An expanded succession of organic-rich marlstones and limestones deposited in the Tarfaya Basin provides an outstanding opportunity to closely retrace climate evolution and sea-level changes during the Cretaceous greenhouse period. We present high-resolution X-ray fluorescence (XRF) scanning and bulk carbon- and oxygen-isotope records from two newly drilled sediment cores in the Tarfaya Atlantic coastal basin, which recovered a continuous Upper Turonian to Campanian succession of ∼290 m thickness. The XRF core scanning records reveal three long-term oscillations in the abundance of terrigenous elements (increase of Al, Ti, K, Si, and Fe normalized against Ca), which correspond to progressive transgressive phases followed by rapid regressions during the Coniacian and early Santonian. Sea-level highstands during this interval corresponding to the Coniacian–Santonian oceanic anoxic event 3 (OAE 3) are characterized by overall oxygen-depleted to anoxic conditions at the seafloor (indicated by the high organic carbon content, the presence of laminations, and low log[Mn/S], high log[V/Ca], and high log[Br/Ca]). The upper Santonian interval marks the transition from anoxic to oxic bottom-water conditions, prevalent through the early Campanian. The composite bulk carbonate δ13C curve exhibits strong similarities to the global stacked δ13C reference curve, characterized by negative excursions in the early Coniacian (Navigation and East Cliff events) and late Santonian (bracketed by the Haven Brow and Buckle events) and by positive excursions in the latest Turonian (Hitchwood event), middle Coniacian (Wight Fall event), and at the Santonian-Campanian boundary. During the early Campanian, enhanced accumulation of fine-grained carbonate and clay-rich hemipelagic sediments, increasing bulk carbonate δ18O, and low log(Br/Ca) and log(V/Ca) values indicate climate cooling, associated with a substantial improvement in bottom-water ventilation. Two long-term δ13C cycles of ∼2 m.y. duration, probably related to variations in Earth’s orbital eccentricity, are associated with the long-term cooling trend initiating the Campanian–Maastrichtian climate transition toward a cool greenhouse state.
The Late Cretaceous was characterized by several oceanic anoxic events (OAEs), marked by the widespread deposition of organic-rich black shales in coastal and open-ocean areas (Schlanger and Jenkyns, 1976). These events have been associated with periods of rising sea-level and positive carbon-isotope (δ13C) excursions related to global enhanced burial of 12C-enriched organic matter (Arthur et al., 1990; Calvert and Pedersen, 1993). The paleoenvironmental conditions during these episodes of organic-rich deposition, as well as the causes and consequences of attendant perturbations in biogeochemical cycles, have been the subject of intense and controversial debate over several decades (e.g., Gautier, 1987; Arthur et al., 1988; Arthur and Sageman, 1994; Brumsack, 2006). Unresolved issues include the global character of anoxic events, their sensitivity/response to orbital climate forcing, the frequency and amplitude of associated sea-level fluctuations, the possible triggering mechanisms, including voluminous volcanic CO2 release, the nature of subsequent perturbations in the weathering and nutrient cycles, and the origin of black shale facies (high primary productivity vs. enhanced preservation through bottom-water anoxia).
Reconstructions of eustatic sea-level changes during the Cretaceous were initially based on global syntheses of sequence-stratigraphic data, which are still widely used as an exploration and first-order global correlation tool (Haq et al., 1988; Hardenbol et al., 1998; Haq, 2014). Eustatic curves were derived from the analysis and correlation of regional relative sea-level changes along continental margins, resulting in the identification of as many as 58 third-order eustatic events (Haq, 2014). These potentially eustatic fluctuations are mainly expressed as relatively rapid (<1 m.y.) and large-amplitude (>25 m) sea-level falls, and most of them are documented in several basins (Cloetingh and Haq, 2015), although the precise stratigraphic correlation of many events remains ambiguous (Miall, 2009; Haq, 2014). Although the causes for such rapid sea-level falls during the largely ice-free Cretaceous remain a matter of vigorous discussion (e.g., Sames et al., 2016), the buildup of transient ice sheets in Antarctica has been proposed as a likely explanation (Miller et al., 2005; Flögel et al., 2011; Haq, 2014).
The scarcity of continuous expanded successions has, however, severely hampered the establishment of a global orbitally calibrated biostratigraphic and lithostratigraphic framework. The expanded succession of organic-rich marlstones and limestones within the Tarfaya Basin provides an excellent opportunity to reconstruct climate evolution and sea-level changes through the late Turonian to early Campanian. This marginal basin along the east Atlantic passive continental margin was continuously subsiding during the late Albian to early Campanian, resulting in the deposition of an ∼700-m-thick series of hemipelagic marlstones and limestones in the distal part of the basin, close to the present-day coastline near the town of Tarfaya (Choubert et al., 1966; Wiedmann et al., 1978; Aquit et al., 2013). The sedimentary succession can be subdivided into a hierarchy of lithological units, including units with thicknesses of several tens of meters, beds on the scale of decimeters to meters, and fine laminations on the millimeter scale. Several Upper Cretaceous sequences (unconformity-bounded units) have been discriminated in outcrop sections (Choubert et al., 1966; Aquit et al., 2013), and excellent age control can be achieved with abundant, relatively well-preserved planktonic foraminifera, ammonites, and calcareous nannofossils (Choubert et al., 1966; Lehmann, 1966; Wiedmann et al., 1978; Kuhnt and Wiedmann, 1995; Kuhnt et al., 1997, 2005, 2009). Previous studies have mainly focused on the intensity of anoxia, the magnitude and nature of the δ13C excursion, the biotic effects on benthic and planktonic foraminifera, biostratigraphic records, and paleoenvironmental evolution during the Cenomanian-Turonian OAE 2 event (e.g., El Albani et al., 1999a, 1999b; Holbourn et al., 1999; Kuhnt et al., 1997, 2005, 2009; Kolonic et al., 2002, 2005; Mort et al., 2007, 2008; Gertsch et al., 2010; Aquit et al., 2013).
Here, we analyzed a 290 m composite record of late Turonian to early Campanian sedimentation from two drill cores recovered in 2009 within the central part of the Tarfaya Basin, close to the Sebkha Tah (SN 1) and the town of Tarfaya (SN 2). Our main objectives were: (1) to reconstruct the paleoenvironmental evolution and sea-level changes of the Tarfaya Basin, based on visual core description, bulk carbonate stable isotopes, and X-ray fluorescence (XRF) scanner-derived elemental distribution data; (2) to correlate Upper Cretaceous successions in the newly drilled cores with more proximal outcrop sections (Aquit et al., 2013); (3) to date and correlate observed unconformities (based on carbon isotope and elemental ratio fluctuations and inferred sequence boundaries) to unconformities in other sedimentary basins; and (4) to compare the local sea-level record to other regional and global sea-level curves (Hardenbol et al., 1998; Miller et al., 2004; Haq, 2014) in order to investigate the influence of eustatic sea-level fluctuations on sedimentation in the Tarfaya Basin.
