On the the origin of the Yermak Plateau north of Svalbard, Arctic Ocean

On the the Plateau in the Arctic north The Yermak Plateau north of Spitsbergen and Morris Jesup Spur and rise north of Greenland relate to the Late Cretaceous-early Cenozoic interaction between an independent Greenland plate and the larger North American and European plates. We have recovered 21 new dredge hauls from three locations on the Yermak Plateau with an abundance of metasedimentary and gneissic rocks with strong affinities to known lithologies from northwest Spitsbergen. The continental outlier requires Paleogene dextral shear close to the coast of West Spitsbergen to accommodate opening of the Sophia Basin between the plateau and the continental margin. The postulated large-offset (100–150 km) shear zone (de Geer Fault) is supported by seismic velocity anomalies down to mid-crustal levels, a ubiquitous feature of known large-offset continental transform faults regardless of crustal rock composition. A continental sliver including the Yermak Plateau and Prins Karls Forland initially moved with Greenland along the de Geer Fault during the early Eocene stage of Eurasia Basin opening and facilitated opening of the Sophia Basin north of Spitsbergen by crustal extension. Later offset of the de Geer Fault north of Spitsbergen and formation of the Danskøya Basin in a transfer zone was probably induced by a restraining bend in the Hornsund Fault Zone active at the same time. The 65 km-wide, circular-shaped, northeastern tip of the Yermak Plateau is a young volcanic feature formed between Chron 22 and Chron 18 at the junction between the Gakkel Ridge and the Yermak continental block before separation of the Morris Jesup Spur and Yermak Plateau. The Yermak Plateau became part of the European plate prior to Chron 13 as the Gakkel Ridge propagated into the Northeast Greenland margin and the subsequent dextral motion shifted west to the Hornsund Fault Zone. The de Geer Fault and the Hornsund Fault Zone may have been in existence at the same time. E-mail (Yngve


Introduction
Arctic geoscientific research over the last fifty years has documented a general relationship between the independent Paleogene motion of Greenland and tectonic events in the Canadian Arctic (Eurekan Orogeny) and on Svalbard (West Spitsbergen Thrust and Fold Belt) (Fig. 1). While the geology of the respective land areas has been fairly well explored (CASE Team, 2001;Henriksen, 2005;Dallmann, 2015;Piepjohn et al., 2016), the offshore links are uncertain; in particular, the origin of outlying structures such as the Yermak Plateau north of Svalbard and the Morris Jesup Spur north of Greenland (Fig. 1). It was early recognised that reconstructions of the position of Greenland and the Lomonosov Ridge back to the time of onset of E-mail corresponding author (Yngve Kristoffersen): Yngve.kristoffersen@geo.uib.no seafloor spreading in the Eurasia Basin produced overlaps if the Morris Jesup Spur and Yermak Plateau were continental fragments (LePichon et al., 1977;Feden et al., 1979;Vogt et al., 1979). In this contribution, we first review the state of knowledge of the relevant tectonic features, report on an extensive effort to dredge known basement outcrops at three sites on the Yermak Plateau and reconsider published crustal velocity information. The objective is to develop a working hypothesis which integrates transform plate motion west of Spitsbergen with the formation of the Sophia Basin and the Yermak Plateau to the north.  Greenland with domains (red and yellow) and major faults (white) of the land geology involved in the early Cenozoic Eurekan Orogeny. Bathymetry from Jakobsson et al. (2012), geological information from Svalbard from Dallmann (2015), from Ellesmere Island from Piepjohn et al. (2016) and from North Greenland from Pedersen & Håkanson (1999), von Gosen & Piepjohn (2003) and Svennevig et al. (2016). Abbreviations: EGR -East Greenland Ridge, HR -Hovgard Ridge.

Outline of the geological framework
The plateau north of Spitsbergen is named after Yermak, the first icebreaker in the world. Yermak, with is 9000 horsepower, reached up to 81 o 21' N during its first season in 1899 (www.prlib.ru/en/ history/619672). The plateau (water depth 500-1000 m) has a western NNW-trending part which extends north about 200 km from the shelf break north of Svalbard, and a northeastern part parallel to the Gakkel spreading centre (Figs. 1 & 2). The eastern end (east of 82⁰ 30' N, 14⁰ E) is a circular-shaped, 65 km-wide feature about 500 metres deeper than the adjacent part of the plateau. The northeastern part of the Yermak Plateau is separated from the margin north of Spitsbergen by the Sophia Basin ( Fig. 2). The relevant geological framework includes; i) the Yermak Plateau, ii) the Sophia Basin and the continental margin north of Spitsbergen, iii) the West-Spitsbergen Fold and Thrust belt and structures on the continental shelf west of Spitsbergen.

