Reactivation of Magma Pathways: Insights From Field Observations, Geochronology, Geomechanical Tests, and Numerical Models

Abstract

Field observations and unmanned aerial vehicle surveys from Caldera Taburiente (La Palma, Canary Islands, Spain) show that pre-existing dykes can capture and re-direct younger ones to form multiple dyke composites. Chill margins suggest that the older dykes were solidified and cooled when this occurred. In one multiple dyke example, an 40Ar/39Ar age difference of 200 kyr was determined between co-located dykes. Petrography and geomechanical measurements (ultrasonic pulse and Brazilian disc tests) show that a microscopic preferred alignment of plagioclase laths and sheet-like structures formed by non-randomly distributed vesicles give the solidified dykes anisotropic elastic moduli and fracture toughness. We hypothesize that this anisotropy led to the development of margin-parallel joints within the dykes, during subsequent volcanic loading. Finite element models also suggest that the elastic contrast between solidified dykes and their host rock elevated and re-oriented the stresses that governed subsequent dyke propagation. Thus, the margin-parallel joints, combined with local concentration and rotation of stresses, favored the deflection of subsequent magma-filled fractures by up to 60° to form the multiple dykes. At the edifice scale, the capture and deflection of active intrusions by older ones could change the organization of volcanic magma plumbing systems and cause unexpected propagation paths relative to the regional stress. We suggest that reactivation of older dykes by this mechanism gives the volcanic edifice a structural memory of past stress states, potentially encouraging the re-use of older vents and deflecting intrusions along volcanic rift zones or toward shallow magma reservoirs.

