor1995; Nunn et al. Reference Nunn, Price, Hart, Page and Leng2009). Morton (Reference Morton1987) proposed that the initial phase of crustal stretching responsible for the Hebrides basin system started during Late Triassic times. This explains why the Jurassic sediments of the Hebrides basin were deposited on tilted blocks.
Lower Kimmeridgian palaeogeographic map adapted from Cecca et al. (Reference Cecca, Azema, Fourcade, Baudin, Guiraud, Ricou, Wever, Dercourt, Ricou and Vrielynck1993) and palaeobiogeographical provinces.
At the Isle of Skye, the Upper Oxfordian–Lower Kimmeridgian sedimentary series belongs to the Flodigarry Shale Member of the Staffin Shale Formation. The stratigraphy of the Flodigarry Shale Member was first defined by Sykes & Callomon (Reference Sykes and Callomon1979) then précised by Morton & Hudson (Reference Morton, Hudson and Taylor1995), Hesketh & Underhill (Reference Hesketh and Underhill2002) and Matyja, Wierzbowski & Wright (Reference Matyja, Wierzbowski and Wright2006). The Flodigarry Shale Member is 25 m thick and is composed of alternating strata of dark grey, silty and bituminous claystones with limestone beds. Presently, these strata are almost vertical and the differential erosion has emphasized the limestone beds, which are then used as marker-beds. Each stratum of the Staffin Shale Formation was identified by a number (after Sykes & Callomon, Reference Sykes and Callomon1979). Strata of the Flodigarry Shale Member range from number 33 (Upper Oxfordian) to number 45 (Lower Kimmeridgian). The Mesozoic deposits are cut by small faults. These Cenozoic faults determine eight sections (F1 to F8) (Fig. 3). Cenozoic dolerite dykes and sills also intrude the sequence. However, metamorphism and strong diagenetic alteration are visually restricted to the contact of the Mesozoic deposits.
Map of the foreshore at Flodigarry and details on beds and section (see text).
The palaeontological content (mostly ammonites and belemnites) of the Flodigarry Shale Member has been intensively studied in the past (Sykes & Callomon, Reference Sykes and Callomon1979; Wright, Reference Wright1989; Wright & Cox, Reference Wright and Cox2001). Ammonites found in this member belong to both Boreal and Sub-Boreal provinces. Biostratigraphic zonation (Fig. 4) of this section is therefore very accurate and spreads over three or four ammonite zones (for Sub-Boreal and Boreal zonations, respectively). The Oxfordian/Kimmeridgian boundary was placed at the base of the Bauhini (Boreal)/Baylei (Sub-Boreal) zones. The boundary is characterized by the first appearance of Pictonia with Prorasenia replacing Ringsteadia and Microbiplices (Sub-Boreal), as well as the first occurrence of Amoeboceras praebauhini (Boreal). The boundary is located ~ 1 m below the nodular limestone bed 36 (Wierzbowski et al. Reference Wierzbowski, Coe, Hounslow, Matyja, Ogg, Page, Wierzbowski and Wright2006). A problem still not resolved is the precise place of the boundary (Wierzbowski, Reference Wierzbowski2010): at the base of the flodigarriensis horizon or at the base of the densicostata horizon. The boundary was also dated by Re–Os geochronology at 154.1 ± 2.2 Ma (Selby, Reference Selby2007). New data concerning geochemistry and magnetostratigraphy were also recently published (Wierzbowski, Reference Wierzbowski2004; Nunn et al. Reference Nunn, Price, Hart, Page and Leng2009; Przybylski et al. Reference Przybylski, Ogg, Wierzbowski, Coe, Hounslow, Wright, Atrops and Settles2010).
Biostratigraphic framework of the Flodigarry Shale Member and stratigraphical position of the beds.
3. Material and methods
3.a. Samples
Laminated bituminous shales and limestone beds of Flodigarry were sampled during the summer of 2009. Fourteen samples were submitted to molecular organic analysis. Nine samples come from section 5 (bed nos. 37, 39, 40, 41, 42, 43, 44, 45a & 45b) and four samples from section 6 (beds 33, two samples of 36, 37 & 38). Two samples were taken in beds 36 (section F6: 36 in shale, 36c in limestone concretion) and 45 (section F5) in order to study potential lateral evolution of the molecular content. This sampling enables the study of strata of the Flodigarry Shale Member, from bed 33 to bed 45 (Fig. 4). Strata nos. 36, 40 and 44 represent limestone beds or nodule concretions. Others represent shales. Bed 38 is clearly the most bituminous. A major drawback of this outcrop is that the sedimentary rocks are covered by green and brown seaweeds (Fig. 5). The two most abundant taxa are Fucus vesiculosus (brown algae) and Enteromorpha intestinalis (green algae) and represent the main part of the visible biomass of the foreshore. In order to ensure that no molecular contamination was induced by these algae, including the fauna they shelter like microencrusters, bacteria, microalgae, etc., they were also separately sampled. Rock and algae samples were carefully stored in solvent-prewashed glass bottles before analysis to avoid further contamination. The two species of algae sampled on the premises were analysed using the same protocol as for the rock samples after crushing and drying for several days under a hood at room temperature.
