Abstract

The Mourne Mountains magmatic center in Northern Ireland consists of five successively intruded granites emplaced in the upper crust. The Mourne granite pluton has classically been viewed as a type locality of a magma body emplaced by cauldron subsidence. Cauldron subsidence makes space for magma through the emplacement of ring dikes and floor subsidence. However, the Mourne granites were more recently re-interpreted as laccoliths and bysmaliths. Laccolith intrusions form by inflation and dome their host rock. Here we perform a detailed study of the deformation in the host rock to the Mourne granite pluton in order to test its emplacement mechanism. We use the host-rock fracture pattern as a passive marker and microstructures in the contact-metamorphic aureole to constrain large-scale magma emplacement-related deformation. The dip and azimuth of the fractures are very consistent on the roof of the intrusion and can be separated into four steeply inclined sets dominantly striking SE, S, NE, and E, which rules out pluton-wide doming. In contrast, fracture orientations in the northeastern wall to the granites suggest shear parallel to the contact. Additionally, contact-metamorphic segregations along the northeastern contact are brecciated. Based on the host-rock fracture pattern, the contact aureole deformation, and the north-eastward–inclined granite-granite contacts, we propose that mechanisms involving either asymmetric “trap-door” floor subsidence or laccolith and bysmalith intrusion along an inclined or curved floor accommodated the emplacement of the granites and led to deflection of the northeastern wall of the intrusion.

1. INTRODUCTION

The emplacement of large volumes of granitic magma in Earth’s crust is a long-standing problem that has led to much debate since the Earth sciences emerged as a modern scientific discipline (e.g., Cloos, 1922; Read, 1957; Hutton et al., 1990; Pitcher, 1997; Tikoff et al., 1999; Petford et al., 2000; Burchardt, 2018). This so-called “space problem” (e.g., Daly, 1903; Bowen, 1948; Read, 1957; Buddington, 1959) is not only of significance for the question of how the continental crust formed, but is also of importance for understanding hazards related to the emplacement of magma bodies under active volcanoes and the formation of geothermal heat sources and ore deposits (e.g., Richards, 2003; Kennedy et al., 2012; van Wyk de Vries et al., 2014; Kennedy et al., 2018). A better understanding of the mechanisms of granitic magma emplacement can, for example, help decipher the relationship between magma movement in the shallow crust and ground deformation at active volcanoes.

The granites of the Mourne Mountains in the northeastern part of the island of Ireland represent one of the final iterations of the British and Irish Paleogene Igneous Province (Fig. 1A) (Thompson et al., 1987; Gibson et al., 1987, 1995; Gamble et al., 1999; Wilkinson et al., 2017). The Mourne Mountains magmatic center consists of five granites emplaced into Silurian Hawick Group sediments of the Caledonian Southern Uplands–Down-Longford terrane (Anderson, 2000, 2004; Cooper and Johnston, 2004) (Figs. 1 and 2). However, Hood (1981) noted that it is possible to divide the pluton into more units. The study of the Mourne granite pluton by Richey (1927) introduced the idea that pluton formation can be incremental and popularized the magma emplacement mechanism of cauldron subsidence (Fig. 3A). Cauldron subsidence encompasses subsidence of a block of country rock along ring faults within the crust, while magma is transferred along the ring-fault system into the space created above the subsiding block (Clough et al., 1909). Richey (1927) based his interpretation of the emplacement mechanism on the observations that the host rock close to the Mourne granite pluton did not display distinct deformation related to granite emplacement and that the intrusion contacts are mainly discordant to the host-rock bedding (i.e., the granites truncate the bedding). Thus, doming of the roof was viewed as unlikely.

The host rock to the Mourne granites has experienced multi-phase deformation during the Caledonian orogeny (Anderson, 1987, 2000) and is generally not well exposed in the Mourne Mountains. Consequently, there has been some debate about the mechanism of granite emplacement. Several authors pointed out caveats in Richey’s (1927) model after remapping the granites and the contacts to and within the intrusion, which led to refinements of the cauldron subsidence emplacement model for the Mourne granite pluton (Emeleus, 1955; Robbie, 1955; Hood, 1981; Gibson, 1984; Meighan et al., 1984). However, other models for the emplacement of the Mourne granites have also been presented. Rohleder (1931), for example, interpreted the granites as discordant laccoliths. Laccolith intrusions form in the shallow crust by deforming and uplifting their host rock (Gilbert, 1877; Johnson and Pollard, 1973). Walker (1975) proposed that the Mourne granites were fed laterally from a source in the NE into space created by an internally sagging dense mafic body. This process evokes subsidence of the underlying material, but not along ring faults. Two more recent anisotropy of magnetic susceptibility (AMS) studies on the Mourne granite intrusions by Stevenson et al. (2007) and Stevenson and Bennett (2011) interpreted that the granites of the eastern part of the pluton were emplaced both as laccoliths and bysmaliths, while the granites intruded as laccoliths in the western part of the pluton (Fig. 3B). They based their interpretation on a subhorizontal magnetic fabric in the center of the intrusion, magnetic fabrics with shallow dips at the contacts in the eastern part of the pluton, the divergent strikes of the Southern Uplands–Down-Longford terrane bedding along the granite pluton’s northeastern boundary (Fig. 2A) (Hood, 1981), and remapping of granites (Meighan et al., 1984).

Because of the preexisting deformation from the Caledonian orogeny (e.g., Anderson, 1987, 2000) and sparse exposures, the deformation of the host rock associated with the emplacement of the Mourne granite pluton remains poorly understood. In this study, we investigate the emplacement mechanism of the Mourne granite pluton using fractures in the granites and the fracture patterns in the host rock as proxies for deformation. Our approach includes fracture data rotation analysis based on the dip and strike of the contact to shed more light on granite emplacement from the perspective of host-rock deformation. We also provide observations on the contact aureole to the Mourne granite pluton in order to constrain the timing of deformation relative to granite emplacement. By combining previous research on the Mourne granite pluton with our new data, we discuss models for the emplacement of the granites.

Our analyses reveal that the intrusion roof does not display pluton-wide doming, while the northeastern wall of the intrusion records significant deformation, which suggests shear parallel to the strike of the contact. We propose that the Mourne granite pluton was emplaced by asymmetric floor subsidence or laccolith intrusion along a curved floor, which could explain the lack of doming of the roof.

1.1. Laccoliths

Laccolith intrusion involves making space for magma by pushing upwards and deforming the host and the magma body (e.g., Gilbert, 1877; Hawkes and Hawkes, 1933; Pollard and Johnson, 1973; Hutton, 1988). Laccolith growth can be divided into three stages: (1) Magma intrudes and forms a sill, followed by (2) inflation of the sill to form a laccolith, and finally (3) the bysmalith stage where the host rock is uplifted along faults (de Saint-Blanquat et al., 2006). The growth of the magma body may terminate at any of the above-mentioned stages yielding different types of magmatic intrusions (namely sills, laccoliths, and bysmaliths). Deformation of the host rock associated with laccolith growth involves rotation and uplift, faulting, and minor bending (de Saint-Blanquat et al., 2006; Morgan et al., 2008; van Wyk de Vries et al., 2014; Wilson et al., 2016; Mattsson et al., 2018). In case of the inflated-sill laccolith type, the host rock should largely drape the dome-shaped laccolith (Gilbert, 1877; Hawkes and Hawkes, 1933). In contrast, the host rock to stacked sill- and bysmalith-type laccoliths should overall retain its original inclination except at the edges of the laccolith where it should be rotated and bent and/or faulted (Corry, 1988; de Saint-Blanquat et al., 2006; Morgan et al., 2008; Wilson et al., 2016). Even if a laccolith has a flat roof and the host was uplifted in a piston-like fashion, some discernible deformation should occur at the flanks of the intrusion.

1.2. Cauldron Subsidence

Cauldron subsidence was first described by Clough et al. (1909) for Glen Coe, Scotland, and is a type of subterranean caldera collapse (Richey, 1927; Paterson and Fowler, 1993; Cruden and McCaffrey, 2001; Yoshinobu et al., 2003; Burchardt et al., 2012). Cauldron subsidence can be grouped into magma emplacement mechanisms that involve floor subsidence (e.g., Cruden, 1998; Cruden and McCaffrey, 2001). Here, cauldron subsidence is distinguished as floor subsidence facilitated by faults that are related to magmatic activity such as outward inclined ring faults with a bell-jar geometry (cf. Anderson, 1937), as opposed to floor subsidence that is controlled by regional faults (e.g., Holohan et al., 2005). The mechanisms causing cauldron subsidence have been likened with caldera collapse, in which removal of magma from an underlying reservoir induces subsidence at Earth’s surface (e.g., Cruden, 1998; Cruden and McCaffrey, 2001; Branney and Acocella, 2015).