MATERIAL AND METHODS
Drilling of Cores Tarfaya SN 1 and SN 2
The east Atlantic passive continental margin, and in particular the Tarfaya Basin, with its rapid subsidence during the Late Cretaceous, provides an excellent location at which to recover Cretaceous deposits affected by minimal regional tectonic influences (Choubert et al., 1966). Two drill cores, Tarfaya SN 1 and SN 2 (∼30 km apart), were recovered with the help of the Office National des Hydrocarbures et des Mines (ONHYM, Morocco) during October–December 2009 (Fig. 1). Core Tarfaya SN 1 was located relatively close to the Sebkha Tah, ∼30 km southeast of the town of Tarfaya (27°42′36.6″N, 12°56′39.0″W). Core Tarfaya SN 2 was drilled closer to the coastline, ∼10 km east of the town of Tarfaya (27°57′43.1″N, 12°48′37.0″W). In total, a 550 m succession of sediment was recovered with a Salzgitter WD3500 hydraulic drilling system from these two drill sites (350 m at Tarfaya SN 1 and 200 m at Tarfaya SN 2). Metal core barrels (3.05 m long) were used, with a diameter of 8.8 cm in the upper part of each hole (from 0 to 147 m for core Tarfaya SN 1 and from 0 to 57.3 m for core Tarfaya SN 2) and a diameter of 6.8 cm in the lower part of each hole. Drilled sections of 3.05 m were divided into segments (∼80 cm), which were inserted into plastic sleeves, sealed to avoid desiccation, and then stored in wooden boxes for transport. Segments were subsequently split into archive and working halves with a high-precision Kaufmann-Titan diamond rock saw. Detailed core descriptions made on working-half segments prior to XRF are presented in the Supplementary Material (Figs. DR1 and DR2).1
In total, 73 (42 samples from SN 1 and 31 samples from SN 2) micropaleontological samples from bituminous marls with high organic matter content were crushed and processed using an alcoholic solution of anionic tensides (REWOQUAT by REWO-Chemie, Steinau, Germany), which helped to break down indurated samples. Around 50 g of dry sediments from each sample were washed and sieved into 63–150, 150–250, and 250–630 µm fractions. Planktonic foraminifera from the 250–630 µm fraction were picked, and an initial stratigraphy was established based on the biozonation scheme of Robaszynski and Caron (1995). The Robaszynski and Caron (1995) scheme defines four global planktonic foraminiferal zones between the early Campanian and late Turonian that can be clearly discriminated in cores Tarfaya SN 1 and SN 2 (Fig. 2):
(1) Globotruncanita elevata zone (early Campanian);
(2) Dicarinella asymetrica zone (late Coniacian to around the Santonian-Campanian boundary in terms of planktonic foraminiferal biostratigraphy);
(3) Dicarinella concavata zone (early to late Coniacian); and
(4) Marginotruncana sigali zone (late Turonian to early Coniacian).
In total, 18 subsamples from core Tarfaya SN 1 and 38 subsamples from core Tarfaya SN 2 were additionally taken for nannofossil investigation to support and refine the planktonic foraminiferal biostratigraphy using the Upper Cretaceous (UC) biozonation scheme of Burnett et al. (1998).
Line Scanning and Core Photography
Line-scan measurements and photographs were acquired with a Jai CV-L107 3 charge coupled device (CCD) color line-scan camera with three sensors of 2048 pixels and a dichroic red-green-blue (RGB) beam-splitter prism (RGB channels at 630 nm, 535 nm, and 450 nm) at the Institute of Geosciences, Christian-Albrechts-University in Kiel. Color measurement in L*a*b* units are from RGB digital images. Scanning was performed (resolution of 143 pixels per centimeter) on the polished surfaces of oriented cores.
X-Ray Fluorescence (XRF) Core Scanning
Elemental composition of the sediment was analyzed on the working-half core surfaces using a second-generation Avaatech XRF core scanner at the Institute of Geosciences, Christian-Albrechts-University in Kiel. The core surface was covered with a 4-μm-thick Ultralene plastic film to avoid contamination and to protect the detector. Measurements were taken continuously at 1 cm intervals with a down-core slit size of 10 mm over a 1 cm2 area. A few short intervals that were too disturbed to obtain smooth surfaces, mainly in the clay-rich upper part of core Tarfaya SN 1, could not be scanned and appear as gaps in the plots. Tube voltage settings of 10, 30, and 50 kV were used with a sampling time of 10 s to analyze the following elements: Al, Si, Ti, V, Fe, Mn, S, K, Ba, Zr, Sr, Br, and Ca. Raw data spectra were processed by the analysis of X-ray spectra with the iterative least-square software (WIN AXIL) package from Canberra Eurisys. The analytical precision of the second-generation Avaatech XRF scanner depends on the physical properties of the analyzed core material, the element concentration, the energy level of the X-ray source, and the count time. Replicate measurements of typical core material indicated deviations around the mean of less than ±10% for element intensities of 1000 and more (P, V, Mn, Zr), ±5% for element intensities of 3000 and more (Al, S, Ti, Fe), and ±2% for element intensities of 20,000 and more (Ca, Si, K; i.e., Tjallingii et al., 2007; Kuhnt et al., our data, 2015). Results are reported in the logarithms of elemental ratios, which provide the most easily interpretable signals of relative changes in chemical composition down-core and minimize the risk of measurement artifacts from variable signal intensities and matrix effects (Weltje and Tjallingii, 2008).
Variations in the abundance of the major elements K, Fe, Ti, Si, and Al are commonly used to reconstruct abundance changes in the terrigenous components of marine sediments (Peterson et al., 2000; Haug et al., 2001; Jaeschke et al., 2007; Mulitza et al., 2008; Tisserand et al., 2009; Govin et al., 2012). We normalized these records against Ca, mainly derived from the biogenic carbonate of marine organisms, and express the ratio of terrestrial-derived elements versus marine carbonate as log([Al + Ti + Fe + K + Si]/Ca). The K/Al ratio has been suggested as a proxy indicator for the clay mineral composition (Boyle, 1983; Weaver, 1967, 1989; Niebuhr, 2005). Low log(K/Al) ratios are characteristic of kaolinite-rich and/or smectite-rich assemblages, which preferentially form under warm humid (kaolinite) or warm arid (smectite) climatic conditions. The presence of smectite is very variable in Upper Cretaceous sediments of the Tarfaya Basin, with highest percentages in the Coniacian–Santonian (El Albani et al., 1999a). High log(K/Al) ratios indicate dominance of illite in the clay mineral assemblages, which is generally a weathering product in more arid, temperate climates, where physical weathering dominates over chemical weathering. However, in short intervals of the core, the K counts exhibit unusual spikes, which are related to precipitation of small salt crystals from the seawater drilling liquid on the core surface. We masked these spikes in the plots of XRF-scanner-derived log(K/Al).
