Devonian–Mississippian collapse and core complex exhumation, and partial decoupling and partitioning of Eurekan deformation as alternatives to the Ellesmerian

In the Late Devonian, Svalbard was affected by a short-lived episode of contraction called the Ellesmerian (Svalbardian) Orogeny, which resulted in top-west thrusting of Proterozoic 15 basement rocks onto Devonian sedimentary strata along the Balliolbreen Fault, a major fault segment of the east-dipping Billefjorden Fault Zone, and juxtaposition of undeformed Mississippian–Permian strata against intensely folded Devonian rocks. The present study of field and seismic data shows that backward-dipping duplexes comprised of phyllitic coal and beddingparallel décollements and thrusts localized along lithological transitions in thickened uppermost 20 Devonian–Mississippian coals and coaly shales of the Billefjorden Group partially decoupled uppermost Devonian–Permian sedimentary rocks of the Billefjorden and Gipsdalen groups from Devonian rocks during Cenozoic contraction–transpression. In addition, Devonian strata probably experienced syn-depositional, post-Caledonian, extensional, detachment-related folding. Seismic data in Sassenfjorden and Reindalspasset show the presence of Cenozoic duplexes and bedding25 parallel décollements within Lower–Middle Devonian, uppermost Devonian–Mississippian and uppermost Pennsylvanian–lowermost Permian sedimentary strata of the Wood Bay and/or Widje Bay and/or Grey Hoek formations, of the Billefjorden Group and of the Wördiekammen Formation respectively, which further decoupled stratigraphic units during Eurekan deformation. Beddingparallel décollements and thrusts are possibly related to shortcut faulting, a roof décollement of a 30 https://doi.org/10.5194/se-2019-200 Preprint. Discussion started: 13 January 2020 c © Author(s) 2020. CC BY 4.0 License.

However, despite the numerous works showing evidence in favor of such contractionaltranspressional tectonic event, a few studies highlight evident shortcomings of this tectonic model.
Notably, Rippington et al. (2010) noticed that this contractional episode, though well-studied through the Arctic (e.g., in Arctic Canada, northern Greenland, and central Spitsbergen) currently 70 lacks time constraints and, in places, structures ascribed to this event might partly belong to the early Paleozoic Caledonian Orogeny and/or to the Cenozoic Eurekan deformation event.
Furthermore, in western Norway, east-to northeast-plunging folds trending parallel to the latepost-orogenic extension direction in Middle Devonian collapse basins initially interpreted as contractional-transpressional structures formed during an episode of (Ellesmerian) Late 75 Devonian-Mississippian contraction (Roberts, 1983) were reinterpreted as transtensional folds formed during a late phase of the extensional collapse of the Caledonides (Chauvet and Séranne, 1994;Osmundsen and Andersen, 1994;Fossen et al., 2013).
The presented evidences question the occurrence of the Ellesmerian Orogeny in Svalbard, and broader implications of the proposed tectonic model include the occurrence of the Ellesmerian Orogeny in other Arctic regions like northeastern Greenland, northern Canada, northern Alaska, and northeastern Russia, and the actual timing of formation of fold and thrust structures ascribed 105 to this tectonic event, which might have formed (and were reactivated/overprinted) during Caledonian or Eurekan tectonism and/or during Devonian extensional collapse of the Caledonides.
The present study simplifies the tectonic history of Spitsbergen, arguing for a continuous episode of late-post-Caledonian extensional collapse-rifting throughout the Devonian-Carboniferous. However, future work is required to further test the proposed tectonic model and constrain the 110 tectonic setting during the deposition of Devonian-Mississippian sedimentary rocks throughout the Arctic, notably off the coasts of Alaska and northeastern Russia (Endicott Group), and of northern Canada (Emma Fiord Formation). 115