2. GEOLOGICAL SETTING

The following section summarizes (1) the deformation of the Hawick Group that occurred during the Caledonian orogeny and in the Paleogene and (2) previous research on the Mourne granite pluton.

2.1. Southern Uplands–Down-Longford Terrane

The Southern Uplands–Down-Longford terrane trends ENE-WSW through southern Scotland, Northern Ireland, and Ireland and is made up of a succession of isoclinally folded turbidite tracts or blocks bounded by major thrust faults (Anderson and Cameron, 1979; Anderson, 2000). The terrane youngs toward the SE and has been interpreted as a relict accretionary prism that formed during the Caledonian orogeny (Anderson, 2004). The Hawick Group is situated at the front of that accretionary prism and consists of a fine-grained graywacke with interbedded sandstone, siltstone, and dark shale layers (Anderson, 2004). The graywacke consists of quartz, chlorite, epidote, minor amounts of iron oxides, sericite, and altered feldspars (Emeleus, 1955). Carbonates are also abundant in the groundmass, as well as in concretions and nodules in Hawick Group rocks (up to 20 vol%, Emeleus, 1955; Anderson, 2004).

The axial surface of isoclinal folds in the Hawick Group strikes 067°, which is parallel to the dominant strike of the bedding (Fig. 1A) (Anderson, 1987). North of the Mourne Mountains, the dominant strike of the Hawick Group (and the Southern Uplands–Down-Longford terrane) bedding changes to ∼037° (Beamish et al., 2010). However, no distinct change to bedding strikes within the Hawick Group has been reported in the vicinity of the Mourne Mountains (Hood, 1981; Gibson, 1984). Structures related to three phases of deformation (D1 to D3) have been mapped in the Southern Uplands–Down-Longford terrane. D1 (cleavage) developed parallel to the axial plane of the isoclinal folds (and the fault-bounded blocks). The fold axis of F1 folds is generally shallowly plunging (Anderson, 1987, 2000). D2 created a crenulation on the D1 cleavage that verges to the south, and D3 generated steeply plunging folds (Anderson, 2000). The D1 to D3 deformation has been interpreted to be caused by regional sinistral transpression (Anderson, 1987, 2000). Brittle deformation occurred during the waning stages of the Caledonian orogeny (Anderson, 2000). Two sets of sinistral strike-slip faults and a set of dextral strike-slip faults displace the Hawick Group and record dominant strikes of 195°, 355°, and 125°, respectively (Anderson, 1987). The slip on these faults ranges from a few centimeters to 100 m. The regional metamorphic grade of the Southern Uplands–Down-Longford terrane varies between zeolite and prehnite-pumpellyite facies caused by burial during accretion down to 12 km depth (Oliver and Leggett, 1980; Merriman and Roberts, 2001).

2.2. Paleogene Tectonics

During the early Paleogene, the principal tectonic stresses affecting the Southern Uplands–Down-Longford terrane of Northern Ireland were caused by the far-field Alpine orogeny that created two large strike-slip conjugate fault sets that trend NNW-SSE and ENE-WSW (Fig. 1A) (Cooper et al., 2012). At the onset of the opening of the North Atlantic during the formation of the British and Irish Paleogene Igneous Province, mantle plume–related tectonics became dominant (Lawver and Muller, 1994; Emeleus and Bell, 2005). Four major dike swarms display the stress field of plume-related rifting (Fig. 1A). The two oldest dike swarms, (1) the Erne dike swarm that strikes primarily WNW-ESE and (2) the Donegal–Kingscourt dike swarm that strikes NW-SE, are overlain by the Antrim lavas and therefore intruded before ca. 62 Ma (Cooper, 2004; Ganerød et al., 2010; Cooper et al., 2012). (3) The Ardglass-Ballycastle dike swarm that strikes from NNW-SSE to WNW-ESE has a bracketed age between ca. 62 and 59 Ma (Cooper, 2004; Ganerød et al., 2010; Cooper et al., 2012). The youngest dike swarm, (4) the St. John’s Point–Lisburn dike swarm, consists of NNW-SSE–striking dikes that intrude the Antrim lavas but are truncated by the Mourne granite pluton. This suggests an age of the St. John’s Point–Lisburn dike swarm between ca. 59 and 56 Ma (Cooper, 2004; Cooper and Johnston, 2004; Ganerød et al., 2010; Cooper et al., 2012; see also section below). The stress fields of the igneous centers in northeastern Ireland have locally affected the trend of the dike swarms (Cooper et al., 2012). There is evidence of reactivation of the Paleogene faults in Northern Ireland after the end of magmatic activity, but no major tectonic event occurred in the area of the Mourne Mountains after the emplacement of the granites (Johnston, 2004; Quinn, 2006; Cooper et al., 2012).

2.3. The Mourne Mountains Magmatic Center

The Mourne granite pluton is exposed in a mountainous area along the southeastern coast of Northern Ireland (Fig. 1). The granite pluton has been treated as two separate magmatic centers, an eastern and western center (Fig. 2A) (Emeleus, 1955; Hood, 1981; Gibson, 1984; Meighan et al., 1984; Cooper and Johnston, 2004; Stevenson et al., 2007; Stevenson and Bennett, 2011). However, mapping has shown that the granites form a connected pluton (Richey, 1927; Emeleus, 1955; Hood, 1981; Gibson, 1984). The oldest granites, G1 to G3, form the eastern part of the pluton, and the younger granites, G4 and G5, form the western part (Fig. 2B) (Cooper and Johnston, 2004). Radiometric ages of granites G1 to G4 overlap within the error of the applied methods, suggesting that the bulk of the pluton was assembled rapidly at ca. 56 Ma (cf. Gibson et al., 1987, 1995; Gamble et al., 1999; Wilkinson et al., 2017), while G5 yields a slightly younger age of 54.6 ± 1.0 Ma (40Ar/39Ar, Gibson et al., 1995). In the eastern Mourne Mountains, the granites are exposed at higher elevations (up to 850 m) than the western Mourne Mountains (up to 638 m), which can be interpreted, together with estimates of contact location and dip, as due to ∼200 m or more difference in intrusion depth (cf. Gibson, 1984). Allen et al. (2002) estimated that ∼3000 m of overlying rock have been eroded from the Mourne granites since their emplacement.

The Mourne dike swarm is mafic in composition and located to the east and SE of the Mourne granite pluton (Akiman, 1971). The Mourne dike swam displays strike orientations (NW-SE and NNW-SSE) similar to the regional Donegal and Kingscourt and St. John’s point and Lisburn dike swarm (Tomkeieff and Marshall, 1935; Cooper et al., 2012) and might therefore be an extension of these two swarms (Fig. 1). The Mourne granite pluton is only sparsely intruded by dikes and interpreted to be younger than the dike swarms (Akiman, 1971; Hood, 1981; Cooper et al., 2012). Additionally, several moderately inclined magmatic sheets that dip toward the Mourne granite pluton are exposed in the surrounding graywacke. These magmatic sheets are mafic to felsic in composition and commonly composite. Composite sheet exposures occur at localities surrounding the eastern and western part of the Mourne Mountains (Fig. 1B). The sheets are generally inclined at an angle of ∼35° toward the granite pluton and might therefore be a circumferential exposure of the same (cone) sheet that originated from beneath the current exposure of the Mourne granite pluton (cf. Richey et al., 1930; Anderson, 1937; Akiman, 1971; Burchardt et al., 2013). The granites in the western part of the Mourne Mountains truncate the composite sheet. The “Mourne” dike swarm is cut by a composite magmatic sheet at the mouth of the Bloody Bridge River and is therefore interpreted to be older than the observed composite sheets close to the Mourne granite pluton (Emeleus, 1955; Akiman, 1971).

The graywacke next to the Mourne granites has been altered to hornfels (Richey, 1927; Emeleus, 1955). Emeleus (1955) distinguished two different types of hornfels: one type consisting of quartz, plagioclase, and oxides, the second type is banded with biotite (blue bands) and diopside (green bands), where Emeleus (1955) interpreted the banding as a relict sedimentary feature.

In the northeastern part of the Eastern Mourne pluton, the contact with the Silurian graywacke dips moderately outward from the center of the pluton (35–50° NE) and has been suggested to represent the wall (or a lateral contact) of the granite pluton (Richey, 1927; Hood, 1981). This contact is terminated at both ends by granitic dike-like tips parallel to the contact (Figs. 2A and 4C). Richey (1927) therefore interpreted the northeastern contact to the Eastern Mourne pluton to represent part of the feeding ring dike to the granite pluton.

