Tephra horizons identi ﬁ ed in the western North Atlantic and Nordic Seas during the Last Glacial Period: Extending the marine tephra framework Quaternary Science

Geochemically distinct volcanic ash (tephra) deposits are increasingly acknowledged as a key geochro- nological tool to synchronize independent paleoclimate archives. Recent advances in the detection of invisible (crypto) tephra have led to the ongoing establishment, development and integration of regional tephra lattices. These frameworks offer an overview of the spatial extent of geochemically characterized tephra from dated eruptions e a valuable tool for precise correlation of paleorecords within these areas. Here, we harness cryptotephra analysis to investigate the occurrence of two well-known tephra markers from the Last Glacial Period (i.e. FMAZ II-1 (26.7 ka b2k) and NAAZ II (II-RHY-1) (55.3 ka b2k)), in marine sediment cores from the Nordic, Irminger and Labrador Seas. In addition, we assess the imprint of bioturbation on two of these tephra deposits using Computed Tomography (CT) imagery. We have successfully identi ﬁ ed FMAZ II-1 in the Nordic and Irminger Seas. The tephra deposit is a visible deposit in the Nordic Seas, whereas it appears as a single high concentration peak within the ﬁ ne-grained shard size fraction (i.e. 25-80 m m) in the Irminger Sea. Both horizons are primary airfall deposits, and this study is the ﬁ rst to identify a FMAZ II-1 deposit of isochronous nature in the Irminger Sea region. In addition, we have identi ﬁ ed a new tephra horizon in the Irminger Sea, which is stratigraphically associated with FMAZ II-1, and geochemically similar to the known 2-JPC-192-1 population. We discuss its potential to serve as a new reference tie-point for correlations in the region. Lastly, we have successfully identi ﬁ ed NAAZ II (II-RHY-1) of isochronous nature in both the Irminger and Labrador Sea. The layers are inter- preted to be deposited by either direct airfall or by sea-ice drifting past the sites. Compared to the existing frameworks, which previously mainly focused on sites east of Iceland, our ﬁ ndings expand the knowledge and utility of the FMAZ II-1 and NAAZ II (II-RHY-1) horizons.


Introduction
Tephrochronology, the use of synchronously deposited and geochemically fingerprinted ash horizons as time markers across geological archives, has become an increasingly recognized tool for correlating Late Quaternary climate records. Fundamentally, the detection of well-dated and geochemically distinct tephra horizons within disparate and/or distant records allows for an assessment of the synchronicity of change during abrupt climate transitions in the past (Austin et al., 2012). Recent advances in cryptotephra (invisible to the naked eye) analysis (Davies, 2015), have resulted in the discovery of new chronological tie-points at more distal localities, further promoting the development of more detailed tephra frameworks (Bourne et al., 2015;Abbott et al., 2018a).
Tephra frameworks are a compilation of both visible and cryptotephra occurrences in distal and proximal settings. In addition, they provide an overview of the dispersal area of certain volcanic eruptions and of the past eruptive frequency of volcanoes in the region. Several tephra frameworks from the North Atlantic region already exist, such as the overview of tephra horizons identified in marine, terrestrial and ice core records from the last 128-8 ka BP by Blockley et al. (2014). As for tephra frameworks focusing solely on the Last Glacial Period (60-25 ka b2k), Bourne et al. (2015) showed that close to 100 Icelandic eruptions between 45 and 25 ka b2k can be traced in the Greenlandic ice cores DYE-3, NEEM, NGRIP and GRIP. Furthermore, Abbott et al. (2018a) recently compiled information of 14 tephra horizons from ten different marine sediment cores in the North Atlantic, covering the period between 60 and 25 ka b2k. However, only two marine tephra horizons, notably the Faroe Marine Ash Zone (FMAZ) II (26.5 ka b2k) and North Atlantic Ash Zone (NAAZ) II (55.3 ka b2k) , have so far been confidently correlated to the records from the Greenland icecores, which provide a precise geochronological control on the timing of these eruptions (Haflidason et al., 2000;Wastegård et al., 2006;Griggs et al., 2014;Abbott et al., 2018a). This study focuses on these two significant tephra isochrons.
Faroe Marine Ash Zone II, also called Fugloyarbanki (FMAZ II-1), is a basaltic tephra first identified as a visible primary airfall deposit in the Faroe Island region by Rasmussen et al. (2003). The only report of a FMAZ II-1 deposit outside the Faroe Island region, is from the Labrador Sea by Wastegård et al. (2006) whichidentified a deposit with a geochemical composition consisting of a mixture of both FMAZ II-1 and another layer with a different geochemical signature (i.e. . The FMAZ II-1 horizon has since been correlated to the Greenland ice-core NGRIP ( Fig. 1) . The horizon was assigned an age of 26 740 ± 390 years b2k derived from the GICC05 chronology (Svensson et al., 2006) and an origin from the Icelandic Hekla-Vatnafj€ oll volcanic system, in the Eastern Volcanic Zone (EVZ) was suggested (Wastegård et al., 2006;Davies et al., 2008). Stratigraphically, the layer was deposited during Greenland Stadial (GS) 3, about 1000 years after the onset of the warmer Greenland Interstadial (GI) 3 period and marks the transition between Marine Isotope Stages (MIS) 3 and MIS 2 . This is consistent with its position within the marine realm, where it was recorded just after the warmest reconstructed temperature peak assigned to be a marine counterpart to Greenland interstadial 3 (Rasmussen et al., 2003).
The objective of this study is to further develop the existing North Atlantic tephra framework between 60 and 25 ka b2k previously presented by Abbott et al. (2018a). This objective is aimed for by examining the occurrence of FMAZ II-1 and NAAZ II (II-RHY-1) in marine sediment cores from the eastern (Nordic Seas) and western (Irminger and Labrador Sea) North Atlantic Ocean. In addition, we will assess the isochronous nature of these tephra layers and, as such, investigate whether they can act as independent time-markers (isochrons) for future correlation to other records.

