Methane seepage in a Cretaceous greenhouse world recorded by an unusual carbonate deposit from the Tarfaya Basin, Morocco

During the Cretaceous major episodes of oceanic anoxic conditions triggered large scale deposition of marine black shales rich in organic carbon. Several oceanic anoxic events (OAEs) have been documented including the Cenomanian to Turonian OAE 2, which is among the best studied examples to date. This study reports on a large limestone body that occurs within a black shale succession exposed in a coastal section of the Tarfaya Basin, Morocco. The black shales were deposited in the aftermath of OAE 2 in a shallow continental sea. To decipher the mode and causes of carbonate formation in black shales, a combination of element geochemistry, palaeontology, thin section petrography, carbon and oxygen stable isotope geochemistry and lipid biomarkers are used. The 13C‐depleted biphytanic diacids reveal that the carbonate deposit resulted, at least in part, from microbially mediated anaerobic oxidation of methane in the shallow subseafloor at a hydrocarbon seep. The lowest obtained δ13Ccarbonate values of −23·5‰ are not low enough to exclude other carbon sources than methane apart from admixed marine carbonate, indicating a potential contribution from in situ remineralization of organic matter contained in the black shales. Nannofossil and trace metal inventories of the black shales and the macrofaunal assemblage of the carbonate body reveal that environmental conditions became less reducing during the deposition of the background shales that enclose the carbonate body, but the palaeoenvironment was overall mostly characterized by high productivity and episodically euxinic bottom waters. This study reconstructs the evolution of a hydrocarbon seep that was situated within a shallow continental sea in the aftermath of OAE 2, and sheds light on how these environmental factors influenced carbonate formation and the ecology at the seep site.


INTRODUCTION
The Cretaceous was a period in Earth history when major episodes of oceanic anoxic conditions led to large scale accumulation of organic matter in marine sediments (Arthur & Sageman, 1994;Larson & Erba, 1999;Jones & Jenkyns, 2001). These organic-rich deposits have been termed black shales, and phases of their coeval and repeated occurrence in the Phanerozoic rock record have been referred to as oceanic anoxic events (OAEs, Schlanger & Jenkyns, 1976;Weissert et al., 1979;Schlanger et al., 1987;Kerr, 1998;Jenkyns, 2003;Meyer & Kump, 2008). Sea-level was markedly elevated during some periods of the Cretaceous compared to the present due to the warm, equable climate (Sames et al., 2016 and references therein). The expansion of shallow epicontinental seas, where upwelling and enhanced nutrient flux from the continents increased primary production, enabled the accumulation of organic carbon-rich sediments in water depths of several hundreds of metres (Hallam & Bradshaw, 1977;Weissert, 1981;Kuhnt et al., 1990;Arthur & Sageman, 1994). Two processes held primarily responsible for the deposition of black shales are (1) pulses of enhanced primary productivity accompanied by elevated atmospheric carbon dioxide concentrations and sea surface temperatures (Barron, 1983;Ingall et al., 1993;Forster et al., 2007), and (2) the increased preservation of organic matter within restricted, stratified, oxygen-deficient basins (Mascle et al., 1997;Meyer & Kump, 2008). With supporting evidence for both scenarios, it is likely that both production and preservation in stratified basins increased sequestration rates of organic carbon in Cretaceous marine sediments (Jenkyns et al., 2007;Owens et al., 2013).
Several OAEs have been documented during the Cretaceous including the Aptian to Albian OAE 1, the Cenomanian to Turonian OAE 2 and the Coniacian to Santonian OAE 3 (Jenkyns, 2010 and references therein). The OAE 2, or Bonarelli event (Bonarelli, 1891;Schlanger et al., 1987), is among the best studied examples to date. High rates of sea floor spreading and volcanic activityboth linked to the breakup of Gondwana with the formation of early ocean basins as carbon sinks and the evolution of ocean gatewaysled to increasing atmospheric carbon dioxide concentrations followed by a significant turnover of ocean chemistry and extinctions of marine biota (Brumsack, 1980;Arthur et al., 1985;Larson, 1991;Kaiho, 1994;Gale et al., 2000;McAnena et al., 2013). Black shales related to OAE 2 have been reported from many parts of the world including the proto southern North Atlantic and eastern Tethys Oceans (L€ uning et al., 2004). During OAE 2 the Tethys and Atlantic Oceans were affected by sluggish circulation of bottom waters caused by the shallow and partly closed connections to neighbouring oceans . Replenishment of oxygen was impeded within these partially isolated basins, favouring the development of anoxic and, possibly, euxinic conditions even in the photic zone (Herbin et al., 1986;Baudin, 1995;Sinninghe Damst e & K€ oster, 1998;Handoh et al., 1999). Large areas of the continental shelf and slope along the northern parts of Algeria, Libya, Tunisia and Morocco were deprived of oxygen, and enabled the rapid accumulation of shales and limestones rich in organic carbon linked to cyclic perturbations of redox-nutrient cycles and orbital forcing (Leine, 1986;El Albani et al., 1999a;Nederbragt & Fiorentino, 1999;Kolonic et al., 2005;Poulton et al., 2015; for Atlantic-wide teleconnections see Wagner et al., 2013 and references therein).
The Cretaceous has yielded several unusual carbonate deposits occurring before, during and after the OAEs. Some of these deposits have been ascribed to carbonate precipitation related to chemosynthesis-based microbial activity, and include examples from the Western Interior Seaway (Kauffman et al., 1996;Metz, 2010;Kiel et al., 2012) and the northern Tethyan and North African continental margins (Layeb et al., 2012(Layeb et al., , 2014. Limestones from the Western Interior Seaway feature shallow-water methane-seep deposits that formed during OAE 2 in warm bottom waters , and are characterized by 13 C-depleted carbonate mineral phases and a high-abundance but low-diversity faunal assemblage. The Campanian organic-rich Pierre Shales comprise numerous peculiar carbonate bodies, collectively known as the Tepee Buttes (Shapiro & Fricke, 2002;Metz, 2010). These deposits are rich in chemosymbiotic macrofaunal communities and exhibit diagnostic 13 C-depleted mineral fabrics and 13 C-depleted molecular fossils of the anaerobic oxidation of methane (AOM) consortium, confirming that carbonate precipitation was driven by microbial oxidation of methane (Kauffman et al., 1996;Birgel et al., 2006). The fossiliferous Tepee Butte limestones possibly formed during periods of enhanced bottom water oxygenation, supported by the presence of molecular fossils of aerobic methanotrophic bacteria (Birgel et al., 2006). Layeb et al. (2012) documented carbonate build-ups from the Lower Cretaceous Fahdene Basin in Tunisia. These limestones reveal a thrombolitic fabric and formed within organic-rich black shales deposited during the Late Albian OAE 1d. The fabric and texture of these limestones suggest the involvement of microbes during their formation below a water body with eutrophic, anoxic waters. Layeb et al. (2014) further reported on mound-shaped carbonate structures associated with OAE 2 within Late Cenomanian black shales from the Bahloul Formation in Northern Tunisia. The genesis of the Tunisian limestones possibly involved microorganisms mineralizing large amounts of organic matter, as suggested by 13 C-depleted carbonate cements and thrombolitic fabrics (Layeb et al., 2014).
This study reports on a large limestone body associated with black shales deposited in the aftermath of OAE 2 exposed along a coastal section of the Tarfaya Basin, Morocco. A combination of petrography, macro-and micropalaeontology, carbon and oxygen stable isotope and trace element geochemistry, as well as lipid biomarkers has been used to decipher the mode of carbonate accretion that led to the unusual morphology of this limestone body. The potential triggers for carbonate precipitation within the mid-Cretaceous organic-rich deposits, and the impact of the prevailing oceanographic and geochemical conditions on the benthonic environment and faunal communities are assessed. Although the Amma Fatma limestone body reveals only a limited number of lithological features typical of hydrocarbon seep carbonates and apparently no endemic seep fauna, it is shown to have resulted from methane seepage. The lack of obligate, endemic seep fauna is interpreted to have been caused by (1) shallow water deptha factor known to discriminate against seep taxa (Kiel, 2010) and (2) the eutrophic palaeoenvironment in the aftermath of OAE 2 that favoured photosynthetic primary production and subsequent consumption of organic matter by heterotrophs. This work reveals that seep deposits are difficult to identify in high productivity settings, where chemosynthesis-based seep fauna is likely to be outcompeted by heterotrophic background fauna.