Stevenson et al. (2007) and Stevenson and Bennett (2011) analyzed magnetic (AMS) fabrics in the Mourne Mountains granites. The magnetic fabric in the Mourne granite pluton is dominantly oblate to triaxial. In G1 and the bulk of G2, the magnetic fabric dips outward from the center of the pluton. The dip of the magnetic fabric is moderate to steep in G1 and shallow in G2. In the southwestern part of the Eastern Mourne pluton, on the other hand, the fabric is largely triaxial to oblate, subhorizontal, and no dominant dip direction can be observed. In G3, the fabric is dominantly triaxial to weakly oblate with a subhorizontal foliation and lineation. In some parts of G3, the foliation is subvertical and strikes NE-SW. Stevenson et al. (2007) interpreted the generally subhorizontal fabric in the interior of the Eastern Mourne pluton and contact parallel fabric in the eastern and northeastern part of the pluton as evidence that the granites were fed laterally either from the NE or SW. They further interpreted the northeastern contact as a fault, attributing its formation to roof uplift during bysmalith inflation (Figs. 2A and 3B). In the central and southern part of the Western Mourne pluton, the magnetic fabric is subhorizontal, and the magnetic lineation trends NNE-SSW. In the western, northeastern, and eastern parts of the Western Mourne pluton, the magnetic fabric dips shallowly outward from the center of the pluton. Some of the magnetic lineations in the eastern part of the pluton are subhorizontal and trend N-S. Based on the subhorizontal NNE-SSW trends, magnetic lineation, and contact parallel magnetic foliation in the Western Mourne pluton, Stevenson and Bennett (2011) favored an emplacement scenario where all the Mourne granites were fed from the SW. A large gravity anomaly beneath the Carlingford Lough, SW of the Mourne Mountains, is inferred to represent the mafic source magma chamber to the Mourne granite pluton and to the Slieve Gullion and Carlingford igneous centers (Fig. 1) (Reay, 2004; Stevenson et al., 2007; Stevenson and Bennett, 2011). Notably, the Mourne Mountains are characterized by the absence of a negative gravity anomaly (Reay, 2004).

2.3.1. Contact Relationships in the Eastern Mourne Pluton

Apart from the moderately inclined northeastern contact, all other exposed contacts to the Eastern Mourne pluton in general dip shallowly (10°–30°) away from the center of the granite body, and because the host rock overlies the granite at these localities, they are thought to represent roof contacts (Hood, 1981). The exposed contacts to the Eastern Mourne pluton are always discordant to the bedding of the graywacke (Richey, 1927; Hood, 1981). Fine-grained chill zones occur in the granites, both at the external (granite-graywacke) and internal (granite-granite) contacts (Richey, 1927; Gibson, 1984).

G1 is the least voluminous, the oldest, and geochemically most primitive of the Mourne granites and crops out only on the highest peaks in the northeastern part of the Mourne Mountains (Meighan et al., 1984) (Figs. 2A and 4F). G1 has therefore been assumed to be a relatively thin granite sheet, although its roof is not exposed (Hood, 1981; Meighan et al., 1984). G2 underlies G1, and the contact between the two granites dips ∼30° toward NE (Fig. 5). Stevenson et al. (2007) argued that the inclination of the G1-G2 contact is an evidence for up-doming (i.e., laccolith inflation). G2 is composed of a finer- and a coarser-grained variety named G2 outer and G2 inner, respectively. The orientation of the G2 inner-G2 outer contact varies within the granite pluton. In the northwestern and southeastern part of the Eastern Mourne pluton, the contact dips shallowly outward (20°), whereas in the east, the contact dips moderately outward from the center of the pluton (Fig. 5 and Supplemental Fig. S11) (50°, Hood, 1981). G3 underlies and is finer grained than G2 (Fig. 4G). The contact between G2 and G3 is clearly exposed on crags and cliffs in the central part of the eastern Mourne Mountains surrounding the Ben Crom reservoir and dips shallowly (<10°) to the NE (Hood, 1981). The contact is in places wavy, but sharp, suggesting that the not-yet-solid G2 crystal mush was eroded by the intruding G3 magma. G3 displays a fine-grained chilled zone at the contact, while G2 is coarse grained (Fig. 4G). Richey (1927) argued that G2 should display a chill zone as well, if G3 intruded along the G2-graywacke floor interface and therefore interpreted that G3 split G2 and that the bottom of G2 subsided as G3 intruded. Close to the G2-G3 contact, centimeter to decimeter in diameter patches of coarser granite (interpreted as older granites of G3 composition or alternatively xenoliths of G2) occur, as well as pegmatite veins, separated leucocratic bands, and mica bands (Hood, 1981). Hood (1981) inferred that the floor to the granites was gently inclined toward the east caused by asymmetric floor subsidence during the emplacement of the granites (see also Cooper and Johnston, 2004). Stevenson et al. (2007) also considered the possibility of a curved floor that dips to the NE in one of their proposed emplacement models for the Eastern Mourne pluton.

2.3.2. Contact Relationships in the Western Mourne Pluton

The graywacke-granite contact in the Western Mourne pluton is generally discordant to the host-rock bedding (Emeleus, 1955; Gibson, 1984). The only location where the contact is concordant to the graywacke bedding is on Slievenamiskan, in the northern part of the Western Mourne pluton (Gibson, 1984). There, the granite-graywacke contact dips between ∼20°–35° toward the SE. The chill zone next to the graywacke-granite contact is thinner (5 mm) where abundant granite dikes have intruded the graywacke, while it is several meters thick where granitic dikes occurs sparsely in the host rock (Gibson, 1984). Pegmatites and drusy cavities are particularly common directly adjacent to the internal (granite-granite) and external (granite-graywacke) contacts. The G4 granite body is dome-like in shape and exposed at the rim of the Western Mourne pluton and at elevations above 310 m above sea level (asl) in the central part of the Western Mourne pluton (Figs. 5 and S1 [footnote 1]) (Gibson, 1984). Outcrop relationships in the area between the Eastern and Western Mourne plutons indicate that G4 truncates G3 and is therefore younger (see also Fig. 5) (Gibson, 1984). G4 is geochemically very similar to granites G2 and G3, but with a distinct texture (Hood, 1981; Gibson, 1984). G5 crops out below G4 in the central parts of the Western Mourne pluton, and the internal contact between G4 and G5 dips shallowly outward (10°) from the center of the intrusion (Figs. 5 and S1 [footnote 1]). Xenoliths of G4 (several decimeters in diameter) occur adjacent to the granite-granite contact in G5 in some places, and G5 aplites intrude the G4 joint system (Fig. 4H) (Gibson, 1984). Younger regional faults downthrow the western parts of the Western Mourne granite pluton in several grabens (Fig. 5) (Emeleus, 1955).

3. METHODS

3.1. Field Campaign

Structural orientation data were collected on fractures, contacts, bedding and joints during three field campaigns in April 2015, May 2016, and September 2016 using the FieldMove clino Pro application (www.mve.com/digital-mapping/) with an Iphone 6® in the coordinate system UTM Zone 29N and analogue compasses. The smartphone measurements were regularly double-checked with an analogue compass. The Fieldmove clino pro application corrected the smartphone measurements for the magnetic declination in the field, while the orientation measurements with an analogue compass were corrected for the local magnetic declination (3.3°W) after the fieldwork. The smartphone compass together with the Fieldmove clino pro application are generally accurate to ±10° (Scott et al., 2016; Allmendinger et al., 2017), and that the error is overcome by the larger amount of data that can be collected compared with an analogue compass (www.mve.com/digital-mapping/). Measurements were collected at localities close to the contact between the graywacke and the granites. Due to sporadic and often small outcrop exposures, systematic line, circle or square fracture mapping was not practical at most localities. We therefore aimed to measure all fractures (i.e., where strike and dip were attainable) over a smaller area (25 × 25 m) to avoid sampling bias. We did not apply any weighting due to exposure biases for in situ measured fractures. However, certain fracture sets had naturally more relief and might therefore have been favored due to visibility and the orientation of the outcrop surface(s).

To complement the data measured in the field, fractures and bedding in the Hawick Group were measured by marking natural lineaments on outcrops on satellite images along the southeastern coast of Northern Ireland (GLOBALearth) in ESRI ArcGIS™ (https://www.esri.com/en-us/arcgis/). The lineament orientations were analyzed in MOVE 2017.2™ (https://www.mve.com/).

Contact aureole samples were collected during the 2015 field campaign in traverses from the contact up to 1.7 km from the granite in the wall to the Eastern Mourne pluton to constrain the timing of deformation compared to contact alteration. Representative samples were selected for thin section preparation (see Supplemental Table S1 [footnote 1]). Several of the thin sections were analyzed with electron dispersive spectroscopy (EDS) using the field emission source JEOL JXA-8530F Hyperprobe at Uppsala University, Sweden, to identify mineral phases.

