Sandbian) K-Bentonites in Oslo, Norway T ⁎ Eirik G

Palaeogeography, Palaeoclimatology, Palaeoecology 520 (2019) 203–213

Contents lists available at ScienceDirect

Palaeogeography, Palaeoclimatology, Palaeoecology

journal homepage: www.elsevier.com/locate/palaeo

A new age model for the Ordovician (Sandbian) K-bentonites in Oslo, Norway T ⁎ Eirik G. Balloa, , Lars Eivind Auglanda, Øyvind Hammerb, Henrik H. Svensena a Centre for Earth Evolution and Dynamics (CEED), University of Oslo, Pb. 1028, 0316 Oslo, Norway b Natural History Museum, University of Oslo, Pb. 1172, 0318 Oslo, Norway

ARTICLE INFO ABSTRACT

Keywords: During the Late Ordovician, large explosive volcanic eruptions deposited worldwide K-bentonites, including the Chronostratigraphy Millbrig and Deicke K-bentonites in North America and the Kinnekulle K-bentonite in Scandinavia. We have Sandbian-Katian boundary studied a classical locality in Oslo containing one of the most complete sections of K-bentonites in Europe. U-Pb dating In a 53 m section of Sandbian age, we discovered 33 individual K-bentonite beds, the most notable beds being Milankovitch the Kinnekulle and the upper Grimstorp K-bentonite. Magnetic susceptibility (MS) measurements on two in- Age model tervals show signiﬁcant periodicity peaks interpreted as Milankovitch cycles and thus astronomically forced Kinnekulle changes in sediment supply and composition. These cycles ﬁt remarkably well with both the expected Milankovitch periodicities for the Ordovician as well as the radiometric ages presented in this study and may represent one of the most convincing demonstrations of Milankovitch cycles from the lower Paleozoic so far. Five of the K-bentonites have been dated by high-precision chemical abrasion-thermal ionization mass spectrometry (CA-TIMS) U-Pb zircon geochronology, where the Kinnekulle K-bentonite gives an age of 454.06 ± 0.43 Ma. We have integrated the new data and calculated an age model showing the sedimentation rates through the section and thereby the ages of each of the 33 K-bentonites. Using the age model, we further present a new age for the Sandbian-Katian stage boundary. The section in Oslo provides the highest resolution window into the Upper Ordovician K-bentonite succession so far and helps shed more light on the chronology of one of the most intense volcanic periods of the Paleozoic and the relationship with the global carbon cycle changes that followed.

1. Introduction recent advances in geochronological methods provide the possibility to obtain precise age constraints (Condon et al., 2007; Mattinson, 2005; The Ordovician was one of the most environmentally, biologically, Schmitz and Schoene, 2007). and climatologically dynamic periods in Earth history with events such Subduction-related island arc volcanism during the Ordovician de- as the first occurrence of land plants (Lenton et al., 2012), and the posited volcanic tephras, now preserved as K-bentonites (Huff, 2016; extensive Great Ordovician Biodiversification Event (GOBE) that led to Huff et al., 2010). K-bentonites represent excellent stratigraphic mar- a surge of marine life (Servais et al., 2009; Webby et al., 2004). kers of the Phanerozoic, as they cover widespread areas (up to Moreover, a general long-term cooling trend is documented through the 2.2 × 106 km2 in Europe and 6.9 × 105 km2 in North America; Huff entire period (Nardin et al., 2011), culminating in the Hirnantian gla- et al., 2010). Sedimentary successions can record environmental, bio- ciation at the end of the period—accumulating ice sheets on the logical, and climatological data—providing a timeline of geochemical Gondwanan continent and consequently lowering of the global sea level information on these events, with interlayered K-bentonite beds (Armstrong, 2007; Brenchley et al., 1994; Herrmann et al., 2004). The working as age markers. Radiometric dating of well-preserved zircons first of the ‘Big Five’ mass extinctions of the Phanerozoic, killing about in K-bentonites is therefore an important tool in refining the geological 85% of marine species, ended the Ordovician Period (Sheehan, 2001). timescale and defining large-scale geochemical perturbations. Still, the causes of these events remain elusive (Munnecke et al., 2010). Milankovitch cycles are associated with astronomically forced The general lack of precise age determinations is a key factor in un- changes in sediment supply, composition and climate change (Elrick certainties in the cause-and-effect relationships responsible for these et al., 2013; Fang et al., 2016; Sutcliffe et al., 2000; Williams, 1991). To intervals of significant global change (Cramer et al., 2015). However, identify astronomically forced cycles, magnetic susceptibility (MS) is

⁎ Corresponding author. E-mail address: [email protected] (E.G. Ballo). https://doi.org/10.1016/j.palaeo.2019.01.016 Received 1 October 2018; Received in revised form 10 January 2019; Accepted 10 January 2019 Available online 01 February 2019 0031-0182/ © 2019 Elsevier B.V. All rights reserved. E.G. Ballo, et al. Palaeogeography, Palaeoclimatology, Palaeoecology 520 (2019) 203–213

Fig. 1. (A) Southern Scandinavia with box showing the Oslo Region and logs from the Vollen, Sinsen and Kinnekulle localities show the Kinnekulle, lower and upper Grimstorp K-bentonite beds illustrating an apparent westward increase in sedimentation rate. The map is adapted from Svensen et al. (2015). (B) Simplified geological map of the Oslo Region showing the studied localities at Sinsen (including the Storoveien locality; 59°56′13″ N, 10°47′13″ E), Carl Berner (59°55′39″ N, 10°46′48″ E) and Vollen (59°48′11.4″ N, 10°29′12.6″ E). The map is modified from Lutro and Nordgulen (2008). (C) Satellite overview of the Sinsen area showing the main Sinsen section (black box), the stratigraphic up direction (arrow) and strike and dip of the bedding. The image is adapted from Google Maps (2017). frequently used, as it is a low-cost logging parameter that allows for 2. Upper Ordovician stratigraphy of the Oslo Region field measurements (Svensen et al., 2015; Fang et al., 2016; Plado et al., 2016). Climate-driven changes in weathering, erosion and sea level lead In the Oslo Region, the Ordovician successions are characterized by to a variation in sediment influx, reflected in the concentration of for epicontinental, subtidal marine depositional environments, as de- instance iron (Ellwood et al., 2000). Combining high-precision radio- scribed in classical contributions by Kjerulf (1857) and Kiær (1897), metric ages with cyclostratigraphic spectral analyses may test whether and summarized by Nielsen (2004), Owen et al. (1990), and Worsley sediments provide a record of astronomical forcing. and Nakrem (2008). The Middle Ordovician stratigraphy is dominated The first combined radiometric dating and cyclostratigraphic study by shales of the Tøyen and Elnes formations separated by the massive of K-bentonites in Norway was conducted by Svensen et al. (2015) at limestones of the Huk Formation (Bruton et al., 2010; Fig. 2). The Vollen south of Oslo (Fig. 1B). Svensen et al. (2015) dated two of the Upper Ordovician is characterized by the alternation of shales (e.g. key K-bentonites (the Kinnekulle and the upper Grimstorp K-bentonite) Arnestad and Nakkholmen formations) and nodular and tabular lime- and identified cyclicities interpreted to represent the short eccentricity stones (e.g. Vollen and Frognerkilen formations; Bjørlykke, 1974; (100 kyr) and obliquity (30.3 kyr) components. However, as the out- Bruton et al., 2010; Möller and Kvingan, 1988; Fig. 2). The Ordovician crop at Vollen is limited to three K-bentonites, a high-resolution age successions in Oslo are found within the Oslo Graben, a Carboniferous model was not calculated. to Permian rift structure extending about 200 km from the Skien-Lan- The aim of this study is to provide a new age model for this part of gesund area in the south to the Mjøsa district in the north (e.g. the Ordovician stratigraphy, based on a section in Oslo. We have re- Neumann et al., 1992; Størmer, 1953; Sundvoll and Larsen, 1994). visited the classic locality of Hagemann and Spjeldnæs (1955), re- The K-bentonites in this study are all located in the Arnestad vealing 33 K-bentonite beds of which 11 are not previously reported. Formation, predominantly consisting of thick shale beds (30–40 cm), New high-precision ages of the K-bentonites may help constrain the interlayered by thin bedded or nodular horizons of early diagenetic relationships between large explosive eruptions and environmental calcareous limestone rarely exceeding 10–20 cm in thickness. In the changes. upper part of the formation, the limestone horizons develop from