We used the relative abundance of V (expressed as log[V/Ca]) as a qualitative proxy for enhanced organic matter accumulation and reducing conditions in the sediments, since V is commonly associated with organic matter in marine sediments, either by direct incorporation in organic complexes or by absorption on particulate organic matter during scavenging (Lewan and Maynard, 1982; Prange and Kremling, 1985). Vanadium in deposited particulate organic matter is also more stable under anoxic conditions and thus indicates not only enhanced accumulation of organic matter but also oxygen-depleted conditions at the seafloor (Shaw et al., 1990; Tribovillard et al., 2006). Above 158.8 m (top of the Santonian), the absolute values and amplitude variability of log(V/Ca) remain low, indicating generally better-oxygenated seafloor conditions with reduced organic matter and associated V accumulation.
Log(Mn/S) additionally provides information about deep-water oxygenation, as the sedimentary Mn content in organic-rich environments is often dependent upon redox conditions. Minerals containing reduced Mn are rare in marine sediments (Calvert and Pedersen, 1996; Tribovillard et al., 2006), and modern and Cretaceous upwelling deposits are commonly depleted in Mn (Brumsack, 1986). Sulfur can be used to evaluate the degree of pyritization in sediments, which is commonly related to redox conditions and the preservation of organic carbon (total organic carbon [TOC]). In general, low pyritization associated with low TOC content indicates well-oxygenated conditions during sedimentation (Gautier, 1987; Rachold and Brumsack, 2001). We thus used elevated log(Mn/S) values as a paleoredox indicator for better-oxygenated pore/bottom waters.
Log(Zr/Rb) is related to the grain-size variation in the sediments, depending on the distance from the clastic source. Liu et al. (2004) and Chen et al. (2006) used the Zr/Rb ratio as a grain-size proxy in the Loess Sequence in China. Zirconium is bound to zircon, a heavy mineral, generally transported over short distances, whereas Rb represents a light element derived from biotite and transported in clay minerals over long distances. High log(Zr/Rb) values indicate an increase in grain size and/or a proximal source and sediment transport over a smaller distance.
Stable-Isotope Analysis of Bulk Carbonate
We analyzed 800 samples (500 samples from core Tarfaya SN 1 and 300 samples from core Tarfaya SN 2) for stable isotopes of bulk carbonates. Measurements were made with Finnigan MAT 251 and MAT 253 mass spectrometers at the Leibniz Laboratory for Radiometric Dating and Stable Isotope Research at the Christian-Albrechts-University in Kiel. The instruments were coupled online to a Carbo-Kiel device for automated CO2 preparation of carbonate samples. Samples were reacted by individual acid addition. The systems have an accuracy (on the delta scale) of ±0.05‰ for carbon and ±0.08‰ for oxygen isotopes. The results were calibrated using the National Bureau of Standards and Technology (Gaithersburg, Maryland) carbonate isotope standard NBS 20, internal standards, and NBS 19 and are reported as δ18Ocarb and δ13Ccarb relative to Peedee belemnite (PDB).
The biostratigraphy of cores Tarfaya SN 1 and SN 2 is based on planktonic foraminiferal datums, supplemented by calcareous nannofossil datums. The top of the G. elevata zone was not reached in core Tarfaya SN 1, and the topmost Cretaceous sediment belongs to nannofossil zone UC15cTP (late early Campanian). The base of typically early Campanian planktonic foraminiferal assemblages, with G. elevata and Globotruncanita stuartiformis, in core Tarfaya SN 1 is at 158.24 m, the top of D. asymetrica is at 158.80 m, the base of G. elevata (in association with D. asymetrica) is at 159.03 m, and the highest sample without early Campanian Globotruncanita and common D. asymetrica is at 159.41 m, which indicates that the Santonian-Campanian boundary interval lies between 159.03 and 159.41 m, based on a planktonic foraminiferal approximation of this boundary (Gale et al., 2008). This boundary is located in the upper part of the positive carbon-isotope shift, at or very close to a distinct sediment change at 159.20 m (Fig. 2). On the basis of nannoplankton datums in core Tarfaya SN 1, this interval lies within zone UC14a, which corresponds to the early Campanian.
Based on nannofossils, the Santonian-Campanian boundary lies in zone UC13 (Gale et al., 2008), between 170.5 and 179.30 m, ∼10 m deeper than that interpreted from the foraminifera (Fig. 2). This discrepancy has two possible explanations: (1) a delayed occupation of the deep-dwelling G. elevata in the Tarfaya Basin, analogous to similar late occurrences of Helvetoglobotruncana (H.) helvetica and D. concavata in the Tarfaya Basin (Aquit et al., 2013), and/or (2) poor calibration of planktonic foraminiferal and nannoplankton datums in the Santonian-Campanian interval, partly resulting from lack of a ratified boundary-stratotype section and level for the base of the Campanian.
The base of the D. asymetrica zone is commonly used to approximately define the Coniacian-Santonian boundary (Robaszynski and Caron, 1995; Gradstein et al., 2012), even though the first appearance of D. asymetrica occurs below the boundary in the stratotype at Olazagutia, Spain (Lamolda et al., 2014). However, the stratigraphic placement of this datum may vary between workers and sections because of the continuous evolution from D. concavata into D. asymetrica and the relatively rare occurrence of D. asymetrica in the early part of its range. We previously redefined the base of this zone using the occurrence of “typical” specimens of D. asymetrica with five or more chambers in the last whorl, a wide umbilicus, and distinct umbilical ridges (Aquit et al., 2013). However, these typical D. asymetrica specimens are rare in the lower part of our D. asymetrica zone, where D. concavata and intermediate forms between D. concavata and D. asymetrica dominate the assemblages of umbilicoconvex Dicarinella. Based on the lowest, but rare, occurrences of D. asymetrica in core Tarfaya SN 2, the base of the D. asymetrica zone (57.20 m) lies within nannofossil subzone UC11b, above the base of Lithastrinus grillii and the top of Quadrum gartneri, and below the top of Lithastrinus septenarius and Flabelites oblongus, a position that corresponds to a late Coniacian age. We therefore place the Coniacian-Santonian boundary above this datum (Fig. 2).
The lower boundary of the D. concavata zone is well defined by the base of D. concavata at 104.16 m in core Tarfaya SN 2. However, the base of D. concavata appears somewhat later in the Tarfaya Basin (in nannofossil zone UC10, Coniacian) than its global base, which is in the late Turonian (Gradstein et al., 2012). In Tarfaya outcrop sections, D. concavata first occurs above the “astarte-lumachelle” marker beds (Aquit et al., 2013), dated by ammonites as early Coniacian (Choubert et al., 1966; Wiedmann et al., 1978). A Coniacian age for the first occurrence of D. concavata in the Tarfaya Basin is supported by our correlation to the calcareous nannofossils, where it lies in zone UC10 (Middle–Upper Coniacian), above the base of Micula staurophora.
