Abstract
The Etive Dyke Swarm is the largest Caledonian dyke swarm in Britain and Ireland. Field and petrophysical (AMS–anisotropy of magnetic susceptibility) data are presented that indicate the dykes resulted from passive upwelling of magmas into fractures created by regional tectonic stresses. During the period of dyke intrusion (around 415 Ma) periodic episodes of sinistral transpression affected the region and were probably associated with sinistral movement on the Great Glen and other regional faults that lie subparallel (approximately 5° clockwise) from the long axis of the swarm. AMS data show that magma upwelled subvertically across the swarm and suggests that the dykes fed a volcanic field much larger than the now preserved Glencoe Volcano complex.
Dykes provide a means for transport and emplacement of igneous rocks within the Earth's crust, forming feeders to both volcanic and intrusive bodies. In many complexes, dykes form a significant percentage of the subvolcanic crust and can, in some cases, constitute 100% of a granitoid pluton. The process of dyke formation is therefore important in the formation and evolution of many igneous complexes. In this paper we use field and petrophysical observations from the late Caledonian Etive Dyke Swarm to examine the formation of the fractures that host the dykes, and the flow of magma within these fractures.
The Etive Dyke Swarm (Fig. 1), southwestern Grampian Highlands of Scotland, is the largest Caledonian dyke swarm in Britain and Ireland. It covers an area in excess of 100×20 km in a NNE–SSW-trending zone centred on the Etive Igneous Complex and the associated Glencoe Volcano (e.g. Bailey 1960) (Fig. 1). Dyke orientation is predominantly swarm-parallel, and the swarm as a whole lies subparallel (oriented approximately 5° counter-clockwise) to major regional faults such as the Great Glen Fault to the NW and the Ericht-Laidon and Tyndrum faults to the SE. Dykes can account for up to 42% extension within a single outcrop (Clough et al. 1909), although the overall estimate of 10–15% extension, orthogonal to dyke orientation, within the Glencoe area (Kokelaar & Moore 2006) is probably more representative of the swarm as a whole.
Location of the Etive Dyke Swarm relative to regional faults and the Etive Igneous Complex (dark grey). Adapted from Bailey (1960). Insert (top left): location of the study area within Scotland.
Bailey (1960) classifies the rocks of the Etive Dyke Swarm as microdiorites, porphyrites and quartz-porphyries. Our observations show that there is no clear division between the three rock types and that, both petrologically and chemically, the dykes form a continuum. Composition ranges from basalt to rhyolite (SiO2=51–77 wt%, MgO=6–0 wt%), with the intermediate to more evolved dykes being more numerous. The basalts and basaltic andesites are fine-grained and equigranular (corresponding to the microdiorites of Bailey). Hornblende is the main ferromagnesian phase, with pyroxene the dominant phase in rare examples. These dykes are generally confined to the centre of the swarm in the Glencoe and Glen Etive areas and are generally narrower (0.1–2.0 m) than the more evolved dykes. Andesitic–rhyolitic dykes are porphyritic, with euhedral plagioclase phenocrysts between 1 and (rarely) 10 mm in size. Plagioclase shows complex cyclic zoning, with multiple resorption surfaces within crystals, and range from calcic oligoclase to calcic andesine (An25–50) with the majority being sodic andesine (An30–40). The major ferromagnesian phase is hornblende, with some samples containing significant quantities of biotite. These dykes correspond to Bailey's porphyrites and are found throughout the swarm. The most evolved dykes in the Etive Dyke Swarm contain phenocrysts of quartz and albite pure anorthoclase (Or00–25). These correspond to the quartz-porphyries of Bailey and are generally restricted towards the centre of the swarm.
Ample evidence exists of the co-existence, mingling and mixing of two or more magmas. Disequilibrium assemblages of pyroxene and amphibole phenocrysts, with pyroxenes showing reacted and resorbed crystal margins, are observed in many of the basaltic–dacitic dykes. Composite dykes with sharp internal contacts separating an internal mafic phase from a silicic external phase are also observed.
