1. Introduction
The middle Miocene period was one of profound climate change (e.g. Flower & Kennett, Reference Flower and Kennett1994; Zachos et al. Reference Zachos, Pagani, Sloan, Thomas and Billups2001). A relatively warm period, the Middle Miocene Climatic Optimum (MMCO; 16–14.5 Ma according to Abels et al. Reference Abels, Hilgen, Krijgsman, Kruk, Raffi, Turco and Zachariasse2005; 17–15 Ma according to Shevenell & Kennett, Reference Shevenell and Kennett2004) terminated during the Middle Miocene Climate Transition (MMCT; c. 14 Ma, Langhian–Serravallian), the second-largest climate glaciation after the Eocene–Oligocene Transition (Zachos et al. Reference Zachos, Pagani, Sloan, Thomas and Billups2001). Global comparison between stable isotopic records is essential in understanding the palaeoceanographic and climatic changes during the MMCO and MMCT. Several high-resolution isotope records spanning the MMCO and/or MMCT have become available over the last years, but three large areas remain uncovered: the Indian Ocean, the North Atlantic Ocean and the Arctic Ocean. The Arctic Ocean is extremely difficult to drill, and the few recovered cores have a low recovery (Moran et al. Reference Moran, Backman, Brinkhuis, Clemens, Cronin, Dickens, Eynaud, Gattacceca, Jakobsson, Jordan, Kaminski, King, Koc, Krylov, Martinez, Matthiessen, McInroy, Moore, Onodera, O'Regan, Pälike, Rea, Rio, Sakamoto, Smith, Stein, John, Suto, Suzuki, Takahashi, Watanabe, Yamamoto, Farrell, Frank, Kubik, Jokat, Kristoffersen, St John, Suto, Suzuki, Takahashi, Watanabe, Yamamoto, Farrell, Frank, Kubik, Jokat and Kristoffersen2006). The North Atlantic Ocean has numerous drill sites, however. Nonetheless, the studies from these locations often suffer from hiatuses and generally have a lower resolution than the existing Pacific and Southern Ocean records.
Louwye et al. (Reference Louwye, Foubert, Mertens and Van Rooij2008) and Quaijtaal et al. (Reference Quaijtaal, Donders, Persico and Louwye2014) presented palynological data from the Porcupine Basin in the eastern North Atlantic (Fig. 1). Sediments from Integrated Ocean Drilling Program (IODP) Site U1318 are of late early Miocene – middle Miocene age and contained well-preserved organic-walled dinoflagellate cysts (dinocysts) and other marine palynomorphs (Louwye et al. Reference Louwye, Foubert, Mertens and Van Rooij2008). Louwye et al. (Reference Louwye, Foubert, Mertens and Van Rooij2008) proposed a first age model for IODP Site U1318 based on magnetostratigraphy, biostratigraphic dinocyst data and the Sr-dating of a mollusc (Kano et al. Reference Kano, Ferdelman, Williams, Henriet, Ishikawa, Kawagoe, Takashima, Kakizaki, Abe, Sakai, Browing, Li, Andres, Bjerager, Cragg, De Mol, Dorschel, Foubert, Frank, Fuwa, Gaillot, Gharib, Gregg, Huvenne, Léonide, Mangelsdorf, Monteys, Novosel, O'Donnell, Rüggeberg, Samarkin, Sasaki, Spivack, Tanaka, Titschack, van Rooij and Wheeler2007). As reported by Expedition 307 Scientists (Reference Ferlman, Kano, Williams and Henriet2006), the shipboard palaeomagnetic measurements at Site U1318 were affected by magnetic overprint and by noise in the cryogenic magnetometer, rendering the identification of the weak original palaeomagnetic signature difficult. The measured magnetic intensity was considered very low and several measurements have an indistinct inclination between –20° and 20° rather than a distinct value near –90° or 90° (Louwye et al. Reference Louwye, Foubert, Mertens and Van Rooij2008). To overcome these problems, Louwye et al. (Reference Louwye, Foubert, Mertens and Van Rooij2008) measured the magnetic signature on several discrete samples since these measurements would not be affected by a magnetic overprint. Quaijtaal et al. (Reference Quaijtaal, Donders, Persico and Louwye2014) updated the age model of Louwye et al. (Reference Louwye, Foubert, Mertens and Van Rooij2008) with the aid of additional biostratigraphic dinocyst and calcareous nannoplankton data. However, biostratigraphic tie-points can be diachronic and correlation to global events is necessary to place the ‘barcode’ of magnetostratigraphy in time correctly.
Location of Integrated Ocean Drilling Program (IODP) drill Site U1318, as well as the location of DSDP Site 574 and IODP Site U1338. Figure adapted from IODP drill site maps (http://iodp.tamu.edu/scienceops/maps.html).