204 E.G. Ballo, et al. Palaeogeography, Palaeoclimatology, Palaeoecology 520 (2019) 203–213

Fig. 2. Global distribution of Middle and Upper Ordovician K-bentonites modified from Huff (2008) using ages from Gradstein et al. (2012). nodular to massive (Owen et al., 1990). A westward shallowing en- two beds originated from different arc settings. The large magnitude of vironment is suggested by faunal and sedimentological indicators these Ordovician volcanic eruptions has brought up questions about (Hansen and Harper, 2006), which is reflected in a westward increase their potential climatic and environmental effects. Perrier et al. (2012) in sedimentation rate (Fig. 1A). suggested that the Kinnekulle K-bentonite was associated with re- The position of the Sandbian-Katian boundary in the Oslo Region organization of ostracod assemblages and strong, long-term perturba- has been discussed by several authors based on trilobite occurrences tions of diversity and abundance. The climatic effects of these eruptions (e.g. Owen et al., 1990), chitinozoan biostratigraphy (Grahn et al., are, however, poorly resolved (Buggisch et al., 2010; Herrmann et al., 1994), conodont biostratigraphy (Hamar, 1966), and U-Pb dating 2010, 2011). Therefore, precise ages for these K-bentonites, especially (Svensen et al., 2015). Bergström et al. (2011) placed the Sandbian- the Kinnekulle K-bentonite, are necessary in order to define the accu- Katian boundary at the onset of the δ13C excursion known as the Gut- rate temporal scale needed for understanding the links between large tenberg isotope carbon excursion (GICE), which in the Oslo Region volcanic eruptions and their impacts. coincides with the Arnestad-Frognerkilen formation boundary. This boundary marker is adopted in this study. The boundary has been estimated at 453.0 ± 0.7 Ma by Cohen et al. (2013). Svensen et al. 4. Material and methods (2015) also suggested an age for the boundary, however, the interpretation required extrapolating sedimentation rates over an interval of We have logged 53 m of stratigraphy at the Sinsen section. The 18.7 m, giving a potentially large error. section is located around 600 m along the railroad tracks from the Grefsen Station (Fig. 1C). A total of 33 K-bentonites were identified and seven of these were processed for zircons. The K-bentonite ages were 3. The Ordovician K-bentonites analyzed using chemical abrasion-thermal ionization mass spectrometry (CA-TIMS). The Sinsen section in Oslo is one of the most complete sections of Magnetic susceptibility (MS) was measured in the field using a Ordovician K-bentonites in Europe, providing unique possibilities to handheld Terraplus KT-10 MS meter with a sampling diameter of 6 cm − study the interplay and timing of geochemical events of the Late and a lower detection limit of 10 6 SI units. Three measurements were Ordovician. First described by Hagemann and Spjeldnæs (1955), the taken at each level and averaged to reduce noise. Before and after each section was divided into four complexes of K-bentonite beds comprising measurement the instrument was calibrated in free air. Details of the a total of 24 beds, ranging in thickness from 0.5 to 65 cm. Bergström MS techniques used are presented in Appendix A. et al. (1995) correlated the K-bentonites at Sinsen with other localities For zircon extraction, the samples were washed and sieved before in Northern Europe (Sweden, Denmark, Norway, Iceland, Finland, Es- separation of heavy minerals using standard heavy liquid techniques at tonia, Latvia, and Lithuania) and introduced a new nomenclature that the University of Oslo. Zircons were selected under an optical micro- enabled correlation across regions. The paleogeographic reconstruction scope, annealed and chemically abraded (Huyskens et al., 2016; in Fig. 3 shows the position of these regions during the Late Ordovician. Mattinson, 2005). The zircon grains chosen for analyses were spiked Bergström et al. (1995) identified the most prominent bed at Sinsen with a tracer that has been calibrated to the EARTHTIME (ET) 100 Ma as the well-studied Kinnekulle K-bentonite that can be traced across solution (appendix A) and measured on a Finnigan MAT 262 thermal many localities in Northern Europe (Bauert et al., 2014; Jürgenson, ionization mass spectrometer (TIMS). Further details are presented in 1958; Kiipli et al., 2014b; Svensen et al., 2015). A possible common appendix A. volcanic source for the Kinnekulle and Millbrig K-bentonites (North The spectral analysis of the MS data used the REDFIT procedure America) was proposed by Huff et al. (1992). However, this transat- (Schulz and Mudelsee, 2002) implemented in PAST, version 3.13 lantic connection has later been discarded by several authors (Haynes (Hammer et al., 2001), which allows for uneven spacing of data points et al., 1995; Min et al., 2001; Sell and Samson, 2011) based on non- and includes a significance test using the null hypothesis of auto- overlapping ages and geochemical fingerprinting, suggesting that the correlated noise. We used the ‘Welch’ window and the 95% significance

