1. Introduction
The continental blocks bordering the North Atlantic Ocean today contain remnants of several large-scale orogenic belts ranging in age from the Archaean to the Palaeozoic (Fig. 1). One of these belts, the Caledonian orogen, formed in the Ordovician to Devonian owing to the collision of the Laurentian, Avalonian and Baltican continents, and was later disrupted during the opening of the North Atlantic Ocean (Fig. 1; e.g. Roberts & Gee, Reference Roberts, Gee, Gee and Sturt1985; Torsvik et al. Reference Torsvik, Smethurst, Meert, Van der Voo, McKerrow, Brasier, Sturt and Walderhaug1996; McKerrow, Niocaill & Dewey, Reference McKerrow, Niocaill and Dewey2000; Higgins & Leslie, Reference Higgins, Leslie, Higgins, Gilotti and Smith2008; Gee et al. Reference Gee, Fossen, Henriksen and Higgins2008). The Caledonian orogeny fundamentally altered the margins of the involved continents and led to large-scale nappe- and terrane-displacements (e.g. Roberts, Nordgulen & Melezhik, Reference Roberts, Nordgulen, Melezhik, Hatcher, Carlson, McBride and Martínez Catalán2007; Augland, Andresen & Corfu, Reference Augland, Andresen and Corfu2011).
The role and place of the different pre-Devonian parts of Svalbard before, during and after the Caledonian orogeny have been a matter of debate (e.g. Harland, Reference Harland1971, Reference Harland, Gee and Sturt1985; Ohta, Reference Ohta1994; Gee & Teben'kov, Reference Gee, Teben'kov, Gee and Pease2004; Labrousse et al. Reference Labrousse, Elvevold, Lepvrier and Agard2008; Mazur et al. Reference Mazur, Czerny, Majka, Manecki, Holm, Smyrak and Wypych2009; Petterson, Pease & Frei, Reference Petterson, Pease and Frei2010). Pre-Devonian rocks occur on Nordaustlandet, Ny Friesland, northwestern Spitsbergen north of Kongsfjorden and along the west coast of Spitsbergen (Figs 1, 2). The rocks from Nordaustlandet, Ny Friesland and northwestern Spitsbergen (Fig. 2) are lithologically similar to rocks within the Caledonian nappes of East Greenland south of 76°N (e.g. Gee & Teben'kov, Reference Gee, Teben'kov, Gee and Pease2004). U–Pb dating of magmatic and detrital zircons from various units in these areas has shown that the rocks of Nordaustlandet, Ny Friesland, northwestern Spitsbergen and the Caledonian nappes of East Greenland indeed share a common tectonic history of latest Mesoproterozoic sedimentation, c. 950 Ma granitic magmatism, Neoproterozoic to Ordovician sedimentation and widespread Scandian (c. 440–410 Ma) magmatism and metamorphism (Fig. 2; Johansson et al. Reference Johansson, Gee, Larionov, Ohta and Tebenkov2005; Myhre, Corfu & Andresen, Reference Myhre, Corfu and Andresen2009; Petterson, Pease & Frei, Reference Petterson, Tebenkov, Larionov, Andresen and Pease2009; Petterson et al. Reference Petterson, Pease and Frei2009). In contrast, the pre-Devonian rocks south of Kongsfjorden (Fig. 2) are different: they include a c. 1100–1200 Ma magmatic complex, Neoproterozoic rocks metamorphosed during a c. 650 Ma tectonometamorphic event, a Lower to Middle Ordovician blueschist to eclogite-facies complex, at least three marked unconformities of Neoproterozoic, Cambrian and Ordovician age, and no evidence for Scandian magmatism or high-grade metamorphism (Figs 2–5; e.g. Gee & Teben'kov, Reference Gee, Teben'kov, Gee and Pease2004; Labrousse et al. Reference Labrousse, Elvevold, Lepvrier and Agard2008; Majka et al. Reference Majka, Mazur, Manecki, Czerny and Holm2008; Mazur et al. Reference Mazur, Czerny, Majka, Manecki, Holm, Smyrak and Wypych2009; Czerny et al. Reference Czerny, Majka, Gee, Manecki and Manecki2010). How and where the tectonostratigraphy of the western coast of Spitsbergen formed within the framework of the Mesoproterozoic to Carboniferous evolution of the North Atlantic region, including its evolution during the Caledonian orogeny, is unknown so far.
Figure 2. Geological overview of the Svalbard archipelago, with the pre-Devonian rocks highlighted. Insets: schematic representation of the pre-Devonian geology of different parts of Svalbard. Abbreviations: BBF – Breibogen–Bockfjorden Fault; BF – Billefjorden Fault; LF – Lomfjorden Fault; RF – Raudfjorden Fault; VKZ – Vimsodden–Kosibapasset shear zone. Numbered units: 1 – Brennevinsfjorden/Helvetesflya units; 2 – Murchisonfjorden and Lomfjorden groups; 3 – Hinlopenstretet Supergroup; 4 – Atomfjella Complex; 5 – Mont Blanc and Biscayarhuken units and Richarddalen Complex; 6 – Krossfjorden/Smeerenburg Complex; 7 – Bullbreen Group; 8 – Vestgötabreen Complex; 9 – Comfortlessbreen diamictites; 10 – Daudmannsodden unit; 11 – St Jonsfjorden unit; 12 – Kapp Lyell diamictites; 13 – Sofiebogen Group; 14 – Deilegga/Nordbukta unit and Magnethøgda/Berzelius igneous suite, probably tectonically intercalated; 15 – Eimfjellet Complex; 16 – Isbjørnhamna Group. Small numbers in insets refer to age data in Ma. References are given in the text.
Figure 3. Geological map of pre-Devonian rocks of the western coast of Spitsbergen south of Kongsfjorden (simplified after Dallmann et al. Reference Dallmann, Ohta, Elvevold and Blomeier2002). The location of Figure 3 is indicated on Figure 2. The white areas within the pre-Devonian rocks correspond to glaciated areas. Numbers in boxes correspond to lithotectonic units from the map of Dallmann et al. (Reference Dallmann, Ohta, Elvevold and Blomeier2002), which are also used in Figure 4.
Figure 4. Tectonostratigraphic sketch for pre-Devonian rocks, western coast of Spitsbergen south of Kongsfjorden. Black circles represent samples analyzed in this study. Note that sample 6 is from Carboniferous strata overlying the pre-Devonian rocks and is not shown in this figure. The location of sample 7 is not given either since its stratigraphic position is unclear (see text). Note the gap in the timescale to the left. The numbers in boxes are explained in the legend of Figure 3. Geochronological information is given in Ma, with the references given in brackets: (1) Bernard-Griffiths, Peucat & Ohta (Reference Bernard-Griffiths, Peucat and Ohta1993); (2) Dallmann et al. (1990); (3) Bjørnerud (Reference Bjørnerud2010); (4) Czerny et al. (Reference Czerny, Majka, Gee, Manecki and Manecki2010); (5) Balashov et al. (Reference Balashov, Tebenkov, Ohta, Larionov, Sirotkin, Gannibal and Ryungenen1995); (6) Mazur et al. (Reference Mazur, Czerny, Majka, Manecki, Holm, Smyrak and Wypych2009); (7) Majka et al. (Reference Majka, Mazur, Manecki, Czerny and Holm2008); (8) Majka et al. (Reference Majka, Czerny, Mazur, Holm and Manecki2010); (9) Manecki et al. (Reference Manecki, Holm, Czerny and Lux1998); (10) Larionov et al. (Reference Larionov, Tebenkov, Gee, Czerny and Majka2010). Arrows with question marks indicate poorly defined depositional ages. Abbreviations: NJWL – Northern Wedel Jarlsberg Land; SWJL – Southern Wedel Jarlsberg Land; VKZ – Vimsodden–Kosibapasset shear zone. Abbreviations within units: (34) HP-LT – high-pressure low-temperature metamorphic event; (40) C – Comfortlessbreen, T – Trondheimfjella, L – Lågneset, K – Kapp Lyell; (48) D – Daudmannsodden, L – Lågnesbukta, S – Sofiebogen; (50) D – Daudmannsodden, M – Moefjellet, L – Lågnesrabbane, S – Slettfjelldalen; (51) S – Sofiebogen; (52) M – Kapp Martin, S – Slyngfjellet.
Figure 5. Geological map of the study area according to Bergh et al. (Reference Bergh, Ohta, Andresen, Maher, Braathen and Dallmann2003) with sample locations indicated.
Detrital zircon geochronology is widely used for characterizing the depositional age and potential source areas of non-fossiliferous sediments in various tectonic settings (e.g. Fedo, Sircombe & Rainbird, Reference Fedo, Sircombe, Rainbird, Hanchar and Hoskin2003; Carrapa, Reference Carrapa2010). Detrital zircon data now exist for many (meta)-sedimentary sequences within the North Atlantic region and have been useful for tracking the tectonic history of various continental fragments and sedimentary basins throughout the Proterozoic to Palaeozoic (e.g. Cawood et al. Reference Cawood, Nemchin, Smith and Loewy2003, Reference Cawood, Nemchin and Strachan2007; Cawood, Nemchin & Strachan, Reference Cawood, Nemchin and Strachan2007; Kirkland, Daly & Whitehouse, Reference Kirkland, Daly and Whitehouse2007; Petterson, Pease & Frei, Reference Petterson, Pease and Frei2009, Reference Petterson, Pease and Frei2010; Be'eri-Shlevin et al. Reference Be'eri-Shlevin, Gee, Claesson, Ladenberger, Majka, Kirkland, Robinson and Frei2011; Bingen, Belousova & Griffin, Reference Bingen, Belousova and Griffin2011; Kirkland et al. Reference Kirkland, Bingen, Whitehouse, Beyer and Griffin2011; Slama et al. Reference Slama, Walderhaug, Fonneland, Kosler and Pedersen2011). In this contribution we explore the potential of detrital zircon dating for constraining the tectonic evolution and provenance of Proterozoic to Carboniferous strata from the western coast of Spitsbergen south of Kongsfjorden (Fig. 2). As stated above, the tectonic history of this area differs from other pre-Devonian rocks of Spitsbergen, and so far no detrital zircon data is available from this c. 300 km long segment of pre-Devonian rocks. We present data from six samples covering the entire stratigraphic range available in the northern part of the area below and above the Ordovician high-pressure complex. In addition, we present one U–Pb zircon thermal ionization mass spectrometry (TIMS) age of an augengneiss occurring in this area (Figs 2–5). Based on these data we then discuss whether and how these data can be used to better constrain the evolution of this part of Svalbard from the latest Mesoproterozoic to the Carboniferous relative to other units exposed in the North Atlantic region.