Selection of marine core depth-intervals
Previously developed chronologies for all investigated cores allowed us to target the time intervals, and thus core depths, where we expect FMAZ II-1 and NAAZ II (II-RHY-1) to be deposited.

MD99-2284
The position of a black tephra layer between 1404 and 1409 cm in MD99-2284, stratigraphically possibly correlating to FMAZ II-1, was first visually identified and reported by Dokken et al. (2013). Nonetheless, its geochemical composition has, so far, never been analyzed. The first visual appearance of ash at the base of the layer was used as the tephra marker horizon (i.e. 1408-1409 cm).

GS16-204-18CC
Samples with the potential of containing FMAZ II-1 and NAAZ II (II-RHY-1) material in GS16-204-18CC were carefully selected using the magnetic susceptibility record (Dokken and Cruise-Members, 2016) which records cycles of Greenland Interstadials (GI) and Greenland Stadials (GS) (Voelker and Haflidason, 2015). To further support the sampling interval selection, we analyzed concentrations of ice rafted debris (IRD) and planktic foraminifera d 18 O values (Lisa Griem pers. commun 20.08.2018). Subsequently, light isotope events that mark the stratigraphic position of Heinrich events were used, supporting a preliminary age model. Based on this evidence, we selected the intervals 210e250 cm and 505e525 cm, stratigraphically located between GI-3 and GI-2, after Heinrich event 3 and GI-15, respectively, for tephra analysis.

Tephra analysis
Sediment samples from depth-intervals that fall within the age range of the targeted tephra deposits were sampled as 0.5 cm (GS16-204-18/22CC) and 1 cm (MD99-2284) slices at 1 cm intervals. An exception was made for core GS16-204-18CC (250-210 cm) that initially was sampled at 2 cm intervals, and in the case of increased tephra shard concentrations at 1 cm. The samples were first freeze-dried and homogenized. Subsequently, ca 0.5 g dry weight of material from each sample was prepared for tephra analysis following the methodology for marine tephra deposits (Abbott et al., , 2018b. To remove carbonate material, dilute (10%) hydrochloric acid (Hcl) was added to each sample and left overnight (~12 h). Samples were subsequently sieved into three size fractions (i.e. >125 mm, 80e125 mm and 25e80 mm). The finegrained size fraction (25e80 mm) was then separated into different density fractions (i.e. >2.5 g/cm 3 , 2.3e2.5 g/cm 3 and <2.3 g/cm 3 ) using heavy liquid flotation with sodium polytungstate (SPT). This technique is applied to separate rhyolitic (2.3e2.5 g/cm 3 ) from basaltic (>2.5 g/cm 3 ) glass shards (Turney, 1998;Blockley et al., 2005). Using the methodology from Griggs et al. (2014), the >2.5 g/cm 3 fraction was magnetically separated using a Frantz Isodynamic Magnetic Separator in an effort to separate the paramagnetic basaltic shards from the non-magnetic minerogenic material. Finally, each sample was mounted on glass slides using Canada Balsam. If tephra shard concentrations exceeded 10.000 shards/g, the preparation steps described previously were repeated and lycopodium spore tablet(s) were added to the 25e80 mm fraction after the final density separation step. The required number of tablets varied (1e2 tablets) but the aim was to ensure that >300 spores were represented on each microscope slide. Then, to allow total dissolution of the spore tablet(s) the sample was soaked in ca. 5 ml Hcl, where after it was washed and rinsed three times to ensure that the remaining Hcl was completely removed. To ensure a representative range of the sample, three drops of the sample, with the material in suspension, were mounted on a microscope slide. Eventually, all tephra shards and lycopodium spores on the microscope slide were counted after Gehrels et al. (2006). The relative amount of tephra shards was calculated using equation (2.1), where l is the number of lycopodium spores in each tablet (n ¼ 18 584 ± 354, batch no. 177745).
glass shard count Lycopodium spore count Â sample dry weight Þ (2.1) For each analyzed depth interval, peaks in the tephra shard concentration profiles were selected for major element analysis. After repeating the previously described preparation steps, the samples were embedded in epoxy resin on frosted microprobe slides. To expose the surface of the glass tephra shards, the mounted material was ground using p1000 silicon carbide paper and polished using ¼ mm diamond polycrystalline suspension. Individual tephra shards were analyzed using Electron-probe microanalysis (EPMA). These measurements were performed at the Tephrochronological Analytical Unit at the University of Edinburgh using a Cameca SX100 electron microprobe with five vertical wavelength dispersive spectrometers, providing oxide values (wt. %) for 10 major elements (see supplementary information). Approximately 20e40 individual shards per sample were analyzed following the protocols outlined by Hayward (2012). Based on the sample's shard size, a 3 mm or 5 mm beam diameter was used (see supplementary information). To monitor analytical precision, glass standards (Lipari Obsidian (rhyolitic) and BCR2g (basaltic)) were measured regularly. For geochemical data comparison, the data was normalized to 100% total oxides. All raw data values are given in the supplementary information. Totals below 94% and 97% for rhyolitic and basaltic material, respectively, were rejected.
The major element data (oxides expressed as wt. %) was statistically compared to previously published geochemical populations using statistical distance (SD) and similarity coefficient (SC) tests following the methods outlined in Perkins et al. (1995) and Borchardt et al. (1972), respectively. In addition, graphical examination using bi-plots was carried out. When calculating the SC's, we only included elements with concentrations >1 wt %. Traditionally, values between 0.95 and 1 have been interpreted as identical dataset; whereas values between 0.90 and 0.95 as not identical dataset, but most likely originate from the same volcanic source (Davis, 1985;Beget et al., 1992). However, it should be noted that for Icelandic Volcanic systems, Abbott et al. (2018a) only accept SC's higher than 0.97 as identical geochemical compositions. The SD function considers the differences between two datasets and can only be used to assess if two populations are different and thus, not that they are the same. The calculated values are compared to critical values (¼18.48 (rhyolitic) and ¼ 23.21 (basaltic)) at a 99% confidence level. The difference in the critical values between rhyolitic and basaltic material is a result of comparing major element oxides with an average value higher than 0.1 wt % (10 elements for basaltic and 7 elements for rhyolitic material). If the statistical distance value is higher than the critical value, the datasets are considered to be different. If the value is lower than the critical value the datasets are not considered to be different, but not necessarily identical (Pearce et al., 2008). In addition to statistical tests, the stratigraphic position of the tephra horizons in the different marine sequences were also considered when correlating deposits.

