December 2024 LIP of the Month

Controls on long-range lateral dyke propagation: insights from the Mull Dyke Swarm, NW Europe

Joe Cartwright,

Department of Earth Sciences, University of Oxford, South Parks Road, Oxford OX1 3AN, UK

joe.cartwright@earth.ox.ac.uk

Introduction

Individual dykes within giant dyke swarms commonly exceed lengths of 100 km and in numerous cases exceed 1000 km (El Bilali and Ernst, 2024). These extraordinary dimensions for magma pressure-driven fractures prompt questions regarding the controls on lateral dyke propagation in the crust of the terrestrial planets and the means by which the propagation is regulated. Long-range lateral propagation requires that the dykes in question do not approach the surface and with sufficient magma pressure to erupt, since were eruption to happen vigorously the dyke would lose its necessary magma supply to continue to propagate laterally (Ernst et al. 1995; Parfitt and Wilson, 2009).

One of the reasons that this problem of controls on upper tip propagation remains open is that the upper tip regions of dykes in giant dyke swarms are either eroded away (typically the case on Earth) or are inaccessible to direct observation because they are concealed beneath the modern surface (e.g. on Venus and Mars). However, the wider accessibility of reflection seismic data acquired originally for oil exploration means that giant dyke swarms can be investigated in situ and in rare cases with their upper tips preserved (Wall et al. 2010; Magee et al. 2022). In this summary, we report some observations from the Mull Dyke Swarm (NW Europe) made using petroleum exploration data (seismic and wells) that may have a fundamental bearing on this problem, or at the very least, serve as an example of the great potential of integrating classical field and geochemically-based methods with seismic interpretation.

The Mull Dyke Swarm (MDS)

The MDS is one of the most intensely studied of the terrestrial giant dyke swarms is the Mull Dyke Swarm (MDS) and is probably the source of the term ‘dyke’ itself, having been used first by Playfair in the late 1790s to describe vertical walls of dark rock that were reminiscent of the stone walls built by landowners across northern Britain. The word ‘dyke’ itself derives from the Old Norse word for wall or ditch.

The MDS was emplaced during the late Palaeocene/early Eocene from a central source region on the island of Mull, one of the main igneous centres for the Hebridean Magmatic Province (part of the British and Irish Palaeogene Igneous Province) (Bailey et al., 1924; Kerr et al., 1999). This swarm consists of hundreds of individual dykes intruded in an approximately NW-SE zone some tens of kilometres wide with its locus in the three aligned sub-volcanic centres on Mull (Fig. 1). These centres provided the magma reservoirs for the development of the giant swarm (Tyrrell, 1917; Ritchie, 1939; MacDonald et al. 1988; Kerr et al., 1999).

Early mapping suggested that constituent dykes of the swarm had propagated across the Midland Valley and the Southern Uplands as far as northern England (Geikie 1897; Tyrrell, 1917; Ritchie, 1939). Dykes of post-Carboniferous age had long been recognised and mapped in detail in the coalfields of northern England (Winch, 1815; Teall, 1884). More modern mapping, geochemical fingerprinting and age dating (MacDonald et al. 2015, Ishizuka et al. 2017) has confirmed the earlier inferred connection between the Tertiary dykes of northern England and the Mull Centre (Geikie, 1897; Tyrrell, 1917; Holmes and Harwood, 1929).

The continuation of the MDS much further to the south-east into the North Sea was first established using seismic and aeromagnetic data in a pioneering study by Kirton and Donato (1985). This was substantiated by additional geophysical mapping of several dykes (Brown et al. 1994; Gauer et al.2004 and Wall et al. 2010).

More recently, interpretation of a large 3D seismic survey of the southern North Sea led to updated mapping of individual dykes for distances of over 200 km to their lateral termination.

The extended mapping of the MDS from onshore outcrop into the subsurface of the southern North Sea Basin offers an exceptional opportunity to examine the geometry of a buried giant dyke swarm preserved from exhumation and erosion with its upper structure intact by virtue of post-emplacement subsidence and burial under the Cenozoic basin fill.

Main Observations

A combination of 2D and 3D seismic, well-log and aeromagnetic data was used to identify the dykes (Carver et al. 2023). Individual dykes were interpreted from their association with vertical to sub-vertical seismic disturbance zones (SDZs) (Fig. 2). Seismically mapped dykes coincided closely with prominent linear, negative aeromagnetic anomalies, modelled with best fits corresponding to vertical to steeply dipping negatively magnetised dykes with thicknesses of the order of 10 m, terminating upwards within the Upper Cretaceous (Brown et al. 1994; Wall et al. 2010; Carver et al. 2023).

The SDZs are prominent wherever the upper dyke tips occur beneath erosive craters (Fig. 2). These craters can be mapped almost continuously for over 200 km along strike of the central sub-swarm of the MDS. The dyke trajectories are best revealed by the alignment of the craters, or by amalgamation of individual craters into more linear troughs (Fig. 3) (Pryce et al. 2024). Several important borehole calibrations have been found that provide clear evidence that the crater fills comprise tuffs and reworked chalky sediments. In addition, we mapped a well-defined tephra apron flanking the craters for over 150 km of strike length, calibrated by boreholes to show that it consists of reworked Upper Cretaceous sediments, excavated to form the craters (Pryce et al. 2024) (Fig. 3).