205 E.G. Ballo, et al. Palaeogeography, Palaeoclimatology, Palaeoecology 520 (2019) 203–213

Fig. 3. Paleogeographic map with stars illustrating where the K-bentonite occurrences shown in Fig. 2 were located at 450 Ma. The map is modified from Torsvik and Cocks (2016). level was calculated using Monte Carlo simulation. Potential problems limestone layers. Herein, the base of the Frognerkilen Formation was with significance testing using a red noise model are well known (e.g. set at the base of the fourth massive limestone layer below the rubbly Smith and Bailey, 2017) and the significance level must therefore be limestone. regarded as indicative only. For calculating an age model, we adopted the Bayesian method of De Vleeschouwer and Parnell (2014) implemented in the program 5.2. U-Pb zircon geochronology Bchron. A total of five K-bentonites from the Sinsen section were dated. The ages for each individual K-bentonite are given as weighted average 5. Results 206Pb/238U-ages based on the concordant youngest zircon analyses, following the interpretation strategy of e.g. Schoene et al. (2010) for 5.1. K-bentonite stratigraphy complex age distributions in volcanic ash layers. Errors are reported at 2σ as ± x/y/z, where x includes only the analytical uncertainties, y is A total of 33 K-bentonite beds were discovered at the Sinsen lo- the error including the analytical and tracer calibration uncertainties cality. A few meters below the base of the lowermost K-bentonite (bed and z is the error including the analytical uncertainties, tracer un- #1, 15 mm thick; Fig. 4) an anticlinal fold leaves the Vollen-Arnestad certainties and uncertainty in the decay constant. formation boundary unobservable. In the lowest 5 m of the logged Bed #1 (EB16-S-0) shows a spread in single zircon 206Pb/238U-ages section five K-bentonite beds were found, evenly spaced with 1–2cm from 461.43 to 456.67 Ma, with a dominant equivalent and concordant thickness. K-bentonite #6 is located at 10.00 m, typically eroded a few (within errors) population of 6 grains yielding a weighted average centimeters into the outcrop. A 3 m thick sill is present at 10.90 m in the 206Pb/238U-age of 456.81 ± 0.30/0.38/0.61 Ma (2σ; MSWD = 0.070) section without visible metamorphic effects in the adjacent nodular interpreted to reflect the emplacement age of the K-bentonite (Figs. 6, limestone. Bed #7 and #8 are both 1 cm thick, positioned at 15.65 m A1E). and 16.75 m respectively. Following bed #8, there is a 7 m interval of The interpreted emplacement age of Bed #16 (EB16-S-28.63) is shale and nodular limestone, without K-bentonites (Fig. 4). 456.13 ± 0.39/0.45/0.66 Ma (2σ; MSWD = 0.95), based on the two A complex of closely spaced K-bentonites follows, referred to as the youngest equivalent and concordant grains (Fig. 6). Individual grains Grefsen K-bentonite complex by Bergström et al. (1995). The complex is from the sample show a spread in single zircon 206Pb/238U-ages from separated into a lower (beds #9–#14) and upper part (beds #15–#26; 461.6 to 456.11 Ma. The discordant older zircons (no. 1–8) are likely Fig. 4), separated by a 1.83 m K-bentonite hiatus. The entire complex of antecrysts (Figs. 6, A1D). This is supported by CL-imaging of zircon 12 K-bentonite beds spans 3.8 m. One of the thinnest beds at the lo- crystals from the sample that reveal complex zonation with resorption cality, bed #27, was discovered 1.30 m above bed #26, with a thickness and truncating overgrowth (Fig. A1D). of around 0.1 cm. Above this, the two beds #28 and #29 have pre- Bed #30 (EB16-S-39.60, the Kinnekulle K-bentonite) shows a spread viously been referred to as the Sinsen K-bentonite beds by Bergström in single zircon 206Pb/238U-ages from 458.20 to 453.89 Ma, with the et al. (1995), however we chose not to adopt this name to avoid con- two youngest concordant (within errors) grains yielding a weighted fusion. average 206Pb/238U-age of 454.06 ± 0.43/0.49/0.68 Ma (2 sigma; The Kinnekulle K-bentonite is positioned at 39.60 m and is the MSWD = 0.33) interpreted to reflect the eruption and depositional age thickest bed (35 cm) (Fig. 5). At 42.23 and 42.71 m, the lower (#31) of the K-bentonite (Fig. 6). The nine older grains are likely antecrysts, and upper (#32) Grimstorp K-bentonites were found (Fig. 5), with which is supported by CL-imaging confirming the presence of grains thicknesses of 5 and 13 cm respectively. Almost at the top of the Ar- with several growth domains (Fig. A1C). nestad Formation at 51.68 m, the uppermost K-bentonite of the section Bed #32 (EB16-S-42.71) yielded a spread in single zircon (bed #33) has a thickness of 2 cm. 206Pb/238U-ages from 458.20 to 453.48 Ma (Figs. 6, A1B). The youngest The base of the Frognerkilen Formation was defined following concordant analysis yields an age of 453.48 ± 0.74/0.78/0.91 Ma Owen et al. (1990). At the hypostratotype at Rabben in Asker south of (2σ), which is considered to represent the eruption and depositional age Oslo, the Arnestad-Frognerkilen formation boundary is set at the base of of the bentonite. The older zircons likely represent antecrysts, which is the fourth double massive limestone layer below the characteristic also supported by CL-imaged zircons, some of which show the presence rubbly limestone. At Sinsen, it appears as if the individual double of resorption textures and multiple growth domains (Fig. A1B). massive limestone layers have blended into four single massive Bed #33 (EB16-S-51.68) shows a spread in single zircon 206Pb/238U-

206 E.G. Ballo, et al. Palaeogeography, Palaeoclimatology, Palaeoecology 520 (2019) 203–213

Fig. 4. Log of the Sinsen section showing the distribution of K-bentonites (black lines) in the Arnestad Formation. Grey areas = shale; white ellipses = limestone nodules (not to scale). The dashed line in the upper part of the log shows the interpreted position of the Arnestad-Frognerkilen formation boundary, herein assumed to approximately coincide with the Sandbian-Katian boundary. Oslo-Asker stratigraphy (left) is based on Bruton et al. (2010). The interval column illustrates the complexes of beds that are discussed in this study. Figure illustrates the stratigraphic positions of the other ﬁgures in this study.

