The origin of the Cantabrian orocline of the Variscan orogen in NW Iberia remains a topic of debate. We present a structural study of the Ponga Unit, a Cambrian to Carboniferous tectonostratigraphic package within the West European Variscan belt foreland fold-and-thrust belt that lies within the core region of the orocline. Our primary goal was to determine if W-plunging folds of the fold-and-thrust belt are attributable to formation of the Cantabrian orocline, or if they reflect lateral ramps in the underlying Variscan thrust faults.

The major lithologic units of the Ponga Unit are the rheologically competent Lower Ordovician Barrios quartzite, and the less-competent, Carboniferous Barcaliente limestone and Beleño shale and sandstone formation. Our mapping and structural analysis within the Ponga Unit focused on the Laviana, Rioseco, and Campo de Caso thrust sheets, and associated bounding thrusts. Over 800 structural orientation measurements were collected across the study area. These data, coupled with data compiled from regional geological maps, allow for analysis of the crustal structure. West-plunging folds of the Laviana, Rioseco, and Campo de Caso thrust sheets form kilometer-scale anticline-syncline pairs, producing a complex fold interference pattern that is characteristic of the Ponga Unit. Our analysis shows that: (1) the geometry of the W-plunging folds is inconsistent with a lateral ramp model; (2) the map pattern defines a mushroom-type fold interference pattern, indicating two distinct deformational events characterized by principal compressive stresses oriented at a high angle (perpendicular) to one another; and (3) paleomagnetic data from the study area are consistent with the secondary model of orocline formation and indicate that there was a short window of time between the end of Variscan orogenesis and the onset of oroclinal buckling. Our results indicate that early N-S–trending folds, which resulted from Variscan orogenesis, were refolded during a post-Variscan orogen-parallel compression event attributable to formation of the Cantabrian orocline.


The origin of bends, as observed in map view, of mountain systems and orogenic belts is debated. End-member progressive and secondary models make testable predictions about the timing and processes involved in the development of such bends. In progressive models, curvature is thin-skinned, develops progressively during orogenesis due to local vertical-axis rotations, and is driven by the same stress field responsible for orogenesis. Indentation by continental promontories is probably the most common progressive model applied to explain curved continental collisional orogens. Examples include the Appalachian Kinston orocline (Marshak and Tabor, 1989) and the Cordilleran Wyoming Salient (Weil et al., 2010). In contrast, secondary models explain bends as thick-skinned, vertical-axis buckles that accommodate strike-parallel shortening of a preexisting orogen. Secondary bends, which are referred to as oroclines, necessarily postdate orogenesis and require a principal compressive stress orthogonal to the initial orogenic stress. Examples include the Alaskan oroclines that characterize the Cordillera of western North America (Johnston, 2001) and the Carpathian-Balkan oroclines of the Alpine orogen (Shaw and Johnston, 2012).

The collisional Variscan orogen of northern Iberia is characterized by a convex-to-the-W hairpin bend that is commonly referred to as the Cantabrian orocline (Fig. 1; Suess, 1909). Progressive and secondary models have been proposed to explain Cantabrian orocline formation. Progressive explanations include the involvement of a continental promontory (Lefort, 1979; Murphy et al., 2016) or corner (Brun and Burg, 1982); noncylindrical deformation (Pérez-Estaún et al., 1988; Martínez-Catalán, 1990); and synorogenic strike-slip shearing (Aerden, 2004; Martínez-Catalán, 2011). Secondary models include bending of a linear Variscan orogen in response to buckling of the continental core of Pangea about a pole of rotation at the western end of the Tethyan embayment (Gutiérrez-Alonso et al., 2008); an unexplained 90° rotation of the stress field (Ries and Shackleton, 1976; Van der Voo et al., 1997; Weil et al., 2001; Aerden, 2004); a clockwise rotation of Gondwana, which produced a 90° rotation of the stress field (Pastor-Galán et al., 2015); and buckling of a ribbon continent (Shaw and Johnston, 2016).