Here we present new benthic foraminiferal stable carbon and oxygen isotope data from intermediate water depths in the eastern North Atlantic Ocean. These data from a new, relatively complete record fill one of the data gaps in the North Atlantic Ocean. Additionally, the benthic δ18O and δ13C stable isotope records presented in this paper aid in confirming or adapting the magnetostratigraphy of Quaijtaal et al. (Reference Quaijtaal, Donders, Persico and Louwye2014) using magnetostratigraphic tie-points provided by Miller, Wright & Fairbanks (Reference Miller, Wright and Fairbanks1991) and Miller et al. (Reference Miller and Mountain1996, Reference Miller, Mountain, Browning, Kominz, Sugarman, Christie-Blick, Katz and Wright1998) and globally recognizable isotope stratigraphic events defined by Woodruff & Savin (Reference Woodruff and Savin1991).
2. Material
A total of 145 samples from IODP Leg 307, Site U1318 was selected for stable carbon and oxygen isotope measurements on benthic foraminifera (Fig. 2; online supplementary Table S1, available at http://journals.cambridge.org/geo). Site U1318 (water depth 409 m) was drilled in May 2005 and a composite record was established from Holes U1318B (51°26.148′N, 11°33.019′W) and U1318C (51°26.150.4358′N, 11°33.040′W) based on physical properties (Expedition 307 Scientists, Reference Ferlman, Kano, Williams and Henriet2006). The composite record spans seismic units P1, P2 and P3 (Van Rooij et al. Reference Van Rooij, De Mol, Huvenne, Ivanov and Henriet2003) that correspond to (parts of) lithostratigraphic units 1–3. The seismic units and lithostratigraphy are described in more detail in Expedition 307 Scientists (Reference Ferlman, Kano, Williams and Henriet2006) and Quaijtaal et al. (Reference Quaijtaal, Donders, Persico and Louwye2014) (Fig. 2). The samples for this study were selected from lithostratigraphic units 3A, 3B and 3C, which mainly consist of greenish-grey clay and are divided into subunits based on their calcium carbonate contents. Subunit 3A and 3B correspond to seismic unit P2; subunit 3C corresponds to seismic unit P1 (Expedition 307 Scientists, Reference Ferlman, Kano, Williams and Henriet2006). Sample spacing in the composite record is c. 1 m on average.
Core recovery (in black), lithology, lithostratigraphy of holes U1318B and U1318C and position of samples analysed for stable isotopes.
3. Methodology
3.a. Strategy for isotope analysis
Benthic isotopic records are preferably generated from epifaunal benthic species, since the relative change in δ13C values of epifaunal foraminifer tests of similar-sized specimens from the same species are dominantly influenced by the δ13C values of seawater (e.g. Jorissen, Reference Jorissen and Gupta2003; Fontanier et al. Reference Fontanier, Jorissen, Michel, Cortijo, Vidal and Anschutz2008). The most abundant epifaunal species in the record is Cibicidoides pachyderma. This species is morphologically similar to the often-used C. mundulus (ex. C. kullenbergi) and is known to occur from early Oligocene time until today (Holbourn, Henderson & MacLeod, Reference Holbourn, Henderson and MacLeod2013). Unfortunately C. pachyderma was not present in every sample, and measurements on a mixture of several other epifaunal benthic foraminifera were not justified due to preservational problems. The infaunal taxon Uvigerina sp. was therefore selected for measurements in addition to C. pachyderma. Uvigerina sp. specimens were present almost consistently from 196.54 mcd upwards (Fig. 2; online Supplementary Table S1) with a relatively high abundance and a general good to excellent preservation. The species was absent below 197.59 mcd. The δ13C value of infaunal species is not only influenced by the δ13C value of the seawater at a given offset, unlike for epibenthic species, but is also influenced by the microhabitat. In general, a deeper microhabitat results in a more depleted δ13C signature, which is a potential consequence of the pore-water δ13C dissolved inorganic carbon (DIC) gradient and/or a consequence of microhabitat effects (McCorkle, Emerson & Quay, Reference McCorkle, Emerson and Quay1985; Fontanier et al. Reference Fontanier, Jorissen, Michel, Cortijo, Vidal and Anschutz2008). Despite this drawback, the advantages of measuring the isotopes on the well-preserved Uvigerina sp. specimens is that: (1) preservation issues that could affect the δ18O values of less-well-preserved C. pachyderma specimens (Sexton & Wilson, Reference Sexton and Wilson2009) can be identified and prevented; and (2) a mixture of epibenthic specimens with possibly different isotope disequilibrium correction factors can be avoided. Changes in the δ18O values of Uvigerina sp. are, as well as the δ18O values of C. pachyderma, expected to reflect changes in the δ18O value of the seawater (e.g. by glaciation) and the prevailing temperature. There is no known microhabitat influence on the δ18O isotopes (Fontanier et al. Reference Fontanier, Jorissen, Michel, Cortijo, Vidal and Anschutz2008).
