Abstract
Detailed geological mapping of the La Cerdanya area (Canigó unit, Eastern Pyrenees) provides new data characterizing the different structural styles exhibited by Cambrian–Lower Ordovician (Jujols Group) and Upper Ordovician successions. Their unconformable contact, related to the Sardic Phase, ranges from 0° (paraconformity) to 90° (angular discordance). The Jujols Group rocks topped by the unconformity are affected by Sardic foliation-free open folds. The pre-Sardic succession, the Sardic Unconformity and the lower part of the post-Sardic succession (Rabassa Conglomerate and Cava formations) are cut and offset by several Late Ordovician NNE–SSW-trending synsedimentary extensional faults associated with hydrothermal activity, which dramatically affected the thickness of the lower part of the Upper Ordovician succession. We relate (1) the Mid-Ordovician Sardic uplift and erosion, and (2) a Mid- to Late-Ordovician upward propagating extensional fault system bounding the outline of half-grabens, subsequently infilled by alluvial deposits, to a thermal doming event (about 475–450 Ma) that led to the uplift and stretching of the Ordovician lithosphere. Thermal doming may be caused by mafic magma underplating and responsible for the coeval calc-alkaline magmatic activity broadly developed in the Eastern Pyrenees. We discuss the similarities between the Mid-Ordovician Sardic Unconformity and other Early Paleozoic unconformities described in neighbouring areas. Finally, we suggest a geodynamic scenario in which a regional-scale thermal event was related to the opening of the Rheic Ocean that led to the drift of Avalonia from the SW European margin of Gondwana.
The Pyrenees are a WNW–ESE-trending Alpine fold and thrust belt that contains pre-Variscan basement rocks of late Neoproterozoic to Carboniferous age (Fig. 1). The pre-Variscan basement forms a large strip in the core of the cordillera and provides evidence of Cadomian, Sardic and Variscan magmatic episodes (Cocherie et al. 2005; Castiñeiras et al. 2008; Casas et al. 2010, 2015; Navidad et al. 2010; Pereira et al. 2014; Martínez et al. 2016; Padel et al. 2018a), and Variscan and Alpine deformation events (Guitard 1970; Zwart 1979; Muñoz 1992). The onset and significance of the Ordovician deformation has been matter of debate since the pioneer work of Llopis Lladó (1965). After Santanach (1972), it is widely accepted that the Upper Ordovician succession unconformably overlies either the Cambrian–Lower Ordovician Jujols Group or the Ediacaran–Terreneuvian Canaveilles Group (Den Brok 1989; García-Sansegundo & Alonso 1989; Kriegsman et al. 1989; Poblet 1991; Muñoz & Casas 1996; García-Sansegundo et al. 2004; Casas & Fernández 2007). However, the origin of this unconformity has been the subject of several interpretations. Santanach (1972) in the Canigó massif and García-Sansegundo et al. (2004) in the Garona dome attributed the unconformity to basement tilting, related to a Late Ordovician faulting episode and subsequent erosion. In the Lys–Caillaouas massif, Den Brok (1989) and Kriegsman et al. (1989) proposed the existence of a pre-Variscan deformation event. A pre-Late Ordovician folding episode has been also suggested as related to the unconformity on the southern Canigó massif (Casas 2010; Casas et al. 2012). However, the geodynamic significance of this deformation episode is not yet well established: it is not related to metamorphism or cleavage development, although it seems to be related to uplift, widespread emersion and considerable erosion before the onset of Late Ordovician deposition (Casas 2010). As a result, the Upper Ordovician rocks onlap a palaeorelief formed by different pre-Upper Ordovician formations, ranging from the upper Neoproterozoic to the Lower Ordovician in the Central and Eastern Pyrenees (Santanach 1972; Laumonier & Guitard 1986; Cirés et al. 1994). This intra-Ordovician unconformity may be interpreted as equivalent to the ‘Sardic Phase’ described in Sardinia (Teichmüller 1931; Stille 1939), which separates the Cambrian–Ordovician from the Upper Ordovician successions (Pillola et al. 2008).
Geological sketch of the Eastern Pyrenees with the location of the study area.
In this paper we present new data on the structural style exhibited by the Cambrian–Ordovician and Upper Ordovician successions of the La Cerdanya area, in the Eastern Pyrenees. The data provide clear evidence of well-expressed Ordovician deformations in the form of extensional faults and folds, the latter affecting only the Cambrian–Ordovician succession. The interpretation of these pre-Variscan deformation structures should contribute to a better understanding of the origin and meaning of the Sardic Unconformity, and thus to discussion of the geodynamic setting of this fragment of North Gondwana during Ordovician times, bracketed between the Cadomian and Variscan orogenies.
Geological setting
The study area crops out in the southern flank of the Canigó Unit, which represents an antiformal Alpine structure with exposures ranging in age from Late Neoproterozoic to Carboniferous (Fig. 1). The Alpine floor thrust of this unit is the Orri thrust (Muñoz 1992).
In the study area (Fig. 2), the Cambrian–Lower Ordovician succession consists of a thick package (c. 1500 m) of grey to greenish shales alternating with subsidiary centimetre-scale sandstone interbeds. This succession represents the Serdinya Formation (Padel et al. 2018b) of the Jujols Group (Laumonier 1988). Acritarchs recovered from the uppermost part of this succession, close to the Sardic Unconformity at La Molina locality (Fig. 1) (southern slope of the Canigó massif), indicate a Furongian–earliest Ordovician age (Casas & Palacios 2012). This age is coincident with a maximum depositional age of c. 475 Ma for the uppermost part of the Jujols Group in the La Rabassa dome, on the basis of the youngest detrital zircon population (Margalef et al. 2016). To the east, in the Albera massif, metapelites and metapsammites from the uppermost part of a metasedimentary succession that can be correlated with the Jujols Group are crosscut by acidic subvolcanic dykes, which constrain its minimum depositional age to 465–472 Ma (Liesa et al. 2011). All these data suggest a depositional age for the uppermost part of the Jujols Group at c. 475 Ma. On the other hand, a c. 455 Ma U–Pb age for the Upper Ordovician volcanic rocks directly overlying the Sardic Unconformity has been proposed in the Eastern Pyrenees (Martí et al. 2014), in the Les Gavarres (455 ± 1.8 Ma, Navidad et al. 2010) and Les Guilleries (452 ± 4 Ma, Martínez et al. 2011) areas. This suggests a time gap of about 20 myr for the Sardic Phase in the Pyrenees, similar to the gap found in SW Sardinia, the type area where the original unconformity was described; there, the unconformity separates the Cambrian–Ordovician from the Upper Ordovician successions, marking a stratigraphic gap of c. 18 myr that includes part of the Floian (Pillola et al. 2008), the entire Dapingian and Darriwilian, and part of the Sandbian (Hammann 1992; Leone et al. 2002).