Flodigarry exposure: shale (foreground), dolerite boulders and brown and green seaweeds.
3.b. Analytical procedure
Samples were crushed and finely powdered. Soluble organic matter was extracted with dichloromethane at 100 bars and 80 °C from 80 to 140 g of rock powder using an accelerated solvent extractor (DIONEX ASE 350). Steel cells filled with pulverized samples were prewashed. The purge gas was nitrogen. Two extraction cycles were performed to ensure that the soluble fraction was completely extracted. Elemental sulphur was removed by introduction of HCl-activated copper chips in vials containing the solvent and the extract. Dichloromethane was evaporated using a Zymark TurboVap LV and the extract was left to dry overnight and weighed. The hydrocarbon fraction was separated from the polar fraction using column chromatography on alumina, eluting successively with dichloromethane and methanol/dichloromethane (50/50 by volume), respectively. Hydrocarbons were fractionated using silica gel column chromatography to recover aliphatic, aromatic and a more polar fraction by successive elution with pentane, pentane/dichloromethane (65/35 by volume) and methanol/dichloromethane (50/50 by volume). The two polar fractions were both recovered in the same vial. For each polar fraction, an aliquot was silylated by reaction with N,O-bis-(trimethylsilyl)trifluoroacetamide (BSTFA) at 60 °C for 3 hours. Both non-silylated and silylated polar fractions were analysed by gas chromatography–mass spectrometry (GC-MS). Aliphatic, aromatic and polar fractions were analysed using GC-MS (HP 5890 Series II GC coupled to a HP 5971 MS; full scan and SIM modes used, with an ionizing voltage of 70 eV). The column was a DB-5 J&W capillary (length 60 m, internal diameter 0.25 mm with 0.1 mm of film thickness). The temperature programme was 70–315 °C at 15 °C min−1 to 130 °C, then 3 °C min−1 followed by an isothermal stage at 315 °C for 15 min. Helium was the carrier gas (1 ml min−1). All fractions were diluted in hexane (4 mg ml−1) and 1 μl of solution was injected for GC-MS analysis. Individual compounds were identified by comparison of their mass spectra with published spectra and by interpretation of mass spectrometric fragmentation patterns.
4. Results
4.a. Algae molecular signature
Both species of algae, Fucus vesiculosus and Enteromorpha intestinalis, present the same molecular fingerprint though they belong to distinct algae groups (brown and green algae, respectively). The aliphatic fraction is characterized by the abundance of short-chain n‑alkenes, mostly corresponding to heptadecenes, and long-chain n‑alkanes (Fig. 6b). The n‑alkanes range from n‑C22 to n‑C41, and show a unimodal distribution with a mode at n‑C36; longer n-alkanes are probably present but were not eluted during gas chromatography. A very slight odd-over-even predominance is observed between n‑C27 and n‑C32 (Fig. 6b). The aromatic fraction also presents some short-chain n‑alkenes as well as squalene, a common compound synthesized by most living organisms. Numerous compounds are present in the polar fraction. The two most abundant are phytol and fucosterol. This last compound is by far the most predominant solvent-soluble molecule synthesized by these two species of algae (Fig. 6a). Fucosterol seems to be one of the main components of algal tissues (e.g. Komura, Reference Komura1974).
Algal signature: (a) partial chromatogram of the polar fraction of the sample of the alga Fucus vesiculosus and mass spectrum of the fucosterol peak; (b) chromatogram of the aliphatic fraction of the sample of the alga Enteromorpha intestinalis, showing the n-alkane distribution.
4.b. Extraction yields and bitumen composition
Extraction yields were calculated for each sample. These yields represent the weight of solvent-soluble organic matter per weight of rock sample (expressed in mg g−1). So, they do not directly reflect the content of organic matter since only labile organic compounds are here recovered. However, their changes are generally correlated with organic matter content and therefore to the nature of the rock. Yields of soluble organic matter significantly fluctuate between 0.03 and 0.49 mg g−1 of rock (Fig. 7). No general trend can be recognized through the Flodigarry Shale Member but yields are lower in carbonated beds (e.g. beds 40 and 44) than in shale beds. Furthermore, the highest value occurs in bed 38, which was previously described as the more bituminous layer. Weight proportions of aliphatic, aromatic and polar fractions are quite similar in all samples (respectively, an average of 31, 16 and 63 %). Figure 8 represents these proportions in a ternary plot. All samples belong to a single cluster, except sample 38, which differs from the others in having a lower proportion of the aliphatic fraction.
Soluble yield of the organic matter (OM) (in mg of organic extract per g of rock), Carbon Preference Index (CPI), n-C24 +/n-C24 − ratio, pristane/phytane ratio and δ18O (values of Wierzbowski, Reference Wierzbowski2004) of the samples of clay and limestone beds of the Flodigarry section. The bed number rimmed by the grey circle is the bituminous bed; bed 36 rimmed in orange is the boundary bed between the Oxfordian and Kimmeridgian stage proposed by Wierzbowski et al. (Reference Wierzbowski, Coe, Hounslow, Matyja, Ogg, Page, Wierzbowski and Wright2006).