3.b. Sample preparation and measurements
For isotope measurements the samples were oven dried at 60°C for 24 hours, weighed and subsequently soaked in a tetra-natriumdiphosphat-decahydrat solution (5 g L–1) for at least 24 hours in order to deflocculate the clay (after Snyder & Waters, Reference Snyder, Waters, de Graciansky and Poag1984). The samples were then washed over a sieve with a 63 µm mesh size. The remaining >63 µm fraction was removed from the sieve with distilled water, oven-dried and weighed. The samples were then dry-sieved into three fractions: the 63–180 µm fraction; the 180–250 µm fraction; and the >250 µm fraction. Foraminifera were preferentially picked in the 180–250 µm fraction and complemented with foraminifera from the >250 µm fraction. Two to six specimens (c. 40–70 µg) were required for each measurement. The preservation of the picked foraminifera was evaluated relative to each other with the binocular microscope by breaking the foraminifera and assessing the transparency or the fading of textures. The preservation was generally good to moderate for the C. pachyderma specimens. Large secondary crystals were never observed with the binocular microscope; scanning electron microscopy showed that some specimens had minor secondary crystals on their test walls, however, which might have influenced the isotopic values (see online Supplementary Plate S1, available at http://journals.cambridge.org/geo). However, there are no indications that the isotopes from these specimens deviate from the others. The Uvigerina sp. specimens were generally well preserved.
The δ18O and the δ13C stable isotope ratios were measured with a ThermoFinnigan Deltaplus XL Mass spectrometer at the Vrije Universiteit Brussel (VUB) in Belgium. This continuous-flow isotope ratio mass spectrometer (CF-IRMS) is equipped with an automated ThermoFinnigan Kiel III carbonate preparation line. Accuracy corrections were made with an in-house standard called Nu Carrara Marble (NCM). The analytical precision averages 0.031‰ (1σ) for δ13C and 0.080‰ (1σ) for δ18O. All data (see online Supplementary Table S1) are reported against the Vienna Pee Dee Belemnite (VPDB) standard after calibration of the in-house standard with NBS-19. Duplicate measurements were performed on eight samples. Stable isotope data are used here for stratigraphic purposes only.
3.c. Correction factors for isotopic measurements
The δ13C and δ18O measurements of modern benthic foraminifera often have a consistent offset from calcite precipitated in equilibrium, and between different species. This is considered a consequence of microhabitat preferences and vital effects (e.g. Katz et al. Reference Katz, Katz, Wright, Miller, Pak, Shackleton and Thomas2003). Despite their different habitat, changes in the δ18O value of C. pachyderma and Uvigerina sp. are both expected to reflect the prevailing temperature and the δ18O value of the seawater at the time of test precipitation. The δ18O values are therefore assumed to correlate linearly. In 59 samples, both C. pachyderma and Uvigerina sp. were analysed for stable isotopes. The δ18O values of both species show a near 1:1 relation (see Fig. 3a). The linear regression resulted in the following equation: δ18OC. pachyderma = 0.984 × δ18OUvigerina sp. – 0.296‰ with a coefficient of determination (R²) of 0.94 and a prediction error of 0.096‰ (1σ, root mean square error of prediction). This good correlation confirms that the δ18O value of Uvigerina sp. can predict the δ18O value of C. pachyderma. The equation can be simplified with the assumption of a slope of 1, which results in the formula: δ18OC. pachyderma = δ18OUvigerina sp. – 0.311‰ with a prediction error of 0.097‰ (1σ, root mean square error of prediction). This prediction error is considered a consequence of the average analytical precision of 0.080‰. The good correlation furthermore suggests that the moderate preservation of some C. pachyderma specimens only has a minor or an insignificant impact on the measured δ18O values in this study.
Correlation between uncorrected stable isotope measurements of C. pachyderma and Uvigerina sp. (a) Correlation of δ18O. The correlation between the two species is δ18OC. pachyderma = 0.984×δ18OUvigerina sp. – 0.296‰; R 2 is 0.94, prediction error is 0.096‰ (1σ, root mean square error of prediction). (b) Correlation of δ13C. The correlation between the two species is δ13CC. pachyderma = 0.692×δ13CUvigerina sp. + 1.137‰; R 2 is 0.5172, prediction error is 0.170‰ (1σ, root mean square error of prediction). Error bars depict one standard deviation (1σ).