2. Geological background
The archipelago of Svalbard is located on the northwestern corner of the Barents Sea Shelf, and was uplifted during Late Mesozoic to Cenozoic tectonic events (Fig. 1). It exposes various rocks ranging in age from the Archaean to Neogene, which can be grouped into the following successions (Fig. 2): (1) pre-Devonian rocks, (2) Devonian sedimentary rocks, (3) Carboniferous to Permian sedimentary rocks, (4) Mesozoic sedimentary and volcanic rocks, and (5) Cenozoic sedimentary and volcanic rocks (e.g. Dallmann et al. Reference Dallmann, Ohta, Elvevold and Blomeier2002). The pre-Devonian rocks relevant for understanding the Caledonian tectonic history crop out in several distinct regions: Nordaustlandet, Ny Friesland, the northwestern corner of Spitsbergen north of Kongsfjorden, and along the southwestern coast of Spitsbergen south of Kongsfjorden (Figs 2, 3).
2.a. Pre-Devonian rocks of Nordaustlandet, Ny Friesland and northwestern Spitsbergen
In Nordaustlandet and eastern Ny Friesland, Meso- to Neoproterozoic sediments (Brennevinsfjorden and Helvetesflya units) are unconformably overlain by c. 960 Ma andesitic and rhyolitic volcanic rocks and are intruded by 960–940 Ma granites (Fig. 2; Gee & Teben'kov, Reference Gee, Teben'kov, Gee and Pease2004; Johansson et al. Reference Johansson, Gee, Larionov, Ohta and Tebenkov2005). A Neoproterozoic platform succession (Murchisonfjorden and Lomfjorden groups), as well as Late Neoproterozoic (‘Vendian’) tillites and Cambro-Ordovician sediments (Hinlopenstretet Supergroup) occur unconformably on top of the deformed and metamorphosed Meso- to Neoproterozoic basement (Fig. 2). These successions were then folded and metamorphosed during the main Scandian phase of the Caledonian orogeny (at c. 450–410 Ma, Fig. 2; Gee & Teben'kov, Reference Gee, Teben'kov, Gee and Pease2004; Johansson et al. Reference Johansson, Gee, Larionov, Ohta and Tebenkov2005). In western Ny Friesland, an amphibolite-facies stack of Palaeoproterozoic to Archaean orthogneisses with interlayered Palaeo- to Mesoproterozoic sedimentary and volcanic rocks is exposed in a large-scale Caledonian antiform, the Atomfjella anticlinorium (Fig. 2; Hellman et al. Reference Hellman, Gee, Johansson and Witt-Nilsson1997; Hellman, Gee & Witt-Nilsson, Reference Hellman, Gee and Witt-Nilsson2001; Gee & Teben'kov, Reference Gee, Teben'kov, Gee and Pease2004). In the northwestern corner of Spitsbergen, c. 960 Ma and 435–415 Ma (Scandian) granites occur in migmatized Meso- to Early Neoproterozoic sediments (Krossfjorden Group and Smeerenburg Complex, Fig. 2; Myhre, Corfu & Andresen, Reference Myhre, Corfu and Andresen2009; Petterson, Pease & Frei, Reference Petterson, Tebenkov, Larionov, Andresen and Pease2009; Petterson et al. Reference Petterson, Pease and Frei2009). On Biscayarhalvøya (Fig. 2), mica schists and phyllites (Mont Blanc and Biscayarhuken units) as well as eclogite-bearing gneisses, marbles and schists (Richarddalen Complex) tectonically overlie the migmatized sediments (Fig. 2; Gee & Teben'kov, Reference Gee, Teben'kov, Gee and Pease2004; Labrousse et al. Reference Labrousse, Elvevold, Lepvrier and Agard2008). The Richarddalen Complex contains c. 960 Ma and probably 650 Ma orthogneisses, eclogite-facies metamorphism occurred prior to c. 455 Ma, and Scandian metamorphic ages have not been reported from this complex (Gromet & Gee, Reference Gromet and Gee1998).
2.b. Pre-Devonian rocks of western Spitsbergen south of Kongsfjorden
Since the tectonostratigraphy of pre-Devonian rocks of the western coast of Spitsbergen south of Kongsfjorden (Figs 2, 3) is especially relevant for the current study, it is treated in more detail below and is summarized in Figures 3 and 4. Unravelling the geological evolution of this region is complicated by the fact that this area lies within the Palaeogene West Spitsbergen fold-and-thrust-belt (WSFTB, Fig. 2; e.g. Bergh, Braathen & Andresen, Reference Bergh, Braathen and Andresen1997; Braathen, Bergh & Maher, Reference Braathen, Bergh and Maher1999). There is currently no agreement about whether all these rocks belong to the same pre-Devonian tectonostratigraphic sequence, or whether they represent different terranes juxtaposed before, during or after the Caledonian orogeny. In the traditional terrane model of Harland (Reference Harland, Gee and Sturt1985), he separated the area into a northwestern and southeastern terrane along a proposed strike-slip fault running below the Recherchebreen–Torellbreen glaciers on northern Wedel Jarlsberg Land (Fig. 3). To the northwest of this proposed fault, Upper Neoproterozoic tillites are present whereas Cambrian–Ordovician sediments are absent, opposite to what is present to the southeast of this proposed fault (Figs 3, 4). However, according to Bjørnerud, Decker & Craddock (Reference Bjørnerud, Decker and Craddock1991) and Gee & Teben'kov (Reference Gee, Teben'kov, Gee and Pease2004), there is no field evidence for the existence of such a major fault running below these glaciers. Apart from the differences in stratigraphy, the rocks on both sides of this proposed fault are similar, suggesting that they belong to the same terrane.
Recently, Mazur et al. (Reference Mazur, Czerny, Majka, Manecki, Holm, Smyrak and Wypych2009) proposed the existence of another (sinistral) strike-slip terrane boundary separating the southwestern corner of Wedel Jarlsberg Land from the rocks further to the northeast, the Vimsodden–Kosibapasset shear zone (VKZ, Figs 2, 3). To the southwest of this shear zone, a c. 1200 Ma igneous complex is exposed, which probably experienced a metamorphic overprint at c. 930 Ma (Eimfjellet Complex; Balashov et al. Reference Balashov, Tebenkov, Ohta, Larionov, Sirotkin, Gannibal and Ryungenen1995, Reference Balashov, Peucat, Tebenkov, Ohta, Larionov, Sirotkin and Bjørnerud1996; Larionov et al. Reference Larionov, Tebenkov, Gee, Czerny and Majka2010). This complex is thrust on top of Neoproterozoic (< 695 Ma) sediments (Isbjørnhamna Group; Larionov et al. Reference Larionov, Tebenkov, Gee, Czerny and Majka2010). Coeval with thrusting, the Isbjørnhamna Group was metamorphosed up to amphibolite-facies grade during the Late Neoproterozoic (c. 650 Ma), followed by slow cooling until c. 480 Ma (Manecki et al. Reference Manecki, Holm, Czerny and Lux1998; Majka et al. Reference Majka, Mazur, Manecki, Czerny and Holm2008, Reference Majka, Czerny, Mazur, Holm and Manecki2010, Reference Majka, Larionov, Gee, Czerny and Prsek2012; Mazur et al. Reference Mazur, Czerny, Majka, Manecki, Holm, Smyrak and Wypych2009).
The rocks to the northeast of the Vimsodden–Kosibapasset shear zone, particularly well studied between Bellsund and Hornsund on Wedel Jarlsberg Land (Fig. 3), consist of an older and a younger sequence separated by a major unconformity, the Torrelian unconformity (e.g. Bjørnerud, Reference Bjørnerud1990; Bjørnerud, Craddock & Wills, Reference Bjørnerud1990; Bjørnerud, Decker & Craddock, Reference Bjørnerud, Decker and Craddock1991). Below this unconformity, various deformed metasedimentary and metavolcanic rocks are exposed (the Nordbukta, Deilegga and Magnethøgda sequences; Figs 3, 4; Bjørnerud, Reference Bjørnerud1990; Dallmann et al. Reference Dallmann, Ohta, Elvevold and Blomeier2002; Mazur et al. Reference Mazur, Czerny, Majka, Manecki, Holm, Smyrak and Wypych2009). The metamorphic grade of these rocks varies from greenschist to amphibolite facies. Mazur et al. (Reference Mazur, Czerny, Majka, Manecki, Holm, Smyrak and Wypych2009) mentioned unpublished detrital monazite ages of c. 1098 Ma from the Deilegga Group, indicating that deposition of these rocks post-dates this age. According to Czerny et al. (Reference Czerny, Majka, Gee, Manecki and Manecki2010), the Magnethøgda sequence contains c. 950 Ma porphyritic rocks metamorphosed at c. 660 Ma. Detrital monazite of c. 650 Ma in the rocks overlying the unconformity (unpublished results mentioned by Mazur et al. Reference Mazur, Czerny, Majka, Manecki, Holm, Smyrak and Wypych2009; Czerny et al. Reference Czerny, Majka, Gee, Manecki and Manecki2010) indicate that the unconformity corresponds to the c. 650 Ma metamorphic event recorded in the rocks below. Above the unconformity, a series of conglomerates, limestones, dolomites and phyllites, intercalated with mafic magmatic and volcanic rocks, is exposed (Figs 3, 4; e.g. Harland, Hambrey & Waddans, Reference Harland, Hambrey and Waddans1993; Harland, Anderson & Manasrah, Reference Harland, Anderson and Manasrah1997; Dallmann et al. Reference Dallmann, Ohta, Elvevold and Blomeier2002). Above these deposits, there are thick diamictites, intercalated with phyllites and psammites (Fig. 4), which were probably deposited during the Late Neoproterozoic (c. 635 Ma; Harland, Hambrey & Waddans, Reference Harland, Hambrey and Waddans1993; Harland, Anderson & Manasrah, Reference Harland, Anderson and Manasrah1997; Bjørnerud, Reference Bjørnerud2010). These tillites appear to be missing south of central Wedel Jarlsberg Land (Figs 3, 4), where Cambro-Ordovician sediments of the Sofiekammen and Sørkapp Land groups directly and unconformably lie on top of Neoproterozoic phyllites (Figs 3, 4; Birkenmajer, Reference Birkenmajer1991). The Ordovician sediments of the Sørkapp Land Group contain faunas of unambiguous Laurentian affinity (e.g. Gee & Teben'kov, Reference Gee, Teben'kov, Gee and Pease2004).
