Magnetostratigraphy of U-Pb–dated boreholes in Svalbard, Norway, implies that magnetochron M0r (a proposed Barremian-Aptian boundary marker) begins at 121.2 ± 0.4 Ma

Rates of plate tectonic motions, biologic evolution, geochemical excursions, and other processes in Earth's history depend on an accurate geologic time scale. The high-resolution time scale for the Late Jurassic through Early Cretaceous is compiled mainly from the correlation of biostratigraphy to the M sequence of magnetic polarity chrons, and the durations of many of those biozones and polarity chrons have been derived from cyclostratigraphy on reference sections (e.g., Channell et al., 1995; Sprovieri et al., 2006; Malinverno et al., 2012).

In particular, the age model for the Aptian Stage in the Geologic Time Scale 2012 (GTS2012; Gradstein et al., 2012) had used ca. 126 Ma for the beginning of Chron M0r, the proposed magnetozone marker for the base of the Aptian Stage (Erba et al., 1996). That age estimate was based on (1) the consistency of ⁴⁰Ar/³⁹Ar and U-Pb radioisotopic dating of oceanic basalts and volcanic ash beds from Ocean Drilling Program (ODP) sites, from the Great Valley Group in California (USA), and from Argentina (Fig. 1A); and (2) an assumed duration of 13 m.y. for the Aptian Stage according to cyclostratigraphic interpretation of the Piobbico core of central Italy (Huang et al., 2010) relative to a U-Pb date of 113.1 ± 0.3 Ma near the Aptian-Albian boundary (Selby et al., 2009). This 126 Ma age was significantly older than the ca. 121.5 Ma age suggested in earlier scales (e.g., Channell et al., 1995). A younger age was partly supported by ⁴⁰Ar/³⁹Ar dates of 122.0 ± 0.5 Ma on basalt flows yielding reversed polarity in northeastern China that were interpreted as belonging to magnetozone M0r (He et al., 2008), although biostratigraphic constraints and bounding magnetozones were lacking.

However, U-Pb dates published after 2012, which used an isotope dilution–thermal ionization mass spectrometry (ID-TIMS) method coupled with new techniques for processing of zircons and EARTHTIME standards (http://www.earthtimetestsite.com/working-groups/upb-isotope-dilution/), seem to suggest that nearly all of the ⁴⁰Ar/³⁹Ar dates derived from ODP basalt sites and most of the pre-2012 U-Pb dating of volcanic ashes were systematically too old by an average of ∼4 m.y. (Fig. 1B). For example, a bentonite bed in the uppermost Hauterivian of Argentina that had yielded a U-Pb date of 132.5 ± 1.3 Ma was re-dated by the ID-TIMS method at 129.09 ± 0.16 Ma (Aguirre-Urreta et al., 2015).

Critical to this study is a bentonite layer in the Helvetiafjellet Formation in Svalbard, Norway, dated using ID-TIMS U-Pb at 123.1 ± 0.3 Ma (Corfu et al., 2013; Midtkandal et al., 2016). Based on regional palynology and a negative carbon-isotopic (δ¹³C_org, org—organic carbon) excursion in the overlying Carolinefjellet Formation (Midtkandal et al., 2016), which was proposed to be equivalent to the δ¹³C excursion segment “C3” accompanying the onset of Oceanic Anoxic Event 1a (OAE1a) of the midle– early Aptian (e.g., Menegatti et al., 1998; Erba et al., 1999), the bentonite was interpreted as being of midle–late Barremian age. A postulated triggering cause of that global negative δ¹³C excursion is the eruption of the immense Ontong Java Plateau large igneous province, which rapidly released isotopically light carbon into the ocean-atmosphere system (e.g., Erba et al., 2015).

We collected magnetostratigraphic data from a core (DH1; Fig. 2) penetrating the Helvetiafjellet Formation to accurately place the U-Pb–dated 123.1 Ma bentonite into the global framework of the Barremian through earliest Aptian magnetic polarity time scale, thereby helping to resolve the disputed age model for parts of the Early Cretaceous.