GEOLOGICAL SETTING
Deposits from the Cenomanian-Turonian (C/T) boundary in Northern Africa have been intensively studied in Tunisia and Morocco. This is partly due to the fact that these regions feature up to 800 m thick successions of organicrich shales and limestones Kolonic et al., 2002), making them prolific petroleum source rocks (Leine, 1986) and high resolution climate archives (Kolonic et al., 2005;Kuhnt et al., 2005). The southern Moroccan passive continental margin includes the Tarfaya Basin, one of several palaeo-shelf basins located on the coasts of Senegal, Morocco and Tunisia (Figs 1 and 2A;L€ uning et al., 2004;Kolonic et al., 2005). It covers an area of around 170 000 km 2 and features the highest accumulation rates of organic matter from the C/T boundary known to date, as well as one of the largest oil shale deposits in the world (Ambl es et al., 1994;Kuhnt et al., 2001). The Tarfaya Basin is situated to the south of the Anti-Atlas Mountains and extends into the Tindouf Basin to the East, the Senegal Basin to the South, and is bordered by the East Canary Ridge to the West (Kolonic et al., 2002). It has been suggested that the basin was a distal section of the proto southern North Atlantic with outer shelf to upper bathyal water depths of 200 to 300 m (Figs 1 and 2B;El Albani et al., 1999b;Kuhnt et al., 2001). It experienced intense upwelling, extended oxygen minimum zones and recurrent pulses of high primary production since at least the Late Cenomanian (Kuhnt & Wiedmann, 1995;El Albani et al., 1999a,b;Kolonic et al., 2002Kolonic et al., , 2005, leading to pronounced redox cycles on orbital time scales that alternated between sulphidic and anoxic ferruginous water-column conditions (Poulton et al., 2015). The carbonate body studied here crops out at Amma Fatma Plage, which is a coastal to middle shelf section at the northern edge of the Tarfaya Basin (Fig. 2). Amma Fatma Plage is located ca 8 km to the south-west of Mohammed Plage, the latter being a well-studied coastal section of the basin where deposition of organic-rich shales and carbonate beds occurred close to the palaeocoastline (El Albani et al., 1999b;Kolonic et al., 2005). Both sites experienced a series of transgressional cycles during the mid to late Cretaceous that caused repeated and widespread flooding, and initiated the deposition of laminated biogenic sediments associated with OAE 2 (Kolonic et al., 2002(Kolonic et al., , 2005Gebhardt et al., 2004;Kuhnt et al., 2009;Gertsch et al., 2010). Amma Fatma Plage lies along a 175 km long palaeotransect through the Tarfaya Basin, together with shallow exploration wells S13, S57, S75 and Mohammed Plage (cf. Kolonic et al., 2002Kolonic et al., , 2005. The transect runs in SW to NE direction and connects its deepest and most distal part at wells S13 and S57 to its most proximal part at Mohammed Plage (Fig. 2B). Exploration well S13 is interpreted as an openmarine shelf setting with water depths of 200 to 300 m, whereas Mohammed Plage and Amma Fatma Plage represent shallower, proximal successions deposited at the northwestern margin of the Sahara platform (Kolonic et al., 2002(Kolonic et al., , 2005Kuhnt et al., 2009).

Petrography, mineralogy and palaeontology
Rock samples from the carbonate body and from a vertical profile at Amma Fatma Plage were collected during a field trip in 2003. Samples from the studied sedimentary succession were recovered from a fresh coastal exposure in direct proximity to the carbonate body, and taken at regular intervals of 5 to 10 cm across the profile. A total of 20 samples were collected from various zones of the carbonate body, two further samples were collected from a carbonate bank covering the structure (hereafter referred to as cover bed). Thin sections (150 9 100 mm, 100 9 75 mm) were made at the University of Bremen and the University of Vienna. Thin section microscopy was conducted using a Leica DM4500 P polarization microscope (Wetzlar, Germany). Photographs were taken with a Leica DFC 450 C camera using the Leica Application Suite 4.4.0 software. Additionally, a Nikon SMZ 1500 stereomicroscope coupled to a Prog.Res Speed XT core 5 camera, as well as a Nikon Optiphot-2 optical microscope (Tokyo, Japan) were used for petrographic observations using the software Prog.Res Capture pro 2.8 for image analysis and camera control. To discriminate the different carbonate minerals under the microscope, several thin sections were stained with a mixture of potassium ferricyanide and alizarin red S solution dissolved in 0Á1% hydrochloric acid (cf. Dickson, 1966). Powdered samples for X-ray diffraction (XRD) and carbon and oxygen stable isotope analyses were obtained from polished rock slabs using a hand-held microdrill at low to medium rotational speed. For XRD analysis, powdered samples were analysed with a Panalytical PW 3040/60 X'Pert PRO (CuKa radiation, 40 kV, 40 mA, step size 0Á0167Á5 s per step) diffractometer at the University of Vienna. X-ray diffraction patterns were interpreted using the Panalytical software "X 0 Pert High score plus". Calcareous nannoplankton  Kolonic et al., 2005). Yellow stars indicating the Amma Fatma section in (A), (B) and the small inset map in the upper left, which indicates the location of the studied section in Africa. (B) Palaeogeographic map of the Tarfaya Basin (modified from Kolonic et al., 2005); AF = Amma Fatma, MB = Mohammed Plage, S13, S57 and S75 = drill cores described by Leine (1986) and Kolonic et al. (2005). assemblages were studied in 150 samples with a highresolution spacing of about 4 to 7 cm within the slabbed intervals. For the calcareous nannoplankton, smear slides were prepared using techniques described in Bramlette & Sullivan (1961) and Hay (1961Hay ( , 1965. The slides were examined at a magnification of about 1250 times under both cross-polarized and phase-contrast illumination. Inorganic geochemistry -Total organic carbon, carbon and oxygen stable isotopes and trace elements Total carbon (TC) and total organic carbon (TOC) contents of the Amma Fatma carbonate body were determined using a LECO RC-612 Carbon Analyzer (St. Joseph, MI, USA), equipped with a solid-state infrared detector at the Department of Environmental Geosciences at the University of Vienna. All samples were ground to fine powder in an agate mortar prior to measurements. To measure TC, carbon dioxide was released at a temperature of 1000°C. Prior to sample measurements a pure CaCO 3 standard (Co. Merck, Kenilworth, NJ, USA) was measured, revealing a relative standard deviation of 0Á06%. The TOC contents were determined at 550°C. A Synthetic Carbon Leco 502-029 (1Á01 AE 0Á02 carbon%) standard was measured prior to sample analyses and a standard deviation of 0Á005% was calculated. Carbonate (CaCO 3 %) contents were calculated according to the equation: CaCO 3 % = (TC À TOC) 9 8Á333. All black shale samples were carefully dried and ground in an agate mortar. The TIC and TOC contents of black shales were measured using a LECO CS-300 carbon-sulphur analyser at the Geoscience Department, University of Bremen. Before TOC determination, inorganic carbon was carefully removed by repetitive addition of 0Á25 N HCl. The CaCO 3 contents were calculated as above (precision of measurements AE3%). The d 13 C TOC values (AE0Á1& vs. Vienna Pee Dee belemnite, V-PDB) were determined on decalcified sediment samples using automated on-line combustion followed by conventional isotope-ratio mass spectrometry.
A total of 48 sample powders from 12 different samples throughout the Amma Fatma carbonate body were analysed for their stable carbon and oxygen isotope content. Sample powders were obtained from polished rock slabs using a hand-held microdrill. Stable carbon and oxygen isotope analyses were performed at the light stable isotope laboratory of the Institute of Earth Sciences, Karl-Franzens University, Graz. Effective accuracy was AE0Á05 for d 13 C values and AE0Á11 for d 18 O values. The d 13 C and d 18 O values are reported relative to V-PDB (standard deviation smaller than 0Á04&), and appropriate correction factors were applied.
Trace metal compositions of black shales and limestones were determined using X-ray fluorescence (XRF) spectrometry on fused-borate glass beads, and by inductively coupled plasma atomic emission spectrometry (ICP-AES). The XRF calibration curves were based on international standard reference material. For ICP-AES, about 50 mg of sample powder was digested in a mixture of 3 ml HNO 3 (65%), 2 ml HF (40%) and 2 ml HCl (36%) of suprapure quality at 200°C and 30 kbar in closed Teflon vessels (Heinrichs et al., 1986). Sample powders were then dried by evaporation, and the residue was re-dissolved in 0Á5 ml HNO 3 (65%) and 4Á5 ml deionized water. The solutions were analysed with ICP-AES, and the results were checked with international standard reference material. Relative standard deviations in duplicate measurement were below 3%. Enrichment factors (EFs) of trace elements were calculated according to Brumsack (2006): EF element ¼ ðelement=AlÞ sample =ðelement=AlÞ average shale Average shale refers to the composition of collected shale samples compiled by Wedepohl (1969). An element with an EF either above or below 1 is considered to be enriched or depleted compared to average shale, respectively.