Yermak Plateau
The geophysical results reported by Jokat et al. (2008) and Geissler et al. (2011) show that the smooth and rounded cross-section of the Yermak Plateau is due to sediments which cover basement topography with up to 2 km of elevation differences over a distance of 10-15 km in the northwest (Fig.   3A). Basement in the south forms a large, about 180 km-wide block which becomes narrower (~100 km) and more dissected towards the north. The Sverdrup Bank is part of the high eastern side of this large block and bedrock is exposed at the sea bed in at least two local areas. Another basement exposure is off this block farther to the northeast (Figs. 2, 3B).
An unreversed seismic refraction line (line 4, Fig. 2) shows velocities of 4.3, 6.0 and 8.0 km/s which was interpreted as a 20 km-thick crust of continental rocks (Jackson et al., 1984). Riefstahl et al. (2013) reported the first effort to dredge basement outcrops on the Yermak Plateau apart from a single gneiss boulder recovered in a dredge of opportunity in the late 1970s (Jackson et al., 1984). More than half of the rocks recovered from a dredge on Sverdrup Bank and another site 25 km to the north (Fig. 2) were magmatic rocks, mostly alkaline basalts. It was concluded that the alkaline dolerites are related to rift magmatism ( ̴ 51 Ma) and the metamorphic rocks comparable to the Devonian and older basement rocks of northern Spitsbergen (Riefstahl et al., 2013).
The northeastern part of the Yermak Plateau (north of 82⁰ N) is associated with relatively high magnetic amplitudes (up to 1000 nT) which stand out compared to the quiet magnetic field associated with the western and southern part (Jackson et al., 1984;Brozena et al., 2003;Jokat et al., 2008). Several investigators have linked the magnetic anomaly amplitude to a relatively high content of magnetic minerals in volcanic source rocks (Feden et al., 1979;Brozena et al., 2003). A 100 km-long seismic refraction line shot in 1981 (line 3, Fig. 2) to investigate the deeper crustal structure had very sparse spatial shot-point intervals (15-22 km) beyond 50 km offset (Jackson et al., 1984). Nevertheless, the observed amplitudes appeared to match a synthetic seismogram representing an upper 8 km-thick layer of 5.0 km/s velocity over rocks with velocities in the range of 6.7 -7.2 km/s typical of oceanic layer 3, but thicker. Their preferred interpretation is thickened oceanic crust, in line with the proposal of Feden et al. (1979) and Vogt et al. (1979); the magnetic part of the Yermak Plateau may have been Figure 2. Overview of locations of seismic profiles and major basement structures north of 79°N. Basement ridges outlined by contours for sediment thickness <1.5 km (brown areas) with data from northwest of Spitsbergen from Eiken (1992) and Opsahl (1997), and data from the Yermak Plateau from Jokat et al. (2008) and Geissler et al. (2011). Dashed white line outlines the northeastern tip of the Yermak Plateau which is suggested to be a volcanic construction. Seismic reflection lines (red) are from Geissler et al. (2011). The crustal transition from continent to ocean (COB) west of Spitsbergen is from Engen et al. (2008), the location of Hornsund Fault Zone from Eiken (1994), Jokat et al. (2008) and Blinova et al. (2009), and the Bouguer gravity gradient (red area) is adapted from Minakov et al. (2012, fig. 6). The simplified geology of Spitsbergen and Nordaustlandet is from Dallmann (2015). Our dredge locations on the Yermak Plateau are shown by white Xs and the seismic refraction profiles (3 and 4) are from Jackson et al. (1984) and shown by heavy black lines. Abbreviations: AWI -Alfred Wegener Institute, Germany, BF -Billefjorden Fault, IF -Isfjorden, KF -Kongsfjorden, LF -Lomfjorden Fault, NBP -Northwestern Basement Province, NFT -New Friesland Terrane, NPD -Norwegian Petroleum Directorate, SB -Sverdrup Bank, PKF -Prins Karls Forland, YBF -Ymerbukta Fault. Jokat et al. (1995) and Gjengedal (2004), and from the western part from Geissler et al. (2011). Sediments interpreted as younger than mid-Miocene  are shown in green colour and acoustic basement in brown colour. Seismic velocities are in km/sec. The locations of seismic profiles are shown in Fig. 2. (B) Line drawing of composite seismic profiles across the Sophia Basin based on data from Geissler et al. (2011) with the same colour scheme as in Fig. 3A. Abbreviation: MS -Mosby Seamount. Abbreviation: AWI -Alfred Wegener Institute, Germany, ODP -Ocean Drilling Program. generated by seafloor spreading with excessive outpouring of basalt starting at Chron 20 and reaching a maximum between Chrons 18 and 13. Other investigators favour a northeastern volcanic part generated by high basalt production contemporaneously with formation of the Morris Jesup Rise, and a northwestern and southern part of the Yermak Plateau comprising: i) thinned continental crust based on crustal seismic velocity structure and magnetic signature (Jackson et al., 1984) or a downfaulted splinter of the Svalbard shelf (Birkenmajer, 1972).
ii) a transform-related ridge formed by massive intrusions of basalt (post Chron 7) along the western margin to account for high heat flow (Crane et al., 1982(Crane et al., , 1988Okay & Crane, 1993).
A third alternative has been advocated by Jokat et al. (2008) based on the combined seismic reflection and potential field data. They rejected a potential plate boundary between the two parts of the plateau and suggested that the structure is rather a collection of blocks of stretched continental crust formed by a broadening of the Hornsund Fracture Zone in the north. The high intensity of the magnetic field over the northeastern part was considered to be a result of intrusions at depth.