Caledonian Orogeny
The Svalbard Archipelago is composed of three terranes (excluding Bjørnøya) with significantly different geological histories. These terranes are thought to have started to assemble during Caledonian contraction-transpression and were juxtaposed against one another by N-Sstriking crustal faults like the Billefjorden Fault Zone (Harland and Wright, 1979;Ohta et al., 1989, 120 1995; Gee and Page, 1994). Caledonian deformation was accompanied by tectonothermal events with high-grade (eclogite and blueschist) metamorphism from mid-Cambrian to late Silurian times https://doi.org/10.5194/se-2019-200 Preprint. Discussion started: 13 January 2020 c Author(s) 2020. CC BY 4.0 License. that occurred during subduction and closure of the Iapetus Ocean and that are partly preserved in northwestern (Ohta et al., 1989) and western Spitsbergen (Horsfield, 1972;Kośmińska et al., 2014).
In Pyramiden, in Dickson Land (northern-central Spitsbergen; Figure 1b), Proterozoic basement rocks were thrusted top-west onto Lower Devonian sedimentary rocks of the Wood Bay Formation along the Balliolbreen Fault (Piepjohn, 2000;Bergh et al., 2011;Braathen et al., 2011) in Late Devonian times, and presumably undeformed Mississippian clastic and coal-bearing 175 sedimentary deposits of the Billefjorden Group overlie folded Lower Devonian metasedimentary rocks that were involved in Ellesmerian contraction (Piepjohn, 2000). West of Pyramiden, the lower Munindalen thrust transported Lower Devonian sedimentary rocks of the Wood Bay Formation over lower Frasnian Newman et al., 2019) sedimentary strata of the Plantekløfta Formation and is presumably unconformably overlain by flat-lying, 180 undeformed, uppermost Pennsylvanian-lower Permian strata of the Wördiekammen Formation (McCann and Dallmann, 1996;MIchaelsen et al., 1997;Piepjohn, 2000;Bergh et al., 2011). In Triungen (Figure 1a- and Dallmann, 1996). In Sentinelfjellet and Odellfjellet (Figure 1b), the Balliolbreen Fault juxtaposes Proterozoic basement rocks in the hanging wall against Devonian sedimentary rocks in the footwall and is thought to be unconformably overlain by undeformed, uppermost Devonian-Mississippian sedimentary rocks of the Billefjorden Group, thus suggesting Late Devonian contraction (Friend and Moody-Stuart, 1972;Lamar et al., 1986). Farther 190 north, in Andrée Land (Figure 1a), west-verging (i.e., opposite to Cenozoic folds in western and southern Spitsbergen) folds are believed to represent evidence of Ellesmerian tectonic movements, although timing remains speculative because most post-Devonian sedimentary rocks in the area were eroded (Piepjohn, 2000).

Restoration of the Adriabukta cross-section in southern Spitsbergen
The Adriabukta section (see figures 4 and 5 in Bergh et al., 2011) is a c. 5 km-wide, E-W 290 transect in southern Spitsbergen where erosion exposed Precambrian basement and Devonian-Cretaceous sedimentary rocks (Bergh et al., 2011). In order to investigate the initial geometry of Devonian-Mississippian sedimentary rocks and faults prior to Cenozoic transpression, the unconformity surface between the Hyrnefjellet and Adriabukta formations was restored to a horizontal geometry. In addition, Triassic-Cretaceous sedimentary rocks were removed together 295 with the effect of Mesozoic-Cenozoic normal faulting. A major change applied to the section was the rotation of the Mariekammen Shear Zone (Bergh et al., 2011), which crosscuts the Adriabukta Formation, by an angle equal to the rotation applied to the unconformity at the contact between the Adriabukta and Hyrnefjellet formations (c. 50° counterclockwise). There is no trace of Hecla Hoek basement in this area although field studies and geological maps suggest that Proterozoic basement were thrusted over Lower Devonian strata along the Balliolbreen Fault (McCann and Dallmann, 1996;Piepjohn et al., 1997;Dallmann, 1999;Dallmann et al., 2004;Bergh et al., 2011;Braathen et al., 2011;svalbardkartet.npolar.no). Sample preparation 315 for thin sectioning actually proved problematic for Devonian sedimentary rocks located in the hanging wall of the presumed fault, which resulted in a misleading thick section (supplement 1).
Thus, it is more likely that earlier maps showing exclusively Devonian-Mississippian sedimentary rocks of the Wood Bay Formation and Billefjorden Group below the mine entrance by , Lamar et al. (1986), andArktikugol (1988;Sirotkin, pers. comm. 2019) are correct.