Correlation of Cores Tarfaya SN 1 and SN 2
The correlation of cores Tarfaya SN 1 and SN 2 is based on a combination of elemental ratios from XRF scanning, lithological descriptions, line-scan records, bulk carbonate stable isotopes, and planktonic foraminiferal biostratigraphy (Fig. 2; Fig. DR3 [see footnote 1]). We define the tie point for correlation at 191.6 m in core Tarfaya SN 1 and at 38.85 m in core Tarfaya SN 2, corresponding to a lithological change from (lower) laminated black shales with nodular limestones to (upper) light-brown marlstones in both cores. This correlation point is further supported by a downhole decrease in lightness (L*), a major increase in Zr/Rb, and a long-term decrease in Al/Ca from the base of the D. asymetrica zone in both cores. The negative excursion in the bulk carbonate stable isotopes at 39.49 m in core Tarfaya SN 2 can be further correlated to the negative excursion at 192.73 m in core Tarfaya SN 1, supporting the lithological correlation. Both negative excursions have values ∼–4‰.
We constructed a composite depth scale (referred to as mcd) using the upper 191.6 m of core Tarfaya SN 1 and the part below 38.85 m in core Tarfaya SN 2. The total composite length of the two cores is 290 mcd. Down-core depths are given as mcd, when referring to the composite section, whereas down-core depths in individual cores are given as m. Composite depths in core Tarfaya SN 2 can be calculated as:
Lithology and Regional Unconformities
Cores Tarfaya SN 1 and SN 2 were divided into lithological units based on visual core descriptions and XRF-derived elemental data. Detailed lithological descriptions are provided in Figures DR1 and DR2 (see footnote 1). The upper 14 m section in core Tarfaya SN 1 and the upper 24.6 m section in core Tarfaya SN 2 consist of Pliocene–Pleistocene (“Moghrabien”) sediments, including sandstones, limestones, and lumachelles along with shell fragments. Core Tarfaya SN 1 is not described between 14 and 30 m due to intense fragmentation.
Based on the XRF scanning (Al/Ca and Zr/Rb), line-scan records (L*), and lithological changes, two major unconformities are identified at 156.6–158.8 m (U1a, 156.6 mcd; U1b, 158.8 mcd) in core Tarfaya SN 1 and at 134 m (U2, 286.75 mcd) in core Tarfaya SN 2 (Figs. DR1, DR2, and DR4 [see footnote 1]). Unconformity U1 lies in the lowermost Campanian, in the lowermost part of the G. elevata zone (U1a) and at the boundary between the D. asymetrica and G. elevata zones (U1b; and nannofossil subzone UC14a). Unconformity U2 occurs in the M. sigali zone (latest Turonian; nannofossil zone UC9, middle Turonian–early Coniacian).
Unconformities U1a and U1b correspond to a stepwise lithological change from brown shale to gray marl with heavy bioturbation, an increase in lightness (L*), and a decrease in Al/Ca and Zr/Rb (Figs. DR1, DR2, and DR4 [see footnote 1]). In contrast, unconformity U2 forms an erosive horizon at the base of a light-olive-green siltstone, characterized by oblique laminations (Fig. DR4 [see footnote 1]). This unconformity occurs above an interval with alternating laminated black shales and nodular limestones and is characterized by a marked increase in lightness (L*), a significant decrease in log(Al/Ca), and a slight decrease in log(Zr/Rb).
Bulk Carbonate δ18O
Bulk carbonate δ18O in cores Tarfaya SN 1 and SN 2 ranges between –2‰ and –5‰ (Fig. 3), reflecting a robust cooling trend in the Lower Campanian (mean = –4.19‰, standard deviation [SD] = 0.59 for the Upper Turonian–Santonian below 160 m; mean = –3.12‰, SD = 0.83 for the Lower Campanian, from 28.5 to 160 m). Cross-plots of δ13C and δ18O do not show significant correlation (Fig. 4), ruling out the presence of a late diagenetic cement to the primary skeletal calcite (Nelson and Abigail, 1996). We estimated sea-surface temperatures (SSTs) from the bulk (mainly coccoliths with fragments of nektonic larger organisms, planktonic foraminifera, and carbonate micro-particles so-called “micarb”) δ18O using the equation of Epstein et al. (1953):When using a δ18Owater of –1‰ (standard mean ocean water [SMOW]) for an ice-free world (Shackleton and Kennett, 1975), which corresponds to –1.27‰ on the PDB scale, the resulting SST estimates are ∼30 °C (36 °C) for the late Turonian to Santonian and ∼25 °C (30 °C) for the early Campanian. Values in brackets, which include vital effects corrections for coccolith δ18O, by subtracting 1.1‰ for the vital effects of coccolithophorid calcite formation (Dudley et al., 1986; Ennyu et al., 2002), are in agreement with other subtropical Late Cretaceous SST estimates (Pearson et al., 2001; Schouten et al., 2003; Alsenz et al., 2013; MacLeod et al., 2013; Linnert et al., 2014). Even though these absolute values may be strongly biased by issues related to coccolith vital effects and possible dilution of the sea-surface signal by fragments of deeper-dwelling nektonic and benthic organisms, the substantial increase in bulk δ18O indicates a major cooling event close to the Santonian-Campanian boundary. Increasingly arid conditions during the early Campanian may have additionally contributed to the seawater δ18O increase through enhanced evaporation and net transfer of water to the continental groundwater reservoirs.
Bulk Carbonate δ13C
Bulk carbonate δ13C values range between +1.7‰ and –2.4‰ (Fig. 5). In the upper part of the section (28.5–160 mcd, Lower Campanian), mean values (0.28‰, 0.46 SD) are lower than in European reference sections (Jarvis et al., 2006) but are comparable to records from eastern Tethys (Li et al., 2006; Wendler et al., 2009; Wendler, 2013) or Demerara Rise (MacLeod, 2006). The unusually low values (mean –0.77‰) and high scatter (SD 0.94) of δ13C below 160 mcd suggest either the presence of secondary calcite with a remineralized organic carbon component, or the growth of biogenic (mainly coccolith) calcite in unusually nutrient-rich and δ13C-depleted water masses. Previous studies in the Tarfaya Basin showed that δ13C is considerably lowered (to –11‰) within calcareous nodular concretions (El Albani et al., 2001). We avoided these lithologies when sampling for bulk isotopes, focusing on intervals with a relatively high clay-mineral content.
The composite δ13C record is characterized by a series of five low-amplitude positive carbon-isotope excursions with wavelengths in the range of 30–70 m, separated by transient high-amplitude negative excursions. We identified a first maximum (labeled 1 in Fig. 5) at ∼75 mcd, a second maximum at ∼155 mcd (labeled 2 in Fig. 5) in the Lower Campanian, a third, poorly developed maximum with significantly lower values at ∼175 mcd (labeled 3 in Fig. 5) in the Middle Santonian, a fourth maximum at ∼215 mcd (labeled 4 in Fig. 5) in the Upper Coniacian, and a fifth maximum at ∼245 mcd (labeled 5 in Fig. 5) in the Middle Coniacian.
