Smøla island, located in central Norway, is set along the mid-Norwegian passive margin (Fig. 1a), a region shaped by a prolonged and complex geological history (Peron-Pinvidic and Osmundsen, 2020). The island (Fig. 1b) comprises 445.7±3.8 Ma plutonic magmatic-arc rocks (Tucker et al., 2004; Slagstad et al., 2011) and oceanic crustal units, all of which were emplaced and accreted just prior to the Silurian–Devonian (∼430 to 400 Ma) Scandian Phase of the Caledonian Orogeny (Tucker et al., 2004; Gee et al., 2008; Slagstad and Kirkland, 2018; Corfu et al., 2014). Caledonian shortening led to nappe and thrust fault development, with the transposition of the Smøla rocks along major detachments over both the allochthonous continental Neoproterozoic Baltoscandian nappes and the underlying autochthonous Palaeo- to Mesoproterozoic Baltican crystalline basement of the Western Gneiss Region (WGR) (Tucker et al., 2004; Corfu et al., 2014; Fossen, 1992) (Fig. 1c).

During the Early Devonian collapse of the by then overthickened orogen resulted in significant crustal extension by both the reactivation of favourably oriented contractional structures (Fossen et al., 2017; Fossen, 2010) and the development of new crustal-scale extensional shear zones and regional exhumation (Fig. 1) (Osmundsen et al., 2006). Devonian extension also caused the Smøla rocks to be translated in response to both strike-slip and down-faulting across the Møre–Trøndelag Fault Complex (MTFC), a regional fault system with complex kinematics that was repeatedly reactivated through time (Redfield et al., 2004; Watts et al., 2023; Grønlie and Roberts, 1989; Grønlie et al., 1994). Smøla experienced limited metamorphic overprint (on average sub-greenschist facies conditions) and only localised ductile deformation (Tucker et al., 2004). Associated with the Devonian tectonic inversion, extensional (possible pull-apart) basins formed in response to brittle extensional faulting, such as the MTFC-related Edøyfjorden basin proximal to Smøla (Fig. 1b) (Fossen et al., 2017; Osmundsen et al., 2006; Bøe and Bjerkli, 1989; Bøe et al., 1989).

Post-Caledonian extension and rifting continued to the Late Cretaceous/Early Palaeocene, eventually leading to the Greenland–Eurasia plate break-up at ∼54 Ma and opening of the North Atlantic, forming the Norwegian passive margin (Peron-Pinvidic and Osmundsen, 2018; Bunkholt et al., 2022). This prolonged process involved multiple tectonic phases: rift initiation during the Carboniferous, Permo-Triassic crustal stretching and rifting, Late Triassic to Jurassic crustal thinning, and renewed Mid-Jurassic to Early Cretaceous rifting, with the extension of the margin during the Late Cretaceous to Palaeocene and early seafloor formation and spreading during the Palaeocene–Eocene (Mosar et al., 2002; Faleide et al., 2008; Peron-Pinvidic and Osmundsen, 2018; Gernigon et al., 2020).

These events produced discrete generations of basins (such as the Froan and Møre basins) along the newly forming Norwegian shelf, which progressively infilled with synrift sediments (Fig. 1a and c) (Faleide et al., 2008; Zastrozhnov et al., 2020). In places, structural highs of Caledonian-aged and underlying Baltica crystalline basement, such as the Frøya High to the NW of Smøla formed through rift-related exhumation between these basins, leading to a remarkable host and graben series (Muñoz-Barrera et al., 2020). Bounding the structural highs and the different basins are crustal-scale extensional faults (Fig. 1a and c) (Skilbrei et al., 2002; Mosar et al., 2002). Overall, this protracted rifting and extension history resulted in a highly attenuated offshore domain reactivating and overprinting the existing Caledonian deformation features (Fig. 1a and c) (Zastrozhnov et al., 2020; Mosar et al., 2002).