The cores investigated in this study were retrieved from a series of research wells (wells DH1, DH3, and DH7; Fig. 2D) drilled in relation to a carbon-sequestration project (Longyearbyen CO₂ Lab, http://co2-ccs.unis.no/) and are stored in the University Centre in Svalbard (UNIS; Longyearbyen, Svalbard). In these wells, the Helvetiafjellet Formation is an ∼75-m-thick paralic succession sandwiched between open-marine shelfal facies of the lower Valanginian to lowermost Barremian Rurikfjellet Formation and the Aptian to Albian Carolinefjellet Formation (Grundvåg et al., 2019). The base of the Helvetiafjellet Formation is on a subaerial unconformity surface of early Barremian age (Fig. 2B; Śliwińska et al., 2020). A 20-cm-thick bentonite layer occurs in boreholes DH3 and DH7 (Corfu et al., 2013; Midtkandal et al., 2016) and projects to just below a fluvial sandstone unit at about level 180 m in borehole DH1 (see Section S1 in the Supplemental Material¹). The bentonites are considered to be a product of the regional High Arctic large igneous province (e.g., Polteau et al., 2016). The transgressive base of the overlying mudstone-rich Caroline-fjellet Formation is a reworked deposit marking the flooding of the Helvetiafjellet Formation coastal plain (Midtkandal et al., 2016; Grundvåg et al., 2019). Details are discussed in the Supplemental Material.

Paleomagnetic minicores were drill-pressed at ∼1 m spacing from levels 130 m to 215 m of core DH1, spanning the entire Helvetiafjellet and the lowermost part of the Carolinefjellet Formations (Fig. 2). The paleomagnetic directions of the 110 minicores were obtained using a composite scheme of thermal demagnetization to 200–300 °C (as dictated by lithology) followed by alternating field (AF) treatments. This scheme was guided by rock magnetic experiments including thermal demagnetization of orthogonal isothermal remanent magnetization (IRM) and rotational remanent magnetization (RRM). Additional details are discussed in the Supplemental Material. The characteristic remanent magnetization (ChRM) for each sample was computed by a three-dimensional “least-squares fitting” technique (Kirschvink, 1980) using the public software PaleoMagX (Jones, 2002), and a subset was analyzed using LINEFIND (Kent et al., 1983) utilizing the variance of each measurement. Quality ratings on each ChRM and polarity interpretation were assigned [N (or R)—confident; NP (RP)—valid; NPP (RPP)—probable; N? (R?)—possible; or INT—uncertain; where N refers to normal polarity, and R to reversed polarity] based on the stability of the magnetic vectors during the progressive demagnetization (Tables S2 and S3 in the Supplemental Material). Examples of the quality ratings are illustrated in Figure S3.

The coercivity and IRM results (Fig. S1) indicate mixed magnetic mineralogy in the samples, with magnetite being the main phase of primary remanent magnetization carriers with some minor contributions from detrital pyrrhotite or maghemite.

The magnetostratigraphy of the Helvetiafjellet Formation in core DH1 yielded three pairs of reversed- and normal-polarity magnetozones (Hv0 to Hv2) based on the higher-quality-rated samples (Fig. 2). Reversed-polarity magnetozone Hv0r in the uppermost Helvetiafjellet Formation occurs just below the negative δ¹³C_org excursion documented by Midtkandal et al. (2016; Fig. 2). This relationship is apparently very similar to the stratigraphic position of magnetozone M0r below the OAE1a δ¹³C_carb (carb—carbonate) excursion within the Aptian reference sections in northern Italy (e.g., the Cismon core; Erba et al.,1999, 2015; Fig. 2).

Guided by recent constraints from palynology and dinocysts (Śliwińska et al., 2020), the base of the Helvetiafjellet Formation appears to be no older than middle–early Barremian. Therefore, the magnetozones Hv1 and Hv2 of paired normal and reversed polarity underlying Hv0r are correlated to magnetochrons M1n, M1r, M3n, and uppermost M3r, respectively (blue guidelines in Fig. 2). This implies that the bentonite dated at 123.1 ± 0.3 Ma by Corfu et al. (2013) occurs in the uppermost part of magnetozone M1r.

Cyclostratigraphy of Italian sections yields durations for magnetochrons M1r and M1n of 0.20 m.y. and 1.85 m.y., respectively (Sprovieri et al., 2006). Therefore, the placement of the bentonite bed (123.1 Ma) within magnetozone M1r is 1.9 m.y. (±0.1 m.y.) prior to the onset of magnetozone M0r, implying that Chron M0r begins at 121.2 Ma. This interpolated age has an uncertainty of ∼0.4 m.y. from its relative placement within magnetozone Hv1r (±0.1 m.y.) and the total uncertainty on the U-Pb date (±0.3 m.y., which includes external uncertainties). This derived 121.2 Ma age for the onset of Chron M0r is significantly younger than the ca. 126 Ma age used in GTS2012 and younger than a recent range estimate of between 123.8 and 121.8 Ma determined by reevaluation of the constraints from published radioisotopic dates (Olierook et al., 2019).