3.2. Virtual Outcrops

A virtual outcrop is a georeferenced photomosaic-textured 3D model of an outcrop, which can be used to remotely study structural features, such as fracture orientations (cf. Senger et al., 2015). To examine the fracture pattern near the granite-graywacke contact, we applied structure from motion (SfM) photogrammetry on the host rock to the Western Mourne pluton. Three localities with exposed outcrops at different vertical and horizontal distances to the graywacke-granite contact were chosen as targets for photogrammetry to generate the virtual outcrops.

We performed photogrammetry on nine outcrops in total. Photographs were taken with a 24-megapixel DSLR camera with GPS. We also implemented ground control points, the XYZ (UTM) coordinates of which were saved with a hand-held GPS. One photograph was taken with the camera at chest height and one above the head every meter along the outcrop length. The camera was kept level at an equal distance to the rock face (allowing a fixed focal distance). The images were then processed using the software Agisoft Photoscan™ (http://www.agisoft.com/). The quality of photographs was checked with the function “Estimate Image Quality,” and photographs with a quality <0.5 were removed from the image set. First, the photographs were automatically aligned, and tie points in the photographs were identified. The tie points were then used to calculate a point cloud and a mesh surface. The mesh surface was draped with a photomosaic texture created from the photographs generating a 3D “virtual” outcrop. The virtual outcrops were georeferenced using the location information in the metadata of the photographs and the ground control points. In situ fracture measurements were collected at every locality to check the accuracy of the virtual outcrop fracture measurements.

3.3. Three-Dimensional Reconstruction of the Mourne Mountains Granite Pluton Roof

Determining the shape of the roof and wall contacts of the intrusion is important to the understanding of its emplacement (e.g., Burchardt et al., 2010). For example, the dip of the contact should represent the degree of rotation of the host if the intrusion was emplaced by laccolith inflation (e.g., Gilbert, 1877). Since not all contacts could be measured in situ, we used the software MOVE 2017.2™ to perform map analysis for reconstruction of the orientation of the roof contact to the Mourne granite pluton. A topographical surface of the Mourne Mountains was created from the Ordnance Survey of Northern Ireland (OSNI) open data digital-terrain model (sheets 253–285, 10 m resolution, http://www.nationalarchives.gov.uk/doc/open-government-licence/version/3/). The contacts of the Mourne granites from the geological maps of Hood (1981) and Gibson (1984) were traced manually using the line tool in MOVE™ and projected onto the topographical surface. Contact dip data were compiled from previously published data (Hood, 1981; Gibson, 1984), our new field measurements, and structure contours (i.e., the projected contact line). The dip of the contact was then projected as a surface (down-dip) from the contact line. To reconstruct the shape of the Mourne granite pluton’s wall and roof, 14 cross-sections were created in an E-W direction and 23 cross sections in an N-S direction. In each cross section, a line representing the intrusion roof was drawn manually taking into consideration the projected dip of contacts and topography, i.e., where the pluton roof was eroded, the contact was assumed to lie close to the granite peaks to avoid overestimating the dip of the contact (Fig. S2 [footnote 1]). A roof surface of the pluton was created with the Ordinary kriging method using the contacts and the cross-section lines. The dip of the roof surface near the present-day contacts and the roof pendant is well constrained by in situ measurements and by map analysis. However, the shape of the Mourne granite pluton roof far from a present-day exposed contact is uncertain.

3.4. Fracture Orientation Analysis

Fracture measurements were sorted into 16 host-rock localities to discern variability in the fracture pattern around the pluton, as well as one locality with measurements of fractures collected farther than two kilometers from the granite contact (Fig. 2B). The host-rock fracture localities were defined based on measurement location and contact orientation. Fourteen of the localities are located in the pluton roof, where the host rock overlies the granite, and two localities are located in the intrusion wall, where the host rock is laterally adjacent to the granites. Data from the localities were analyzed with k-mean cluster analysis in the software SG2PS (Sasvári and Baharev, 2014). The software was used to separate fracture sets with k-mean clustering, which limits some of the bias in determining which fractures belong to the same set. We used the Fisher k values generated in MOVE™ to determine dispersion within the fracture clusters. Higher Fisher k values indicate a more narrowly clustered set of fractures (Fisher et al., 1993). Fracture data were plotted using the Software Stereonet 10 (Allmendinger et al., 2012). Strike data are given in the right-hand rule convention, i.e., strike is oriented 90° anti-clockwise from the dip direction. The mean principal orientation of a fracture set was calculated in MOVE™. The mean principal orientation is determined by the maximum eigenvector of the poles to the fractures of a set and therefore gives a good estimate of the mean orientation of bimodal data (cf. Mardia, 1975; Fisher et al., 1993).

3.4.1. Virtual Outcrop Analysis and Scan Lines

To analyze the strike and dip (3D orientation) of fractures on the virtual outcrop, the textured surface was imported into the LIME v. 1.0 software (Buckley et al., 2019, http://virtualoutcrop.com/lime) in a local coordinate system, but with accurate scale and orientation. The “Structural data from 3 points” tool was used on the outcrop surface to attain the orientations of measurable fractures. Fracture orientations in the virtual outcrops were analyzed using (virtual) scan lines in MOVE™, which enable comparison of fracture intensity and spacing between different areas around the pluton (cf. Zeeb et al., 2013). The scan lines were drawn manually with a start and end point placed horizontally along the long axis of the virtual outcrop. Fracture intersections on the scan line were collected by converting the scan lines to wells and counting plane intersections on the well track. Fracture spacing was analyzed with 1 m moving windows in MATLAB™. When measuring fractures along a scan line, a bias is inherently introduced due to the orientation of the scan line. To correct for this bias when comparing fracture intensities between scan line localities, we applied a weighting factor (w) (Equation 1): 
graphic
where β is the angle between the pole of the fracture and the azimuth of the scan line (Terzaghi, 1965). When the angle between the pole to the fracture and the scan line was ∼75°, we did not apply the weighting factor to avoid large and possibly unrepresentative weighting factors (cf. Terzaghi, 1965; Park and West, 2002; Lato et al., 2010).

4. RESULTS

4.1. Contact Relationships to the Mourne Granite Pluton

The exposed contacts between the granite and the Silurian graywacke in the eastern Mourne Mountains are sharp and straight (Figs. 4B–4D). Host-rock xenoliths and granitic dikes intruding the host rock are rare in the eastern part of the Eastern Mourne pluton. However, granitic dikes are common in the host rock on Slieve Muck, next to the western part of the Eastern Mourne pluton. The contact between the host rock and the Western Mourne pluton is sharp and generally either wavy or stepped. Centimeter to decimeter in diameter angular host-rock xenoliths occur abundantly in the granite (Fig. 4D), and granitic veins and/or dikes often intrude bedding and preexisting fracture planes within 50 m from the contact. The internal granite-granite contacts are sharp. G1 displays conjugate fractures above the contact with G2 (Fig. 4F), and the G2-G3 granite contact is wavy at several localities (Fig. 4G).

4.2. Structures in the Contact-Metamorphic Aureole to the Mourne Granite Pluton

The following sections present descriptions of samples collected along traverses in the contact-metamorphic aureole to the Mourne granite pluton (Fig. 2A). See Table S1 (footnote 1) for detailed descriptions of the samples collected in the aureole.

4.2.1. Tollymore Traverse

The Tollymore traverse extends from 100 m to 1660 m north of the granite (Fig. 2A). The rocks in the traverse consist of siltstone and shale altered to hornfels. In sample TM-3, ∼60 m from the granite-graywacke contact, diopside and biotite are separated in different layers in the rock. In samples TM-1 and 2 (∼700 m from the contact), actinolite and tremolite form layer-parallel segregations that are cut and displaced along chlorite-filled veins (Fig. 6C). In samples >1000 m from the granite-graywacke contact (e.g., TM-7), chlorite and actinolite veins cut the diopside, actinolite, and biotite groundmass of the sample.