The Sophia Basin and the continental margin north of Spitsbergen
The Sophia Basin constitutes a more than 2000 m-deep reentrant in the continental margin north of Spitsbergen (Figs. 1 & 2). A northeast-trending basement ridge runs along the axis of the basin (Figs. 2 & 3B) and reaches 700 metres above the basin floor in two places (Geissler & Jokat, 2004), but any buried continuity between the two is hitherto unknown. The basement ridge (Mosby Seamount) is flanked to the north by rotated basement blocks (Fig. 3B), and basement to the south is undetected below a thick sediment accumulation which may reach a thickness of 9 km (Geissler & Jokat, 2004). The magnetic field intensity over the deep part of the Sophia Basin has no distinct linear features except for local maxima  Minakov). The direction of initial opening in the Eurasia Basin is show by the black arrow based on data given by Gaina et al. (2002) and Glebovsky et al. (2006). The dashed white line outlines the northeastern tip of the Yermak Plateau which is suggested to be a volcanic construction. (Berglar et al., 2016;Jokat et al., 2016). The widest part of basin is related to an offset in the continental slope at 81°N, 17°E north of Hinlopen Strait, accompanied by a steep change in crustal thickness as defined by the Bouguer gravity gradient (red shaded area in Fig. 2). The Bouguer gravity field is obtained by correcting the observed Free Air Gravity values for the density deficit of water with respect to sediments across a continental margin, and the horizontal Bouguer gravity gradient will reflect the slope of the crust-mantle interface. The Bouguer gravity values indicate an abrupt crustal thinning at the northeastern end of the Sophia Basin (Fig. 4). The crust below the basin is interpreted from gravity modelling to be dense, about 5 km thick and oceanic-like (Geissler & Jokat, 2004). The combined geophysical data suggest that the basin is floored by attenuated and heavily intruded continental crust (Geissler & Jokat, 2004;Engen et al., 2008).
The continental basement on the shelf north of Spitsbergen drops by about 1.5 km at the Moffen Fault (Eiken, 1992). Deep seismic data combined with gravity modelling suggest that the continental crust shallows to the north from 24 km to 16 km across the fault and the modelled average density of the crust to the north is higher than below the shelf (Sundvor & Austegard, 1990). Also, Geissler & Jokat

The plate boundary between Greenland and Svalbard
The first continental reconstruction presented by Wegener (1912) sought to minimise the misfit of approximate shelf edge geometries. His proposal implied a straight E-W directed separation between Greenland and Svalbard. At the same time, de Geer (1912de Geer ( , 1919 Eiken (1994), Blinova et al. (2009) and Gabrielsen et al. (1992) and the structural data south of Sørkapp from Bergh & Grogan (2003). Basement ridges NW of Spitsbergen as in Fig. 2. The location of the Hornsund Fault is from Eiken (1994) and Jokat et al. (2008) and the continent-ocean crustal boundary from Engen et al. (2008) early Tertiary Scandic land area of de Geer (1926). Wilson (1965) associated the de Geer Line with a dextral transform fault which has been referred to by many authors as the "de Geer Shear Zone".
We use the term 'de Geer Fault' for a postulated major dextral transform close to the coast of West Spitsbergen with active displacement during formation of the Sophia Basin in the Paleocene-early Eocene (Fig. 2).
The main tectonic structures along the continental shelf west of West Spitsbergen are a near-shore graben and the Hornsund Fault Zone (Fig. 5). The graben is 15-20 km wide and appears to open up north of 79 o N, but its southern continuation is unclear (Bergh & Grogan, 2003). The main graben boundary fault changes polarity between Isfjorden and Bellsund ( Fig. 5) and is filled with 1-2 km of weakly deformed sediments (Blinova et al., 2009). A drillhole in Forlandsundet (Fig. 5,star symbol) reached metamorphic basement at 1.05 km. Kleinspehn & Teyssier (2016) suggested that the Forlandsundet graben was part of a broader piggy-back basin within the active West Spitsbergen Fold and Thrust Belt and the oldest graben sediments were deposited during the late Eocene (after 38 Ma).
The main graben formation came later in the early Oligocene.
The Hornsund Fault was first detected in single-channel seismic data and sonobuoy measurements on the outer shelf from Bear Island to Hornsund (77 o N). It was characterised by a province of high seismic velocities (3.8-4.2 km/s) at the sea bed to the east and a wedge of low-velocity sediments (1.7-1.8 km/s) below the upper continental slope to the west (Sundvor & Eldholm, 1976). The boundary between shallow basement and the depositional wedge was named the Hornsund Fault. The feature was later extended to 79 o N Sundvor et al., 1978;Myhre et al., 1982) and farther north by Jokat et al. (2008). Subsequent surveys between Bear Island and the southern tip of Spitsbergen revealed an up to 50 km-wide fault complex where the listric eastern master fault (Knølegga Fault) has a throw of >3 km (Gabrielsen et al., 1990;Rehman, 2012). The fault complex appears to terminate just south of 76 o N and its NNW-ward extension is unclear (Bergh & Grogan, 2003).
The subsurface structure below the continental shelf north of 76o 30' N is poorly surveyed. The position of the Hornsund Fault used in most publications is not a distinct master fault, but the eastern fault of a >20 km-wide, complex, coast-parallel zone of down-dropped blocks to the west Eiken & Austegard, 1987;Eldholm et al, 1987;Eiken, 1994;Bergh & Grogan, 2003;Faleide et al., 2008;Blinova et al., 2009;Faleide et al., 2010). Jokat et al. (2008) suggested a northward extension to 81o 30' N ( Fig. 2). The Hornsund Fault does not represent the boundary between continental and oceanic crust west of Spitsbergen as Breivik et al. (1999) have used the Bouguer gravity calibrated by seismic velocity data to outline the crustal transition from continent to ocean to within a zone of 20 km width west of Spitsbergen (Figs. 2 & 5). The implication is that up to 40 km of extended continental crust may be present in a transtensional transition zone between the Hornsund Fault and the inferred boundary between continental and oceanic crust. We use the term 'Hornsund Fault Zone' for this transition zone.