(a) Detailed geological map of the study area. (b) Geological cross-section of the study area. Location shown in Figure 1.
In the Eastern Pyrenees, the overlying Upper Ordovician succession (Cavet 1957; Hartevelt 1970) forms a broad fining-upward siliciclastic package that comprises some limestone key levels and displays significant thickness variations, ranging from 100 to 1000 m (Fig. 3). This succession was originally described by Hartevelt (1970), who distinguished five formations: from base to top, the Rabassa Conglomerate, Cava, Estana, Ansobell and Bar Quartzite formations. The Rabassa Conglomerate Formation, up to 100 m thick, consists of reddish-purple, polygenetic conglomerates with heterometric clasts composed of quartzite and slate derived from underlying rocks and vein quartz. The overlying Cava Formation, 0–850 m thick, is made up of conglomerates, sandstones and shales with volcanic intercalations; one coquina punctuating the shale interbeds has yielded brachiopods (Svobodaina havliceki, Rostricellula sp. and Rafinesquina sp.) and bryozoans, suggesting a Katian (former late Caradoc–early Ashgill) age (Hartevelt 1970; Gil-Peña et al. 2004; Puddu & Casas 2011). The Estana Formation, 5–200 m thick, is composed of limestone and marly limestone rich in brachiopods, bryozoans, echinoderms and conodonts of late Katian age (Gil-Peña et al. 2004). The top of the carbonate succession is unconformably capped by the black–grey shales of the Hirnantian Ansobell Formation, 20–320 m thick, which in turn is overlain by the Bar Quartzite, 2–20 m thick and dated as Hirnantian on the basis of its fossiliferous content (Sanz-López & Sarmiento 1995; Sanz-López et al. 2002; Roqué Bernal et al. 2017; Štorch et al. 2018).
Cambrian–Ordovician and Upper Ordovician stratigraphic log measured near Ordén. Location shown in Figure 2. 1, shales; 2, siltstones; 3, sandstones; 4, conglomerates; 5, quartzites; 6, marlstones; 7, fossils; 8, ripples; 9, plane lamination; 10, bioturbation.
In several areas of the Pyrenees, from Pierrefite to the west (Calvet et al. 1988) to the Ribes de Freser area to the east (Robert & Thiébaut 1976; Ayora 1980; Robert 1980; Martí et al. 2014) volcanic rocks have been reported interbedded with the Upper Ordovician sediments. These volcanic rocks consist of pyroclastic deposits with subsidiary lavas and subvolcanic intrusive rocks (Martí et al. 1986, 2019). They include andesite, rhyodacite and rhyolite, which make up a small fraction of the corresponding Ordovician succession. Additionally, the Ribes area includes a subvolcanic granitic body, the Ribes granophyre, emplaced in the lower part of the Sandbian–lower Katian succession and dated at 458 ± 3 Ma by Martínez et al. (2011), and several granitic orthogneissic bodies emplaced in the lower part of the pre-Variscan succession (the Núria gneiss with a protolith age of 457 ± 4 Ma, Martínez et al. 2011; and the Queralbs gneiss, with a protolith age of 457 ± 5 Ma, Martínez et al. 2011).
All these pre-Variscan rocks display a polyphase deformation linked to the Variscan deformation and a low-pressure–high-temperature metamorphism (Guitard 1970; Zwart 1979). A pervasive foliation is the main deformational structure in the medium- to high-grade metasediments and the gneissic bodies derived from Ordovician and Cadomian granitoids. This foliation (S1–2) is a composite fabric developed prior to or coeval with the Variscan regional metamorphism, being associated with an east–west- to NE–SW-oriented stretching lineation. S1–2 is folded by later east–west to NW–SE upright folds, forming the mesostructure that depicts the Aston–Hospitalet, Canigó, Roc de Frausa and Albera gneissic domes along the backbone of the chain. According to Clariana & García-Sansegundo (2009) and García-Sansegundo et al. (2011), in the Garona dome and the Pallaresa massif, S1 is a slaty cleavage developed only in the Cambro-Ordovician succession and not related to folds, and S2 is associated with north-verging recumbent folds formed at Carboniferous time. In the eastern part of Pallaresa massif (central Pyrenees), south of Aston massif and north of the Tor–Casamanya syncline, Clariana & García-Sansegundo (2009) and Clariana et al. (2009) described a subvertical S3 crenulation cleavage, axial plane of east–west-trending upright folds, deforming S1–2. This S3 becomes the main cleavage in the Pallaresa massif, the Tor–Casamanya and Llavorsí synclines, and the Rabassa and Orri domes (Speksnijder 1986; Poblet 1991; Clariana & García-Sansegundo 2009; Margalef & Casas 2016). S3 exhibits a fan-like attitude close to the southern contact with the post-Variscan rocks owing to the Alpine rotations linked to thrust sheet stacking (Muñoz 1992) The Devonian and the Carboniferous Culm rocks of the study area exhibit a well-developed cleavage that can be correlated with this S3 mesostructure (Martín-Closas et al. 2018). Additionally, to the south of the study area, several thrust sheets involving Devonian and Carboniferous rocks can be observed (Hartevelt 1970; Domingo et al. 1988; Casas et al. 1989; Martín-Closas et al. 2018). Their basal detachment is located at the base of the Silurian black shales; although most of the thrusts are south-directed, some north-directed thrusts can be observed. In the southern limb of the D3 megastructures, such as the Rabassa dome, thrusts and the lower detachment dip to the south. In turn, thrusts cut the fold-related cleavage. Thus, thrust development in the post-Silurian rocks must have been broadly coeval with the onset of D3 folds in the low-grade pre-Silurian rocks (Martín-Closas et al. 2018). In the Eastern and Central Pyrenees this deformation episode occurred at 330–319 Ma, according to the age of synorogenic Carboniferous Culm deposits (decribed as latest Visean to Serpukhovian by Sanz-López et al. 2006; Namurian by Delvolvé 1981; Delvolvé & Perret 1989; Delvolvé et al. 1993; Martín-Closas et al. 2018). This age is coincident with the maximum deposition ages of 336 ± 2 and 356 ± 9 Ma obtained from U–Pb dating on detrital zircon from Carboniferous rocks (Martínez et al. 2016). A minimum age, in turn, is provided by the age of the crosscutting Andorra–Montlluís granodiorite, dated as 305 ± 3 Ma by Romer & Soler (1995), 305 ± 5 Ma by Maurel et al. (2004) and 301.5 ± 1.9 Ma by Pereira et al. (2014).