Organic geochemistry -Lipid biomarkers
Four carbonate rock samples were chosen for biomarker analysis, one from the cover bed and one from each of the zones 4, 3 and 1. The samples were cleaned repeatedly by washing with 10% hydrochloric acid (HCl) and then dichloromethane (DCM) to avoid any contamination, and then ground to fine powder. All carbonates were treated with a 'two-step extraction procedure' to discriminate preferentially intercrystalline-bound from preferentially intracrystalline-bound molecular fossils (cf. Thiel et al., 1999;Guido et al., 2013), and to separate compounds tightly associated with the carbonate matrix from easily extractable compounds. To obtain preferentially intercrystalline-bound compounds, the ground samples were extracted by ultrasonication in DCM : MeOH (3 : 1) nine times until the solvents became colourless to get the first total lipid extract (TLE). This extract is referred to as the extract before decalcification below. To obtain preferentially intracrystalline-bound molecular fossils, the residual sediment of the first extraction step was decalcified and then extracted again. In detail, doubly distilled water was added and 10% HCl was slowly poured onto the samples. To avoid transesterification reactions at low pH, HCl addition was stopped when ca 75% of the carbonate was dissolved. The remaining sample powder was centrifuged and extracted as described above to obtain a second TLE. This extract is referred to as the extract after decalcification below. From here on, the extracts before and the extract after decalcification are treated in the same fashion. After washing with clean water, the samples were saponified with 6% potassium hydroxide (KOH) in methanol (MeOH) to release bond carboxylic acids. The supernatants were decanted and the residues were subsequently extracted via ultrasonication in DCM/MeOH (3 : 1; v : v) four times until the solvents became colourless. The combined extracts were then treated with 10% HCl to transfer the free fatty acids (FAs) to the organic solvent phase. For gas chromatography (GC) analyses the extracts were dried with sodium sulphate and the TLE of both extraction procedures were separated by column chromatography. Alcohols were derivatized by adding 100 ll pyridine and 100 ll N,O-bis(trimethylsilyl) trifluoracetamide (BSTFA) to the alcohol fraction at 70°C for 30 min. The derivatized fraction was dried under a stream of nitrogen (N 2 ) and dissolved in n-hexane prior to injection. The alcohol fractions are not described below due to the lack of pristine compounds. Free carboxylic acids were reacted with 1 ml 14% boron trifluoride (BF 3 ) in MeOH at 70°C for 1 h to form FA methyl esters. After cooling, the mixture was extracted four times with 2 ml n-hexane. Combined extracts were evaporated under a stream of N 2 and redissolved in n-hexane before injection. The derivatized carboxylic acid fractions were analysed using GC-mass spectrometry (GC-MS) with an Agilent 7890 A GC system coupled to an Agilent 5975C inert MSD mass spectrometer. Quantification was done using GC-flame ionization detection (GC-FID) with an Agilent 7820 A GC system. Internal standards used were 5a-cholestane for the hydrocarbon and 2-Me-C 18 FA for the carboxylic acid fractions. Both GC systems were equipped with HP-5 MS UI fused silica columns (30 m 9 0Á25 mm i.d., 0Á25 lm film thickness). Helium was used as the carrier gas. The GC temperature program for hydrocarbons and carboxylic acids was 60°C (1 min) to 150°C at 10°C min À1 , then to 320°C (held 25 min) at 4°C min À1 . Compound assignment was based on retention times and published mass spectra.
Compound-specific stable carbon isotope analysis was performed using a Thermo Fisher Trace GC Ultra connected via a Thermo Fisher GC Isolink interface to a Thermo Fisher Delta V Advantage spectrometer at the Department of Terrestrial Ecosystem Research, University of Vienna. Compound-specific carbon isotope values are given as d values in per mil (&) relative to V-PDB. The d 13 C values of carboxylic acids (methyl esters) were corrected for additional carbon introduced after derivatization. Each measurement was calibrated using several pulses of carbon dioxide (CO 2 ) of known isotopic composition prior to and after the run. Precision was checked with an n-alkane mixture (C 14 to C 38 ) of known isotopic composition. Analytical standard deviation was <0Á7&.

RESULTS
Amma Fatma section -Lithology, biostratigraphy and inorganic geochemistry The complete section at Amma Fatma Plage measures ca 14 m from top to bottom and features cyclic alternations of black shales, siltstones and organic-rich, marly limestone beds (Figs 3 and 4). The limestones are represented by (1) carbonate nodules mainly composed of mudstones and wackestones that are aligned parallel to stratification and (2) discontinuous bioclastic storm beds with hummocky cross-stratification (cf. El Albani et al., 1999b). The limestones contain Foraminifera, calcispheres, molluscs including ammonites, brachiopods, variable amounts of quartz grains and are partly bioturbated. A detailed microfacies analysis is included in El Albani et al. (1997, 1999a). The succession of 22 beds is shown by the stratigraphic chart of the outcrop in Figs 3 and 5. Limestone and black shale beds alternate regularly, whereby their thicknesses measure 20 to 80 cm and 60 to 300 cm, respectively. Beds comprising limestone nodules with chert measure up to 20 cm in thickness. One exception is a 1 m thick bed 3 m above the base of the profile featuring large carbonate nodules with a corona made of chert. Calcium carbonate and TOC contents of the black shale beds vary from 60 to 90% and 4 to 9%, respectively. Calcium carbonate and TOC values throughout the section are shown in Table 1. The characteristic positive d 13 C TOC excursion that is prominently observed across the C/T boundary in adjacent sites is not found at Amma Fatma Plage, supporting a post OAE period of deposition. The measured d 13 C TOC values vary slightly between À26Á2 and À27Á8&.
Biostratigraphic data confirm the positioning relative to the OAE 2 carbon isotope excursion, placing the investigated sedimentary succession within the H. helvetica planktonic foraminiferal zone and the coloradoense and nodosoides ammonite zone of the upper part of the Lower Turonian stage (cf. Marzouk & L€ uning, 2005). The position of the nannofossil zonal datum marker CC11/CC12 is placed at about 9Á5 m above the helvetica and nodosoides zones. Planktonic Foraminifera are present in very high numbers and benthonic foraminiferal faunas are of low diversity. A highly diverse and highly abundant calcareous nannofossil flora was observed in all samples studied. Preservation is moderate, except in the uppermost part of the section (from 13Á5 m onwards), where preservation is good. Watznaueria barnesae and Eiffelithus turriseffeilii are abundant throughout the section (except in two carbonate beds), and are known to be very resistant to solution. Although frequent, the abundance of Watznaueria barnesae does not surpass 40%, a cut-off value proposed by Roth & Krumbach (1986) to indicate that nannofossil distribution patterns can still largely be used for palaeoecological interpretations, despite diagenetic and syndepositional dissolution. Notably, Cyclagelosphaera margerelii shows a similarly abundant distribution and, therefore, appears to be also solution resistant.
The nannofossil distribution in the Amma Fatma section is characterized by many species having longer term, gradual changes, with only some species exhibiting stronger, high-frequency abundance fluctuations over a few centimetres (Tranolithus orionatus, Zeugrhabdus diplogrammus). Abundant species throughout the section include Zygodiscus erectus, Rhagodiscus splendens and Eprolithus floralis. R. splendens occurs in moderate and high abundances in the lowermost and uppermost parts of the studied section, and is absent in the interval in-between. Notably, the distribution of E. floralis in the section is characterized by similar abundance peaks at the base and top of the section. Other identified species Broisonia spp. (except B. enormis) and Gartnerago spp. (Thierstein, 1976;Hattner et al., 1980) are notably rare throughout the section. An exception is a Gartnerago segmentatum abundance peak between 7Á5 and 8 m. Abundance and distribution of all identified nannofossil species and phosphorus content throughout the Amma Fatma section are shown in Fig. 5.
Total contents of all measured elements are shown in Table 1. Prominent trace metal enrichment is restricted to stratigraphic intervals that correspond to the organic carbon-rich black shales. Trace metal patterns vary across the section between alternating black shales and limestone beds, and also within the black shale successions themselves. Calculated EFs of trace metals whose deposition is influenced by bottom and pore water redox conditions are plotted with depth in Fig. 6. Mean trace metal concentrations and EFs of black shale and limestone beds that correspond to zones 1 to 4 and the cover bed are shown in Tables 2 and 3, respectively. Zone 1 shows the lowest concentration of several major elements including Al, Mg, Fe and Mn (Tables 1 and 2). Particularly, zone 1 has an Al content that is half that of zone 4, which has a considerable effect on calculated EFs. Zone 4 reveals relatively high Al contents, and all EFs are higher in the lower zones 1 and 2 compared to zone 4 ( Table 3). As EFs may be exaggerated due to varying Al content, they must be considered together with absolute contents of the respective elements. This is particularly the case for bed 14, which consists mainly of marly limestone and exhibits low Al concentrations. Chromium, Co and U are the only trace metals whose concentrations increase from zone 1 to 4 (Table 2), whereas Mo concentrations in zone 1 and 4 are almost equal. There is a distinct peak of all trace metals except Cr and Co in zone 3, which is not as apparent in respective EFs due to high Al contents in this zone. Bed 14 exhibits the lowest contents of most trace metals with the exception of U. Molybdenum/U ratios decrease steadily from bottom to top and fall below unity in bed 14. Correlation coefficients between trace metals and Al, as well as between trace metals themselves were calculated and are shown in Table 4. The only trace metals that exhibit correlation with Al are Cr and Co, with the latter correlated better than the former. Trace metal correlations can be distinguished into three groups of elements that correlate well among each other, these being (1) Cr with Cu, (2) Mo with Ni, Zn and V and (3) Zn with Ni and Cd (Table 4).

