Interactive comment on “ Neoproterozoic and post-Caledonian exhumation and shallow faulting in NW Finnmark from K / Ar dating and p / T analysis of fault-rocks ”

shorted to make the salient points more accessible to the readership. Comment 2: The ages are distinct from other high temperature and Ar/Ar texturally constrained results from Finnmark which illustrate the importance of Caledonian processes. Much better evaluation of the results in the context of these other geochronometers is needed. Specifically, why are the ages so different, is it simply due to different fabrics being dated, a different closure temperature of the system and mineral, or are secondary alteration processes at play? It is trivial to address this question but is needed to place the work in the current regional context. The reported ages may very well be correct but then more detail is required to put them in the context of post Caledonian process (in the main). Comment 3: More evidence is required to demonstrate that the clay minerals all formed at the same time and that they have not been subsequently modified with later fluid movement on the “dated” structures. For example, more SEM and EBSD textural work would be a distinct advantage in addressing how many different fabrics are associated with each of the sampled structures. Comment 4: What is the potential of fine fraction feldspar modifying the K/Ar results? Is there evidence of significant fluid alteration on these structures? Comment 5: I am somewhat concerned with the number of references to unpublished works that are cited as “submitted” (e.g. Koehl et al., Davids et al.,). I do not think it acceptable to heavily rely on such currently unpublished work so I would suggest that a summary of the salient points in those unpublished works is presented in this paper so all the evidence for statements in this work is available to readers. Comment 6: Line 16: sentence structure I think you mean “during the opening of”: : :. Line 34: replace “which” with “whose”. Comment 7: Line 50; you probably should mention the timing of collisional phases as constrained this region prior to discussing post-Caledonian extension. See Kirkland et al., 2006, The structure and timing of lateral escape during the Scandian Orogeny: A combined strain and geochronological investigation in Finnmark, Arctic Norwegian Caledonides, which discusses the constraint on the timing of the collisional phase (431-428 Ma). Comment 8: Line 65: you need to present the evidence or at least discuss (if it is already published) the evidence for Timanian deformation in northern Finnmark. Comment 9: More

Abstract.Well-preserved fault gouge along brittle faults in Paleoproterozoic, volcano-sedimentary rocks of the Raipas Supergroup exposed in the Alta-Kvaenangen tectonic window in northern Norway yielded latest Mesoproterozoic (approximately 1050 ± 15 Ma) to mid-Neoproterozoic (approximately 825-810 ± 18 Ma) K-Ar ages.Pressure-temperature estimates from microtextural and mineralogy analyses of fault rocks indicate that brittle faulting may have initiated at a depth of 5-10 km during the opening of the Asgard Sea in the latest Mesoproterozoic-early Neoproterozoic (approximately 1050-945 Ma) and continued with a phase of shallow faulting to the opening of the Iapetus Ocean-AEgir Sea and the initial breakup of Rodinia in the mid-Neoproterozoic (approximately 825-810 Ma).The predominance and preservation of synkinematic smectite and subsidiary illite in cohesive and non-cohesive fault rocks indicate that Paleoproterozoic basement rocks of the Alta-Kvaenangen tectonic window remained at shallow crustal levels (< 3.5 km) and were not reactivated since mid-Neoproterozoic times.Slow exhumation rate estimates for the early-mid-Neoproterozoic (approximately 10-75 m Myr −1 ) suggest a period of tectonic quiescence between the opening of the Asgard Sea and the breakup of Rodinia.In the Paleozoic, basement rocks in NW Finnmark were overthrusted by Caledonian nappes along low-angle thrust detachments during the closing of the Iapetus Ocean-AEgir Sea.K-Ar dating of non-cohesive fault rocks and microtexture mineralogy of cohesive fault rock truncating Caledonian nappe units show that brittle (reverse) faulting potentially initi-ated along low-angle Caledonian thrusts during the latest stages of the Caledonian Orogeny in the Silurian (approximately 425 Ma) and was accompanied by epidote-chloriterich, stilpnomelane-bearing cataclasite (type 1) indicative of a faulting depth of 10-16 km.Caledonian thrusts were inverted (e.g., Talvik fault) and later truncated by high-angle normal faults (e.g., Langfjorden-Vargsundet fault) during subsequent, late Paleozoic, collapse-related widespread extension in the Late Devonian-early Carboniferous (approximately 375-325 Ma).This faulting period was accompanied by quartz-(type 2), calcite-(type 3) and laumontite-rich cataclasites (type 4), whose cross-cutting relationships indicate a progressive exhumation of Caledonian rocks to zeolitefacies conditions (i.e., depth of 2-8 km).An ultimate period of minor faulting occurred in the late Carboniferousmid-Permian (315-265 Ma) and exhumed Caledonian rocks to shallow depth at 1-3.5 km.Alternatively, late Carboniferous (?) to early-mid-Permian K-Ar ages may reflect late Paleozoic weathering of the margin.Exhumation rates estimates indicate rapid Silurian-early Carboniferous exhumation and slow exhumation in the late Carboniferous-mid-Permian, supporting decreasing faulting activity from the mid-Carboniferous.NW Finnmark remained tectonically quiet in the Mesozoic-Cenozoic.
The main goal of this work is to constrain the timing of brittle fault initiation (Proterozoic and/or post-Caledonian) and to discuss the reactivation and exhumation history of faults in basement rocks and Caledonian units of NW Finnmark.We focus on margin-parallel brittle faults in basement rocks of the Alta-Kvaenangen tectonic window in the Altafjorden area, for comparison with exposed fault segments of the Neoproterozoic, margin-oblique TKFZ (Fig. 1).We also mapped and analyzed post-Caledonian fault segments and splays of the margin-parallel LVF to compare with the age of similar, offshore, basin-bounding faults (e.g., Troms-Finnmark and Måsøy fault complexes; Fig. 1) and associated syntectonic sedimentary rocks on the Finnmark Platform and in offshore basins, e.g., the Hammerfest and Nordkapp basins (Gabrielsen et al., 1990;Indrevaer et al., 2013).
A second goal is to evaluate the amount of exhumation the margin underwent from the Neoproterozoic to the Caledonian Orogeny and from the collapse of the Caledonides to present times.Thus, we sampled cataclastic fault rocks along multiple brittle faults including brittle faults that crosscut Archean-Paleoproterozoic rocks in fresh road cuts in Altafjorden, and faults like the LVF and TKFZ in Caledonian thrust nappe rocks (Fig. 1).Then, we analyzed characteristic mineral assemblages for each cataclastic fault rock, i.e., each faulting event recorded, which we used in conjunction with cross-cutting relationships between fault rocks to reconstruct the evolution of p/T conditions (i.e., depth) during faulting and thus resolve the exhumation history of the margin.We compare our results with those from analog studies in western Troms (Davids et al., 2013(Davids et al., , 2018;;Indrevaer et al., 2014) and Finnmark (Torgersen et al., 2014), and discuss regional implications for the tectonic evolution of the Troms-Finnmark margin during post-Caledonian extension.
The three fault sets commonly form two major fault systems.On the one hand, NNE-SSW-and ENE-WSW-striking faults commonly interact to produce zigzag-shaped fault complexes.An example in NW Finnmark is the LVF, which is made up of alternating ENE-WSW-and NNE-SSWstriking fault segments (Zwaan and Roberts, 1978;Lippard and Roberts, 1987;Koehl et al., 2018b) and resembles the offshore, basin-bounding, zigzag-shaped Troms-Finnmark Fault Complex in map view (Fig. 1; Indrevaer et al., 2013).This resemblance is verified by offshore prolongation of the LVF into shallow Carboniferous half-graben on the eastern Finnmark Platform (Koehl et al., 2018a).On the other hand, WNW-ESE-striking faults are usually observed as swarms of high-frequency, (sub)parallel fractures.A key example is the TKFZ, a major Neoproterozoic fault zone that extends from the Varanger Peninsula in the east and crops out onshore the island of Magerøya in the west (Fig. 1).This fault complex is made up with multiple segments of subparallel, WNW-ESE-striking brittle faults that die out just west of Magerøya (Koehl et al., 2018b).Several of these fault segments were intruded by highly magnetic dolerite dykes during an early Carboniferous extensional event (Roberts et al., 1991;Lippard and Prestvik, 1997;Nasuti et al., 2015), when the TKFZ acted as a strike-slip transfer fault and segmented onshore-nearshore areas of NW Finnmark from the offshore eastern Finnmark Platform during late/post-Caledonian extension (Koehl et al., 2018a, b).

Dating of brittle faults in north Norway
Previous K-Ar dating of synkinematic illite/muscovite in fault gouge in NW Finnmark shows that brittle faults in the Repparfjord-Komagfjord tectonic window, like the Kvenklubben fault, formed in Precambrian times and were reactivated as Caledonian thrusts and, later on, as normal faults during late/post-orogenic extension and subsequent rifting (Torgersen et al., 2014).Non-cohesive fault rocks sampled by Torgersen et al. (2014) showed enrichment in authigenic smectite and chlorite clay minerals and a rather low content of illite/muscovite.Farther southwest along the margin, in western Troms, similar coastal brittle faults display cohesive cataclastic fault rocks with epidote, chlorite and pumpellyite (Indrevaer et al., 2014), and these faults are juxtaposed against amphibolitefacies Precambrian rocks of the West Troms Basement Complex (Zwaan, 1995;Bergh et al., 2010).These faults are interpreted to have formed during late Paleozoic extension at depth > 10 km and to have been exhumed to shallower crustal level < 8.5 km, as shown by the widespread occurrence of pumpellyite mineral in fault rocks.These faults yielded late Paleozoic K-Ar ages and are possibly associated with post-Caledonian extension in the Devonian-Carboniferous (Davids et al., 2013).