Smøla is composed of several Caledonian-aged bedrock types traditionally attributed to the Caledonian Upper Allochthon (Roberts and Gee, 1985) (Fig. 1b), including diorite (the most common lithology), gabbro, tonalite, alkali granite, quartz monzonite to monzogranite, quartz–feldspar and amphibolite–biotite gneiss, and limestone and volcanic rocks (andesitic to rhyolitic in composition in a faulted block on the SE portion of Smøla). Late felsic to mafic dykes crosscut all other rock types (Gautneb, 1988; Gautneb and Roberts, 1989; Fediuk and Siedlecki, 1977; Roberts, 1980; Bruton and Bockelie, 1979). In addition to the 445.7±3.8 Ma (1σ) U/Pb intrusive age of a dated granodiorite (Tucker et al., 2004), a granitic dyke was dated to 428±10 Ma (2σ) (Rb/Sr) (Gautneb, 1988), providing an Ordovician–Silurian age envelope for the plutonic rocks on the island. To the S and SE of Smøla, the Caledonian bedrock is unconformably overlain by the late Silurian–Mid-Devonian sedimentary cover of the Smøla group and later Jurassic–Cretaceous sedimentary successions (Bøe et al., 1989; Fediuk and Siedlecki, 1977; Bøe and Bjerkli, 1989). These include siliciclastic deposits of red-oxidised polymictic conglomerate, siltstone, and sandstone (Bøe et al., 1989) which fill in Devonian extensional basins (e.g. Edøyfjorden basin; Fig. 1b) along the MTFC (Bøe and Bjerkli, 1989).

Further to the S and SE of Smøla (and bounding the Devonian basins), the ∼ NE–SW-striking MTFC is the most proximal and dominant regional-scale structure to the island. This wide (10 to 50 km) structure extends >600 km from central Trøndelag into the northern part of the North Sea and branches into two sub-parallel strands, namely the Hitra–Snåsa Fault (HSF) to the NW (the most proximal to Smøla) and the Verran Fault (VF) to the SE (Fig. 1a–c) (Olsen et al., 2007; Watts et al., 2023; Watts, 2001). The structural evolution of the MTFC is complex, with a polyphase deformation history from the Devonian until the Cenozoic (e.g. Grønlie and Roberts, 1989; Seranne, 1992; Watts, 2001; Redfield et al., 2004; Tartaglia et al., 2023; Watts et al., 2023).

The geophysical and DTM data used in this study were acquired by the Geological Survey of Norway (NGU) (Nasuti et al., 2015), with details on survey parameters and data processing available in the Supplement. Our lineament picking and mapping followed White (2014) and Scheiber and Viola (2018), making use of both the magnetic geophysical and DTM imagery within a geographic information system (GIS) software platform. The description of the methodology used to identify and place the lineaments is also available in the Supplement.

The field-based structural data acquisition involved standard geological field methods, including the systematic mapping of representative outcrops and ground-truthing of possible exposed deformation zones identified in the lineament mapping. In total, 66 outcrop sites were studied, with data collected recording deformation feature types, fracture/vein mineral infill, fault slip information, host rock lithology, and crosscutting/genetic relationships between the various features. The four diamond drill holes used in this study (stored in Trondheim, Norway) were drilled in 2019 and were designed to target major structures and sample weathered basement rocks. Geological features in the drill cores were systematically logged downhole (for a total investigated length of 364.9 m of diamond drill core), recording lithology and rock alteration, deformation types, fracture/fault characteristics, mineral infill, and crosscutting relationships. These data were also used for the 3D modelling of intersected deformation zones in drill holes BH1 and BH2. The methodology used to measure structural data in the drill core followed both the methods of Holcombe (2013) and Blenkinsop et al. (2015). A more detailed description of the methodology is available in the Supplement.

Seven structurally controlled fault gouge and breccia samples were collected from fault and deformation zones in both drill core and outcrop and subsequently processed at the NGU laboratory in Trondheim, Norway. Initially, all samples were disintegrated and separated into <0.1, 0.1–0.4, 0.4–2, 2–6, and 6–10 µm size fractions. Each of these size fractions then underwent quantitative analysis for both potassium (K) and argon (Ar) using total digested ICP-OES (inductively coupled plasma–optical emission spectrometry) and an Isotopx NGX multicollector noble gas mass spectrometer system, respectively. X-ray diffraction (XRD) analysis was performed on each fraction to determine and quantify the mineralogical composition. A full description of both the XRD and K–Ar analysis methodology is available in the Supplement and in Viola et al. (2018).