207 E.G. Ballo, et al. Palaeogeography, Palaeoclimatology, Palaeoecology 520 (2019) 203–213

Fig. 5. (A) The upper part of the Sinsen section (ca. 37-44 m in the log). The outcropping K-bentonites in the interval are shown with white arrows. The dashed lines show the lower and upper boundary of the MS measured interval. (B) Bed #33 is shown in the attached image as a thin cut in the stratigraphy. The simplified log is shown with green boxes illustrating the stratigraphic position of the images. ages from 457.38 to 448.66 Ma, where the youngest, strongly reversely a new constraint of the age span of the upper Sandbian volcanism. discordant grain clearly show that some analytical error occurred Furthermore, our new U-Pb ages of the Kinnekulle K-bentonite (bed during analysis. The youngest concordant age that gives a 206Pb/238U- #30) can be used for better correlation with other localities. age of 452.76 ± 0.86/0.89/1.0 Ma (2σ) is interpreted to reflect the Bed #1 is the oldest Ordovician K-bentonite analyzed in this study, emplacement age of the K-bentonite (Figs. 6, A1A). The older analyses with an age of 456.81 ± 0.30 Ma. Bed #16 represents the lowest dated are interpreted to represent antecrysts. bed of the #15–#26 interval. The interval is characterized by high- frequency volcanic eruptions—preceded, and followed by, prolonged 5.3. Magnetic susceptibility in the Arnestad Formation periods of non-deposition of ash. Our age of bed #16, 456.32 ± 0.39 Ma, is the first radiometric dating of any K-bentonite The lower MS measured interval (28.75–32.00 m) consists of 11% bed in the Grefsen K-bentonite complex. The age of the bed provides a K-bentonite and 89% shale and limestone in terms of stratigraphic precise age constraint for the onset of this high-frequency volcanic in- thickness. The K-bentonites are distributed evenly across the interval terval (Fig. 6). Bergström et al. (1995) suggested that K-bentonites from and yield considerable low MS excursions, compared to the relatively this part of the stratigraphy are equivalent to K-bentonites in Estonia stable signal for shales, as seen in Fig. 6. The spectrum in Fig. 7A shows and Sweden, which can be tested by further U-Pb geochronology. only two peaks above the 95% significance curve: (i) 3.6 cycles/m, and The Kinnekulle K-bentonite is one of the most studied Ordovician K- (ii) 4.5 cycles/m. The peak at 3.6 cycles/m has the highest intensity. bentonites worldwide (Huff et al., 2010), and has been subject to many The peak at 4.5 cycles/m is the second strongest (Fig. 7A). dating efforts, yet its precise age is still under discussion. Huff (2008) The upper MS measured interval (39.00–42.50 m) spans 3.50 m in summarized at least 14 separate attempts at dating the K-bentonite the stratigraphy and is confined by the two dated beds #30 (Kinnekulle using various methods, and many have followed in the years since. K-bentonite, 39.60 m) and #32 (upper Grimstorp, 42.71 m). From the Bauert et al. (2014) reported laser ablation-inductively coupled mass base of bed #30 to the top of bed #32 the interval consists of 12% K- spectrometer (LA-ICPMS) ages of K-bentonites from Mossen in Sweden bentonite and 88% shale and limestone, however, in contrast to the first (453.6 ± 6.6 Ma) and northern Estonia (454.9 ± 4.9 Ma) that are MS-measured interval, only three K-bentonites are present (Fig. 6). The convincingly correlated to the Kinnekulle bed (Bergström et al., 2016). three most dominant low excursions in the curve coincide with the K- Svensen et al. (2015) reported a CA-TIMS U-Pb zircon weighted average 206 238 bentonites (Fig. 6). The REDFIT spectrum in Fig. 7B shows two peaks Pb/ U-age of 454.52 ± 0.50 Ma ( ± 0.70 Ma including tracer ca- beyond the 95% significance line: (i) 2.7 cycles/m, and (ii) 6.3 cycles/ libration errors) from Vollen south of Oslo in Norway, considered as the m. most precise age published (Bergström et al., 2016). As this study and that of Svensen et al. (2015) have used the same U-Pb tracer the ages 6. Discussion can be directly compared after applying the same spike calibration as implemented in this study (Appendix A3). In fact, our new age for the Bergström et al. (1995) and Hagemann and Spjeldnæs (1955) re- Kinnekulle K-bentonite of 454.06 ± 0.43 Ma ( ± 0.49 Ma including ported a total of 24 K-bentonite beds at Sinsen. We have identified 11 tracer calibration errors) is equivalent within error to that of Svensen new beds nearly doubling the stratigraphic span of the already re- et al. (2015) but with improved precision. Further, as our new age does 206 238 markable K-bentonite outcrop at Sinsen. The locality therefore stands as not overlap with the established weighted average Pb/ U-age for a uniquely preserved Ordovician K-bentonite section where the number the Millbrig K-bentonite (452.86 ± 0.29 Ma (CA-TIMS); Sell et al., — of data points allow for age modeling with a high resolution. In the 2013) we here support the hypothesis that the two K-bentonites are following, we discuss the U-Pb ages, the cyclostratigraphic results and not identical. the robustness and implications of the age model. The upper Grimstorp K-bentonite has also been subject to several dating attempts. Sell et al. (2013) reported a weighted average 206 238 6.1. K-bentonite geochronology Pb/ U-age of 454.41 ± 0.17 Ma (CA-TIMS) for the K-bentonite from the Billegrav-2 drill core in Bornholm, Denmark. The dated K- We have dated beds #1, #16 and #33 for the first time, allowing for bentonite was assumed to represent the Kinnekulle K-bentonite,

208 E.G. Ballo, et al. Palaeogeography, Palaeoclimatology, Palaeoecology 520 (2019) 203–213

Fig. 6. The Sinsen log with the resulting age (left) and MS (right) data. Geochron plots for the ﬁve dated K-bentonite beds show individual zircon analyses (vertical bars) with concordant and equivalent zircon analyses covered by a horizontal grey bar illustrating the calculated mean of each sample. MSWD = Mean Square Weighted Deviation. In the log black lines = K-bentonite beds; grey areas = shale; red dashed line = interpreted position of the Arnestad-Frognerkilen formation boundary. however Kiipli et al. (2014b) suggest that the dated K-bentonite re- of bed #1 (456.81 ± 0.30 Ma), the entire volcanic activity recorded in presents the Grimstorp K-bentonites (preserved as one single bed). This the Ordovician at Sinsen has a duration of 4.05 ± 0.91 myr. interpretation was supported by Bergström et al. (2016). Our new age Taking into account the presence of antecrysts in all the analyzed of 453.48 ± 0.74 Ma indicates that the upper Grimstorp K-bentonite is samples in this study, we have documented a substantial spread in 0.93 Ma younger than the published age from Sell et al. (2013). Again, single zircon ages that could indicate a long-lived volcanic arc system using the new tracer calibration (Appendix A3), the age of the upper with sporadic but large scale explosive eruptions. Kiipli et al. (2014a) Grimstorp K-bentonite from Svensen et al. (2015) is equivalent within argued that the Kinnekulle K-bentonite most probably originated from a error with our new age. large and slowly cooling magma chamber, supporting this argument. The stratigraphically youngest K-bentonite at Sinsen is dated to We see a comparable distribution of single zircon ages in our samples, 452.76 ± 0.86 Ma (bed #33). This K-bentonite is signiﬁcant due to the which is underlined by the overlapping ages of one of the oldest ana- proximity to the Sandbian-Katian boundary, which we have interpreted lyzed grains (457.38 ± 0.93 Ma) in the youngest dated sample (bed to be 0.82 m above the K-bentonite. Moreover, combined with our age #33) and the age of the oldest K-bentonite sample (bed #1;

209 E.G. Ballo, et al. Palaeogeography, Palaeoclimatology, Palaeoecology 520 (2019) 203–213

Fig. 7. Time series REDFIT spectrum of (A) the lower MS measured interval and (B) the upper MS measured interval at Sinsen. The interpreted periods for the peaks are shown next to the grey vertical bars.