Salient features are the positive shifts in the Middle Coniacian (starting at 260 mcd and peaking at 243 mcd) and around the Santonian-Campanian boundary interval (starting at 168 m and reaching a maximum at 153 m). These positive excursions appear as robust features in other Late Cretaceous δ13C records (i.e., corresponding to the δ13C increase between the Buckle/Foreness and the Santonian-Campanian boundary event and between the East Cliff to White Fall events in the English Chalk records; Jarvis et al., 2006). The amplitude of these events exceeds 1.5‰ in Tarfaya, which is more than double that in the English Chalk and the Niobrara Formation of the U.S. Western Interior (Locklair and Sageman, 2008), but it is similar to records from the eastern Tethys margin in Tibet (Li et al., 2006; Wendler et al., 2009; Wendler, 2013).
XRF Scanning Elemental Distribution
Terrigenous Elements versus Marine Carbonate—Log([Al + Ti + Fe + K + Si]/Ca)
The ratios of terrigenous elements to Ca derived from marine carbonate reach highest values (mean = –0.99, SD = 0.23) in the Santonian to Coniacian (158.8–286.7 mcd) and lowest values (mean = –1.58) in the Upper Turonian (286.7–290 mcd; Fig. 6). Terrigenous elements display the highest variability (SD = 0.35) within the Upper Turonian. Three sequences exhibiting an overall increasing trend in terrigenous elements can be identified in the Lower Turonian to Santonian from 190.5 to 221.5 mcd, from 221.5 to 253.5 mcd, and from 253.5 to 286.7 mcd (Fig. 5). The Lower Campanian (33.5–158.8 mcd) exhibits intermediate values (mean = –1.22), with comparatively low variability (SD = 0.19).
Proximity of Clastic Source and Sorting during Transport—Log(Zr/Rb)
The log(Zr/Rb) curve displays considerably higher values in the Middle and Upper Santonian (mean = 0.48, SD = 0.13), indicating either a more proximal sediment source or intense winnowing of fine-grained clastic sediment (clay minerals) on the Tarfaya shelf (Fig. 6). Both possibilities are compatible with a significant and persistent lowering of sea level associated with the sequence boundary at 190.5 mcd in the Lower–Middle Santonian. Lowest log(Zr/Rb) values occur in the Upper Turonian, between ∼268 and 290 mcd (mean = 0.02, SD = 0.13), probably associated with a distant sediment source and/or reduced fluvial sediment discharge. Intermediate values of log(Zr/Rb) characterize the Lower Campanian (28.5–158.8 mcd, sedimentary unit I; mean = 0.15, SD = 0.13) and between ∼268 and 190.5 mcd in the Coniacian and Lower Santonian interval (Fig. 6).
Geochemical Indicators of Organic Matter Accumulation and Bottom-Water Oxygenation—Log(Mn/S) and Log(V/Ca)
The Lower Campanian exhibits the highest log(Mn/S) values (mean = –1.33, SD = 0.23), indicating improved bottom-water ventilation, which is also evident from benthic foraminiferal assemblages in outcrop sections (Aquit et al., 2013). Between 190.5 and 158.8 mcd (Middle to Upper Santonian), log(Mn/S) increases from an average of –1.6, reaching –1.25 in the Lower Campanian (Fig. 6). The Lower Santonian to Upper Turonian interval is characterized by generally low, but highly variable, log(Mn/S) values (mean = –1.61, SD = 0.18), indicating overall oxygen-depleted to anoxic conditions at the seafloor, punctuated by short ventilation events. The most prominent of these events was associated with a regressive phase (shown by a decrease in log[Al + Ti + Fe + K + Si]/Ca) in the latest D. concavata zone (above the unconformity at 221.5 mcd).
The log(V/Ca) curve generally exhibits an inverse relationship to log(Mn/S) (Fig. 6), reflecting the affinity of V to organic matter. Lowest values characterize the Lower Campanian (mean = –3.31, SD = 0.17). The Santonian interval between 158.8 and 190.5 mcd exhibits intermediate values (mean = –3.03, SD = 0.17). Extremely high values occur in the Lower Santonian to Upper Turonian (mean = –2.86, SD = 0.32), with maxima in the highstand sediments below the sequence boundaries at 190.5, 221.5, and 253.5 mcd. The entire interval shows high-amplitude fluctuations, with high (>–2.5 log(V/Ca)) V accumulation peaks in laminated black shale intervals.
Latest Turonian to Early Campanian Sedimentary Evolution of the Tarfaya Basin
Processes Controlling Deposition of Organic-Rich and Carbonate-Rich Sediments
The Upper Cretaceous sediments of the Tarfaya Basin were deposited on an open-marine shelf, in water depths of ∼100–150 m (Wiedmann et al., 1978; El Albani et al., 1999a; Kuhnt et al., 2009). Following the depositional concept used for the analysis of Cenomanian sequences in the Tarfaya Basin (Kuhnt et al., 2009), lowstands and sequence boundaries are characterized by carbonate-rich sediments, including redeposited, shallow-water carbonate sands and pebbles. In contrast, transgressive system tracts coincide with the deposition of organic-rich sediments, which are commonly laminated, relatively low in carbonate, and enriched in Al, K, S, and Si (partly biogenic). The episodic occurrence of organic-rich sediments has been previously explained by the impingement of an expanded and intensified oxygen minimum zone onto the shelf during transgressive and maximum flooding phases (Kuhnt et al., 2009). This pattern of relatively coarse-grained lowstand carbonates and organic-rich, fine-grained transgressive and highstand sedimentation bears similarities to modern “shaved shelf” depositional models (James et al., 1994; Brachert et al., 2003). These models, developed for modern high-productivity, temperate carbonate shelves such as the Great Australian Bight, indicate that carbonate-rich sediments form during sea-level lowstands, when the wave-abrasion depth intersects with the seafloor. Wave abrasion during lowstands then winnows and exports the fine terrigenous clastic material (clay minerals) into deeper parts of the basin. Thus, the residual coarser material on the shelf mainly consists of carbonate grains (foraminiferal tests, shell fragments, calcareous dinoflagellate cysts).
Organic-rich sediments were recovered in cores Tarfaya SN 1 and SN 2 as laminated or homogeneous black, brown, or gray marlstones, characterized by high log(V/Ca) and log(Br/Ca) and low log(Mn/S). These elemental ratios exhibit obvious correlation (Fig. DR5 [see footnote 1]) to sulfur and TOC weight percentages analyzed in core Tarfaya SN 2 (Sachse et al., 2012). The organic matter is mainly derived from marine primary producers (phytoplankton and marine algae), with a small contribution from terrestrial particles (e.g., vitrinite and inertinite; Sachse et al., 2012). These organic- and pyrite-rich sediments are commonly laminated, pointing to oxygen-depleted bottom-water conditions, with enhanced flux and preservation of organic matter, reflecting episodic encroachment of an expanded and intensified oxygen minimum zone on the middle to outer shelf of the Tarfaya Basin. During these transgressive phases and highstands, fine-grained terrigenous sediments were remobilized close to the coastline and transported across the shelf, finally accumulating on the middle to outer shelf. Intervals of laminated, organic-rich sediments prominently occur near the top of sequences, in particular, below carbonate-rich lowstand sediments near sequence boundaries at 286.7, 253.5, 221.5, 190.5, and 158.8 mcd (Fig. 6; Fig. DR4 [see footnote 1]).