North of Bellsund (Fig. 3), the pre-Devonian rocks are much less studied. Geological mapping indicates that the Torrelian unconformity continues northwards at least until Nordenskjöld Land (Fig. 3), where conglomerates occur on top of phyllites and limestones (Fig. 4; Harland, Hambrey & Waddans, Reference Harland, Hambrey and Waddans1993; Harland, Anderson & Manasrah, Reference Harland, Anderson and Manasrah1997; Dallmann et al. Reference Dallmann, Ohta, Elvevold and Blomeier2002). North of Isfjorden, a series of phyllites, limestones and intercalated volcanic rocks overlies a mafic volcanic suite (Trollheimen volcanics; Ohta, Reference Ohta1985; Fig. 4). On top of this sequence, diamictites interpreted as Neoproterozoic tillites occur (Fig. 4; Kanat & Morris, Reference Kanat and Morris1988; Harland, Hambrey & Waddans, Reference Harland, Hambrey and Waddans1993; Harland, Anderson & Manasrah, Reference Harland, Anderson and Manasrah1997; Dallmann et al. Reference Dallmann, Ohta, Elvevold and Blomeier2002). An exotic tectonic unit is thrust on top of these tillites: the blueschist to eclogite-facies Vestgötabreen Complex and the unconformably overlying Ordovician to Silurian Bullbreen Group (Figs 3–5; Ohta, Hiroi & Hirajima, Reference Ohta, Hiroi and Hirajima1983; Hirajima et al. Reference Hirajima, Banno, Hiroi and Ohta1988; Ohta et al. Reference Ohta, Krasilscikov, Lepvrier and Tebenkov1995; Agard et al. Reference Agard, Labrousse, Elvevold and Lepvrier2005; Labrousse et al. Reference Labrousse, Elvevold, Lepvrier and Agard2008). The protolith ages of the metamorphic rocks in this complex are not well known: the mafic protolith of an eclogite body yielded an upper intercept zircon age of 2121 ± 50 Ma and a Sm–Nd whole-rock model age of 2100 Ma, whereas the surrounding schists yielded a Sm–Nd model age of 1100–970 Ma, which has been interpreted as representing Palaeoproterozoic basalts within a matrix of sediments with a Neoproterozoic maximum depositional age (Bernard-Griffiths, Peucat & Ohta, Reference Bernard-Griffiths, Peucat and Ohta1993). However, the Palaeoproterozoic upper intercept and model ages also could represent crustal assimilation/inheritance and do not necessarily reflect the timing of basalt crystallization. A U–Pb lower intercept zircon age and Rb–Sr and 40Ar–39Ar mica ages constrain the timing of blueschist to eclogite-facies metamorphism to c. 490–450 Ma (Dallmeyer et al. Reference Dallmeyer, Peucat, Hirajima and Ohta1990; Bernard-Griffiths, Peucat & Ohta, Reference Bernard-Griffiths, Peucat and Ohta1993). Limestones and dolomites (Motalafjella Fm), conglomerates (Bulltinden Fm) and turbidites (Holmesletfjella Fm) of the Bullbreen Group unconformably overlie the metamorphic rocks of the Vestgötabreen Complex (e.g. Kanat & Morris, Reference Kanat and Morris1988; Harland, Anderson & Manasrah, Reference Harland, Anderson and Manasrah1997). The Bulltinden conglomerates contain clasts of the underlying limestones as well as of the metamorphic rocks, indicating that the latter were exhumed during deposition of the sediments. Conodonts, corals and brachiopods from the Bulltinden Formation constrain the depositional age of the Bullbreen Group to the Ordovician (Caradoc age: ~ 460–450 Ma; Armstrong, Nakrem & Ohta, Reference Armstrong, Nakrem and Ohta1986) to Silurian (Llandovery to Wenlock, ~ 440–420 Ma; Scrutton, Horsfield & Harland, Reference Scrutton, Horsfield and Harland1976). Along the eastern boundary of the pre-Devonian rocks of the west coast of Spitsbergen, slivers of Carboniferous sandstones (Billefjorden Group) occur tectonically intercalated with the Proterozoic to Silurian rocks (Figs 2, 5).
In summary, the pre-Devonian rocks of the western coast of Spitsbergen south of Kongsfjorden are characterized by the occurrence of c. 1200 Ma and c. 950 Ma orthogneisses, pre-650 Ma sedimentary rocks, amphibolites-facies metamorphism at c. 650 Ma, post-650 Ma Neoproterozoic sedimentary rocks including diamictites, Cambro-Ordovician sedimentary rocks, a c. 490–450 Ma blueschist to eclogite-facies complex and unconformably overlying Ordovician to Silurian rocks (Fig. 4). There is no evidence for classical Scandian (mainly Silurian) Caledonian metamorphism within this area. This is in strong contrast to the pre-Devonian rocks exposed on Nordaustlandet, Ny Friesland and northwestern Spitsbergen (Fig. 2).
3. Sampling strategy
In order to characterize the various sedimentary rocks under- and overlying the exotic Vestgötabreen Complex in the Motalafjella area on Oscar II Land in terms of detrital zircon geochronology, we selected six sedimentary samples covering the tectonostratigraphy from the oldest to the youngest units (latest Mesoproterozoic to Carboniferous; Figs 4, 5). In addition, we selected one deformed augengneiss and three samples of mafic metavolcanic rocks from the Trollheimen Suite for U–Pb TIMS zircon dating. Five of the metasedimentary samples, the augengneiss and the three metavolcanitic rocks were collected by Y. Ohta during field campaigns in 1972 and 1973 and were stored in the rock archive of the Norwegian Polar Institute in Tromsø. Sample 5 was collected by A. Andresen during a field trip to the Motalafjella area in 2005. The mineral separates from the three mafic metavolcanitic rocks did not contain any zircons, and dating was therefore not possible. These three samples are not considered further in the manuscript and the age of the Trollheimen volcanics is still unknown (Fig. 4).
4. Analytical methods
Zircons were separated from the samples by using standard crushing and mineral separation techniques, followed by hand-picking of the grains under a binocular microscope. We did not use magnetic separation for the sedimentary samples in order to avoid bias in the zircon selection. Zircons from the sedimentary samples were then mounted in epoxy, polished to expose the grain centres, and cathodoluminescence (CL) images were obtained on a JEOL scanning electron microscope at the Department of Geosciences, University of Oslo.
U–Pb laser-ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) analyses on detrital zircons were performed using a Nu Plasma HR multicollector ICP-MS system coupled to a New Wave/Merchantek LUV-213 laser at the Department of Geosciences, University of Oslo, using a standardization method described by Rosa et al. (Reference Rosa, Finch, Andersen and Inverno2009). Masses 204, 206, 207 and 238 were measured simultaneously and 235U was calculated from 238U using a natural 238U/235U = 137.88. Mass number 204 was used as a monitor for 204Pb. The reference zircons GJ-01 (609 ± 1 Ma; Belousova, Griffin & O'Reilly, Reference Belousova, Griffin and O'Reilly2006), 91500 (1065 ± 1 Ma; Wiedenbeck et al. Reference Wiedenbeck, Alle, Corfu, Griffin, Meier, Oberli, Quadt, Roddick and Spiegel1995) and the in-house standard A382 (1876 ± 5 Ma) were used for standardization. Age calculation was done using Isoplot version 4.1 (Ludwig, pers. comm.) and normalized probability plots were done using the Excel macro NORMALIZED_PROB_PLOT of the Arizona LaserChron Center. For age probability plots, 207Pb/206Pb ages are used for analyses older than 700 Ma, and 206Pb/238U ages for analyses younger than 700 Ma. We chose this cut-off age since there are very few analyses between c. 900 and 700 Ma in our dataset. For statistical analysis, a Kolmogorov–Smirnov (K–S) statistical test was applied using the Excel macro developed by the Arizona LaserChron Center.
The U–Pb TIMS analyses were carried out at the Department of Geosciences at the University of Oslo. The zircons selected for ID-TIMS analyses were chemically abraded in order to remove areas affected by Pb loss (Mattinson, Reference Mattinson2005). These zircons were first annealed for c. 60 hours at c. 900°C and were then etched in 48% HF (+HNO3) during c. 16 hours at c. 195°C. The grains were then washed in dilute HNO3, ionized water and acetone using an ultrasonic bath to remove any contamination. Each sample was weighed on a microbalance and spiked with a mixed 202Pb–205Pb–235U tracer. The samples were then dissolved in HF and a drop of HNO3 in Teflon bombs at c. 195°C for five days. Dissolved samples weighing more than 0.005 mg were chemically separated using micro-columns and anion-exchange resin in order to remove cations that may inhibit ionization (Krogh, Reference Krogh1973). U/Pb-solutions were dried down and loaded on degassed single Re filaments with silica gel and measured on a Finnigan MAT 262 mass spectrometer. Details on the measuring process are given in Appendix A of Augland, Andresen & Corfu (Reference Augland, Andresen and Corfu2010). The analytical errors and corrections were incorporated and propagated using the ROMAGE 6.3 program, originally developed by T. E. Krogh. Common Pb corrections were employed using the Pb-evolution model of Stacey & Kramers (Reference Stacey and Kramers1975) at the age in question. Ages were calculated using Isoplot version 4.1 (Ludwig, pers. comm.) and the decay constants referred to in Steiger & Jäger (Reference Steiger and Jäger1977).
5. Sample description and U–Pb results
The metasedimentary samples are first described from oldest to youngest, followed by a description of the augengneiss sample dated by TIMS. All U–Pb LA-ICP-MS analyses are given in Table S1 in the online Supplementary Material at http://journals.cambridge.org/geo and the TIMS data are given in Table S2 in the online Supplementary Material at http://journals.cambridge.org/geo. Results of K–S statistical tests can be found in Table S3 in the online Supplementary Material at http://journals.cambridge.org/geo.
5.a. Sample 1
This sample is a white quartzite which occurs within phyllites on the west coast south of St Jonsfjorden (Fig. 5; Y. Ohta, unpub. field notes and map, Norwegian Polar Institute, 1973). It belongs to the Løvlibreen subunit, which is interbedded with the Trollheimen Suite volcanics (Figs 4, 5). It is part of the oldest unit in the area, the possibly Mesoproterozoic St Jonsfjorden unit (Bergh et al. Reference Bergh, Ohta, Andresen, Maher, Braathen and Dallmann2003). The zircons from this sample are mostly < 125 μm sized, clear to yellowish, rounded fragments. CL images reveal unzoned, patchy or oscillatory zoned grains. We performed 126 U–Pb LA-ICP-MS analyses on 124 grains. Eleven analyses show > 5% central discordance and are excluded from age probability plots. The 115 analyses < 5% discordant belong to three main groups: (1) ~ 1600–1750 Ma, with a major peak at ~ 1630–1640 Ma; (2) ~ 1300–1500 Ma with a peak at ~ 1440 Ma; and (3) ~ 1000–1190 Ma, with a peak at ~ 1100 Ma (Figs 6a, 7; Table S1 in the online Supplementary Material at http://journals.cambridge.org/geo). The three youngest grains give a weighted mean age of 1016 ± 13 Ma, which is an upper limit for the depositional age of this sample (e.g. Dickinson & Gehrels, Reference Dickinson and Gehrels2009). In addition to the three main groups, three zircons give ages in the range 1850–2100 Ma. Only one discordant Archaean grain was found (Fig. 6a).
Figure 6. Concordia plots of the six detrital samples analysed by LA-ICP-MS.
Figure 7. Probability and frequency plot of the six detrital samples analysed by LA-ICP-MS. Only analyses with < 5% central discordance are included in the plot.