Structural field data
In summer 2015, we acquired extensive structural field data along brittle faults in NW Finnmark, which we compiled, interpreted and discussed in earlier contributions (Bergø, 2016;Lea, 2016;Koehl et al., 2018b).Among the numerous brittle faults cropping out in NW Finnmark, we selected 10 based on their proximity to major faults (e.g., LVF, TKFZ), location relative to major faults (footwall/hanging wall) and according to their strike (parallel to dominant fault trends in Finnmark), which we briefly describe from a structural perspective at outcrop scale.Fault geometries and kinematic indicators will be used in conjunction with cohesive and noncohesive fault-rock compositions and with the results of K-Ar dating of fault gouge to propose an evolutionary model for the tectonic evolution and exhumation history of the SW Barents Sea margin and NW Finnmark.

Microscopic analysis of onshore cohesive fault rock
We collected cohesive fault-rock samples along numerous brittle faults that we encountered in NW Finnmark (including the 10 dated faults), and we used them to investigate kinematic indicators along the selected brittle faults at microscale.We also studied mineral assemblages included in brittle fault rocks in order to constrain metamorphic facies (p/T ) conditions during faulting, therefore adding to the understanding of the exhumation and uplift history of the SW Barents Sea margin.When needed, thin sections were analyzed through an optical microscope and a scanning electron microscope (SEM) at the University of Tromsø to obtain more detailed information about mineral composition.

K-Ar dating and mineralogical analysis of fault gouge
We sampled non-cohesive fault rock along brittle faults in NW Finnmark and attempted to date authigenic (i.e., synkinematic) illite clay mineral formed during faulting events (e.g., Vrolijk and van der Pluijm, 1999;Davids et al., 2013;Torgersen et al., 2014;Ksienzyk et al., 2016), whose platy crystal shape differs from their irregular detrital counterpart (e.g., Davids et al., 2013;Torgersen et al., 2014).The 10 dated samples of non-cohesive fault rock are referred to as samples 1-10 in the text and figures.K-Ar dating of fault gouge was carried out in the K-Ar laboratory facility at the University of Göttingen, Germany.Three grain-size fractions were analyzed for each sample: "2-6 µm", "< 2 µm" and "< 0.2 µm".Clay-rich fault gouge samples were resolved in water and wet-sieved using a 63 µm sieve.The fraction < 63 µm was used to extract the clay fractions < 2 µm by settling in Atterberg cylinders.The fractions < 0.2 µm have been separated using an ultra-centrifuge.All these fine fractions were examined using X-ray diffraction (XRD) for mineralogical composition and determination of the illite crystallinity using a Philips PW 1800 diffractometer.Illite crystallinity, the peak width at half height of the 10-Å peak, was determined using a computer program developed at the University of Göttingen.Digital measurement of illite crystallinity was carried out by step scan (301 points, 7-10 • 2 , scan step 0.010 • 2 , integration time 4 s, receiving slit 0.1 mm, automatic divergence slit).Illite crystallinity determinations have been shown to be a sensitive indicator for the degree of very low-grade metamorphism in clastic sedimentary rocks.Reviews of the preparation techniques and the interpretation have been given, for example, by Kisch (1991) and Krumm (1992).All samples have been investigated in duplicates (A and B).The measurements were carried out in the "air dry" status and the "ethylene glycol saturated" status in order to detect expandable layers of smectite-type minerals.Smectite classification (Reichweite) has been determined following Moore and Reynolds (1997).Illite crystallinity is expressed as the Kübler index (KI, • 2 ); the limits for diagenesis/anchizone (approximately 200 • C) and anchizone/epizone (300 • C) are 0.420 and 0.250 • • 2 (Kübler, 1967(Kübler, , 1968(Kübler, , 1984)), respectively.
The argon isotopic composition was measured in a pyrex glass extraction and purification line coupled to a Thermo Scientific ARGUS VI ™ noble gas mass spectrometer operating in static mode.The amount of radiogenic 40 Ar was determined by an isotope dilution method using a highly enriched 38 Ar spike from Schumacher, Bern (Schumacher, 1975).The spike is calibrated against the biotite standard HD-B1 (Fuhrmann et al., 1987).The age calculations are based on the constants recommended by the IUGS quoted in Steiger and Jäger (1977).Potassium was determined in duplicate by flame photometry using a BWB-XP flame photometer ™ .The samples were dissolved in a mixture of hydrogen fluoride (HF) and HNO 3 according to the technique of Heinrichs and Herrmann (1990).The analytical error for the K-Ar age calculations is given on a 95 % confidence level (2σ ).Details of argon and potassium analyses for the laboratory in Göttingen are given in Wemmer (1991).

Temperature constraints from illite-smectite clay minerals
Since the dominant synkinematic clay mineral in the analyzed fault rocks are smectite and subsidiary interlayered illite-smectite clay, we use the smectite-illite clay mineral reaction to infer maximum/minimum temperature estimates for faulting events in NW Finnmark (Eberl et al., 1993;Huang et al., 1993;Morley et al., 2018).Synkinematic illite commonly grows due to illitization of smectite and, alternatively, due to dissolution-precipitation of existing clay minerals of the bedrock (Vrolijk and van der Pluijm, 1999).Illitization along fault surfaces is enhanced by temperature increase, e.g., related to frictional heating, hydrothermal processes or burial, grain comminution, strain, changes in fluid composition and fluid-rock ratio (Vrolijk and van der Pluijm, 1999).Illitization of smectite is commonly thought to begin at a temperature range of 40-70 • C (Jennings and Thompson, 1986;Harvey and Browne, 2000;Ksienzyk et al., 2016).

Interpretation of inclined age spectra
K-Ar dating requires targeted minerals to behave as "closed systems" with no loss of argon or potassium (Lyons and Snellenburg, 1971).Mineral closure temperature varies with grain size and is lower for finer grains.For example, aggregates of fine grains may accidentally be incorporated and dated as part of coarser fractions of fault gouge, thus causing coarse fractions to yield younger ages than finer fractions (Hamilton et al., 1989;Heizler and Harrison, 1991).More specifically along shallow faults, illite grains < 2 µm crystallize below the closure temperature of the K-Ar system (> 250 • C; Velde, 1965) and thus yield robust, synkinematic crystallization ages rather than less accurate, generally younger cooling ages obtained along deeper faults (Hunziker et al., 1986;Ksienzyk et al., 2016).Further, contrary to metamorphism-related heating, the short heating time associated with hydrothermal events or frictional heating along brittle faults is unlikely to reset illite ages, which would require longer exposure to temperatures > 250 • C (Torgersen et al., 2014), thus suggesting that K-Ar ages on illite along shallow faults provide reasonable estimates of the age of faulting.Nevertheless, we emphasize that the analytical data presented as ages cannot be treated like high-precision ages from, e.g., modern U-Pb zircon dating, but rather point to a time interval that can, in some cases, be much larger than the analytical error given in the 2σ error interval.
Mixing of host-rock-inherited minerals, e.g., detrital illitemuscovite, with authigenic illite may influence K-Ar ages and cause age dispersion in faults, notably in the coarser fractions dated (Hower et al., 1963;Vrolijk and van der Pluijm, 1999).However, inherited illite-muscovite may be distinguished from authigenic clay minerals as they display more irregular shapes than their generally platy authigenic counterparts (e.g., Torgersen et al., 2014).In addition, faulting may even isotopically reset fine-grained, host-rock illitemuscovite, thus yielding ages bearing no influence of inherited, older minerals (Vrolijk and van der Pluijm, 1999).In addition, fault-inherited illite may also affect K-Ar ages, which is especially verified along repeatedly active, progressively exhumed faults because high-temperature illite may survive low-temperature reactivation of the faults (Davids www.solid-earth.net/9/923/2018/Solid Earth, 9, 923-951, 2018et al., 2013;Viola et al., 2013).Another mineral that may have a significant impact on K-Ar ages is host-rock-inherited K-feldspar.Most importantly, K-feldspar has a significantly lower closure temperature (350-150 • C) than illite clay mineral (> 250 • C), hence yielding younger ages than the actual age of faulting, particularly for finer fractions in fault gouges (Lovera et al., 1989).Hornblende may also affect K-Ar dating of illite (Torgersen et al., 2015b) but was not encountered in any fault-rock samples in NW Finnmark and its effects are therefore not considered here.

Results
We sampled brittle fault rocks for K-Ar dating and microtextural analysis from several dominant fault systems and fault trends in NW Finnmark (Fig. 1; Koehl et al., 2018b).The sampling sites include (i) faults in Paleoproterozoic rocks of the Raipas Supergroup (Zwaan and Gautier, 1980) in Altafjorden (samples 3 and 4), (ii) faults in the Caledonian Kalak Nappe Complex along segments and splay faults of the LVF (samples 1, 2, 5, 6 and 7), and (iii) faults in rocks of the Kalak Nappe Complex and Magerøy Nappe (samples 8, 9 and 10) adjacent to segments of the TKFZ (Fig. 1).For the sampled faults, we first describe trends, field relations and kinematics of the faults.Second, we describe the mineral assemblages and microstructures of sampled cohesive fault rocks in order to infer deformation mechanisms and estimate the p/T conditions during faulting and exhumation.Third, we present the K-Ar data and mineralogical results obtained on fault gouge.