Full text

1. Introduction
Dykes are a common means of magma transport in magma plumbing systems (Rivalta etal.,2015). Individ-
ual dyking events can transport magma laterally up to tens to hundreds of kilometers from its point of origin
(e.g., Neal etal.,2018; Sigmundsson etal.,2014), and vertical magma transport through dykes is often in-
ferred to explain recharge of shallow magma chambers (Karlstrom etal.,2010). Understanding likely prop-
agation paths of future dykes is therefore an important component of hazard assessment in volcanic areas.
Dykes propagate as fluid-driven fractures (Gudmundsson,2011). Their propagation and or arrest is con-
trolled by: (1) the magma pressure and buoyancy; (2) the mechanical properties and structure of the host
rock; (3) the stress field into which they intrude; and (4) the temperature and rheology of the magma (Ri-
valta etal.,2015). Theoretical and field studies of dyke propagation have highlighted the importance of
pre-existing structures in the host rock. For example, it is well established that contacts between units
of different elastic stiffness (e.g., lava flows vs. pyroclastic layers) promote dyke arrest and the formation
of sills (Gudmundsson,2005a, 2011; Kavanagh et al.,2006). Similarly, many authors have suggested that
dykes tend to propagate along regional and volcanic faults (e.g., Browning & Gudmundsson,2015; Gaffney
etal.,2007; Valentine & Krogh,2006; van den Hove etal.,2017).
Dykes formed from two or more parallel and co-located intrusions are also commonly observed in volcanic
regions. Extreme examples of these include sheeted dyke complexes exposed in ophiolites (Gass, 1968),
mid-ocean ridges (Stewart etal.,2002), and ocean-island volcanoes (Walker,1992), in which dykes intrude
so densely that they comprise >90% of the rock mass, often making it impossible to distinguish individual
Abstract Field observations and unmanned aerial vehicle surveys from Caldera Taburiente (La
Palma, Canary Islands, Spain) show that pre-existing dykes can capture and re-direct younger ones to
form multiple dyke composites. Chill margins suggest that the older dykes were solidified and cooled
when this occurred. In one multiple dyke example, an 40Ar/39Ar age difference of 200 kyr was determined
between co-located dykes. Petrography and geomechanical measurements (ultrasonic pulse and Brazilian
disc tests) show that a microscopic preferred alignment of plagioclase laths and sheet-like structures
formed by non-randomly distributed vesicles give the solidified dykes anisotropic elastic moduli and
fracture toughness. We hypothesize that this anisotropy led to the development of margin-parallel joints
within the dykes, during subsequent volcanic loading. Finite element models also suggest that the elastic
contrast between solidified dykes and their host rock elevated and re-oriented the stresses that governed
subsequent dyke propagation. Thus, the margin-parallel joints, combined with local concentration and
rotation of stresses, favored the deflection of subsequent magma-filled fractures by up to 60° to form the
multiple dykes. At the edifice scale, the capture and deflection of active intrusions by older ones could
change the organization of volcanic magma plumbing systems and cause unexpected propagation paths
relative to the regional stress. We suggest that reactivation of older dykes by this mechanism gives the
volcanic edifice a structural memory of past stress states, potentially encouraging the re-use of older vents
and deflecting intrusions along volcanic rift zones or toward shallow magma reservoirs.
THIELE ET AL.
© 2021. The Authors.
This is an open access article under
the terms of the Creative Commons
Attribution-NonCommercial License,
which permits use, distribution and
reproduction in any medium, provided
the original work is properly cited and
is not used for commercial purposes.
Reactivation of Magma Pathways: Insights From Field
Observations, Geochronology, Geomechanical Tests, and
Numerical Models
Samuel T. Thiele1 , Alexander R. Cruden2 , Xi Zhang3 , Steven Micklethwaite4 , and
Erin L. Matchan5
1Helmholtz Institute Freiberg for Resource Technology, Helmholtz-Zentrum Dresden-Rossendorf, Freiberg, Germany,
2School of Earth, Atmosphere and Environment, Monash University, Melbourne, VIC, Australia, 3SCT Operations Pty
Ltd, Wollongong, NSW, Australia, 4Sustainable Minerals Institute, University of Queensland, Brisbane, QLD, Australia,
5School of Earth Sciences, The University of Melbourne, Parkville, VIC, Australia
Key Points:
• Dykes form significant and highly
oblique mechanical discontinuities
in volcanic edifices
• Local stresses and mechanical
anisotropy within solidified dykes
can capture and deflect subsequent
intrusions to form a multiple-dyke
• Reactivation of magma pathways by
these mechanisms influences the
emergent organization of magma
plumbing systems
Supporting Information:
Supporting Information may be found
in the online version of this article.
Correspondence to:
S. T. Thiele,
s.thiele@hzdr.de
Citation:
Thiele, S. T., Cruden, A. R., Zhang,
X., Micklethwaite, S., & Matchan,
E. L. (2021). Reactivation of
magma pathways: Insights from
field observations, geochronology,
geomechanical tests, and
numerical models. Journal of
Geophysical Research: Solid Earth,
126, e2020JB021477. https://doi.
org/10.1029/2020JB021477
Received 4 DEC 2020
Accepted 19 APR 2021
10.1029/2020JB021477
RESEARCH ARTICLE
1 of 17