4.2.2. Bloody Bridge River Traverse

The Bloody Bridge River traverse extends from the sedimentary screen to 610 m east of the granite contact (Fig. 2A). The samples in the traverse consist of siltstone altered to hornfels. In samples collected in the sedimentary screen and next to the granite-graywacke contact (SC-1, BBT-4, 5, and 8), the groundmass is relatively fine grained compared to samples collected farther from the contact in the traverse (BBT-9–16). Samples SC-1 and BBT-8 display diopside segregations, and samples BBT-4 and 5 display diopside and actinolite segregations that are cut by quartz and actinolite veins. The sedimentary screen samples are either brecciated or folded (Fig. 6B), and sample BBT-3 has an intrusive breccia vein (tuffisite with cataclasite texture) with biotite-rich hornfels and feldspar clasts (Fig. 6A). The granite to the west of the sedimentary screen in the eastern part of the Eastern Mourne pluton is also brecciated and displays several millimeters to centimeter thick contact-parallel cataclasite bands (Fig. 4E). In sample BBT-12 (100 m from the contact), diopside segregations are cut by thin quartz veins, which are transgressed by quartz and actinolite veins. Sample BBT-14 (210 m from the contact) is brecciated, and biotite and quartz segregations are displaced by fractures. Sample BBT-16 (610 m from the contact), located in the far end of the traverse, shows bedding-parallel diopside and actinolite segregations.

4.2.3. Glen River Traverse

The Glen River traverse extends from 100 m to 770 m NE of the granite contact (Fig. 2A). The rocks in the traverse consist of siltstone but also shale altered to hornfels. Porphyroblasts of amphibole (likely kaersutite) and actinolite have crystallized in the biotite-dominated groundmass of sample GRT-4, which is transgressed by quartz and diopside segregations. In the samples collected farther than 200 m from the granite-graywacke contact, the porphyroblasts disappear, but plagioclase, quartz, and diopside segregations are still present in the rock and are cut by later quartz veins. The samples GRT-10–15 in the far end of the traverse are banded (Fig. 6D). In the darker layers, the quartz is clotted, while quartz clasts are more abundant in the lighter layers of the hornfels. The layers are displaced (≤1 mm) along fractures. Several thin cataclasite shear zones occur parallel to the layers (Fig. 6D), and GRT-13 and 15 are brecciated.

4.2.4. Roof Sample

In a siltstone (hornfels) sample from Shan Slieve (SS-1) next to G1, the groundmass is biotite-rich and cut by quartz veins, diopside segregations, and quartz and actinolite veins. Within a shear zone, both amphibole and sigmoidal porphyroblasts are replaced by quartz, biotite, and alkali feldspar. The mineral assemblage suggests that the retrogressed porphyroblasts were originally cordierite (cf. Pattison and Tracy, 1991).

4.3. Bedding Orientation and Roof Shape Reconstruction

The inclination of the graywacke bedding varies in the Mourne Mountains, ranging from subhorizontal to subvertical (8°–88°) with an average dip of 29° and a mean principal plane that strikes 054° (Table 1; Fig. 7). Several open folds are observed in the host rock on Slieve Muck (cf. Richey, 1927; Hood, 1981). The reconstructed fold axis of the folds on northern Muck trends 238/21°, and on southern Muck, the reconstructed fold axis trend and plunge 229/21°. The reconstructed axial surfaces of the folds at different localities are also similar with orientations of 226/62° on northern Muck and 218/71° (strike/dip) on southern Muck, respectively. The reconstructed roof of the Eastern Mourne granite pluton is slightly dome shaped (Figs. 5, 8A, and 8B; Fig. S1 [footnote 1]), and the edges of the pluton roof dip shallowly to moderately outwards from the center of the pluton (10°–50°) (Figs. 8C and 8D). Map analysis indicates that the roof to the eastern part of the Western Mourne pluton dips shallowly (10°) to the east and suggests that the inclination of the roof to the western part of the Western Mourne pluton is marginally steeper inclined (15°) (Figs. 5, 8C, and 8D; Fig. S1).

4.4. Fracture Pattern

In total, 4549 measurements were collected on structural features including contacts, bedding, fractures, dikes, and lineaments in the Mourne Mountains and the surrounding area. The data for fractures, dikes, and lineaments are presented below and in Supplemental Table S2 (footnote 1).

4.4.1. Regional Fractures

In total, 505 lineaments interpreted to represent faults and fractures were remotely measured on beach exposures of the Hawick Group east of the Mourne Mountains (Fig. 1B and Supplemental Fig. S3). The measured lineaments trend primarily 140° and secondarily 035°. Measured lineaments interpreted to represent bedding trend on average 060° (n = 88). Orientations of fracture and bedding lineaments on beach exposures adjacent to the Mourne Mountains are offset compared to lineaments measured >10 km from the intrusion, where the fracture and fault lineament sets trend 100° and 160°. All fracture data are given in Supplemental Table S2 (footnote 1).

4.4.2. Fractures in the Intrusion Roof

In total, 2094 fractures were measured in host-rock outcrops surrounding the Mourne granite pluton (Fig. 9A). Four subvertical fractures sets are observed in the studied roof outcrop localities striking dominantly SE, S, NE, and E (see Table 1 for detailed orientation of each fracture set; Figs. 2B, 4B, and 10A). It was only possible to distinguish a maximum of three sets of subvertical fractures at a single locality, which could be due to outcrop orientation. In several localities, a set of subhorizontal fractures is also identified. In the southwestern part of the Western Mourne pluton, the dominant strike of the subvertical fracture sets is NW, N, and E. There, the calculated mean principal fracture plane of the NW(-SE)–striking fracture set dips toward the NE as opposed to the SW, which is the dominant dip-direction in the other studied localities (Table 1). Limited fracture plane exposures made it hard to observe slickenlines or other displacement indicators. All fracture data and k-mean cluster groupings are given in Supplemental Table S2 (footnote 1).

4.4.3. Fractures in the Intrusion Wall

In the wall rock to the Mourne granite pluton at Glen River (Fig. 2A), the mean strike of the bedding is 298°, and two dominant fracture sets were identified striking 209° and 271° (Table 1; Figs. 7 and 10A). In the Bloody Bridge River, the mean strike of the bedding is 017°, and two fracture sets were identified that strike 165° and 274° (Table 1; Figs. 7 and 10A). All fracture data and k-mean cluster groupings are given in Supplemental Table S2.

4.4.4. Joints in the Granites

In total, 1411 fractures were measured within the Mourne granite pluton (Fig. 10B). We determined these fractures to overwhelmingly be joints since displacement could rarely be observed on them (cf. Pollard and Aydin, 1988). The G1 joints were studied at three different localities and are primarily subvertical and dip toward the S and SW. The jointing in G1 is visibly denser spaced compared to the more voluminous younger granites. Conjugate fractures with moderate dips were observed in the easternmost studied outcrop of G1. A subhorizontal set of joints in G1 was also identified; the set dips on average 20° (n = 30) to the NNE (Fig. 10B).

The studied outcrops of G2 and G3 in the central part of the Eastern Mourne pluton have similar joint sets with two subvertical joint sets striking SW and ESE and a subhorizontal set of joints. At Bloody Bridge River, a pervasive set of subvertical joints in G2 outer strikes to the S. At Glen River, joints in G2 are steep and dip dominantly to the SW. The G4 joint pattern is similar to those of G2 and G3 and includes subvertical and subhorizontal joint sets. At Gruggandoo, two steeply inclined joint sets occur in G4; these sets strike NW and NE (Fig. 10B). Next to the granite-graywacke contact, a set of fractures in G4 is oriented 319/80° (n = 20). Four striations were measured on the fracture surfaces, which trend and plunge 135/37° on average. The G5 joint pattern largely resembles the G4 joint pattern (Fig. 10B). All joint data are given in Supplemental Table S2 (footnote 1).

4.4.5. Granitic Dikes

In total, 243 granitic dikes that range from decimeter to millimeter in thickness were measured in the host rock to the Mourne granites (Fig. 9B). The granitic dikes in the Mourne Mountains parallel the fractures in the host rock and strike dominantly SE and secondarily S and NE (Fig. 9B). All dike data are given in Supplemental Table S2.

4.5. Scan Lines

The weighted (bias-corrected) average fracture intensity on the studied scan lines ranges from 2.2 to 5.5 fractures/m, while the uncorrected fracture intensity ranges from 0 to 11 fractures/m in the studied outcrops. Scan line locations are given in Figure 10A; see Figure 11 and Table 2 for results and Supplemental Table S3 for weighting factors of individual fractures collected on the scan lines and the supplementary information for individual scan line data. The scan lines show that the NW-striking fracture set is the most dominant set in the host rock. This is corroborated by relatively low fracture intensity observed along the scan lines at Pigeon Rock; these lines trend toward the NW and are roughly parallel to the NW-striking fracture set. The scan lines record no distinct decrease in fracture intensity with distance from the contact (Fig. 11; Table 2). Fracture intensity may instead be controlled by lithology in the Hawick Group (cf. Anderson, 1987). However, our data cannot resolve this further. The fracture data collected on the virtual scan lines compare well with the data measured in situ at the scan line localities (albeit not along a scan line) (see Supplemental Fig. S4).