Ice rafted debris
In ice-proximal areas like the studied region, icebergs provide a possible transport pathway for tephra shards that eventually hamper the isochronous nature of a tephra horizon. Therefore, to evaluate the influence of tephra transported to the region by icebergs, we generated IRD records for the same depth intervals as investigated for the tephra analysis. This combination of IRD and tephra shard concentration profiles will offer insight to whether or not a tephra peak results from ice-rafting transport to the study sites (represented by increased IRD) (Griggs et al., 2014;Abbott et al., 2018b). For GS16-204-18CC we constructed a new IRD record (lithic grains between 150 and 500 mm) using the standard method of split counting (Heinrich, 1988;Bond and Lotti, 1995). For GS16-204-22CC, we used the IRD record presented in Griem et al. (2019).

Computed Tomography (CT) scanning
We visualize the imprint of bioturbation and IRD on parts of the cores GS16-204-18CC (510e529.5 cm) and GS16-204-22CC (452.5e488.5 cm), using Computed Tomography (CT) after e.g. Griggs et al. (2015). To do so, we employed a ProCon X-Ray CT-ALPHA scanner, operated at 100 kV and 850 mA and using a 267 ms exposure time. To capture sub-millimeter scale features, we minimized the distance between the detector and source by scanning 2 cm wide u-channels. Reconstructed 16 bit scans were processed with the Thermo Scientific™ Avizo™ 9.1.1 software suite. First, we selected specific CT grayscale ranges with the threshold tool to highlight the density of hollow burrows (air) and ice rafted debris (clastic). For this purpose, we relied on the grayscale intensity histograms of our scans after Griggs et al. (2015): the lightest (low grayscale) peak corresponds to air, while the densest (high grayscale) peak reflects clastic material. To warrant correct interpretation, the outcome of this iterative process was compared to visual evidence of hollows and rock particles in the scanned core segments. Next, we created 3-D visualizations of thresholded (highlighted) CT grayscale ranges (features) using the volume rendering tool. Finally, we used the sieving tool to remove isolated voxels with a diameter smaller than 400 mm to reduce cluttering (noise), before visualizing highlighted features using a combination of 2-D ortho slices and 3-D visualizations after Van der bilt et al. (2018).

Evaluating the isochronous nature of tephra deposits
Tephra shards can be transported to the marine realm by a range of different pathways and during the Last Glacial Period, in our study areas, this was predominantly via direct aerial ash fall-out, by icebergs or sea-ice (Griggs et al., 2014). However, tephra layers are also susceptible to secondary depositional mechanisms such as remobilization of material by bioturbation and/or strong bottom currents. To be able to use tephra layers as time-markers, they need to be deposited and incorporated in the sediment sequence nearinstantaneously following an eruption. In this study, we evaluate tephra layers with respect to their potential primary and secondary depositional mechanisms following the newly introduced classification scheme on deposit types outlined by Abbott et al. (2018b). They classified tephra deposits into five different types common in North Atlantic marine sequences. The five deposit types are summarized in Table 1. A type 1 deposit is defined as a well-constrained, low concentration shard peak with homogeneous composition, representing one single depositional event most likely deposited by primary airfall. A type 2 deposits reveals a distinct high concentration peak in shard concentration with an upward or downward spanning of shards and is either geochemically homogeneous (2A) or heterogeneous (2B). The deposit represents one single depositional event, but can be subjected to secondary reworking. The transport mechanism of this deposit type must be evaluated based on the geochemical composition as primary airfall, sea-ice rafting and iceberg rafting is possible. A type 3 deposit typically shows a flat bottom profile with an upward tailing of shards, a very high shard concentration and a geochemically homogeneous composition. Secondary reworking and/or bioturbation cause the gradational upward tailing and the most likely transport mechanism is either primary airfall or sea-ice rafting. A type 4 deposit has high shard concentrations and reveals multiple peaks over a large spread (10s of cm). Such a deposit type likely represents either several closely spaced eruptions or deposition by icebergs. A type 5 deposit has a wide spread of consistent shard concentrations, which typically represents a background signal. These shards are most likely reworked and remobilized within the ocean system, which could mask low concentration peaks representing single volcanic events.