On the other hand, there is a weaker correlation for δ13C between C. pachyderma and Uvigerina sp. (see Fig. 3b). The linear regression resulted in the equation: δ13CC. pachyderma = 0.692 × δ13CUvigerina sp. + 1.137‰ with a coefficient of determination (R²) of 0.52 and a prediction error of 0.170‰ (1σ, root mean square error of prediction). This lower correlation cannot be considered a consequence of preservation since δ13C has been demonstrated to be more robust to recrystallization than δ18O (e.g. Sexton, Wilson & Pearson, Reference Sexton, Wilson and Pearson2006). The δ13C values of Uvigerina sp. are on average 1.076‰ more depleted than the δ13C values of C. pachyderma. This is likely the result of the infaunal habitat of Uvigerina sp. Similar offsets of c. 0.9‰ between Uvigerina and Cibicidoides were found by Shackleton & Hall (Reference Shackleton and Hall1984) and Shackleton, Hall & Boersma (Reference Shackleton, Hall and Boersma1984). The lower δ13C correlation between both taxa is suggested to be a result of the variation of the δ13CDIC pore-water gradient with time and the possible migration of the infaunal species within the sediment. For correlation of globally recognizable δ13C maxima, however, this should be of no consequence.
3.d. Isotope stratigraphy
Woodruff & Savin (Reference Woodruff and Savin1991) demonstrated that specific events in the benthic δ18O and δ13C isotope records from the Antarctic, Atlantic, Indian and Pacific oceans could be correlated. They defined seven δ13C maxima (CM1–7) and positive or negative δ18O excursions (A–G). They ‘relied heavily’ (Woodruff & Savin, Reference Woodruff and Savin1991, p. 762) on the high-resolution isotope record of DSDP Site 574 in the Pacific Ocean to define distinct isotope features. We therefore correlate our δ13C and δ18O records of IODP Site U1318 (Porcupine Basin) with the isotopic records of DSDP Site 574 composed of data from Pisias, Shackleton & Hall (Reference Pisias, Shackleton, Hall, Mayer and Theyer1985), Shackleton (unpublished data, 1985, as tabulated by Woodruff & Savin, Reference Woodruff and Savin1989) and Woodruff & Savin (Reference Woodruff and Savin1989, Reference Woodruff and Savin1991) using the terminology of Woodruff & Savin (Reference Woodruff and Savin1991), plotted against depth because the initial data were plotted against depth and to avoid problems with different versions of time scales. Furthermore, we include a correlation with IODP Site U1338 in the eastern equatorial Pacific using a recent high-resolution stable isotope study with an astronomically tuned time scale by Holbourn et al. (Reference Holbourn, Kuhnt, Lyle, Schneider, Romero and Andersen2014). However, the isotope stratigraphy of Woodruff & Savin (Reference Woodruff and Savin1991) is not calibrated against magnetostratigraphy, a correlative tool that is available in the Porcupine Basin (Louwye et al. Reference Louwye, Foubert, Mertens and Van Rooij2008). We have therefore chosen to additionally compare the δ18O against the Mi-(Miocene isotope) zonation of Miller, Wright & Fairbanks (Reference Miller, Wright and Fairbanks1991) and Miller et al. (Reference Miller and Mountain1996, Reference Miller, Mountain, Browning, Kominz, Sugarman, Christie-Blick, Katz and Wright1998) using the updated ages of Boulila et al. (Reference Boulila, Galbrun, Miller, Pekar, Browning, Laskar and Wright2011). The base of these Mi-zones, or Mi-events, is defined by the maximum δ18O value of one of the originally nine prominent positive δ18O excursions, possibly related to periods of cryosphere expansion (Miller, Wright & Fairbanks, Reference Miller, Wright and Fairbanks1991). By linking the isotope events of Woodruff & Savin (Reference Woodruff and Savin1991) to those of Miller, Wright & Fairbanks (Reference Miller, Wright and Fairbanks1991) and Miller et al. (Reference Miller and Mountain1996, Reference Miller, Mountain, Browning, Kominz, Sugarman, Christie-Blick, Katz and Wright1998), we can link our isotope stratigraphy to the correct magnetochron.
4. Results
The isotope stratigraphic correlation between the three compared records is straightforward, despite the fact that IODP Site U1338 and DSDP Site 574 are located in the Pacific Ocean and that IODP Site U1318 is located in the North Atlantic (Fig. 4). If we compare the absolute isotopic δ18O and δ13C values between the three records, the δ18O values from IODP Site U1318 are approximately 1.5‰ lower than the δ18O values of IODP Site U1338 and DSDP Site 574 (Fig. 4). It is assumed that the significantly more depleted δ18O values in the Porcupine Basin record are mainly a reflection of the warmer temperature conditions as a consequence of the shallower water depth. The absolute δ13C values of IODP Sites U1318, U1338 and DSDP Site 574 are remarkably similar.