Materials and methods
We have used a hovercraft research platform (Kristoffersen & Hall, 2014) to recover a total of 14 successful rock dredges from two groups of localities separated by -15 km in a north-south direction on Sverdrup Bank and one group of 7 dredges from a site on the northeastern part of the plateau (Figs. Note that orthogneisses may also be of Caledonian age (e.g., Petterson et al., 2009).

Sverdrup Bank
The highest part of the Sverdrup Bank is bevelled with a slightly elevated western rim (Figs. 2 & 6,bottom). The opaque acoustic character of the high is interpreted as basement outcrop and weak internal reflections have a northeast component of dip . A small prograding wedge of sediments is present on the southwest side of the high (Fig. 6, bottom), but such a wedge is not apparent on the western and northwestern sides . The dredging operation was concentrated over the high on the eastern rim (Fig. 6). In one instance the dredge fastened on the bottom and the kevlar line (breaking strength 2.8 ton) parted.
The results presented in Table 1 and Fig. 6  Thin-sections from seven of the samples were studied and the results will be discussed as part of the interpretation.

Northeastern Basement High
The successful dredge hauls and seismic reflection measurements define a 4.5 km-long and 1 km-wide, E-W-trending, bedrock outcrop at the seabed (Fig. 6) at a location termed the Northern Basement High of Riefstahl et al. (2013). The range of recovered lithologies includes a plate fragment of bedded black shale with plant fragments, silicified limestone, chert and a reduced abundance (20%) of metamorphic rocks (Fig. 6, upper right). Dredge # 10 (not shown in Fig. 6) became temporarily stuck on the seabed but eventually recovered large fragments of basalts with fresh broken surfaces comprising 24% of all dredged material recovered at the site. The basalts are holocrystalline with prismatic plagioclase.
Hornblende and pyroxene are visible to the eye and no recrystallisation features are observed.