Results
The tephra deposits identified in this study are summarized in Table 2.
The tephra deposit in core MD99-2284 has a sharp visual boundary between the tephra layer and the underlying sediments as well as a visible upward tailing of decreasing tephra shards (Fig. 2B). Tephra shards from the base (1408e1409 cm) of the visible layer have been geochemically analyzed for major elements.
The stratigraphic and geochemical features of the MD99-2284 (1408e1409 cm) deposit are consistent with a deposit type 3 (Table 1), which is most likely transported via primary airfall or seaice rafting. The visibility of the layer and thus the immense Table 1 Overview of the tephra deposit type classification scheme used in this study after Abbott et al. (2018b Yes, if peaks can be tied to the Greenland tephra framework. The tephra peaks also have potential as regional marine tie-lines.
Type 5 Background signal of consistent low shard concentration. Geochemically heterogeneous Reworking and remobilization within the ocean system.
No, but potential isochrons could be masked by the background signal.

Table 2
Summary of tephra deposits investigated in this study with respect to their isochronous integrity, deposit type, climatic event, correlative isochrons and volcanic source. GS ¼ Greenland stadial, GI ¼ Greenland interstadial, H¼Heinrich event. References are as follows: (1)   concentration of shards argue for a primary airfall deposition although sea-ice rafting cannot be fully excluded. Nonetheless, as the potential temporal delay by sea-ice rafting (months to years) is shorter than the chronological resolution within marine sequences (Brendryen et al., 2010), neither of these potential transport processes are considered to cause a significant temporal delay. Thus, this deposit contains all the required characteristics to be defined as an isochron.  (Wastegård et al., 2006;Griggs et al., 2014) and from the Greenland ice core NGRIP (blue circle), . In addition, the tephra shard geochemistry from GS16-204-18CC (225.5e226 cm, >125 mm) is compared to the 2-JPC-192-1 geochemical data (blue area) from Wastegård et al., (2006). Error bars represents 2 standard deviations of replicate analyses of BC2rg reference glass. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