The upper dyke tips reach within 200-500m of the contemporaneous surface, tipping out in highly porous Chalk Group sediments (Fig. 2). Graben structures similar to those reported from a NW Australian dyke swarm by Magee et al. (2022) are interpreted along portions of the MDS but account for less than 25% of the strike length of the dykes mapped using 3D seismic data. These typically detach into underlying evaporite sequences and their geometry does not provide a reliable guide to upper tip positions.

The combination of seismic and aeromagnetic mapping resulted in a revised map of the MDS (Fig. 4). Of the three main sub-swarms, the central sub-swarm (the Blyth Group) extends furthest from the source in Mull, with the most distal dykes mapped 672 km from the centre of the putative feeder beneath the sub-volcanic ring complexes of central Mull. The new mapping shows that the distal termini do not correspond to any significant upper crustal structure. All three sub-swarms contain dykes that cross all the major crustal obstacles such as terrane boundaries, major thrusts or strike-slip faults and even the Iapetus Suture Zone. The gross curvature of the dyke swarm seen on the map has changes in orientation of about 30 degrees in two positions, one on land near the Southern Uplands Fault, and the second offshore close to the northern tip of the major Dowsing Fault Zone.

The 3D seismic mapping in particular shows some of the details of sub-swarm architecture, including clear examples of segmentation and downstream bifurcation. More significantly, the two longest dykes belonging to the central sub-swarm terminate within a few kilometres of each other, even though they are typically 10 km or more apart for most of their offshore trajectory, suggesting that these two prominent dykes were sourced at exactly the same time.

Discussion

Relationship to basement structure

The observation that all three component sub-swarms cross a series of major crustal boundaries without any significant deviation in strike or position demonstrates the tremendous potency of magma pressure-driven fractures. The major dykes of the MDS evidently propagated with well-defined trajectories defined by the local intermediate compressive stress direction (σ₂ ) (England, 1997). Curvature in σ₂ direction may reflect the influence of the magma source closer to Mull (R. Ernst, pers. comm., 2024), but the more distal change in strike may reflect the change to a dominant WNW-ESE crustal fabric evident in many of the upper crustal faults in the southern North Sea.

Timing of Dyke Intrusion

Only limited direct radiometric dating has been undertaken on dykes linked to the MDS. K-Ar dates of 58.4 +/- 1.1 Ma (Evans et al. 1973) and 59.3 +/- 2.0 Ma (Mitchell et al. 1989). Magnetic modelling of dykes in the MDS on the Scottish mainland (Dagley et al. 2008; Busby et al. 2009), northern England (Robson, 1964) and in the North Sea (Kirton and Donato, 1985; Brown et al. 1994; Gauer et al. (2004) and Wall et al. (2010) show that the dykes are all negatively magnetized, and must therefore have been intruded during a period of reversal of the magnetic field. From a synthesis of Ar-Ar dating of extrusive and intrusive components of the Mull Centre, Chambers and Pringle (2001) have argued convincingly that all the igneous activity occurred between c. 61 Ma and 58 Ma, within Chron C26R. It seems most likely therefore that all three sub-swarms were emplaced at some point during C26R.

Independent corroboration of this age range comes from biostratigraphic analysis of boreholes penetrating one of the craters overlying a major dyke of the central sub-swarm and its associated tephra apron (Pryce et al. 2024). These boreholes yielded a Late Palaeocene age of c. 58 Ma, consistent with radiometric and magnetostratigraphic evidence. This age is also consistent with a recent review of the geochemical evidence for linking the distal dykes of the MDS with the main igneous components on Mull. In this review, Ishizuka et al. (2017) present a detailed petrogenetic model for the MDS in which they argue that the distal dykes (their Types 1-3) were contemporaneous with the activity of Mull Centre 1 (Kerr et al. 1999), and that all the distal dykes formed after the Mull Plateau Group lavas. The wide spatial separation of the three sub-swarms implies at least three separate intrusive episodes. From the chronology established by Chambers and Pringle (2001), this would suggest an age range encompassing the three sub-swarms a maximum of a million years between 59 and 58 Ma.

Craters above upper dyke tips: implications for dyke arrest

Craters above some of the major dykes in the central sub-swarm have now been mapped for over 200 km along strike. Wall et al. (2010) argued that the craters could have formed either phreatomagmatically (they invoked the Maar-Diatreme model of Lorenz, 1986) or by escape of juvenile volatiles from the dyke tip (c.f. Scott and Wilson, 2002). Based on detailed borehole calibration of craters and associated tephra apron, Pryce et al. (2024) argued that a phreatomagmatic eruptive mechanism best explains the genesis of the crater arrays. A phreatomagmatic eruptive mechanism neatly explains all the key observations namely (1) the highly erosive nature of the craters themselves, (2) the thick development of tuffs in the crater fill and (3) the reworked chalks in the tephra apron. The newly mapped tephra apron cannot be explained by any collapse mechanism to form the craters but would be expected in any eruptive mechanism where magma fragmentation combined with explosive expansion of gas phases was involved (Lorenz, 1986).