The analysis of regional-scale features such as geophysical lineaments and topographic features (Fig. 2) formed the initial stage of the study. These features occur on the kilometre scale and, using both geophysical and DTM imagery, can be traced over both off- and onshore areas. The >2000 mapped lineaments (available in the data repository in the Supplement) are linear and curvilinear to irregular features which either crosscut or terminate against one another (Fig. 2a and b). The lineaments are classified based on the length and hierarchy of the formation: first-order lineaments are potentially major regional structures; second-order lineaments, commonly >10 km in length, are typically major splays or secondary features off the first-order lineaments, while third-order lineaments, typically <10 km in length, constitute the remaining lineaments.

The first-, second-, and third-order lineaments are subdivided into four main orientation trends, based on azimuth frequency maxima, which are labelled (L1) N–S, (L2) E–W, (L3) NE–SW, and (L4) NW–SE (Fig. 2b) (hereinafter L1 to L4). Determining precise crosscutting relationships between these lineament trends is often challenging due to the imagery resolution and the difficulty of defining lineament end points. However, where possible, interpretive crosscutting relationships provide a relative timing of lineament formation. The age, crosscutting, and termination relationships of the specific lineament trends are summarised in Table 1 and selectively illustrated in Fig. 2b.

L1 is most common in the western part of Smøla and offshore in areas A and B in Fig. 2a. Its lineaments are typically second- to third-order and are crosscut by all other lineaments (Fig. 2bI). Local mutual crosscutting relationships between L1 and L4 suggest they are likely coeval/conjugates. L1, L2, and L3 lineaments have complex crosscutting relationships, with L1 being potentially horizontally offset (dextrally and sinistrally) with multiple possible reactivations in places (Fig. 2bI and bII).

The L2 lineaments are most common in central Smøla, particularly in areas B, D, and E, clustering on the NW side of the MTFC (Fig. 2a) and occurring as second-order lineaments and possible splays off the first-order MTFC (HSF) lineament (Fig. 2a). L2 is crosscut by L1 and L4 (Fig. 2bII) and also exhibits mutual crosscutting (potentially coeval) relationships with the L3 lineaments (which, in turn, typically splay off the E–W structures) (Fig. 2bI). Frequently, L1, L2, and L4 (third-order lineaments only) are similarly crosscut by the same lineament trends, indicating that these lineaments may have formed coevally (Fig. 2bII). The longer L2 second-order lineaments do, however, crosscut most other lineaments (Fig. 2bI), suggesting either a later formation or a longer period of activity/reactivation.

The L3 lineaments are pervasive across Smøla within areas A, D, E, B, and F (Fig. 2a). These lineaments are typically third-order lineaments and sub-parallel to the MTFC (HSF) trend. In the eastern part of Smøla (areas E and D), they cluster on and terminate against the N/NE sides of L2 (Fig. 2bI), possibly representing third-order Riedel shears or splays off the second-order L2. In the western part of Smøla, they form prominent elongated topographical valleys in the DTM and terminate less commonly against L2. These lineaments crosscut and dextrally offset both the L1 and L4 (less distinctly) and terminate against L2 (infrequently also L4) (Fig. 2bII). L3 lineaments are, however, in turn either locally crosscut by or are terminated up against by L1 and rarely L4 (mutual terminating relationship) (Fig. 2bII).

L4 is common to the north of the MTFC over most parts of Smøla and offshore (areas A, B, D, and E in Fig. 2a). L4 is similarly crosscut by many of the other lineament sets (Fig. 2b), although locally L4 crosscuts L2 and the L3 lineaments, as well as the magnetic grain (Fig. 2b), but this may be due to later reactivations. The L4 and L1 lineaments have rare mutual crosscutting relationships, although in general L1 tends to terminate up against L4.

Representing a later subset of L4, the WNW–ESE-striking portion of L4 is also present throughout the study area, although mostly NW of the MTFC and are either second- or third-order lineaments. This late L4 crosscuts all other lineaments, with only rare examples of L1 crosscutting it (Fig. 2bII). Similarly, L1 has a younger, possibly reactivated component. These younger L1 lineaments are most abundant in the areas D and B (more NNE–SSW-striking), and areas F and E (more NNW–SSE-striking) over Smøla and the WGR. Overall, these lineaments terminate against and crosscut all the other lineaments (Fig. 2bI and bII), suggesting an early formation and a possible late reactivation.