5. BRITTLE DEFORMATION FEATURES IN AND AROUND THE MOURNE GRANITE PLUTON

5.1. Joint Orientations in the Granites

G1 exhibits a conjugate set of fractures (Fig. 4F), which suggests that G1 was deformed in a brittle manner when G2 intruded or at separate deformation event. The highly systematic vertical and subhorizontal fractures in G2 to G5 are interpreted as cooling joints (cf. Cloos, 1922; Price and Cosgrove, 1990). In the Eastern Mourne pluton, the subvertical joints that strike parallel to the internal contacts (G1-G2 and G2-G3) dip toward the SW. Cooling joints in magmatic intrusions are generally oriented at a high angle to the contacts (cf. Cloos, 1922; Price and Cosgrove, 1990; Bergbauer and Martel, 1999) and therefore support the interpretation that the external and internal contacts in the Eastern Mourne pluton dip at a shallow angle toward the NE (Fig. 5; see also section 2.3.1; Hood, 1981). Moreover, the subhorizontal joints dip dominantly toward the NE. In the Western Mourne pluton, the orientation of the subhorizontal and subvertical joints are largely parallel and perpendicular to the orientation of the host-rock contact (Fig. 10B), respectively, which is inclined outward from the center of the Western Mourne pluton (cf. Figs. 8B–8D).

5.2. Fractures in the Host Rock to the Mourne Granite Pluton

Three possible scenarios can be envisaged regarding the timing of the different fracture sets in the host rock to the Mourne granite pluton. The fractures either (1) pre-date, (2) formed syn-emplacement, or (3) post-date the emplacement of the granites. We now discuss how to evaluate the timing of formation of the fractures measured in the Mourne granite pluton and its host rock.

  • (1) The Southern Upland–Down-Longford terrane records both ductile and brittle deformation phases associated to the Caledonian orogeny (Anderson and Cameron, 1979; Anderson, 1987). The recorded regional fracture orientations overlap with the three sets of strike-slip faults recorded by Anderson (1987) on the Ardglass peninsula that strike dominantly 195°, 355° and 125°. However, our fracture data present different dominant orientations. Fractures measured farther than 2 km from the granite contact and remotely measured lineaments on the coast NE of the Mourne granite pluton strike SE and SW (Table 1; Figs. 1B and S3 [footnote 1]). Notably, the remotely measured regional fractures show strong correlation with two of the measured fracture sets in the host rock to the Mourne granite pluton (Table 1).

  • (2) Fractures related to the emplacement of shallow magmatic intrusions are commonly parallel or perpendicular to the intrusion contact (Johnson and Pollard, 1973; de Saint-Blanquat et al., 2006; Morgan et al., 2008; Feng et al., 2011; Wilson et al., 2016). Wilson et al. (2016) also reported later tensile joints in the host rock to the Trachyte mesa intrusion in the roof and at the margins that are parallel and perpendicular to the intrusion margin and interpreted them to be related to extension during the final stage of magma emplacement. In the roof to the Black Mesa pluton, USA, steeply inclined joints are oriented perpendicular to the direction of strain (flow of magma) during magma emplacement (de Saint-Blanquat et al., 2006). Additionally, the fracture intensity has been reported to increase towards the intrusion contact (Woodcock and Underhill, 1987; de Saint-Blanquat et al., 2006; Morgan et al., 2008; Wilson et al., 2016). We therefore assume that fracture sets created by deformation during the emplacement of the Mourne granite pluton should be parallel and perpendicular to the pluton contacts. This is supported by the orientation of the joints within the Mourne pluton, which indicates extensional stresses parallel and perpendicular to the contact (see section 5.1.; Fig. 10B). In contrast, the orientations of the steeply inclined fracture sets in most roof outcrops around the Mourne granite pluton are strikingly similar, even where the strike of the contact is different (see Fig. 10A and Table 1 for a comparison of the fracture sets at the different studied host-rock localities). For example, at localities where the contact between granite and host rock dips >20°, such as on Shan Slieve and Thomas Mountain, next to G1, the fractures dip steeply and with orientations similar to the regional fracture sets. Also, on Southern Slieve Muck, where the contact dips >20°, the fracture sets are steep with strikes similar to the regional fracture sets. Moreover, the scan lines show that observed fracture pattern differences (for example, between the Pigeon rock and Gruggandoo scan lines) are most likely due to the orientation of the studied outcrops (see section 4.5; Fig. 11; Table 2). Additionally, the scan lines do not record a decrease in fracture intensity with increasing distance from the granite contact and exhibit quite consistent fracture orientations (Fig. 11; Table 2). These observations suggest that the most dominant fracture sets in the host rock to the Mourne pluton formed before the emplacement of the granites. However, at localities where the roof is relatively flat (e.g., Upper Pigeon rock, 4°W) some fracturing might have been caused by magma flow. The magnetic fabric close to the Pigeon Rock localities indicates flow toward the NE (Stevenson and Bennett, 2011), roughly perpendicular to the strike of the contact and the SE-striking dominant fracture set. The wall-rock outcrops at Bloody Bridge and Glen River show roughly contact-parallel and contact-perpendicular fracture orientations, which imply that these fractures could have formed during granite emplacement. Additionally, the mean orientation of the dominant fracture set in the scan line Gruggandoo 1 (312/79°, n = 65) is parallel to fractures in granite G4 (319/80°, n = 20), located just below the granite-graywacke contact at Gruggandoo. Both fracture sets strike perpendicular to the contact. The striations on the G4 fractures indicate displacement parallel to the dip direction of the contact. The other Gruggandoo scan lines do not correlate with the G4 fractures (Fig. 11A). The Gruggandoo 1 scan line might therefore display an emplacement-related fracture set located close to the contact.

  • (3) Fractures formed after the emplacement of a granite are expected to have similar orientations whether they develop within the granite or the host rock (assuming that the difference in rheology does not have an influence). In our case study, the orientations of the dominant fracture sets in the host rock and joints in the granite do not correlate (Fig. 10). Any abundant postemplacement fracturing in the Mourne Mountains is therefore deemed unlikely.

While a syn-emplacement origin to some of the fractures measured in the roof to the Mourne granite pluton is hard to rule out, the two most dominant SE- and NE-striking fracture sets can be observed in the roof all around the pluton. We therefore assume that these fracture sets formed prior to the emplacement of the granites. The granitic dikes are largely parallel to the most dominant NE-SW–striking fracture set, which shows that the magma intrusion utilized these fractures in the roof. The presence of abundant preexisting fracturing may therefore have inhibited the formation of new fractures in the host rock during emplacement of the Mourne pluton.

5.3. Quantifying Rotational Host-Rock Deformation around the Mourne Granite Pluton

The shallow outward inclination of the granite-granite and granite-graywacke contacts, i.e., the dome shape of the Mourne granite pluton, was proposed by Stevenson et al. (2007) to reflect laccolith inflation (see also Figs. 8C and 8D). Hence, if we assume that the Mourne granites were initially horizontal sills that inflated into laccoliths, the dip of the contact should correspond to the doming of the host rock, i.e., the dip of the contact is the angle of rotation of the host rock caused by up-doming (cf. Fig. 3B).

Our results indicate that the orientation of the bedding at the same locality and between localities where the contact has similar orientation and dip (Fig. 7; Table 1). Moreover, the limited exposure of the host rock combined with the preexisting deformation (see section 2.1.) make the Hawick Group bedding unsuitable to test for up-doming and deformation within roof rock by magma emplacement. Instead, we employ the pervasive regional fracture sets in the host rock that formed prior to or locally during the granite emplacement (see section 5.2) as passive markers. Passive markers are expected to be deformed together with the host rock (cf. Fossen, 2010, p. 62); therefore, the fracture orientations should reflect the dip of the pluton contact if rotational deformation occurred during granite emplacement (Nelson and Sylvester, 1971; Feng et al., 2011).

In order to test if the host rock to the Mourne granite pluton has been deformed by laccolith inflation, we performed rotations of steeply dipping regional fractures and compared these rotated fracture sets with our in situ–measured host-rock fractures and dikes. The scan line data were not used for this comparison. The regional fractures used for the rotation were compiled from fracture data collected at distances of >2 km from the granite pluton. In the host rock to shallowly emplaced laccoliths, emplacement-related deformation can generally only be observed within a couple hundred meters from the lateral boundary of the intrusion (cf. de Saint-Blanquat et al., 2006; Morgan et al., 2008; van Wyk de Vries et al., 2014; Wilson et al., 2016; Mattsson et al., 2018). Laccoliths inflate by uplifting and rotating the host rock on the flanks of the intrusion (e.g., Pollard and Johnson, 1973; van Wyk de Vries et al., 2014). The accompanied rotation of the host rock during laccolith growth should be clockwise around a horizontal axis that trends parallel to the strike of the contact to the pluton. A clockwise rotation would either steepen or shallow regional fractures depending on the relative location on the pluton roof and the orientation of the preexisting fracture sets. The degree of rotation was assumed to be the same as the dip of the roof contact at the different roof localities (see above). To complement the rotation analyses of the regional fractures (Fig. 12; Supplemental Table S4 [footnote 1]), the in situ–measured fracture sets were rotated by subtracting the dip of the contact (“back-rotated”) at the respective host-rock locality (see Supplemental Fig. S5).