The Cantabrian orocline is cored by the Variscan E-verging, foreland fold-and-thrust belt within which a Precambrian to Carboniferous stratigraphic sequence is imbricated and folded. Thrust faults strike south and verge east within the Cantabrian orocline core. Folds include subhorizontal fault-bend folds that are inferred to reflect a stair-step, flat-ramp-flat fault geometry and a second set of folds that plunge west and are characterized by steep to vertical, E-W–striking axial planes. The second set of folds is responsible for imparting a complex, sinuous map pattern on the region. These folds potentially provide us with the means for testing the progressive and secondary models of the Cantabrian orocline. Interpretation of the E-W–trending folds as reflecting lateral ramps within the underlying thrust faults (Pérez-Estaún et al., 1988; Alvarez-Marrón, 1995) explains all the faults and folds as products of E-W convergence, consistent with progressive models of Cantabrian orocline development. Alternatively, interpretation of the E-W–trending folds as post-Variscan structures developed during strike-parallel, N-S shortening of the Variscan orogeny (e.g., Weil et al., 2013a) requires a principal compressive stress orthogonal to the Variscan stress field, consistent with models of the Cantabrian orocline as a secondary orocline.

In order to distinguish between these contrasting explanations, we undertook a detailed structural study of the Ponga Unit, a thrust nappe within the fold-and-thrust belt in the core of the Cantabrian orocline. Here, we review the geological setting of the Variscan orogen and summarize the stratigraphy and regional structure of the Ponga Unit prior to presenting our structural analysis of the study area. Our data allowed us to construct detailed down-plunge projections that we have used to constrain the geometry of the thrust sheets that characterize the Ponga Unit, and the geometry of the E-W–trending folds of the thrust sheets. Aerden (2004) interpreted the folds of the Ponga Unit to be representative of progressive changes in principal compressive stress during the Variscan orogeny. However, due to lack of observed evidence of progressive superpositions of folding during the Variscan orogeny, the high amplitude and short, regular wavelength of the E-W–trending folds require significant N-S shortening of the entire fold-and-thrust belt. Therefore, we propose that these folds formed during a post-Variscan event characterized by a principal compressive stress orthogonal to the previous Variscan stress field.


The Devonian–Carboniferous Variscan orogen of western Europe is interpreted as a product of the Pangea-forming continental collision of Gondwana and Laurussia upon closure of the intervening Rheic Ocean (e.g., Nance et al., 2010). The Variscan foreland fold-and-thrust belt is referred to as the Cantabrian zone in Iberia. The zone is characterized by a Paleozoic sedimentary succession including Cambrian to Ordovician passive margin strata, and a younger Carboniferous foreland basin sequence. The passive-margin sedimentary succession consists of carbonates and siliciclastics that are interpreted to have been deposited along the north margin of Gondwana (Shaw et al., 2014) during opening of the Rheic Ocean. The Rheic Ocean is inferred to have opened between a more northerly series of peri-Gondwanan terranes that drifted north from Gondwana, starting in the Cambrian (Nance et al., 2010), and autochthonous Gondwana to the south. The Carboniferous sequence consists of syntectonic carbonates and siliciclastic sedimentary rocks that were deposited into the Variscan foreland basin (Julivert, 1971). The fold-and-thrust belt is characterized by thrust faults that cut up section and verge to the east (Julivert, 1971). As in classic fold-and-thrust belts, the sedimentary sequences of the foreland basin strata thin toward the foreland, resulting in a décollement surface that dips west (Pérez-Estaún et al., 1994, 1995; Gallastegui et al., 1997). Synkinematic remagnetization and structural data restrict deformation in the fold-and-thrust belt to between 320 and 310 Ma (Pérez-Estaún et al., 1988; Parés et al., 1994; Van der Voo et al., 1997; Weil et al., 2000, 2001, 2010, 2013b; Weil, 2006).