Field relations of sampled brittle faults in NW Finnmark
4.1.1Brittle faults in Paleoproterozoic basement rocks (samples 3 and 4) Samples 3 and 4 (Fig. 2a and b) are from Altafjorden faults 1 and 2, NW-dipping brittle faults in new, fresh road cuts along the western shore of Altafjorden, truncating metaarkoses of the Paleoproterozoic Raipas Supergroup (Skoadduvarri Sandstone Formation) of the Alta-Kvaenangen tectonic window (Zwaan and Gautier, 1980;Bergh andTorske, 1986, 1988).The two sampled faults are located a few tens of meters away from each other, display meter-thick fault cores mostly composed of grey-colored clay particles in non-cohesive fault gouge and partly cohesive cataclasites (Fig. 2a and b).The Altafjorden faults' cores also show multiple slip surfaces cemented by centimeter-thick quartz grains (Fig. 2a).Slickenside lineations on fault surfaces indicate normal dip-slip movement.A mafic band is bent into (drag folded) the fault core facilitating normal down-NW sense of shear (Fig. 2a).The absence of this mafic bed in the footwall of the Altafjorden fault suggests that the fault accommodated vertical displacement > 5-6 m.
4.1.2Brittle faults along the Langfjorden-Vargsundet fault (samples 1, 2, 5, 6 and 7) We sampled fault rocks along several segments/splay faults of the LVF (Torgersen et al., 2014;Koehl et al., 2018b).This regional fault complex can be traced from Sørkjosen in the south (location of samples 1 and 2 in Fig. 1) to the Porsanger Peninsula in the north (Fig. 1) and defines a zigzag-shaped pattern of alternating, NNE-SSW-to ENE-WSW-striking fault segments that dominantly dip WNW and NNW, respectively (Koehl et al., 2018b).Samples 1 and 2 are taken from two minor NW-dipping fault splays (Fig. 2c and d) in the footwall of the Sørkjosen fault, a major fault segment of the LVF (Fig. 1; Koehl et al., 2018b).The faults cross-cut granodioritic gneisses of the Kalak Nappe Complex and display thin, 10-40 cm thick fault cores with lenses of dark clayish gouge material (Fig. 2c and d).The northern fault accommodates approximately 5-8 m top-NW, normal displacement of a 20-30 cm thick layer of mafic amphibolite (Fig. 2c), while the southern fault offsets the same mafic layer by approximately 2 m of top-NW motion (Fig. 2d).Slickenside lineations along these two faults support normal dip-slip sense of shear (Fig. 2c and d).
Along strike to the northeast, we sampled another subsidiary fault linked to the LVF along the western shore of Altafjorden, the Talvik fault (sample 5; Fig. 1), a low-angle, north-dipping fault that cross-cuts arkosic psammites of the Kalak Nappe Complex (Fig. 2e).The Talvik fault shows evidence of both brittle and ductile faulting in an approximately 1 m thick fault core composed of semi-ductile, mylonitic fault rock, overprinted by calcite-and quartz-bearing cataclasite, as well as thin layers of non-cohesive fault gouge along distinct fault surfaces (Fig. 2e).Field observations of quartz sigma clasts and S-C fabrics in the mylonites reveal top-south thrusting, whereas slicken grooves and asperities are present along distinct brittle fault surfaces, implying topnorth motion along the Talvik fault (Fig. 2e).Brittle offset of ductile quartz sigma clasts confirms that brittle fabrics are younger than ductile fabrics.
Sample 6 was taken along a high-angle WNW-dipping brittle fault that crops out along the eastern shore of Altafjorden (Figs. 1 and 2f).This fault defines the southeastern boundary of a graben structure in garnet-rich psammite of the Kalak Nappe Complex and is characterized by an approximately 1 m thick fault core made of non-cohesive clayish fault gouge and adjacent epidote-rich cataclasite (Fig. 2f).Slickenside lineations indicate down-WNW, dip-slip normal movement, which is consistent with normal dip-slip offsets of boudinaged mafic dykes across nearby NNE-SSWstriking faults (Fig. 2f).
The final fault-rock sample along LVF segments and splays (sample 7; Fig. 1) is from the steep SE-dipping Snøfjorden-Slatten fault on the Porsanger Peninsula (Figs. 1 and 2g;Passe, 1978;Townsend, 1987b), which represents a major, antithetic fault segment/splay of the LVF (Koehl et al., 2018b).This fault cross-cuts felsic metasedimentary rocks and micaschists of the Kalak Nappe Complex and displays a several-meter-wide fault core that is made of non-cohesive iron-and quartz-rich fault gouge, including a few lenses of cohesive fault rock (Fig. 2g).Slickensided fault surfaces reveal down-SE, normal dip-slip movement, probably a few meters to a few tens of meters due to the presence of the same host rock on both sides of the fault (Fig. 2g).jacent to a lens of preserved Magerøya Nappe rocks on the Porsanger Peninsula (Kirkland et al., 2007).This fault comprises thin, decimeter-scale lenses of dark clay particles along the fault core and damage zone with splaying fault geometries (Fig. 2h).Oblique normal-sinistral movement is inferred from slickenside lineations (Fig. 2h), and the amount of displacement probably does not exceed a few tens of meters because garnet-bearing gneisses occur on both sides of the fault.Farther north, on Magerøya, sample 9 corresponds to a steep, WNW-ESE-to E-W-striking, south-dipping fault (Figs. 1 and 2i) in a new quarry within a suite of weakly foliated gabbroic rocks of the Honningsvåg Igneous Complex (Robins, 1998;Corfu et al., 2006).The sampled fault core includes two thin, 5-10 cm thick layers of light-colored, non-cohesive clay particles (Fig. 2i).
The last fault we sampled for K-Ar dating (sample 10) was taken along a steep NNE-dipping brittle fault in the western part of Magerøya (Figs. 1 and 2j).This fault is part of a high-frequency, WNW-ESE-striking lineament and brittle fault system that corresponds to fault segments of the TKFZ, which pervasively truncate metasedimentary rocks of the Kalak Nappe Complex in western Magerøya (Koehl et  al., 2018b) and displays an approximately 0.5 m thick fault core made up with light-colored clay particles (Fig. 2j).

Mineralogy and microtextural analysis of cohesive
fault rock 4.2.1 Cohesive fault rocks within the Alta-Kvaenangen tectonic window In order to describe and analyze cohesive brittle fault-rock characters and mineralogical and textural changes during cataclasis, the host rock characters are used as frame.Cohesive fault rock was found along most exposed brittle faults cross-cutting Precambrian volcano-sedimentary rocks of the Alta-Kvaenangen tectonic window (Fig. 2a and b).The host rocks are fairly undeformed but underwent low-grade (greenschist-facies) metamorphic conditions during the Svecofennian Orogeny, and developed a weak, bed-parallel foliation (Bøe and Gautier, 1978;Zwaan and Gautier, 1980;Bergh and Torske, 1988).This foliation comprises partly recrystallized, sigma-shaped grains of quartz and feldspar locally incorporated into an S-C foliation made of elongated crystals of white mica (Fig. 3a).
Brittle faults analyzed in the present study cross-cut both metasandstones and metamorphosed carbonate-rich host rocks, and include three types of cataclasites and mineral precipitation.The first type shows a matrix of finely crushed clasts of quartz (Fig. 3b).The second type is made of a partly healed, calcite-cemented cataclasite, whose crystals display both type II and IV twinning (Ferrill, 1991;Burkhard, 1993;Fig. 3c).The third type includes cataclasite with abundant brownish to reddish matrix of very fine-grained clay-(smectite and subsidiary illite) and iron-rich minerals associated with iron-bearing precipitation, commonly truncating veins of recrystallized quartz and quartz-rich cataclasite (Fig. 3b).Iron-bearing precipitation appears to localize along fractures developed parallel to preexisting, white-mica, S-C-C foliation (Fig. 3b).Relative timing of quartz-, calcite-and clayrich cataclasite could not be directly resolved from crosscutting relationships.

Cohesive fault rocks within Caledonian nappes
Metamorphic Caledonian host rocks consist of a variety of granodioritic gneisses, metasandstones, metapelites, amphibolites/metavolcanites, gabbros and mica schists.When truncated by brittle faults, most rocks are altered and/or display retrograde mineral assemblages.Notably, in mafic/granodioritic host rocks, biotite is systematically retrograded into chlorite in the vicinity of brittle faults (Fig. 3d), host rocks are generally enriched in epidote at the expense of amphibole, and garnet porphyroblasts are highly fractured (Fig. 3e).Brittle faults cross-cutting Caledonian rocks comprise up to 1-2 m wide lenses of fractured host rocks showing preserved ductile fabrics, such as widespread muscovite-biotite and S-C-C foliation with feldspar sigma clasts partly recrystallized into quartz (e.g., along the Sørkjosen and Snøfjorden-Slatten faults; Fig. 3d and f).Typically, S, C and C foliation surfaces are partly replaced by cataclastic fault rocks and thus may have localized subsequent brittle faulting (Fig. 3g), as observed along the brittle-ductile Talvik fault (Fig. 2e).
We identified five texturally different types of cataclasites cross-cutting each other in a systematic order.The first type of cataclasite (type 1) is composed of fine-grained, rounded to subrounded clasts of epidote and chlorite (Fig. 3e and h), commonly reworked into large, angular clasts incorporated into subsequent cement or cataclastic matrix (Fig. 3e and  i).Epidote-chlorite fracture precipitation sometimes appears undeformed (Fig. 3j).In places, epidote-chlorite-bearing veins comprise rounded clasts of a brownish mineral showing moderate to strong relief (Fig. 3h).Energy dispersive spectroscopy (EDS) analysis reveals that this mineral is enriched in calcium, iron, silica, aluminium, magnesium and oxygen (Fig. 3k), and thus may correspond to stilpnomelane (Eggleton, 1972).
The second type of cataclasite (type 2) is composed of very fine-grained, rounded to subrounded clasts of quartz (Fig. 3h).This type of cataclasite is commonly observed adjacent to host rock clasts (Fig. 3e), regularly incorporates angular clasts of epidote-rich cataclasite (Fig. 3e and h) and occurs in conjunction with veins of recrystallized quartz (Fig. 3h).
The fourth type of cataclasite (type 4) consists of abundant, new-grown, mildly cataclased, prismatic/columnar mineral grains with acute edges and steep-oblique terminations showing low relief, low refraction index and three sets of cleavage (Fig. 3e and m).We interpret this mineral as laumontite, i.e., a high-temperature zeolite mineral (Dill et al., 2007;Triana et al., 2012).In the study area, laumontite crystals commonly grew with their long-crystallographic axis perpendicular to brittle fractures (Fig. 3m) and consistently cross-cut epidote-and quartz-rich cataclasites and epidotechlorite-and calcite-filled veins (Fig. 3e).
The fifth type of cataclasite (type 5) shows enrichment in very fine-grained, iron-oxide-bearing mineral precipitation and an even more fine-grained, microscopic matrix of brownish and greyish clay minerals (Fig. 3e, h, l and n).These cataclasites truncate and commonly incorporate clasts of epidote-chlorite-, quartz-and calcite-rich cataclasite and associated mineral veins (Fig. 3e, l and n).