Ternary plot and relative proportions of aliphatic (ALI), aromatic (ARO) and polar (POL) fractions for each sample of the Flodigarry shale section.
4.c. n- and iso-alkanes
n-Alkanes (key ion m/z 57 in the aliphatic fraction) are major components of non-biodegraded sedimentary organic matter. They are not biomarkers sensu stricto because they can be produced by non-biological processes (as well as their precursors) and they have many potential biological origins (Eglinton & Hamilton, Reference Eglinton and Hamilton1967; Tissot et al. Reference Tissot, Pelet, Roucoche, Combaz, Campo and Goni1977; Brocks & Summons, Reference Brocks, Summons, Holland and Turekian2003). Nevertheless, n-alkanes can be useful to discriminate marine and continental contributions (Eglinton & Hamilton, Reference Eglinton and Hamilton1967; Caldicott & Eglinton, Reference Caldicott, Eglinton and Miller1973; Welte & Waples, Reference Welte and Waples1973). Except for sample 38, which will be discussed later, all samples present similar n-alkanes profiles. These profiles are characterized by a bimodal distribution (Fig. 9). The first mode includes n-C13 to n-C24 alkanes (short-chain n-alkanes), while the second includes n-C25 to n-C40 alkanes (long-chain n-alkanes). The first mode is generally attributed to algae and bacteria (Tissot et al. Reference Tissot, Pelet, Roucoche, Combaz, Campo and Goni1977). The second mode clearly displays an odd-over-even predominance. This imparity of long-chain n-alkanes is quantified by the Carbon Preference Index (CPI, calculation after Bray & Evans, Reference Bray and Evans1961). CPI values range between 1.8 and 2.9 (Fig. 7). Most of them are higher than 2 and the lowest value remains significantly higher than 1. These values indicate that the second mode can be assigned to epicuticular waxes of continental plants (Eglinton & Hamilton, Reference Eglinton and Hamilton1967; Caldicott & Eglinton, Reference Caldicott, Eglinton and Miller1973). This marked imparity was previously noticed by Foster et al. (Reference Foster, Grice, Riding and Greenwood2003). Thus, the organic matter preserved in the Flodigarry Shale Member appears to have both marine and terrestrial origins.
Bimodal distribution of the n-alkanes: (a) example of sample 39 (m/z 57); (b) example of sample 38 (m/z 57).
Bed 38 displays a different n-alkane profile. It is characterized by a unimodal distribution and the larger predominance of short-chain n-alkanes over the long-chain n-alkanes. No significant odd-over-even predominance is observed for long-chain n-alkanes, with a CPI close to 1. This indicates a very low, if not absent, terrigenous supply of organic matter at the time of its deposition (Fig. 7).
Since the first and second modes can be respectively attributed to marine and terrestrial origins, we calculated the n-C24 +/n-C24 − ratio in order to trace the stratigraphic evolution of the continental versus marine contribution. This ratio is calculated using the formula: n-C24 +/n-C24 − = (Σn-Ci/ Σn-Cj) with i = 25 to 34 and j = 17 to 24. This ratio fluctuates between 0.2 and 2.5, with the lowest value concerning bed 38. The lower part of the section (beds 33 to 38) is characterized by the lowest values without significant evolution in this part of the section (from 0.2 for bed 38 to 1.2 for bed 36). Then, the upper part of the section, from bed 39 upwards, is characterized by an increase, with the maximum in bed 42, followed by a decrease in the ratio (Fig. 7). Figure 7 presents the n-C24 +/n-C24 − ratio together with the isotopic curve of δ18O published by Wierzbowski (Reference Wierzbowski2004). Curiously, there is a correlation between these two curves. Indeed, in both cases, bed 38 presents the lowest values of the succession while bed 42 shows the highest values of n-C24 +/n-C24 − and δ18O.
The values of the pristane/phytane ratio (Pr/Ph) range between 0.7 and 2.3 and thus present significant evolution (Fig. 7). The lower part of the section (from bed 33 to bed 40) is characterized by values around 0.7 to 1.8. In the upper part of the section, Pr/Ph ratio progressively decreases from 2.3 (the maximum value) to 0.9 at the top of the Flodigarry Shale Member.
4.d. Hopanes
Hopanoids are triterpenoids derived from bacteriohopanepolyols, diplopterol and diploptene, which act as rigidifiers in the cell walls of bacteria. These last biomolecules are then degraded during diagenesis into unsaturated, saturated and aromatic hopanoids (Ourisson, Albrecht & Rohmer, Reference Ourisson, Albrecht and Rohmer1979). Hopanoids are essentially recovered in the aliphatic fraction (key ion m/z 191) and are widely used as markers for bacterial contribution. All the samples present a similar distribution of hopanoids, which range from C27 to C34. Their distribution in aliphatic fractions is characterized by three main features (Fig. 10; Table 1): (1) the abundance of hopenes like 30-norneohop-13(18)-ene, neohop-13(18)-ene and homohop-17(21)-ene; (2) the large predominance of ββ isomers followed by the βα (intermediate conformation or moretane) over αβ isomers; (3) the absence of 22S-homohopane epimers. Indeed, αβ-homohopanes are only represented by 22R epimers. Benzohopanes are also present in the aromatic fraction.