Structure
We made a detailed (1:5000 scale) geological map of the Talltendre area and measured a stratigraphic log of the Upper Ordovician succession cropping out along the road near Ordén village (Fig. 2). This study area constitutes the WNW–ESE-oriented vertical to subvertical southern flank of the Canigó antiform that joins southwards with a synformal megastructure, where Silurian, Devonian and pre-Variscan Carboniferous successions are extensively exposed (Fig. 1). Together with this subvertical arrangement, excellent outcrop conditions allow the recognition of the Sardic Unconformity, a system of extensional faults and several deformation mesostructures chronologically subdivided into three episodes. To better visualize and reconstruct the fault geometry, cartographic data were integrated into a 3D environment (MOVE software), which also facilitated the geometric analysis of the Sardic Unconformity.
Sardic Unconformity
The Cambrian–Lower Ordovician outcrops are onlapped, at angles ranging from a few degrees to 90°, by both the Rabassa Conglomerate and the Cava formations (Figs 2–4). The angular discordance separates the Upper Ordovician succession on its southern side from underlying Cambrian–Ordovician strata on its northern side. This contact shows a planar to weakly undulated, locally irregular surface capping the Cambrian–Ordovician strata, and can be considered as an unconformity laterally ranging from a paraconformity to an angular discordance. The unconformity exhibits a roughly WNW–ESE trend (194/75), although locally the orientation ranges from east–west to NW–SE (Fig. 5a and b).
Examples of the Upper Ordovician unconformity in the study area. UO, Upper Ordovician; CO, Cambrian–Ordovician. Angular unconformity is 78° (a), 69° (b) and 34° (c), respectively. Location shown in Figure 2.
(a) A 3D representation of the study area with estimation of the orientation of the unconformity surface and the extensional faults; (b) equal area lower hemisphere stereoplot of the unconformity; (c) equal area lower hemisphere stereoplot of extensional faults from cartographic data using MOVE software.
The Upper Ordovician succession shows a NW–SE trend of vertical to subvertical strata. Upper Ordovician beds may be either overturned or strongly dipping northwards or dip eastwards in the eastern part of the study area (Fig. 6). In contrast, Cambrian–Ordovician bedding surfaces exhibit a wide range of orientations (see below).
(a–f) Equal area lower hemisphere stereoplots of the disposition of the Upper Ordovician bedding surfaces in the six domains (A1, A2, A3, A4, A5 and A6) defined from west to east between the extensional faults. The Upper Ordovician unconformity surface is shown (red dashed line) in the stereoplots together with the average bedding planes (black lines) in (a) to (d), or the best-fit girdle (blue line) in (e) and (f). n indicates the number of measurements.
Polyphase folding
The most visible mesostructures of the Cambrian–Ordovician and Upper Ordovician rocks are bedding surfaces. Field data show that bedding surfaces exhibit different orientations (Fig. 7). Orientation of the Cambrian–Ordovician succession presents a marked dispersion and dip values ranging from subvertical to subhorizontal (Fig. 7a). In contrast, the Upper Ordovician bedding poles exhibit a more regular NW–SE trend (Fig. 7b). As we will discuss below, the marked dispersion of the bedding poles in Cambrian–Ordovician strata suggests the superposition of several fold systems. In the description that follows, L1 and L3 refer to different fold axes, and L0–2 and L0–3 to bedding–cleavage intersection lineations.
Equal area lower hemisphere stereoplots of bedding (S0), D2-related mesostructures (S2 cleavage and L0–2 intersection lineation), and D3-related mesostructures (S3 cleavage and L0–3 intersection lineation) from the Cambrian–Ordovician (a, c, e) and Upper Ordovician successions (b, d, f) in the study area. In (c) and (d), the L0–2 intersection lineations are distributed along a great circle that is coincident with the S2 average plane (in green), and in (e) and (f) the L0–3 intersection lineations are distributed along a great circle that is coincident with the S3 average plane (in red). n indicates the number of measurements.