Mineralogy
X-ray fluorescence (XRF) analyses of various grain-size fractions of the sampled fault gouges in NW Finnmark consistently show (1) high smectite content commonly associated with chlorite (mixed-layer chlorite-smectite), e.g., samples 1, 2, 6 and 8 (Fig. 1), (2) a relatively low content in illite and (3) variable amount of residual quartz (Table 1 and Appendix A).The only exception is sample 8 from the Porsanger Peninsula (Fig. 2h), which contains higher amounts of illite and quartz together with smectite, chlorite and kaolinite clay minerals (Table 1 and Appendix A).Since the mineralogical composition is dominated by smectite and chlorite, the specific peaks for the different illite polytypes are not recognizable due to peak overlap.The analysis of the diffraction spectrum for all three grain-size fractions indicates that the fault gouge from western Magerøya (sample 10) is composed of almost pure smectite (Fig.  1 and Appendix A).Sample 9 from the Honningsvåg Igneous Complex on Magerøya (Fig. 2i) shows very low potassium content (Table 2), which resulted in high 2σ errors associated with the K-Ar ages obtained for all three fractions (Table 2).We also noticed the presence of possible laumontite and/or stilbite in this sample (Table 1 and Appendix A; Triana et al., 2012).

Precambrian ages
All three dated fractions of samples 3 and 4 (Figs.theless, these K-Ar ages from well-preserved, non-cohesive fault gouges suggest that Altafjorden fault 1 (Fig. 2a) formed in late Mesoproterozoic times and was reactivated at least once in the Neoproterozoic (early Cryogenian).Younger, post-Caledonian reactivation seems unlikely, as the sensitivity of the K-Ar geochronometer would certainly have recorded subsequent reactivation by yielding younger ages.Similar Precambrian K-Ar ages were obtained for the three fractions of sample 4 from Altafjorden fault 2 (Figs.  5).Considering the proximity of Altafjorden faults 1 and 2 (samples 3 and 4; Fig. 1), and the comparable ages obtained for these faults, we consider the K-Ar ages to be reliable and likely reflecting synkinematic crystallization of authigenic illite during protracted Mesoproterozoic-Neoproterozoic tectonic events.These results further suggest that Altafjorden faults 1 and 2 were not reactivated in the Phanerozoic, as further reactivation would necessarily have been recorded and resulted in younger ages.

Late Paleozoic ages
Most dated fault gouges from segments of the LVF and TKFZ in the Kalak Nappe Complex and Magerøy Nappe on the Porsanger Peninsula and Magerøya yielded late Paleozoic ages ( The intermediate and fine-grained fractions of sample 6 along the eastern shore of Altafjorden (Figs. 1 and 2f) yielded earliest Permian and latest Carboniferous synkinematic crystallization ages, respectively (292.6 ± 4.0; 298.5 ± 5.0 Ma), while the coarse-grained fraction yielded a Jurassic age of 208.5 ± 3.1 Ma (Table 2, Figs. 4 and 6).A possible explanation for this discrepancy is that the coarse fraction of sam-   1 and 4. Columns to the right show extensional tectonic events related to the formation of the NW Baltoscandian basins during the opening of the Asgard Sea (Siedlecka et al., 2004;Nystuen et al., 2008;Cawood et al., 2010;Cawood and Pisarevsky, 2017) and to the earliest phase of rifting of the Iapetus Ocean-AEgir Sea (Li et al., 1999(Li et al., , 2008;;Torsvik and Rehnström, 2001;Hartz and Torsvik, 2002), and associated geological periods and eras.ple 6 partly consists of aggregates of smaller grains reflecting a much younger faulting event (Hamilton et al., 1989;Heizler and Harrison, 1991).However, XRF analysis of the sample suggests that the anomalous, younger age obtained for the coarse fraction may be the product of excess potassium due to the presence of K-feldspar in the sample (Table 1 and Appendix A; Lovera et al., 1989).
The sample taken along the low-angle, WNW-ESEstriking fault on the Porsanger Peninsula (sample 8; Figs. 1  and 2h) yielded similar latest Carboniferous-earliest Permian ages all included within a 5 Ma time span of 302.3 ± 6.5, 297.6 ± 7.7 and 296.6 ± 3.8 Ma (Table 2, Figs. 4  and 6).These synkinematic crystallization ages suggest that the fault was not reactivated after the earliest Permian, as the finest fraction would have recorded a younger faulting event and thus yielded a younger age.It is, however, possible that the fault accommodated earlier faulting events.
Fault gouge sampled along a WNW-ESE-striking fault segment of the TKFZ (Koehl et al., 2018b) near Gjesvaer in western Magerøya (sample 10; Figs. 1 and 2j) yielded late Carboniferous (312.5 ± 8.7 Ma), early Permian (284.0 ± 6.2 Ma) and mid-Permian (270.8 ± 6.0 Ma) ages, respectively, for the coarse, intermediate and finest fractions, which we all interpret as syntectonic crystallization ages.These ages suggest that the fault experienced multiple extensional faulting events from the late Carboniferous to mid-Permian (Table 2, Figs. 4 and 6).and 4. Columns to the right of the graph display major contractional (based on Ramberg et al., 2008;Vetti 2008;Corfu et al., 2014) and extensional events (Faleide et al., 1993;Steltenpohl et al., 2011;Koehl et al., 2018a) 2).This, together with a high contamination of atmospheric argon, resulted in high errors for all three dated fractions.Considering the high 2σ error percentage associated with the K-Ar ages obtained for this sample, the coarse fraction may cover a time span of faulting from the mid-Carboniferous (early Serpukhovian) to the Late Pennsylvanian (Gzhelian), 315.6 ± 13.6 Ma (Table 2, Figs. 4 and 6).The finest fraction exhibits an even higher 2σ error percentage, and the age window included within 2σ interval spans from the early Permian to the Middle Triassic (Table 2, Figs. 4 and 6).The intermediate fraction yielded a younger age (234.7 ± 18.0 Ma) than the finest fraction (265.2 ± 23.6 Ma), which is considered to be erroneous.The ages obtained for the coarse and fine fractions are interpreted to represent synkinematic crystallization of authigenic illite.Brittle faulting along this fault most likely initiated in the mid-Carboniferous and the fault was later reactivated in the Permian.

Mesozoic ages
Mesozoic K-Ar ages were obtained for the Snøfjorden-Slatten fault on the Porsanger Peninsula (sample 7; Figs. 1  and 2g), which yielded Middle (238.0 ± 5.4 Ma) to Late Triassic (227.4 ± 5.3 Ma) to Hettangian (earliest Jurassic; 200.4 ± 6.0 Ma) ages (Table 2, Figs. 4 and 6).This sample, however, contains minor K-feldspar in the coarse and intermediate grain-size fractions, which may have induced an excess of potassium, hence yielding ages younger than the actual age of faulting (Lovera et al., 1989).Nonetheless, the Hettangian age obtained for the finest fraction seems reasonable and most likely reflects synkinematic crystallization during an actual faulting event.
The fault gouge of the coarse, intermediate and fine grainsize fractions of sample 1 taken near the Sørkjosen fault segment of the LVF (Figs. 1 and 2c) yielded a latest Triassic (Rhaetian) age of 206.8 ± 2.6 Ma, and Late Jurassic (late Kimmeridgian) ages of 153.2 ± 3.7 and 153.4 ± 1.9 Ma, respectively (Table 2, Figs. 4 and 6), possibly suggesting that this fault splay of the LVF formed in the latest Triassic and was reactivated in the Late Jurassic.Similar K-Ar ages were obtained for another splay fault of the LVF in Sørkjosen (sample 2; Figs. 1 and 2d), i.e., Olenekian 247.6 ± 3.7 and 249.4 ± 3.3 Ma (Early Triassic) ages for the coarse and intermediate fractions and a latest Middle Jurassic age of 164.4 ± 4.5 Ma for the finest fraction (Table 2, Figs. 4  and 6).Similarly to the other fault in Sørkjosen (sample 1), it is possible that the gouge in sample 2 formed in the Early Triassic and was reactivated during the Middle Jurassic (Table 2, Figs. 4 and 6).However, a minor K-feldspar content observed in the diffraction spectra of all three grain-size fractions of both of these faults suggests that the K-Ar ages probably postdate the actual faulting, but it is uncertain by how much time (Table 1 and Appendix A).

Discussion
We combine mineral assemblages in cohesive and noncohesive fault rocks to reconstruct the faulting and burialexhumation history of the NW Finnmark margin, using an average geothermal gradient of 30 • C km −1 based on well data in adjacent portions of the SW Barents Sea (Bugge et al., 2002;Chand et al., 2008;Vadakkepuliyambatta et al., 2015), and utilizing the K-Ar dating results of authigenic illites in non-cohesive fault rocks to constrain the timing of faulting.The discussion starts with the estimated p/T conditions and the Mesoproterozoic-Neoproterozoic K-Ar ages obtained for Altafjorden faults 1 and 2, and proceeds with mid-late Paleozoic, and finally, Mesozoic exhumation (p/T ) and faulting data obtained from the LVF and TKFZ as the basis for comparison with p/T constraints and K-Ar faulting ages from western Troms.