Stable benthic oxygen (δ18O) and carbon isotopes (δ13C). Oxygen isotopes are represented by blue squares, carbon isotopes by red dots. The red dashed line is the CM line of 1.6‰ from Woodruff & Savin (Reference Woodruff and Savin1991) as set from Site 574. (a) Isotopes from DSDP Site 574 versus depth (Pisias, Shackleton & Hall, Reference Pisias, Shackleton, Hall, Mayer and Theyer1985; Woodruff & Savin, Reference Woodruff and Savin1989; Woodruff & Savin, Reference Woodruff and Savin1991; Shackleton, unpublished data tabulated by Woodruff & Savin, Reference Woodruff and Savin1989). δ13C values were not corrected for vital effects. (b) Isotopes from IODP Site U1318 versus depth. Dark blue squares and dark red dots are uncorrected measurements on C. pachyderma, light blue crosses and red plus signs are corrected measurements on Uvigerina sp. using the following formulas: δ18OC. pachyderma = 0.984×δ18OUvigerina sp. – 0.296‰; δ13CC. pachyderma = 0.692×δ13CUvigerina sp. + 1.137‰. (c) Isotopes from IODP Site U1338 versus age. Data from Holbourn et al. (Reference Holbourn, Kuhnt, Lyle, Schneider, Romero and Andersen2014). δ13C values were not corrected for vital effects.
4.a. Determination of isotope stratigraphic tie-points
Discussed from major to minor events, the most prominent event in the record is δ18O event E, which is defined as an abrupt δ18O increase of approximately 0.5‰. Event E is present in all three records, facilitating further comparison and aiding in the recognition of other isotope stratigraphic tie-points (Fig. 4). At Site U1318, event E can be observed from 143.51 to 133.17 mcd. This event is immediately followed by CM6, subdivided into CM6a and CM6b, at 131.81 and 126.61 mcd, respectively. CM6 is coeval with the base of Mi-zone Mi-3, tied to magnetosubchron C5ABr (Miller, Wright & Fairbanks, Reference Miller, Wright and Fairbanks1991). The δ18O maximum that will be the base of Mi-3 at 128.11 mcd lies just within an interval of normal polarity. According to the age model of Quaijtaal et al. (Reference Quaijtaal, Donders, Persico and Louwye2014) this is C5AAn (Fig. 5), but because of the δ18O maximum corresponding to Mi-3 it is more likely that this normal subchron represents C5ABn. Chron C5ABr then lies just below Mi-3 in our record, within the undefined polarity interval between 154.85 and 130.02 mcd.
Updated age model for IODP Site U1318 (in black); the tie-points based on the isotope stratigraphy of Site 1338 (in blue). Pink diamonds represent the Mi-events versus depth at Site U1318 and ages from Boulila et al. (Reference Boulila, Galbrun, Miller, Pekar, Browning, Laskar and Wright2011), connected to the magnetostratigraphy at Site U1318 with grey dashed lines. The black dashed line indicates the interval where ages were interpolated because of the uncertainty in the magnetostratigraphy. Black squares are magnetostratigraphic tie-points, red dashed lines indicate hiatuses, red arrows indicate dinocyst biostratigraphic tie-points, yellow arrows indicate nannoplankton presence or absence, grey mollusc indicates Sr-isotopic age from Kano et al. (Reference Kano, Ferdelman, Williams, Henriet, Ishikawa, Kawagoe, Takashima, Kakizaki, Abe, Sakai, Browing, Li, Andres, Bjerager, Cragg, De Mol, Dorschel, Foubert, Frank, Fuwa, Gaillot, Gharib, Gregg, Huvenne, Léonide, Mangelsdorf, Monteys, Novosel, O'Donnell, Rüggeberg, Samarkin, Sasaki, Spivack, Tanaka, Titschack, van Rooij and Wheeler2007).
Working downcore, the next stratigraphic tie-point should be CM5. In our record CM5, at c. 149.7 mcd, is less pronounced than in the records of DSDP Site 574 and IODP Site U1338 (Fig. 4). A problem with correlating CM5 is that there appear to be three peaks with increased δ13C values around the interval that Woodruff & Savin (Reference Woodruff and Savin1991) indicated. The middle peak is highest at DSDP Site 574, whereas in IODP Sites U1318 and U1338 the uppermost peak is heaviest. These peaks are tentatively named 5A, 5B and 5C (155.05, 148.51 and 145.51 mcd, respectively, at Site U1318). However, these peaks may also be artefacts of the different resolution between the three records. For depth correlation, the average depth/age value of the three peaks was used.
The next carbon isotope maximum at Site U1318 downcore, CM4, is located at 203.55 mcd (Fig. 4). CM4 appears to be more pronounced at the Porcupine Basin in comparison to the other two records. This δ13C maximum is higher than CM6, which is not the case in the Pacific records.