The Cantabrian zone is divisible into six tectonic units that are defined by stratigraphy and structure: (1) the Narcea; (2) the Somiedo/Correcillas; (3) the Central Coal Basin; (4) the Ponga Nappe/Ponga Unit; (5) Picos de Europa; and (6) Pisuerga-Carrion units. The Ponga Unit lies east of the Central Coal Basin and west of the Pisuerga-Carrión and Picos de Europa units. It contains several thrust sheets, most notably, the Laviana, Rioseco, and Campo de Caso thrust sheets. These thrust sheets are up to 4 km thick and carry Cambrian, Ordovician, and Carboniferous strata (Julivert, 1971).


From oldest to youngest, the main Cambrian–Ordovician formations that characterize the Ponga Unit include the Láncara, Oville, and Barrios Formations. The Middle Cambrian Láncara Formation is a 50–150-m-thick carbonate sequence (Julivert, 1971). The overlying Upper Cambrian to Lower Tremadocian Oville Formation consists of shales and sandstones (Julivert, 1971), which are in turn overlain by quartzite of the Barrios Formation (hereafter referred to as the Barrios quartzite). The Barrios quartzite is locally up to 750 m thick in places and consists of pure white quartzite, with thin volcanic ash layers, beds of phyllitic siltstones, and minor shale layers (Shaw et al., 2012). The Barrios quartzite is an important stratigraphic and structural marker horizon within the study area. It is correlative with the Armorican quartzite, an extensive quartzite unit that can be traced across West Africa, Armorica, Western Europe, and as far east as Serbia (Gutiérrez-Alonso et al., 2007). In the study area, the Barrios quartzite consistently underlies the topographically highest terrain and forms near-vertical faces that give the Ponga Unit mountains their dramatic appearance.

A thin Upper Ordovician black shale unit (“Pizarras del Sueve”) and conglomerate and sandstone of the early Carboniferous Ermita Formation (∼50 m thick) are locally present above the Barrios quartzite. These two units are only preserved in the northern syncline of the Laviana thrust sheet (Gutiérrez-Marco et al., 2002, and references therein). There are no Devonian strata preserved within the study area.

The main Carboniferous foreland basin units are, in ascending order, the Alba, Barcaliente, Beleño/Fresnedo, and Redonda/Escalada Formations, and the Fito/Lena Group. The Upper Famennian–Lower Tournaisian Alba Formation (Colmenero et al., 2002) is an ∼20–40-m-thick nodular, iron-bearing, red limestone unit. Above the Alba Formation, there is the 200–500-m-thick black, azoic, Serpukhovian Barcaliente Formation limestone (also known as the Caliza de Montaña or Mountain Limestone Formation).

The overlying Moscovian Beleño Formation (named Fresnedo in the Laviana thrust sheet) is commonly over 500 m thick and consists of shales and channelized turbiditic sandstones, with thin, interbedded limestones, rooted horizons, and coal seams near the top of the unit (Colmenero et al., 2002). Above the Beleño Formation, there is the only cliff-forming Carboniferous unit, the 200–300-m-thick, Moscovian, gray to white, micritic and skeletal limestone. It is named the Redonda or Escalada Formation (in the Laviana and Campo de Caso thrust sheets, respectively; Colmenero et al., 2002).

The uppermost unit, the Moscovian Fito or Lena Group (in the Laviana and Campo de Caso thrust sheets, respectively) consists of over 1000 m of shales with interbedded limestones, sandstones, and coal seams. The Beleño Formation and Fito/Lena Group are marine basin-fill sequences that thin landward toward the east (Bahamonde, 1990; Bahamonde and Colmenero, 1993).