Interpretation
In situ material or not?
The level top surface of the Sverdrup Bank and exposure of basement are most likely due to erosion by deep draft icebergs and/or ice shelves through the Pleistocene (Kristoffersen et al., 2004;Gebhardt et al., 2011). Moving ice resting on bedrock plucks off material and moves debris in a basal zone where isometric shapes rotate more easily and lead to greater abrasion and more equidimensional debris which quickly attains a sub-angular roundness mode (Boulton, 1978;Bennett et al., 1997). Clast shape is influenced by the lithology for rocks with a distinct foliation whereas blocky samples are more robust (Bennett et al., 1997).
Rocks recovered by dredging always raise the question of genuine in situ representation.  (14) and Northern Basement High (7). The results of Riefstahl et al. (2013) from a single dredge given in parenthesis.
logies and metamorphic assemblages were dominant with no exceptions (Fig. 6, right panel).
The eleven attempts to dredge on the eastern perimeter of the bank swept over an area of more than 40 square kilometres and the recovery from each deployment ranged from 0 (2 hauls) to more than 15 (3 hauls) rock samples. Maximum recovery was narrowed down to a 5 km-long and 1.5 km-wide NWtrending area where also one dredge got stuck and was lost (Fig. 6). All three short dredge hauls at the southern site on Svalbard Bank were successful (10-16 samples). The local character of rock concentration on the seabed is also illustrated by the fact that four attempts to dredge on the western slope of Sverdrup Bank covering a distance of 13 kilometres were unsuccessful except for a single gneiss sample ( Fig. 6, dredges 11-14/2011). The dredge site of Riefstahl et al. (2013) is within a kilometre of where one of these unsuccessful attempts was terminated (Fig. 6).
On the Northern Basement High, we had maximum rock recovery at a basement ridge also documented from multibeam bathymetry by Riefstahl et al. (2013).
If we classify the shapes of our rock samples into four categories; well-rounded, rounded, sub-angular and angular, an average of 13% of the rocks from Yermak Plateau may be categorised as rounded (Fig. 7  northern sites on the Yermak Plateau are generally twice as abundant as the shapes of clasts entrained in the same glacial sediment wedge mentioned above. In particular, we note a platy rock specimen in one dredge (4-2010) on Sverdrup Bank (Fig. 7, upper panel). The southern site on Sverdrup Bank, however, has a relatively low abundance of angular shapes, but the sum of angular and sub-angular shapes is just as high as at the two other sites to the north (Fig. 7, lower panel).
Another aspect is repeated recovery. Dolerite rock samples are present in all dredges from the main Sverdrup Bank which stongly suggests proximity to a local source rather than random glacial erractics.
Two of the specimens recovered from the main Sverdrup Bank outcrop were considered 'exotic' rocks ( Fig. 6). Thin-sections of these samples reveal a medium-grained, weakly schistose calcareous schist where centimetre-long radial bundles of wollastonite overgrow the elongate calcite grains. Although this lithology has no known analogue in the geology of northern Svalbard or Nordaustlandet, we note that the samples were recovered in two different dredge hauls and therefore may be less 'exotic' than originally thought. We conclude that a contribution from ice-rafted material may be present in the rock populations recovered from the Sverdrup Bank, but not of any magnitude which could significantly influence our conclusions.

Sverdrup Bank: geological interpretation of dredged material
The second author has made a qualitative visual assessment of each rock specimen in the total inventory of 269 samples based on a personal knowledge base gained from more than twenty seasons of geological field mapping of the basement rocks of Spitsbergen and Nordaustlandet (Fig. 2). Thinsections were made from 17 samples for closer inspection. The samples were grouped following the criteria given in Table 2. suggests that the entire southern and northwestern part of the Yermak Plateau is part of a continental crustal structure (Feden et al., 1979;Jackson et al., 1984;Jokat et al., 2008).
Sedimentary rocks (53%) dominate the recovery at the Northern Basement High on the Yermak Plateau where the metamorphic contribution is reduced to 20% (Fig. 6). This implies that some of the cover rocks above metamorphic basement are retained at this location and the presence of volcanic rocks (24%) is significant.

Yermak Plateau -a continental outlier
Our successful dredge hauls from the top of the Sverdrup Bank traversed a more than 40 square kilometre area and returned internally consistent lithological representations (Fig. 6). As shown in Table   1, our results contrast with the relative abundances recovered by a single dredge in each area reported by Riefstahl et al. (2013). Most notably is a scarcity of metamorphic rocks (7% vs. our 63-84%) and abundance of igneous rocks (63% vs. our 3-13%) in the dredges recovered by Riefstahl et al. (2013) from the Sverdrup Bank. Also, the complete lack of consolidated sediments (0% vs. our 53%) in their dredge from the Northern Basement High. All our recovered lithologies have analogues in the geology of The continental rocks on Yermak Plateau are separated from the continental margin north of Spitsbergen by the deep-water embayment which forms the Sophia Basin (Fig. 2). The gravity data suggest the basin is floored by a high-density (2.9 g/cm3) crust interpreted as extended and intruded continental rocks (Geissler & Jokat, 2004). If the Sophia Basin formed by crustal extension, we need to consider at least two issues: i) First, an overlap arises between the northeastern tip of Yermak Plateau and the continental margin north of Nordaustlandet if we close the Sophia Basin by back-tracking the present northeast-trending part of the plateau along a direction either parallel to the postulated de Geer Fault or the initial opening of the Eurasian Basin (Fig. 2).
ii) Secondly, a plate geometry which facilitates opening of the Sophia Basin requires initial dextral shear motion along the west coast of Spitsbergen through Forlandsundet on the order of 100-150 km.
This offset range is estimated from the separation of acoustic basement of the plateau from an uncertain basement configuration of the continental margin north of Spitsbergen (Fig. 3B) and assuming that extension in the deep basin was compensated by thinning of continental crust and intrusions. We use an offset range because of the uncertain constraints on acoustic basement north of Spitsbergen and the crust below the Sophia Basin.
The circular structure of the northeasternmost tip of Yermak Plateau presents an apparent enigma unless we postulate the feature to be a younger post-rift structure, possibly constructed by volcanism between Chron 22 and Chron 18 based on the age of the adjacent oceanic crust. South-dipping reflections and seismic velocities (>3.2 km/s) of the acoustic basement (Fig. 3A), as well as a partial magnetic anomaly signature (Jokat et al., 2008), may be circumstantial evidence for the northeasternmost part of the plateau being a volcanic wedge.
The startling lack of exposures of major strike-slip faults along the west coast of Spitsbergen parallel to the relative plate motion has long remained a puzzle (Craddock et al., 1985;Dallmann et al., 1988;Maher & Craddock, 1988;CASE Team, 2001;Dallmann, 2015). Exceptions are the >70 km-long Svartfjella, Eidembukta and Daudmannsodden lineaments NW of Isfjorden (Fig. 5) with several kilometres of offset (Maher et al., 1997). The proposed solution to the conundrum is decoupling of stresses in the brittle upper crust between a normal component creating the fold-and-thrust belt and a tangential component responsible for simple shear motion in a proposed zone off the coast of West Spitsbergen (Maher & Craddock, 1988;Nøttvedt et al., 1988). Geological field studies of partitioning of slip in the upper crust along obliquely convergent plate boundaries have been published by Lee et al. (1998), McCaffrey et al. (2000 and , and the spatial scale over which partitioning often occurs is documented by geodetic and seismic networks (Lettis & Hanson, 1991;Wdowinski et al., 2001). Strain partitioning requires anisotropic rocks (Jones and Tanner, 1995).