The δ18O event D is less easily characterized. The definition of Woodruff & Savin (Reference Woodruff and Savin1991) states that it is a period of low δ18O between CM4 and CM5 that is longer in duration than events A–C. However, the interval they appoint to D in their table 11 is rather short. This interval is the first interval of lower δ18O values, but at both Sites 574 and U1338 five peaks with decreased values follow. At Site U1318 three intervals with lighter values can be found, of which we defined the first as the D-event at 195.03–192.51 mcd (Fig. 4). After this lighter interval at Site U1338, at c. 14.7 Ma, a decrease in the amplitude of the δ18O cycli can be observed (Holbourn et al. Reference Holbourn, Kuhnt, Lyle, Schneider, Romero and Andersen2014). This tipping point can also be identified at IODP Site U1318 around c. 162 mcd, where the δ18O minima of the δ18O cycli decrease from –0.3‰ to 0.0‰, while the δ18O maxima remain c. 0.5‰.
The identification of CM3 is not straightforward. At first the heavy values at c. 250–240 mcd were thought to be CM3. However, CM3 is contemporaneous with the base of Mi-zone Mi-2, tied to the base of magnetosubchron C5Br (Miller, Wright & Fairbanks, Reference Miller, Wright and Fairbanks1991). According to the age model of Quaijtaal et al. (Reference Quaijtaal, Donders, Persico and Louwye2014) the initial interval for CM3 should correspond to subchron C5Cr.3n, and the base of C5Br should be at 184.15 mcd (Fig. 5). However, at the latter depth no clear δ18O maximum and corresponding carbon isotope maximum can be observed. Another carbon isotope maximum can be observed just directly below CM4 at 211.57 mcd. This maximum is close to a δ18O maximum at 209.55 mcd, and just below an interval of reversed polarity. Although this does not fit the description of Mi-2 entirely, the resolution of the isotope records in this interval is somewhat lower than higher up the composite record and it is to be expected that the highest maxima lie within the non-sampled part. We therefore tentatively appoint a depth of 211.57 mcd to CM3 and a depth of 209.55 mcd to Mi-2. There are no decreased δ18O values between CM3 and CM4, implying that isotope event C of Woodruff & Savin (Reference Woodruff and Savin1991) is not present in the Porcupine Basin. This event however is ‘often exhibited as a single data point’ (Woodruff & Savin, Reference Woodruff and Savin1991) and it might therefore be missing due to the lower sampling resolution in this part of the composite.
Due to the successful identification of Mi-zones Mi-2 and Mi-3, zone Mi-2a could also be identified. According to Miller et al. (Reference Miller, Mountain, Browning, Kominz, Sugarman, Christie-Blick, Katz and Wright1998) zone Mi-2a is tied to the base of magnetosubchron C5ADr. The fourth reversal above 207.15 mcd (base C5Br according to the location of Mi-2) is at 166.25 mcd (Fig. 4). There are two δ18O maxima surrounding this point, at 159.98 mcd and 173.08 mcd. Both points are located within an interval of normal polarity, contradictory to the indication of Miller et al. (Reference Miller, Mountain, Browning, Kominz, Sugarman, Christie-Blick, Katz and Wright1998). We have tentatively placed Mi-2a at 173.08 mcd because of the slightly heavier δ18O values.
The remaining CM at 237.1 mcd in IODP Site U1318 is CM2 (Fig. 4). It is plausible that the low δ18O value at c. 232.045 mcd then corresponds to δ18O-event B, although this depth should be treated with care since this is a single measurement and the resolution between 220 mcd and 240 mcd is very low.
The remaining δ18O event of Woodruff & Savin (Reference Woodruff and Savin1991) is event F. This event is coeval with the base of Mi-zone Mi-4 at the base of (undivided) C5Ar (Miller, Wright & Fairbanks, Reference Miller, Wright and Fairbanks1991). Several δ18O maxima are present above the base of zone Mi-3 at the Porcupine Basin. The fourth reversal above the top of C5ABr/base of C5ABn (as indicated by the location of Mi-3) should mark the base of C5Ar, the location of Mi-4 according to Miller, Wright & Fairbanks (Reference Miller, Wright and Fairbanks1991) and Miller et al. (Reference Miller, Mountain, Browning, Kominz, Sugarman, Christie-Blick, Katz and Wright1998), and is located at 116.51 mcd (Figs 4, 5). Around that depth a smaller δ18O maximum can be observed. However, according to Holbourn et al. (Reference Holbourn, Kuhnt, Schulz and Erlenkeuser2005, Reference Holbourn, Kuhnt, Schulz, Flores and Andersen2007) and Holbourn et al. (Reference Holbourn, Kuhnt, Frank and Haley2013), there is an interval of dominance of the 100 ka eccentricity cycle during 13.5–13.1 Ma, which at the Porcupine Basin should be between c. 128 and 102 mcd (13.1–12.5 Ma according to Quaijtaal et al. Reference Quaijtaal, Donders, Persico and Louwye2014; A. Holbourn, pers. comm., 2016). The δ18O maximum at the end of this eccentricity-dominated interval should mark Mi-4, and we place the base of zone Mi-4 at 101.95 mcd. Isotope event F spans 121.76–109.11 mcd. Event F is defined as the start of an episode of maximum middle Miocene δ18O values (Woodruff & Savin, Reference Woodruff and Savin1991). It also corresponds to one of the major incremental steps of increasing δ18O values and decreasing amplitude in the δ18O variability recognized by Holbourn et al. (Reference Holbourn, Kuhnt, Schulz, Flores and Andersen2007). Furthermore, the relative increase in δ 18O maxima after δ18O event E and F is approximately 0.3‰ in all three records.