Evolution of temperature conditions in Precambrian rocks
Microtextural and mineralogical analysis of cohesive fault rocks along Altafjorden faults 1 and 2 (Fig. 2a and b) shows that brittle faulting initiated with the formation of quartz-and calcite-rich cataclasites (Fig. 3b and c).On the one hand, quartz-rich cataclasite was derived from a foliated psammitic host rock with quartz/feldspar sigma clasts (Figs.2a, b and  3a).On the other hand, calcite-cemented cataclasite commonly incorporates crystals with type II and IV twinning, which indicates that these crystals were subjected to temperature ranges of 150-300 • C (i.e., 5-10 km depth) and > 250 • C (depth > 8 km), respectively (Fig. 3c;Ferrill, 1991;Burkhard, 1993).Quartz-rich and calcite-cemented cataclasites are truncated and, in places, incorporated into subsequent iron-/clayrich cataclasites (Fig. 3b and c).XRF analyses of noncohesive fault rocks sampled along Altafjorden faults 1 and 2 (samples 3 and 4) show a dominance of smectite (Table 1 and Appendix A), which suggests that the dominant clay mineral in related iron-/clay-rich cohesive fault rock shown in Fig. 3b is smectite.Considering such a predominance of authigenic smectite in both non-cohesive and cohesive, iron-/clay-rich fault rocks (Fig. 3b, Table 1 and Appendix A), and assuming a complete diagenetic transformation of smectite into illite at approximately 105 • C (Morley et al., 2018) and a complete absence of authigenic illite at temperature < 35 • C (Eberl et al., 1993), we propose that clay-rich fault rocks along brittle faults in the Alta-Kvaenangen tectonic window formed at temperature conditions comprised between 35 and 105 • C (i.e., 1-3.5 km depth).Although cross-cutting relationships of calcite-rich cataclasite with quartz-and iron-/clay-rich fault rocks are unknown, the irreversibility of the diagenetic transformation of smectite into illite (Eberl et al., 1993) suggests that calcite-cemented cataclasite, which formed at 5-10 km depth, is older than the iron-/clay-rich (cohesive and non-cohesive) fault rocks, which formed at shallow depth 1-3.5 km.Hence, we argue that Precambrian basement rocks in Altafjorden experienced at least three brittle faulting events, starting when quartz-rich cataclasite (Fig. 3b) and/or calcitecemented cataclasite formed at a depth of 5-10 km (Ferrill, 1991;Burkhard, 1993;Fig. 3c).Subsequently, basement rocks were exhumed to a shallow crustal level < 3.5 km (Morley et al., 2018) when the final iron-and smectite-rich faulting event occurred (Fig. 3b).

Timing of faulting and exhumation of Precambrian rocks
The latest Mesoproterozoic (approximately 1050 Ma) -early Neoproterozoic ages (approximately 945 Ma) obtained for the coarse fraction of sample 3 and the coarse and intermediate fractions of samples 4 (Table 2, Figs. 4 and 5), and the slickenside lineations and drag-folded foliation indicating down-NNW normal motions along both faults (Fig. 2a and  b) suggest that Altafjorden faults 1 and 2 contributed to the initial stages of formation of the NW Baltoscandian basins (Siedlecka et al., 2004;Nystuen et al., 2008) during the rifting of the Asgard Sea (Cawood et al., 2010;Cawood and Pisarevsky, 2017).Possible driving mechanisms for the formation of these basins and faults are a far-field influence of the coeval, basin-oblique/orthogonal, Sveconorwegian contraction, i.e., a formation as impactogenic rift basins (Barberi et al., 1982), and/or a possible influence of late/post-orogenic collapse of the Sveconorwegian Orogeny (Bingen et al., 2008;Viola et al., 2013).The latest Mesoproterozoic-early Neoproterozoic ages (approximately 1050-945 Ma) obtained on illite in coarsest and intermediate fractions of brittle fault rocks (Table 2, Figs. 4 and 5), which we interpreted as synkinematic crystallization ages, suggest that basement rocks of the Alta-Kvaenangen tectonic window were already exhumed above the brittle-ductile transition at that time, and may provide a maximum estimate for the age of quartz-and calcite-rich cataclasites formed at a depth of 5-10 km along these faults (Fig. 3b and c).
The finest fractions of both samples and intermediate fraction of sample 3 of non-cohesive fault rocks in basement rocks yielded mid-Neoproterozoic ages (approximately 825-810 Ma; Table 2, Figs. 4 and 5), which we interpreted as crystallization ages.Combining these ages with normal shearsense indicators observed in the field (Fig. 2a and b), we propose that they represent the onset of rifting of the Iapetus Ocean-AEgir Sea during the breakup of Rodinia between 825 and 740 Ma (Torsvik and Rehnström, 2001;Hartz and Torsvik, 2002;Li et al., 2008).Similar Neoproterozoic K-Ar ages of approximately 790-780 and 740-735 Ma are reported from dating of authigenic illite/muscovite along the Kvenklubben and Porsavannet faults in the adjacent www.solid-earth.net/9/923/2018/Solid Earth, 9, 923-951, 2018 Repparfjord-Komagfjord tectonic window in NW Finnmark (Torgersen et al., 2014;Fig. 1).
Altafjorden faults 1 and 2, although located close and oriented subparallel to Caledonian thrust faults (e.g., the Talvik fault; Fig. 2e) and major, post-Caledonian normal faults (e.g., the LVF; Fig. 1), were most likely not reactivated after approximately 810 Ma (mid-Neoproterozoic; Table 2, Figs. 4 and 5), as subsequent faulting would have triggered younger mineral assemblages and ages.Possible explanations for the non-reactivation of these faults include a north-to westward (basinward?) migration of rifting to areas adjacent to the LVF, e.g., the Kvenklubben and Porsavannet faults dated at approximately 790-735 Ma (Torgersen et al., 2014), and to faults in Troms and northern Finnmark, where Ediacaran metadolerite dykes intruded basement rocks during the breakup of the Iapetus Ocean-AEgir Sea (Zwaan and van Roermund, 1980;Siedlecka et al., 2004;Nasuti et al., 2015).The lack of reactivation of Altafjorden faults 1 and 2, predominance of authigenic smectite clay mineral in non-cohesive fault rock (samples 3 and 4 in Table 1 and Appendix A) and the irreversibility of smectite-illite transformation (Eberl et al., 1993) suggest that Precambrian rocks of the Alta-Kvaenangen tectonic window were exhumed and have remained at shallow depth < 3.5 km since the mid-Neoproterozoic (approximately 825 Ma; Table 2, Figs. 4  and 5).This conclusion is supported by predominance and preservation of authigenic smectite in similar non-cohesive fault rocks in the Repparfjord-Komagfjord tectonic window (Torgersen et al., 2014).
Exhumation rates during latest Mesoproterozoic-early Neoproterozoic normal faulting are unknown.However, exhumation rates from the early (approximately 945 Ma and 5-10 km depth) to mid-Neoproterozoic (approximately 825 Ma and 1-3.5 km depth) were probably in the range of approximately 10-75 m Myr −1 , i.e., comparable to what is expected from average continental erosion rates (10-100 m Myr −1 ; Schaller et al., 2002;Eppes and Keanini, 2017, their Fig. 5).This therefore suggests a period of tectonic quiescence between opening of the Asgard Sea (first two, quartz/calciterich faulting events) and the onset of Iapetus rifting (final, smectite-rich faulting event).

Evolution of temperature conditions in Caledonian rocks
We described five types of cohesive cataclastic fault rocks in NW Finnmark based on mineralogical and textural descriptions.First, epidote-and chlorite-rich, stilpnomelanebearing cataclasite (Fig. 3e and h-j) formed by faulting of Caledonian mafic schists and gneisses (amphibolites; Fig. 3d and e; Ramsay et al., 1979Ramsay et al., , 1985;;Gayer et al., 1985) and is consistently truncated by and incorporated into the other four types of cataclasites (Fig. 3e and h-j), suggesting that epidote-chlorite-rich cataclasites correspond to the earliest stage of brittle faulting recorded by Caledonian rocks.The epidote plus chlorite plus stilpnomelane plus/minus biotite mineral assemblages present both in the epidote-chloriterich cataclasites and adjacent host rocks, where biotite is almost completely recrystallized into chlorite (Fig. 3d), indicate lower greenschist-facies conditions during this faulting event, which constrain the minimum temperature during faulting to approximately 300 • C (i.e., 10 km depth).Further, rounded clasts of stilpnomelane in epidote-rich cataclastic veins onshore Magerøya (Fig. 3h and k) suggest faulting temperatures at prehnite-pumpellyite-to lower greenschist-facies conditions comprised between 300 and 470 • C (Miyano and Klein, 1989), i.e., a depth range of 10-16 km, which is consistent with pseudosection thermobarometry and U-Pb ages on titanite constraining retrograde Caledonian shearing < 550 • C (i.e., < 18 km depth) in the Kalak Nappe Complex in northern Troms to 440-420 Ma (Silurian; Gasser et al., 2015).
The second type of cataclasite corresponds to very finegrained, quartz-rich cataclasite and veins of recrystallized quartz that commonly truncate and incorporate clasts of (type 1) epidote-and chlorite-rich cataclastic veins (Fig. 3e  and h).Since quartz dissolution only occurs at temperatures > 90 • C (Worley and Tester, 1995) and deforms plastically at temperature > 300 • C (Tullis and Yund, 1977;Scholz, 1988;Hirth and Tullis, 1989), we argue that the quartz-rich cataclasite and associated quartz veins were formed during a discrete, second faulting event at depths between 3 and 10 km.The transition from early, deep (10-16 km), epidote-chloriterich faulting to subsequent, shallower (3-10 km) quartz-rich cataclasis indicates that Caledonian rocks were partly exhumed between the two faulting events.
The third type of cataclasite is made up of widespread calcite both as clasts, cement and precipitation (Fig. 3e  and i).This type of cataclasite cross-cuts epidote-chlorite-(type 1) and quartz-rich cataclasites (type 2), hence suggesting calcite-rich cataclasite (type 3) formed during a younger (reactivation) faulting event (Fig. 3e and i).Since calcite crystals of the cataclasite display characteristic twinning types I and II (Fig. 3i), we inferred a temperature range of 150-200 • C (Ferrill, 1991;Burkhard, 1993) and a depth of 5-7 km during this tentative, third faulting event.
A fourth type of cataclasite is present along fault segments of the LVF and TKFZ, showing pervasive laumontite clasts and precipitation (Fig. 3m), and consistently truncates greenschist-facies cataclasites (types 1, 2 and 3).Laumontite crystals commonly are undeformed and appear as elongated crystals with their long axis perpendicular to fracture boundaries (Fig. 3m).These observations suggest that laumontite formed as late growth along opening extensional cracks or in tension veins, most likely at a later stage of faulting than minerals in the greenschist-facies cataclasites.Laumontite crystals themselves commonly appear mildly cataclased (Fig. 3e, l and m), thus indicating that faulting persisted after the growth of laumontite.The temperature stability range of laumontite is 50-230 • C (Jové and Hacker, 1997), which suggests that syn-post-laumontite faulting occurred at a depth range of approximately 2-8 km, i.e., probably shallower than the faulting event associated with epidote-chlorite-(type 1), quartz-(type 2) and calcite-rich (type 3) cataclasites.
The fifth type of cataclasite is composed of iron oxide and clay minerals that cross-cut all other cataclasite types and vein minerals (Fig. 3h, l and n).XRF analyses of related non-cohesive fault rock show a consistent dominance of authigenic smectite with subsidiary mixed-layer chloritesmectite and minor illite (Table 1 and Appendix A), suggesting that the dominant clay mineral observed in the fifth type of cohesive cataclastic fault rocks is smectite.Based on the preservation of abundant authigenic smectite (Table 1 and Appendix A) and on the irreversibility of the smectiteillite diagenetic transformation (Eberl et al., 1993), we propose that the ultimate (fifth) faulting event(s) in Caledonian rocks in NW Finnmark occurred at temperatures < 105 • C (Morley et al., 2018), i.e., depth < 3.5 km, and that Caledonian rocks have remained at such shallow depth through Mesozoic-Cenozoic times.
Locally, XRF analysis of cohesive fault rocks along WNW-ESE-striking fault segments of the TKFZ in western Magerøya (sample 10; Fig. 1) shows almost pure authigenic smectite (Table 1 and Appendix A).This constrains temperature during faulting to a minimum of 35-65 • C (approximately 1-2 km depth) at which small amounts of illite may form (Eberl et al., 1993;Huang et al., 1993;Morley et al., 2018).Shallow faulting is further supported by the presence of mixed-layer chlorite-smectite clays (in samples 1, 2, 6 and 8; Table 1 and Appendix A), which suggests that smectite (and mixed-layer chlorite-smectite) authigenic clays formed by retrograde diagenesis (i.e., exhumation) of crushed chlorite during faulting (Warr and Cox, 2001;Nieto et al., 2005;Haines and van der Pluijm, 2012).Thus, we argue that Caledonian rocks along the LVF and TKFZ experienced another (late-stage) phase of uplift/faulting and exhumation from zeolite-facies conditions (2-8 km) to diagenetic conditions (1-3.5 km).Alternatively, chlorite-smectite in non-cohesive fault rocks along fault segments of the LVF (e.g., samples 1 and 2; Fig. 1, Table 1 and Appendix A) formed during the shallowest phase of smectite-illite clay mineral reaction, while interlayered illite-smectite (sample 8; Table 1 and Appendix A) formed during deeper phases of this reaction due to higher, normal faulting-related burial in the hanging wall of the LVF (Whitney and Northrop, 1988).However, more samples are needed in the hanging wall of the LVF to verify this hypothesis (only sample 8; Fig. 1).