An initial dextral fault off the coast of West Spitsbergen -the de Geer Fault?
If the Hornsund Fault Zone indeed represented the early Cenozoic strike-slip plate boundary between Svalbard and Greenland, the crustal extension in the Sophia Basin should have propagated westwards to meet the fault and isolated the Yermak Plateau as a continental fragment to the north (Figs. 2 & 8).
Instead, the Sophia Basin is terminated to the west by NNW-trending basement ridges which continue north from Spitsbergen. From this, we infer that the eastern flank of the largest basement ridge on the southern Yermak Plateau (Fig. 2) must have represented the plate boundary at the time and was connected to the south to a coast-parallel transform fault (de Geer Fault) which due to later deformation was offset by a transfer zone (Ritzmann & Jokat, 2003) now occupied by the Danskøya Basin (Fig. 2). We present a new geophysical argument for the past existence of the de Geer Fault On the origin of the Yermark Plateau north of Svalbard  Brevik et al. (2003) , transect Hornsund from Ljones et al. (2004 ), transect Bellsund from Ritzmann et al. (2002), the transect Prins Karls Forland North from , from 79°45'N from Czuba et al. (1999) and from NW Spitsbergen North from Ritzmann & Jokat (2003).
based on the fact that major strike-slip faults in the continental crust are associated with a geophysically defined fault damage zone roughly proportional to the fault length and/or displacement of the fault (Ben-Zion & Sammis, 2003;Faulkner et al., 2003;Sibson, 2003;Mooney et al., 2007). The geophysical signature of the damage zone extends to at least 3-5 km depth, and is associated with 20-50% lower seismic velocity than the neighbouring rocks regardless of crustal rock composition (Ben-Zion et al., 2007;Mooney et al., 2007).
Published transects of the seismic velocity distribution in the crust offshore West Spitsbergen all show significant velocity perturbations in the upper brittle crust (Fig. 8). All investigators interpret these deep velocity anomalies as domains of intense faulting (Ritzmann et al., 2002;Breivik et al., 2003;Ritzmann and Jokat, 2003;Ljones et al., 2004;Czuba et al.,2005). The perturbation north of Spitsbergen below the Danskøya Basin is considered to represent a south-plunging crustal-scale detachment (Ritzmann & Jokat, 2003), and to the west (Fig. 8, Transect 79° 45'N) an old transform zone is manifested as a ˞1 km/s seaward velocity increase at mid-crustal level (Czuba et al., 2008). Lower velocities are associated with a <20 km-wide zone down to 15 km depth in the Prins Karls Forland North and Bellsund transects, while the Hornsund transect shows a graben underlain by a high-velocity ridge reaching 5 km depth (Figs. 8 & 9). To the south of Spitsbergen (75⁰ 30' N), the coast-parallel trend of crustal velocity perturbations merges with the Hornsund Fault Zone and the continent-ocean boundary (Fig. 8).
We suggest that the significant variations in seismic velocities in the upper crust along the coast of West Spitsbergen landward of the Hornsund Fault Zone compare with other large-offset strike-slip environments and support past strain-partitioned, orogen-parallel dextral displacements along the coast of Spitsbergen of a magnitude not hitherto considered (Maher & Craddock, 1988;Nøttvedt et al., 1988;Morris, 1989;Gabrielsen et al., 1992;Dallmann et al., 1993;Ohta et al., 1995;Maher et al., 1997;Kleinsphen & Teyssier, 2016). Also, this coast-parallel shear motion satisfies a plate boundary geometry which allows for opening of the Sophia Basin, because the basin is bypassed to the west by activity along the Hornsund Fault Zone (Fig. 2). We note, however, that the shear motion associated with the opening of the Sophia Basin (100-150 km) roughly 2/3 half of the predicted relative motion (200 km)  , from Bellsund from Ritzmann et al., (2002), from the San Andreas Fault from Murphy et al. (2010) and Wallace (1991), from the Dead Sea Transform from ten Brink et al. (2006) and DESERT Group (2004 between Greenland and Svalbard during the seven million year time span between Chron 24 and Chron 21 (Fig. 10). Further displacement must have bypassed the basin and occurred farther west along the Hornsund Fault Zone (Fig. 2). Continued activity on the de Geer Fault was possible if the fault continued north across the Yermak Plateau to link up with the Ellesmere domain (Fig. 12).