The Amma Fatma carbonate body
The carbonate body associated with black shales (Figs 3 and 4) is enclosed by beds 6 to 14 of the Amma Fatma section, and measures 5Á5 m in height and 3Á6 m in width. The carbonate body was divided into 4 zones, illustrated in a schematic sketch (Fig. 4B). Zones 1 and 2 comprise the base and lowermost part of the structure and feature concretionary and nodular carbonates penetrated by numerous veins made of secondary precipitates (Fig. 7). A carbonate bank measuring 17 to 19 cm in thickness separates zone 1 from 2 ( Fig. 7B). Zone 3 comprises the middle part of the body, and features massive limestone blocks and nodules. Zone 4 is the topmost part of the carbonate body and yielded abundant molluscan fossils, as well as cylindrical to tubular trace fossils (Fig. 8C). The top of the body is capped by a 15 to 20 cm thick carbonate bank, referred to as a cover bed (Figs 4 and 8A to C), which corresponds to bed 14 of the section (Figs 3 and 4A).

The Amma Fatma carbonate body -Petrography and mineralogy
Micrite (i.e. microcrystalline calcite with a crystal size below 10 lm) is the dominant primary mineral that   makes up the matrix of the Amma Fatma carbonate body. Matrix micrite and peloidal micrite occur, whereby the former is the most abundant carbonate phase in zones 1 and 2, making up at least 80% of the bulk rock volume (Figs 9 and 10A). Matrix micrite is light to dark brown in colour, exhibits a very heterogeneous and partly clotted texture and is rich in detrital material comprising large amounts of recrystallized bivalve and gastropod shell fragments (Fig. 9A), benthonic Foraminifera (Fig. 9B), Palaxius faecal pellets (Fig. 10B), radiolarians (Fig. 11A) and    (Figs 10C and 11B). This micrite is characterized by abundant oval to spherically shaped peloids, which measure 100 to 500 lm in diameter. The peloids do not show any obvious internal structure and exhibit a cloudy and heterogeneous fabric. The space between the peloids is occupied by fine, dispersed equant calcite cement. Similar to the matrix micrite, the peloidal micrite exhibits scattered occurrences of Palaxius faecal pellets, gastropod, bivalve and ammonite shell fragments, as well as Foraminifera and radiolarian tests. The peloids and Palaxius faecal pellets are fluorescent under UV light, suggesting that they are enriched in organic matter compared to the surrounding micrite. Palaxius faecal pellets are present in all zones but are particularly common in zones 1 and 2 (Fig. 10B). The crescent-shaped canals are arranged in two groups around a symmetry plane, and are oriented in different directions towards and away from the symmetry plane. They are oval to spherical in shape and measure up to 500 lm in diameter. Their internal structure clearly distinguishes them from the aforementioned micritic peloids. Pyrite occurs as very small and finely dispersed grains measuring only several microns in size. In some rare cases pyrite also forms aggregates and clusters of idiomorphic crystals (Fig. 9B). Dolomite occurs within zones 1 and 2. The dolomite crystals are pale grey in colour and show well-defined crystal faces (Figs 9B and 10B). The Amma Fatma limestone is penetrated by numerous veins, cracks and large cavities throughout zones 1 to 4. Cavities and veins show the same paragenetic mineral sequence in all zones: Matrix micrite or peloidal micrite represents the primary fabric of the rock. Cavities and veins are rimmed by microquartz, followed by equant calcite spar and megaquartz (Figs 10C, D, 11B, C, 12A and B). Microquartz (sensu Knauth, 1994) is pale brown to beige in colour (Figs 10C and 11B,C), and typified by very small grain sizes reaching no more than several microns in diameter. In places it can also be found finely dispersed among the equant calcite crystals. Microquartz occasionally forms crystal fans showing oscillatory extinction patterns under crossed polarized light. Volumetrically, equant calcite is the dominant void filling phase in all four zones (Figs 10A, 11C, 12A and B); its crystals measure up to 1 mm in diameter. The last phase in the paragenetic phase sequence is megaquartz (sensu Knauth, 1994). Megaquartz crystals reach up to 1 mm or more in diameter ( Fig. 10C and D). An accessory peculiar carbonate phase present within fractures and voids is an authigenic rim of microcrystalline calcite referred to as seam micrite (Fig. 12). It is present in all zones and occurs exclusively along the outer rims of veins and cavities,   where it forms highly irregular patches and fringes ( Fig. 12A and B). Seam micrite is dark brown, almost black in colour and is very homogenous, showing no detrital inclusions. It shows strong epifluorescence when excited with UV light (Fig. 12C). The carbonate cover bed (Fig. 4) shows a quite similar paragenetic phase sequence to zones 1 to 4 of the carbonate body (Fig. 4). The matrix comprises micrite, which, compared to the matrix micrite of zones 1 to 4, contains less detrital material. It is relatively homogeneous showing no clotted or peloidal textures and contains minor dolomite crystals (Fig. 13). The fossil content is low, represented by mainly radiolarian and foraminiferan tests (Fig. 13A). Similar to the carbonate body, the cover bed is penetrated by numerous cracks and veins, which have been filled by secondary precipitates. Cavities contain microquartz and later formed equant calcite spar ( Fig. 13B and C).

Macrofauna
The macrofossils collected from Zone 4 of the carbonate body comprise a low diversity molluscan assemblage of gastropods, bivalves and ammonites, together with Palaxius faecal pellets and burrows. The gastropods comprise 23 specimens of the aporrhaid Drepanocheilus sp. (Fig. 14A and B), mostly juveniles, but also a few adults up to 21 mm high, and a single, partially fragmented specimen that might represent a neogastropod belonging to the genera Drilluta, Bellifusus or Paleopsephaea (Fig. 14C). The bivalves belong to three taxa: a single articulated lucinid specimen (Fig. 14E), juvenile  inoceramids (n = 4; Fig. 14D) and gryphaeid oysters (n = 6), which could belong to the genus Pycnodonte. The small specimen sizes (<10 mm) of the two latter taxa and the articulation of the lucinid precludes further identification. The ammonites (n = 5) are juveniles and adults belonging to the genus Benueites, including Benueites cf. benueensis.

Isotope geochemistry, TC and TOC
Results of carbon and oxygen stable isotope analyses are shown in Fig. 15. Matrix micrite is 13 C-depleted to different degrees throughout all zones of the carbonate body, including the cover bed. It revealed the lowest d 13 C values  in zone 1, ranging from À23Á5 to À21Á3&. Zone 2 micrite yielded a wider range of values from À23Á1 to À7Á9&, and zone 3 micrite more positive values between À11Á0 to À0Á3&. Zone 4 micrite shows the largest scatter of d 13 C values, falling between À22Á9 and À1Á3&. The matrix micrite of the cover bed yielded d 13 C values from À2Á3 to À0Á4&. The d 13 C values of equant calcite were analysed from zones 3, 4 and the cover bed. Generally, equant calcite is less 13 C-depleted than matrix micrite. A sample from zone 3 revealed the only positive d 13 C value of 1Á6&. Samples from zone 4 and the cover bed show little heterogeneity and fall between À3Á7 to À1Á5& and À0Á9 to À0Á7&, respectively.
The d 18 O values of matrix micrite and equant calcite from all sections of the carbonate body are exclusively negative. Matrix micrites from zones 1, 2 and 3 exhibit values from À2Á9 to À2Á3&, À3Á2 to À2Á1& and À3Á4 to À1Á8&, respectively. Zone 4 matrix micrite displays more scattered d 18 O values from À4Á3 to À1Á7&. Matrix micrite of the cover bed yielded values from À3Á3 to À2Á6&. The d 18 O value of equant calcite from zone 3 and 2 is À3Á4&. Equant calcite from zone 4 shows the most negative d 18 O values from À6Á2 to À2Á6. Equant calcite of the cover bed is less 18 O depleted (À2Á3 and À1Á7&).
A total of six samples from the Amma Fatma carbonate body were analysed for their TC and TOC contents ( Table 5). The TOC content increases significantly from top to bottom, from 0Á1% in zone 4 to 0Á7% in zone 1. The cover bed exhibits a similar TOC content as the topmost zone 4 of 0Á1%.