320
Farther up the gully, a one-two meter-thick succession of interbedded sandstone and coal is juxtaposed against steeply east-dipping Devonian strata to the west and overlain by a (at least three meter) thick layer of uppermost Devonian-Mississippian coals of the Billefjorden Group that shows phyllitic shear fabrics ( Figure 2 and Figure 3b). Bedding surfaces within the one-two meterthick succession dip gently-steeply to the east ( Fig. Figure 3a), display sigmoidal geometries with 325 Z-like shapes, and terminate abruptly against the three meter-thick layer of uppermost Devonian-Mississippian phyllitic coal upwards and against Devonian rocks downwards (dashed yellow lines in Fig. Figure 3b). In addition, coaly shales within this succession display phyllitic fabrics similar to those observed within overlying coals, and seem to form repeated successions of alternating beds of sandstone and coaly shale truncated by steeply east-dipping sigmoidal fault surfaces (thin dashed 330 red lines in Fig. Figure 3b). The Z-like sigmoidal shape of bedding surfaces, phyllitic shear fabrics of the coaly shales, and possible repetitions of the succession suggest that the steeply east-dipping, sigmoidal faults crosscutting the succession are imbricate thrust faults (stereonet 3 in Figure 2 1992) with top-west to top-WNW sense of shear. In cross-section, the interaction of intrasuccession, steeply east-dipping link thrusts and inter-succession, moderate-low-angle floor-and 340 roof-thrusts defines an east-dipping duplex structure (Boyer and Elliott, 1982) of imbricate thrusts bounded upwards and downwards by potential décollements and/or detachments parallel to original (i.e., prior to deformation) bedding surfaces (e.g., transition from interbedded coaly shales and sandstone to coal, and from coal to sandstone; Figure 3b). The nomenclature of hindward/forwarddipping duplexes of Boyer and Elliott (1982) does not apply here since the foreland of the West 345 Spitsbergen Fold-and-Thrust Belt (Tertiary Central Basin) is located southeast of Pyramiden. Thus, the term "backward" is used to describe the east-dipping character of the duplexes, i.e., opposite to the inferred transport direction.
Above the mine entrance, sedimentary rocks of the Billefjorden Group are dominated by yellow sandstone that are crosscut by dominant WNW-ESE-striking fractures and subsidiary N-350 S-and ENE-WSW-striking fractures (stereonets 1 and 2 in Figure 2) showing oblique-slip kinematics. Poorly preserved slickenside lineations did not yield any information on relative displacement between footwall and hanging wall. In the west, dark sandstone and quartzite crop out and contain wood fossils, which are probably Devonian in age. The contact between the Devonian dark sandstone and uppermost Devonian-Mississippian yellow sandstone of the 355 Billefjorden Group, and intra-Devonian lithological contacts (e.g., between Devonian quartzite and dark sandstone; Fig. Figure 3a), although partly covered by screes and/or mostly made of loose blocks, do not appear to be faulted and trend c. WNW-ESE to NW-SE as bedding surfaces appear to change from moderately-steeply east-dipping below the mine entrance to gently NNE-dipping Devonian rocks and the overlying succession of uppermost Devonian-Mississippian sandstone, coaly shale and coal ( Fig. Figure 3b) correspond to the upward-flattening continuation of this fault.
However, no fault was observed between Devonian rocks and sandstones of the Billefjorden Group above the mine, and lithogical and stratigraphic contacts there display significantly different trends (WNW-ESE to NW-SE; Fig. Figure 3a). 375

Triungen
Satellite images in Triungen (Figure 1a-b) show that the Triungen-Grønhorgdalen Fault Zone (McCann and Dallmann, 1996)  Group in the hanging wall are largely covered by dark screes (Fig. Figure 3c). Based on the presence of thick, flat-lying, coal-rich strata in the lower part of the Billefjorden Group overlying Lower Devonian sedimentary strata in the hanging wall of the fault, the dark screes along the fault trace (right hand-side inset in Fig. Figure 3c) are believed to represent uppermost Devonian-Mississippian coals-coaly shales that might have been dragged along the Triungen-Grønhorgdalen 385 Fault Zone during tectonic movements.
In Reindalspasset (Figure 1a-b), potential Devonian rocks display partly disrupted, semicontinuous, sub-parallel to chaotic, moderate-to low-amplitude seismic reflections (Figure 4g). amplitude seismic reflections that are most likely the product of acoustic impedance contrast between low density coal seams interbedded with clastic deposits. Such seismic facies is relatively common for uppermost Devonian-Mississippian sedimentary rocks in the Norwegian Barents Sea (Koehl et al., 2018;Tonstad, 2018). In Reindalspasset, uppermost Devonian-Mississippian, phyllitic, coal-rich deposits of the Billefjorden Group were penetrated by exploration well 7816/12- identified on seismic data because they crop out at sea level along the northern shore of Sassenfjorden and Tempelfjorden and, hence, can be directly tied to onshore geology (Dallmann et al., 2004(Dallmann et al., , 2009Dallmann, 2015). Mesozoic sedimentary rocks are not the focus of the present study and were therefore not described.