D1 structures
The existence of D1 folds can be inferred from the arrangement of bedding surfaces at outcrop and different map scales, and from the pattern displayed by D3 linear elements, intersection lineations and fold axes in the Cambrian–Ordovician sequence. That is, D3 folds developed over initially non-horizontal surfaces and D1 folds control the geometry of D3 mesoscale folds (e.g. Turner & Weiss 1963, p. 130; Mey 1967, p. 194; Ramsay 1967, p. 539). The existence of previous D1 folds may also explain the distribution of D2 mesostructures in the Cambrian–Ordovician rocks. Concerning the initial orientation of D1 fold axes, the map analysis suggests the following: (1) despite dispersion, Cambrian–Ordovician bedding exhibits two maxima in orientations at WNW–ESE and NE–SW (Fig. 7a); (2) WNW–ESE orientations predominate over NE–SW ones; (3) especially interesting are the outcrops in which the Cambrian–Ordovician beds are NE–SW oriented, and are both vertical and normal to the unconformity horizon (Fig. 8a). In such a situation, after unfolding the Upper Ordovician sequence using a WNW–ESE horizontal rotational axis parallel to the Upper Ordovician unconformity, the NE–SW vertical Cambrian–Ordovician beds still remain vertical and with a similar NE–SW orientation (Fig. 8b). This observation suggests that D1 fold axes presented an initial subhorizontal NE–SW orientation, with a short limb oriented NE–SW subvertical or strongly dipping, and a long limb subhorizontal or gently dipping towards the SE or NW. However, the effect of superimposed D3 folds precludes assessment of whether D1 folds have a NW or SE vergence. Analysis of the areas where subvertical WNW–ESE and NE–SW orientations coexist allows the estimation of 10–100 m scale dimensions of D1 folds (Fig. 9). It should be noted that there is no evidence of an axial planar cleavage development or other mesostructures related to D1 fold formation.
(a) Sketch and equal area lower hemisphere stereoplot of the Upper Ordovician unconformity (033/83) overlying Cambrian–Ordovician subvertical beds (296/77 and 178/86); (b) disposition of the Cambrian–Ordovician beds after a c. 90° rotation about a horizontal axis parallel to the unconformity (00/123).
(a–d) Detailed sketches showing the disposition of S0 bedding planes in the Cambrian–Ordovician rocks. Inset map shows location of the different zones.
D2 structures
A not very pervasive cleavage (S2) is observable at outcrop scale in the Cambrian–Ordovician rocks, whereas S2 is well developed in the fine fraction of the Upper Ordovician, being deflected by coarser deposits (such as conglomerates and sandstones), where it develops irregularly. S2 surfaces generally dip 20–80° towards the east or ENE (Fig. 7c and d) and are commonly deformed by D3 folds. In the Upper Ordovician rocks, S2 exhibits dispersion along a north–south-oriented girdle (Fig. 7d). Neither folds nor other mesostructures related to this event have been clearly identified in the field. As for D3 linear mesostructures, L0–2 exhibits a wider dispersion in the pre- than in the post-unconformity rocks (Fig. 7c and d).
D3 structures
A north-dipping S3 cleavage is the dominant deformation structure recognizable in the exposures of the study area. S3 cleavage is observable from outcrop down to microscopic scale, and is correlated with the dominant cleavage observed in the Silurian and Devonian rocks cropping out to the south of the area. S3 is oriented NW–SE with variable dips towards the NE. In the eastern area, it is subvertical or strongly dipping, whereas in the western sector it dips gently to moderately towards the NE (Fig. 2).
S3 is related to D3 folds affecting both the pre- and post-unconformity rocks. As for the S3 surfaces, the D3 fold axial surfaces are generally gently dipping (10–45°) towards the north or NE in the western part of the study area, whereas to the east they have the same orientation, although steeply dipping or subvertical. D3 folds are SW vergent, open to tight and generally 1–100 m in scale with centimetre-scale minor folds (second- or third-order folds) on the limbs of the main folds. D3 folds show mildly curved hinges that generally developed harmonic folding, especially in the Upper Ordovician rocks. D3 fold axes exhibit different orientations in the pre- and post-unconformity rocks (Fig. 7e and f). In the Upper Ordovician succession, L3 axes and L0–3 intersection lineations are grouped forming a maximum with a moderate plunge to the WNW or ESE (Fig. 7f). The bedding–cleavage relationship indicates that the study area constitutes the southern flank of a south-facing first-order D3 fold oriented WNW–ESE (Fig. 2b).
In contrast, the minor L3 axes and L0–3 intersection lineations of the Cambro-Ordovician rocks display a wide range of orientations. L0–3 plunge ranges from subhorizontal to 45° (Fig. 7e), and no simple distribution pattern may be deduced from map analysis (Fig. 2). If L0–3 and L3 are plotted together, they are mainly distributed along a great circle coincident with the S3 average plane, although two maxima. oriented c. 20°–40°/000 and 25°/115, can be envisaged with predominance of low plunge values (Fig. 7e).
Extensional faults and tectonically induced sedimentation
At the map scale, several extensional faults affecting the Cambrian–Ordovician strata, the Sardic Unconformity and the lowermost part of the Upper Ordovician succession can be recognized (Figs 2–5) (Casas & Fernández 2007; Casas 2010). At present, the faults trend broadly north–south to NNE–SSW, and their hanging wall generally appears to be the eastern block, despite the presence of some antithetic faults (Fig. 2). The faults are steep, domino-style and their surface dips range from 48 to 90° towards the WNW and ESE. This variation in orientation together with the changes in strike along a fault plane detected in 3D analysis suggest that the faults are slightly folded around a subhorizontal NNE–SSW axis (189/00) (Fig. 5c). Maximum throws vary between 200 and 750 m. The NNE–SSW fault orientation could have prevented the inversion of the faults during the Variscan and Alpine deformation events, but the orientation did not prevent their rotation following horizontal east–west axes during the former events. The faults are associated with synsedimentary hydrothermal activity, represented by dense networks of quartz veins encased in the uppermost part of the Jujols Group and subparallel to the trace of the extensional faults. These served as the source for huge amounts of vein quartz clasts that form the Rabassa Conglomerate (Fig. 10a and b).
Breccia levels of the Rabassa Conglomerate Formation neighbouring synsedimentary normal faults at La Molina. (a) Clast-supported angular cobbles of sandstone and shale (derived from the underlying Jujols Group) mixed with hydrothermal quartz clasts sourced from synsedimentary veins (arrowed). (b) Thin-section photomicrograph of previous sample showing reworked hydrothermal polycrystalline quartz (pQ), siltstone (s), claystone (cl) and mafic phenocryst (v) clasts embedded in an unsorted sandy-to-clayey matrix; scale represents 1 mm.