Partial chromatogram from the aliphatic fraction of sample 45a (m/z 191).
Peak assignments for hopanes and hopenes (m/z 191) represented in Figure 10
4.e. Steroids
Steroids are ubiquitous in sedimentary organic matter and derive from sterols, which are components of eukaryotic cell walls (Volkman & Maxwell, Reference Volkman, Maxwell and Johns1986). In the biosphere, they essentially contain 27, 28 and 29 carbon atoms. C27-sterols are common in animals and red algae. C28-sterols are generally produced by microalgae (such as diatoms) and fungi while C29-sterols are mostly synthesized by land plants. C29-sterols have also been observed in several diatom species (Rampen et al. Reference Rampen, Abbas, Schouten and Sinninghe Damste2010). Because the number of carbon atoms is often preserved during diagenesis, steroid biomarkers from ancient sediments are potentially good source indicators (Huang & Meinschein, Reference Huang and Meinschein1979).
In the Flodigarry Shale Member samples, geosteroids are essentially present in the aliphatic and polar fractions. In the aliphatic fraction, they are represented by steranes (key ions m/z 217 & 218), diasterenes (key ion m/z 257) and methyldiasterenes (key ion m/z 271) (Fig. 11). Furthermore, C29-aliphatic steroids are clearly more abundant than C27 and C28. Polar steroids are also recovered in the polar fraction (Fig. 12). The sample from bed 38 is characterized by a very low abundance of steroids, below the detection limit. Once more, all samples, except bed 38, display a similar distribution of aliphatic steroids. Unsaturated rearranged steroids (diasterenes and methyldiasterenes) are by far more abundant than steranes and all these compounds range from C27 to C29 (Fig. 11).
Chromatogram of the aliphatic fraction of sample 45a (m/z 217, 257, 271). Diasterenes and methyldiasterenes peaks are predominant. C26 to C29 – regular steranes having from 26 to 29 carbon atoms; Di-C27 to Di-C29 – diasterenes having from 27 to 29 carbon atoms; Me-Di-C27 to Me-Di-C29 – methyldiasterenes having from 27 to 29 carbon atoms.
Chromatogram of the polar fraction of sample 45 showing oxygenated plant biomarkers. The dot-dash box marks the magnified peak displayed under the whole chromatogram. Mass spectra of compounds a (podocarpa-8,11,13-trien-12-ol), b and c are also pointed out.
4.f. Higher plant biomarkers
Many cyclic sesqui- and diterpenoids preserved in sedimentary rocks are land plant biomarkers. Sesquiterpenoids are produced by most vascular plants (e.g. compounds of the cadinane class) while diterpenoids essentially come from pinophytes (e.g. compounds of the abietane class; Simoneit, Reference Simoneit and John1986). These plant biomarkers received very special attention for a decade because they are useful palaeofloristic proxies (Otto & Simoneit, Reference Otto and Simoneit2001). On the Isle of Skye, they were the main interest of the study of Foster et al. (Reference Foster, Grice, Riding and Greenwood2003) but it is important to note also that Riding & Thomas (Reference Riding and Thomas1997) indicated the presence of floristic evidence through pollens and fossil wood. A careful consideration of all aliphatic fractions failed to find any cyclic sesquiterpanes and diterpanes (key ions m/z 109 and 123). Thus, there is no trace of plant biomarkers in these fractions. However, aromatic and polar fractions are known to contain significant proportions of plant biomarkers. In the aromatic fractions, cadalene appears to be the most abundant compound. This diaromatic sesquiterpenoid derives from the diagenesis of cadinenes and cadinols produced by most vascular plants (Otto & Simoneit, Reference Otto and Simoneit2001). Systematically, cadalene is associated with its two monoaromatic precursors, calamene and calamenene. Aromatic abietanes are also present in these fractions. They are represented by retene (the triaromatic abietane) and its precursors like simonellite, dehydroabietane and tetrahydroretene (mono- and diaromatic abietanes). They mostly derive from the diagenesis of biological abietanes, which are produced by conifers (Otto & Simoneit, Reference Otto and Simoneit2001; Otto & Wilde, Reference Otto and Wilde2001). This is why cadalene versus retene ratios (retene/cadalene, HPI, etc.) were used to trace palaeoflora changes (Van Aarssen, Alexander & Kagi, Reference Van Aarssen, Alexander and Kagi2000; Hautevelle et al. Reference Hautevelle, Michels, Malartre and Trouiller2006). The retene/cadalene ratio was calculated for each sample. It is very constant from one sample to another and close to ~ 0.05 (Fig. 13).
Partial chromatogram from the aromatic fraction of sample 39 (m/z = 219, 234, 183 and 198).
The study of the polar fractions reveals the abundance of oxygenated plant biomarkers. One of its main components is the 3-ethyl-4-methyl-1H-pyrrol-2,5-dione. This maleimide is a very common component of sedimentary organic matter since it derives from chlorophyll a of algae and land plants (Grice et al. Reference Grice, Gibbison, Atkinson, Schwark, Eckardt and Maxwell1996).