Figure 10. A compilation of plate motion parameters relevant to early Cenozoic tectonic events north of Svalbard and
Greenland. Spreading half-rates and the direction of opening in the Norwegian-Greenland Sea keeping Europe fixed are from Gaina et al. (2019). The spreading rates from the Labrador Sea published by  and Kristoffersen & Talwani (1977) are corrected for the difference relative to the geomagnetic time scale of Gradstein et al. (2012) used here. The stratigraphy of the Central Basin is from Dallmann (2015) and from the Forlandsundet basin from Kleinspehn & Teyssier (2016

Plate interaction north of Greenland and Svalbard -the state of knowledge
The overlap created by the Yermak Plateau and Morris Jesup Rise evident in early reconstructions (LePichon et al., 1977;Feden et al., 1979) of the Eurasia Basin has remained a puzzle and recent plate tectonic models are vague with respect to details of Paleogene tectonic scenarios north of Svalbard and Greenland (Fig. 11). Minakov et al. (2012) focus on explanations for a relatively narrow (c. 100 km) transition between the point of rapid thinning of the continental crust at the margins of the Eurasia Basin as determined by the horizontal gradient in the Bouguer gravity field and oceanic crust associated with the first identifiable magnetic isochron (Chron 24). Little attention is given to events which involved the crustal domain north of Grenland and Svalbard (Fig. 11, panel A). Similarly, Døssing et al. (2013b and leave out the details of the Yermak Plateau and Morris Jesup Spur and rise in their reconstructions of the Eurasia Basin north of Greenland (Fig. 11, panel B). A gravity low north of Greenland highlighted by Brozena et al. (2003) as a loading effect generated by crustal shortening is also postulated by Døssing et al. (2014) to relate to a Lincoln Sea -Klenova Valley Fault Zone (LKFZ). The fault zone extends from the western Lincoln Sea shelf and passes eastwards at the foot of the slope north of the Morris Jesup rise and spur (Fig. 11, panel B). The LKFZ projects the domain of Eurekan deformation from the southern boundary of the Pearya terrane offshore, and the Morris Jesup structure is interpreted as a volcanic province formed on an extended continental margin during peak-Eurekan  Greenland (GRN) relative to Svalbard (European plate fixed; EUR) at Chron 24 using Gplates software (www.gplates.org) and the rotation parameters from Gaina et al., (2002) and Barnett-Moore et al. (2016), respectively. The conceptual outline of tectonic blocks on Ellesmere Island involved in the Eurekan Orogeny (brown colour) are adapted from Piepjohn et al. (2016). Active faults shown by red colour. (B) Reconstruction for Chron 21 (47 Ma.) using parameters from Gaina et al. (2002) for the rotation of North America relative to Europe and from Gaina et al., (2009) Gaina et al. (2002) for the rotation of North America relative to Europe and from Gaina et al., (2009) Fig. 11, panel D. We note their position of Greenland derived from the rotation parameters of Srivastava & Tapscott (1986) is about 100 km to the south relative to Spitsbergen when compared to more recent plate rotation parameters, e.g., Gaina et al. (2009). In Ellesmere Island, motion on the major faults was sinistral during  Lyberis & Manby (1993). Extensive shortening (>42%) documented along the inner part of St. Jonsfjorden ( Fig. 5) amounts to about 13 km (Welbon & Maher, 1992), and Bergh et al. (1997) estimated 45% shortening over a 13 km section north of Isfjorden to be a minimum of 20 km. South of Bellsund (Fig. 5), the shortening is 8 km over a 5 km section (von Gosen & Piepjohn, 2001) and >8 km across the fold belt in the Hornsund (Fig. 5) area (Dallmann, 1992). In summary, there appears to be reasonable agreement between the predicted and observed magnitudes of contraction across the West Spitsbergen Fold and Thrust Belt. However, the sum of strike-slip components observed from the land geology are more than an order of magnitude short of the value predicted from plate motion. We have argued for a coastparallel major dislocation (de Geer Fault) with an offset of 100-150 km required for opening of the Sophia Basin (Figs. 2 & 12). Continued relative motion between Greenland and Svalbard bypassed the Sophia Basin either by northward continuation of the de Geer Fault and/or motion along the a proto-Hornsund Fault Zone to the west (Fig. 2).
Following Srivastava & Tapscott (1986) and Gaina et al. (2002), we assume the Lomonosov Ridge and the Pearya terrane were part of the North American plate during opening of the Eurasia Basin (Fig. 12).