Removal of >200 m of chalky sediments during erosional excavation of the craters (Fig. 2) is best explained by a highly vigorous eruption. High pressure venting of supercritical steam generated by magmatic flash heating of pore water in contact with the upper tips of the dykes as they entered the highly porous Chalk Group would be capable of such deep excavation (Delaney, 1982). Volcaniclastic debris mixed with excavated chalky sediments would be blasted upwards into the water column, with some of this material settling out to form tephra aprons. Dyke cooling and pore pressure deflation would have led to localised compaction and subsidence of brecciated sediments beneath the crater along with crater widening by failure of crater flank deposits.

The rapid cooling and fragmentation of the magma in the upper dyke tips involved in a phreatomagmatic eruption may have been a crucial factor in the upward arrest of the dykes below the surface and the prevention of surface eruption of magma. We do not know if the craters originally formed along the entire 650 km lateral extent of the MDS since the upper tips are only preserved for the most distal 250 km. However, their ubiquitous development wherever the upper tips entered porous sedimentary layers points to an important but as yet unquantified role for rapid cooling in upward dyke arrest. Other factors may be primarily responsible for upward arrest, such as the loss of buoyancy drive, or magma pressure losses due to narrowing of dyke aperture (Lister and Kerr, 1991; Fialko and Rubin, 1999). However, if phreatomagmatic craters are proven to occur more widely above giant dyke swarms elsewhere, as demonstrated here for the MDS, then near surface cooling through groundwater/magma interactions may warrant more serious consideration as a prime control on long-range lateral propagation.

Volumetric considerations and implications for long-range lateral dyke propagation

MacDonald et al. (1988) provided the first volumetric estimate for one of the major dykes suggesting a volume of 85 km³ based on physical arguments and surface observations. We updated these volumes based on the new mapping but following their assumptions regarding dyke height and thickness. Based on these assumptions, we obtain values ranging from 90 km³, 97 km³ and 202 km³ for three main intrusive ‘events’. Note that these are volumes solely for the long-range dykes, and exclude the hundreds of thinner, shorter dykes emplaced within a 50-100 km radius from Mull (Jolly and Sanderson, 1995) and also the longer dykes propagated to the northwest of Mull (Fig. 1). These intruded volumes are also conservative given the likelihood that near surface dyke widths are probably not maximum values (see discussion in MacDonald et al. 1988).

MacDonald et al. (1988), Kerr et al. (1999) and Ishizuka et al. (2017) have all suggested that the emplacement of the large volumes of magma to drive the propagation of the long-range dykes would have possibly involved caldera collapse of the Mull volcano. It is not clear whether such a caldera collapse occurred during the phase of activity linked to specific sub-volcanic centres on Mull, but from the most recent geochemical fingerprinting (Ishizuka et al. 2017), it seems likeliest that the distal dykes were related to Centres 1 or 2, which have both been linked to caldera collapse (Kerr et al. 1999). For scaling purposes, the three long-range sub-swarms could have been supplied by magma by complete evacuation of a 1 km thick tabular magma chamber that ranged between 8-14 km in diameter (neglecting compressibility). Kerr (1999) and MacDonald et al. (2015) have also suggested that such large magma volumes depleted from any sub-volcanic magma reservoir would probably have involved a highly explosive eruption on a large scale in addition to the caldera collapse (c.f. Geshi et al. 2020). For comparison, the intrusion of the Bardarbungar dyke some 48 km away from the source, resulted in formation of a caldera with an area of 110 km² (Gudmundsson et al., 2016; Woods et al. 2019).

Such large volumes for single intrusive events are challenging to explain with current models of magma chamber dynamics. Considerable progress has been made in recent years on the interplay between magma chamber failure by dyke propagation, but has only been applied to short range dyke propagation with modest intrusive volumes (McLeod and Tait, 1999; Pinel and Jaupart, 2004; Buck et al. 2006; Gudmundsson et al. 2016). The requirement for long-range propagation of dykes within giant dyke swarms in general is a relatively uniform flux of magma into the conduit localised at the margins of the magma reservoir over a period which may extend to weeks or months but not years (MacDonald et al. 1988; Ishizuka et al. 2017). The documentation here of large volume, multiple intrusions of long-range dykes from a single volcanic complex may hopefully help to stimulate further research into the interplay of the complex factors governing the formation of giant radial dyke swarms on Earth and the terrestrial planets.

Conclusions

1. New mapping of the Mull Dyke Swarm in the southern North Sea using seismic and aeromagnetic data has proven a continuation of the major component dykes for distances of up to 672 km from their source on Mull. Three distinct sub-swarms are recognised in the mapping of the offshore region in the southern North Sea, from north to south referred to as the Acklington, Blyth and Cleveland Dyke Groups.

2. The three main axes of intrusion probably formed in different intrusive events within a c. 1 million year period , from 59 to 58 Ma, during magnetic chron C26R.

3. Intrusive volumes were of the order of 100 to 200 km³, suggesting caldera collapse of the Mull Volcano as the likeliest driving mechanism.

References