Structures in Sassenfjorden-Tempelfjorden
Seismic data in Sassenfjorden-Tempelfjorden (Figure 1a tentatively related to tectonic thickening due to Cenozoic thrusting and, potentially, to the presence of partially mobile evaporite within the Gipshuken Formation (Dallmann, 1999).

Discussion
In Reindalspasset, potential décollements and low-angle thrusts folded into a gentle upright Hoek formations in Andrée Land (Roy, 2007Roy et al., unpublished).

595
Based on the significant differences in deformation styles, it is probable that the décollements and backward-dipping duplexes in sheared uppermost Devonian-Mississippian coals-coaly shales decoupled Cenozoic deformation between tightly folded, shale-rich, Lower  known to be able to decouple deformation both in contractional (Frodsham and Gayer, 1999, their figures 1b, 2, 7 and 9) and extensional settings (Wilson and Wojtal, 1986, their figures 7 and 10).
Uppermost Devonian-Mississippian coal-rich strata are locally thicker in Pyramiden, thus 625 resulting in their exploitation by the Russian until the early 90s (Livshitz, 1966;Cutbill et al., 1976) Based on field and seismic data in central Spitsbergen (present study;Koehl and Muñoz-Barrera, 2018) and on analog modelling (Bonini, 2001), it is probable that Devonian sedimentary deposits were folded in Cenozoic times since the differences in deformation style and intensity  (Boyer and Elliott, 1982) due to lateral lithological variations within Pennsylvanian formations (Ringset and Andresen, 1988). In addition, in western Spitsbergen, Maher (1988), Saalmann and Thiedig (2000) and Bergh and Andresen

Mississippian rocks in Pyramiden
Backward-dipping duplexes in Pyramiden are located along the eastern limb of a major N-680 S-trending fold in Devonian sedimentary rocks (Figure 2), thus far ascribed to the Ellesmerian Orogeny (Piepjohn, 2000;Bergh et al., 2011). It is possible that, during Cenozoic folding, Devonian rocks in the west may have acted as a rigid buttress that localized the formation of duplexes and décollements within relatively soft, uppermost Devonian-Mississippian coals and coaly shales of the Billefjorden Group and allowed these structures to ramp upwards to the west.

685
This is supported by field studies (Fard et al., 2006) and analog modelling (Bahroudi and Koyi, 2003)   Above the coal mine in Pyramiden, the contact between Devonian sedimentary strata and uppermost Devonian-Mississippian sedimentary rocks is not clearly exposed (partly loose blocks) and its nature is speculative. It may be (1) a (folded?) stratigraphic unconformity and/or (2) a bedding-parallel décollement. Based on the internal geometry of bedding surfaces and deformation 790 state of uppermost Devonian-Mississippian sedimentary strata of the Billefjorden Group, which are arranged into contractional, west-verging duplexes separated by low-angle, bedding-parallel décollements ( Fig. Figure 3b), it is possible that the stratigraphic contact hosts a décollement, e.g., the potential prolongation of one of the décollements within coal-and coaly shale-rich deposits of the Billefjorden Group (Figure 2 and Figure 3b). However, Mississippian deposits above the coal 795 mine appear to consist only of clastic deposits and, hence, lack soft coals-coaly shales into which décollements preferentially form. Thus, the contact between Lower Devonian and uppermost (fold-limb-parallel) fault related to post-Caledonian gravitional collapse processes (Chorowicz, 1992) and low-angle detachments (e.g., the Woodfjorden detachment in Andrée Land; Roy, 2007Roy et al., unpublished; Figure 1a) in Devonian sedimentary rocks in northern Spitsbergen, or formed as a minor, bedding-parallel Cenozoic accommodation thrust (e.g., Cosgrove, 2015). Devonian times as proposed by previous works (Vogt, 1938;Friend, 1961;Piepjohn, 2000;Dallmann and Piepjohn, submitted).