Our focus is on the Laviana, Rioseco, and Campo de Caso thrust sheets (Fig. 2). It is the surface traces of these thrust sheets that form a series of four folds (kilometer-scale), two of which are anticlines, and two of which are synclines. The vector of motion of the thrust sheets during emplacement, as defined by cut-off lines, fold axes, minor shear bands, and the orientation of map-scale fault-bend folds, is 090 (Alvarez Marrón, 1989, 1995). Therefore, the vector of thrust sheet emplacement lies in the E-W–striking axial plane of kilometer-scale regional folds that deform the thrust sheets. The thrust sheets are characterized by a stair-step (flat-ramp-flat) geometry (Pérez-Estaún et al., 1988) and are locally characterized by recumbent, foreland-verging folds in which the fold axes parallel the strike of the thrust faults. Alvarez-Marón and Pérez Estaún (1988) concluded, through palinspastic restoration of E-W–oriented balanced cross sections, that the minimum amount of accumulated transport by these three thrust sheets was 62 km. The Laviana, Rioseco, and Campo de Caso thrusts root to the west into a décollement located at the base of the Lower–Middle Cambrian Lancara Formation. The regional folds affecting the thrust sheets include the Beleño and Tarna synclines and the Rio Monasterio and San Isidro anticlines (Fig. 2). Siliciclastic shale and sandstone units, such as the Fito Formation and the Beleño Formation, are characterized by a steeply dipping, irregular, scaly cleavage that is commonly steep to vertical, strikes parallel to adjacent thrust faults, and is assumed to have formed during Variscan thrusting.


The majority of our data were collected across the Tarna syncline and the Rio Monasterio anticline.

Bedding and Minor Folds

Over 800 structural orientation measurements on bedding planes and minor fold axes were collected across the 400 km2 study area. In addition, we compiled existing data from regional geological maps. The majority of bedding measurements were collected from the ridge-forming Barrios quartzite, outcrops of which are commonly well bedded. Bedding data were also collected from the Escalada/Redonda limestone, the Barcaliente limestone, the Beleño Formation, and the Fito/Lena Group. The incompetent and readily deformable shales and sandstones of the Beleño Formation rarely preserve primary bedding structures; however, this unit was host to many (∼42%, n = 18) of the measurable minor, centimeter- to decimeter-scale fold axes. The minor folds resulted from folding of the steeply dipping scaly cleavage. The Fito/Lena Group, which is similarly incompetent, hosted ∼33% (n = 15) of the minor folds of the scaly cleavage. Neither the massive limestone units nor the Barrios quartzite exhibited evidence of pervasive deformation, which we attribute to their rheological competency.

To facilitate our structural analysis, we defined six structural domains. Domain boundaries included major thrust faults, an E-W–striking line across the inflection point between the Rio Monasterio anticline and the Tarna syncline, and a late strike-slip fault that cuts across the map area (Fig. 3). As indicated by the map pattern and by the great circle distribution of the poles to bedding in each of the six domains, the thrust sheets are folded. The folds plunge westward, and the magnitude of the plunge decreases regularly from west to east. Cylindrical best-fit analyses for bedding data in the six domains are displayed in Figure 3 and Table 1.

The plunge of W-plunging fold axes does not vary more than 1° within thrust sheets. A cylindrical best-fit analysis for bedding data showed: the Laviana thrust sheet fold axis plunging 60° toward 266°; the Rioseco thrust sheet fold axis plunging 41° toward 267°; and the Campo de Caso thrust sheet fold axis plunging 33° toward 264°. A stereonet plot of poles to bedding for the entire study area yielded a regional fold axis plunging 48° toward 265°.

Analysis of the minor fold axes, measured in the scaly cleavage, yielded a high concentration (31%) of hinge lines measured on minor folds that plunged over 60° (Fig. 4). A comparison of the orientation of minor fold axes with regional folds showed that steeply plunging fold axes are pervasive in the study area. The variation in fold axes in the scaly cleavage is influenced by three factors: (1) Variscan folding; (2) vertical-axis rotation during oroclinal buckling; and (3) flow of less-competent units toward the noses of regional folds.