The most prominent examples of carbonate-rich intervals, depleted in terrigenous elements, occur above the unconformity at 286.7 mcd (U2) in the uppermost Turonian. The carbonates are characterized by low log(Al/Ca), log([Al + Ti + Fe + K + Si]/Ca), and log(Zr/Rb) (Fig. 6; Fig. DR4 [see footnote 1]), intense bioturbation, low log(V/Ca,) and high log(Mn/S), indicating improved oxygenation at the seafloor. Sediment structures (inverse grading, cross-lamination) indicate the influence of storm-induced bottom currents or storm wave currents (Fig. DR4 [see footnote 1]). We relate these limestone beds, which rest unconformably over organic-rich, fine-grained black shales, to major regressive events, which brought the seafloor of the central part of the Tarfaya shelf into the wave-abrasion zone for the first time since the Cenomanian-Turonian sea-level maximum.
Evolution of Depositional Environments
The latest Turonian to early Santonian was generally characterized by oxygen-depleted to anoxic conditions at the seafloor, punctuated by short-lived ventilation events (low log[Mn/S] episodes). The impingement of an expanded and intensified oxygen minimum zone on the shelf exhibits a periodic pattern, with three cycles of increasing organic matter and clay to carbonate ratios (286.7–253.5 mcd, 253.5–221.5 mcd, and 221.5–190.5 mcd; Fig. 6). This period of organic-rich sedimentation on the Tarfaya shelf corresponds, in a very broad sense, to OAE 3, which marks a significant transition in the long-term global climate evolution from Cretaceous greenhouse conditions, characterized by high global temperatures and reduced equator-to-pole thermal gradients, to the Late Cretaceous–Paleogene cooling (Wagner et al., 2004). However, the organic-rich deposits associated with OAE 3 occur in several marginal basins of the Western Tethys and Central Atlantic at different time intervals between the early Coniacian and the Santonian-Campanian boundary. Prominent organic-rich deposits occur in marginal basins of the eastern equatorial Atlantic (Holbourn and Kuhnt, 1998; Wagner, 2002; Hofmann et al., 2003), the Casamance Shelf offshore Senegal (Ly and Kuhnt, 1994), the La Luna Formation in Venezuela (Erlich et al., 2000; Rey et al., 2004), the Brazilian shelf (Mello et al., 1989), Mexico (Sohl et al., 1991; Nuñez-Useche et al., 2014), and the Western Interior Seaway of North America (Dean and Arthur, 1998; Locklair et al., 2011). The asynchronous occurrence and the absence of a distinct positive global δ13C excursion associated with OAE 3 indicate that the deposition of organic-rich sediments represents regional rather than global events (Wagner, 2002; Wagner et al., 2004; Wagreich, 2009, 2012). For instance, the bottom waters of lower bathyal and abyssal sites in the Western Tethys and within the Central Atlantic were fully oxygenated during OAE 3, as shown by the deposition of red sediments containing abundant and diverse agglutinated benthic foraminiferal assemblages (Kuhnt, 1990; Kuhnt and Moullade, 1991). Thus, the spatial and temporal distribution pattern of OAE 3 organic-rich sediments supports a scenario of several marginal basins along the Atlantic margin maintaining intense oxygen minimum zones, whereas the deep basins were already well oxygenated (Arthur et al., 1990; Ly and Kuhnt, 1994; Holbourn et al., 1999; Wagreich, 2012).
A major environmental change occurred around the Santonian-Campanian boundary interval (unconformities U1a and U1b) in the Tarfaya Basin. The early Campanian sedimentary environment was characterized by enhanced accumulation of fine-grained carbonate and clay-rich hemipelagic sediments. Benthic foraminiferal assemblages in outcrop sections close to core Tarfaya SN 1 indicate a depositional environment well below storm wave base (Holbourn et al., 1999; Aquit et al., 2013). Although the deepening of the Tarfaya Basin in the early Campanian may reflect local tectonic activity (Choubert et al., 1966; Lancelot and Winterer, 1980; Davison, 2005), it probably also relates to eustatic transgressive events. The increased water depth resulted in an improvement of oxygenation at the seafloor (increase in log[Mn/S]), as demonstrated by the high diversity and abundance of benthic foraminiferal assemblages (Holbourn et al., 1999; Aquit et al., 2013). The end of the Cretaceous sedimentary succession in the Sebkha Tah and Tisfourine sections, as well as in core Tarfaya SN 1, is marked by a major hiatus, which spans the late Campanian, Maastrichtian, and Paleogene.
Comparison of Tarfaya Depositional Sequences to the Eustatic Record from the New Jersey Margin
The 290-mcd-thick composite succession recovered in cores Tarfaya SN 1 and SN 2 allows us to discriminate five main sedimentary sequences, from the latest Turonian to early Campanian, with boundaries mainly defined by log(Al/Ca) data (Fig. 6). We correlated these sedimentary sequences to the New Jersey margin (Miller et al., 2003, 2004; Mizintseva et al., 2009), based on biostratigraphy and carbon-isotope chemostratigraphy (Fig. 3).
The sequence boundary at 286.7 mcd (unconformity U2 in core Tarfaya SN 2) marks the base of the Magothy III sequence of Miller et al. (2004) in the Upper Turonian (Fig. 3). This sequence boundary separates sediments deposited above the storm wave base following the latest Cenomanian–Turonian sea-level highstand, as this part of the Tarfaya Basin remained in an open-marine, middle to outer shelf setting throughout this period. A clear effect of lowered sea level is not evident on the depositional environment, except for the occurrence of common lumachelle beds in the early Coniacian, probably related to the action of storm waves.
During the Coniacian, three distinct cycles are recognized, based on the XRF records (e.g., log[Al/Ca]; Fig. 3). Sequence boundaries are placed at the abrupt change from organic carbon–rich laminated sediments to coarser, carbonate-rich sediments at 253.5, 221.5, and 190.5 mcd, which form the bases of stacked cycles. According to its stratigraphic position, the sequence boundary at 253.5 mcd corresponds to the base of the Cheesequake Sequence of Miller et al. (2004) and Mizintseva et al. (2009) in the Lower Coniacian (Fig. 3). The sequence boundary at 221.5 mcd can be correlated to the boundary between the Cheesequake and Merchantville Formations (Miller et al., 2004; Mizintseva et al., 2009) in the Coniacian–Santonian interval. The sequence boundary at 190.5 mcd is traced by XRF data (log[Al/Ca]) in core Tarfaya SN 2 (Fig. 2), but it could not be clearly discerned in core Tarfaya SN 1. This boundary corresponds to a regressive phase in the late Santonian and is possibly the equivalent of the base of the Merchantville II Formation of Miller et al. (2004) and Mizintseva et al. (2009).