Late Paleozoic inversion of brittle-ductile Caledonian thrusts
The coarse fraction of the Talvik fault (sample 5; Fig. 1), a south-verging Caledonian thrust in rocks of the Kalak Nappe Complex in Altafjorden, yielded a mid-/late Silurian age (427.3 ± 8.4 Ma) suggesting that brittle faulting along this fault initiated during the latest stage (Scandian) of the Caledonian Orogeny (Table 2, Figs. 4 and 6).Movement along the Talvik fault started with top-south ductile thrusting as shown by quartz sigma clasts and shear bands in mylonitic foliation, likely at a depth > 10 km (Tullis and Yund, 1977;Scholz, 1988;Hirth and Tullis, 1989), and continued with down-north brittle normal dip-slip faulting truncating ductile fabrics (Fig. 2e).The earliest indications of late-post-Caledonian normal faulting in northern Norway are Early Devonian ages obtained for inverted shear zones in Vesterålen (Steltenpohl et al., 2011).This suggests that the mid-late Silurian faulting event recorded along the Talvik fault (crystallization age of coarse fraction) might represent a phase of top-south Caledonian brittle thrusting, rather than late-post-Caledonian normal faulting (Fig. 2e).This is consistent with thermobarometry and U-Pb ages constraining Caledonian retrograde shearing at temperature < 550 • C to the Silurian at approximately 440-420 Ma (Gasser et al., 2015).Exhumation of the Talvik fault to brittle depth < 10 km in the mid-late Silurian was most likely due to combined thrusting and erosion.Alternatively, the obtained Silurian age may reflect input from an inherited illite/muscovite component as shown by a small illite peak with epizonal KI (< 0.25) in the coarse fraction of this sample (Appendix A), suggesting that brittle faulting initiated after Silurian times.Top-south Silurian brittle thrusting along the Talvik fault was followed by successive early Carboniferous (Tournaisian), 353.7 ± 4.1 Ma and early Permian, 282.1 ± 6.3 Ma faulting events obtained from the intermediate and finest gouge fractions, respectively (Table 2,Figs. 4 and 6), and interpreted as synkinematic crystallization ages.These events likely reflect post-Caledonian, down-north reactivation as a normal fault during the collapse of the Caledonides.Extensional reactivation is supported by normal dip-slip slickensides along the Talvik fault truncating the initial ductile fabrics (Fig. 2e).Further support appears from the early Permian inversion of an analog Caledonian thrust in the Repparfjord-Komagfjord tectonic window, the Kvenklubben fault (Torgersen et al., 2014), and from offshore seismic studies on the Finnmark Platform, where a major Caledonian thrust, the Sørøya-Ingøya shear zone, was inverted in the Middle to Late Devonian-early Carboniferous (Fig. 1; Koehl et al., 2018a).(Torgersen et al., 2014).In addition, the Laksvatn fault in western Troms (Fig. 1) yielded two Late Devonian ages (Davids et al., 2013).This down-NW normal fault is interpreted as a major, inverted Caledonian thrust possibly merging with the southwestern continuation of the LVF (Koehl et al., 2018b), thus suggesting that post-Caledonian, normal brittle faulting along the LVF initiated in the Late Devonian.Further support of Devonian faulting is found offshore, where potential Middle-Late Devonian sedimentary rocks were deposited along inverted Caledonian thrusts on the Finnmark Platform, and where the offshore segments of the LVF on the western Finnmark Platform east (Koehl et al., 2018a) bound a major (half-) graben filled with Carboniferous (Bugge et al., 1995) and, conceivably, uppermost Devonian clastic sedimentary deposits (Roberts et al., 2011).The dated faults in the Porsanger Peninsula (sample 8; Fig. 2h), Talvik (sample 5; Fig. 2e) and Storekorsnes (sample 6; Fig. 2f) show down-north, normal dip-slip to obliqueslip movements, suggesting that they are all related to late Paleozoic, post-Caledonian extension.The long time spread of the obtained late Paleozoic, post-Caledonian, K-Ar syntectonic crystallization ages (Table 2, Figs. 4 and 6) suggests either a long-term progressive or two discrete faulting periods.From our results, we favor two discrete periods, one in the Late Devonian-early Carboniferous at approximately 375-325 Ma (based on one age in this study, four from Torgersen et al., 2014 and two from Davids et al., 2013) and one in the late Carboniferous-mid-Permian at approximately 315-265 Ma (11 ages from the present study and two from Torgersen et al., 2014).
The obtained earliest Permian (Asselian) synkinematic crystallization ages for three different fractions of the same cataclasite in the fault in the hanging wall of the LVF on the Porsanger Peninsula (sample 8, Figs. 1, 4 and 6, Table 2), verified within 2σ error range, may reflect a single faulting event.The short time span suggests that the fault was not reactivated later, since further faulting would have been recorded in the finest grain-size fraction.This conclusion is supported by offshore seismic data on the Finnmark Platform, showing that the thickness of Permian sedimen-tary rocks is constant across brittle normal faults, such as the LVF and Måsøy Fault Complex (Fig. 1), and that most brittle faults die out within the Carboniferous and lower part of the Permian sedimentary successions (Koehl et al., 2018a).However, offshore seismic data on the Finnmark Platform suggest that early-mid-Permian, K-Ar ages (Table 2, Figs. 4 and 6 and Torgersen et al., 2014) obtained onshore NW Finnmark may represent only minor tectonic adjustments rather than major faulting events (Koehl et al., 2018a).Thus, an alternative interpretation for the dominance of Permian ages is due to partial overprinting/resetting of authigenic illite from the main early (-late?)Carboniferous faulting period.
Of the five dated faults that yielded late Paleozoic, post-Caledonian ages (samples 5, 6, 8, 9 and 10; Table 2 and Fig. 4), only two of them included cohesive fault rock (samples 5 and 6; Fig. 1).For those that only comprised noncohesive gouge (samples 8, 9 and 10) with predominance of authigenic smectite clay mineral and subsidiary authigenic illite (Table 1 and Appendix A), it seems reasonable to conclude that faulting occurred at shallow depth between 1 and 3.5 km (Eberl et al., 1993;Morley et al., 2018) in the late Carboniferous-mid-Permian (Figs. 4 and 6).The other two faults that yielded late Paleozoic ages (samples 5 and 6; Fig. 1) and comprise both cohesive and non-cohesive fault rocks (Fig. 2e and f), likely formed at deeper crustal levels and higher p/T conditions.Non-cohesive fault rocks along the Talvik fault (sample 5), consisting of authigenic smectite with minor illite, yielded early Carboniferous (intermediate fraction) and early Permian (finest fraction) ages (Table 2), while cohesive fault rocks along this fault are characterized by both quartz-and calcite-rich cataclasites (types 2 and 3).These data, backed by cross-cutting relationships between quartz-, calcite-and clay-rich cohesive fault rocks (types 2, 3 and 5; Fig. 3e), suggest that quartz-and calcite-rich cataclasites (types 2 and 3) along the Talvik fault formed in the Late Devonian (?) -early Carboniferous at a depth range of approximately 3-10 km (Scholz, 1988;Hirth and Tullis, 1989;Ferrill, 1991;Burkhard, 1993;Worley and Tester, 1995).Later on, these cataclasites were overprinted by noncohesive, smectite-rich, cohesive (type 5) and non-cohesive fault rocks generated at a depth of 1-3.5 km in the early Permian (Fig. 6; Eberl et al., 1993;Morley et al., 2018).These data are consistent with a partial exhumation of the Talvik fault and nearby Caledonian host rocks from the early Carboniferous to early Permian, with the average exhumation rate along this fault varying from < 185 (Silurian-early Carboniferous) to < 125 m Myr −1 (early Carboniferous-early Permian).
Of importance in restoring the exhumation history of NW Finnmark is that authigenic smectite is particularly dominant in non-cohesive fault rocks in the footwall and along fault segments of the LVF, although generally associated with large amounts of chlorite along fault segments of the LVF (e.g., Talvik fault and Sørkjosen faults; see samples 1, 2 and 5 in Table 1 and Appendix A), whereas interlayered illite-smectite dominates in non-cohesive fault rocks in the hanging wall of the LVF, e.g., sample 8 (Fig. 1, Table 1, and Appendix A).A plausible interpretation is that kilometer-scale, post-Caledonian down throw to the northwest along the LVF (partly) enhanced the exhumation of brittle faults and Caledonian rocks in the footwall, while hanging wall segments of the LVF remained at deeper levels, producing interlayered illite-smectite (8; Table 1, and Appendix A), a deeper endmember product of the smectite-illite reaction (Whitney and Northrop, 1988).This conclusion may be partly falsified by the illite-smectite-rich composition of fault gouges along the Porsavannet and Markopp faults in the footwall of the LVF (Torgersen et al., 2014).However, these faults yielded significantly older ages (respectively, mid-Neoproterozoic-mid-Paleozoic and early Carboniferous) than the late Carboniferous-early/mid-Permian ages of smectite/chlorite-smectite-rich fault rocks from our study (Table 2).It is therefore possible that the authigenic, interlayered illite-smectite in fault rocks along the Porsavannet and Markopp faults reflects earlier, deeper faulting periods.
Furthermore, XRF analyses of fault gouge from a segment of the TKFZ in western Magerøya (sample 10; Figs. 1  and 2j), yielding late Carboniferous to mid-Permian K-Ar ages (Table 2,Figs. 4 and 6), show almost pure authigenic smectite (Table 1 and Appendix A).This suggests that Permian faulting occurred at very low temperatures of 35-65 • C (Eberl et a., 1993;Huang et al., 1993;Morley et al., 2018), i.e., depth of approximately 1-2 km, and remained at shallow crustal levels until the present, thus preventing transformation of smectite to illite through diagenesis.Late Paleozoic exhumation and shallow faulting in NW Finnmark and along other portions of the Barents Sea margin, e.g., Lofoten-Vesterålen, are supported by apatite fission-track data, indicating that exposed rocks in northern Norway have remained at relatively low temperature < 120 • C, i.e., depth < 4 km, since mid-Permian times (Hendriks et al., 2007).
Considering K-Ar synkinematic crystallization ages obtained on non-cohesive fault rocks, and mineral assemblages and cross-cutting relationships of the five types of cohesive fault rocks, we estimate exhumation rates from the Silurian (approximately 425 Ma and 10-16 km depth) to Late Devonian (approximately 375 Ma and minimum 5-10 km depth) to be < 220 m Myr −1 , i.e., analogous to the estimate along the Talvik fault (< 185 m Myr −1 ).Similarly, exhumation rates through the Late Devonian (approximately 375 Ma and minimum 5-10 km depth) -early Carboniferous (approximately 325 Ma and 2-8 km depth) faulting period were < 160 m, and < 115 m Myr −1 through the ultimate faulting period from the end of the early Carboniferous (approximately 325 Ma and maximum 2-8 km depth) to the mid-Permian (approximately 265 Ma and 1-3.5 km depth), i.e., similar to exhumation rates obtained along the Talvik fault (< 125 m Myr −1 ).The decreasing exhumation rate from Silurian to mid-Permian times might indicate progressively milder faulting activity along the margin, with exhumation rates in the late Carboniferousmid-Permian being comparable/slightly higher than average continental erosion rates thought to be on the order of 10-100 m Myr −1 (Schaller et al., 2002;Eppes and Keanini, 2017, their Fig. 5).Thus, we propose that exhumation in the Silurian-early Carboniferous was driven by a combination of continental erosion, thrusting and, later on, normal faulting, while exhumation in the late Carboniferous-mid-Permian was mostly due to continental erosion with only a limited contribution of normal faulting.
Alternatively, the predominance of Permian faulting ages obtained onshore NW Finnmark may be attributed to a period of extensive weathering in the (late Carboniferous-?) earlymid-Permian (Table 2, Figs. 4 and 6), possibly reflected by highly weathered host rocks and brittle fault surfaces showing no kinematic indicators onshore Magerøya (samples 9 and 10; Figs. 1, and 2i and j).Although this weathering may be related to much younger processes (Olesen et al., 2012(Olesen et al., , 2013)), Carboniferous-Permian, (sub)tropical climate conditions prevailed in Baltica (Stemmerik, 2000;Larssen et al., 2002;Samuelsberg et al., 2003) and hence may have initiated weathering of exposed, uplifted footwall blocks along major faults like the LVF, and Måsøy and Troms-Finnmark fault complexes offshore (Fig. 1).This is supported by widespread www.solid-earth.net/9/923/2018/Solid Earth, 9, 923-951, 2018 exhumation and erosional truncation of pre-Pennsylvanian rocks in the footwall of the Troms-Finnmark Fault Complex, on the Finnmark Platform (Koehl et al., 2018a), linked to a mid-Carboniferous phase of eustatic sea-level fall (Saunders and Ramsbottom, 1986).This exhumation/weathering event is also consistent with the Early Mesozoic, minimum age estimate of weathering of basement rocks along the Norwegian continental shelf (Olesen et al., 2012(Olesen et al., , 2013)).Although the dated faults were still buried to depth > 1 km in the mid-Permian, as shown by the presence of (minor) authigenic illite (Eberl et al., 1993) used for K-Ar age dating of the faults, field studies in onshore tunnels in Norway show that weathering processes related to percolation of acidic water may penetrate the bedrock > 200 m along fault surfaces (Olesen et al., 2012(Olesen et al., , 2013)), thus making this alternative explanation possible, though unlikely.Another obstacle to this interpretation is the lack of a major erosional unconformity/truncation in upper Carboniferous-lower/mid-Permian sedimentary rocks on the Finnmark Platform offshore (Larssen et al., 2002;Samuelsberg et al., 2003;Koehl et al., 2018a).