The displacement of some of these faults diminishes progressively up-section and vanishes in the upper part of the Upper Ordovician rocks, especially in the lower part of the Cava Formation, indicating that the faults became inactive before deposition of the overlying Estana strata (Fig. 2). Thus, the Rabassa Conglomerate and the lower part of the Cava Formation are syntectonic sediments, whereas the upper part of the Cava Formation and the overlying Estana, Ansobell and Bar formations are post-tectonic sediments. As a result, the whole Upper Ordovician succession shows sharp changes in total thickness ranging between 190 and 900 m, with a general increase from west to east. Perpendicular to these faults, the Rabassa Conglomerate Formation displays sharp changes in thickness (from 0 to 100 m) and facies associations. Deposits of the Rabassa Conglomerate Formation neighbouring the extensional faults consist of chaotic breccias ranging both laterally and vertically to matrix- to clast-supported conglomerates (Fig. 11d and e). These include massive to weakly stratified, unsorted, pebble-dominated deposits of sandstone, shale, vein quartz and mafic phenocrysts (Fig. 11a and b). Contacts are commonly scoured and grading is both normal and inverse. Breccias and matrix-supported conglomerates show either chaotic or amalgamated patterns; they interfinger laterally with packages of clast-supported conglomerates and pebbly sandstones that pinch out or change abruptly, both laterally and vertically, into sandstone–shale alternations (Cava Formation). Clast-supported conglomerates show massive to weakly imbricated pebbles and cobbles (Fig. 11f), which are subrounded, moderately sorted and grade upsection and laterally into litharenites; the latter display centimetre- to decimetre-scale trough and low-angle cross-stratified sets alternating with shales (Fig. 11g), which are irregularly burrowed. Lithostratigraphically, the occurrence of shale interbeds marks the base of the Cava Formation. Three facies associations can be recognized, as follows.
Breccias and matrix-supported conglomerates are interpreted as having undergone different types of downslope transport by basal sliding of plastic to semi-rigid sediment masses that included contemporaneous hydrothermal dykes of quartz (Fig. 11e). Amalgamation of debris-flow deposits indicates the presence of neighbouring slopes close to synsedimentary extensional faults.
Packages of clast-supported conglomerates and pebbly sandstones represent in-channel gravel bars that graded into straight to sinuous crested bedforms in shallow, braided systems. During high-discharge events, sediments were reworked and accumulated as channelized conglomerates.
Finally, sandstone–shale alternations reflect the influence of marine shoreface-to-offshore conditions with local development of burrowing by soft-bodied metazoans.
(a) Sardic angular discordance separating the Jujols shales (Ju; left) from the Rabassa breccia deposits (Ra; right) at Talltendre. (b) Subvertical dyke of hydrothermal quartz representing a synsedimentary fault marking the Jujols–Rabassa contact at La Molina; scale represents 10 cm. (c) Folded dykes of hydrothermal quartz encased in the Jujols shale neighbouring a synsedimentary normal fault that separates it from the Rabassa Formation at Talltendre. (d) Basal part of the Rabassa Formation close to the unconformity of (a), showing a breccia of clast-supported polymictic blocks rich in shales and sandstones of the underlying Jujols Group and hydrothermal quartz clasts sourced from the dykes (arrowed) that crosscut the outcrop at Talltendre. (e) Matrix-supported breccia of the Rabassa Formation with folded dykes of hydrothermal quartz (arrowed) at Talltendre. (f) Typical clast-supported polymictic conglomerate of the Rabassa Formation at La Molina. (g) Amalgamated trough cross-stratified sets of a lenticular sandstone package, up to 2.2 m thick and with 8 m of lateral continuity, embedded in shales; lower part of Cava Formation at Talltendre.
It should be noted that as the Upper Ordovician unconformity is cut by these normal faults, significant pre-Upper Ordovician erosive processes took place preceding the normal faulting event.
Synsedimentary faults are planar, either non-rotational or with small amounts of rotation affecting the beds in the hanging wall. The geographical distribution of these normal faults allows subdivision of the Upper Ordovician succession into six domains (Fig. 6). In the four western domains (A1–A4), the Upper Ordovician bedding planes exhibit an east–west- or NW–SE-trending orientation, ranging from N100E to N120E with subvertical or strongly dipping attitude. In domain A5, the Upper Ordovician bedding planes exhibit a wider range of dipping values, with bedding poles distributed in a NE–SW girdle, as a result of a 119/28-oriented fold axis (Fig. 6). In the easternmost domain A6, the Upper Ordovician bedding planes show a roughly NW–SE trend with variable dip as a result of a 115/16-oriented fold axis. The divergence between the beds next to the hanging wall varies between a minimum value of 10° (A5) and a maximum of 24° (A4), with a main value of 18° (A2) and 20° (A3), which allows us to infer a c. 20° fault rotation.
It is difficult to identify the faults and to recognize an eventual detachment horizon in the Cambrian–Ordovician rocks. This and the steep arrangement of the fault surfaces preclude an estimate of the amount of extension linked to fault movements. In any case, the small amount of bed rotation in the hanging wall of the faults, c. 20°, indicates a small amount of east–west extension.