Hydrocarbons
Organic matter enclosed in the sediments of the Tarfaya Basin has been shown to be of low maturity (Kolonic et al., 2002;Sachse et al., 2012). The hydrocarbon fraction before decalcification from the cover bed is characterized by short and long chain n-alkanes that range from C 16 to C 34 . The shorter-chain C 16 to C 20 n-alkanes dominate over longer chain n-alkanes, whereby n-alkanes peak at n-C 20 . Apart from n-alkanes, the only compounds detected are minor amounts of the regular isoprenoids pristane and phytane, with phytane more abundant than pristane. The hydrocarbon fraction after decalcification from the cover bed shows a very similar pattern, except that the contents are much reduced compared to the fraction before decalcification.
Hydrocarbons in the extract before decalcification from zone 1 are characterized by the dominance of n-alkanes from C 16 to C 30 . In contrast with the cover bed, C 16 and C 18 dominate over C 20 . Pristane and phytane are the only non-alkane compounds detected in zone 1, with phytane slightly more abundant than pristane. The hydrocarbon fraction after decalcification is similar to the fraction before decalcification except that the overall intensities of all compounds are greatly reduced. Very similar results were obtained for two carbonate samples of zones 3 and 4, and are not further detailed here.

Carboxylic acids
Carboxylic acid fractions before and after decalcification from the cover bed and zone 1 limestone are shown in Fig. 16. The carboxylic acid fraction from the cover bed is dominated by saturated short chain n-fatty acids ranging from C 12 to C 18 . Most abundant n-fatty acids are C 16 and C 18 , which dominate over C 12 . Unsaturated n-fatty acids and long-chained saturated n-fatty acids are present in very small amounts (Fig. 16A). The carboxylic acid fraction after decalcification varies only slightly from the fraction before decalcification, with medium-chain n-fatty acids up to C 23 more abundant in the former (Fig. 16B). The contents of short chained n-fatty acids are markedly decreased (ca 50% of all carboxylic acids) compared to the fraction before decalcification (Table 6), however, the distribution of C 12 , C 16 and C 18 is similar.
The total amount of carboxylic acids varies strongly between the samples. For the cover bed the amount of acids in the extract after decalcification is half that of the fraction before decalcification. For the samples from zones 4, 3 and 1 the ratio is reversed, where the carboxylic acids were found to be more tightly bound to the carbonate lattice. The amount of acids increases from zone 4 to zone 1. Contents and phase-specific stable carbon isotope values of the carboxylic acids are shown in Table 6.

INTERPRETATION AND DISCUSSION
The Amma Fatma carbonate body -A hydrocarbon seep deposit The comprehensive data set obtained for the Amma Fatma carbonate body allows constraining of the geological setting and the processes that triggered its formation. Although this deposit is unlike many hydrocarbon-seep deposits in terms of its faunal content (cf. Campbell, 2006), there is multiple evidence that it formed at a shallow-water seep in a high productivity realm. The strongest indication for methane seepage is the presence of 13 C-depleted biphytanic diacids, which are molecular fossils of anaerobic methane oxidizing archaea (ANMEs; Birgel et al., 2008a). These compounds are putative diagenetic breakdown products of glycerol dibiphytanyl glycerol tetraethers (GDGTs; Liu et al., 2016) and are membrane lipids of archaea including ANMEs (see Schouten et al., 2013 for a review). Biphytanic diacids with extreme 13 C-depletions have been reported from various methane-rich environments (Schouten et al., 2003;Fig. 15. Carbon and oxygen stable isotope cross-plot of carbonate phases from zones 1 to 4 (z1 to z4) and the cover bed (cb). All values are relative to V-PDB.   Birgel et al., 2008a;De Boever et al., 2009;Natalicchio et al., 2012). GDGTs and biphytanic diacids are uncommon in Mesozoic and Palaeozoic seep limestones, and the more persistent irregular isoprenoids crocetane and pentamethylicosane (PMI) have been used to track methane oxidation in the rock record (Birgel et al., 2008b), but neither of these were found in the Amma Fatma deposit. The most abundant biphytanic diacid in the Amma Fatma limestone is BP-2. Values as low as À96& for BP-2 in the Amma Fatma deposit have only been reported from seep-dwelling ANMEs. A similar predominance of 13 C-depleted BP-2 among the biphytanic diacids was also reported for other Cenozoic and modern methane-seep carbonates (Birgel et al., 2008a). In contrast with BP-2, BP-3 is derived from other sources, as indicated by its d 13 C value similar to that of crenarchaeol-derived tricyclic biphytanes of planktonic Thaumarchaeota (K€ onneke et al., 2012;Schouten et al., 2013;Pearson et al., 2016). Interestingly, biphytanic diacids are only present in the lowermost zone 1, suggesting a heterogeneous composition of the Amma Fatma carbonate body. In zones 2 to 4, however, biphytanic diacids and other molecular fossils of hydrocarbon seepage such as isoprenoid hydrocarbons are possibly obscured by hydrocarbons and unresolvable complex mixtures, similar to some modern seep carbonates, particularly oil-seep carbonates (Naehr et al., 2009;Feng et al., 2014). Since biphytanic diacids were only detected within the carboxylic acids fraction after decalcification, these compounds are tightly bound to the carbonate lattice and consequently reflect environmental conditions during authigenesis, consistent with anaerobic methanotrophy contributing to carbonate formation.
Many ancient seep limestones are typified by a distinct paragenetic sequence of mineral phases that enables the reconstruction of processes that governed their formation (Beauchamp & Savard, 1992;Campbell et al., 2002;Peckmann et al., 2002;Hagemann et al., 2013;Zwicker et al., 2015). Microcrystalline calcite (i.e. micrite) forms the matrix of most seep limestones, and several types can be distinguished including clotted (Peckmann et al., 2002), peloidal (Roberts & Aharon, 1994) and detritus-rich matrix micrite Tong et al., 2016). Micrite of the Amma Fatma deposit exhibit a wide range of d 13 C values, a pattern commonly found for seep limestones (cf. Greinert et al., 2001). The composition of the carbon pool from which carbonates precipitate at seeps varies over time and depends on the degree of mixing between various carbon sources (Suess & Whiticar, 1989;Peckmann & Thiel, 2004). Biogenic and thermogenic methane, higher hydrocarbon compounds including crude oil and dissolved inorganic carbon from sea water all contribute to the carbon pool of the parent fluid, and determining the contribution of each source to carbonate precipitation at ancient seeps thus remains problematic. Sulphate-dependent AOM is considered to be the main process triggering carbonate precipitation at seeps, whereby the carbonates inherit the 13 C-depletion of the organic carbon sources (Ritger et al., 1987;Peckmann & Thiel, 2004;Campbell, 2006). Regardless of the carbon source controlling the isotopic composition of the carbon pool, the carbonates that precipitate at seeps are usually more 13 C-enriched than the organic carbon sources due to an admixture of marine carbonate (Peckmann & Thiel, 2004).
The lowest d 13 C value of the Amma Fatma matrix micrite of À23Á5& therefore suggests that AOM did not necessarily contribute substantially to carbonate precipitation, although 13 C-depleted biphytanic diacids testify to the presence of microbial communities performing AOM. Slightly lower d 13 C values were obtained from seep deposits associated with Jurassic black shales from Kilve, UK, and were attributed to biogenic methane production in the host sediments and its oxidation close to the sea floor (Xu et al., 2016). The oxidation of higher hydrocarbon compounds may have contributed to carbonate formation at Amma Fatma. This notion, however, is rather speculative as evidence for the presence of oil at the ancient seep is missing, and it is only supported by d 13 C values that are within the range of carbonates precipitating at modern oil seeps (cf. Anderson et al., 1983;Sassen et al., 1999;Formolo et al., 2004). A new approach using trace and rare earth metal contents may shed light on the composition of fluids of the Amma Fatma seep (cf. Smrzka et al., 2016), and future work may assess the impact of crude oil at ancient seeps with more confidence. Finally, the Amma Fatma black shales themselves might have acted as a readily available source of organic carbon, causing microbially remineralization-driven carbonate formation. The d 13 C TOC values of the Amma Fatma succession are slightly lower than the d 13 C values of the authigenic carbonates. Remineralization of large amounts of organic carbon buried during OAE 2 may have consequently contributed to carbonate precipitation, but it does not explain the presence of 13 C-depleted biphytanic diacids.
Seam micrite is a peculiar carbonate phase of the Amma Fatma deposit. Peckmann et al. (2003) described a similar phase in Eocene seep deposits from Washington State that also lines cavity walls and predates later-stage calcite cements. The purity and microfabric of the Amma Fatma seam micrite are indicative of an in situ precipitation, and its intense epifluorescence (Fig. 12C) distinguishes it from matrix and peloidal micrite. The latter observation suggests an increased content of organic residue, agreeing with an involvement of microbial activity in its formation (cf. Peckmann & Thiel, 2004). Equant calcite is a typical, later-stage mineral phase of seeps limestones (Campbell et al., 2002;Peckmann et al., 2003), and is interpreted as an abiotically formed cement that precipitated during increasing burial and progressive diagenesis, which agrees with the only moderate 13 C-depletion of the Amma Fatma equant calcite.
Authigenic silica minerals typify many Mesozoic and Cenozoic seep limestones (Goedert et al., 2000;Himmler et al., 2008;Peckmann et al., 2011;Kuechler et al., 2012;Kiel et al., 2013;Smrzka et al., 2015). Most of the dissolved silica required for quartz precipitation in seep and non-seep environments is derived from the dissolution of radiolarian tests, diatom frustules and sponge spicules (Lancelot, 1973;Von Rad et al., 1977;Bohrmann et al., 1994;DeMaster, 2002;Kuechler et al., 2012), although dissolution of volcanic glass and clay minerals may also become relevant during late burial diagenesis (Keene, 1975). The abundance of radiolarian tests within the Amma Fatma carbonate body (Fig. 11A) suggests a biogenic source of silica. The sequence of microquartz occurring as thin linings along crack and cavity walls (Figs 11B, C and 12) and later megaquartz occluding large cavities (Fig. 10C and D) is typical of seep limestones . The presence of silica polymorphs suggests that changing silicic acid concentrations controlled the mineralogy of the precipitate over time (Heany, 1993), whereby microquartz precipitated from fluids with higher concentrations of silicic acid and megaquartz formed during a later stage when silica concentrations had decreased.
Although the Amma Fatma deposit features mineral phases that are common in ancient and modern seep limestones, it lacks a carbonate phase that typifies many seep deposits, namely banded and botryoidal aggregates of fibrous aragonite cement. This type of cement typically exhibits low d 13 C values, reflecting its origin from AOM (Beauchamp & Savard, 1992;Roberts et al., 1993;Savard et al., 1996). The lack of this phase together with the dominance of micrite suggests that the Amma Fatma seep was characterized by diffusive rather than advective seepage of hydrocarbon-rich fluids (cf. Peckmann et al., 2009).