456.81 ± 0.30 Ma). This overlap may be explained by a similar source is assumed that the four peaks in Fig. 7B represent the expected periods volcano for the K-bentonites in Oslo. (100, 30.3, 19.2 and 16.3 kyr) from Berger et al. (1992), as indicated by the ratios between the observed peaks, the averaged sedimentation rate for this interval inferred from the periodicities of the four peaks is 6.2. Cyclostratigraphy 1.12 cm/kyr which is consistent within the age model confidence intervals (further discussed below). From this we propose that the first The spectra for the two selected intervals in the Sinsen section peak at 0.8 cycles/m (1.25 m/cycle) represents a period of 111.6 kyr, (Fig. 7) show several peaks above the 95% significance level. Ideally, thus in good compliance with Ordovician eccentricity. The peak at the ratios of Ordovician orbital frequencies from e.g. Berger et al. 2.7 cycles/m (0.37 m/cycle) corresponds to a period of 33.1 kyr, fitting (1992) should be recovered in the depth domain, if the periodicity is well with the expected period of obliquity (30.3 kyr). The peak at indeed orbitally forced, the sediments are undisturbed, and with a 5.1 cycles/m (0.20 m/cycle) corresponds to a period of 17.5 kyr, fitting constant sedimentation rate (Rodionov et al., 2003). For an age of with the expected period of long precession (19.2 kyr). Furthermore, 450 Ma, Berger et al. (1992) and Rodionov et al. (2003) suggested the the peak at 6.3 cycles/m (0.16 m/cycle) corresponds to a period of following orbital periodicities: precession 16.3 and 19.2 kyr, obliquity 14.2 kyr, and is close to the expected short precession component of 30.3 (strong) and 36.8 (weak) kyr. Short eccentricity cycles drift be- 16.3 kyr. The expected Ordovician cycles represent estimates, and also tween 100 kyr and 125 kyr through time (Berger et al., 1992). the observed periodicities are associated with both measurement errors Following Berger et al. (1992) and Rodionov et al. (2003), the and analytical errors. Therefore the correspondence between field-de- strongest expected signal for the Ordovician is eccentricity (period of rived periods and expected periods (Berger et al., 1992) may vary. ca. 100 kyr), however, interpreting the 3.6 cycles/m peak in the lower Our Milankovitch interpretation of the cyclicity also fits with the interval (28.75–32.00 m; Fig. 7A) as eccentricity yields a sedimentation results of Svensen et al. (2015) from the corresponding stratigraphic rate of 0.28 cm/kyr—unreasonably low compared to what previous interval at Vollen where both eccentricity and obliquity were detected. studies on the Arnestad Formation suggest (Svensen et al., 2015). The However, our data now suggest that the periodicities and sedimentation measured interval (3.25 m) was REDFIT-processed as two 1.63 m ‘seg- rate (0.92 cm/kyr) presented by Svensen et al. (2015) from Vollen ments’, thus the stratigraphic span of each segment is very close to the should be revised. As shown in Fig. 1A, the difference in stratigraphic lower limit needed to record eccentricity signals. We here argue that position (neglecting the thickness added by K-bentonite) between the this could be a possible explanation for why eccentricity has not been upper Grimstorp and Kinnekulle K-bentonite at Vollen (6.89 m) com- recorded in the lower MS measured interval. If the peak at 3.6 cycles/m pared to Sinsen (2.71 m) implies that the sedimentation rate at Vollen instead is tuned to the 30.3 kyr period (obliquity) expected for the should be 2.5 times higher than at Sinsen. Taking a sedimentation rate Ordovician, the sedimentation rate for the interval yields 0.73 cm/kyr. of 2.8 cm/kyr into account, the two peaks in Fig. 5 from Svensen et al. From this, the peak at 4.5 cycles/m corresponds to a periodicity of (2015) yield periodicities of 35.7 kyr and 9.9 kyr respectively. Our new 37.9kyr, which may be interpreted as the second obliquity periodicity data therefore suggest that the peaks found by Svensen et al. (2015) at in the Ordovician, at 36.8 kyr. The two peaks in Fig. 7A can therefore be Vollen may rather represent obliquity and precession. interpreted as the strong and weak obliquity periods. A possible source of inaccuracy lies in the high amount of K-bentonite through the measured interval (11 beds in total) that could disrupt the signal of the 6.3. Age model and Arnestad Formation sedimentation rates shortest Milankovitch periods (precession) from the sediments. The REDFIT spectrum for the upper MS measured interval The volcanic eruptions that deposited the individual K-bentonites (39.00–42.50 m) is presented in Fig. 7B and clearly shows two peaks may have lasted on a scale of weeks (Bergström et al., 1995), whereas a above the 95% significance curve in addition to well-defined peaks at similar thickness of shale would have needed thousands of years to 0.8 and 5.1 cycles/m, not reaching the significance level. By using the accumulate. Therefore, the cumulative stratigraphic thickness of the K- sedimentation rate at 36.0 m derived from the radiometric dates and bentonites was subtracted and approximated as points in the age model. the age model described below (0.43 cm/kyr) the four peaks in Fig. 7B The relationship between the corrected level in the age model (Fig. 8) represent periods of 290.7, 86.1, 45.6 and 36.9 kyr from left to right and the logged level (Fig. 4) is shown in Table A1. respectively. However, the Bayesian age model-derived sedimentation The red curve in Fig. 8 shows the median Bchron age model. The rate for the upper interval (0.43 cm/kyr) is notably lower than what is gradual upwards decrease in sedimentation rate in the Arnestad For- expected for the Oslo Region (Bjørlykke, 1974). If, on the other hand, it mation, as shown in Fig. 8, may reflect rising relative sea level. This

210 E.G. Ballo, et al. Palaeogeography, Palaeoclimatology, Palaeoecology 520 (2019) 203–213

Fig. 8. The Bayesian Bchron age model for the Sandbian aged Arnestad Formation at Sinsen (red curve). Black curves are the upper and lower limits of 95% confidence interval. The five radiometric dated beds are shown as black dots, with 2σ error bars (horizontal black lines). Interpreted sedimentation rates from this study are included in grey boxes for two approaches: 1) Bchron age model, 2) age difference between dated levels (dashed black lines between the age anchors show linear inter- polations). Ordovician stage boundaries are from the latest revision of the ICS International Chronostratigraphic Chart Cohen et al. (2013), where the Sandbian-Katian boundary is set at 453.0 ± 0.7 Ma. Dashed lines next to the stage boundaries illustrate the uncertainties for the boundary ages. Cyclicity-derived sedimentation rate of 1.12 cm/kyr is shown in blue.

interpretation is supported by Nielsen (2004) who described the so- 6.4. Implications for the Sandbian-Katian boundary called Arnestad Drowning. Moreover, Nielsen (2004) described the onset of a lowstand event at the base of the Frognerkilen Formation, At Sinsen, the Arnestad-Frognerkilen formation boundary is set at which is in good agreement with the increase in sedimentation rate at 52.50 m in the section, which is 0.82 m above the base of the uppermost the top part of the section. Note that the regional versus global nature of dated K-bentonite (bed #33). Herein, the Arnestad-Frognerkilen for- these sea level changes is debated (Bergström et al., 2010, 2011). mation boundary is assumed to coincide with the Sandbian-Katian stage In the upper MS-interval the sedimentation rate proposed by the boundary and the onset of GICE (Bergström et al., 2011). Using the Bayesian age model (0.43 cm/kyr; Fig. 8) does not coincide with the onset of GICE as a proxy for the stage boundary is controversial for cyclicity-derived sedimentation rate (1.12 cm/kyr). We suggest that this several reasons, and thus caution should be exercised in doing so problem is caused by weak constraints on the age model in this interval (Bergström et al., 2016). The absence of a continuous succession of the (Telford et al., 2004). The variation in sedimentation rate between graptolite Diplacanthograptus caudatus in Baltoscandia, which defines radiometric dates in age models can never be accurately estimated by the base of Katian, complicates direct correlation to the GSSP any statistical method. Therefore, the purpose of age models based on (Bergström et al., 2016). In Oslo, GICE has an excellent development radiometric dates should be to provide the most parsimonious (i.e. but often consists of numerous smaller ‘wiggles’ that further challenge simplest possible) hypothesis. A consequence of this philosophy of identifying its precise stratigraphic position (Bergström et al., 2011). parsimony is to assume slowly varying or constant sedimentation rates Based on our new data the age of the stage boundary can be estimated between age anchors. As the cyclicity-derived sedimentation rate (blue by the age model by extrapolating the age model curve up to the in- line in Fig. 8) considers both the age and cyclostratigraphic data, it terpreted boundary position. The age model gives an estimated should be preferred as the most likely sedimentation rate interpretation boundary age of 452.62 ± 0.39 Ma. for the upper interval. This interpretation is further strengthened by the Our new age estimate overlaps within error with the newest age for cyclicity-derived curve fitting within the confidence intervals of the age the boundary, 453.0 ± 0.7 Ma, given by Cooper et al. (2012) for model. GTS2012. However, the boundary age from Cooper et al. (2012) is The presence of nodular early diagenetic limestone throughout the based on a smoothing spline poorly constrained by younger radiometric succession is likely to have an effect on the estimated sedimentation ages. Note that the interpreted position of the Sandbian-Katian rates due to differential compaction (Westphal et al., 2008, 2010). As boundary presented here is debatable and the effect of differential the limestone content generally increases upwards through the section, compaction is a possible source of inaccuracy. A precise age determi- the pre-compaction sedimentation rate is therefore expected to de- nation of the boundary is also affected by how well the onset of GICE crease slightly upwards compared to the post-compaction sedimenta- itself coincides with the base of the Frognerkilen Formation, errors in tion rate shown in the age model (Fig. 8). the age analyses and uncertainties of the age model.