Down-Plunge Projections

Surface map patterns of deformed terranes provide a distorted view of deformation patterns within the subsurface (Johnston, 1999). In our map area, where structures are plunging and topography represents less than 5% of the depth to which structures extend into the subsurface, it is appropriate to create down-plunge projections in order to view structures in profile. By constructing down-plunge projections of the folds that affect the thrust sheets, we are able to view the folds in profile, which allows us to better constrain fold geometry. Because the folds in the thrust sheets are not cylindrical (i.e., the fold axes steepen toward the west and shallow with depth), we divided the map area into six smaller domains that we could treat as cylindrical. We then created down-plunge projections for the Laviana, Rioseco, and Campo de Caso thrust sheet (Figs. 5, 6, 7, and 8) by stitching together down-plunge projections that were created for each of the six domains. The down-plunge projections were created according to the method defined by Johnston (1999). Using down-plunge projections of the steep, westward-trending regional folds, we were able to calculate the amount of N-S shortening that each thrust sheet accommodated. Shortening was calculated by measuring the distance across the folds from the hinge of the Beleño syncline to the hinge of the San Isidro anticline in each thrust sheet and comparing it with the measured restored length of the thrust sheets prior to folding. The length across the folds in the Laviana thrust sheet after folding is 10.6 km, and the length of the thrust sheet prior to folding is 18.8 km. By dividing the difference of these measurements by the length of the restored thrust sheet, we calculate that the Laviana thrust sheet was shortened by 44%. We used the same technique to calculate the shortening in the Rioseco and Campo de Caso thrust sheets. The distance across the folds of the Rioseco thrust sheet is 9 km, the restored length of the sheet is 21 km, and the shortening is 57%. The distance across the folds of the Campo de Caso thrust sheet is 8.8 km, the restored length of the sheet is 18.4 km, and the shortening is 52%. We also used the down-plunge projections to measure the wavelength and amplitude of the folds. The wavelength and amplitude of the folds in the Laviana thrust sheet are 12 km and 3.5 km, respectively. The wavelength and amplitude of the folds in the Rioseco thrust sheet are 11.6 km and 4.4 km, respectively. We also observe that the folds become asymmetric in the younger thrust sheets. Furthermore, the northward-dipping axial planes of folds that shallow with depth indicate northward vergence of the folds.

Vertical Cross Sections

We also created a balanced (as per Dahlstrom, 1969) cross section (Fig. 2) that contains the average vector of thrust sheet emplacement, trending 090 degrees, in order to analyze the amount of shortening attributable to Variscan thrusting (Fig. 9). This palinspastic cross section indicates >16 km of shortening was taken up just by the Laviana and Rioseco thrust sheets, and this is consistent with previous estimates of ∼60 km of shortening (Alvarez-Marón and Pérez Estaún, 1988) accommodated by the fold-and-thrust belt in total. In addition, the cross section shows that, along the axial plane of the Tarna syncline, the Rioseco thrust merges with the Laviana at a depth of ∼4 km, implying 6 km of structural relief relative to the San Isidro anticline to the south and 9 km with respect to the Rio Monasterio anticline to the north, which is consistent with our down-plunge projections (Figs. 5, 6, and 7).


West-plunging folds of thrust faults and related fault-bend folds within the E-verging Cantabrian zone thrust belt of the Variscan orogen have been attributed to lateral ramps in underlying thrust faults (Alvarez-Marron and Perez-Estaun, 1988; Alvarez-Marrón, 1995). An alternative interpretation is that the folds are post-Variscan structures that overprinted the thrust belt. Our structural analysis of thrusts and folds in the Ponga Unit and detailed down-plunge projections place constraints on the formation mechanism of the W-plunging folds, as discussed further here. We start with a brief review of the typical geometry of lateral ramp–related folds and then demonstrate how the geometry of the W-plunging folds is inconsistent with this model and instead requires superposed folding of the thrust belt. Finally, we comment on the significance of our findings for models of the hairpin Cantabrian orocline in the Variscan orogen of northern Iberia.