Field observations (Figs. 1b and 3) and four diamond drill holes (Figs. 1b and 4) offer a novel 3D perspective on Smøla's geology. Various minerals decorating fracture surfaces, fault rocks, and veins are systematically documented, showing coeval or crosscutting relationships and orientation trends (Figs. 3, 5, and 6). As such, assemblages of coeval infill minerals have been used to group the deformation features based on the relative timing (labelled D1 to D5). The coeval mineral assemblages, with corresponding consistent deformation styles and orientation trends, may represent distinct mineralisation events associated with discrete deformation episodes (e.g. Drake et al., 2009; Viola et al., 2009).

We now systematically describe the progressive deformation and mineralisation types, starting from the apparent earliest to the latest features. As part of the structural characterisation, we have included both the field and drill hole structural data (Fig. 3). However, the drill hole data were localised in specific parts of the island and biased towards including shallowly dipping features (e.g. Terzaghi, 1965). Therefore, field data primarily inform the structural descriptions, as they illustrate typical geometries and kinematics across Smøla.

The four oriented drill holes (Fig. 4, with locations shown in Fig. 1b) provide a 3D view of Smøla's geology. Using Leapfrog Works (Seequent, 2022) 3D modelling software, we could delineate the hanging wall and footwall boundaries of deformation zones using logged downhole intersections and orientation data. Implicit volumes were then generated to reproduce the actual thickness, orientation, and extent of zones through 3D space. The modelling of different zones in one volume provides insights into the continuity and extent of the deformation zones, as well as potential fracture connectivity. Furthermore, by incorporating both relative (Fig. 7) and absolute deformation ages, the modelling also includes a temporal dimension.

In this study, we have specifically used drill holes BH1 and BH2 for the modelling. These drill holes are close enough together to allow confident correlations between the drill holes. These drill holes intersected deformation zones, veined intervals, and shear zones/fault cores (Fig. 4). Considering the geological and deformation characteristics of each of these zones, particularly the fault core intersections (assumed to have the greatest strike continuity) and the associated damage zones, we have correlated at least four zones between the drill holes. The correlations and modelling also used the orientations of bounding deformation features to resolve the geometries of these zones. Examples of modelled zones which are correlated between BH1 and BH2 (labelled I–IV in Fig. 8a) are the following: I.

BH1 (37.46–37.81 m) and BH2 (12.94–23.59 m) intersect a S-dipping and E–W-striking zone characterised by early D1 epidote veins and hydraulic breccia (Ep 2) crosscut by D2 chloritic/sericitic fractures and breccia and late D5 quartz–calcite veins (Cal 1 and Cal 2).

The modelled zones crosscut each other, as assumed here with no offsets or lateral terminations. Their true thicknesses range from 1 to 6 m and exhibit some dip and thickness variation (as shown in the N–S and E–W sections in Fig. 8a). The zones display different orientation trends, with the D2-related zones moderately S to SE-dipping and D3-related zones shallowly W-dipping.

To explore Smøla's deformation features across scales, 3D models were created for selected intervals, such as D5 quartz–calcite veins in BH2 (∼10 m scale) or epidote–prehnite (D1) veins in BH1 (∼1 m scale) (Fig. 8b). These models, along with Fig. 8a (∼100 m scale), demonstrate the systematic to semi-systematic fracture arrays at various scales. Additionally, they illustrate differences in aperture widths and potential strike extents among structures associated with different deformation episodes; D2 structures, for example, typically have the widest apertures and longest strike extents, while D1 features exhibit the smallest apertures and the least lateral continuity on average but have the highest number of fractures per modelled volume.

Seven fault gouge samples (Table 2; Fig. 9) were collected from drill holes BH2 (SK2008, SK2012, and SK2015), BH4 (120714) (Fig. 4), and outcrops on Smøla island (SK1024_1, SK1029_1, and SK1033_1) to date potential authigenic and synkinematic potassium (K)-bearing clays. The dated samples compositions by X-ray diffraction (XRD) are shown in Fig. 10, with the K–Ar dating results in Table 3 and Fig. 11a subdivided by the grain size fraction for each sample (<0.1 to 6–10 µm). The orientations of the sampled structures (Fig. 11b) is provided in the text as dip inclination (0–90°) and dip direction (0–360°). None of the samples has associated kinematics, apart from sample SK1033_1. The sample descriptions are ordered by interpreted deformation episode, with the full sample characterisations available in the Supplement.