Mesozoic faulting
A reliable Mesozoic K-Ar syntectonic crystallization age was obtained only for the finest fraction of the Snøfjorden-Slatten fault (sample 7; Fig. 1 and Table 2), yielding a Hettangian age.All the other faults yielding Mesozoic ages (all three fractions in samples 1 and 2, coarse fraction of sample 6, coarse and intermediate fractions of sample 7 and intermediate fraction of sample 9) comprise subsidiary K-feldspar (Table 1 and Appendix A), which provided additional potassium and thus yielded younger ages than the actual age of faulting (Table 2; red ages in Fig. 4).Thus, we disregard these ages because of their high uncertainty.Considering the scarcity of Mesozoic-Cenozoic ages, we argue that NW Finnmark, as well as adjacent offshore areas of the Finnmark Platform (Koehl et al., 2018a) were tectonically quiet after late Paleozoic (Devonian-mid-Permian) extension and were only subjected to minor, local extensional faulting events, e.g., in the earliest Jurassic (Hettangian) for the Snøfjorden-Slatten fault (Figs. 1, 4 and 6, Table 2) and Early Cretaceous for the Kvenklubben fault (Torgersen et al., 2014).

Regional implications
An implication of the latest Mesoproterozoic-Neoproterozoic K-Ar ages obtained for ENE-WSW-striking Altafjorden faults 1 and 2 in the Alta-Kvaenangen tectonic window is that they partly support the interpretation of Koehl et al. (2018b), suggesting that ENE-WSW-striking faults represent inherited Precambrian fault fabrics.However, the inferred normal sense of shear and latest Mesoproterozoicmid-Neoproterozoic K-Ar faulting ages obtained for Altafjorden faults 1 and 2 (Fig. 2a and b) suggest that these faults formed as extensional normal faults rather than conjugate strike-slip faults to WNW-ESE-striking faults like the TKFZ as suggested by Koehl et al. (2018b).Instead, latest Mesoproterozoic-mid-Neoproterozoic brittle faults might have provided preferentially oriented weakness zones for the formation of subparallel, subsequent and adjacent Caledonian thrusts (e.g., Talvik fault) and post-Caledonian normal faults (e.g., LVF; Fig. 1).Nevertheless, conjugate strike-slip faults have been reported from NW Finnmark (Roberts, 1971;Worthing, 1984), but these display subvertical geometries and significant lateral displacement, and may have formed during (E-W-to) ENE-WSW-directed, Timanian contraction in the late Neoproterozoic, e.g., TKFZ (Siedlecka et al., 2004;Herrevold et al., 2009) and Akkarfjord fault (Roberts, 1971;Koehl et al., 2018b).Analogous studies of post-Caledonian brittle faults in western Troms show that post-Caledonian extensional faulting initiated at depth > 10 km at greenschist-facies conditions and continued under pumpellyite-prehnite-facies conditions at depth < 8.5 km, thus supporting a gradual exhumation of the margin (Indrevaer et al., 2013(Indrevaer et al., , 2014)).More detailed mineralogic-textural analysis of clay-rich non-cohesive fault rocks of the Vannareid-Burøysund, Sifjord and Laksvatn faults revealed dominance of smectite and chlorite clay minerals (Davids et al., 2013), suggesting that brittle faults in western Troms were exhumed to low temperature conditions (35-105 • C; Eberl et al., 1993;Morley et al., 2018) and shallow depths (1-3.5 km) comparable the LVF and TKFZ in NW Finnmark.Furthermore, fault gouge along the SSE-dipping Vannareid-Burøysund and Sifjord faults yielded similar early Carboniferous (intermediate fractions) and early Permian K-Ar ages (finest fractions; Davids et al., 2013) compatible with the proposed Late Devonian-early Carboniferous and late Carboniferous-mid-Permian stages of post-Caledonian brittle faulting in NW Finnmark (Table 2, Figs. 4, 6 and Torgersen et al., 2014).
A major contrast in K-Ar ages in western Troms and NW Finnmark is occurrence of latest Mesoproterozoic-mid-Neoproterozoic ages for gouges of the southeasternmost normal faults within basement rocks of the Alta-Kvaenangen tectonic window in NW Finnmark (Figs. 4 and 5), while analogous faults in Archean-Paleoproterozoic rocks of the West Troms Basement Complex (Zwaan, 1995;Bergh et al., 2010) yielded Carboniferous-Permian ages (Davids et al., 2013).Another mild contrast is the occurrence of slightly younger, late Permian-Early Triassic, K-Ar faulting ages for brittle faults in western Troms (Davids et al., 2018), suggesting that extension migrated westwards after the late Carboniferousmid-Permian and persisted until the Early Triassic in coastal areas of western Troms.Westwards younging of K-Ar faulting ages is further supported by Mesozoic ages obtained for three faults in western Lofoten (Davids et al., 2013).Nonetheless, widespread Late Devonian-early Carboniferous and late Carboniferous-mid-Permian ages in NW Finnmark, western Troms and Lofoten-Vesterålen suggest that the main episode of extension and exhumation along the mar-gin occurred in the late Paleozoic and was probably related to the collapse of the Caledonides (Davids et al., 2013(Davids et al., , 2018;;Torgersen et al., 2014;Koehl et al., 2018a).Apatite fissiontrack data in western Troms and Lofoten-Vesterålen also indicate a period of rapid cooling (1-2 • C Myr −1 ) in the late Paleozoic, possibly due to combined extensive normal faulting and erosion, followed by a period of relatively slow cooling (< 0.2 • C Myr −1 ) in Mesozoic times, likely suggesting a tectonically quiet time period (Davids et al., 2013).