GS16-204-18CC
The tephrostratigraphy for GS16-204-18CC (210e250 cm) is presented in Fig. 4. Between 225.5 cm and 229 cm a basaltic tephra horizon is observed in all size fractions. However, within the finegrained size fraction (i.e. 25-80 mm), the basaltic tephra concentration peak occurs between 228.5 and 229 cm. In contrast, within the coarser-grained size fractions (i.e. >125 mm and 80e125 mm), both concentration peaks appear 3 cm upwards, between 225.5 and 226 cm. Tephra shards from all size fractions from the depth intervals that capture a concentration peak were geochemically analyzed for major elements.
One single high concentration peak with a homogeneous geochemistry and no up-or downward tailing of shards characterizes the fine-grained (i.e. 25-80 mm) tephra deposit from 228.5 to 229 cm. In addition, the concentration peak does not co-occur with any peaks in the IRD record (Fig. 4). This evidence is most consistent with a type 2A deposit (Table 1). Such a deposit is most likely transported and deposited by primary airfall or sea-ice rafting, which cause no significant temporal delay after the eruption. Therefore, this deposit is defined as an isochron.
3.1.2.2. GS16-204-18CC: 225.5e226 cm. Concerning the tephra deposit found 3 cm upwards between 225.5 and 226 cm, the tephra shard geochemistry from the two smallest size fractions (i.e. 80-125 mm and 25e80 mm) is basaltic and heterogeneous. For instance, SiO 2 values range from 48 to 51.3 wt % (Fig. 3A). Of these shards, six (of 39) have a geochemistry similar to FMAZ II-1, which is recorded 3 cm earlier. These FMAZ II-1 shards are likely deposited as a result of secondary transport mechanisms such as reworking and/or iceberg rafting. On the other hand, the geochemistry from the coarser-grained shards (i.e. GS16-204-18CC, 225.5e226 cm; >125 mm) shows, with the exception of two outliers, a fairly homogeneous basaltic composition. Representative features are normalized values of ca. 49.5e51 wt. % SiO 2 , ca. 1e1.4 wt. % TiO 2 , ca. 7.4e8.6 wt. % MgO, ca. 12e13.7 wt. % CaO, ca. 0.06e0.16 wt. % K 2 O and FeO/MgO ratios between 1.2 and 1.6 (Fig. 3B). This geochemical Table 3 Statistical comparison of the geochemical compositions from the basaltic layers of MD99-2284 (1408e1409 cm) and GS16-204-18CC (228.5e229 cm, 25e80 mm) with the FMAZ II-1 population in North Atlantic marine records and the Greenland ice-core record. In addition, statistical comparison of the geochemical compositions of the rhyolitic layers in GS16-204-18CC (512.5e513 cm) and GS16-204-22CC (474e475 cm) with the NAAZ II (II-RHY-1) population in North Atlantic marine records and the Greenland icecore record. composition is distinctly different from the FMAZ II-1 population identified 3 cm lower in the core, respectively at 228.5e229 cm. This geochemical signature suggests an origin from either the Bardarbunga-Veidiv€ otn volcanic system in the Eastern Volcanic Zone (EVZ) (Jakobsson, 1979;Olad ottir et al., 2011) or the Reykjanes volcanic system in the Western Volcanic Zone (WVZ) (Jakobsson et al., 1978) (see supplementary figures). The only report of a tephra deposit that is stratigraphically related to the FMAZ II-1 isochron, but has a distinctly different geochemistry is the 2-JPC-192-1 deposit from the Labrador Sea (core EW9302-2JPC) (Wastegård et al., 2006). Statistical comparison between this deposit and the GS16-204-18CC (225.5e226 cm, >125 mm) geochemical population reveals a SC of 0.90 and a SD of 7.02. The low SC value of 0.90 indicates that the geochemical signatures are not similar. The small number of measurements (n ¼ 7) from the 2-JPC-192-1 layer (Wastegård et al., 2006) offers a limited dataset for statistical comparison which might explain the low SC. Nonetheless, the 2-JPC-192-1 and GS16-204-18CC (225.5e226 cm, >125 mm) populations could also represent different, but closely spaced eruptions from the same volcanic center. The tephra shard concentration profile of this deposit shows a high concentration peak with tailing of shards a few centimeters downwards (Fig. 4). In addition, the deposit co-occurs with increased peaks in IRD concentration, which is indicative of iceberg transport to the site.
However, geochemical data from the >125 mm fraction reveals a fairly homogeneous population. This particular deposit was most likely not deposited by icebergs as iceberg-rafted deposits often exhibit a heterogeneous geochemistry. These results argue for a type 2A deposit (Table 1), which most likely is transported to the site by primary airfall or sea-ice rafting. These are transport and depositional mechanisms that do not affect the isochronous integrity of the deposit and therefore, the deposit is defined as an isochron.

GS16-204-22CC
The tephra shard concentration profile from GS16-204-22CC (191e210 cm) reveals in all size fractions a continuous background signal of basaltic tephra shards (<500 shards/g) (Fig. 5b). A minor tephra shard concentration peak was observed between 196 and 196.5 cm in all size fractions and therefore, shard material from this depth was prepared for geochemical analysis. In addition, based on increasing concentrations in the >125 mm size fraction, tephra shards from the 201e201.5 cm interval was geochemically analyzed.  (Larsen, 1981;Jakobsson, 1979). A statistical comparison between this homogeneous sub-population and the FMAZ II-1 geochemistry from a Greenland ice-core  and several North Atlantic marine records (Wastegård et al., 2006;Griggs et al., 2014) reveals SC's between 0.93 and 0.98 and SD's between 0.25 and 1.15, which are indicative of a correlation (Table 3). Geochemical analyses of shards from the two coarsergrained size fractions (i.e. >125 mm and 80e125 mm) of GS16-204-22CC (201e201.5 cm) reveal a basaltic heterogeneous geochemistry (Fig. 3A). Five (of 15) of these shards correlate to the FMAZ II-1 geochemical suite and are likely deposited as a product of secondary transport mechanisms. and thus, this heterogeneous geochemistry represents tephra material derived from a mix of volcanic sources (Fig. 3A). One dominant source is the FMAZ II-1 eruption as 19 (of 67) shards correlate to this isochron. The GS16-204-22CC (191e210 cm) tephrostratigraphy is characterized by a consistent background concentration of tephra shards with no clear concentration peak. In addition, the tephra shard concentration is low (<500 shards/g) and generally the geochemistry of the measured intervals is heterogeneous. The upper part of the deposit (191e196 cm) coincides with increased levels of IRD concentrations. This evidence argues for a type 5 deposit (Table 1), which has most likely been influenced by postdepositional reworking and remobilization, potentially masking smaller tephra shard concentration peaks. Possibly, the analyzed tephra shards from the fine-grained size fraction between 201 and 201.5 cm that correlate to the FMAZ II-1 horizon are such a masked deposit. However, due to the remobilization of this material, the deposit cannot be convincingly correlated to the FMAZ II-1 isochron.