Benthonic fauna
The benthonic macrofossil assemblage of the top part of the Amma Fatma carbonate body includes aporrhaid (Drepanocheilus sp.) and possible neogastropod gastropods, lucinid, inoceramid and gryphaeid (Pycnodonte sp.) bivalves. The small inoceramids and gryphaeids were epifaunal filter feeders, and the lucinid was an infaunal taxon that almost certainly had sulphide-oxidizing chemosymbionts, by comparison with modern members Table 6. Contents and compound-specific stable carbon isotope values of carboxylic acids. Summed up values are in ng g À1 rock, grouped lipid contents are in relative percentages of all carboxylic acids, of this family (Dando et al., 1986). The aporrhaid was an epifaunal/shallow infaunal suspension or deposit feeder. The possible neogastropod was most likely epifaunal, but without further taxonomic information its diet cannot be inferred. The presence of this fauna suggests that oxygenated conditions were at least episodically present in the water column and sediment during the formation of at least the upper part of the carbonate body, although the absence of large inoceramid and oyster specimens may hint at challenging environmental conditions. A taxonomically very similar low diversity fauna is present in shell-rich micritic carbonate concretions from below the carbonate body in bed 2 at Amma Fatma Plage, collected by Andrew Gale in 2003 and now housed in the collections of the Oxford University Museum, UK. This includes the same species of lucinid ( Fig. 14F and G) and Drepanocheilus sp. (Fig. 14I), but also much larger inoceramids (up to 35 mm long, Fig. 14H). Furthermore, species of Pycnodonte, inoceramids, lucinids and aporrhaid gastropods are common elements in Late Cretaceous faunas from across North Africa (Busson et al., 1999;El Qot, 2006;Ayoub-Hannaa et al., 2014). This indicates that the Amma Fatma seep fauna contains mostly elements of normal marine background faunas, and lacks specialized obligate taxa, possibly apart from the lucinid, which might belong to the seep obligate genus Nymphalucina (Kiel, 2013), although is also similar to the lucind figured in El Qot (2006, plate 14, figs 7 to 10) as Lucina fallax Forbes, 1846 from the Late Cenomanian of Egypt. This paucity of obligate seep taxa is a common feature of shallow water seep faunas in the Mesozoic and Cenozoic (Kiel, 2010;Little et al., 2015). Other examples include the slightly older Cenomanian Tropic Shale seep faunas from southern Utah, USA , which are dominated by mostly small specimens of the bivalve genera Inoceramus and Nymphalucina, the gastropod genus Drepanochilus, rarer examples of the neogastropod Paleopsephaea and Callianassa decapod claws. A younger example is the Paleocene Panoche Hills seep deposit from California, USA (Schwartz et al., 2003), which contains numerous specimens of Nymphalucina and less common aporrhaid gastropods (Kiel, 2013).
The Palaxius faecal pellets found throughout the Amma Fatma carbonate body are microcoprolites, which are today produced by callianassid decapod crustaceans. Palaxius microcoprolites are common in Palaeozoic and Mesozoic shallow water limestones Schweigert et al., 1997) and other methaneseep deposits Senowbari-Daryan et al., 2007;Zwicker et al., 2015). The number and shape of the crescent shaped canals in the Amma Fatma microcoprolites strongly resemble Palaxius decemlunulatus, which has previously been found in Cretaceous shallow water carbonate deposits of northern Africa (Senowbari-Daryan & Kuss, 1992). The high density of Palaxius microcoprolites in the Amma Fatma body suggests high productivity and organic matter availability, sustaining dense populations of crustaceans.