5.b. Sample 2
This sample is a laminated shale–sandstone alternation from north of St Jonsfjorden (Fig. 5; Y. Ohta, unpub. field notes and map, Norwegian Polar Institute, 1972). It belongs to the Sparrefjellet subunit of the Daudmannsodden unit (Bergh et al. Reference Bergh, Ohta, Andresen, Maher, Braathen and Dallmann2003; Fig. 5). This unit is a clast-free, non-tillitic shallow-marine succession of lower metamorphic grade than the underlying St Jonsfjorden unit (Y. Ohta, pers. comm. 2011). The zircons from this sample are < 125 μm sized, clear-reddish to dull-brownish, rounded to long-prismatic fragments. CL images reveal variable internal zoning. We performed 186 analyses on 181 grains. Forty-seven analyses show > 5% central discordance. The reason for this discordance is probably Pb loss some time after the original crystallization of the grains. However, the significance or timing of this Pb loss cannot be resolved from our dataset, and it is not correlated with U-content of the grains. The 139 analyses < 5% discordant belong to three main groups: (1) ~ 910–1200 Ma, with a major peak at ~ 1100 Ma; (2) ~ 1200–1600 Ma, with a major peak at ~ 1440 Ma; and (3) ~ 1600–1800 Ma, with a major peak at ~ 1650 Ma (Figs 6b, 7; Table S1 in the online Supplementary Material at http://journals.cambridge.org/geo). The three youngest grains give a weighted mean age of 948 ± 65 Ma, which represents the maximum depositional age for this sample. In addition, there are six discordant analyses which point to upper intercept ages of > 2600 Ma (Fig. 6b).
5.c. Sample 3
This sample is a tillitoid sandstone with dolomitic pebbles from north of St Jonsfjorden close to Ankerbreen (Fig. 5; Y. Ohta, unpub. field notes and map, Norwegian Polar Institute, 1972). It belongs to the Comfortlessbreen (or West Coast diamictite) unit of probable Late Neoproterozoic (‘Vendian’) age and tectonically overlies the Daudmannsodden unit (Fig. 5; Bergh et al. Reference Bergh, Ohta, Andresen, Maher, Braathen and Dallmann2003). The zircons from this sample are < 250 μm in size, clear to pinkish, rounded fragments. CL images reveal mostly unzoned or simply and patchy zoned grains; oscillatory zoning is rare. We performed 131 analyses on 129 grains. Twenty-four analyses show > 5% central discordance; the remaining 107 analyses are < 5% discordant. The latter show the following pattern (Figs 6c, 7): (1) the largest age group is ~ 940–1190 Ma, with an apparent hiatus around 1100 Ma; (2) the second biggest group is represented by ages of ~ 1300–1500 Ma; and (3) there are two smaller peaks at ~ 1630 Ma and ~ 1830 Ma. The two youngest concordant ages give a weighted mean age of 713 ± 7 Ma (Fig. 6c; Table S1 in the online Supplementary Material at http://journals.cambridge.org/geo). There is one concordant Archaean grain of ~ 2700 Ma and several of the discordant ages point to Archaean upper intercepts (Fig. 6c).
5.d. Sample 4
This sample is a medium-grained sandstone intercalated with conglomerates from east of Bullbreen south of St Jonsfjorden (Fig. 5). The conglomerates contain pebbles of quartzite, mica schist, gneissic diorite and limestones and belong to the Bulltinden Formation of the Bullbreen Group (Figs 4, 5; Y. Ohta, unpub. field notes and map, Norwegian Polar Institute, 1973; Bergh et al. Reference Bergh, Ohta, Andresen, Maher, Braathen and Dallmann2003). The zircons from this sample are mostly < 125 μm in size, clear to pinkish crystals. CL images reveal variable zoning patterns with some core-rim structures. We performed 121 analyses on 121 grains. Forty analyses show > 5% central discordance; the remaining 81 analyses are < 5% discordant. The latter all lie in the range of ~ 920–1700 Ma, with the exception of one grain of ~ 1960 Ma (Figs 6d, 7). Within the main age group, the main peaks occur at ~ 1400, ~ 1300, ~ 1500, 1650 and ~ 1180 (Figs 6d, 7). The three youngest grains give a weighted average age of 933 ± 47 Ma. Five discordant grains point to Archaean upper intercepts (Fig. 6d).
5.e. Sample 5
This sample is a turbiditic calcareous slate from the Motallafjella area (Fig. 5). It belongs to the Holmesletfjella Formation of the Bullbreen Group (Figs 4, 5; Bergh et al. Reference Bergh, Ohta, Andresen, Maher, Braathen and Dallmann2003). The zircons from this sample are mostly < 125 μm in size, clear to reddish, long-prismatic to rounded fragments. CL images reveal variable zoning and some grains show thin, bright-CL overgrowths. We performed 126 analyses on 125 grains. Twenty analyses show > 5% central discordance. The remaining 106 analyses < 5% discordant form two main groups: (1) the majority of analyses fall in the age range ~ 940–1120 Ma, with a major peak at ~ 950–980 Ma and a second peak at ~ 1080 Ma; (2) a second group occurs at ~ 640–760 Ma, with a peak at ~ 670 Ma (Figs 6e, 7; Table S1 in the online Supplementary Material at http://journals.cambridge.org/geo). The eight youngest grains give a weighted mean age of 675 ± 8 Ma. The remaining analyses scatter between ~ 1200–1840 Ma, and one single discordant grain points to an Archaean upper intercept age (Fig. 6e).
5.f. Sample 6
This sample is a white, coarse-grained sandstone with black fossil wood fragments from north of Charlesbreen (Fig. 5; Y. Ohta, unpub. field notes and map, Norwegian Polar Institute, 1973). It belongs to the Orustdalen Formation of the Billefjorden Group (Fig. 5; Bergh et al. Reference Bergh, Ohta, Andresen, Maher, Braathen and Dallmann2003). The zircons from this sample are rounded fragments of variable size with clear, pinkish or yellowish colours. Most grains show variable zoning on CL images. We performed 138 analyses on 132 grains. Forty-one analyses show > 5% central discordance; the remaining 97 analyses < 5% discordant fall into several groups: (1) the largest group contains ages spanning ~ 950–2100 Ma, with peaks at ~ 1000–1100 Ma and 1800–1900 Ma; (2) four grains give ages of ~ 360–400 Ma; (3) four grains give ages of ~ 480–690 Ma; and (4) five grains give ages of ~ 2450–2600 Ma (Figs 6f, 7; Table S1 in the online Supplementary Material at http://journals.cambridge.org/geo). Most of the discordant ages are pointing to upper intercepts of > 2100 Ma (Fig. 6f).
5.g. Sample 7
This sample comes from south of Bullbreen and was originally described as ‘feldspar spotted quartzite, like siliceous gneiss’ (Fig. 5; Y. Ohta, unpub. field notes and map, Norwegian Polar Institute, 1973). On the map of Bergh et al. (Reference Bergh, Ohta, Andresen, Maher, Braathen and Dallmann2003), it is indicated as quartzite bands within the Sparrefjell subunit of the Daudmannsodden unit (Fig. 5). However, the spotty nature of the feldspar and the strong foliation led us to suspect that it could represent a foliated augengneiss. Indeed, the zircon population from this sample is very homogenous and consists of long-prismatic, euhedral grains with dirty cores surrounded by clear tips (Fig. 8a), which is atypical for a quartzite but more typical for a granitic/pegmatitic gneiss. We analysed six fractions of one to four tips and two fractions of one to six clear grains by TIMS. The results are given in Table S2 in the online Supplementary Material at http://journals.cambridge.org/geo and plotted in Figure 8d. They indicate that this sample most likely represents an orthogneiss. One fraction is clearly discordant, which could be the result of recent Pb loss. Two fractions are slightly older than the main cluster of five fractions, which might be the result of partly analysing inherited cores. Five of the eight fractions define a tight cluster on the Concordia diagram and a Concordia age of 927 ± 3 Ma can be calculated, which we interpret as representing the time of melt crystallization.
Figure 8. (a) Photograph of zircon population from orthogneiss sample 7. Note the clear core-tip relationships and the uniform zircon morphology, pointing to a magmatic origin. (b) Tips broken off for TIMS analysis from same sample. (c) Tips after chemical abrasion. (d) Concordia plot of the eight single and multigrain fractions analysed by TIMS (entire dataset in online Supplementary Material Table S3 at http://journals.cambridge.org/geo). Stippled analyses were excluded from final Concordia age calculation. The Concordia age calculated is given at the 95% confidence interval with decay constant errors included.
6. Discussion
Based on the U–Pb zircon data presented above, we discuss the provenance and tectonic evolution of the pre-Devonian rocks of western Spitsbergen in the Motalafjella area from the oldest to the youngest events.
6.a. Latest Mesoproterozoic sedimentation
Sample 1 has a latest Mesoproterozoic maximum depositional age of ≤ 1016 ± 13 Ma and a detrital age spectrum dominated by Meso- to Late Palaeoproterozoic ages, with a major peak between 1600 and 1700 Ma (Figs 6a, 7). Statistically, it is likely that sample 1 comes from a different source than samples 2 and 3 (p < 0.05; Table S3a in the online Supplementary Material at http://journals.cambridge.org/geo). Additionally, sample 1 is lithologically different and of higher metamorphic grade (quartzites, phyllites intercalated with volcanic rocks, the Løvlibreen subunit; Fig. 5) than samples 2 and 3 (tillitoid sandstones and shales, Comfortlessbreen and Daudmannsodden units; Fig. 5, Harland, Hambrey & Waddans, Reference Harland, Hambrey and Waddans1993). We therefore discuss the provenance and tectonic setting of sample 1 separately from samples 2 and 3.
The age spectrum of sample 1 has a statistically low probability of being different from age spectra observed in other latest Mesoproterozoic to Neoproterozoic sequences with similar maximum depositional ages within the North Atlantic region (p = 0.057–0.825; Table S3b in the online Supplementary Material at http://journals.cambridge.org/geo; Figs 1, 9): the Krossfjorden Complex of northwestern Spitsbergen (deposition between 1020–995 Ma; Petterson, Pease & Frei, Reference Petterson, Pease and Frei2009), the Sværholt succession of northern Norway (deposition between 1030–980 Ma; Kirkland, Daly & Whitehouse, Reference Kirkland, Daly and Whitehouse2007) and parts of the Heggmovatn supracrustals of central–northern Norway (deposition between 1050–930 Ma; Agyei-Dwarko, Augland & Andresen, Reference Agyei-Dwarko, Augland and Andresen2012). Other units with similar maximum depositional ages but a statistically high probability of having a different age spectrum include the Brennevinsfjorden Group of Nordaustlandet (deposition between 1000–950 Ma; Johannson et al. Reference Johansson, Gee, Larionov, Ohta and Tebenkov2005), the Krummedal/Smallefjord sequence of eastern Greenland (deposition between 1000–950 Ma; Strachan, Nutman & Friderichsen, Reference Strachan, Nutman and Friderichsen1995; Kalsbeek et al. Reference Kalsbeek, Thrane, Nutman and Jespen2000; Watt, Kinny & Friderichsen, Reference Watt, Kinny and Friderichsen2000; Watt & Thrane, Reference Watt and Thrane2001; Leslie & Nutman, Reference Leslie and Nutman2003), the Torridon Group of Scotland (deposition between 1060–970 Ma; Rainbird, Hamilton & Young, Reference Rainbird, Hamilton and Young2001), the Morar Group of Scotland (deposition between 1020–840 Ma; Friend et al. Reference Friend, Strachan, Kinny and Watt2003; Kirkland, Strachan & Prave, Reference Kirkland, Strachan and Prave2008) and the Westing Group of Shetland (deposition between 1030–930 Ma; Cutts et al. Reference Cutts, Hand, Kelsey, Wade, Strachan, Clark and Netting2009; Fig. 1). All these units are part of the lower lithotectonic assemblage 3 of Cawood et al. (Reference Cawood, Nemchin and Strachan2007) and represent the lithotectonic assemblage 1 of Kirkland, Strachan & Prave (Reference Kirkland, Strachan and Prave2008) and Bingen, Belousova & Griffin (Reference Bingen, Belousova and Griffin2011).