Lateral ramp–related folds (Fig. 10): (1) have axes that trend perpendicular to the thrust traces and plunge toward the hinterland; (2) are typically open and homoclinal with upright limbs and verge along strike toward the thrust tips; (3) tend to converge into parallelism with fault-bend folds that are attributable to frontal ramps, due to linking of lateral and frontal ramps through oblique structures; (4) are irregularly developed but are more commonly developed away from the center of the thrust sheet; (5) are restricted in relief to be less than or equal to the height of the lateral ramps, restricting them to a maximum of a few kilometers of structural relief; and (6) typically result in minor to insignificant (no more than a few percent) orogen-parallel shortening of the thrust sheet. Classic examples of thrust sheets characterized by lateral ramp–related folds include the Moine thrust belt (Boyer and Elliott, 1982) and the Rundle thrust (Wilkerson et al., 2002), among others.

Consistent with their interpretation as lateral ramp–related folds (Pérez-Estaún et al., 1988; Alvarez-Marrón, 1995), the W-plunging folds of the Ponga Unit trend perpendicular to the regional strike of the Ponga Unit thrust faults. They are, however, otherwise unlike lateral ramp–related folds. The folds plunge consistently westward without variation in plunge direction; oblique ramps are absent, and nowhere are the W-plunging folds observed to merge via oblique structures into fault-bend folds attributable to frontal ramps. The W-plunging folds are symmetric, not homoclinal, and they are characterized by a regular, predictable pattern with consistent wavelengths and amplitudes throughout. Fold amplitudes average 3.5 km, and hence structural relief averages ∼7 km. The structural relief exceeds the thickness of any of the Ponga Unit thrust sheets (Fig. 9), and hence it exceeds the maximum height of any possible Ponga Unit lateral ramp. Far from being open, the W-plunging folds are tight and locally isoclinal and have steeply dipping to locally overturned limbs. Their axial planes are vertical to steeply S-dipping, yielding a consistent N-verging geometry. Steeply plunging to vertical folds of the scaly Variscan cleavage developed in the siliciclastic units demonstrate that folding was post-Variscan and imply that the W-plunging folds were generated during buckling of the thrust belt about a vertical axis of rotation.

Interpretation of the W-plunging folds as being superposed on the main post-Variscan thrusts implies that there were two distinct orogenic events. A testable prediction of this model is that folding of the Variscan fold-and-thrust belt should have yielded a fold interference pattern (e.g., Ramsay and Huber, 1987). Indeed, the Ponga Unit map pattern resembles a classic mushroom-type fold interference pattern (Fig. 11). Geometric modeling shows that mushroom-type fold interference patterns result from successive deformation events characterized by principal compressive stresses at high angle or even perpendicular to each other (Ramsay type 2; Fig. 11C). Recognition of the Ponga Unit mushroom-type fold interference pattern is attributable to Julivert and Marcos (1973). The mushroom-type interference pattern represents a positive test for a model of the W-plunging folds as products of orogen-parallel principal compressive stress after Variscan fold-and-thrust belt development. An additional argument in favor of superposed folding is that our data show a progressive steepening of fold axes toward the west (Fig. 3), which is predicted in a typical fold-and-thrust belt where folds and bedding progressively steepen toward the metamorphic hinterland. A superposed perpendicular shortening event will create a second generation of folds that also progressively steepen toward the hinterland, which is observed in the field area. Oroclinal buckling resulted in local thrust reactivation and the development of numerous brittle and brittle-ductile strike-slip faults with complex movement histories within the orocline hinge region. These steeply dipping faults accommodated tangential shortening during oroclinal buckling (Weil et al., 2013a; Gutiérrez-Alonso et al., 2015).

Paleomagnetic data have been used to argue that the W-plunging folds are at least in part attributable to lateral ramps. Declination data from a secondary “B” remanence (Weil et al., 2013a) show deflections from a regional magnetic trend that were attributed to a postremanence, but pre-oroclinal folding that Weil (2006) interpreted as Variscan deformation related to lateral and oblique ramps in the Ponga Unit thrust faults. Our structural study shows that the W-plunging folds are unlikely to be related to lateral ramps and are instead better interpreted as the result of post-Variscan deformation, and this suggests that an alternative interpretation of the paleomagnetic data should be entertained.