Journal of Geophysical Research: Solid Earth
dykes. In less extreme cases, internal chilled margins or compositional variations are most simply explained
by multiple co-located magma injections. These dykes are commonly referred to as “multiple dykes”
(known as “composite dykes” when the different injections have contrasting compositions), although for
brevity and to avoid grammatical confusion we prefer to use the term “multi-dyke.
While geochemical variations within multi-dykes have been widely studied (e.g., Ehlers & Ehlers,1977;
Esteve etal.,2014; Gibson & Walker,1963; Kanaris-Sotiriou & Gibb,1985; Sun et al.,2013), literature on
the mechanics of their formation is lacking. Indeed, it is commonly assumed that multi-dykes form either
because the initial dyke did not have time to solidify completely before the subsequent injection, or because
the solidified dyke or its chilled margin were weaker than the host rock it intruded (Ehlers & Ehlers,1977;
Esteve etal.,2014; Gudmundsson,1984; Guppy & Hawkes,1925; Marinoni & Gudmundsson,2000; Plat-
ten,2000; Walker,1992).
Here we present observations of exceptionally well-exposed multi-dykes on the island of La Palma (Canary
Islands, Spain) that appear not to have formed by either of these mechanisms.
2. The Taburiente Dyke Swarm
The Taburiente dyke swarm is found within the ∼2–0.5Ma Volcán Taburiente (Carracedo etal.,1999), a
large basaltic edifice that forms the northern portion of the island of La Palma (Canary Islands, Spain; Fig-
ure1a). At the heart of this edifice a deeply incised and eroded collapse-scarp forms a bowl-like depression,
Caldera Taburiente, bounded by spectacular cliffs up to ∼1km high. This highly eroded landscape provides
exceptional exposure of the volcano's shallow plumbing system, revealing a complete stratigraphic section
through the most recent eruptive products to the edifice basement.
The Taburiente dykes radiate from a focal point in the southern part of the caldera, as described in detail in
Thiele etal.(2020). Along the northern side of the caldera these dykes crosscut an earlier NE striking and
shallower-dipping (∼45–60°) dyke set interpreted to have formed relatively early in the growth of Volcán
Taburiente (Thiele etal.,2020). Flow lineations indicate sub-horizontal flow both toward and away from the
focal point, suggesting that ascending dykes became radially oriented as they interacted with topographic
loading below the Taburiente edifice and then began to flow laterally to form blade-shaped dykes (Thiele
etal.,2020).
3. Field Observations and Mapping
To gain access to the steep and unstable exposures within Caldera Taburiente, images collected via un-
manned aerial vehicle (UAV) were used to construct three-dimensional (3-D) digital outcrop models using
a structure-from-motion multi-view-stereo photogrammetric workflow (SfM-MvS; cf. Bemis et al.,2014;
Dering et al.,2019). These digital outcrop models and details of the methods used to construct them are
described in (Thiele etal.,2019,2020). For this study, we focus on three surveys from a site known locally
as Hoyo Verde (Figure1a), where dykes of different orientations intersect to form multi-dykes (Figure1b).
The dykes at Hoyo Verde intrude volcaniclastic breccia deposits (Figure1c), welded scoria (Figure1d) and
finely laminated palagonite tuffs (Figure1e). Multi-dykes were observed in all of these lithologies. The
volcaniclastic breccias occur in >10m thick, polylithic, poorly sorted, and matrix-rich beds toward the
bottom of the section, but become finer-grained and more well-sorted toward the middle of the section. The
breccias are overlain by thick-bedded scoria deposits. Beds of laminated palagonite tuff of 1–20m thickness
occur in both the breccia and scoria units, and dip ∼20–30° north-west.
Mapping of the dyke network (Figure2) shows that older, more shallowly dipping (∼45°) dykes appear to
capture younger intrusions and re-orient them by up to 60°, forming thick multi-dyke bundles typically
characterized by complex cross-cutting relationships. While some of the captured dykes propagate along the
contact of the older dyke, many also propagate along dyke cores. Although the resolution of the models is
not sufficient to accurately track individual dykes through these bundles, dykes are observed to re-emerge
from the tops of the bundles after tens of meters, suggesting that in some cases the capture is transient.
THIELE ET AL.
10.1029/2020JB021477
2 of 17