Mode of formation of the Amma Fatma carbonate deposit
The Amma Fatma carbonate body is not older than Lower Turonian in age. The presence of ammonites belonging to the genus Benueites in zone 4 shows that the upper part of the seep deposit belongs to the nodosoides Zone, the youngest zone of the Lower Turonian, as elsewhere in the Tarfaya Basin. Benueites species, including benueensis, are associated with Mammites nodosoides (Collignon, 1966). El Albani et al. (2001) described the laminated marls at Amma Fatma and deduced an early Turonian age based on the ammonite fauna and planktonic Foraminifera. The macrofaunal assemblage of the upper part of the Amma Fatma deposit suggests that the entire water column was at least episodically oxygenated at that time. The absence of the diagnostic d 13 C isotope excursion observed for OAEs confirms that the black shales and the carbonate body formed in the aftermath of OAE 2, and is consistent with biostratigraphic evidence.
The Amma Fatma deposit is a tall, concretionary carbonate body, and in this respect resembles some of the exhumed Oxfordian seep deposits from Beauvoisin  and Miocene deposits of Italy (Conti et al., 2004). Concretions of various shapes and sizes are common in Mesozoic strata, and typically stand out from the host sediment after erosion (Coleman, 1993;Coleman & Raiswell, 1995;Sell es-Mart ınez, 1996;Seilacher, 2001;Marshall & Pirrie, 2013;Liang et al., 2016). What is most indicative of concretionary growth is that the Amma Fatma carbonate body clearly stands out from the host successions. Apart from this large carbonate deposit, laminated strata from Amma Fatma and Mohammed Plage feature isolated, nodular, diagenetic concretions that precipitated within the organic carbon-rich limestones and black shales (El Albani et al., 1997, 1999a. El  described lower Turonian, elongated carbonate concretions in direct proximity to the Amma Fatma carbonate body. These concretions formed during early diagenesis before significant compaction occurred, as fossils were well-preserved and surrounding beds of the host marls wrap above and below the outside of the concretions. The Amma Fatma deposit crosses almost the entire section of laminated shales and limestones, making it difficult to ascertain if the carbonates precipitated during or after their deposition. The cover bed capping the Amma Fatma deposit sheds light on the mode of accretion. It thins out around the top of the carbonate body ( Fig. 8A and B), which agrees with the presence of a relief on the sea floor during the time of deposition. Such a scenario is in contrast with a thickening of a cover bed as seen during pockmark formation (cf. Agirrezabala et al., 2013). The circumstance that the cover bed is much thinner directly on top of the seep deposit, and thickens readily away from it, indicates that the deposit was elevated above the sea floor. Given the changing environmental conditions, the carbonate body may have projected into the water column during episodes of sulphidic bottom waters as observed today in the Black Sea (Peckmann et al., 2001). However, there is no evidence for an ancient hardground or firmground, as cementing epifauna are absent. It is conceivable that concretionary growth accompanied by compaction of the host sediment were the main modes of accretion of the Amma Fatma carbonate body, and that carbonate precipitation occurred in the subsurface but close to the sea floor. Large concretionary bodies require a constant and large supply of pore water that transports the necessary solutes to the locus of precipitation (Raiswell & Fischer, 2004;Mozley & Davis, 2005). Diffusive and advective transport of solutes by seeping fluids is the most likely mechanism that sustained concretionary growth over long timescales that subsequently fostered compaction and cementation of the Turonian black shales and marly limestones.
Mean d 18 O values of the predominantly authigenic matrix micrite ranging between À2Á6 and À2Á9& also point towards precipitation close to the sea floor, as they are within the range of values for limestones that precipitated in equilibrium with Cretaceous sea water at Mohammed Plage (cf. Kuhnt et al., 2009). They show no appreciable deviation throughout the body, suggesting that pore waters did not experience large temperature shifts or meteoric water input that would have produced more negative d 18 O values (Hudson, 1978;Astin & Scotchman, 1988). Such fairly homogeneous oxygen isotope compositions are common for concretions that grew in shallow sediments close to the sea floor (Hudson & Friedman, 1974;Raiswell & Fisher, 2000). The d 18 O values of the later-stage equant calcite do not differ substantially from those of the matrix micrite, suggesting that neither phase experienced much elevated temperatures during burial diagenesis (cf. Campbell et al., 2002).
In contrast, the trend towards lower d 13 C values from the topmost zone 4 to the lowermost zone 1 (Fig. 15) is interesting. The bottom zone 1 not only shows the highest 13 C depletion, but also the sole occurrence of biphytanic diacids and the highest TOC content. Explaining these patterns is challenging, as the preferential growth direction of the Amma Fatma deposit is not obvious. Neither pervasive growth, outward, downward, nor upward growth can be excluded, and assumptions on the mode of accretion remain speculative. Seep carbonate formation is fuelled by a source of reduced, hydrocarbonrich fluids from below. Zone 1 was possibly affected most by these fluids, which drove microbial activity as evidenced by the occurrence of 13 C-depleted biphytanic diacids. Fluid seepage apparently continued during ongoing sedimentation of organic material, and zone 1 remained the hotspot of hydrocarbon-driven authigenesis that clogged up the pore space and reduced permeability. The impact of methane-bearing fluids on carbonate precipitation in the upper zones 3 and 4 was apparently lower, resulting in gradually less negative d 13 C carbonate values as the dominant carbon source changed from isotopically lighter, seepage-derived hydrocarbons to heavier organic matter derived from the host black shales associated with OAE 2.
Methane seepage and carbonate precipitation in the aftermath of OAE 2 The depositional environment of Amma Fatma Plage was most likely an outer shelf setting, as suggested by the foraminiferal assemblages and smectite-dominated clay mineral assemblages (Holbourn et al., 1999) as well as the macrofossil content of the seep deposit (see above). The three palaeoecologically significant Foraminifera species identified at Amma Fatma are Zygodiscus (=Zeugrhabdotus) erectus, Rhagodiscus splendens and Eprolithus floralis. Generally, temperature and trophic resources are considered the most important factors affecting the distribution of calcareous nannoplankton, whereby ecological preferences of some species are already partly known (Eshet & Almogi-Labin, 1996;Mattioli, 1997;L€ uning et al., 1998;Herrle, 2003). The foraminiferal assemblages in combination with the occurrence of radiolarians and diatoms suggest high surface water productivity (El Albani et al., 1999b;Holbourn et al., 1999). Specifically, Zygodiscus (=Zeugrhabdotus) erectus is interpreted as an indicator for high productivity (Roth, 1981;Roth & Krumbach, 1986;Erba et al., 1992;Herrle, 2003). This species is abundant throughout the Amma Fatma section and lacks any systematic distributional changes, suggesting persistent high productivity conditions throughout the entire section consistent with overall high carbon burial. Another species identified as high productivity-prone is Rhagodiscus splendens. This may indicate further intensification of upwelling conditions, possibly marking high productivity phases at the base and the top of the section within a generally highly productive setting. An additional argument for upwelling conditions is the presence of Eprolithus floralis, which is also considered to be an indicator for cooler waters . The cooccurrence of these three species suggests that intensified upwelling of relatively cool, intermediate water occurred during the lower Turonian along the Tarfaya shelf. This interpretation is supported by pronounced peaks in phosphorus content that coincide with higher abundances of Eprolithus floralis and Rhagodiscus splendens at the bottom and top parts of the Amma Fatma section (Fig. 5). These observations, together with the characterization of the macrofaunal assemblage, help to reconstruct the environmental conditions in which the Amma Fatma seep deposit formed.
Carbonates from zone 4 and the cover bed formed under the influence of more prolonged periods of oxygenated conditions. The occurrence of marine background fauna suggests that it dominated and outcompeted any specialized seep-specific taxa in this shallow water and high productivity environment. The absence of benthonic macrofauna in the lower part of the body might suggest lower levels of oxygen compared to the upper part. The presence of Palaxius in the lower part does not contradict this notion, as callianassid decapod crustaceans can tolerate extended periods of anoxia (Forster & Graf, 1992;Anderson et al., 1994;Bromley, 1996). The carbonate body formed either during or shortly after deposition of the host strata, and vital information on the environmental conditions can be derived by analysing the trace metal content of host black shales and limestones. Organic-rich shales are distinctly enriched in trace metals including Cr, Mo, U, Ni, Zn, Cu, V, Cd, Co and Re, whereby the modes and mechanisms of enrichment differ somewhat between these elements (Calvert & Pedersen, 1993;Algeo & Maynard, 2004;Brumsack, 2006). The fact that all trace metalswith the exception of Cr and Codo not correlate with Al confirms that their authigenic enrichment is due to geochemical changes in bottom and pore waters. The use of Cr in palaeoenvironmental analysis has been cautioned, as it is to a large degree of detrital provenance and is highly mobile in marine sediments (Tribovillard et al., 2006). Cobalt mainly resides within the alumosilicate fraction of the black shales as it shows the strongest correlation with Al of all trace metals, and will be consequently omitted from the discussion below. Black shale beds 7 to 13 show EFs of varying magnitudes, which argue for changing environmental conditions during deposition (Table 3). Molybdenum, Cd, Re and U exhibit the highest EFs. Molybdenum behaviour in sedimentary pore waters is linked to the presence of hydrogen sulphide that causes a speciation change from oxidized Mo to reduced thiomolybdate, which can be much more readily deposited either on organic species, or be incorporated and co-precipitated by iron sulphides (Helz et al., 1996). Cadmium exhibits a similar behaviour towards dissolved hydrogen sulphide, except that it accumulates as a separate sulphide phase and can be just as easily enriched in barely reducing conditions at very low hydrogen sulphide concentrations (Rosenthal et al., 1995). The ease of Cd enrichment in anoxic sediments and its high sensitivity towards hydrogen sulphide leads to high EFs, as well as to the large spread in EFs and total concentrations between zones 1 and 4 (Tables 2 and 3). Rhenium EFs are the highest on average of any trace metal. This is due to its very low crustal concentrations and the ease with which it accumulates in anoxic sediments (Crusius et al., 1996). Rhenium enrichment does not require the presence of free hydrogen sulphide or Fe or Mn (oxy)hydroxides, and is solely controlled by the redox potential of the ambient water (Morford et al., 2005). This suggests that the sediments in which carbonates of zones 1 to 4 formed were all characterized by a very shallow oxygen penetration depth or, more likely, periodically anoxic bottom waters.
In modern sediments U enrichment is observed at redox potentials present within the Fe reduction zone, before Mo accumulation occurs under more reducing conditions (Morford & Emerson, 1999). It has been argued that preferential U over Mo uptake is an indicator for suboxic conditions, whereas the reverse is the case when pore waters are strongly sulphidic (Algeo & Tribovillard, 2009). The decrease in Mo/U ratios, total Mo concentrations and Mo, Cd and Re EFs, combined with the concomitant increase in U content from zone 1 to 4 suggests a trend from euxinic conditions in zone 1 to more mildly anoxic conditions in zone 4. Nickel and Zn accumulate most effectively under euxinic conditions and reside in Ni-and Fe-sulphides, respectively (Algeo & Maynard, 2004). Vanadium can also accumulate efficiently within the nitrate reduction zone of marine sediments, and does not necessarily require dissolved sulphide. However, the combined enrichment of V, Zn, Ni and Mo suggest at least temporal euxinia at the sediment-water interface throughout zones 1 to 4 (cf. Hetzel et al., 2009). The notion that ambient water mass chemistry lead to authigenic enrichment of Mo, Ni, Zn, V and Cd is supported by the strong positive correlation between these elements, and their lack of correlation to Al (Table 4).
Trace metal contents of black shales and limestones are governed by bottom water redox conditions, as well as sedimentary redox conditions. Redox-sensitive metals precipitate at redox boundaries in the sediments that often lie substantially below the sediment surface, which may lead to a strong offset between the age of authigenic precipitates and the age of the sediment in which the elements are enriched. It can therefore be difficult to discriminate between enrichment of trace metals via direct precipitation from bottom waters, and enrichment via the precipitation of solid phases within redox zones in the sediment (cf. Kasten et al., 2003;Reitz et al., 2004). However, this age offset between authigenic minerals and the host sediment decreases the more the redox zonation is compressed in the sediment like in the case of anoxic bottom waters. Analyses of trace metal enrichment argue for a formation of black shale and marly limestone beds 7 to 14 under episodic euxinic bottom waters. More importantly, the distribution of Mo, U, Cd and Re implies changes in hydrogen sulphide content of the water mass, from higher contents in zone 1 to lower contents in zone 4. This questions the proposition that the carbonate body grew in its entirety and simultaneously within the shales and limestones. Zone 4 carbonates formed under at least episodically oxygenated conditions, which would have hampered the formation of black shales in bed 13.
Despite the occurrence of macrofossils in zone 4, the Amma Fatma body is poor in fossils compared to most Phanerozoic seep deposits, which are usually replete in shelly macrofauna (Campbell & Bottjer, 1995;Kiel & Peckmann, 2007;Little et al., 2015). Most of the shells identified in thin sections are severely fragmented and were most likely transported to the site over some distance, indicating that most of these molluscs did not live at the seep. Environmental conditions for macrofaunal species at the Amma Fatma seep were at the very least challenging, if not precluding their survival altogether. Euxinic conditions occurred during all stages of carbonate precipitation, interrupted by sporadic oxygenation during formation of zone 4 that enabled macrofaunal assemblages to temporarily colonize the seep. Bottom water conditions were temporarily sulphidic, as were the shallow sediments where carbonate precipitation occurred, consistent with reconstructions for the wider Tarfaya palaeo-shelf area (Kolonic et al., 2005;Poulton et al., 2015). The carbonate mineralogy of the Amma Fatma seep deposits is almost exclusively calcite, lacking any banded and botryoidal aragonite cement. Calcite precipitation at seeps is favoured over aragonite precipitation under the influence of strongly reducing pore waters with low sulphate concentrations (Haas et al., 2010;N€ othen & Kasten, 2011). The environmental conditions as constrained by trace metal analyses of the black shales did consequently not only govern the mode of accretion, but apparently influenced the mineralogy of authigenic carbonate at the Amma Fatma seep as well.
At passive continental margins various processes control the mechanism of fluid flow from sea floor sediments, including tectonic processes, thermohaline convection, sea-level changes and changes in sedimentation rates (Judd & Hovland, 2007;Suess, 2014). Relevant factors that may have influenced the mode of seepage in the Cretaceous Tarfaya Basin include differential compaction and pore water overpressure, both of which are in turn controlled by sedimentation rates and changes in sea-level. Low sea levels promote seepage of fluids through the sediment column (Teichert et al., 2003), as the hydraulic pressure needs to exceed the pressure of the water column to enable upward fluid flow towards the sea floor (Carson & Screaton, 1998). Highfrequency sea-level changes at Amma Fatma Plage are indicated by intercalations of bioclastic storm beds into the marly, hemipelagic background sediments (cf. El Albani et al., 1999b). The Amma Fatma coastal section is early Turonian in age, and studies from the Tarfaya Basin have demonstrated that the sea-level rose continuously during the Cenomanian due to a combination of eustatic sea-level rise and shelf subsidence (Gebhardt et al., 2004;Mort et al., 2008;Kuhnt et al., 2009). This sea-level rise culminated during the earliest Turonian where sea-levels were possibly up to 250 m higher than today, the highest during the Cretaceous (Haq, 2014). However, Kuhnt et al. (2009) inferred that the lowermost Turonian at Mohammed Plage was characterized by a sea-level lowstand, linked to the ephemeral build-up of Antarctic ice sheets during that time. This is in accord with the inferred outer shelf water depths at Amma Fatma, which should have promoted seepage of hydrocarbon-rich fluids due to the comparatively low pressure of the water column. The presence of shallow water species Broisonia enormis and the abundance peak of Gartnerago segmentatum (Thierstein, 1976;Hattner et al., 1980) further suggests that a short-term shallowing event may have occurred. Finally, El Albani et al. (1997) noted an upward increase in the frequency and thickness of the storm beds, and interpreted this observation as evidence for a falling sea-level through the section corresponding to the eustatic sea-level development during the late Early Turonian also suggested by Hardenbol et al. (1998).
Apart from sea-level changes, sedimentation rates also affect seepage dynamics in continental margin sediments. Sedimentation rates in the Tarfaya Basin during OAE 2 were estimated to have been above 10 cm kyr À1 , or correspond to 4Á8 cm kyr À1 at well S13 (see Fig. 2) and 1Á6 cm kyr À1 at Mohammed Plage (Kolonic et al., 2005). Despite the differences in these reconstructions, all estimates reflect high sedimentation rates. Such high sedimentation rates increase sediment loading on the continental slope that can lead to pore water overpressure, which squeezes reduced fluids from the underlying organic carbon-rich source rocks towards the sea floor (cf. Suess, 2014). Taken together, favourable conditions for seepage over long timescales were apparently favoured by a relative sea-level fall and high sedimentation rates, which facilitated the precipitation of large amounts of authigenic carbonate at the Amma Fatma seep.