Discussion
Structure: age and significance of folds D1, D2 and D3
D1 folds affecting the Cambrian–Ordovician succession are sealed by Upper Ordovician strata. Thus, it is assumed that the episode D1 is earlier than the successions deposited above. An Early–Mid-Ordovician age can be proposed for this folding event. Moreover, D1 fold axes are parallel to subparallel to the direction of the Late Ordovician extensional faults. We noted the absence of microscale D1 folds and the lack of either associated cleavage or metamorphism. We interpret that D1 folds can form at the hanging-wall blocks of extensional faults. Folds related to extensional faults have been extensively described, caused by slip on nonplanar normal faults (Xiao & Suppe 1992), on faults with ramp–flat–ramp fault geometry (Rotevatn & Jackson 2014), on a localized zone of shortening between two fault trends (Abu Sharib et al. 2017) or as fault-propagation folds (Sharp et al. 2000). From these possibilities, we favour that D1 folds formed as fault-propagation folds of upward-propagating normal faults. In this situation, vertical or steeply dipping pre-Sardic beds in the hanging wall near the fault surface are common as propagation of the fault tip to the surface results in breaching of a previous fold (Sharp et al. 2000). Therefore, two possibilities can be envisaged: either the fold formation is linked to pre-Sardic (Early–Mid-Ordovician) movements of the mapped extensional faults, or they resulted from the movement of other extensional faults, not detected in the area because of the monotonous character of the Cambrian–Ordovician succession and the lack of cartographic marker beds. According to the first possibility, post-Sardic (Late Ordovician) upward fault propagation should lead to normal faults cutting previous folds, the topographic surface and the unconformity, and post-Sardic sediments filling the hanging walls thinning away from fault surfaces (Fig. 2). According to the second option, the faults would not reach the topographic surface, and normal faults would remain as blind faults in the pre-Sardic sequence. In any case, the Upper Ordovician unconformity changes from a paraconformity in the long limb of the D1 folds or in the footwall of the faults, to an angular unconformity in the hinge zone or in the vertical limbs of the D1 folds in the hanging wall of the faults (Fig. 9). As stated before, the outcrop conditions, together with the monotonous character and the lack of cartographic key levels in the pre-Sardic sequence, make it difficult to identify these blind faults as well as the continuation in the pre-unconformity sequence of the normal faults cutting the unconformity. The situation described above argues against a contractional origin for the Ordovician (Sardic) folds in the Eastern Pyrenees as previously proposed by Casas (2010) and Casas et al. (2012).
Equivalent D1 pre-main cleavage folds were described by Casas (2010) to the north of the Canigó unit, in the El Conflent area, and further to the east, in the La Molina area. The former exhibit a similar north–south direction with axes strongly dipping to the north, whereas in the La Molina area, pre-main cleavage folds are mainly oriented NNW–SSE. As in the study area, the combination of pre-main cleavage folds and syn-foliation folds with variable orientations gives rise to a complex disposition of bedding planes (Casas 2010).
The interpretation of the deformation episode D2 is more difficult owing to the lack of related mesostructures. It could be correlated with the S2 cleavage associated with north-verging recumbent folds described in the Garona and Pallaresa domes by Clariana & García-Sansegundo (2009) and Clariana et al. (2009). However, the lack of related folds in the study area makes this correlation speculative. In any case, the age of this deformation should be post-Ordovician (Variscan) because it affects both the pre- and post-unconformity deposits.
The main foliation (S3) recognized in the study area can be correlated with the pervasive axial plane crenulation cleavage (S3) of east–west-trending upright folds, which represents the main cleavage preserved in the Pallaresa massif, the Tor–Casamanya and Llavorsí synclines, and the Rabassa and Orri domes (Speksnijder 1986; Poblet 1991; Clariana & García-Sansegundo 2009; Margalef & Casas 2016). It should be noted that the Cambrian–Ordovician succession of these Pyrenean massifs shares a common structural arrangement: a regional crenulation cleavage, regularly oriented and dipping to the north, associated with intersection lineations and minor fold axes displaying a significant dispersion. This arrangement, described, for example, in the Rabassa dome (Poblet 1991; Capellà & Bou 1997), the Massana anticline (Hartevelt 1970; Poblet 1991; Casas et al. 1998), the Orri dome (Hartevelt 1970; Speksnijder 1986; Poblet 1991) and the Lys–Caillaouas massif (Den Brok 1989), among others, has been attributed to the presence of pre-foliation (pre-D3) deformations.
Origin of the Sardic Unconformity, extensional faults and D1 folds
The angular unconformity described above represents the prolongation of those described westwards by Casas & Fernández (2007), and in the La Molina area by Santanach (1972) and Casas (2010). The unconformity can be attributed to the doming, tilting and erosion of a pre-Sardic succession, which preceded the deposition of the Upper Ordovician conglomerates and sandstones (Santanach 1972; García-Sansegundo et al. 2004). Doming, tilting and truncation of Cambrian–Ordovician strata suggest relative uplift, perhaps associated with a fall in the eustatic sea-level.
The Sardic Unconformity was affected by a Late Ordovician extensional pulsation, which was responsible for the movement of the NNE–SSW-oriented normal faults that led to the contemporaneous opening of grabens and half-grabens. These were infilled with the alluvial, fluvial and volcano-sedimentary deposits of the Upper Ordovician Rabassa Conglomerate and Cava formations, which also reflected a significant and stepwise hydrothermal activity. Simultaneously, a Late Ordovician magmatic pulse yielded a varied suite of magmatic rocks: small granitic bodies are emplaced in the pre-unconformity strata of the Canigó massif and form the protoliths of the Cadí, Casemí, Núria and Canigó G-1 type gneisses (c. 457–446 Ma, Casas et al. 2010; Martínez et al. 2011; Navidad et al. 2018), together with metre-scale thick bodies of metadiorite (c. 453 Ma, Casas et al. 2010). Coeval calc-alkaline volcanic rocks (ignimbrites, andesites and volcaniclastic rocks) are interbedded in the Upper Ordovician succession of the Ribes de Freser area (Robert & Thiébaut 1976; Ayora 1980; Robert 1980; Martí et al. 1986, 2019); a granitic body at Ribes with granophyric texture, dated at c. 458 Ma by Martínez et al. (2011), intrudes at the base of the Upper Ordovician succession. Metre-scale rhyodacitic to dacitic subvolcanic sills intercalated within the pre-unconformity succession close to the base of the Upper Ordovician have also yielded Late Ordovician ages in the La Pallaresa dome (c. 453 Ma, Clariana et al. 2018).