GS16-204-18CC
The tephra shard concentration profile of GS16-204-18CC (505e525 cm) reveals a rhyolitic deposit in the 25e80 mm and >125 mm size fractions between 511 and 518 cm (Fig. 6). We find a distinct high concentration peak between 512.5 cm and 513 cm (Fig. 6) and analyzed tephra shards from this high concentration peak as well as from the base of the deposit between 517.5 cm and 518 cm. Due to the immense shard concentrations the size fraction 80e125 mm was not counted.
On either side of the main concentration peak we identify a high number of rhyolitic tephra shards (Fig. 6). Evidence from the highresolution CT-scan between 510 and 529.5 cm, which identified 0.5e1 cm elongated burrows positioned just below the main concentration peak between 512.5 and 513 cm, indicate that bioturbation has been an active process in this section of the core and the downward tailing of shards could be a product of this activity (Fig. 8A). The geochemistry of the deposit is homogeneous, and there is no IRD peak coinciding with the concentration peak. These results are indicative of a type 2A deposit (Table 1) hinting at two possible transport mechanisms: (1) the tephra was transported to the site directly by airfall or (2) the tephra was transported by primary airfall onto sea-ice that most likely drifted to the site along the East Greenland Current (EGC) (Fig. 1). Although the proximity to the Icelandic source and the presence of a relatively high concentration of coarse shards (>125 mm) argue stronger for sea-ice rafting, deciphering between the two transport mechanisms is at this point not possible. However, in either scenario, there is no significant temporal delay of deposition after the eruption that would affect the integrity of the isochron.