Furthermore, a cadinol, the 7-hydroxycadalene, is also detected in all the polar fractions. This compound is reported in many extant plants like liverworts, conifers and angiosperms (e.g. Lindgren & Svahn, Reference Lindgren and Svahn1968; Rowe et al. Reference Rowe, Seikel, Roy and Jorgensen1972; Sankaram, Reddy & Shoolery, Reference Sankaram, Reddy and Shoolery1981; Nabeta et al. Reference Nabeta, Katayama, Nakagawara and Katoh1992). Owing to its cadinane structure, this compound is typically a biological precursor of cadalene.
Oxygenated diterpenoids are also present in high abundance. They are mainly represented by a phenolic and a keto-phenolic abietane, ferruginol and sugiol, respectively. Ferruginol appears to be the most abundant compound in the polar fractions. These oxygenated abietanes are known as the biological precursors of aromatic abietanes detected in our aromatic fractions. They are also specific to pinophytes (Otto & Simoneit, Reference Otto and Simoneit2001). Two other compounds are associated with these oxygenated abietanes; they also present a molecular weight of 286 Da, like ferruginol, which suggests a C20H30O formula. Furthermore, their mass spectra show peaks at m/z 175, 189, 201 and 271 (compounds b and c in Fig. 12), which suggests a structure close to that of ferruginol since the mass spectrum of the latter is also characterized by these same peaks. Unfortunately, up-to-date mass spectra interpretation and careful comparison with published mass spectra do not lead to a clear identification of these compounds. Nevertheless, they are probably functionalized tricyclic diterpenoids like ferruginol. Furthermore, podocarpa-8,11,13-trien-12-ol, an oxygenated diterpenoid of the podocarpane class (compound a in Fig. 12), is also detected in all the polar fractions (identification made by comparison with spectra of the Wiley database). Because podocarpanes also originate from pinophytes (Otto & Wilde, Reference Otto and Wilde2001), podocarpa-8,11,13-trien-12-ol probably has the same origin as oxygenated abietanes.
Careful attention was paid to chromatograms of the silylated polar fractions. The presence of ferruginol and sugiol is also highlighted as well as other oxygenated diterpenoids presenting a molecular weight of 358 Da (like the trimethylsilyl derivative of ferruginol). A few other diterpenoids are also detected but, as in the non-derivatized polar fractions, it was not possible to clearly identify these compounds.
4.g. Other biomarkers
Polycyclic aromatic hydrocarbons (PAHs) are considered markers of palaeovegetation fire events since they are produced during incomplete combustion of vegetal matter (e.g. Killops & Massoud, Reference Killops and Massoud1992; Jiang et al. Reference Jiang, Alexander, Kagi and Murray1998). In the Flodigarry Shale Member samples, PAHs are essentially represented by phenanthrene, fluoranthene and chrysene. PAHs of higher molecular weight are only present in trace amounts in the aromatic fractions of all samples. This low abundance, close to the detection limit, is typical of the ‘background noise’ of PAHs in sedimentary organic matter and cannot be linked to important fire activity.
Because marine sedimentary series dated from the Kimmeridgian were often deposited under anoxic conditions (e.g. The Kimmeridge Clay Formation, which is a major source rock in the North Sea; e.g. Gallois, Reference Gallois1976; Tyson, Wilson & Downie, Reference Tyson, Wilson and Downie1979; Oschmann, Reference Oschmann1988; Miller, Reference Miller and Huc1990; Powell, Reference Powell2010; see Chambers, Reference Chambers2000 for more references), biological markers of anoxia were carefully checked. For instance, isorenieratene is a carotenoid specifically synthesized by Chlorobiaceae (anoxygenic green sulphur bacteria that only live in a euxinic photic zone). They produce specific compounds like isorenieratene, which is degraded into isorenieratane, 2,3,6-trimethylalkylbenzenes and many other aromatics during diagenesis. Thus, these last biomarkers are well-known tracers for photic zone anoxia (Summons & Powell, Reference Summons and Powell1987). These compounds are not detected in the Flodigarry Shale Member. Other molecular features typical of anoxia were also checked (abundance of perylene, C35-hopanes > C34-hopanes, presence of C28-hopanes and gammacerane). Like other five-ring PAHs, perylene is close to the detection limit while C28- and C35-hopanes, as well as gammacerane, are not detected.
5. Discussion
5.a. Possible contamination by extant algae
The molecular analysis of extant seaweeds that cover the studied outcrop (Fig. 5) shows that they both synthesize high amounts of fucosterol. This biosteroid is typical of extant algae and has never been, to our knowledge, detected in rocks. This is probably owing to the fact that it is quickly transformed into geosteroids during diagenesis. Thus, in this case, fucosterol can be used as a specific marker to trace potential contamination of fossil organic matter by recent organic matter. No trace of fucosterol coming from the extant seaweeds has been detected in the organic extracts of all rock samples. Furthermore, both species of algae exhibit an n-alkane pattern which is clearly different to those of the rock samples (Figs 6, 9) The former is characterized by a mode at C35 and a low odd-over-even predominance from n-C27 to n-C32 while the latter presents a mode for long-chain n-alkanes at C27 and high odd-over-even predominance. These differences in the n-alkane patterns also support that the n-alkanes recovered in the rock extracts do not originate from the seaweeds that cover the outcrop. Moreover, many major compounds present in the algae extracts, like n-alkenes, are not recovered in the rock extracts. Therefore, all these significant differences between the molecular signature of the seaweeds and those of the rocks indicate that the samples of the Flodigarry Shale Member are not contaminated by organic matter coming from the extant seaweeds.