Previously, a Floian–Dapingian magmatic pulse, coeval with the uplift and subsequent extensional activity related to D1 fold formation, took place. This pulse gave rise to the intrusion of voluminous peraluminous granites, about 500–3000 m thick and emplaced into the pre-unconformity strata. They form the protoliths of the large laccolith-shaped, orthogneissic bodies that punctuate the backbone of the Pyrenees, from west to east, the Aston (c. 470–467 Ma, Denèle et al. 2009; Mezger & Gerdes 2016), Hospitalet (c. 472 Ma, Denèle et al. 2009), Canigó (c. 472–462 Ma, Cocherie et al. 2005; Navidad et al. 2018), Roc de Frausa (c. 477–476 Ma, Cocherie et al. 2005; Castiñeiras et al. 2008) and Albera (c. 470 Ma, Liesa et al. 2011) massifs. As a whole, the Ordovician magmatism in the Pyrenees lasted about 30 myr, from c. 477 to 446 Ma, a time span contemporaneous with the Ordovician deformations described above (Fig. 12).
Relative probability plot of the geochronological ages (U–Pb on zircon) of the Ordovician magmatism of the Pyrenees. Data after Deloule et al. (2002), Cocherie et al. (2005), Castiñeiras et al. (2008), Denèle et al. (2009), Casas et al. (2010), Liesa et al. (2011), Martínez et al. (2011), Mezger & Gerdes (2016), Navidad et al. (2018) and Liesa et al. (unpublished data); n = 25.
Taken altogether, uplift, denudation, extensional tectonics and magmatic activity may be related to a thermal doming caused by the underplating of hot mafic magmas (Fig. 13). This scenario has been invoked as the origin of the Furongian–Early Ordovician magmatism in the Central Iberian Zone of the Iberian Massif related to the so-called Toledanian Phase (Bea et al. 2007; Montero et al. 2009; Díez Montes et al. 2010). The latter is characterized by a Furongian gap associated with the onset of a remarkable magmatic activity of calc-alkaline affinity, giving rise to the Ollo de Sapo and Urra formations, c. 495–470 Ma (see summaries by Álvaro et al. 2007; Casas & Murphy 2018; García-Arias et al. 2018; Sánchez-García et al., in press). In the Pyrenees, doming started at c. 475 Ma, causing regional uplift and emersion responsible for the end of the marine Jujols Group sedimentation, its erosion and the emplacement of a widespread calc-alkaline granitic magmatism (Fig. 13). Subsequent extensional episodes were related to the opening of half-grabens and their infill with the Rabassa Conglomerate Formation and the lower part of the Cava Formation. The latter sedimentation was coeval with magmatic activity until c. 450 Ma, when thermal relaxation led to marine transgression and sedimentation (middle part of the Cava Formation) that onlapped an inherited Sardic-related palaeorelief. After a glaciogenic transgression represented by the blanketing of the basin by the Ansobell Formation, isostatic rebound at c. 445 Ma may be signalled by the progradation of the Bar Quartzite Formation, coincident with the end of the magmatic activity and with the final sealing of inherited Sardic palaeotopographies in Silurian times. Later, the Variscan deformation overprinted all these rocks and structures in Mid-Carboniferous times.
Sketch showing the geodynamic evolution described in the text (not to scale).
Comparison with neighbouring areas
In Sardinia, the Sardic Unconformity is associated with an intra-Ordovician stratigraphic gap in the autochthonous Paleozoic basement of the Iglesiente–Sulcis area (southwestern Sardinia), which corresponds to the External Zone of the South European Variscan Chain. Similar gaps (named the Sarrabese Unconformity) are found in the External nappe Zone, associated with coeval igneous activities in the Sarrabus–Gerrei area (c. 465 Ma: Buzzi et al. 2007a, 2007b; Oggiano et al. 2010). The post-unconformity succession is related to Late Ordovician magmatism in SW Sardinia (c. 457 Ma: Pavanetto et al. 2012) and central–northern Sardinia (c. 469 Ma to c. 456 Ma: Helbing & Tiepolo 2005; Giacomini et al. 2006; Boriani 2008). Indeed, in Sardinia, the pre-Sardic sequence is affected by different east–west-striking structures with neither regional metamorphism nor cleavage, displaying 1–100 m scale folds without cleavage and faults, subsequently sealed by Upper Ordovician strata (Pasci et al. 2008; Cocco & Funedda 2012, 2019; Cocco et al. 2018). We recall, for instance, the structure of the Su Scoffoni area in southeastern Sardinia, where Cocco & Funedda (2019) described the Sardic Unconformity sealing folds affecting only the pre-Upper Ordovician succession. As a result, the Upper Ordovician succession in SE Sardinia overlies Cambrian–Ordovician beds with different attitudes, in a similar way to that described above for the Eastern Pyrenees. The occurrence of faults and an environment of half-grabens have been supposed by Brusca & Dessau (1968) and Martini et al. (1991), who associated them with a post-unconformity deposit (Monte Argentu Formation) in SW Sardinia.
In the Occitan Domain (Mouthoumet and Montagne Noire massifs) of southern France, the Ordovician is punctuated by two major gaps: a Sardic Unconformity (ranging from paraconformable to angular discordant contacts) and a diachronous Silurian erosive unconformity that represents blanketing of an inherited glaciogenic palaeorelief (Álvaro et al. 2016). As in the case for the Eastern Pyrenees and Sardinia, the Montagne Noire includes emplacement of massive (Sardic-related) Middle Ordovician granitic bodies. The migmatitic Somail Orthogneiss of the southern Axial Zone was derived from an Ordovician metagranite that was emplaced at 471 ± 4 Ma and 450 ± 6 Ma (U–Pb zircon ages; Roger et al. 2004; Cocherie et al. 2005). Recent geochronological dating of these augen orthogneisses has updated the story of various generations of granitoids that form the core of the Axial Zone (e.g. Cocherie 2003; Faure et al. 2004; Roger et al. 2004; Bé Mézème 2005; Charles et al. 2009). A post-Sardic rifting reactivation is recorded in the Mouthoumet massif and the Cabrières klippes (southern Montagne Noire), where Late Ordovician fault-controlled subsidence and the record of rift-related continental tholeiitic lavas were coeval, in some areas, with the onset of the Hirnantian glaciation (Álvaro et al. 2016, 2018). Reopening of rifting branches was followed by onlapping patterns and final sealing of Sardic palaeotopographies during Silurian and Early Devonian times. It led to fault-controlled subsidence and the generation of structurally controlled depocentres, at least in the Cabrières sector of the Montagne Noire and the Mouthoumet area. It was accompanied by marine transgression and extensional pulses that gradually led to flooding and onlapping on the shoulders of the rifting branches, which were finally sealed during Early Devonian times.