The duration of Chron M0r is estimated as 0.5 m.y. (e.g., Huang et al., 2010), whereas the onset of OAE1a occurred 0.3 m.y. after the end of Chron M0r (e.g., Malinverno et al., 2010). Several latest Barremian through earliest Aptian ammonite zones and microfossil datums of the Tethyan and Subboreal regions have calibrations relative to Chron M0r and to carbon-isotope trends, and some ammonite zones have durations derived directly from cyclostratigraphy (e.g., Frau et al., 2018; Frau, 2020; Luber et al., 2019; Gale et al., 2020; Martinez et al., 2020). The interpolated 121.2 Ma age for the base of Chron M0r enables the assignment of an age model for this integrated bio-magneto-isotopic stratigraphic scale (Fig. 3).

Three of the markers currently under discussion for assigning the base of the Aptian Stage (i.e., meeting minutes of the International Sub-commission on Cretaceous Stratigraphy, Third International Congress on Stratigraphy, STRATI 2019, 4 July 2019, Milan, Italy) are (1) the base of magnetozone M0r, (2) the onset of negative δ¹³C excursion “C3” or the beginning of OAE1a, or (3) the base of ammonite zone Deshayesites oglanlensis or another biological datum. The indicated 121.2 Ma age in the DH1 core (Fig. 2) for the base of Chron M0r thus enables age estimates for other markers (Fig. 3).

According to the estimated 121.2 Ma age for the beginning of Chron M0r, the Cretaceous Normal Superchron spanning the Aptian to the beginning of Chron C33r (beginning of the Campanian Stage) becomes 5 m.y. shorter than its estimated span in GTS2012. This revised age model indicates a commensurate ∼12% increase in average global oceanic spreading rates during the Aptian–Santonian interval. A similar conclusion, but of slightly lesser magnitude (∼6%), was reached by Olierook et al. (2019) from their reevaluation of published radioisotopic dates.

The implied ∼5 m.y. shortening of the duration of the Aptian Stage relative to its span in GTS2012 does not necessarily imply that the duration of the underlying Barremian Stage becomes longer. Instead, it appears that the array of new U-Pb dates (Fig. 1) requires that an expansion by ∼4 m.y. should be distributed across portions of the age model for much of the Middle Jurassic through Barremian (i.e., a time period of ∼50 m.y.). Numerous consistent cyclostratigraphy studies of the durations of Oxfordian through Barremian biozones, substages, and magnetic polarity zones (e.g., Martinez et al., 2020; and reviews in Gradstein et al. [2012], and in Gale et al. [2020]) imply that many of these must retain approximately the same durations as compiled in GTS2012. We thus suggest that a significant portion of the required expansion of the age model would be for the relatively brief Middle Jurassic stages, which were constrained by the ⁴⁰Ar/³⁹Ar date of 168.7 ± 1.7 Ma on ODP Site 801C on Pacific magnetic anomaly M42n.4r (Koppers et al., 2003; Tominaga et al., 2008).

In the future, the distribution of the suggested ∼4 m.y. expansion and upward shifting of the age models for the Middle Jurassic through Barremian and of the commensurate shortening of the Aptian Stage can be resolved by applying ID-TIMS methods with EARTHTIME standards to re-date the California volcanic ash layers; by acquiring and verifying cyclostratigraphy of the Kimmeridgian through Aptian from outcrops, ODP cores and other boreholes; and by acquiring additional radioisotopic dates from successions that have precise age frameworks established from bio-, magneto- and cyclostratigraphy.

We thank editor James Schmitt, Elisabetta Erba, Helmut Weissert, Andrew Gale, Sten-Andreas Grundvåg, Camille Frau, Kenneth Kodama, and an anonymous reviewer for their constructive suggestions on this paper. This study was inspired by numerous discussions with Beatriz Aguirre-Urreta, Víctor A. Ramos, and Elizabeth Johnson. We thank UNIS CO₂ LAB (University Centre Svalbard, Longyearbyen CO₂ Lab) for access to core material and logistics, and Julian Janocha for his careful drill pressing; Anita de Chiara and Vassil Karloukovski for their help in the Lancaster Environment Centre (UK) lab work; and sponsorship from the LoCrA (Lower Cretaceous basin studies in the Arctic) consortium managed by the University of Stavanger (Stavanger, Norway) and the UNIS. Funding was provided by research grants from the Geologic TimeScale Foundation, Chevron Energy Technology Company, Chengdu University of Technology, and the Research Council of Norway.