The age of the secondary “B” remanence is “latest Stephanian to Early Permian, after initial D1 thrusting but prior to major secondary rotation” (Weil, 2006, p.1). However, the constraints on the timing of oroclinal bending are commonly given as being Stephanian to Early Permian (see, for example, Gutiérrez-Alonso et al., 2012; Weil et al., 2013b, their fig. 7). Given these constraints, there is only a small window between the end of Variscan thrusting and the onset of oroclinal buckling within which we have to place the secondary “B” remanence. We suggest that instead of overlapping with and recording deformation attributable to Variscan thrusting (as per Weil, 2006), the “B” remanence in the Ponga Unit likely postdates Variscan deformation and overlaps with and records initial deformation attributable to oroclinal buckling. Our interpretation is consistent with the “B” remanence being an entirely post-Variscan remanence, as it is to the west of the Ponga Unit (Weil, 2006); this interpretation is compatible with the available age constraints, which require that orocline formation began almost immediately after the cessation of Variscan deformation. Weil (2016, personal commun.) has suggested that the Ponga Unit magnetization is not the “B” magnetization recorded west of the study area, but it is instead a “B2” magnetization that was acquired later, during early oroclinal bending. Syndeformation magnetization can be protracted and have systematic spatial acquisition, with more recent magnetization acquisition the farther one progresses into the foreland (e.g., Enkin et al., 1997). Interpretation of the onset of Cantabrian orocline formation as being immediately post-Variscan and overlapping with a “B2” magnetization within the Variscan foreland reconciles structural and paleomagnetic data within the Ponga Unit, confirms the latest Stephanian to earliest Permian onset of oroclinal buckling, and implies that the switch from Variscan deformation to oroclinal buckling was geologically instantaneous.

A secondary orocline model for the Cantabrian orocline requires a post-Variscan change in the orientation of the principal compressive stress from orogen-perpendicular to orogen-parallel orientation (Gutiérrez-Alonso et al., 2012; Johnston et al., 2013; Weil et al., 2013b). Our analysis shows that the W-plunging folds of the Ponga Unit affected the Variscan thrust faults and formed in response to an orogen-parallel principal compressive stress. Either the stress field rotated 90° about a vertical axis of rotation, or the maximum and intermediate stress axes, both of which lay in the horizontal plane, switched places. Either way, the principal compressive stress switched from an orogen-perpendicular orientation to an orogen-parallel orientation (Pastor-Galán et al., 2015). Our findings are congruent with the Cantabrian orocline being a true orocline formed by buckling of preexisting linear structures. Furthermore, paleomagnetic data record the onset of orocline-related folding immediately after the cessation of Variscan orogenesis, providing only a small window of time between the end of Variscan orogenesis and the beginning of oroclinal buckling (Johnston et al., 2013).


We investigated the complex fold pattern of the Ponga Unit in the hinge of the Cantabrian orocline in order to test progressive versus secondary models of orocline formation. Our data show that the observed map pattern is attributable to post-Variscan folding of the Cantabrian fold-and-thrust belt in response to an orogen-parallel principal compressive stress. The post-Variscan deformation that we describe is consistent with the predications of models of the Cantabrian orocline as a secondary orocline. Timing constraints provided by paleomagnetic data indicate that orocline formation began immediately after the cessation of Variscan deformation.

Research was supported by a University of Victoria Graduate Studies Award and a Natural Sciences and Engineering Research Council of Canada (NSERC) Discovery Grant to Stephen T. Johnston. Arlo Weil is thanked for helping us understand and interpret the paleomagnetic data from the Ponga Unit. Theron Finley is thanked for his tireless efforts as a field assistant. Insightful and constructive reviews provided by Domingo Aerden and an anonymous reviewer significantly improved the manuscript.

Stephen Johnston