Biotites from two dykes were recently dated using the laser stepwise heating 40Ar–39Ar technique at 415±1.8 and 414±2 Ma (Morris et al. 2005). While these dates do not necessarily contradict an earlier published date of 411±5.1 Ma by Rb–Sr whole rock–plagioclase–bioite (Thirlwall 1988), the age error on the latter date cannot be ignored and it is feasible that the dyke swarm was longer lived than indicated by the small range on the 40Ar–39Ar data.
Dyke-wall morphology and fracture formation
The assumption that fractures form perpendicular to the least compressive stress orientation (Anderson 1938) has been shown to be a good basis for examining the morphology of fracture-hosted geological features such as hydrothermal veins and dykes. Fractures can propagate as a series of fingers or segments, ahead of the main fracture, that are rotated out of the main plane of that fracture, appearing en echelon in cross-section. These later join by breaking of bridges between segments to form the main fracture (Pollard 1987). The orientation of these bridges is perpendicular to the propagation direction of the fracture, a feature used by some workers to examine the orientation of the opening of horizontal fractures that host sills (e.g. Hutton et al. 2006). In the Etive Dyke Swarm observed bridges are always horizontal, indicating vertical fracture propagation.
The perturbation of the stress field at the propagating tips of fractures can give clues as to the nature, and relative dominance, of the stresses that control their formation (e.g. Nicholson & Pollard 1985; Delaney et al. 1986; Pollard 1987). The interaction of the propagating tips of en echelon segments is controlled by the relative proportions of near-field (magma pressure) and far-field (tectonic) stresses. Nicholson & Pollard (1985) established two end-member morphological situations (Fig. 2). In the first end member, the propagation paths of fracture tips are straight even when en echelon segments overlap. Dilation of the fracture is accommodated by mechanical bending of bridges until the bridge fails, thereby joining the segments to form the main fracture (Fig. 2a). This end member occurs when far-field (i.e. regional or tectonic) stresses are dominant and the orientation of the least compressive stress remains perpendicular to the fractures as a whole. In the second end member, local stresses at the propagating fracture tips induced by, for example, fluid or magma pressure deflect the propagation of the fracture tip, resulting in curved propagation paths where segments overlap (Fig. 2b). Crack walls then consist of two curved portions near the tips and a linear central portion. Initial dilation of the fractures is accommodated by rotation of the bridges until one crack tip intersects the planar portion of the next fracture segment.
Schematic illustration of en echelon crack growth and dilation under: (a) far-field (tectonic) stress-dominated conditions; and (b) near-field (e.g. magma or hydrothermal fluid pressure) stress. Adapted from Nicholson & Pollard (1985).
By matching re-entrants on opposing dyke walls, we observe that the majority of Etive dykes dilated perpendicular to their long axis. Bridge and broken bridge structures observed show evidence of mechanical bending of bridges, formation of incipient fractures and eventual failure as dyke segments unite to form the main dyke (Figs 3 and 4). All observed fracture propagation paths were linear, no evidence of curved propagation paths (e.g. Fig. 2b) was observed suggesting that far-field stresses dominated the formation of the Etive Dyke Swarm. While these features are most apparent in basaltic and basaltic andesite dykes, they can be observed in dykes of all rock types in the Etive swarm.
Examples of Etive dykes filling fractures that have grown along straight propagation paths, then jointed by bridge bending and eventual failure. Heavy arrows indicate the direction of fracture dilation. Compare the left-hand photograph with Figure 2a.
Field sketches showing four examples of Etive dykes where fracture formation was dominated by far-field stresses. On the compass rose, the double-headed arrow indicates the orientation of the dyke while the grey arrow shows the direction of dyke dilation.