Journal of Geophysical Research: Solid Earth
The dykes are basaltic, variably vesiculated and sometimes crosscut by 5–15cm spaced cooling joints (Fig-
ure3). Some dykes, typically those with fewer cooling joints, are also crosscut by well-developed sets of
internal margin-parallel joints (MPJs). These MPJs are orthogonal to the cooling joints, and can be observed
in dykes throughout La Palma, though their length and spacing varies greatly: some form shale-like fracture
cleavages (Figure3a) while others persist laterally over many meters and are closely (Figures3b and3c) to
widely spaced (Figure3d). Similar sets of MPJs have also been observed in basaltic dykes from other volcan-
ic islands (Delcamp etal.,2012; Porreca etal.,2006) and the Troodos ophiolite (Kidd & Cann,1974), and are
probably common; our fieldwork suggests they are present in ∼40% of the dykes on La Palma.
THIELE ET AL.
10.1029/2020JB021477
3 of 17
Figure 1. Location of La Palma, Hoyo Verde, and sites where dykes were sampled for microstructural and geomechanical testing (red triangles) (a). Cliffs at
Hoyo Verde are crosscut by moderately to steeply dipping dykes (b) that intrude volcaniclastic breccia (c), bedded scoria (d) and finely laminated palagonite
tuff (e). A ∼30m thick sill crosscuts dipping scoria layers near the top of the section, and breccias at the base of the section contain several ∼1m thick basaltic
units that could be sills or the cores of lava flows. A lower hemisphere stereographic projection (stereonet) with density contours of poles to the older, shallower
dipping dykes (blue) and steeper cross cutting dykes (red) is provided in (b) for reference, along with a stratigraphic log interpreted from UAV images. A total of
46,715 dyke orientation estimates were extracted from the UAV data using the method of Thiele, Grose, etal.(2019). The average orientation of the cliff is given
by the black triangle.

Journal of Geophysical Research: Solid Earth
The margin-parallel orientation of these joints suggests they are not related to the thermal stresses that
form cooling joints, which should propagate inwards and parallel to the thermal gradient to form joints per-
pendicular to the dyke margin (Budkewitsch & Robin,1994). Hence, their formation remains unexplained,
although Porreca etal.(2006) tentatively suggested that similar joints observed on Mount Somma-Vesuvio
could result from elevated stresses within the dykes during volcanic loading.
Many of the dykes contain non-randomly distributed vesicles, which form margin-parallel sheets at regu-
larly spaced intervals near the dyke margins (Figures4a and4b) or as a single sheet in the dyke core (Fig-
ure4c). MPJs commonly link vesicles within these sheets (Figure4d); similar structures in dykes on Tener-
ife have been described by Delcamp etal.(2012). These authors attribute these vesicle sheets to preferential
THIELE ET AL.
10.1029/2020JB021477
4 of 17
Figure 2. Anastomosing dyke network exposed in a cliff face above Hoyo Verde. Mapping of these using the digital
outcrop model (a) shows a set of shallow-dipping (∼45° NW) dykes capturing and re-orienting a later generation of
steeply dipping dykes (b, c), The acute intersection angles between the older intrusions and captured dykes have been
measured using structure-normal estimates (cf. Thiele, Grose, etal.,2019) created during the digital outcrop mapping
and are shown in (a).

Journal of Geophysical Research: Solid Earth
degassing pathways that form along straight bands parallel to the dyke margins, although we speculate that
such sheets could form when vesicles adhere to solidifying dyke margins during periods of lower magma
pressure or flux, and in the case of sheets within dyke cores, during the final stages of dyke activity. Regard-
less of the mechanism of their formation, sheet-like concentrations of vesicles would significantly weaken
these parts of the dyke and promote fracture growth, especially if they have been stretched into non-spher-
ical shapes during cooling (Bubeck etal.,2017; Delcamp etal.,2012).
Finally, internal contacts within multi-dykes at many locations have distinct glassy chilled margins up to
∼5mm thick (Figures5a–5c), suggesting that the external dyke cooled prior to the injection of the subse-
quent intrusion. Some of the external dykes at Hoyo Verde also show peperitic margins (Figure5d), indicat-
ing that the host rocks were water-saturated during the emplacement of the earliest intrusions.
THIELE ET AL.
10.1029/2020JB021477
5 of 17
Figure 3. A selection of different margin-parallel joint styles observed in dykes throughout La Palma, ranging from
shale-like fracture cleavages (a) to persistent (b) and sometimes imbricated joints (c) and more widely spaced but highly
persistent (tens of meters) “tram-track” joints (d). High resolution versions of these images can be downloaded from
FigShare (Thiele,2020) for closer inspection.