The stratigraphically highest sequence boundaries, at 158.8 and 156.5 mcd (unconformities U1a and 1b), are dated as earliest Campanian. These sequence boundaries correspond to regression phases prior to the major Campanian transgression phase and are correlative to the base of the Merchantville III Sequence of Miller et al. (2004) and Mizintseva et al. (2009) (Fig. 3). A marked increase in carbonate and related decrease in the concentration of terrigenous elements and log(Zr/Rb), which occur in the Lower Campanian of core Tarfaya SN 1 at 84.5 mcd, may be correlative to the base of the Englishtown Sequence of Miller et al. (2004).
Sea level during the Cretaceous greenhouse world was substantially higher than at present and exhibited considerable long- and short-term variability (Miller et al., 2005; Müller et al., 2008). Long-term eustatic sea-level changes were probably controlled by plate tectonics. In contrast, relatively rapid variations in the order of tens of meters remain difficult to explain without assuming glacio-eustasy and the presence of ephemeral ice sheets punctuating periods of extreme warmth (Stoll and Schrag, 2000; Miller et al., 2003; Kominz et al., 2008; Kuhnt et al., 2009). Recently, Föllmi (2012), Wagreich et al. (2014), Sames et al. (2016), and Wendler and Wendler (2016), following earlier calculations by Hay and Leslie (1990), proposed aquifer eustasy as a viable alternative to glacio-eustasy to explain rapid sea-level fluctuations in a greenhouse world. In their scenario, net transfer of water to the continent by the infill of dried-out groundwater reservoirs and large inland basins at the (probably orbitally paced) transitions of arid to humid climate periods may have resulted in significant short-term sea-level falls. These rapid aquifer-eustatic forced regressions and subsequent increased runoff of freshwater into the ocean would have resulted in short-term oxygen-isotope fluctuations, which would be hardly distinguishable from rapid cooling/glaciation events followed by warming/freshening of marginal basins.
Correlation to the Global Carbon-Isotope Curve
Variability of the δ13C Record—Primary Signal or Early Diagenesis?
The Tarfaya bulk carbonate δ13C curve exhibits marked differences to published high-resolution Late Cretaceous δ13C records (e.g., Wendler, 2013), which show overall higher δ13C values and lack the high-amplitude fluctuations characterizing the Upper Turonian and Coniacian in Tarfaya. These high-frequency δ13C fluctuations are reflected by lithological changes, as high δ13C generally occurs within organic carbon–depleted intervals, when no upwelling-related oxygen minimum zone was established. In contrast, low δ13C characterizes laminated organic-rich intervals, when nutrient-rich water masses upwelled to the sea surface (Fig. 3). A recent compilation of Late Cretaceous δ13C records (Wendler, 2013) points to substantial differences (1‰–2‰) in the amplitude of δ13C events in the English Chalk reference section, the U.S. Western Interior Niobrara Formation, and Eastern Tethys sections in Tibet, which nevertheless all preserve the general pattern of the major carbon-isotope variations. Interbasinal and latitudinal differences in the amplitude of Cretaceous δ13C variations are likely related to locally differing upwelling regimes of nutrient-rich and δ13C-depleted water masses and to the efficiency of the biological pump. For example, productivity- and pCO2-dependent latitudinal differences in the amplitude of the carbon-isotope excursion in the range of 1‰–4‰ were recorded in biomarkers from photosynthetic algae during the OAE 2 carbon-isotope excursion (van Bentum et al., 2012). Today, in a well-ventilated ocean, the dissolved inorganic carbon of nutrient-rich upwelling water masses can be depleted by more than 1‰ (Berger and Killingley, 1977), and this value may have been significantly higher in the Late Cretaceous Atlantic Ocean, where intermediate water masses were strongly depleted in oxygen and enriched in nutrients and 12C. We, thus, consider the generally lower δ13C in the Tarfaya Basin to be largely a primary signal related to the upwelling of strongly nutrient-enriched and δ13C-depleted intermediate water masses.
The positive shifts in δ13C in the middle Coniacian (starting at 260 mcd and reaching a first peak at 243 mcd) and around the Santonian-Campanian boundary (starting at 168 mcd and reaching a maximum at 155 mcd) are correlated to the globally recognized δ13C increases between the East Cliff minimum and the White Fall event and between the Buckle/Foreness carbon-isotope minimum and the Santonian-Campanian boundary event in the English Chalk (Jarvis et al., 2006). Based on these anchor points, the remaining succession of δ13C fluctuations is tentatively correlated to main isotope events in carbon-isotope reference curves of the U.S. Western Interior Niobrara Formation and the stack of Wendler (2013), which is mainly based on the English Chalk reference section of Jarvis et al. (2006) (Fig. 5). We relate the maximum at ∼155 mcd to the Santonian-Campanian boundary event, the maximum at ∼175 mcd to the Horseshoe Bay event, and the maximum at ∼245 mcd to the Wight Fall event. In addition, the negative excursions at ∼167 mcd, ∼193 mcd, 248 mcd, and ∼270 mcd correspond to the Buckle, Haven Brow, East Cliff, and Navigation events, respectively.
The global positive carbon-isotope shift of the Santonian-Campanian boundary event (Scholle and Arthur, 1980; Jarvis et al., 2006) is well developed between ∼168 and 155 mcd in core Tarfaya SN 1 (Fig. 7). The total amplitude of the shift (∼1.5‰), from average background values of –1‰ at the base (168–170 mcd) to 0.5‰ above the shift (average between 156 and 150 mcd), is substantially higher than in the English Chalk (0.6‰) and in the Scaglia Limestone at Gubbio, Italy (0.4‰). The expanded succession in core Tarfaya SN 1 allows a detailed reconstruction of the δ13C curve across this interval, which reveals a negative excursion with an amplitude of ∼1‰ at the onset of the event (168.5–167.5 mcd) and a second short-lived negative excursion between 162.5 and 162 mcd (0.9‰ amplitude). In general, the variability of the δ13C is considerably higher in the lower part (168–162 mcd) than in the upper part (162–156 mcd) of the excursion. The Lower Campanian interval (between 158.8 and 34 mcd) exhibits two distinct cycles that appear to correspond to the 2–2.4 m.y. eccentricity cycle, bearing similarity to Mesozoic greenhouse sequences (Boulila et al., 2011) and to the third-order sequence record of the New Jersey margin (Miller et al., 2003, 2004, 2005; Van Sickel et al., 2004; Browning et al., 2008; Kominz et al., 2008; Mizintseva et al., 2009).