As has previously been reported (Morris & Hutton 1993), a number of dykes show evidence of oblique opening. Evidence includes offset matching re-entrants on dyke walls (Fig. 5), stepping of dyke orientation (Figs 5 and 6a & b) and asymmetric dyke terminations (Fig. 6a–c). Offset is consistently sinistral and, while such features are only observed in a small percentage of dykes, these dykes cover the entire compositional spectrum of the Etive swarm. Evidence for sinistral shear can also be observed at a larger scale were both individual dykes and portions of the swarm as a whole show counter-clockwise rotation in orientation before returning to swarm-parallel (Morris & Hutton 1993).
A basaltic andesite Etive dyke showing oblique opening (double-ended arrows) and counter-clockwise rotation of the dyke orientation away from swarm-parallel. These features are interpreted as resulting from sinistral transpression during dyke dilation.
Field sketches of dykes showing evidence of oblique opening. On the compass rose, the double-headed arrow indicates the orientation of the dyke while the grey arrow shows the direction of dyke dilation.
Magma flow orientation by AMS
The magnetic fabric of an igneous rock, studied with the anisotropy of magnetic susceptibility (AMS), is considered to be a good indicator of magmatic flow direction in dykes and sills (e.g. Ernst & Baragar 1992; Knight & Walker 1988; Callot et al. 2001; Liss et al. 2002). Magnetic susceptibility is reported as three eigen vectors, Kmax, Kint and Kmin. The intensity, inclination and declination of each vector is measured from a number of subsamples and averaged for each sample. A mean magnetic susceptibility, Km, is also reported for each sample. The degree of anisotropy, the deviation of susceptibility between Kmax and Kmin is expressed by P (P=Kmax/Kmin). R describes the degree of clustering of Kmax data from subsamples about the average Kmax for each sample, with ratios near 0% indicating randomly distributed data and high percentages indicating good clustering of data (Wallbrecher et al. 1986).
Oriented samples were collected from 28 localities across the Etive Dyke Swarm. Samples were reoriented in the laboratory and cored, with an average of 10 subsamples being obtained from each. The bulk susceptibility was measured using a Bartington MS2B, while the anisotropy of magnetic susceptibility was determined with a Molspin Minisep, both in the petrophysical laboratory of the Department of Mineralogy, University of Geneva. Results, averaged for each sample site, are presented in Table 1.
Anisotropy of magnetic susceptibility (AMS) data
The generally strong mean susceptibility (Km) of samples from the Etive Dyke Swarm indicates that the magnetic signal is dominated by ferrimagnetic minerals such as magnetite and ilmenite, an observation confirmed by thin-section examination. Despite this, the degree of anisotropy is low–moderate (P=1.02–1.11, average 1.04). The clustering of Kmax data is, however, generally high (average R=88.4%), indicating a strong linearity to the AMS data within individual samples and suggesting that the low degree of anisotropy is owing to the shape of individual mineral grains.
Three distinct magnetic fabrics were identified in the Etive Dyke Swarm and are illustrated in Figure 7. Type A has Kmax vectors grouped nearly parallel to the plane of the dyke and is found at 18 (64%) of the sites examined. This type of fabric is normally interpreted as resulting from the alignment of the long axis of prolate minerals, such as feldspar, parallel to magma flow and is therefore taken to represent the vector of magma flow within the dyke (e.g. Knight & Walker 1988; Callot et al. 2001). Type B is characterized by the distribution of Kmax and Kint vectors along a plane, and corresponds to oblate minerals such as micas. While the magnetic fabric still lies parallel to the dyke walls, no unique long axis can be defined and these samples are not used for interpreting magma flow directions. Type B fabrics are the rarest in the Etive Dyke Swarm, occurring at only two sites (7% of samples). Type C fabric is defined by strongly clustered Kmax data perpendicular to the dyke wall. There are several possible interpretations for such a fabric. Some minerals, e.g. hornblende, can give an inverse signal, i.e. Kmin would lie parallel to the long axis of the mineral and therefore the direction of magma flow. This interpretation could be valid for samples MK01-E16 and MK01-E53, where the weaker mean susceptibility (Km=3.44×10−3 and 3.25×10−4) suggests that minerals with a lower susceptibility, such as hornblende, dominate the magnetic signal. Other samples with Type C fabrics show strong Km, similar to Type A samples, and, as such, may result from processes other than magma flow. Owing to the ambiguity of interpretation, samples with Type C fabrics were not used in the magma flow interpretation. Four sites (14%) showed Type C fabrics. In addition, four sites (14%) showed fabrics intermediate between two of the above fabrics and were not used in magma flow interpretations. There is no consistant spatial distribution of the three magnetic fabrics (Fig. 8).