Figure 9. Normalized probability plot of Mesoproterozoic basin successions from the North Atlantic region. The locations of the sequences are indicated on Figure 1. Only analyses with < 5% discordance are included. The spectrum of sample 1 shows p-values of 0.05–0.8 when compared with samples from the Krossfjorden Complex, the Sværholt succession and parts of the Heggmovatn unit (online Supplementary Material Table S3 at http://journals.cambridge.org/geo), which indicates similar provenance patterns. Note that all successions are dominated by Neo- to Mesoproterozoic detritus, with only minor Archaean input in the Westing and Torridon groups. Data from: 1 – Cutts et al. (Reference Cutts, Hand, Kelsey, Wade, Strachan, Clark and Netting2009); 2 – Strachan, Nutman & Friderichsen (Reference Strachan, Nutman and Friderichsen1995), Kalsbeek et al. (2001), Watt, Kinny & Friderichsen (2000), Leslie & Nutman (2001) and Kalsbeek et al. (Reference Kalsbeek, Thrane, Nutman and Jespen2000); 3 – Rainbird, Hamilton & Young (2001); 4 – Friend et al. (Reference Friend, Strachan, Kinny and Watt2003) and Kirkland, Strachan & Prave (2008); 5 – Agyei-Dwarko, Augland & Andresen (2012); 6 – Kirkland, Daly & Whitehouse (2007); 7 – Petterson, Pease & Frei (2009).
The similarity in detrital zircon population between sample 1 and the Krossfjorden Group of northwestern Spitsbergen indicates that these deposits might have shared a common history at c. 1000 Ma. Further south along the west coast of Spitsbergen, the Nordbukta and Deilegga units below the Torrelian unconformity (Figs 3, 4) could belong to the same basinal system, but no detrital zircon data are available from these units for comparison. The interpretation of sample 1 as a latest Mesoproterozoic deposit is notably different from the interpretation of Harland, Hambrey & Waddans (1993) and Harland, Anderson & Manasrah (1997), who attributed all Proterozoic strata of Oscar II Land to the Vendian period (Late Neoproterozoic). It is also noteworthy that no similar latest Mesoproterozoic sediments are known from the autochthonous rocks of northern and NE Greenland or the nappes on Pearya (e.g. Trettin, Reference Trettin1987, Reference Trettin and Trettin1991; Collinson et al. Reference Collinson, Kalsbeek, Jespen, Pedersen, Upton, Higgins, Gilotti and Smith2008). In addition, a 1800–2100 Ma detrital zircon peak dominates the older and younger strata in the autochthonous rocks of NE Greenland (Cawood et al. Reference Cawood, Nemchin and Strachan2007; Kirkland et al. Reference Kirkland, Pease, Whitehouse and Ineson2009; Anfinson et al. Reference Anfinson, Leier, Embry and Dewing2012), a peak which is completely absent in sample 1. This indicates that the oldest deposits along the western coast of Spitsbergen cannot be linked to autochthonous rocks of northeastern or northern Greenland during latest Mesoproterozoic time.
The aforementioned latest Mesoproterozoic sediments from the North Atlantic region have been interpreted to represent remnants of successor basins which were deposited on top of the Grenvillian–Sveconorwegian orogen either within or at the border of the supercontinent Rodinia (e.g. Cawood et al. Reference Cawood, Nemchin, Strachan, Kinny and Loewy2004, Reference Cawood, Nemchin and Strachan2007, Reference Cawood, Strachan, Cutts, Kinny, Hand and Pisarevsky2010; Cawood, Nemchin & Strachan, Reference Cawood, Nemchin, Strachan, Prave and Krabbendam2007; Kirkland, Daly & Whitehouse, Reference Kirkland, Daly and Whitehouse2007; Kirkland, Strachan & Prave, Reference Kirkland, Strachan and Prave2008; Petterson et al. Reference Petterson, Pease and Frei2009). These basins were interpreted to have received their detritus mainly from the Palaeo- to Mesoproterozoic orogenic belts of Laurentia and/or Baltica while they were shielded from the Archaean cratons (e.g. Kirkland, Daly & Whitehouse, Reference Kirkland, Daly and Whitehouse2007; Kirkland, Strachan & Prave, Reference Kirkland, Strachan and Prave2008; Petterson et al. Reference Petterson, Pease and Frei2009). However, more detailed interpretations of the detrital age spectra and therewith the origins of these latest Mesoproterozoic basins is hampered by several factors: (1) apart from the Torridon Group of the Laurentian foreland, all these deposits occur within allochthonous units detached from their original basement, and large-scale tectonic transport probably isolated them from their previous source areas (e.g. Kirkland, Daly & Whitehouse, Reference Kirkland, Daly and Whitehouse2007; Petterson et al. Reference Petterson, Pease and Frei2009); (2) long-range sedimentary transport and recycling of depositional units could have mixed the detrital age spectra, so that it is difficult to exactly reconstruct the position of the basins at the time of deposition (e.g. Rainbird et al. Reference Rainbird, McNicoll, Theriault, Heaman, Abbott, Long and Thorkelson1997; Bingen, Belousova & Griffin, Reference Bingen, Belousova and Griffin2011); (3) potential source areas for the wide Palaeo- to Mesoproterozoic age spectra are present within Baltica and Laurentia (and probably even on other continents), and the location and geometry of major source areas, such as the Sveconorwegian–Grenvillian orogenic belt, are debated (e.g. Lorenz et al. Reference Lorenz, Gee, Larionov and Majka2012); (4) palaeogeographic reconstructions of Rodinia for the time of interest show large variations, most of them placing Baltica adjacent to Laurentia, but some with Baltica in an upright position linked by the Grenvillian–Sveconorwegian orogen to Amazonia in the south (Fig. 11a; e.g. Cawood & Pisarevsky, Reference Cawood and Pisarevsky2006; Li et al. Reference Li, Bogdanova, Collins, Davidson, De Waele, Ernst, Fitzismons, Fuck, Gladkochub, Jacobs, Karlstrom, Lu, Natapov, Pease, Pisarevsky, Thrane and Vernikovsky2008) and others with Baltica in an upside-down position, with the Norwegian margin of Baltica facing the opening Ægir Sea (Fig. 11a; e.g. Hartz & Torsvik, Reference Hartz and Torsvik2002; Cocks & Torsvik, Reference Cocks and Torsvik2005, Reference Cocks and Torsvik2011). In our opinion, the available detrital zircon patterns cannot currently resolve these uncertainties in Mesoproterozoic palaeogeography.
6.b. Late Grenvillian magmatism
The Late Grenvillian intrusion age of sample 7 (927 ± 3 Ma; Fig. 8) is the first documentation of Early Neoproterozoic magmatism along the west coast of Spitsbergen south of Kongsfjorden, except for an unpublished age of c. 920 Ma mentioned by Czerny et al. (Reference Czerny, Majka, Gee, Manecki and Manecki2010) from the Magnethøgda sequence (Figs 2, 4). The collector of the sample did not recognize it as an orthogneiss, but the zircon population of the sample clearly identifies it as such (Fig. 8). Because of these difficulties in recognition, the field relationships of this sample are not clear: it could represent (a) a minor intrusion within the surrounding sediments or (b) a tectonic sliver of crystalline basement intercalated with the surrounding sediments. If it represents an intrusion into the surrounding sediments, then the sediments have to be older than c. 930 Ma. According to the map of Bergh et al. (Reference Bergh, Ohta, Andresen, Maher, Braathen and Dallmann2003), the sediments in which sample 7 occurs belong to the Sparrefjell subunit of the Daudmannsodden unit (Fig. 5). Our sample 2 is a correlative of this unit, and has a maximum depositional age of ≤ 948 ± 65 Ma (the three youngest detrital grains), which is within error of the intrusion age of sample 7. So in principle, the Daudmannsodden unit could be older than 930 Ma, and it could be intruded by sample 7. However, based on detrital provenance similarities between samples 2 and 3 and on regional and lithological correlations discussed further below, we suggest that the Daudmannsodden unit is probably younger than the intrusion and that the orthogneiss of sample 7 represents a tectonic sliver of older crystalline basement. Given the strong overprint of the region by the Eocene West Spitsbergen fold-and-thrust belt, tectonic intercalation is likely (e.g. Figs 2, 5). Therefore, future workers in the area should pay attention to the probable occurrence of orthogneisses and their tectonic relationships in that particular area.
Czerny et al. (Reference Czerny, Majka, Gee, Manecki and Manecki2010) reported that the Magnethøgda sequence on Wedel Jarlsberg Land (Figs 2–4) mainly consists of c. 920 Ma porphyritic rocks, and magmatism of similar age is common in many latest Mesoproterozoic to Neoproterozoic basins throughout the North Atlantic region (Fig. 1): in the Krossfjorden Group (c. 990–960 Ma; Petterson, Pease & Frei, Reference Petterson, Tebenkov, Larionov, Andresen and Pease2009; Myhre, Corfu & Andresen, Reference Myhre, Corfu and Andresen2009), in the Brennevinsfjorden Group (960–940 Ma; Johannson et al. 2005; Gee & Teben'kov, Reference Gee, Teben'kov, Gee and Pease2004), in the Sværholt succession (c. 980–970 Ma; Kirkland, Daly & Whitehouse, Reference Kirkland, Daly and Whitehouse2006, Reference Kirkland, Daly and Whitehouse2007), in the Krummedal/Smallefjord sequence (c. 950–910 Ma; Strachan, Nutman & Friderichsen, Reference Strachan, Nutman and Friderichsen1995; Kalsbeek et al. Reference Kalsbeek, Thrane, Nutman and Jespen2000; Watt & Thrane, Reference Watt and Thrane2001; Leslie & Nutman, Reference Leslie and Nutman2003), in the Heggmovatn supracrustals (c. 925 Ma; Agyei-Dwarko, Augland & Andresen, Reference Agyei-Dwarko, Augland and Andresen2012) and in the Westing Group (c. 930 Ma; Cutts et al. Reference Cutts, Hand, Kelsey, Wade, Strachan, Clark and Netting2009).