4.b. Revision of the age model
The astronomically tuned isotopic events recognized at IODP Site 1338 provide a first approximate age for the corresponding events at IODP Site U1318 (see Table 1). According to the isotope stratigraphy, the record at IODP site U1318 encompasses a time interval from somewhat older than 16 Ma to somewhat younger than 13 Ma (Figs 4, 5); this is in contrast to the age model of Quaijtaal et al. (Reference Quaijtaal, Donders, Persico and Louwye2014) that suggests a time interval from c. 17.8 Ma to 12.0 Ma. The approximate ages provided by the isotope stratigraphy correspond better with three biostratigraphic dinocyst ages in the lower part of the record (Fig. 5). The highest occurrence of Cousteaudinium aubryae and Distatodinium paradoxum were calibrated against the top of magnetosubchron C5Br (De Verteuil & Norris, Reference De Verteuil and Norris1996), now at an age of 15.16 Ma (ATNST 2012). The lowest occurrence of Labyrinthodinium truncatum at 237 mcd, in the bottom part of the record, is correlated against the Burdigalian–Langhian boundary according to Williams et al. (Reference Williams, Brinkhuis, Pearce, Fensome, Weegink, Exon, Kennett and Malone2004). This boundary is currently dated at 15.97 Ma, supported by a Sr-isotope dating of <16.6 Ma by Dybkjaer & Piasecki (Reference Dybkjaer and Piasecki2010).
Depths and ages of the isotope stratigraphic events of Woodruff & Savin (Reference Woodruff and Savin1991) for DSDP Site 574, IODP Site U1318 and IODP Site 1338.
The first direct tie-points for magnetostratigraphy were provided by the base of Mi-zones (Table 2, Fig. 5). The identification of the base of Mi-2 led to the interpretation of the reversal at 207.15 mcd as the boundary between magnetosubchrons C5Cn.1n and C5Br. Mi-2a provided the identification of the boundary between C5Bn.1n and C5ADr at 166.25 mcd. This indicates that the age model of Quaijtaal et al. (Reference Quaijtaal, Donders, Persico and Louwye2014) is one magnetochron too old during Mi-2a and two magnetochrons too old during Mi-2. Above the undefined polarity interval between 154.85 and 130.02 mcd we can identify the boundary between C5ABr and C5ABn at 130.02 mcd due to the position of Mi-3 at 128.11 mcd. The boundary between C5Ar.2n and C5Ar.3r could be identified because of Mi-4. The age model of Quaijtaal et al. (Reference Quaijtaal, Donders, Persico and Louwye2014) is one magnetochron too young for the interval spanning Mi-3, and two magnetochrons too young during Mi-4.
Ages, corresponding magnetosubchrons and depths of the Mi-zones of Miller, Wright & Fairbanks (Reference Miller, Wright and Fairbanks1991) and Miller et al. (Reference Miller, Mountain, Browning, Kominz, Sugarman, Christie-Blick, Katz and Wright1998). First ages and magnetosubchrons according to Boulila et al. (Reference Boulila, Galbrun, Miller, Pekar, Browning, Laskar and Wright2011), Miller, Wright & Fairbanks (Reference Miller, Wright and Fairbanks1991) and Miller et al. (Reference Miller, Mountain, Browning, Kominz, Sugarman, Christie-Blick, Katz and Wright1998), following depths, ages and magnetosubchrons according to this study.
The combination of offsets between the different magnetochrons and the differential offsets between the Mi-events indicate that a simple shift in magnetochrons would not provide a solution; we have therefore carefully studied the inclination data of Site U1318. From c. 160 mcd downwards, the divergence between the bio- and isotope stratigraphic tie-points and the age model of Quaijtaal et al. (Reference Quaijtaal, Donders, Persico and Louwye2014) becomes largest. When restudying the inclination data, it appeared that the signal of the discrete samples of Louwye et al. (Reference Louwye, Foubert, Mertens and Van Rooij2008) in the interval between 207.15 and 190.15 mcd is quite irregular (see their fig. 8). Louwye et al. (Reference Louwye, Foubert, Mertens and Van Rooij2008) interpreted this interval as three magnetosubchrons: C5Bn.1r, C5Bn.2n and C5Br. The interpretation of C5Bn.2n is based on two discrete measurements. However, the surrounding measurements are relatively scattered, whereas in most subchrons the measurements tend to cluster together. We have therefore reinterpreted this interval as being one magnetosubchron: C5Br. This reinterpretation is compatible with the isotope stratigraphy, and the longer duration of magnetosubchron C5Br fits better with the sediment thickness (17 m). From the base of the composite record up to 207.15 mcd, the magnetostratigraphy was therefore shifted by two chrons towards a younger age; from 190.15 up to 154.85 mcd, age was shifted one chron upwards.