Conclusions
Three faulting events occurred in the latest Mesoproterozoicmid-Neoproterozoic (approximately 1050-810 Ma), including (i) latest Mesoproterozoic faulting (approximately 1050 Ma.) and (ii) an early Neoproterozoic faulting event (approximately 945 Ma) with quartz-rich and calcitecemented cataclasites formed at a depth of approximately 5-10 km, possibly reflecting the formation of the NW Baltoscandian basins during the opening of the Asgard Sea, and (iii) a shallow (depth 1-3.5 km), mid-Neoproterozoic faulting episode (approximately 825-810 Ma) with abundant authigenic smectite, related to the opening of the Iapetus Ocean-AEgir Sea and breakup of Rodinia between 825 and 740 Ma.
Exhumation rates estimates from 945 to 825 Ma were on the order of 10-75 m Myr −1 , thus indicating that continental erosion alone may account for early-mid-Neoproterozoic exhumation and that tectonic quiescence prevailed between the opening of the Asgard Sea and the opening of the Iapetus Ocean-AEgir Sea.
The preservation of abundant authigenic smectite in cohesive and non-cohesive fault rocks suggests that Paleoproterozoic basement rocks were exhumed to and remained at shallow crustal levels (< 3.5 km depth) since the mid-Neoproterozoic (approximately 825 Ma), and were not reactivated after mid-Neoproterozoic times despite being oriented parallel to major Caledonian thrusts and post-Caledonian normal faults.
Kilometer-scale, down-NW normal faulting and footwall uplift along the Langfjorden-Vargsundet fault may be responsible for local variation of dominant, authigenic clay minerals in type 5 fault rocks (1-3.5 km depth), producing deeper, interlayered illite-smectite in the hanging wall and shallower, smectite and mixed-layer chlorite-smectite in the footwall.
Data availability.Structural field measurements, analyzed thin sections and K-Ar geochronological data may be obtained from the corresponding author.
-ESE-striking brittle faults and fracture systems are widespread within units of the Magerøy Nappe on the Porsanger Peninsula and on the island of Magerøya (Fig.1;Koehl et al., 2018b).Sample 8 is from fault gouge found along an anomalously low-angle NNE-dipping fault on the Porsanger Peninsula (Figs.1 and 2h).The fault crosscuts garnet-mica gneisses of the Kalak Nappe Complex, ad- Figure 2.

Figure 2 .
Figure 2. Outcrop photographs of the dated faults.Each photograph is accompanied by the location of the dated samples (yellow squares; see Fig. 1 for sample location), structural measurements presented in white boxes and in Schmidt stereonets (left-hand-side stereonets show fracture surfaces as great circles and right-hand-side stereonets display slickenside lineations as pole-to-fault surfaces indicating the movement of the hanging wall), and potential kinematic indicators where available.(a) NNW-dipping Altafjorden fault 1 (sample 3) crosscutting Precambrian rocks of the Alta-Kvaenangen tectonic window along the western shore of Altafjorden.A potentially drag-folded mafic bed is shown in blue.Modified after Koehl et al. (2018b); (b) NNW-dipping Altafjorden fault 2 (sample 4) within the Alta-Kvaenangen tectonic window.The country rock displays preserved bedding surfaces (in dotted yellow); (c) NNW-dipping fault in the footwall of the Sørkjosen fault (Sørkjosen 1; sample 1).A potentially offset mafic bed is shown in green and the bedrock fabric in dotted orange.The lower left frame is a zoomed-in image of the fault core displaying the location of the dated sample; (d) NNW-dipping fault in the footwall of the Sørkjosen fault (Sørkjosen 2; sample 2).Dotted yellow lines show normal offsets of mafic beds across brittle faults; (e) north-dipping Talvik fault (sample 5) cross-cutting rocks of the Kalak Nappe Complex along the western shore of Altafjorden.Kinematic indicators include ductile fabrics made of microshears (lower left frame), sigma clasts (middle left frame) and slickenside lineations (upper right frame).The bedrock fabric is shown in green.Modified after Koehl et al. (2018b); (f) NNE-SSW-striking brittle faults cross-cutting rocks of the Kalak Nappe Complex along the eastern shore of Altafjorden (sample 6); see normal offsets of geological markers (in green and blue).Modified after Koehl et al. (2018b); (g) SE-dipping fault segment of the Snøfjorden-Slatten fault (sample 7) in Snøfjorden, on the Porsanger Peninsula.Modified after Koehl et al. (2018b); (h) low-angle NNE-dipping fault (sample 8) cross-cutting rocks of the Kalak Nappe Complex on the Porsanger Peninsula; (i) steep, E-W-to WNW-ESE-striking faults (sample 9) within the gabbroic rocks of the Honningsvåg Igneous Complex; (j) steep, WNW-ESE-striking brittle faults near Gjesvaer (sample 10), in the western part of Magerøya.The faults cross-cut rocks of the Kalak Nappe Complex.

Figure 4 .
Figure 4. Structural map of the study area showing the obtained K-Ar ages for each sample in Ma, including from top to bottom finest, intermediate and coarsest grain-size fraction ages.Ages in red are considered erroneous (see main text).The map is modified after Indrevaer et al. (2013) and Koehl et al. (2018a).Legend and abbreviations as are in Fig. 1.

Figure 5 .
Figure 5. Graph displaying the obtained Precambrian ages for brittle faults in the Alta-Kvaenangen tectonic window.The vertical axis is time in Ma and the horizontal axis shows the dated sample numbers, whose locations are displayed in Figs.1 and 4. Columns to the right show extensional tectonic events related to the formation of the NW Baltoscandian basins during the opening of the Asgard Sea(Siedlecka et al., 2004;Nystuen et al., 2008;Cawood et al., 2010;Cawood and Pisarevsky, 2017) and to the earliest phase of rifting of the Iapetus Ocean-AEgir Sea(Li et al., 1999(Li et al., , 2008;;Torsvik and Rehnström, 2001;Hartz and Torsvik, 2002), and associated geological periods and eras.

Figure 6 .
Figure 6.Graph showing the Phanerozoic K-Ar ages and associated error obtained for fault gouge samples in NW Finnmark.The vertical axis represents time in Ma and the horizontal axis shows the dated sample numbers, whose locations are displayed in Figs. 1and 4. Columns to the right of the graph display major contractional (based onRamberg et al., 2008;Vetti 2008;Corfu et al., 2014) and extensional events(Faleide et al., 1993;Steltenpohl et al., 2011;Koehl et al., 2018a) that affected north Norway in late Paleozoic-Mesozoic times, and associated geological time periods.
Figure 6.Graph showing the Phanerozoic K-Ar ages and associated error obtained for fault gouge samples in NW Finnmark.The vertical axis represents time in Ma and the horizontal axis shows the dated sample numbers, whose locations are displayed in Figs. 1and 4. Columns to the right of the graph display major contractional (based onRamberg et al., 2008;Vetti 2008;Corfu et al., 2014) and extensional events(Faleide et al., 1993;Steltenpohl et al., 2011;Koehl et al., 2018a) that affected north Norway in late Paleozoic-Mesozoic times, and associated geological time periods. ).

Table 1 .
Mineral composition of dated fault gouges."Presumed traces" are based on inclined spectra (Appendix A) showing peaks at 27.3-27.4KeV for K-feldspar, and on the presence of minor illite in smectite, enabling K-Ar dating, e.g., samples 9 and 10.Abbreviations are as follows: Chl is chlorite; Ilt is illite; Kao is kaolinite; Kfs is K-feldspar; Lmt is laumontite; Pl is plagioclase; Qz is quartz; Sme is smectite.The question mark indicates uncertainty around the nature of the mineral here interpreted as laumontite.
respectively(Table 2, and Figs. 4 and 5).The intermediate fraction yielded a slightly younger age compared with the finest fraction; taking the errors into account, both ages do not differ significantly, and we interpret them both as synkinematic crystallization along an active normal fault.Never-**: major component, *: minor component, O: trace, -: presumed trace; all smectites are of R = 0 type, indicating maximum 10 % of illite content.

Table 2 .
K-Ar ages from synkinematic illite in the sampled fault gouges.