GS16-204-22CC
The GS16-204-22CC (455e479 cm, >125 mm and 25e80 mm) The concentration of basaltic shards per gdw (gram dry weight) is quantified in three different size fractions (i.e. >125 mm, 80e125 mm and 25e80 mm (>2.5 g/cm 3 )). c) Ice rafted debris per gdw from the 150e500 mm size fraction of GS16-204-18CC (200e280 cm, every 2 cm). Dotted horizontal lines mark the position of the two depth intervals, 196e196.5 cm and 201e201.5 cm, that were geochemically analyzed. tephrostratigraphy shows a relatively flat-bottomed profile with an upward tailing of tephra shards starting from a high concentration peak between 474 cm and 474.5 cm. The deposit is positioned between 469 and 474.5 cm (Fig. 9b) and the main rhyolitic shard maximum is observed between 474 cm and 474.5 cm in both size fractions (i.e. >125 mm and 25e80 mm). Due the extensive shard concentrations, the size fraction 80e125 mm was not counted.
Tephra shards from the main concentration peak (474e474.5 cm) and from neighbouring samples (473e473.5 cm and 470e470.5 cm) were geochemically analyzed for major elements in order to assess the relationship to the main concentration peak.
The high resolution CT-scan of GS16-204-22CC (452.5e488.5 cm) reveals 1e2 cm elongated burrows upwards from 474 to 474.5 cm (Fig. 8B). The presence of burrows at this level in the core verifies bioturbation as an active process that could cause the upward tailing of tephra shards identified in the tephra shard concentration profile. In addition, there are no IRD peaks coinciding with the tephra deposit. These characteristics indicate a type 3 deposit (Table 1), which was most likely deposited by primary airfall or sea-ice rafting. Subsequently, the tephra deposit is useful as an isochron. Both the visible tephra layer recorded in MD99-2284 (1408e1409 cm) from the Nordic Seas and the fine-grained fraction (25e80 mm) in GS16-204-18CC (228.5e229 cm) are of isochronous nature and can be correlated to the established geochemistry of the FMAZ II-1 horizon in the literature. Similar to our results in MD99-2284 (1408e1409 cm), FMAZ II-1 appears as a thick and visible layer in many records from the Nordic Seas and Faroe region (Kuijpers et al., 1998;Rasmussen et al., 2003;Wastegård et al., 2006;Griggs et al., 2014). In fact, previous marine investigations of airfall deposited FMAZ II-1 have mainly focused on the latter region, and this study is the first to observe primary airfall deposited FMAZ II-1 in the Irminger Sea (GS16-204-18CC). With the new data presented, we expand the known dispersal range of the FMAZ II-1 tephra towards the west (Fig. 1).
Close in depth to the FMAZ II-1 isochron in GS16-204-18CC we find a coarse-grained homogenous basaltic tephra layer with geochemical characteristics similar to the 2-JPC-1-192 layer, previously reported mixed with the FMAZ II-1 horizon in the Labrador Fig. 6. Summary of marine sediment core GS16-204-18CC from the Irminger Sea. a) Magnetic susceptibility (10 À5 Si units) from GS16-204-18CC (500e540 cm) (Dokken and Cruise-Members, 2016). GI ¼ Greenland Interstadial. b) Tephrostratigraphy from GS16-204-18CC (505e525 cm) plotted versus depth (cm). The concentration of rhyolitic shards per gdw (gram dry weight) is quantified in two different size fractions (i.e. >125 mm and 25e80 mm (2.3e2.5 g/cm 3 )). Note that the middle panel refers to the level of shard counts >10.000/g, which were treated with lycopodium to achieve the panel on the right. c) Ice rafted debris per gdw from the 150e500 mm size fraction from GS16-204-18CC (502e522 cm). Grey horizontal line marks the position of the tephra isochron. The dotted horizontal line marks the position of peaks geochemically analyzed in addition to the main peak.
Sea (Wastegård et al., 2006). However, a clear statistical correlation between the 2-JPC-1-192 layer and 225.5e226 cm (>125 mm) layer in GS16-204-18CC cannot be ascertained. Still, with additional geochemical data from these horizons, they may be linked in the future. The tephra from these two horizons appears only as coarser tephra grains (>125/150 mm) and is restricted to the areas south/ southwest of Iceland (i.e. Irminger and Labrador Sea). Nonetheless, attempts have been made to discover the horizon in cores from the Faroe Island region and the Reykjanes ridge (Griggs et al., 2014). The occurrence of coarser grains and the, so far, exclusive recordings southwest of Iceland argue for more local eruption(s) and a regional transport mechanism that transported the tephra material from Iceland and solely to the southwesterly sites. We suggest that the tephra material was predominantly carried westwards by winds and deposited on sea-ice that drifted along the EGC (Fig. 1). In this manner, the material would only be distributed to the southsouthwestern parts of the North Atlantic Ocean. Indeed, it has been suggested that in Greenland Stadials, during which FMAZ II-1 is deposited, a southward shift of the polar front allowed for the EGC to expand and divert southwards, carrying drifting sea-ice to more southerly sites than today (e.g. to core EW 9302-2JPC in the Labrador Sea) (Van Kreveld et al., 2000).
The investigations of FMAZ II-1 in core GS16-204-22CC (191e201 cm) from the Labrador Sea were inconclusive. Although FMAZ II-1 material was present, the core depth-interval that recorded tephra showed evidence for remobilization and reworking of material, and no isochron could be determined. Either the lack of FMAZ II-1 material in GS16-204-22CC (191e201 cm) is a result of local remobilization of sediments or the core site is located outside the western limit of the primary FMAZ II-1 tephra distribution. However, in order to further investigate the FMAZ II-1 airdispersal limits in a southwesterly direction, new efforts might be able to identify primary airfall deposited FMAZ II-1 layers southwest of our findings in the Irminger Sea in marine sediment cores that show no evidence of remobilization.
The largest and most updated MIS 3 and 2 North Atlantic tephra framework was presented by Abbott et al. (2018a) in which they investigated ten North Atlantic marine sediment cores. Within this framework, the FMAZ II-1 horizon is reported in one core from the southeastern Nordic Seas (JM11-19 PC) (Griggs et al., 2014). In the North Atlantic tephra framework by Wastegård et al. (2006), the FMAZ II-1 horizon (>150 mm) is reported in six cores. Five of the six cores within that framework are located in the region around the Faroe Islands whereas only one of them is located in the Labrador Sea. Hence, based on the existing tephra frameworks, there is no comprehensive understanding of the air dispersal pattern of FMAZ II-1 in a southwesterly direction from Iceland. In this study, we show that cryptotephra analysis allows the detection of the FMAZ  (1989). B: Visual biplot comparison of tephra shard analyses (major element oxides) GS16-204-18CC (474-4 cm) and GS16-204-22CC (512.5e513 cm) to the NAAZ II (II-RHY-1) geochemical data from the North Atlantic marine tephra framework (grey shaded area) (Austin et al., 2004;Wastegård et al., 2006;Brendryen et al., 2011;Abbott et al., 2016Abbott et al., , 2018a and from the Greenland ice core GRIP (black line) (Gr€ onvold et al., 1995). Error bars represents 2 standard deviations of replicate analyses of Lipari Obsidian reference glass.
II-1 horizon at sites on the western side of the North Atlantic Ocean. These results expand the tephra framework westwards and allow to link both sides of the North Atlantic Ocean (Fig. 10). Future investigations of FMAZ II-1 should preferentially focus on the western side of the North Atlantic, as existing frameworks by Abbott et al. (2018a) and Wastegård et al. (2006) already cover most of the eastern side. In addition, we cautiously add a new tephra horizon to the framework that is associated with the FMAZ II-1   (Dokken and Cruise-Members, 2016). GI ¼ Greenland Interstadial. b) Tephrostratigraphy from GS16-204-22CC (463e479 cm) plotted versus depth (cm). The concentration of rhyolitic shards per gdw (gram dry weight) is quantified in two different size fractions (i.e. >125 mm and 25e80 mm (2.3e2.5 g/cm 3 )). Note that the middle panel refers to the level of shard counts >10.000/g, which were treated with lycopodium to achieve the right panel. c) Ice rafted debris per gdw from the 150e500 mm size fraction of GS16-204-22CC (460e490 cm). Grey horizontal line marks the position of the tephra isochron. The dotted horizontal lines mark the position of peaks geochemically analyzed in addition to the main peak. deposit, from the Irminger Sea (GS16-204-18CC, 225.5e226 cm, >125 mm). If this new horizon can be detected in more records, there is potential for a new reference horizon that can be used as a correlational tie-point in records from the south and southwest of Iceland. In addition, since this new horizon is stratigraphically linked to the more widespread FMAZ II-1 isochron, the horizon can be used to link records containing either one of the two horizons in future studies.