The geodynamic context of this magmatism is not well understood. It could represent (a) collisional events associated with the latest stages of the Grenvillian–Sveconorwegian orogeny (e.g. Kalsbeek et al. Reference Kalsbeek, Thrane, Nutman and Jespen2000; Watt & Thrane, Reference Watt and Thrane2001; Leslie & Nutman, Reference Leslie and Nutman2003); (b) intracratonic or extension-related magmatism within Rodinia (e.g. Cawood et al. Reference Cawood, Nemchin, Strachan, Kinny and Loewy2004, Reference Cawood, Nemchin and Strachan2007); or (c) accretionary events towards the south (e.g. Kirkland, Daly & Whitehouse, Reference Kirkland, Daly and Whitehouse2007; Kirkland, Strachan & Prave, Reference Kirkland, Strachan and Prave2008) or north (e.g. Cawood et al. Reference Cawood, Strachan, Cutts, Kinny, Hand and Pisarevsky2010) of the Laurentia–Baltica suture (Fig. 11a). Cawood et al. (Reference Cawood, Strachan, Cutts, Kinny, Hand and Pisarevsky2010) proposed to call this accretionary event the Renlandian orogeny. However, geochemical and other evidence for the geodynamic context of this magmatism is still lacking. In addition, the difficulties described above in reconstructing the original position of the sedimentary basins of course also complicate the interpretation of this magmatic event occurring in these basins (Fig. 11a).
6.c. Neoproterozoic sedimentation
Samples 2 and 3 give maximum depositional ages of ≤ 948 ± 65 Ma and ≤ 713 ± 7 Ma, respectively, and they have a low probability of being derived from two statistically distinguishable source areas (p = 0.78; Table S3a in the online Supplementary Material at http://journals.cambridge.org/geo). The difference in maximum depositional age of c. 250 Ma could be the result of either (a) a real difference in depositional age, i.e. sample 2 is older than sample 3 and the similar age spectra could be the result of recycling of sample 2 into sample 3, or (b) owing to scarcity of zircon-producing magmatic events during the Neoproterozoic in the North Atlantic realm (e.g. Bingen & Solli, Reference Bingen and Solli2009; Be'eri-Shlevin et al. Reference Be'eri-Shlevin, Gee, Claesson, Ladenberger, Majka, Kirkland, Robinson and Frei2011; Bingen, Belousova & Griffin, Reference Bingen, Belousova and Griffin2011), sample 2 could be as young as sample 3 but just did not receive any Neoproterozoic zircons. Owing to lithological and regional considerations we favour option (b) as discussed below.
Sample 2 comes from a clast-free shallow-marine unit, whereas sample 3 comes from a diamictitic unit. Both units are of lower metamorphic grade and are less deformed than the underlying St Jonsfjorden unit of sample 1 (Harland, Anderson & Manasrah, Reference Harland, Anderson and Manasrah1997; Y. Ohta, pers. comm. 2011). A similar difference in metamorphic grade is exposed further south on Wedel Jarlsberg Land between the Nordbukta/Deilegga groups below and the Sofiebogen and Kapp Lyell groups above the Torrelian unconformity (Fig. 3; e.g. Bjørnerud, Craddock & Wills, Reference Bjørnerud, Craddock and Wills1990; Bjørnerud, Decker & Craddock, Reference Bjørnerud, Decker and Craddock1991). Even though the Torrelian unconformity is not exposed north of Bellsund (Figs 3, 4), we correlate samples 2 and 3 with rock units above the Torrelian unconformity, restricting their maximum depositional age to ≤ 650 Ma (e.g. Mazur et al. Reference Mazur, Czerny, Majka, Manecki, Holm, Smyrak and Wypych2009; Czerny et al. Reference Czerny, Majka, Gee, Manecki and Manecki2010). The tillites of the Kapp Lyell Group have been correlated, on lithological grounds, with tillites in Nordaustlandet and Greenland of probable Marinoan (c. 650–635 Ma) age (Sønderholm et al. Reference Sønderholm, Frederiksen, Smith, Tirsgaard, Higgins, Gilotti and Smith2008; Bjørnerud, Reference Bjørnerud2010). However, the stratigraphy below and above the tillites in Nordaustlandet and Greenland is different from the stratigraphy along the western coast of Spitsbergen (Figs 2–4; e.g. Gee & Teben'kov, Reference Gee, Teben'kov, Gee and Pease2004), and therefore this correlation is tenuous.
Late Neoproterozoic sediments, including diamictites, are widespread in the North Atlantic region (Fig. 1): they include the Sofiebogen and Kapp Lyell groups of southwestern Spitsbergen (e.g. Bjørnerud, Reference Bjørnerud2010), the Murchisonfjorden and Hinlopenstretet Supergroup of Nordaustlandet (e.g. Johannson et al. Reference Johansson, Gee, Larionov, Ohta and Tebenkov2005), the Eleonore Bay Supergroup and Tillite Group of East Greenland (Dhuime et al. Reference Dhuime, Bosch, Bruguier, Caby and Pourtales2007; Sønderholm et al. Reference Sønderholm, Frederiksen, Smith, Tirsgaard, Higgins, Gilotti and Smith2008), the Morænesø tillites of NE Greenland (e.g. Kirkland et al. Reference Kirkland, Pease, Whitehouse and Ineson2009), the Kennedy Channel Formation of Ellesmere Island (Anfinson et al. Reference Anfinson, Leier, Embry and Dewing2012), the lower part of the Dalradian Supergroup of Scotland (Cawood et al. Reference Cawood, Nemchin, Smith and Loewy2003, Reference Cawood, Strachan, Cutts, Kinny, Hand and Pisarevsky2010), and various Neoproterozoic sediments including glacigenic deposits which are incorporated into the Scandinavian Caledonides (Siedlecka et al. Reference Siedlecka, Roberts, Nystuen, Olovyanishnikov, Gee and Pease2004; Bingen et al. Reference Bingen, Griffin, Torsvik and Saeed2005; Nystuen et al. Reference Nystuen, Andresen, Kumpulainen and Siedlecka2008; Bingen, Belousova & Griffin, Reference Bingen, Belousova and Griffin2011; Be'eri-Shlevin et al. Reference Be'eri-Shlevin, Gee, Claesson, Ladenberger, Majka, Kirkland, Robinson and Frei2011; Kirkland et al. Reference Kirkland, Bingen, Whitehouse, Beyer and Griffin2011). These units belong to the lithotectonic assemblage 3 of Kirkland, Strachan & Prave (Reference Kirkland, Strachan and Prave2008) and Bingen, Belousova & Griffin (Reference Bingen, Belousova and Griffin2011) and were deposited mainly after 800–700 Ma and up into the Late Neoproterozoic/Early Cambrian (e.g. Cawood et al. Reference Cawood, Strachan, Cutts, Kinny, Hand and Pisarevsky2010; Bingen, Belousova & Griffin, Reference Bingen, Belousova and Griffin2011). Basaltic dykes associated with several of these Neoproterozoic sedimentary successions indicate that they were deposited coevally with the first stages of Iapetus rifting (e.g. Nystuen et al. Reference Nystuen, Andresen, Kumpulainen and Siedlecka2008; Bingen, Belousova & Griffin, Reference Bingen, Belousova and Griffin2011).
Detrital zircon age data are available for several of these Late Neoproterozoic deposits in the North Atlantic region (Figs 1, 10). Three broad age ranges dominate the different successions in variable proportions (Fig. 10): (1) Archaean ages (c. 2500–3000 Ma), (2) Late Palaeoproterozoic ages (c. 2100–1600 Ma), and (3) Late Palaeoproterozoic to Neoproterozoic ages (c. 1750–900 Ma). Significant Archaean populations are present in parts of the Dalradian Supergroup (Cawood et al. Reference Cawood, Nemchin, Smith and Loewy2003), as well as in successions from eastern and northern Greenland and Ellesmere Island (Dhuime et al. Reference Dhuime, Bosch, Bruguier, Caby and Pourtales2007; Kirkland et al. Reference Kirkland, Pease, Whitehouse and Ineson2009; Anfinson et al. Reference Anfinson, Leier, Embry and Dewing2012; lower part of Fig. 10). These sequences also contain significant Late Palaeoproterozoic detritus, and the youngest detrital grains range from c. 900–1700 Ma, which is considerably older than their inferred Late Neoproterozoic depositional age. In contrast, Mesoproterozoic detritus dominates the sequences from the Baltoscandian margin, the lower to middle allochthons of the Scandinavian Caledonides, parts of the Eleonore Bay Supergroup (EBSG) of Greenland and samples 2 and 3 from this study, with detritus older than c. 1900 Ma being nearly to completely absent (Bingen et al. Reference Bingen, Griffin, Torsvik and Saeed2005; Bingen, Belousova & Griffin, 2011; Be'eri-Shlevin et al. Reference Be'eri-Shlevin, Gee, Claesson, Ladenberger, Majka, Kirkland, Robinson and Frei2011; Kirkland et al. Reference Kirkland, Bingen, Whitehouse, Beyer and Griffin2011; Slama et al. Reference Slama, Walderhaug, Fonneland, Kosler and Pedersen2011). The youngest detrital grains in these sequences range from c. 900–620 Ma, more closely indicating their Neoproterozoic depositional age.
Figure 10. Normalized probability plot of Neoproterozoic basin successions from the North Atlantic region. The locations of the sequences are indicated on Figure 1. Only analyses with < 5% discordance are included. Data from: 1 – Bingen et al. (Reference Bingen, Griffin, Torsvik and Saeed2005); 2 – Bingen, Belousova & Griffin (2011); 3 – Be'eri-Shlevin et al. (Reference Be'eri-Shlevin, Gee, Claesson, Ladenberger, Majka, Kirkland, Robinson and Frei2011); 4 – Kirkland et al. (Reference Kirkland, Bingen, Whitehouse, Beyer and Griffin2011); 5 – Slama et al. (Reference Slama, Walderhaug, Fonneland, Kosler and Pedersen2011); 6 – Cawood et al. (Reference Cawood, Nemchin, Smith and Loewy2003); 7 – Dhuime et al. (Reference Dhuime, Bosch, Bruguier, Caby and Pourtales2007); 8 – Kirkland et al. (Reference Kirkland, Pease, Whitehouse and Ineson2009); 9 – Anfinson et al. (Reference Anfinson, Leier, Embry and Dewing2012). The sequences are statistically compared in the online Supplementary Material Table S3 at http://journals.cambridge.org/geo.