CONCLUSIONS
The Amma Fatma carbonate body is an unusual seep deposit, as seep-endemic macrofauna that characterizes many other seep deposits is absent. Nonetheless, lipid biomarkers testify to the presence of microbial communities that performed AOM and contributed to authigenic carbonate formation. With d 13 C values of matrix micrite as low as À23Á5& it appears that anaerobic oxidation of methane was not the only trigger for carbonate precipitation. Remineralization of organic matter residing in the host black shales may have been an additional source of carbon for limestone formation. Persistent seepage was probably favoured by a sea-level lowstand and high sedimentation rates. The Amma Fatma deposit formed during the aftermath of OAE 2 over a considerable length of time. High productivity and partly euxinic conditions resulted in the deposition and preservation of black shales rich in organic matter. The Amma Fatma deposit is poor in macrofossils, and those present in the uppermost zone 4 are all Cretaceous shallow water species. Their presence testifies to the sporadic occurrence of oxygenated conditions in the bottom water. However, nannofossil distribution and pronounced enrichments of trace metals within the host strata point towards a high productivity setting with recurrent episodes of euxinia that shaped the shallow marine environment. Identifying seep deposits in such a shallow water setting poses a challenge. Diagnostic signals are masked by high primary production favouring heterotrophic taxa, which take advantage of the vast amounts of organic matter derived from photosynthetic primary production and outcompete obligate, endemic chemosymbiotic seep fauna. The Amma Fatma carbonate body may serve as an example on how to identify seep deposits in high productivity, shallow water settings. In such an environment, only a combination of methods will allow to ascertain that authigenic carbonate deposits formed at hydrocarbon seeps, and to constrain the modes of carbonate formation.