211 E.G. Ballo, et al. Palaeogeography, Palaeoclimatology, Palaeoecology 520 (2019) 203–213

7. Conclusions 1980.tb01026.x. Condon, D., Schoene, B., Bowring, S.A., Parrish, R., McLean, N., Noble, S., Crowley, Q., 2007. EARTHTIME: Isotopic tracers and optimized solutions for high-precision U-Pb A detailed geochronological and cyclostratigraphic study of one of ID-TIMS geochronology. EOS Trans. Am. Geophys. Union 88, 2007. the most complete Ordovician K-bentonite sections in Europe provides Cooper, R.A., Sadler, P.M., Hammer, O., Gradstein, F.M., 2012. In: Gradstein, F., Ogg, J., important constraints on the timing of the high volume explosive Schmitz, M., Ogg, G. (Eds.), The Ordovician period. Elsevier, The Geologic Time Scale, pp. 489–523. https://doi.org/10.1016/B978-0-444-59425-9.00020-2. eruptions leading to massive and worldwide ash deposition. We provide Cramer, B.D., Schmitz, M.D., Huff, W.D., Bergström, S.M., 2015. High-precision U–Pb radiometric ages of five K-bentonites, including the well-studied zircon age constraints on the duration of rapid biogeochemical events during the Kinnekulle K-bentonite. Additionally, MS-based cyclostratigraphy Ludlow Epoch (Silurian Period). J. Geol. Soc. Lond. 172, 157–160. https://doi.org/ shows peaks representing Milankovitch eccentricity, obliquity and two 10.1144/jgs2014-094. De Vleeschouwer, D., Parnell, A.C., 2014. Reducing time-scale uncertainty for the precession components. These Milankovitch interpretations illustrate a Devonian by integrating astrochronology and Bayesian statistics. Geology 42, remarkable fit with both the expected periods for the Ordovician and 491–494. https://doi.org/10.1130/G35618.1. the age data presented in this study. Our cyclostratigraphic data Ellwood, B.B., Crick, R.E., El Hassani, A., Benoist, S.L., Young, R.H., 2000. Magnetosusceptibility event and cyclostratigraphy method applied to marine rocks: therefore represent one of the most precise datasets published to de- detrital input versus carbonate productivity. Geology 28, 1135–1138. https://doi. monstrate Milankovitch periods in the lower Paleozoic. Finally, we org/10.1130/0091-7613(2000)028<1135:MEACMA>2.3.CO;2. have presented an age model for the Arnestad Formation with im- Elrick, M., Reardon, D., Labor, W., Martin, J., Desrochers, A., Pope, M., 2013. Orbital- scale climate change and glacioeustasy during the early Late Ordovician (pre- plications for age determination of the entire sequence of 33 Sandbian Hirnantian) determined from δ18O values in marine apatite. Geology 41, 775–778. K-bentonites, international K-bentonite correlation, and a new age es- https://doi.org/10.1130/G34363.1. timate for the Sandbian-Katian boundary of 452.62 ± 0.39 Ma. This Fang, Q., Wu, H., Hinnov, L.A., Wang, X., Yang, T., Li, H., Zhang, S., 2016. A record of astronomically forced climate change in a late Ordovician (Sandbian) deep marine age is slightly younger than the one proposed by Cooper et al. (2012) sequence, Ordos Basin, North China. Sediment. Geol. 341, 163–174. https://doi.org/ for GTS2012. 10.1016/j.sedgeo.2016.06.002. Supplementary data to this article can be found online at https:// Google Maps, 2017. Map of the Sinsen Area. [online]. URL: (visited on 02/18/2017). https://www.google.com/maps/@59.93607,10.78395,808m/data=!3m1!1e3. doi.org/10.1016/j.palaeo.2019.01.016. Gradstein, F.M., Ogg, J.G., Schmitz, M.D., Ogg, G.M., 2012. The Geologic Time Scale 2012. Elsevierhttps://doi.org/10.1016/C2011-1-08249-8. Acknowledgments Grahn, Y., Idil, S., Østvedt, A.M., 1994. Caradocian and Ashgillian chitinozoan biostratigraphy of the Oslo-Asker and Ringerike districts, Oslo Region, Norway. GFF 116, 147–160. https://doi.org/10.1080/11035899409546178. The authors acknowledge support from The Research Council of Hagemann, F., Spjeldnæs, N., 1955. The Middle Ordovician of the Oslo Region, Norway 6. Norway through its Centres of Excellence funding scheme, project Notes on bentonites (K-bentonites) from the Oslo-Asker District. Nor. Geol. Tidsskr. – 223272. We wish to thank the Norwegian National Rail Administration 35, 29 52. fi Hamar, G., 1966. The Middle Ordovician of the Oslo region, Norway. Nor. Geol. Tidsskr. for providing access to the eld area at Sinsen, and Gunborg B. Fjeld for 46, 27–83. help with sample preparation, Fernando Corfu for valuable insights on Hammer, Ø., Harper, D.A.T., Ryan, P., 2001. PAST: paleontological statistics software zircon chronology and Morgan T. Jones for assistance in the submission package for education and data analysis. Palaeontol. Electron. 4. Hansen, J., Harper, D.A.T., 2006. Brachiopod bio- and ecostratigraphy in the lower part of process. the Arnestad Formation (Upper Ordovician), Oslo Region, Norway. Nor. J. Geol. 86, 403–413. References Haynes, J.T., Melson, W.G., Kunk, M.J., 1995. Composition of biotite phenocrysts in Ordovician tephras casts doubt on the proposed trans-Atlantic correlation of the Millbrig K-bentonite (United States) and the Kinnekulle K-bentonite (Sweden). Armstrong, H.A., 2007. On the cause of the Ordovician glaciation. In: Williams, M., Geology 23, 847–850. https://doi.org/10.1130/0091-7613(1995) Haywood, A.M., Gregory, F.J., Schmidt, D.N. (Eds.), Deep Time Perspectives on 023<0847:COBPIO>2.3.CO;2. Climate Change: Marrying the Signal from Computer Models and Biological Proxies. Herrmann, A.D., Haupt, B.J., Patzkowsky, M.E., Seidov, D., Slingerland, R.L., 2004. The Geological Society, London, pp. 101–121. Response of Late Ordovician paleoceanography to changes in sea level, continental