Based on our new multiscalar and multi-technique results, we propose an evolutionary tectonic history for Smøla (Fig. 12) and discuss its implications for the evolution of the passive margin.

The D1 epidote–prehnite mineralisation, although not directly dated in this study, has been shown in other areas of the Norwegian margin to have formed from syn- to post-Devonian times until ∼290 Ma (Seranne, 1992; Watts et al., 2023; Sherlock et al., 2004; Indrevær et al., 2014). The D1 brittle structures (Figs. 5a, b and 8b) rework primary host rock features and are systematically offset by the D2 to D5 structures, indicating that epidote and prehnite mineralised prior to all other deformation episodes (Fig. 7). The D1 tensile veins or sinistral strike-slip features, striking ENE–WSW (sub-vertical to steep SSE dips and shallow-plunging slip lineations) (Fig. 3), likely formed during post-Caledonian brittle sinistral transtension on the MTFC (Seranne, 1992; Watts et al., 2023) and WNW–ESE extension (represented in Fig. 12).

Later, the NW–SE-striking features (moderately to shallowly NE-dipping), as well as the ∼ NNW–SSE to NNE–SSW-striking features (steeply ∼ E-dipping), being either normal dip-slip features or both dextral and sinistral (conjugate) features, may have formed because of local N–S to NNE–SSW shortening. This shortening direction is also implied by the Mid–Late Devonian development of E–W-trending folds in the Edøyfjorden basin Devonian rocks (Bøe et al., 1989; Fossen, 2010) (Fig. 12). From relative timing and structural data correlations, the lineaments L1, L3, and L4 may have formed during D1, with later reactivations. Overall, despite being the least strike-continuous features (as shown in the 3D model in Fig. 8b), D1 is pervasive across Smøla, the Devonian-aged Edøyfjorden basin, and the WGR to the SE of the MTFC (Fig. 2a and bI, bII).

The D2 sericite–chlorite–calcite structures (Figs. 5c and 6d, e) are dated to ∼204–196 Ma (finest fractions for samples SK2012, SK2015, and SK1029_1 in Fig. 11). However, microstructural and field evidence (Fig. 6d and e) and the plateau in the age spectra graph for the two coarsest fractions for SK2012 and SK2015 (Fig. 11a) suggest a possibility of an earlier inherited deformation episode at ∼300 Ma. Early faulting-related sericitisation of host rock feldspar (e.g. Wibberley, 1999) could explain the coarsest fraction illite/muscovite content and this inherited age. As such, two deformation (sub-)episodes may be associated with the sericite–chlorite–calcite mineral assemblage and thus tentatively to D2. Considering these two possible deformation phases, we can sort the complex array of D2 structures (Fig. 3) into two structural frameworks: (1) the sinistral strike-slip ∼ E–W and NE–SW-striking features (moderately to steeply dipping ∼ N, SSW, and NW), the normal faults striking ∼ N–S (moderately to steeply dipping W), and the ENE–WSW-striking (steeply SSE-dipping) reverse dip-slip features; and (2) the ENE–WSW-striking (steeply SSE-dipping) dextral strike-slip features, the ∼ N–S-striking (steeply E-dipping) reverse faults, and the ∼ E–W to NE–SW-striking (moderately to steeply dipping ∼ N/S to SE) normal dip-slip features.

The first set of structures indicates sinistral strike-slip or transtension conditions along the MTFC, with ∼ WNW–ESE crustal extension and ∼ NNE–SSW shortening. This is largely consistent with the Late Carboniferous–Early Permian MTFC evolution, as proposed by Seranne (1992) and Watts et al. (2023) (Fig. 12). Furthermore, the ∼ N–S-striking normal dip-slip features on Smøla have a similar geometry to the ∼ N–S extensional faults (Raudtinddalen Fault) off the VF, which have been associated with a similar age and kinematics (Watts et al., 2023). The moderately to shallowly south-dipping D2 features in drill holes BH1 and BH2, as shown in 3D modelling (Fig. 8a), likely coincide with the damage zone of the nearby major E–W structure. Adjacent to the drill holes, the major ∼ E–W structure also locally changes the strike to a NW–SE trend (Fig. 2bI). The shallow dipping D2 features, although possibly the result of local stress perturbations (e.g. Kim et al., 2004), may have formed from local shortening associated with the strike rotation of the major E–W structure (forming a restraining bend) (e.g. Cunningham and Mann, 2007), further supporting sinistral faulting.