The results show that the rotated regional fracture sets do not correlate consistently with the measured fractures in the other roof-rock localities around the Mourne granite pluton (Fig. 12; Supplemental Table S4 [footnote 1]). For example, the Southern Slieve Muck, Thomas Mountain, and Slievenamiskan localities do not correlate with the rotated regional fracture sets, which should be the case if the Mourne granites were emplaced as laccoliths. However, at the Gruggandoo locality, the mean principal plane of Fracture set 1 of the rotated regional fractures is oriented 334/89° (strike, Supplemental Table S4) and is similar to the orientation of the mean principal plane of Fracture set 1 of the in situ measured fractures at Gruggandoo (327/77°, strike, Table 1). A similar pattern can also be observed when comparing the in situ–measured fracture sets and the rotated regional sets of the Finlieve and Slievemiskan localities (see section 4.4.2.; Fig. 12; Table 1; Supplemental Table S4). This is consistent with some inflation related to the emplacement of the granite in the western part of the Western Mourne pluton. Although, these fracture patterns require a greater rotation than suggested by the dip of the contact to reproduce the regional fracture pattern (Supplemental Fig. S5).

An inclined and curved geometry of the initial sill(s) (cf. Hood, 1981; Stevenson et al., 2007) could explain the lack of distinct rotation in the roof. For example, if the initial sill(s) was slightly dome-shaped, it would not result in any distinct rotation in the host rock, since the inflation of the sill(s) would be mostly accommodated by vertical uplift of the roof along faults (de Saint-Blanquat et al., 2006; Stevenson et al., 2007, Fig. 12). However, there is no indication that the magma lifted the roof along faults in the southern and western part of the pluton, and the fracture pattern does not suggest a distinct rotation on Slieve Muck in the western part of the Eastern Mourne pluton. On the other hand, if the Western Mourne granites were emplaced as laccoliths with a northeastward-inclined floor, the result would be little to no rotation in localities such as Slievenamiskan and Lower Pigeon Rock, while the host rock at Gruggandoo, Finlieve, and Slievemiskan would be rotated.

The wall rock to the Mourne granites, which was studied in the Glen and Bloody Bridge Rivers, reveals two dominantly steeply inclined fracture sets striking 209° and 271° and 165° and 274°, respectively (Fig. 10A; Table 1). This is roughly parallel and perpendicular to the pluton contact and could indicate that these fracture sets formed during the emplacement of the granites (see section 5.2.). However, the dominant SE-striking fracture set observed in the roof outcrops is seemingly absent. We therefore also tested rotating the fracture data collected in the wall to the pluton (Fig. 12). A rotation around a horizontal axis at an angle of ∼35 to 50° (dip of contact at the wall localities; see Hood, 1981) markedly shallows the dip of the wall-rock fractures. Such a rotation does not resolve the existence of steeply inclined fracture sets at the Glen and Bloody Bridge River localities (Fig. 10A; Table 1). Notably, rotation of the wall-rock fracture sets around a vertical axis at an angle representing the difference between the regional strike of the bedding (067°, Anderson, 1987) and the strike of the bedding at the studied locality aligns the two dominant fracture sets observed in the wall-rock localities parallel to the dominant SE- and NE-striking regional fracture sets. Furthermore, the “best-fit” rotation was clockwise around a vertical axis for the Glen River locality and anti-clockwise for the Bloody Bridge River locality (Table 1)—and not a rotation around a horizontal axis, as expected by laccolith inflation. Additionally, the agreement between the rotated regional fractures and the in situ–measured fracture sets in the wall outcrops suggests that the regional fracture sets in the roof exposures would indeed also have been rotated if laccolith inflation had occurred. See Table 1 for degree of rotation; Figure 12 and Supplemental Table S5 for comparison between the in situ–measured fracture data and the rotated regional fracture set; and Supplemental Figure S4 (footnote 1) for the contact orientation subtracted from the in situ–measured fracture data at each locality (see above for details).

5.4. Deformation in the Contact-Metamorphic Aureole

The small displacements of the contact-metamorphic segregations along fractures in several of the samples show that contact-metamorphic alteration occurred partly before deformation in the aureole (Figs. 6C and 6D). This deformation was likely associated with the emplacement of the Mourne granites, since several fractures are filled with contact-metamorphic minerals (see Table S1). Numerical models indicate that several thousands of years are required for host rock located 1 km from the intrusion to reach peak contact-metamorphic conditions (cf. Cui et al., 2003; Douglas et al., 2016). Notably, the emplacement-related deformation is primarily observed in the distal samples (relative to the granite) of the Glen River traverse (samples GRT-10–15). This deformation pattern suggests that initial emplacement of magma in the Mourne granite pluton did not cause significant deformation of the roof and wall of the pluton, while later granite emplacement is associated with deformation. The aureole to the Eastern Mourne pluton therefore records different stages of emplacement of the granite. A pattern of initial alteration followed by deformation has also been documented in aureoles to other plutons (see Morgan et al., 1998; Morgan, 2018). The deformed contact-metamorphic segregations in the host rock could reflect different laccolith growth stages, where the initial sill emplacement is accommodated by elastic deformation of the host rock (e.g., Pollard and Johnson, 1973; Jackson and Pollard, 1990; Bunger and Cruden, 2011; Michaut, 2011). The deformation of the segregations would then record the inflation of a sill to a laccolith (cf. de Saint-Blanquat et al., 2006). However, the existence of fractures, faults, and folds associated with the growth of sills and stacked-sill–type laccoliths (Pollard et al., 1975; de Saint-Blanquat et al., 2006; Schofield et al., 2012; Schmiedel et al., 2015; Wilson et al., 2016; Spacapan et al., 2017; Galland et al., 2019) have led some authors to question the role of elastic deformation during sill and laccolith emplacement (Haug et al., 2017; Scheibert et al., 2017).

The brecciated contact-metamorphic segregations and cataclasite in the Glen River and Bloody Bridge River traverses show that granite emplacement was associated with significant deformation on the northeastern wall of the pluton (Figs. 6A–6D; Table S1 [footnote 1]). Stevenson et al. (2007) noted a pluton-side-up sense of shear in cataclasite in G2 on Millstone Mountain located ∼1 km NNW of the sedimentary screen. Moreover, the brecciation in the distal samples of the Glen River traverse might mark the location of the transition between deflected host rock and graywacke unaffected by the emplacement of the granites. Our observations in the aureole suggest that neither cauldron subsidence emplacement, which would not cause significant deformation of the roof and wall of the pluton, nor laccolith emplacement, which would likely be associated with deformation throughout granite emplacement, can solely explain the deformation pattern in the aureole.

6. EMPLACING THE MOURNE GRANITES

The conjugate fractures in G1 (Fig. 4F), the rotation of the host rock along the northeastern contact (Figs. 2A and 12), the displaced contact-metamorphic segregations (Fig. 6), and potential uplift and rotation of the roof along the southwestern border of the intrusion point to intrusive deformation of the host during the emplacement of the granites. These observations do not support cauldron subsidence as a mechanism of emplacement; however, the host-rock deformation pattern to the Mourne granite pluton is also largely inconsistent with the expected deformation caused by a laccolith emplaced along a horizontal floor. For example, the regional fracture pattern can be reconstructed by rotation around a vertical axis (as opposed to a horizontal axis) in the wall rock to the Mourne granite pluton in the NE, while the nearby roof exposures (Thomas Mountain and Tollymore) show no distinct rotation (Fig. 12). The reported subhorizontal magnetic fabrics in the interior of the Mourne pluton and the shallowly inclined fabrics close to the contact indicate a largely contact-parallel, magnetic-fabric geometry (Stevenson et al., 2007; Stevenson and Bennett, 2011). The lack of studies of the magmatic fabric in plutons emplaced by floor subsidence makes it hard to pinpoint magma flow patterns related to floor subsidence. Moreover, in plutons, where the magmatic fabric does not record the regional strain, the fabric is commonly parallel to the contact with host rock (Buddington, 1959; Bateman et al., 1963; Paterson and Tobisch, 1988; Paterson et al., 1998). A contact-parallel magmatic fabric is therefore not sufficient evidence of laccolith inflation. The different local responses of the host rock to the emplacement of the granite pluton are hence held to either reflect laccolith (western part of the pluton) and bysmalith (eastern part of pluton) emplacement along a curved floor or multiple mechanisms of emplacement, both floor subsidence and roof uplift and wall deflection.