NAAZ II (II-RHY-1)
We correlate the rhyolitic deposits in GS16-204-18CC (512.5e513 cm) and GS16-204-22CC (474e474.5 cm) to the NAAZ II (II-RHY-1) isochron. The NAAZ II (II-RHY-1) isochron has been identified in several sites across the North Atlantic Ocean and in the Greenland ice-core GRIP (Kvamme et al., 1989;Austin et al., 2004;Wastegård et al., 2006;Brendryen et al., 2011;Abbott et al., 2018a). At some of these sites, predominantly the more eastern ones basaltic/intermediate material is also present within the NAAZ II layer (Abbott et al., 2018a). The basaltic component of NAAZ II appears to be more pronounced at sites closer to the source and localities on the eastern side of the North Atlantic Ocean. Most likely the basaltic/intermediate material was transported to these sites by icebergs calving off the Icelandic ice sheet (Abbott et al., 2018a), which probably completely melted before reaching the more  (Abbott et al., 2018a). EW9302-2JPC, ENAM93-20, ENAM 93-21, ENAM33, LINK17, LINK04 (Wastegård et al., 2006). SO82-05 (Haflidason et al., 2000;Brendryen et al., 2011). JM11-19 PC (Griggs et al., 2014;Abbott et al., 2018a). MD04-2820CQ (Abbott et al., 2016). MD95-2006(Austin et al., 2004. MD99-2289 (Brendryen et al., 2011). Map was generated using the Ocean Data View software (http://odv.awi.de/). B: Schematic of the improved North Atlantic tephra framework after findings in this study, based on Fig. 2 in Abbott et al. (2018a). Newly identified tephra horizons by this study are marked in yellow core schematics, while the previously recorded tephra horizons from the existing North Atlantic tephra framework are marked in grey core schematics. For simplicity marine sediment cores from the existing North Atlantic tephra framework were grouped into geographical areas. Although the 2-JPC-192-1 and GS16-204-18CC (225.5e226 cm) horizons could not be convincingly correlated using statistical tests they most likely originate from the same volcanic source and are therefore here grouped together. Please note that the age scale is approximate. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) south-southwestern core sites. This is consistent with the lack of a basaltic NAAZ II component in our sites in the Labrador and Irminger Sea. Abbott et al. (2018a) report nine recordings of NAAZ II (II-RHY-1) within the North Atlantic marine tephra framework; five of these are from cores located west of the British Isles, two are from cores north and northeast of Iceland, and two are from sites south of Iceland (Gardar Drift and Labrador Sea). Within the North Atlantic tephra framework by Wastegård et al. (2006), the NAAZ II (II-RHY-1) isochron was identified in four cores; three of these cores are located in the Faroe Island region and one is from the Labrador Sea. In addition, Brendryen et al. (2011) contributes to the framework with NAAZ II (II-RHY-1) data from a site on the Reykjanes ridge, south of Iceland, as well as one site in the Nordic Seas (Fig. 10). A NAAZ II (II-RHY-1) horizon has also been inferred in the Irminger Sea (Elliot et al., 1998;Stoner et al., 1998); this reporting is, however, not substantiated with geochemical data. Our results therefore contribute to the North Atlantic tephra framework with new and updated geochemical data of the NAAZ II (II-RHY-1) horizon at localities that either fall outside the existing framework or are sparsely covered (Fig. 10). In addition, through the use of highresolution CT-imagery, we determine that bioturbation most likely attributed to the tailing of shard concentrations in both deposits; this has not convincingly been shown in cores from the existing framework.

Conclusions
We have successfully identified the FMAZ II-1 isochron in MD99-2284 (1408e1409 cm) from the Nordic Seas and in GS16-204-18CC (228.5e229 cm, 25e80 mm) from the Irminger Sea. In contrast with sites in the Nordic Seas and Faroe Islands, where the FMAZ II-1 is recorded as a visible layer, the FMAZ II-1 is observed as a fine-fraction layer in the Irminger Sea (GS16-204-18CC, 228.5e229 cm, 25e80 mm). The discovery of FMAZ II-1 in the Irminger Sea is the first in the region and expands the previously known dispersal of the FMAZ II-1 tephra in a more northwesterly direction than showed by previous studies (Fig. 10A). This result broadens the North Atlantic tephra framework westwards and offers a new strategically located tie-point between the eastern and western side of the North Atlantic Ocean. Close in depth to our discovery of the FMAZ II-1 tephra in the Irminger Sea, a tephra deposit was recorded in the >125 mm size fraction (GS16-204-18CC, 225.5e226 cm). The geochemical composition of this coarser layer shows characteristics similar to 2-JPC-192-1 from the Labrador Sea. This horizon, found in stratigraphic proximity to FMAZeIIe1, has so far only been found southwest of Iceland. We suggest that tephra shards from these horizons were distributed from Iceland on seaice via the EGC, limiting the dispersal in a southwestward direction. Potentially, this new horizon can act as a reference horizon for correlation of records if more occurrences are fingerprinted in the region. We have also successfully identified the NAAZ II (II-RHY-1) in GS16-204-18CC (512.5e513 cm) from the Irminger Sea and in GS16-204-22CC (474e474.5 cm) from the Labrador Sea. These findings contribute to the North Atlantic tephra framework with new and updated geochemical data from the NAAZ II (II-RHY-1) isochron in the region. In addition, in order to better understand secondary reworking processes, we show how high-resolution CTimagery can be used to visualize small-scale bioturbation above and below tephra isochrons.
In total, we report three different tephra horizons in three North Atlantic marine sediment cores that all possess an isochronous nature. All of these layers have potential to be used as time-markers or correlational tie-points in future studies and will aid in unraveling the synchronicity of rapid climatic transitions in different climate archives during the Last Glacial Period.

Declaration of interests
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.