We correlated the carbon-isotope records of outcrop sections in the Tarfaya Basin (Aquit et al., 2013) to our new records from Tarfaya SN 1 and SN 2 in order to improve the stratigraphic assignment of outcrop successions and to assess their stratigraphic completeness. The Tisfourine section (Lower Campanian) corresponds to the upper part (∼30–50 mcd) of Tarfaya SN 1. The interval corresponding to the Lower Campanian and Santonian-Campanian boundary event (in total, more than ∼100 m thickness) is covered by sand dunes in the northern part of the Sebkha Tah, and between the Sebkha Tah and Sebkha Tisfourine, and is only accessible in the drill cores. The Akhfennir (∼35 m of Santonian) and Tah North outcrop sections (∼40 m of Santonian) are found to largely overlap, resulting in an overall thickness of ∼45 m of Santonian sediments. The marked negative excursion in the lower part of the D. concavata zone, originally correlated to the Navigation event in the El Amra section (Aquit et al., 2013), appears stratigraphically slightly younger, and we relate it to the East Cliff event in the English Chalk succession (Fig. 5).
The similarity of the Tarfaya δ13C record to that of the Niobrara Formation in the U.S. Western Interior Seaway (Locklair et al., 2011), which has an orbitally tuned chronology (Locklair and Sageman, 2008; Sageman et al., 2014), and to the global carbon-isotope stack of Wendler (2013), which was correlated to the geological time scale GTS12 (Gradstein et al., 2012) using planktonic foraminiferal zones, allows a correlation of the main isotopic events with the geological time scale. A comparison of the depth/age relationships for the main carbon-isotope events in the Niobrara Formation, the global carbon-isotope stack, and Tarfaya cores is given in Figure 8. According to this compilation, average sedimentation rates in Tarfaya SN 1 and SN 2 were ∼1.8 cm/k.y. in the Coniacian, ∼1.9 cm/k.y. in the Santonian, and ∼2.1 cm/k.y. in the early Campanian (base of UC15cTP around 31.6 mcd). Use of the recalculated ages of stage boundaries from Sageman et al. (2014) results in only slightly higher sedimentation rates over the interval spanning the base of the Coniacian to the Lower Campanian (1.9 cm/k.y. for the Coniacian, and 2.2 cm/k.y. for the Santonian). In contrast, late Turonian sedimentation rates are difficult to estimate, since the only potential tie point is the base of the Magothy III Sequence (in the Turonian of Tarfaya SN 2) at ca. 90.2 Ma in the GTS12 time scale (Gradstein et al., 2012). According to this age, late Turonian sedimentation rates would have been ∼5 cm/k.y., which is comparable to estimates of Turonian sedimentation rates in the nearby S13 well (Kuhnt et al., 1997; Meyers et al., 2012).
Study of newly drilled sedimentary successions in the Tarfaya Basin provides unprecedented insights into climate evolution and sea-level changes during the Cretaceous greenhouse period. Two drill holes recovered a complete spliced record of 290 m of organic-rich marlstones and limestones of late Turonian to early Campanian age. Average sedimentation rates ranged between ∼2.1 cm/k.y. during the Coniacian and early Campanian and ∼1.6 cm/k.y. during the Santonian. High-resolution XRF scanning records, in combination with stable-isotope data and micropaleontological age control, allow us, for the first time, to define and date stratigraphic sequences in this basin. These new ages provide constraints on the frequency and amplitude of local relative changes in sea level and, in combination with other records, contribute to the reconstruction of climate and globally averaged (eustatic) sea-level changes in the Late Cretaceous.
Fluctuations in the abundance of the terrigenous elements Al, Ti, K, Si, and Fe, normalized against Ca, indicate three significant sedimentary cycles of 33.2, 32, and 31 m thickness (286.7–253.5 mcd, 253.5–221.5 mcd, and 221.5–190.5 mcd) during the Coniacian to middle Santonian. These intervals of ∼1.5–2 m.y. duration are also characterized by recurrent impingement of an expanded oxygen minimum zone onto the shelf during transgressive phases and sea-level highstands, as shown by low Mn/S and high V/Ca ratios. Carbonate-rich sediments formed during sea-level lowstands, when the wave-abrasion depth intersected with the seafloor, and fine-grained clay minerals and organic matter were winnowed and exported into deeper parts of the basin. The stacking pattern of increasing carbonate and decreasing oxygenation proxies suggests that the three sedimentary cycles were related to sea-level changes in the order of several tens of meters. The Upper Santonian interval (158.8–190.5 mcd) corresponds to the transition from anoxic to prevalently oxic bottom-water conditions in the early Campanian.
The sea-level record imprinted in depositional sequences of the Tarfaya Basin is correlative to regional sea-level variations along the New Jersey margin (Miller et al., 2004; Mizintseva et al., 2009). In particular, we relate the sequence boundaries at 253.5, 221.5, 190.5, and 84.5 mcd to the bases of the Cheesequake, Merchantville I, Merchantville II, and Englishtown Sequences, respectively, of Miller et al. (2004). The sedimentary unconformities U1a/U1b and U2 are stratigraphically correlative to the base of the Merchantville III Sequence and the base of the Magothy III Sequence of Miller et al. (2004) and Mizintseva et al. (2009), respectively.
The Tarfaya bulk carbon-isotope record closely matches the global carbon-isotope stack of Wendler (2013). Marked events, correlative to the English Chalk (Jarvis et al., 2006) and the Niobrara Formation in the U.S. Western Interior Seaway (Locklair et al., 2011), are the Navigation event in the earliest Coniacian, the Haven Brow, the Horseshoe Bay, and the Buckle events in the Santonian, and the Santonian-Campanian boundary event. The Santonian-Campanian boundary event exhibits a positive carbon-isotope excursion of 1.5‰, followed by long-term cooling through the early Campanian. This trend is associated with cooler and drier conditions in the source area (indicated by an increase in the K/Al ratio related to a change from kaolinite- to illite-dominated clay mineral assemblages; El Albani et al. 1999a), as a first step in the Campanian-Maastrichtian climate transition toward a cool greenhouse state.
This research was funded by DEA Deutsche Erdoel AG (formerly RWE Dea AG) in cooperation with the Office National des Hydrocarbures et des Mines, Morocco, in the framework of the Atlantic Margin Integrated Basin Analysis Project and by the German Research Council (DFG) in the framework of SFB 754, TP A7. We thank Mohammed El Mallali for help with sawing the cores in Rabat. We thank Nils Andersen (Leibniz Laboratory for Radiometric Dating and Stable Isotope Research) for stable-isotope measurements, Dieter Garbe-Schönberg for advice with X-ray fluorescence scanning, and Wolfgang Reimers, Samuel Müller, and Moritz Kuest for technical help. We thank Thierry Adatte and Erik Sperling for critical reviews that helped us to improve the manuscript.
↵1GSA Data Repository item 2016209, lithological descriptions, core photographs with XRF scanner geochemical records of selected intervals and stratigraphic summary tables, is available at http://www.geosociety.org/pubs/ft2016.htm or by request to .
Science Editor: Bradley S. Singer
Associate Editor: Brian R. Pratt
- Received 27 February 2016.
- Revision received 13 June 2016.
- Accepted 13 July 2016.
- © 2016 Geological Society of America