Based on the above interpretation, the ∼ E–W, ENE–WSW, and NE–SW sinistral and reverse features may correspond to L2 and L3 (Fig. 2bI and II) second- to third-order lineaments, and the ∼ N–S-striking normal dip-slip features correlate with the L1 lineaments. Within this setting, the L3 lineaments could represent sinistral R and R′ Riedel shears off the L2 lineaments (Fig. 2bI), with the L1 lineaments predominantly involving extensional features accommodating an ∼ E–W/WNW–ESE extension. Additionally, during this phase, the L2 second-order lineaments which correlate with the major E–W structures and steep sinistral features (with sub-horizontal slip lineations) could also have formed as synthetic P shears off the main MTFC trend (Fig. 12). This earlier episode may therefore correspond to an overall NE–SW crustal stretching and rifting during the Late Carboniferous–Early Permian (timeline in Fig. 12), resulting in the Late Palaeozoic basins such as the Froan Basin and Høybakken-fault-related basin (Kendrick et al., 2004; Faleide et al., 2008; Peron-Pinvidic and Osmundsen, 2018).

The second set of structures may be associated with the ∼200 Ma ages constrained by the finest grain size fractions of the D2-associated samples. These structures suggest dextral strike-slip along the MTFC, with an overall extension direction of ∼ NNE–SSW. Moreover, the steeply dipping ENE–WSW-striking dextral strike-slip features (with a horizontal slip lineation), being sub-parallel to the HSF (MTFC) trend, also correspond to the L3 lineaments. These L3 lineaments do show rare dextral offsets of both the L1 and L4 lineaments, further suggesting that some dextral kinematics is subtlety preserved (Fig. 2bI).

Therefore, reflecting the above, both the steep ∼ N–S-striking reverse faults associated with D2 and the major E–W structures (L2 lineaments) converging on the NE side of the Edøyfjorden basin (Fig. 2a), where there is an ENE–WSW to NE–SW strike rotation of the MTFC, may indicate a possible restraining band (associated with dextral strike-slip; represented in Fig. 12). As such, the E–W structures could represent a positive flower structure forming off the MTFC due to local transpression (e.g. Cunningham and Mann, 2007). The ∼ E–W to NE–SW moderately dipping normal features (with steeply plunging slip lineations), however, indicate that extensional conditions mostly prevailed over Smøla, and many of the ∼ E–W structures formed during earlier D2 times were reactivated as steep normal faults or involve oblique dextral–normal kinematics. Additionally, the shallow D2 structures 3D-modelled in Fig. 8a, dipping south towards the major E–W structures, might represent shallow dipping antithetic or conjugate features formed during an extensional setting (related to late MTFC dextral faulting and extension). The shallow dipping D2 structures in BH3 and BH4, in contrast, are within diorite and gneiss units (Fig. 1b) and may involve the reactivation of foliation planes commonly associated with more micaceous intervals during either early or late D2 times.

Dextral kinematics on the MTFC during the Jurassic were proposed by Grønlie and Roberts (1989), with Watts et al. (2023) suggesting instead that it occurred later along the VF during the Late Jurassic/Early Cretaceous. Dextral strike-slip faulting could have resulted from poorly constrained Triassic–Jurassic far-field compressive forces acting on a mechanically weak MTFC and associated structures. However, at that time, the region was experiencing the start of an extensional phase with regional subsidence (Tsikalas et al., 2012; Peron-Pinvidic and Osmundsen, 2018) in response to ∼ E–W crustal extension (Gernigon et al., 2020; Tartaglia et al., 2023) (shown in the timeline in Fig. 12), which does not fit well with dextral kinematics, suggesting that this tectonic phase remains uncertain.