915
The presence of backward-dipping duplexes and décollements localized within coal seams and coaly shales in the lower part of the Billefjorden Group in Pyramiden (Figure 3b In Triungen (Figure 1a-b), the base of the Billefjorden Group is covered by (black) screes (Playford, 1962; Fig. Figure 3c). Hence, it is conceivable that, there too, the base of the Billefjorden Group consists of highly deformed, coal-rich sedimentary rocks that localized Cenozoic transpression like in Pyramiden, with potential décollements and/or roof-and floor-thrusts decoupling Devonian rocks from overlying Carboniferous-Permian sedimentary strata. The structural setting at the Triungen locality is, indeed, very similar to that in Pyramiden, in that it involves a major east-dipping brittle fault, the Triungen-Grønhorgdalen Fault Zone McCann and Dallmann, 1996). This fault represents a potential analog to the Billefjorden Fault Zone and may have been active during the deposition of thick, uppermost  Mississippian sedimentary rocks Cutbill et al., 1976;McCann and Dallmann, 1996). It is therefore possible that the Triungen-Grønhorgdalen Fault Zone might, just as the Billefjorden Fault Zone in Pyramiden, have accommodated the deposition of thickened, coal-rich sedimentary deposits (Livshitz, 1966;Cutbill et al., 1976), thus making the interpretation of Cenozoic strain (partial) decoupling relevant for the Triungen area as well. Noteworthy, the 935 presence of black screes on the southern flank of Triungen near the presumably faulted contact between Lower Devonian and uppermost Devonian-Mississippian sedimentary strata (Fig. Figure   3c) may indicate the presence of thick coal seams and/or coaly shales along the Triungen-Grønhorgdalen Fault Zone and, conceivably, of bedding-parallel décollements, thrusts and duplexes similar to those in Pyramiden (Fig. Figure 3b). 940 Just west of Pyramiden, two moderately east-dipping faults, the upper and lower Munindalen thrusts, crosscut Devonian sedimentary strata and are possibly of Ellesmerian age (Michaelsen et al., 1997;Figure 1b  , is unconformably overlain by uppermost Pennsylvanian-lower Permian sedimentary strata of the Wördiekammen Formation, and, hence, is more likely to correspond to a Devonian low-angle normal fault or detachment or to an unconformity tilted during core complex exhumation in the west (e.g., Braathen et al., 2018). The lower Munindalen thrust is similar to post-Caledonian extensional detachments mapped in Andrée Land, e.g., the Woodfjorden 950 detachment (Chorowicz, 1992;Roy, 2007Roy et al., unpublished) (Roy, 2007Roy et al., unpublished), thus showing that folding of Devonian strata in Spitsbergen partly occurred prior to Ellesmerian contraction and was partly related to gravitational collapse of the Caledonides. The presence of extensional shear zones in Devonian rocks in central Spitsbergen was also evidenced by Michaelsen et al. (1997, their figure 5a) and Michaelsen (1998, 960 her figures 44 and 45) in Munindalen (Figure 1b).
By contrast, the lower Munindalen thrust McCann and Dallmann, 1996;Piepjohn, 2000;Bergh et al., 2011)   . The Robertsonbreen thrust presents strong similarities to 970 the lower Munindalen thrust, including a NNW-SSE strike, gentle to moderate eastward dip, topwest/southwest reverse sense of shear, and alignment along a NNW-SSE-trending axis (see the alignment in Figure 1b). Shall these two thrusts represent the same fault, it is possible to explain the apparent differences in behavior, i.e., the lower Munindalen thrust potentially not crosscutting the Wördiekammen Formation (Piepjohn, 2000;Dallmann, 2015) and the relatively undeformed  (Ringset and Andresen, 1988;Harland 980 et al., 1988), Brøggerhalvøya (Bergh et al., 2000;Saalmann and Thiedig, 2000) and Oscar II Land (Maher, 1988;Bergh and Andresen, 1990;Bergh et al., 1997), and are especially well illustrated by Cenozoic thrusts within the Gipshuken Formation that flatten downwards and die out within sedimentary strata of the Wördiekammen Formation in Sassenfjorden-Tempelfjorden (Figure 4c and f). In addition, taking into account the evidence for Cenozoic contraction in uppermost 985 Devonian-Mississippian coals-coaly shales in Pyramiden (e.g., Fig. Figure 3b) Spitsbergen used by, e.g., Piepjohn (2000) and Bergh et al. (2011), to infer a Late Devonian age for the Munindalen thrust and identify Late Devonian vertical tectonic movements in the area is regarded as inappropriate since the Wördiekammen Formation in Pyramiden was obviously  (Michaelsen et al., 1997;Piepjohn et al., 1997).
Like for the lower Munindalen thrust, the contact of the Blåvatnet Reverse Fault with Pennsylvanian-Permian sedimentary deposits of the Wördiekammen Formation is not exposed   figure 12), and, thus, the fault may well correspond to a Cenozoic 1005 thrust.
The presence of a "lower décollement level" was also speculated by Bergh and Andresen (1990) in Brøggerhalvøya (Figure 1a), in western Spitsbergen, in order to explain the observed Cenozoic deformation. They speculated that this "lower décollement level", possibly analog to those observed in Pyramiden (Figure 3b), might flatten into syn-rift Carboniferous sedimentary 1015 strata, and uppermost Devonian-Mississippian (coaly) shales and coals of the Billefjorden Group in Brøggerhalvøya (Fairchild, 1982) definitely represent suitable candidates to have localized the formation of such a décollement. Other arguments against the occurrence of the Ellesmerian Orogeny are the restricted extent of presumed Ellesmerian deformation belts (Piepjohn, 2000), and the undeformed character of 1020 slightly tilted Lower Devonian rocks of the Wood Bay Formation in, e.g., Pretender Mountain (western Spitsbergen; Figure 1a), which are unconformably overlain by flat-lying uppermost Carboniferous-lowermost Permian strata of the Wördiekammen Formation (Welbon et al., 1992, unpublished;Dallmann, 2012Dallmann, , 2015). If Ellesmerian contraction-transpression had occurred, it would most likely have folded Devonian strata throughout Spitsbergen, which is not  (Koehl, 2019;Koehl et al., in prep.).