5.b. Maturity and organic matter preservation
Organic matter is very reactive to thermicity and its effects can be efficiently evaluated using many hydrocarbon geochemical parameters. CPI is one of these parameters. Indeed, the odd-over-even predominance of long-chain n-alkanes, inherited from biosynthesis in terrestrial plants, is progressively reduced during thermal diagenesis (Tissot & Welte, Reference Tissot and Welte1984). Most of the CPIs calculated in this study are close to 2 and such values are typical of immature sedimentary rocks.
Hopanoids are among the most sensitive biomarkers to diagenetic stresses like thermicity and oxidation. These stresses progressively hydrogenize hopenes (unsaturated hopanes) into hopanes, transform hopanes in the biological conformation (ββ) into hopanes in the geological conformation (αβ) and induce the appearance of 22S-homohopanes at the expense of 22R-homohopanes (Faure, Landais & Griffault, Reference Faure, Landais and Griffault1999; Seifert & Moldowan, Reference Seifert and Moldowan1981).
The distribution of hopanoids in the Flodigarry Shale Member is characterized by several features, like the abundance of hopenes, the large predominance of ββ-hopanes over the αβ- and βα-hopanes as well as the absence of the R isomers for homohopanes. This distribution indisputably reveals the thermal immaturity of the Flodigarry Shale Member. Steroids also carry information on thermal maturity. In the Flodigarry Shale Member, diasterenes and methyldiasterenes are by far more abundant than steranes. Similarly to hopenes, diasterenes and methyldiasterenes are also reliable indicators of a thermal immaturity owing to the preservation of a double bond during diagenesis. Furthermore, the occurrence of functionalized plant biomarkers like ferruginol and sugiol is another remarkable feature consistent with the thermal immaturity of the Flodigarry Shale Member. Indeed, these molecules are those directly synthesized by living vascular plants and have not yet undergone any change during diagenesis. This also attests to the lack of a significant heat flux that could transform organic matter. Overall, the preservation of still functionalized bioterpenoids (meaning terpenoids directly synthesized by living organisms without any transformation during transport, sedimentation and diagenesis) clearly proves the excellent preservation of this sedimentary organic matter deposited some 150 million years ago. The preservation of such bioterpenoids in sedimentary deposits as old as Jurassic is clearly uncommon. Their recognition in deposits older than 50 Ma are very scarce because they are metastable and, thus, rapidly degraded during the first steps of diagenesis. In fact, to our knowledge, there are only two reports of biomolecules preserved in sedimentary deposits older than Kimmeridgian. One is dated as Callovian and was recently found in the Papile Formation (Poland), which is characterized by a similar sedimentary facies (Marynowski & Zatoń, Reference Marynowski and Zatoń2010). Secondly, the oldest biomolecules known come from the Bajocian to Bathonian fossil wood of Poland (Marynowski et al. Reference Marynowski, Otto, Zatoń, Philippe and Simoneit2007a ,Reference Marynowski, Zaton, Simoneit, Otto, Jedrysek, Grelowski and Kurkiewicz b , Reference Marynowski, Philippe, Zatoń and Hautevelle2008). Thus, this makes the Flodigarry Shale Member the third oldest sedimentary deposit to contain well-preserved biomolecules and emphasizes the exceptional preservation conditions associated with these deposits. Because these molecules are those directly biosynthesized by terrestrial plants and the samples were directly taken from an outcrop in poor conditions, the question may arise as to whether these terpenoids are actually fossils rather than recent molecules. But as indicated earlier in this paper, contamination by algae covering the outcrop, or by other extant organisms located near the outcrop, is unlikely. This makes sure that these bioterpenoids are indeed fossils and not recent. This exceptional preservation can be explained by a physical protection of the organic matter just after its deposition by the enclosing matrix, especially by clay minerals. The organic matter is thereafter isolated from degrading agents of the depositional environment by the fine clayey sediment. Another consequence is the excellent preservation of ammonites with their nacreous skeletal tissues (Sykes & Callomon, Reference Sykes and Callomon1979; A. Lefort, unpub. Ph.D. thesis, pl. 22, Nancy Univ., 2011). The effect of clay minerals on the organic matter is also highlighted by the rearrangement of steroids during early diagenesis. Indeed, diasterenes and methyldiasterenes are typically formed from regular steroids via catalysis on the surface of clay minerals like smectite (Sieskind, Joly & Albrecht, Reference Sieskind, Joly and Albrecht1979). A similar protection effect by clay minerals was previously reported by Hautevelle et al. (Reference Hautevelle, Michels, Malartre, Elie and Trouiller2007) for the Callovo-Oxfordian claystones of the Paris Basin. It is worth noting that the low thermal impact of Cenozoic dolerite dykes and sills observed in the field is confirmed by the excellent preservation of the organic matter.