Magma emplacement through inflation of an inclined and curved (convex) sill and associated uplift of the roof along faults would result in a dome-shaped roof to the pluton. However, sill geometries are usually flat or saucer-shaped (flat base with upwardly inclined sheets; Galland et al., 2018), irrespective of the orientation of host-rock discontinuities (cf. Stephens et al., 2017). Moreover, laccoliths have been suggested to become less common at intrusion depths of more than three kilometers in favor of sills and lopoliths (Corry, 1988; Roman-Berdiel et al., 1995; Cruden, 1998; Cruden and McCaffrey, 2001). The floor to the Mourne pluton was likely located below three kilometers depth for most of the duration of emplacement (see section 2.3.). We therefore deem the inflation of a dome-shaped sill an unlikely emplacement mechanism of the Mourne granite pluton. The southwestern succession of the Mourne granites (G1 to G4) and the shallow northeastward dip of the internal contacts in the Eastern Mourne pluton and granite-graywacke contact in the eastern part of the Western Mourne pluton thus suggest asymmetric floor subsidence. This implies that the area of maximum subsidence is not located in the center of the subsiding floor (Lipman, 1984). The slightly dome-shaped roof of the pluton could therefore be attributed to the bell-jar geometry of the subsidence-related faults that generally form during cauldron subsidence or caldera collapse (Anderson, 1937; Holohan et al., 2015).

A possible mechanism responsible for the deflection of the wall rock parallel to the strike of the contact is viscous indentation of the wall rock during expansion of the pluton (cf. Merle and Donnadieu, 2000; Mathieu et al., 2008; Spacapan et al., 2017; Galland et al., 2018). However, floor subsidence may also have induced shear parallel to the strike of the contact (Fig. 13). Holohan et al. (2013) showed that asymmetric floor subsidence could produce oblique-slip caldera faults. Asymmetric floor subsidence could also explain the opposite rotations inferred (from the rotation model) at the wall-rock localities assuming the area of maximum subsidence is located between the Glen and Bloody Bridge Rivers (Fig. 2A; Table 1), although we lack field observations from the wall-rock outcrops that indicate the sense of shear.

The dike-like tips of the semi-circular northeastern contact could represent part of the feeder to the pluton (Richey, 1927; Walker, 1975; Hood, 1981). The dominantly shallow dips of the magnetic fabric in the area led Stevenson et al. (2007) to favor a feeder in the SW, while they also considered a feeder in the NE a possibility. Since the current exposures of the granite in the northeastern part of the pluton were likely located close to the roof (Fig. 5). The magnetic mineral fabric may therefore record compaction or imbrication of magma as a result of the shear between the magma and the host rock (cf. Knight and Walker, 1988; Geoffroy et al., 2002; Eriksson et al., 2011; Mattsson et al., 2018). We therefore agree with previously presented magma emplacement models (Richey, 1927; Walker, 1975; Hood, 1981; Stevenson et al., 2007), which suggest that the Eastern Mourne granitic magma was fed from the NE and then spread laterally toward the west and SW.

The Western Mourne granite pluton was suggested to be fed from a magma body in the Carlingford Lough with the magma inlet located at the SSW margin of the pluton, based on a NNE-SSW trend of a subhorizontal magnetic lineation in the center of the pluton (Stevenson and Bennett, 2011). The cone sheets surrounding, and pre-dating, the Mourne granite pluton point to the existence of a mafic source magma chamber located below the granites (Emeleus, 1955; Akiman, 1971; Hood, 1981). Such a mafic body could explain the absence of a negative gravity anomaly in the area of the Mourne granite pluton, because the positive gravity anomaly associated with such a mafic pluton could cancel out the common negative gravity signal caused by a granite (Bean, 1953; Bott, 1956; Bott et al., 1958). Notably, the nearby Newry granites show a negative gravity anomaly (cf. Reay, 2004; Stevenson and Bennett, 2011). A mafic source magma body is also consistent with petrological data that indicate that the Mourne granites formed by differentiation from a basaltic magma (Meighan et al., 1984). We therefore propose two possible feeder locations to the Western Mourne pluton. The gently inclined contacts to the E-NE in the eastern part of the Western Mourne pluton indicate floor geometry similar to the Eastern Mourne pluton and together with the observed doming in the west suggest a magma flow from the NE. Alternatively, the NNE-SSW magnetic lineation could indicate a central feeder (Fig. 13).

7. CONCLUSIONS

In this study, we investigated the host-rock deformation related to the emplacement of the Paleogene Mourne granites by using the fracture pattern in the host rock. The fractures in the granite host rock dip steeply and cluster in four major sets striking SE, S, NE, and E. The most dominant fracture sets that strike toward the SE and NE correlate with the regional fracture sets that formed during the Caledonian orogeny. Since rotation of the pluton roof caused by laccolith inflation would have largely shallowed the inclination of the regional fracture sets, the consistently steep dip of the regional fractures in the roof rock does not support significant rotation of the roof, especially where the dip of the contact is ∼20°. However, the wall rock at the northeastern contact to the pluton shows steeply inclined fractures with diverging fracture strikes compared to the regional fracture sets. We propose that their orientations reflect rotation of the wall parallel to the strike of the contact to the Mourne granite pluton.

Several contact-metamorphic segregations in the aureole to the Mourne granites show small displacements (≤1 mm) along fractures. This suggests that the initial emplacement of magma in the Mourne granite pluton caused contact metamorphism but did not cause significant deformation, while later granite emplacement is associated with deformation. Additionally, the displaced and brecciated contact-metamorphic segregations and the occurrence of cataclasite and breccia in the Glen River and Bloody Bridge River traverses indicate that granite emplacement was associated with significant deformation on the northeastern wall of the pluton. The aureole therefore records multiple growth stages of the pluton or different granite pulses emplaced in slightly different fashions.

The lack of pluton-wide rotational deformation in the roof to the granite pluton, the westward younging succession of the granites, and granite-granite contacts inclined toward the NE in the Eastern Mourne pluton suggest that granite emplacement was either (1) facilitated by trap-door floor subsidence combined with local roof uplift in the SW and wall deflection during a later stage of emplacement; or (2) as laccoliths and bysmaliths with a curved floor, which explain the lack of distinct rotation of the roof. The bulk of the uplift would then have been facilitated by a fault. However, we favor the first scenario, because a curved (convex) sill geometry as required for the second emplacement scenario is not commonly observed in nature (see section 6). Furthermore, we suggest that the rotation of the wall rock parallel to the strike of the contact to the pluton, as inferred by the rotation model, was related to asymmetric floor subsidence and/or viscous indentation of the wall rock.

ACKNOWLEDGMENTS

This project is partly financed by the European Union’s INTERREG IVA Cross-border Programme managed by the Special EU Programmes Body via the Mourne Cooley Gullion Research bursary to TM and SB and by the Swedish Research Council (Vetenskapsrådet) grant 2015-03931_VR. Harri Geiger, Frances Deegan, and Valentin Troll collected the samples with the help of TM and SB. Harri Geiger packed, shipped from the Mourne Mountains to Uppsala, and cut the samples with help from TM. We thank Marcus Fiola for assistance in the field and Siobhan Powers for establishing the Mourne Cooley Gullion Research bursary. Discussions with Olivier Galland, Alban Souche, Ole Rabbel, Tobias Schmiedel, and Erika Ronchin, and comments by Sven Morgan and Thomas Blenkinsop improved the manuscript. We thank Eric Horsman, Carl Stevenson, and the Associate Editor Robert B. Miller for informative reviews that greatly improved the manuscript. We thank Ian Meighan for providing a copy of D.N. Hood’s Ph.D. thesis. The authors acknowledge the use of the Move Software Suite granted by Midland Valley’s Academic Software Initiative/Petroleum Experts and LIME provided by Uni Research Centre for Integrated Petroleum Research (CIPR), University of Bergen.

1Supplemental Materials. Text file describing supplemental material (includes five figures). Table S1 and S4 and virtual scan line data. Table S2 provides all structural data (fractures, contacts, bedding, joints, and lineaments) collected during study. Table S3 provides all scan line data. Please visit https://doi.org/10.1130/GES02148.S1 or access the full-text article on www.gsapubs.org to view the Supplemental Materials.
Science Editor: Andrea Hampel
Associate Editor: Robert B. Miller
Gold Open Access: This paper is published under the terms of the CC-BY-NC license.