Bauert, H., Isozaki, Y., Holmer, L.E., Aoki, K., Sakata, S., Hirata, T., 2014. New U-Pb drift, and atmospheric pCO2: potential causes for long-term cooling and glaciation. zircon ages of the Sandbian (Upper Ordovician) “Big K-bentonite” in Baltoscandia Palaeogeogr. Palaeoclimatol. Palaeoecol. 210, 385–401. https://doi.org/10.1016/j. (Estonia and Sweden) by LA-ICPMS. GFF 136, 30–33. https://doi.org/10.1080/ palaeo.2004.02.034. 11035897.2013.862854. Herrmann, A.D., Macleod, K.G., Leslie, S.A., 2010. Did a volcanic mega-eruption cause Berger, A., Loutre, M.F., Laskar, J., 1992. Stability of the astronomical frequencies over global cooling during the Late Ordovician? PALAIOS 25, 831–836. https://doi.org/ the earth's history for paleoclimate studies. Science 255, 560–566. https://doi.org/ 10.2110/palo.2010.p10-069r. 10.1126/science.255.5044.560. Herrmann, A.D., Leslie, S.A., MacLeod, K.G., 2011. Did intense volcanism trigger the first Bergström, S.M., Huff, W.D., Kolata, D.R., Bauert, H., 1995. Nomenclature, stratigraphy, Late Ordovician Icehouse?: Comment. Geology 39, 237. https://doi.org/10.1130/ chemical fingerprinting, and areal distribution of some Middle Ordovician K-bento- G31758C.1. nites in Baltoscandia. GFF 117, 1–13. https://doi.org/10.1080/ Huff, W.D., 2008. Ordovician K-bentonites: issues in interpreting and correlating ancient 11035899509546191. tephras. Quat. Int. 178, 276–287. https://doi.org/10.1016/j.quaint.2007.04.007. Bergström, S.M., Schmitz, B., Saltzman, M.R., Huff, W.D., 2010. The Upper Ordovician Huff, W.D., 2016. K-bentonites: a review. Am. Mineral. 101, 43–70. https://doi.org/10. Guttenberg δ13C excursion (GICE) in North America and Baltoscandia: Occurrence, 2138/am-2016-5339. chronostratigraphic significance, and paleoenvironmental relationships. Geol. Soc. Huff, W.D., Bergström, S.M., Kolata, D.R., 1992. Gigantic Ordovician volcanic ash fall in Am. Spec. Pap. 466, 37–67. https://doi.org/10.1130/2010.2466(04). North America and Europe: biological, tectonomagmatic, and event-stratigraphic Bergström, S.M., Schmitz, B., Young, S.A., Bruton, D.L., 2011. Lower Katian (Upper significance. Geology 20, 875–878. https://doi.org/10.1130/0091-7613(1992) Ordovician) δ13C chemostratigraphy, global correlation and sea-level changes in 020<0875:GOVAFI>2.3.CO;2. Baltoscandia. GFF 133, 31–47. https://doi.org/10.1080/11035897.2011.557162. Huff, W.D., Bergström, S.M., Kolata, D.R., 2010. Ordovician explosive volcanism. Geol. Bergström, S.M., Eriksson, M.E., Schmitz, B., Young, S.A., Ahlberg, P., 2016. Upper Soc. Am. Spec. Pap. 466, 13–28. https://doi.org/10.1130/2010.2466(02). Ordovician δ13Corg chemostratigraphy, K-bentonite stratigraphy, and biostrati- Huyskens, M.H., Zink, S., Amelin, Y., 2016. Evaluation of temperature-time conditions for graphy in southern Scandinavia: a reappraisal. Palaeogeogr. Palaeoclimatol. the chemical abrasion treatment of single zircons for U-Pb geochronology. Chem. Palaeoecol. 454, 175–188. https://doi.org/10.1016/j.palaeo.2016.04.037. Geol. 438, 25–35. https://doi.org/10.1016/j.chemgeo.2016.05.013. Bjørlykke, K., 1974. Depositional history and geochemical composition of Lower Jürgenson, E., 1958. Metabentonites in Estonian SSR. 2. Trudy Instituta Geologii AN Palaeozoic epicontinental sediments from the Oslo Region. Nor. Geol. Unders. 305, ÉSSR, pp. 73–85. 1–81. Kiær, J., 1897. Faunistische Uebersicht der Etage 5 des Norwegischen Silursystems. Skr. Brenchley, P.J., Marshall, J.D., Carden, G.A.F., Robertson, D.B.R., Long, D.G.F., Meidla, Videnskab. Selsk. Kristiania I Mat. Naturv. 3. pp. 76. T., Hint, L., Anderson, T.F., 1994. Bathymetric and isotopic evidence for a short-lived Kiipli, T., Soesoo, A., Kallaste, T., 2014a. Geochemical evolution of caledonian volcanism Late Ordovician glaciation in a greenhouse period. Geology 22, 295–298. https://doi. recorded in the sedimentary rocks of the Eastern Baltic region. Geol. Soc. Spec. Publ. org/10.1130/0091-7613(1994)022<0295:BAIEFA>2.3.CO;2. 390, 177–192. https://doi.org/10.1144/SP390.5. Bruton, D.L., Gabrielsen, R.H., Larsen, B.T., 2010. The Caledonides of the Oslo Region, Kiipli, T., Kallaste, T., Nielsen, A.T., Schovsbo, N.H., Siir, S., 2014b. Geochemical dis- Norway - Stratigraphy and structural elements. Nor. J. Geol. 90, 93–121. crimination of the Upper Ordovician Kinnekulle Bentonite in the Billegrav-2 drill core Buggisch, W., Joachimski, M.M., Lehnert, O., Bergström, S.M., Repetski, J.E., Webers, section, Bornholm, Denmark. Est. J. Earth Sci. 63, 264–270. https://doi.org/10. G.F., 2010. Did intense volcanism trigger the first Late Ordovician icehouse? Geology 3176/earth.2014.29. 38, 327–330. https://doi.org/10.1130/G30577.1. Kjerulf, T., 1857. Ueber die Geologie des südlichen Norwegens. Nyt. Mag. Naturvid. 9, Cohen, K.M., Finney, S.C., Gibbard, P.L., Fan, J.-X., 2013. The ICS International chron- 193–333. ostratigraphic chart. Episodes 36, 199–204. https://doi.org/10.1111/j.1502-3931. Lenton, T.M., Crouch, M., Johnson, M., Pires, N., Dolan, L., 2012. First plants cooled the