5.c. Source of organic matter, palaeoenvironmental and stratigraphic implications
The molecular composition of the organic extracts indicates that the organic component of the Flodigarry Shale Member is a mixture of both autochthonous and allochthonous organic matter. Autochthonous organic matter derives from tissues of marine organisms (algae, plankton) and bacteria that lived within the sedimentary basin. Their contribution is indicated by the presence of short-chain n-alkanes, which typically derive from these organisms. Furthermore, C27- and C28-steroids also originate from marine eukaryotes and microalgae or fungi, respectively. Their presence attests to the occurrence of these organisms in the sedimentary environment. A bacterial contribution is also revealed by the presence of hopanoids. Allochthonous organic matter derives from tissues of terrestrial organisms, which lived outside the sedimentary basin, on continental areas. Their contribution is indicated by the presence of long-chain n-alkanes displaying a significant odd-over-even predominance. These n-alkanes typically derive from the cuticles of terrestrial plants (e.g. Caldicott & Eglinton, Reference Caldicott, Eglinton and Miller1973). The supply of organic matter coming from emerged lands is also supported by the abundance of C29-steroids and vascular plant biomarkers. This mixture of marine and terrestrial organic matter is illustrated here by the bimodal distribution of n-alkanes (Fig. 9). Indeed, the first mode (short-chain n-alkanes) represents the autochthonous contribution while the second one (long-chain n-alkanes) represents the allochthonous one. The n-C24 +/n-C24 − ratio fluctuates significantly from the bottom to the top of the stratigraphic succession (Fig. 7), indicating fluctuations over time of the terrestrial supply in the depositional environment. It is worth noting that the more bituminous bed (bed 38) is characterized by a particularly low terrestrial contribution as indicated by a unimodal distribution of n-alkanes and low amount of plant biomarkers. Apart from bed 38, all the samples from the Flodigarry Shale Member are characterized by a significant terrestrial contribution. Indeed, even if the sedimentary environment was clearly marine, the abundance of long-chain n-alkanes has the same magnitude as those of short-chain n-alkanes, and terrestrial plant biomarkers prevail over other compounds in the aromatic and polar fractions. This can be linked to the palaeogeography of Scotland during the Oxfordian/Kimmeridgian transition. The Flodigarry area was located in a narrow sea channel between the Hebrides and Irish massifs (Fig. 2). From this palaeogeographic location, it is not surprising to record a substantial terrestrial supply in this sedimentary succession. The close position of these two emerged lands can also explain the occurrence of fossil wood and the excellent preservation of plant biomolecules, since they did not have time to be degraded during their transport from the land to the basin.
Globally, the signal is identical in all samples; only bed 38 shows some remarkable features. Total organic carbon (TOC) illustrated in Nunn et al. (Reference Nunn, Price, Hart, Page and Leng2009) indicates values between 1 and 3 wt % in the Flodigarry Shale Member. According to this paper, the δ13Ccarb and δ13Corg curves have a good correlation during Early Kimmeridgian times. The brief positive fluctuation observed during the Cymodoce Zone cannot be compared with our organic matter values and consequently neither connected with the special bed 38.
Nevertheless, the potential position of the boundary stratotype would be placed in bed 36, which does not display significant changes in molecular content. Besides being the more bituminous bed, bed 38 provided the highest value of soluble organic matter and the lowest value of the n-C24 +/n-C24 − index. It exhibits also the lowest value in δ18O of belemnites. These values are interpreted as the co-occurrence of a warming of sea waters (δ18O), a high preservation of organic matter and the lowest contribution of terrestrial plants (n-C24 +/n-C24 −). We see two ways to conciliate these facts:
(1) A purely climatic explanation that changes the temperature of water, the proportion of sources of organic matter and its preservation on the sea bottom.
(2) A combined climatic and eustatic explanation in which the warming initiates a sea level rise that modifies the sources of organic matter in the basin. In this situation, bed 38 becomes a maximum flooding event with a certain interest in terms of correlation. A test could be performed in the palynological assemblages. Quantification of the palynomorphs could contribute to a palaeoclimatic test (cf. Abbink et al. Reference Abbink, Targarona, Brinkhuis and Visscher2001).
The Pr/Ph ratio is widely used as a marker of redox conditions of depositional environments. In oxic conditions, phytol is preferentially degraded into pristane while it is transformed into phytane in reducing conditions (Didyk et al. Reference Didyk, Simoneit, Brassell and Eglinton1978). This is why Pr/Ph < 1 indicates reducing conditions while Pr/Ph > 1 indicates oxic environments. In the case of the Flodigarry Shale Member, most of the Pr/Ph values fluctuate between 1 and 2. Such values are difficult to interpret in terms of redox conditions (Peters et al. Reference Peters, Walters and Moldowan2005). Regardless, as samples display Pr/Ph < 1, it is very likely that reducing conditions prevailed at least during certain periods. This is also supported by sedimentological features like the laminated structure of the sediment, the preservation of fossil wood and a relatively high organic matter content. Furthermore, the scarcity of benthic macrofauna (only some small bivalves and gastropods; Sykes & Callomon, Reference Sykes and Callomon1979) and the absence of microfauna could be also related to poor oxygenation of the lower part of the water column. Astartidae (bivalves) are often associated with a confined environment (Boyer & Droser, Reference Boyer and Droser2007). Their small size and the poor taxonomic diversity indicate that palaeoenvironmental parameters, like salinity or oxygenation, are restrictive (Oschmann, Reference Oschmann, Tyson and Pearson1994). Reducing conditions could be linked to the palaeogeographic configuration of the Isle of Skye. Indeed, the narrow channel in which the sedimentation took place was probably more or less isolated from the open seas that are the Arctic Sea and the Western Tethys (Morton & Hudson, Reference Morton, Hudson and Taylor1995).