The interval between 154.85 and 130.02 mcd proves to be more difficult due to the lack of discrete samples, as well as the hiatus at 133.2 mcd. We have therefore linearly interpolated between the two assigned reversals encompassing this interval.
For the interval from 130.02 mcd to the top of the composite record, we used the same approach as in the interval below 154.85 mcd. Re-evaluation of the inclination data showed that the short reversed interval between 124.32 and 122.81 mcd is misinterpreted. The discrete measurements show no sign of any reversals, and most on-board measurements point towards a normal polarity. We therefore interpret the interval from 130.02 to 116.51 mcd as one instead of three magnetosubchrons. Aided by the location of the Mi-events and their corresponding magnetosubchrons, from 130.02 to 116.51 mcd age was shifted one magnetochron towards older age, and from 116.15 mcd to the top of the composite record age was shifted two magnetochrons towards older age. This implies that the interval between 154.85 and 130.02 mcd contains three magnetosubchrons (Fig. 5).
An overview of all tie-points can be found in Table 3 and a visualization of the new age model is displayed in Figure 5.
Magnetostratigraphic and biostratigraphic tie-points for Site U1318. Tie-points in bold are used to generate the age model.
5. Discussion and conclusions
With the aid of stable isotope stratigraphy we have been able to refine the age model for IODP Site U1318. The sampled interval now encompasses an age of 12.75–16.60 Ma; the average sampling resolution of our stable isotope record with the new age model is c. 29 ka. The estimated duration of the hiatus over 104.60–104.31 mcd is estimated at c. 32 ka. Due to the uncertainties in the magnetostratigraphy over 154.85–130.02 mcd, the duration of the hiatus at 133.20 mcd cannot be determined and the model was linearly interpolated up- and downwards.
Sedimentation rates vary from c. 1.2 to 16.9 cm ka–1. The record can roughly be divided into five sections. The first section, 241.45–207.15 mcd, has an average sedimentation rate of 6.03 cm ka–1 and ends shortly after Mi-2. Sedimentation rates decrease to 2.6 cm ka–1 on average over 207.5–178.35 mcd. Between 178.35 and 160.15 mcd rates increase to c. 6.9 cm ka–1. For the fourth section (160.15–130.02 mcd) the average sedimentation of c. 3 cm ka–1 is a rough estimate, since the age model was linearly interpolated between two recognizable magnetic reversals. The last section (130.02–91.44 mcd) has relatively constant sedimentation rates of c. 4.45 cm ka–1.
The Porcupine Basin benthic foraminiferal stable isotope records show clear imprints of globally recognized stable isotope events and have a relatively large temporal range; they will therefore be significant for further interbasinal comparisons and correlations with surface ocean parameters to elucidate triggers and drivers for Miocene climatic transitions. Some of these stable isotope events were used as guidelines for the correct identification of palaeomagnetic reversals. The other points were compared to the astronomically tuned high-resolution record from IODP Site 1338, and used to improve the age model for IODP Site U1318 further. Furthermore, the oxygen isotope measurements on benthic foraminifer species Cibicidoides pachyderma and Uvigerina sp. correlate very well. Measurements on Uvigerina sp. can be converted to Cibicidoides using the following formula: δ18OCibicidoides pachyderma = 0.984 × δ18OUvigerina sp. – 0.296‰ (R 2 = 0.94), prediction error 0.096‰ (1σ, root mean square error of prediction).
Acknowledgements
The data used for this study can be found in online Supplementary Table S1 (available at http://journals.cambridge.org/geo). The samples for this study were provided by the Integrated Ocean Drilling Program. This work was supported by the Research Foundation-Flanders (FWO) under project number G.0179.11N. Ph.C. thanks the Hercules Foundation Flanders for support of the Stable Isotope facility. The advice of Dr An Holbourn on Miocene stable oxygen isotopes was very valuable for the improvement of the age model. Niels de Winter is thanked for his support during the isotopic measurements in Brussels. Walter Hale and Alex Wülbers kindly supported W.Q. and S.L. during sampling at the Bremen Core Repository. The authors thank Dirk Munsterman and one anonymous reviewer for their useful comments, which have improved the manuscript.
Declaration of interest
The authors declare no conflicts of interest.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/S0016756816001278.