Associated with D3, the hematite-bearing structures (Table 2; Fig. 9c and g) yielded finest size fraction K–Ar ages of between ∼128 and ∼100 Ma (Fig. 11). Importantly, however, the ∼128 Ma age from sample SK1033_1 may reflect local saprolitisation or hydrothermalism instead of deformation linked to D3, even though most Norwegian margin saprolite ages fall within the Triassic/Jurassic time interval (e.g. Fredin et al., 2017; Knies et al., 2022; Olesen et al., 2023). The age spectra plateau for sample SK2008, corresponding to ∼200 Ma (fractions 0.4–2 to 6–10 µm), and the finest fraction K–Ar ages for samples SK2012, SK2015, and SK1029_1 also suggest a D2 inheritance within the coarser fractions of this sample. Thus, the D3 outcrop structures (Fig. 3) can be subdivided into two groups: (1) the majority of the normal to oblique–normal features striking NNE–SSW (dipping steeply W) and NE–SW (dipping moderately to steeply SE), the ∼ ENE–WSW sinistral strike- to oblique-slip features (moderately to steeply NNE and SE-dipping, with inclined to sub-horizontal slip lineations), and the ∼ NW–SE-striking reverse dip-slip features (moderately ENE and steeply SSW-dipping); and (2) ∼ N–S-striking (steeply W-dipping) reverse faults and dextral features striking ENE–WSW (steeply NW-dipping and inclined slip lineation, suggesting oblique–normal kinematics).

The first group of structures (dominated by dip-slip features, with steeply plunging slip lineations), being sub-parallel to the trace of the MTFC and the axis of the Edøyfjorden basin, probably relate to a WNW–ESE to NW–SE extension, half-graben development, and Mesozoic sediment deposition within the Edøyfjorden basin (Bøe and Bjerkli, 1989). A subset of pervasive L1 and L3 lineaments (mostly third-order), also correspond to the moderately SE- or steeply W-dipping normal dip- to oblique-slip faults, either formed or reactivated during D3 (suggested by their lack of offset across the earlier lineaments). The mostly sinistral strike-slip features, however, could involve the coeval reactivation of steeply dipping D2 sinistral features within an extensional setting. Associated with these features, some lineaments also have limited horizontal offsets of the magnetic fabric over the D3 lineaments (Fig. 2bII). Regionally, this relates to Jurassic–Cretaceous extension and exhumation, with regional rifting and basin development (Fig. 12), leading to numerous Cretaceous basins (including the Møre and Vøring basins) (Faleide et al., 2008; Peron-Pinvidic and Osmundsen, 2018).

The second group of structures, particularly the steeply W-dipping ∼ N–S-striking reverse faults, indicate ∼ E–W shortening. This corresponds to gentle folds (N–S-oriented fold axes) in the Mesozoic basin rocks in the Edøyfjorden basin (Bøe and Bjerkli, 1989). The steep NNW–SSE and N–S significant structures (large red circles in the D3 “All planes” stereonets in Fig. 3) and the shallowly SE-/W-dipping thrusts 3D-modelled in Fig. 8a are also associated with this setting. Geometrically, these folds indicate Late Mesozoic dextral transpressional conditions after dip-slip faulting on the HSF/MTFC, with ∼ E–W shortening of the basin sediments (Bøe and Bjerkli, 1989). Watts et al. (2023), however, associated the late dextral strike-slip kinematics with zeolite–calcite mineralisations on the VF (less distinct on the HSF; D4 in this study).

The D4 zeolite and calcite structures were not dated in this study (Table 2; Figs. 9d and 11). However, these veins would have formed after ∼75 Ma, owing to crosscutting relationships (Fig. 9d). Tartaglia et al. (2023) and Watts et al. (2023) noted similar zeolite–calcite mineralisation in other parts of the margin (Hitra and Runde islands) and on the Fosen peninsula associated with the VF (only locally along the HSF). This mineralisation has been attributed to (I) earlier Mesozoic dextral strike-slip to transpressional faulting related to the N–S folds in the Edøyfjorden basin (D3 structures; this study) and (II) later extensional dip-slip faulting on the HSF post-dating the zeolite–calcite features (Watts et al., 2023) or (III) Late Cretaceous extension in the distal part of the rift margin (Gernigon et al., 2020; Tartaglia et al., 2023). The D4 NNE–SSW and ENE–WSW sinistral/dextral faults and associated L1 and L3 lineaments may therefore preserve evidence of the earlier D3 Mesozoic (Jurassic) dextral strike-slip movement on the MTFC, and the NNE–SSW to NE–SW normal features (dipping ESE/W), with corresponding L1 and L4 lineaments, are attributed to a later D4 Cretaceous to Late Cretaceous (NW–SE) extension.