and the SW Barents
Sea (Koehl et al., 2018) and potentially in Northeast Greenland (Sartini-Rideout et al., 2006;Hallett et al., 2014;McClelland et al., 2015), thus making it a reasonable alternative to the Ellesmerian Orogeny. Renewed or continued core complex exhumation (e.g., of the Bockfjorden Anticline) in northern Spitsbergen and associated normal brittle faulting  might also be 1040 responsible for the presence of coarse-grained sedimentary deposits (Piepjohn and Dallmann, 2014) in the lower Frasnian  Planteryggen and Plantekløfta formations in central Spitsbergen.
small-scale structure that is, hence, potentially inappropriate to discuss the occurrence and extent 1050 of regional deformation events. For example, the breccia seems to have escaped Cenozoic deformation (Buggisch et al., 1994;Kempe et al., 1997), which resulted in intense top-NE thrusting and folding in adjacent areas on Brøggerhalvøya, just south/southwest of Blomstrandhalvøya ( Figure 1a). Thus, it is conceivable that the breccia might, as well, have escaped Ellesmerian contraction, making it inappropriate to constrain the timing of thrusting and folding on  To further support Ellesmerian contraction in Svalbard, Thiedig and Manby (1992) and Kempe et al. (1997) argued that the west-and NW-verging thrusts observed on Blomstrandhalvøya are not typical of Cenozoic Eurekan deformation, which produced NE-verging thrusts a few kilometers to the south/southeast on Brøggerhalvøya (Bergh et al., 2000;Piepjohn et al., 2001; Figure 1a), even though NW-verging thrusts seem to have formed in the Cenozoic. Ongoing work shows that 1065 Cenozoic contraction in Spitsbergen was partitioned and that Brøggerhalvøya and Blomstrandhalvøya are separated by a major NW-SE-striking, sinistral-reverse oblique-slip fault that extends from Kongsfjorden to Sassenfjorden and the northern Barents Sea (Koehl, 2019;Koehl et al., in prep.). In addition, on the one hand, basement marbles are poorly deformed on Blomstrandhalvøya, especially in the westernmost part of the peninsula where the Pennsylvanian-

1070
Permian karst breccia crops out, away from the cracked and cataclased contact between basement marbles and Lower Devonian sedimentary rocks of the Red Bay Group. On the other hand, the Pennsylvanian-Permian karst breccia does look mildly deformed (Buggisch et al., 1994, their figure 4a-b). Hence, deformation intensity in the karst breccia and the marbles is not significantly different and, therefore, is not an appropriate indicator to constrain the timing of deformation on 1075 Blomstrandhalvøya.
New paleontological and palynological data cast new light on the potential time span of Ellesmerian contraction, which is believed to have initiated after the deposition of the Fiskekløfta Formation in the late-latest Givetian (i.e., latest Middle Devonian; Berry and Marshall, 2015) and terminated prior to the deposition of middle-late Famenian-Mississippian (Scheibner et al., 2012; 1080 https://doi.org/10.5194/se-2019-200 Preprint. Discussion started: 13 January 2020 c Author(s) 2020. CC BY 4.0 License. Lindemann et al., 2013;Marshall et al., 2015;Würtzen et al., 2019;Lopes, pers. comm. 2019) coalrich sedimentary rocks of the Billefjorden Group (Piepjohn, 2000), which implies a maximum duration of ca. 15-16 Ma or this tectonic event. In addition, Ellesmerian contraction is believed to have been recorded by the deposition of the Planteryggen and Plantekløfta formations in Dickson Land (Piepjohn, 2000;Piepjohn and Dallmann, 2014), which recently yielded early Frasnian (ca.