As a result, the Sardic Phase is shared by the Occitan Domain, the Eastern Pyrenees and Sardinia, all of them located along the eastern branch of the Variscan Ibero-Armorican Arc. According to some researchers (e.g. Stampfli et al. 2002), this margin of Gondwana faced the Proto-Tethys, also called Ran Ocean (Torsvik & Cocks 2009), and later the Paleo-Tethys Ocean (Torsvik & Cocks 2004; Von Raumer & Stampfli 2008). This geodynamic setting is different from that faced by the Iberian Massif, which is located along the western branch of the Ibero-Armorican Arc. In the Central Iberian and Ossa–Morena zones of the Iberian Massif, a similar but earlier unconformity, the Toledanian Unconformity, is represented by an angular discordance that separates variably tilted Ediacaran–Cambrian Series 2 rift-related successions from an overlying Tremadocian–Floian passive-margin succession. The gap involved (at least c. 22 myr) includes, at least, most of the Furongian and basal Ordovician, although the erosion can incise into the entire Cambrian and the upper Ediacaran Cadomian basement (Gutiérrez-Marco et al. 2002). The Toledanian Phase has been recently interpreted (Sánchez-García et al. in press) as representing the rift–drift transition that led to the opening of the Rheic Ocean (Díez Montes et al. 2010; Nance et al. 2010; Thompson et al. 2010; Álvaro et al. 2014, 2018; Casas et al. in press). Neither metamorphism nor penetrative deformation related to this unconformity has been described, except in the Marão Anticline and the Amêndoa–Carvoeiro Synform of the Central Iberian Zone. In these areas, cleavage-bearing folds with steep axial planes at high angles to Variscan structures have been described and related to transient compression and dextral strike-slip features along the Central Iberian-Ossa–Morena contact lineament (Romão & Ribeiro 1993; Correira Romão, et al. 2005; Romão et al. 2013; Amaral et al. 2014). A similar volcanic-free ‘Furongian gap’ related to a break-up unconformity has been reported in the central Anti-Atlas of Morocco, and related to the above-reported Toledanian Phase (Álvaro & Vizcaïno 2018).
The magmatism associated with the Toledanian and Sardic phases is dominantly felsic and calc-alkaline, and exhibits an arc-like signature that some researchers have interpreted as linked to subduction conditions (Castro et al. 2002; Del Greco et al. 2016). However, others have interpreted this signature as an inherited character caused by the fast melting of a subduction-related Neoproterozoic crust (Díez Montes et al. 2010; Navidad et al. 2010; Sánchez-García et al. in press). This fast melting may also explain the great amount of inherited zircons, the highly porphyritic character and the virtual absence of mafic bodies in these magmas. These may have resulted from the partial melting of granitoids or sediments in the continental lower crust caused by the underplating or intrusion of hot mafic magmas that produced a great heat input during an extensional event (Bea et al. 2007; Montero et al. 2009; Díez Montes et al. 2010).
Conclusions
In the Eastern Pyrenees, the Sardic Unconformity separates two successions with different structural features. The Cambrian–Lower Ordovician and the Upper Ordovician successions are affected by Late Ordovician extensional tectonics and by two Variscan deformation events. Moreover, the Cambrian–Lower Ordovician succession is also affected by an Early–Mid-Ordovician fold system that we relate to extensional faults. Upward propagation of some of these extensional faults during Sandbian–Katian times caused the faults to affect the Sardic Unconformity and the lowermost part of the Upper Ordovician succession.
We relate the extensional fault formation to a thermal doming that developed between 475 and 450 Ma. The doming should have produced stretching of the Ordovician lithosphere, together with emersion, uplift and erosion of the Cambrian–Ordovician sequence giving rise to the formation of the Sardic Unconformity. Thermal doming may also be responsible for coeval magmatic activity that is well developed in the same area and caused by hot mafic magma underplating. We compare the Mid-Ordovician Sardic Unconformity in the Pyrenees with contemporaneous gaps recognized in Sardinia and the Occitan Domain, and also discuss the similarities with the Furongian Toledanian gap described in the neighbouring Iberian Massif, which may have originated in similar scenarios.
We suggest a regional-scale thermal event ranging in age from Early–Mid- to Late Ordovician, which affected several areas of the Northern Gondwana margin. Its role in the geodynamic scenario that led to other significant Early Paleozoic events, such as the opening of the Rheic Ocean and the drift of Avalonia (sensu lato) from the western part of North Gondwana, requires investigation.
Acknowledgements
We acknowledge the use of MOVE (Midland Valley) and Stereonet (R. W. Allmendinger) software packages. Detailed reviews by R. Vissers and an anonymous referee greatly improved the first version of the paper.
Funding
Financial support was provided by Spanish MINECO projects CGL2015-66335-C2-1-R and CGL2017-87631-P.
Author contributions
J.M.C.: investigation (equal); C.P.: investigation (equal); J.J.Á.: investigation (equal); N.C.: visualization (equal).
Scientific editing by Stuart Jones
- © 2019 The Author(s). Published by The Geological Society of London. All rights reserved
References
Deciphering the Sardic (Ordovician) and Variscan deformations in the Eastern Pyrenees, SW Europe