Furthermore, certain plant biomarkers have a palaeochemotaxonomic value. This means that they are specific to certain plant taxa. For instance, unlike cadalane sesquiterpenoids (calamenene, calamene, cadalene and 7-hydroxycadalene), which can be linked to all vascular plants, podocarpanes and abietanes are more specifically related to pinophytes (Otto & Wilde, Reference Otto and Wilde2001). As indicated earlier, the molecular composition of the Flodigarry Shale Member is characterized by relatively high amounts of a podocarpane (podocarpa-8,11,13-trien-12-ol) and several abietanes (retene, dehydroabietane, simonellite, tetrahydroretene, ferruginol and sugiol). This indicates the wide occurrence of conifers within the palaeovegetation close to the sedimentary environment at the time of deposition. Fortunately, several biological abietanes have been preserved and kept their initial palaeochemotaxonomic value. This enables us to refine the palaeofloristic composition on the emerged lands. Indeed, the abundance of phenolic and keto-phenolic abietanes indicates the occurrence of a significant proportion of pinaceous conifers, which preferentially synthesize abietane acids. On the other hand, phenolic and keto-phenolic abietanes are produced by Cupressaceae sensu lato, Podocarpaceae as well as the extinct family Cheirolepidiaceae (Nguyen Tu et al. Reference Nguyen Tu, Derenne, Largeau, Mariotti, Bocherens and Pons2000; Otto & Wilde, Reference Otto and Wilde2001; Menor-Salvan et al. Reference Menor-Salván, Najarro, Velasco, Rosales, Tornos and Simoneit2010). The absence of tetracyclic abietanes (e.g. kaurane or phyllocladane) might rather point to the absence of Araucariaceae on continental areas since these conifers essentially produce these kinds of compounds (Otto & Simoneit, Reference Otto and Simoneit2001). Thus, we could imagine forests composed of conifers, and more precisely of Cupressaceae, Podocarpaceae and/or Cheirolepidiaceae, on the emerged lands. This observation is in accordance with the data of Riding & Thomas (Reference Riding and Thomas1997) who studied palynomorphs of the Staffin Shale Member. One of the prominent forms is Classopollis classoides, which is the gamete of Cheirolepidiaceae (Vakhrameev, Reference Vakhrameev, Litvinov and Hughes1991; Alvin, Reference Alvin1982). Indeed, most of the pollen grains are bisaccates. This morphological type of pollen is common in Mesozoic fossils (Traverse, Reference Traverse1988) and is often characteristic of conifers. At this stage, it is important to note that this palaeofloristic interpretation could be improved by the identification of the few compounds that remain still unidentified (e.g. compounds b and c in Fig. 12). This emphasizes the fact that significant progress must be made in the field of palaeochemotaxonomy, notably in the identification of plant biomarkers and in the determination of their botanical origin.
6. Conclusion
The soluble organic matter of the Oxfordian/Kimmeridgian shale samples of the Flodigarry succession is characterized by an exceptional preservation of hopanoids and biomolecules. The geochemical composition of all samples suggests a thermally immature organic matter. The distribution of n-alkanes shows marine and terrestrial contributions with a predominant continental contribution consistent with the presence of terrestrial plants, and with the palaeogeographical situation of the Hebrides islands. The high amount of plant biomarkers like ferruginol, sugiol, cadalene, retene, etc. suggests a palaeovegetation largely composed of conipherophytes on the nearest emerged landmasses (Scottish Massif, Hebrides Platform). From a stratigraphic point of view, neither evolutionary events nor drastic change in palaeoenvironments can be deduced from the molecular content of these sedimentary rocks. Thus, it cannot support any improvement of the Oxfordian/Kimmeridgian boundary location. The global qualitative signal of organic matter in the Flodigarry section does not show any obvious heterogeneity from the bottom to the top, except the bituminous bed 38, which is characterized by a low n-C24 +/n-C24 − ratio, a high Pr/Ph ratio and a small continental contribution. This specific bed, interpreted as a maximum flooding event, could be interesting for sequential correlations.
Acknowledgements
We thank Scottish Natural Heritage who allowed us to sample the protected Flodigarry outcrop. We are grateful to the National Agency for Radioactive Waste Management who funded this project. Also, we thank R. Michels who is in charge of the Analysis and Reactivity of Organic Matter lab (AROM) and Mélanie Gretz for her help in the field. We thank Leszek Marynowski, Armelle Riboulleau and the Geological Magazine editorial committee for their comments and their constructive criticism that improved the manuscript.