Figure 11. Tectonic model in four time steps. (a) Reconstruction of Laurentia, Amazonia and Baltica for the Early Neoproterozoic after Cawood et al. (Reference Cawood, Strachan, Cutts, Kinny, Hand and Pisarevsky2010) and Bingen, Belousova & Griffin (2011), with orientation of Baltica according to Li et al. (Reference Li, Bogdanova, Collins, Davidson, De Waele, Ernst, Fitzismons, Fuck, Gladkochub, Jacobs, Karlstrom, Lu, Natapov, Pease, Pisarevsky, Thrane and Vernikovsky2008). Note a potentially different orientation of Baltica in Hartz & Torsvik (Reference Hartz and Torsvik2002). See text for details. (b) Reconstruction of Laurentia, Amazonia and Baltica for the Late Neoproterozoic. Two potential latest Neoproterozoic basinal systems are indicated. Note that samples 2 and 3 from this study have provenance signatures similar to many Neoproterozoic sediments from the Baltoscandian margin, and are therefore interpreted to have originated close to them. Note that parts of the older Mesoproterozoic basin probably stayed on the Baltican and others on the Laurentian side of Iapetus (e.g. Kirkland et al. Reference Kirkland, Bingen, Whitehouse, Beyer and Griffin2011). See text for details. (c) Possible tectonic situation during the Middle Ordovician. A major arc system was active along the Laurentian margin to the south (Taconian/Grampian arc). Our detrital zircon data do not allow the precise placement of the Vestgötabreen Complex and the overlying Bullbreen Group within this scenario, but an origin offshore northern Greenland is unlikely. A first phase of sinistral shearing could have brought different parts from the Timanian margin / Taconian arc to the north of Greenland. (d) Situation at the end of the Scandian continent–continent collision and after a second phase of strike-slip faulting. See text for details.
Interpreting these Neoproterozoic detrital zircon signatures is difficult. They contain a variety of age signatures with several potential source areas; they could be the result of recycling of the older latest Mesoproterozoic basins (e.g. Bingen, Belousova & Griffin, Reference Bingen, Belousova and Griffin2011) and they could be the result of large-distance pan-continental sedimentary transport in huge depositional systems (e.g. Rainbird et al. Reference Rainbird, McNicoll, Theriault, Heaman, Abbott, Long and Thorkelson1997). In addition, as in the Mesoproterozoic case, there are many uncertainties associated with the underlying plate tectonic reconstructions (e.g. Hartz & Torsvik, Reference Hartz and Torsvik2002; Cocks & Torsvik, Reference Cocks and Torsvik2005, Reference Cocks and Torsvik2011; Cawood & Pisarevsky, Reference Cawood and Pisarevsky2006; Li et al. Reference Li, Bogdanova, Collins, Davidson, De Waele, Ernst, Fitzismons, Fuck, Gladkochub, Jacobs, Karlstrom, Lu, Natapov, Pease, Pisarevsky, Thrane and Vernikovsky2008). Nevertheless, we believe that the following interpretation can be made. At least two different basinal systems existed during the Neoproterozoic: one with prominent Archaean and Late Palaeoproterozoic (c. 1700–2000 Ma) detritus, and one with mainly Mesoproterozoic detritus (Fig. 10). The first basin system includes successions from Scotland, Greenland and Ellesmere Island, and is consistent with a source mainly from the Laurentian Greenland–Labrador shield (and/or the Lewisian complex of Scotland), with only minor input from the Grenvillian–Sveconorwegian orogen (Figs 1, 10, 11b). It probably stretched along the eastern margin of Laurentia during opening of the Iapetus ocean (Fig. 11b). The second basin system includes Neoproterozoic units from the Baltoscandian margin as well as samples 2 and 3 from this study. This basin system was somehow shielded from Archaean cratonic input, received its detritus mainly from the Grenvillian–Sveconorwegian orogen, and probably lay on the Baltican side of Iapetus (Fig. 11b; e.g. Bingen, Belousova & Griffin, Reference Bingen, Belousova and Griffin2011).
Interestingly, samples 2 and 3 from this study clearly resemble the source patterns from known Baltican samples, indicating that they might have originated within the same (large-scale) Neoproterozoic basin, being located on the Baltican side of Iapetus during its opening (Fig. 11b). They clearly do not resemble the patterns of the NE Greenland and Ellesmere Island samples (Morænesø and Kennedy samples, Fig. 10), which today are the closest neighbours (Figs 1, 10). A location on the Baltican side of Iapetus during the Late Neoproterozoic would also fit with the observation of c. 650 Ma amphibolite-facies metamorphism in the Isbjørnhamna block and the corresponding Torrelian unconformity further south on western Spitsbergen, which has been linked with the Timanian orogen (e.g. Mazur et al. Reference Mazur, Czerny, Majka, Manecki, Holm, Smyrak and Wypych2009; Majka et al. Reference Majka, Czerny, Mazur, Holm and Manecki2010). Even though the main phase of the Timanian orogeny is probably somewhat younger (c. 610–500 Ma, e.g. Larionov, Andreichev & Gee, Reference Larionov, Andreichev, Gee, Gee and Pease2004; Pease & Scott, Reference Pease and Scott2009), accreted terranes contain older Neoproterozoic magmatic and metamorphic rocks (e.g. Scarrow et al. Reference Scarrow, Pease, Fleutelot and Dushin2001; Lorentz et al. Reference Lorenz, Pystin, Olovyanishnikov, Gee, Gee and Pease2004; Corfu et al. Reference Corfu, Svensen, Neumann, Nakrem and Planke2010; Kuznetsov et al. Reference Kuznetsov, Natapov, Belousova, O'Reilly and Griffin2010), and more work is needed to reveal the tectonostratigraphy of this orogen and potential links to the evolution of western Svalbard. Alternatively, Petterson, Pease & Frei (Reference Petterson, Pease and Frei2010) proposed an origin of this Late Neoproterozoic metamorphism much further south, in the vicinity of Amazonia/Avalonia (Fig. 11b, c). Other allochthonous units with similar Neoproterozoic metamorphism and magmatism are the Seve (e.g. Rehnström, Corfu & Torsvik, Reference Rehnström, Corfu and Torsvik2002) and Kalak Nappe complexes (Kirkland, Daly & Whitehouse, Reference Kirkland, Daly and Whitehouse2006). Interpreting the origin of these Neoproterozoic magmatic and metamorphic rocks also depends on the choice of plate tectonic reconstructions, which show Baltica in various positions relative to Laurentia and Gondwana in Neoproterozoic and Cambrian times (e.g. Hartz & Torsvik, Reference Hartz and Torsvik2002; Li et al. Reference Li, Bogdanova, Collins, Davidson, De Waele, Ernst, Fitzismons, Fuck, Gladkochub, Jacobs, Karlstrom, Lu, Natapov, Pease, Pisarevsky, Thrane and Vernikovsky2008). In any case, evidence for Late Neoproterozoic tectonic or magmatic activity is completely absent along the Laurentian margin of eastern and northern Greenland, again supporting the interpretation that western Spitsbergen was not located close to this margin during Neoproterozoic times, in contrast to what is often shown in palaeogeographic reconstructions (e.g. Cocks & Torsvik, Reference Cocks and Torsvik2011).
6.d. Ordovician high-pressure metamorphism and Ordovician to Silurian sedimentation
The next tectonic event recorded in the rocks of Oscar II Land of western Spitsbergen is the blueschist to eclogite-facies Ordovician metamorphism in the Vestgötabreen Complex (Figs 3–5; Ohta, Hiroi & Hirajima, Reference Ohta, Hiroi and Hirajima1983; Hirajima et al. Reference Hirajima, Banno, Hiroi and Ohta1988; Dallmeyer et al. Reference Dallmeyer, Peucat, Hirajima and Ohta1990; Bernard-Griffiths, Peucat & Ohta, Reference Bernard-Griffiths, Peucat and Ohta1993; Ohta et al. Reference Ohta, Krasilscikov, Lepvrier and Tebenkov1995; Agard et al. Reference Agard, Labrousse, Elvevold and Lepvrier2005; Labrousse et al. Reference Labrousse, Elvevold, Lepvrier and Agard2008). Unfortunately, no robust age dating for the eclogite-facies metamorphism exists, and available ages scatter over 30 Ma from c. 480 to 450 Ma (see compilation in Labrousse et al. Reference Labrousse, Elvevold, Lepvrier and Agard2008). The complex consists of a tectonic mixture of probably Neoproterozoic volcaniclastic sediments as well as metabasaltic and metagabbroic rocks of unknown origin but with oceanic geochemical affinities (Bernard-Griffiths, Peucat & Ohta, Reference Bernard-Griffiths, Peucat and Ohta1993). The metamorphic complex, together with the overlying Bullbreen Group, is in tectonic contact with the underlying Proterozoic sedimentary rocks (Fig. 5).
The detrital zircon spectra from the unconformably overlying Bullbreen Group give additional evidence for the origin and composition of the Vestgötabreen rocks. Sample 4 comes from the conglomeratic Bulltinden Formation, which contains up to 75% clasts of limestone, sandstone, dolostone and conglomerate, and up to 25% clasts of metamorphic schists and dolerites from the underlying Vestgötabreen Complex (e.g. Kanat & Morris, Reference Kanat and Morris1988; Harland, Anderson & Manasrah, Reference Harland, Anderson and Manasrah1997). It is therefore at least partly sourced from the Vestgötabreen Complex. The age spectrum of sample 4 is dominated by Late Palaeoproterozoic to Early Neoproterozoic detritus, with no concordant ages older than 2 Ga (Figs 6, 7). Since the Vestgötabreen Complex at least partly sourced sample 4, it is likely that also this complex contains mainly Mesoproterozoic detrital zircons, and could therefore mainly consist of Neoproterozoic sediments. In addition, the age spectrum of sample 4 closely resembles the age spectra of samples 2 and 3, indicating that the Vestgötabreen Complex could represent a higher metamorphic equivalent of the underlying Neoproterozoic sediments. Based on these considerations, we interpret the Vestgötabreen Complex as a tectonic melange mainly consisting of Neoproterozoic sediments, possibly of Baltican affinity similar to samples 2 and 3, and basalts of unknown origin, which were subducted and quickly exhumed, eroded and thrust over the underlying Neoproterozoic sediments during the Early Ordovician.