1085
380 Ma) ages based on fossil and spore assemblages for both formations . Based on these new ages, it is possible that Ellesmerian contraction was actually restricted to ca. 383-380 Ma. The intense folding and thrusting ascribed to Ellesmerian contraction by Piepjohn (2000) and other workers could hardly have formed during such a short time interval and are therefore more likely explained by Devonian collapse-related extension and Cenozoic 1090 contraction-transpression.
In Adriabukta, two thin bedding-parallel dolerite sills were intruded within the Adriabukta Formation near the contact with a lens of basement rocks (Birkenmajer, 1964, his figure 2). Since 1120 these two sills probably intruded along Upper Devonian (-Mississippian?) bedding surfaces they are most likely Late Devonian in age or younger. Sills typically intrude the bedrock at depth of 1-5 km (Schmiedel et al., 2017), but in places occur at a few hundreds of meter depths (Bell and Butcher, 2002). Knowing that the Adriabukta Formation is c. 600 m-thick (cumulated) in Adriabukta (Birkenmajer, 1964;Figure 1a), it is possible, though not likely, that the two dolerite 1125 sills are Late Devonian-Mississippian in age. However, the only known Devonian-Mississippian intrusions in Svalbard are dykes (Evdokimov et al., 2006;Senger et al., 2013, their Figure 1c).
Thus, it is more probable that the two dolerite sills in Adriabukta are part of the Cretaceous Diabasodden Suite (Senger et al., 2013). Since sills are generally planar and that two dolerite sills in Adriabukta intruded along bedding surfaces, it is highly probable that Upper Devonian (-1130 Mississippian?) bedding surfaces of the Adriabukta Formation were still planar (i.e., relatively undeformed) in Cretaceous times during the intrusion of the sills, which therefore suggests that folding of the Adriabukta Formation most likely initiated after sill intrusion, i.e., most likely in the early Cenozoic.
In Røkensåta (southernmost Spitsbergen; Figure 1a), gently dipping Lower Triassic 1135 sedimentary rocks overlie folded Devonian strata and, therefore, potentially support Late Devonian Ellesmerian transpression (Dallmann, 1992). However, based on the stratigraphy of the Triassic succession in southern Spitsbergen, which is dominated by interbedded sandstone and shale (Worsley and Mørk, 1978), on the stratigraphic contact between Devonian and Triassic rocks in Røkensåta being covered by screes, on the limited exposure of Triassic rocks in Røkensåta, and on 1140 the presence of sub-horizontal Cenozoic thrust faults/décollements in Triassic sedimentary rocks in Reindalspasset (Figure 4g; Eide et al., 1991), it is conceivable that a potential unidentified  (Maher, 1984;Maher et al., 1986Maher et al., , 1989Andresen et al., 1988;Bergh and Andresen, 1990;Dallmann et al., 1993b;Bergh et al., 1997), the most spectacular examples being the décollement in dark shales on the Midterhuken Peninsula (Maher, 1984;Dallmann et al., 1993b) and the Berzeliustinden thrust in Triassic-Lower Cretaceous bituminous shales (Dallmann, 1988) 1150 in southern Spitsbergen, and the "Lower Décollement Zone" in western (Andresen, 2009) and eastern Spitsbergen . Alternatively, folding in Devonian strata in Røkensåta might reflect upwards propagation of a Cenozoic thrust fault (similar to that of Haremo et al., 1990, their figure 14) that dies out upwards or that flattens into a décollement level at the base of the Triassic succession, or post-Caledonian detachment-related 1155 folding (e.g., in Andrée Land; Roy, 2007Roy et al., unpublished). Another alternative to the model of core complex exhumation and partial decoupling (present study) and to Ellesmerian contraction (Piepjohn, 2000) might be a potential contractional event in the Permian-earliest Triassic (Uralian Orogeny?). However, no such event was ever reported in Svalbard.
The eastern part of the Adriabukta transect was restored prior to Cenozoic deformation and 1160 Mesozoic sedimentation, and the present study reaches substantially different conclusions from Birkenmajer and Turnau (1962, their (Koehl, 2019;Koehl et al., in prep.) and northwestern Spitsbergen (Braathen et al., 2018). Thus, Cambrian basement rock lenses incorporated along the Mariekammen Shear Zone might represent a large clast that was eroded from a basement 1180 culmination (core complex?; Figure 7b1) and/or that was ripped off by excisement and/or incisement processes (Lister and Davis, 1989) along a bedding-parallel, core complex-bounding detachment in the west in Devonian-Mississippian times (Figure 7b2), comparable to the Woodfjorden detachment (Roy et al., unpublished). Exhumation of a N-S-to NNW-SSE-trending core complex would also explain the angular unconformity inferred by Dallmann (1992) between  (Birkenmajer, 1964, his figure 4).