Magmatic fabrics recognized in the Etive Dyke Swarm. Type A shows a prolate fabric with strongly clustered Kmax and Kint parallel to, and Kmin perpendicular to, the dyke plane. This fabric is normally used to interpret magma flow direction parallel to Kmax. Type B shows an oblate fabric with Kmax and Kmin distributed along a plane. With no clear lineation this fabric cannot be used to interpret a magma flow direction. Type C shows a prolate fabric with Kmax perpendicular to the dyke plane. Owing to several possible interpretations of such a fabric, this is not used in magma flow direction interpretations.
Lower-hemisphere projections of Type A mean Kmax (axis of maximum magnetic susceptibility interpreted here as the magma flow fabric). Stereonets with a white background show Type A magnetic fabric, utilized in magmatic flow interpretations; while stereonets with a grey background indicate Type B and Type C magnetic fabric, which, owing to potentially ambiguous interpretations, were not utilized in magmatic flow fabric discussions. Base map adapted from Bailey (1960).
Type A AMS data from the Etive dykes do not show a consistent inclination or declination in any part of the swarm. Rather, magma flow directions interpreted from Kmax data suggest that magma upwelled in a near-vertical orientation over a broad area and spread into the fractures that host the dykes (Fig. 8).
Discussion and conclusions
The spatial and temporal relationship between the Etive Dyke Swarm and the Great Glen Fault has already been pointed out (Morris & Hutton 1993) (Fig. 1). The subparallel nature of the swarm is in itself suggestive of a tectonic rather than magmatic control on the overall distribution of dykes. Morphology of bridges and broken bridges between en echelon segments of dykes supports the conclusion that far-field (i.e. tectonic) stresses dominated the formation of the fracture systems exploited by magmas.
While the majority of dykes show dyke orthogonal opening (Figs 3 and 4), counter-clockwise rotation of dyke orientation coupled with sinistral displacement of dyke wall features on a number of dykes show that transient periods of sinistral shear occurred during the life of the system (Figs 5 and 6) (Morris & Hutton 1993). This would suggest that this region of the British Caledonides was undergoing periodic phases of sinistral transpression in a NNE–SSW orientation during the late Silurian.
AMS studies show that the Type A Kmax orientations within the dykes were generally near-vertical, suggesting that magma flowed up the fracture system that now forms the Etive swarm over a broad area rather than originating from a single-point source such as the Glencoe Volcano. The Etive Dyke swarm could therefore represent the roots of a much wider volcanic field than now preserved in the caldera of the Glencoe Volcano. Fracture morphology shows a dominance of far-field (tectonic) rather than near-field (magma pressure) stresses in fracture formation. The emplacement of the Etive Dyke Swarm is therefore thought to be a passive response to the opening of a series of transtensional fractures created (or possibly reactivated) by sinistral movement of the Great Glen and parallel fault systems.
Acknowledgments
Fieldwork associated with this project was supported by a grant from the Fonds Marc Birkjit to G. A. Morris. We acknowledge the invaluable assistance of Dr I. Hedley, University of Geneva, in obtaining the AMS data presented here. Comments by J. Reavy and C. Kirkland are gratefully acknowledged.
- © The Geological Society of London 2008
References
Emplacement of the Etive Dyke Swarm, Scotland: implications of dyke morphology and AMS data