Whereas the conglomeratic sample 4 likely represents locally sourced sediment from the underlying metamorphic and sedimentary rocks, sample 5 is more fine-grained and of a turbiditic nature, indicating a more distal position relative to its source. It has a very peculiar provenance pattern, with a distinct major peak at c. 980–960 Ma, and a significant number of 740–640 Ma grains (Figs 6, 7). There are many potential source regions for the c. 980–960 Ma grains, spanning from southwestern Baltica to the Renlandian intrusions within the latest Mesoproterozoic basins distributed throughout the Scandinavian, East Greenland and Svalbard Caledonides (Fig. 1; Cawood et al. Reference Cawood, Strachan, Cutts, Kinny, Hand and Pisarevsky2010). The c. 740–640 Ma grains could be derived from further south along the west coast of Spitsbergen below the Torrelian unconformity (Czerny et al. Reference Czerny, Majka, Gee, Manecki and Manecki2010), from terranes accreted to Baltica during the Timanian orogeny (e.g. Lorenz et al. Reference Lorenz, Pystin, Olovyanishnikov, Gee, Gee and Pease2004; Larionov, Andreichev & Gee, Reference Larionov, Andreichev, Gee, Gee and Pease2004; Corfu et al. Reference Corfu, Svensen, Neumann, Nakrem and Planke2010; Kuznetsov et al. Reference Kuznetsov, Natapov, Belousova, O'Reilly and Griffin2010) or from Avalonia-related arcs further south (e.g. Keppie et al. Reference Keppie, Nance, Murphy and Dostal2003; Petterson, Pease & Frei, Reference Petterson, Pease and Frei2010). Another option could be Arctic Alaska, which contains both c. 970 Ma and 750–540 Ma intrusions (e.g. Colpron & Nelson, Reference Colpron, Nelson, Cawood and Kröner2009) and which probably lay somewhere to the north of Laurentia and Baltica in the Ordovician (e.g. Cocks & Torsvik, Reference Cocks and Torsvik2011). Northeastern or northern Greenland are definitively not likely sources for these zircons, since c. 980–960 Ma and 740–640 Ma sources do not exist in northern Greenland (e.g. Collinson et al. Reference Collinson, Kalsbeek, Jespen, Pedersen, Upton, Higgins, Gilotti and Smith2008; Smith & Rasmussen, Reference Smith, Rasmussen, Higgins, Gilotti and Smith2008). We therefore think that the model of Labrousse et al. (Reference Labrousse, Elvevold, Lepvrier and Agard2008), who placed the origin of the Vestgötabreen Complex just offshore northern Greenland, is unlikely.
A different option is the model of Petterson, Pease & Frei (Reference Petterson, Pease and Frei2010): based on the occurrence of Ordovician detrital zircons in a sample of the Silurian Siktefjellet Formation of NW Svalbard, they reconstructed both the Byscayarhalvøya and Vestgötabreen complexes to much further south of their present position in the vicinity of the mainly Ordovician Taconian/Grampian arc (Fig. 11c). Taconian/Grampian-aged detrital zircons are also common in Ordovician to Silurian sediments on Ireland and Scotland, which have been interpreted as being deposited in troughs related to this arc (Fig. 1; e.g. in the Southern Uplands, South Mayo or Midland Valley terranes; Waldron et al. Reference Waldron, Floyd, Simonetti and Heaman2008; Clift et al. Reference Clift, Carter, Draut, Van Long, Chew and Schouten2009; McConnell, Riggs & Crowley, Reference McConnell, Riggs and Crowley2009; Philipps et al. Reference Philipps, Smith, Stone, Pashley and Horstwood2009). Our Ordovician to Silurian samples 4 and 5 from the Bullbreen Group do not contain any Taconian/Grampian-aged zircons (Figs 6, 7), making a direct link of these sediments and the underlying eclogites to the Taconian/Grampian arc more difficult. Nevertheless, in the absence of more unambiguous indicators for provenance, we tentatively place the origin of the Vestgötabreen Complex and the overlying Bullbreen Group at the northern end and outboard of the Taconian/Grampian arc (Fig. 11c).
6.e. Ordovician to Devonian strike-slip assemblage of Svalbard
Independent of the exact location of the latest Mesoproterozoic and Neoproterozoic sediments and the high-pressure complex within the evolving Iapetus ocean, the above presented and discussed data (Figs 11a–c) indicate that early and late Caledonian sinistral strike-slip zones must have been crucial for the final assemblage of the Svalbardian terranes, similar to that originally proposed by Harland (Reference Harland1971, Reference Harland, Gee and Sturt1985; Fig. 11d). Our detrital data, together with the Neoproterozoic and Ordovician metamorphic history of the rocks from the western coast of Spitsbergen, indicate that this part of Svalbard could have originated somewhere proximal to northern Baltica, at least not proximal to northern Greenland (Fig. 11). In addition, these rocks escaped any severe Scandian metamorphic and magmatic overprint, which is omnipresent in Nordaustlandet and northwestern Spitsbergen. One possible explanation is a two-stage strike-slip model (Fig. 11c, d). During a first phase in the Ordovician, shortly after the formation of the Vestgötabreen Complex, the sinistral Vimsodden–Kosibapasset shear zone and potential other sinistral shear zones could have brought fragments from the Timanian margin and the northern Taconian/Grampian arc to the north of Greenland at c. 460 Ma, juxtaposing it with the rocks of the Sofiekammen and (Laurentian) Sørkapp groups (Fig. 11c; e.g. Mazur et al. Reference Mazur, Czerny, Majka, Manecki, Holm, Smyrak and Wypych2009). Tectonic activity roughly at that time is also known from the Timanian region of Novaya Zemlya (e.g. Pease & Scott, Reference Pease and Scott2009). Nordaustlandet and northwestern Spitsbergen (Krossfjorden Complex) were at the same time probably still situated further south offshore NE Greenland (Fig. 11c; e.g. Gee & Teben'kov, Reference Gee, Teben'kov, Gee and Pease2004). In a second sinistral strike-slip phase, mainly after the main Caledonian continent–continent collision occurred, Nordaustlandet and northwestern Spitsbergen were detached from NE Greenland and brought ‘out-of-sequence’ to the outside of the composite western Spitsbergen units. This probably happened in a generally sinistral transpressive to transtensive regime during the Silurian and Devonian, along the major faults bordering the Devonian graben of Svalbard (Figs 1, 11d; e.g. Soper et al. Reference Soper, Strachan, Holdsworth, Gayer and Greiling1992; Dewey & Strachan, Reference Dewey and Strachan2003). These late Caledonian faults probably defined the Caledonian trend now identified by a recent geophysical survey of the Barents Sea showing an anticlockwise-turning Caledonian trend connecting the onshore Scandinavian and Svalbard Caledonides (Fig. 11d; Gernigon & Brönner, Reference Gernigon and Brönner2012).
6.f. Carboniferous sedimentation
No Devonian sediments are exposed in Oscar II Land, and the Carboniferous Billefjorden Group tectonically overlies the underlying strata (Fig. 5). From Devonian outcrops elsewhere on Svalbard, it is known that this period was characterized by latest Caledonian intrusions and extension in parts of the basement, complex continental sedimentation and tectonism within mainly extensional and transtensional grabens, and a shift from a semi-arid to a more humid climate (e.g. Harland, Anderson & Manasrah, Reference Harland, Anderson and Manasrah1997). Detrital zircon data from two Devonian samples of northwestern Svalbard indicate that they are mainly derived from the adjacent Krossfjorden Complex with predominantly Mesoproterozoic detritus and very limited Archaean input (Petterson, Pease & Frei, Reference Petterson, Pease and Frei2010). Our sample 6 from the Carboniferous Billefjorden Group shows a more variable spectrum with four grains as young as 400–360 Ma, a few scattered grains between 500–700 Ma, a majority of 900–2100 Ma and a significant amount of concordant and discordant Archaean grains. Remarkably, no Scandian (c. 440–410 Ma) and no Taconian/Grampian grains are present (Figs 6, 7). Together with plate tectonic reconstructions from the Carboniferous, this supports a position of the now assembled parts of Svalbard to the north or northeast of Greenland, with detritus variably sourced from the underlying Meso- to Neoproterozoic sediments, including the sediments rich in Archaean detritus from northern Greenland (e.g. Kirkland et al. Reference Kirkland, Pease, Whitehouse and Ineson2009; Anfinson et al. Reference Anfinson, Leier, Embry and Dewing2012). Given the considerable amount of Archaean material, which probably was sourced from northern Greenland, the youngest 400–360 Ma zircons could be derived from the NE Greenland eclogite province (e.g. Gilotti, Nutman & Brueckner, Reference Gilotti, Nutman and Brueckner2004).
7. Conclusions
Detrital zircon data from six samples covering the entire available stratigraphic range of Oscar II Land, western Spitsbergen, allow the following conclusions for this area. (1) The oldest sample of the St Jonsfjorden unit gives a Mesoproterozoic maximum depositional age of ≤ 1016 ± 13 Ma and shows an age spectrum which is comparable to other latest Mesoproterozoic deposits in the North Atlantic region. It is therefore interpreted as a remnant of a latest Mesoproterozoic basin system which developed on top of or adjacent to the Grenvillian–Sveconorwegian orogen within the supercontinent Rodinia. Similarly aged rocks are absent in northern and northeastern Greenland. (2) An orthogneiss which probably occurs as a tectonic sliver within the sedimentary rocks of Oscar II Land reveals an intrusion age of 927 ± 3 Ma. It probably belongs to a magmatic event which is widespread in the aforementioned latest Mesoproterozoic basins, but its geodynamic origin is unknown. No intrusions of similar age are known from northern Greenland. (3) Detrital zircon patterns from the Neoproterozoic Daudmannsodden and West Coast diamictite unit indicate that they might have been deposited in a large-scale Neoproterozoic basin on the Baltican side of Iapetus. They show provenance patterns clearly different from similarly aged deposits from northeastern and northern Greenland. (4) Detrital zircon patterns from the Ordovician to Silurian Bullbreen Group indicate that the underlying Vestgötabreen Complex mainly consists of subducted Neoproterozoic sediments, intercalated with basalts of unknown origin. The age spectra do not allow a precise location of the high-pressure Vestgötabreen Complex and the overlying sediments within a Laurentia–Iapetus–Baltica system, but derivation offshore of northern Greenland is unlikely, owing to a prominent input of c. 980–960 Ma and 740–640 Ma zircons in the Ordovician–Silurian Bullbreen Group. (5) Sinistral strike-slip deformation in several episodes probably led to the final assemblage of the different pre-Devonian parts of Svalbard from the Ordovician to the Devonian. Southwestern Spitsbergen escaped the main Scandian continent–continent collision, and was therefore probably already located north of Greenland at the end of the Ordovician. (6) In the Carboniferous, sedimentation was probably dominated by local recycling of mainly Mesoproterozoic detritus as well as an upper Devonian input possibly from the NE Greenland eclogite province and an Archaean input from northern Greenland. Our study supports previous studies which indicated that the different pre-Devonian terranes of Svalbard experienced distinct geologic histories during Proterozoic and early Palaeozoic times and were assembled in their present form during the latest stages of the Caledonian orogeny. Particularly, the western coast of Spitsbergen consists of various pre-Devonian rocks which are difficult to bring into accordance with models deriving this part of Svalbard from just offshore the northern Greenland margin.
Acknowledgements
We thank Wilfried Dallmann and Synnøve Elvevold from the Norwegian Polar Institute for providing samples 1, 2, 3, 4, 5 and 7 from their archive. We thank Fernando Corfu and Gunborg Fjeld for help in the TIMS laboratory and Siri Simonsen for help with the LA-ICP-MS analyses at the University of Oslo. We thank Bernard Bingen, Vicky Pease and an anonymous reviewer for constructive comments.