Late Devonian-earliest Mississippian amphibolite facies metamorphism in western Spitsbergen
The Ellesmerian Orogeny is believed to have halted prior to the onset of deposition of 1220 sedimentary strata of the Billefjorden Group (Piepjohn, 2000). New ages for sedimentary rocks at the base of the Triungen Member (Billefjorden Group) in Triungen (Lindemann et al., 2013; Figure   1a-b) suggest that sedimentation initiated in the middle-late Fammenian, thus implying that Ellesmerian contraction had stopped by ca. 365 Ma. Though, Piepjohn and Dallmann (2014) claim that Fammenian spore assemblages in the Triungen Member described by Lindemann et al. (2013) 1225 were reworked and are older than the actual age of sedimentation, recent data from Marshall et al. (2015) and Lopes (pers. comm. 2019) contradict this claim and further support a middle-late  (Figure 1a) indicate that these rocks were subjected to amphibolite metamorphism during a latest Devonian-earliest Mississippian tectonic event 1240 Kośmińska, 2017). In addition, preliminary K-Ar (Schneider et al., 2018) and 40 Ar-39 Ar (Faehnrich et al., 2017) dating of muscovite yielded 360-355 Ma ages for amphibolite metamorphism and mylonitization along a one-km-thick west-dipping shear zone, the Bouréefjellet fault zone. Another possible way to explain amphibolite facies metamorphism and mylonitization in latest Devonian-earliest Mississippian time is the exhumation of basement rocks in western 1245 Spitsbergen as metamorphic core complex during late-post Caledonian collapse of the Caledonides. Amphibolite facies metamorphism and mylonitization are commonly related to lateorogenic collapse processes (Krabbendam and Dewey, 1998) and core complex exhumation (Snoke, 1980;Lister and Davis, 1989;Beaudoin et al., 2015;Yin et al., 2017). This hypothesis is supported by recent findings in northwestern  and central Spitsbergen 1250 (Koehl, 2019;Koehl et al., in prep.) discussing the exhumation of a N-S-to NNW-SSE-trending core complex, the Bockfjorden Anticline, along the low-angle Devonian Keisarhjelmen detachment, which latest movement was dated to ca. 368 Ma by K-Ar geochronology of a syntectonic granitic dyke in detachment mylonite , and by down-SW to down-NW shear senses obtained from muscovite mica fish and quartz sigma-clasts within amphibolite-  Other common features of metamorphic core complexes are the juxtaposition of high-grade rocks in the footwall with low-grade rocks in the hanging wall of a crustal detachment, and a transition from ductile to brittle deformation from the base to the top of the detachment (Lister and Davis, 1989;Huet et al., 2011). In Prins Karls Forland, the west-dipping Bouréefjellet fault zone (Schneider et al., 2018) separates amphibolite-facies mylonitized metapelites of the Pinkiefjellet 1270 Unit in the footwall in the (south-) east from greenschist-facies siliciclastic metasedimentary strata in the hanging wall in the (north-) west (Manby, 1983;Maraszewska et al., 2016;Schneider et al., 2018), and fault-rocks within the fault zone range from dominantly mylonitic in the lower part to brittle in the upper part (Schneider et al., 2018). In addition, recrystallization processes in quartz within the Bouréefjellet fault zone, including subgrain rotation and bulging (350-450°C) in the 1275 upper part and grain boundary migration (> 500°C) in the lower part (Schneider et al., 2018), indicate temperature ranges comparable to other core complexes around the world, e.g., the hightemperature Ikaria Metamorphic Core Complex in Greece (Beaudoin et al., 2015).

Competing interests
The author declares that he has no conflict of interest. Ph.D. Thesis of John G. Gjelberg (1984) was also digitized and, thanks to the University of Bergen and to John G. Gjelberg's family, is now available from the University of Bergen Library at http://bora.uib.no/handle/1956/20981.