Since a significant tract of crust is present in the Eurasia Basin between Chron 24 and the continentocean transition as defined from gravity and seismic data along the Eurasia Basin margins (Jackson and Gunnarson, 1990; fig. 3 of Glebovsky et al., 2006;Chernykh & Krylov, 2011;Minakov et al., 2012), we assume an initial opening at Chron 25 (56 Ma.). Using the GPlates software (www.gplates.org), we keep Europe fixed and reconstruct the relative position of the Lomonosov Ridge at the time of Chron 25y (56 Ma) using the finite rotation of Gaina et al. (2002)  Terrane (Jokat et al., 2016) projects eastward across the Lincoln Sea continental shelf towards an offset at the continental margin north of Nordaustlandet (Fig. 12). The linking trend may suggest a continuous fault trace. Alternatively, the Mt. Rawlinson Fault may be related to the more northeasterly gravity trend G4 of Døssing et al. (2013b).
ii) During the latest Paleocene/earliest Eocene, several large fault zones on Ellesmere Island show sinistral strike-slip motion in a partly transpressive regime, i.e., Eurekan Stage 1 of Piepjohn et al. (2016).
This indicates that the earliest opening of the Eurasia Basin was partly accommodated by sinistral strike-slip between crustal blocks of the present Ellesmere Island as a result of the North American plate including the Pearya terrane moving relative to Greenland (Fig. 12).
iii) The initial rifting of the Eurasia Basin included the Yermak Plateau north and west of a small  (Kleinspehn & Teyssier, 2016).
The larger picture of the complexity of contemporary parallel faults between the Mohns and Gakkel spreading centres during the Eocene is manifested to the south by detached slivers of continental crust such as the Hovgard Ridge (Johnson & Eckhoff, 1966;Myhre & Eldholm, 1988;Faleide et al., 2008) and the East Greenland Ridge (Døssing & Funck, 2012 . 12) was formed in the Carboniferous (Håkansson & Pedersen, 1982;Pedersen & Håkansson, 1999).
Modelling studies also advocate an early Cenozoic strike-slip boundary proximal to the basementinvolved foreland thrust complexes along the coast of West Spitsbergen (Leever et al., 2011).
An analogue model representing two sediment-covered basement plates forced by dextral motion with 15° convergence to produce a doubly vergent wedge; a largely undeformed retro-wedge to the west, and an internally deformed and tapered pro-wedge towards the east. The surface expression of the shear zone which separates the two parts and their respective differences in the degree of basement-involved thrusting, is positioned slightly to the west of the main dextral plate motion. Leever et al. (2011) suggested that the main shear zone may relate to slip close to the present coasta site which later developed into the coast-parallel graben (Kleinspehn & Teyssier, 2016). Shear at the western boundary of the retro-wedge may represent a zone of weakness co-located with the Hornsund Fault Zone. The model also suggests deformation affected only the Eurasia plate and did not extend significantly into the interior of the juxtaposed Greenland plate.
The continuation of the de Geer Fault north of Spitsbergen interpreted as the eastern edge of the crustal block which includes Sverdrup Bank is offset to the east with respect to the trend of the coast-parallel graben (Figs. 2, 5 & 8). The NE-trending Danskøya Basin which connects these two trends, is located on top of a crustal-scale detachment plane (Fig. 8, northern crustal transect). We suggest this offset in the de Geer Fault and possibly also the Kap Cannon Thrust (von Gosen & Piepjohn, 1999;Tegner et al., 2011) and motion on the Harder Fiord Fault Zone (von Gosen & Piepjohn, 2003) could be the result of stresses caused by a restraining bend of the Hornsund Fault Zone or associated splays connecting to the west with faults in the Ellesmere domain (Fig. 12). The NE-offset of the de Geer Fault north of Spitsbergen is analogous to observations from the exposed thrust-and-fold belt between the Ymerbukta Fault and Kongsfjorden (Bergh et al., 2000;Piepjohn et al., 2001; where the motion is slightly different from the easterly direction of tectonic transport south of Isfjorden (Dallmann, 2015). Here, the possible explanations for the difference in tectonic transport direction along the fold-and-thrust belt are changes in stresses at a restraining bend of the regional transform fault in the Kongsfjorden area (Steel et al., 1985;Gabrielsen et al., 1992), pinning or buttressing against a fault  or basement north of Kongsfjorden (Dallmann et al., 1993;Lyberis & Manby, 1993) or variable degrees of coupled and decoupled transpressional motion (Maher & Craddock, 1988;Maher et al., 1997;Bergh et al., 2000).