212 E.G. Ballo, et al. Palaeogeography, Palaeoclimatology, Palaeoecology 520 (2019) 203–213

Ordovician. Nat. Geosci. 5, 86–89. https://doi.org/10.1038/ngeo1390. miscorrelation between the Millbrig and Kinnekulle K-bentonites of North America Lutro, O., Nordgulen, Ø., 2008. Oslofeltet, berggrunnskart M 1:250 000 (The Oslo Region, and Scandinavia. Geology 39, 303–306. https://doi.org/10.1130/G31425.1. Bedrock Map M 1:250 000). Geological Survey of Norway, Trondheim, Norway. Sell, B., Ainsaar, L., Leslie, S., 2013. Precise timing of the Late Ordovician (Sandbian) Mattinson, J.M., 2005. Zircon U-Pb chemical abrasion (“CA-TIMS”) method: combined super-eruptions and associated environmental, biological, and climatological events. annealing and multi-step partial dissolution analysis for improved precision and ac- J. Geol. Soc. Lond. 170, 711–714. https://doi.org/10.1144/jgs2012-148. curacy of zircon ages. Chem. Geol. 220, 47–66. https://doi.org/10.1016/j.chemgeo. Servais, T., Harper, D.A.T., Li, J., Munnecke, A., Owen, A.W., Sheehan, P.M., 2009. 2005.03.011. Understanding the Great Ordovician Biodiversification event (GOBE): influences of Min, K., Renne, P.R., Huff, W.D., 2001. 40Ar/39Ar dating of Ordovician K-bentonites in paleogeography, paleoclimate, or paleoecology? GSA Today 19, 4–10. https://doi. Laurentia and Baltoscandia. Earth Planet. Sci. Lett. 185, 121–134. https://doi.org/10. org/10.1130/GSATG37A.1. 1016/S0012-821X(00)00365-4. Sheehan, P.M., 2001. The Late Ordovician mass extinction. Annu. Rev. Earth Planet. Sci. Möller, N.K., Kvingan, K., 1988. The genesis of nodular limestones in the Ordovician and 29, 331–364. https://doi.org/10.1146/annurev.earth.29.1.331. Silurian of the Oslo Region (Norway). Sedimentology 35, 405–420. Smith, D.G., Bailey, R.J., 2017. Discussion on “a 2.3 million year lacustrine record of Munnecke, A., Calner, M., Harper, D.A.T., Servais, T., 2010. Ordovician and Silurian sea- orbital forcing from the Devonian of northern Scotland”. J. Geol. Soc. Lond. 175, water chemistry, sea level, and climate: a synopsis. Palaeogeogr. Palaeoclimatol. 474–488. https://doi.org/10.1144/jgs2016-137. Palaeoecol. 296, 389–413. https://doi.org/10.1016/j.palaeo.2010.08.001. Størmer, L., 1953. The Middle Ordovician of the Oslo Region, Norway. Nor. Geol. Tidsskr. Nardin, E., Goddéris, Y., Donnadieu, Y., Le Hir, G., Blakey, R.C., Pucéat, E., Aretz, M., 31, 37–141. 2011. Modeling the early Paleozoic long-term climatic trend. Bull. Geol. Soc. Am. Sundvoll, B., Larsen, B.T., 1994. Architecture and early evolution of the Oslo Rift. 123, 1181–1192. https://doi.org/10.1130/B30364.1. Tectonophysics 240, 173–189. https://doi.org/10.1016/0040-1951(94)90271-2. Neumann, E.R., Olsen, K.H., Baldridge, W.S., Sundvoll, B., 1992. The Oslo Rift: a review. Sutcliffe, O.E., Dowdeswell, J.A., Whittington, R.J., Theron, J.N., Craig, J., 2000. Tectonophysics 208, 1–18. https://doi.org/10.1016/0040-1951(92)90333-2. Calibrating the Late Ordovician glaciation and mass extinction by the eccentricity Nielsen, A.T., 2004. Ordovician sea-level changes: a Baltoscandian perspective. In: cycles of Earth's orbit. Geology 28, 967–970. https://doi.org/10.1130/0091- Webby, B.D., Droser, M.L., Paris, F., Percival, I.G. (Eds.), The Great Biodiversification 7613(2000)28<967:CTLOGA>2.0.CO;2. Event. Columbia University Press, New York, pp. 84–93. Svensen, H.H., Hammer, Ø., Corfu, F., 2015. Astronomically forced cyclicity in the Upper Owen, A.W., Bruton, D.L., Bockelie, J.F., Bockelie, T.G., 1990. The Ordovician successions Ordovician and U-Pb ages of interlayered tephra, Oslo Region. Palaeogeogr. of the Oslo region, Norway. Nor. Geol. Unders. Spec. Publ. 4, 3–54. Palaeoclimatol. Palaeoecol. 418, 150–159. https://doi.org/10.1016/j.palaeo.2014. Perrier, V., Meidla, T., Tinn, O., Ainsaar, L., 2012. Biotic response to explosive volcanism: 11.001. ostracod recovery after Ordovician ash-falls. Palaeogeogr. Palaeoclimatol. Telford, R.J., Heegaard, E., Birks, H.J.B., 2004. All age-depth models are wrong: but how Palaeoecol. 365–366, 166–183. https://doi.org/10.1016/j.palaeo.2012.09.024. badly? Quat. Sci. Rev. 23, 1–5. https://doi.org/10.1016/j.quascirev.2003.11.003. Plado, J., Ainsaar, L., Dmitrijeva, M., Poldsaar, K., Ots, S., Pesonen, L.J., Preeden, U., Torsvik, T.H., Cocks, L.R.M., 2016. Earth History and Palaeogeography. Cambridge 2016. Magnetic susceptibility of Middle Ordovician sedimentary rocks, Pakri University Press, London. https://doi.org/10.1017/9781316225523. Peninsula, NW Estonia. Est. J. Earth Sci. 65, 125. https://doi.org/10.3176/earth. Webby, B.D., Paris, F., Droser, M.L., Percival, I.G., 2004. The Great Ordovician 2016.11. Biodiversification Event. Columbia University Press. Rodionov, V.P., Dekkers, M.J., Khramov, A.N., Gurevich, E.L., Krijgsman, W., Duermeijer, Westphal, H., Munnecke, A., Brandano, M., 2008. Effects of diagenesis on the astro- C.E., Heslop, D., 2003. Paleomagnetism and cyclostratigraphy of the Middle chronological approach of defining stratigraphic boundaries in calcareous rhyth- Ordovician Krivolutsky suite, Krivaya Luka section, Southern Siberian platform: mites: the Tortonian GSSP. Lethaia 41, 461–476. https://doi.org/10.1111/j.1502- Record of non-synchronous NRM-components or a non-axial geomagnetic field? Stud. 3931.2008.00101.x. Geophys. Geod. 47, 255–274. https://doi.org/10.1023/A:1023767523451. Westphal, H., Hilgen, F., Munnecke, A., 2010. An assessment of the suitability of in- Schmitz, M.D., Schoene, B., 2007. Derivation of isotope ratios, errors, and error corre- dividual rhythmic carbonate successions for astrochronological application. Earth lations for U-Pb geochronology using 205Pb-235U-(233U)-spiked isotope dilution Sci. Rev. 99. https://doi.org/10.1016/j.earscirev.2010.02.001. thermal ionization mass spectrometric data. Geochem. Geophys. Geosyst. 8. https:// Williams, G.E., 1991. Milankovitch-band cyclicity in bedded halite deposits con- doi.org/10.1029/2006GC001492. temporaneous with Late Ordovician-Early Silurian glaciation, Canning Basin, Schoene, B., Guex, J., Bartolini, A., Schaltegger, U., Blackburn, T.J., 2010. Correlating the Western Australia. Earth Planet. Sci. Lett. 103, 143–155. https://doi.org/10.1016/ end-Triassic mass extinction and flood basalt volcanism at the 100 ka level. Geology 0012-821X(91)90156-C. 38 (5), 387–390. https://doi.org/10.1130/G30683.1. Worsley, D., Nakrem, H.A., 2008. The Lower Palaeozoic–Cambrian, Ordovician and Schulz, M., Mudelsee, M., 2002. REDFIT: estimating red-noise spectra directly from un- Silurian–the Sea teems with life; 542–416 Ma. In: Ramberg, I.B., Bryhni, I., Nottvedt, evenly spaced paleoclimatic time series. Comput. Geosci. 28, 421–426. https://doi. A., Rangnes, K. (Eds.), The Making of a Land: Geology of Norway. Geological Society org/10.1016/S0098-3004(01)00044-9. of Norway, Trondheim, pp. 148–177. Sell, B.K., Samson, S.D., 2011. Apatite phenocryst compositions demonstrate a

213