Representing the last Cretaceous–Palaeocene deformation episode, the D5 quartz–calcite veins (Figs. 5f and 6h, i) are similar to calcite veins U–Pb dated in western Norway and the WGR (Hestnes et al., 2023). Different veins oriented NE–SW yielded ages between ∼90 and ∼80 Ma, associated with Cretaceous reactivation of the MTFC; ∼ NW–SE (and NE–SW) veins formed between ∼70 to 60 Ma, associated with regional uplift possibly related to the doming of the proto-Icelandic plume; and NE–SW veins forming <50 Ma, associated with different Cenozoic extensional processes (lithospheric flexure and uplift from far-field tectonic stresses) (Hestnes et al., 2023).

The D5 structures, mostly normal dip-slip features or tensile veins, strike NE–SW (dipping steeply to moderately NW/SE) and NW–SE (steeply NE-dipping) (Fig. 3) and correspond to the L3, L4, and late L1 lineaments. These structures and lineaments are mostly late reactivations of older features and are present in all parts of Smøla and SE of the MTFC (HSF) in the WGR (Fig. 12). Overall, they suggest an ∼ E–W to NE–SW crustal extension. Similar to these D5 structures, and likewise indicating a late timing for D5, Bøe and Bjerkli (1989) also documented NW–SE-striking normal faults (SW-/NE-dipping) offsetting Mesozoic sediments within the Edøyfjorden basin. Based on geometry and timing, the D4 and D5 structures potentially relate to both a Cretaceous reactivation of the MTFC and the Cretaceous–Palaeocene rifting preceding the break-up of Greenland and Norway (Faleide et al., 2008). Importantly, the D4 and D5 features are sub-parallel to extensional structures offshore, particularly to the Klakk Fault Complex on the western side of the Frøya High and the Halten Terrace, as well as the Høybakken Detachment (HD) to the NW of Hitra and Frøya, suggesting a temporal and genetic association.

Smøla island represents an ideal onshore analogue for offshore basement highs in the area. The Frøya High, offshore Norway, as an example (owing to it having similar basement rocks to Smøla), is likewise bound by major structures such as the Klakk Fault Complex, a major west-dipping extensional fault zone (Muñoz-Barrera et al., 2020). The Frøya High experienced a comparable polyphase deformation related to crustal stretching and thinning from the mid-Carboniferous to the Late Cretaceous–Palaeocene rifting (Faleide et al., 2008, 2015; Peron-Pinvidic and Osmundsen, 2018). To unravel its complex tectonic history, this structural high would therefore require a similar approach as our Smøla study. Although numerous drill holes have been completed along the mid-Norwegian margin, with at least three drill holes on Frøya (e.g. Slagstad et al., 2008; Bunkholt et al., 2022), the quality of the drill hole material is poor (only rock cuttings or rock plugs) (Muñoz-Barrera et al., 2020). Similarly, margin-wide geophysical (magnetic, gravity, and 2D or 3D seismic reflection surveys) datasets are also available (Skilbrei et al., 2002; Muñoz-Barrera et al., 2020), but most will only afford the ability to interpret at or above the seismic resolution scale. The advantage, therefore, of using an onshore analogue such as Smøla is the access and the availability of numerous datasets without these limitations.

Moreover, investigating fractured and weathered basement is crucial for understanding basement-hosted oil and gas reservoirs or fluid pathways (Trice et al., 2022; Trice, 2014). The petrophysical attributes of various deformation features associated with the D1 to D5 episodes could significantly impact fluid migration or storage. As these features exhibit crosscutting relationships and different orientation trends, structural permeability anisotropy is clearly present in the Smøla basement volume. Therefore, using an onshore analogue to produce high-resolution structural characterisations aided by the absolute time dimension may be crucial in the future understanding of basement-hosted plays offshore.