Journal of Geophysical Research: Solid Earth
3.1. 40Ar/39Ar Geochronology
To test our interpretation that the external dykes cooled completely prior to intrusion of the internal ones,
two samples were collected from an accessible multi-dyke at Hoyo Verde for 40Ar/39Ar geochronology. This
location was chosen because (1) neither dyke showed any sign of alteration and (2) both the internal and ex-
ternal dykes could be sampled several meters from the point where they intersect, minimizing the potential
for thermal overprinting of the external dyke by the internal one. Studies of dyke cooling (e.g., Bonneville &
Capolsini,1999) suggest that temperatures at distances of greater than a few meters from a ∼1m thick dyke
should not exceed the closure temperature of the potassium-argon system, especially if host rock pore-flu-
ids rapidly advect heat away from the dyke.
Following petrographic inspection (Figures 6a and 6b), groundmass concentrates were separated from
crushed samples of HV4 (external dyke) and HV6 (internal dyke) and irradiated according to standard
40Ar/39Ar techniques (see Supporting information for details). Aliquots of irradiated groundmass were ana-
lyzed via the 40Ar/39Ar step-heating method following procedures described by Matchan and Phillips(2014).
Samples from HV4 and HV6 yielded plateau ages of 796±4 ka (MSWD=0.75; P=0.67) and 596±8 ka
(MSWD= 0.53; P=0.71) respectively (Figures6c–6f ). These values are interpreted as the emplacement
ages of the dykes. The ∼200 kyr difference indicates that the external dyke had completely solidified and
cooled before it was intruded by the younger dyke. A detailed analysis of the robustness and significance of
these Ar-Ar plateau ages is included in the Supporting Information.
3.2. Dyke Microstructure and Mechanical Properties
To gain further insight into the formation of the MPJs and multi-dykes, oriented samples from the cores and
margins of 10 different dykes were collected at four locations (Figure1a). Thin sections (n=26) oriented
perpendicular to the dyke margins were prepared, along with ∼22mm3 ultrasonic pulse specimens (n=25)
and pairs of 25mm thick and 45mm diameter Brazilian discs (n=17 pairs). The ultrasonic pulse and Bra-
zilian disc specimens were selected to represent intact rock, taking care to avoid cooling joints and MPJs.
Inspection of the thin sections reveal that most of the dykes contain preferentially aligned and ∼0.5mm
long plagioclase laths (Figure7a). Near the margins of some dykes, rotation of plagioclase lathes relative
to this preferential alignment suggest the presence of small shear-bands (Figure7b). Imbrication of (and
shearing between) domains of aligned minerals is commonly observed at the margins of igneous intrusions,
and generally attributed to shearing during magma flow (Holness & Humphreys,2003).
MPJs are generally parallel to the plagioclase alignment, although in some cases they appear to have formed
along the (older) shear-bands instead (e.g., Figures7a and7b). The MPJs show no evidence for significant
shear offset which, combined with the irregular fracture surfaces, suggests that they formed as predom-
inantly Mode I fractures. Where vesicle bands are present, MPJs link closely spaced vesicles (Figure 7c).
THIELE ET AL.
10.1029/2020JB021477
6 of 17
Figure 4. Alternating vesicle-rich and -poor layers (a) and distinct vesicle sheets (b) commonly observed near dyke margins. Dykes with vesiculated cores
(c) were also observed, although less frequently. Margin-parallel joints appear to sometimes form preferentially in the vesicle-rich layers by linking individual
vesicles together (d). High resolution versions of these images can be downloaded from FigShare (Thiele,2020) for closer inspection.

Journal of Geophysical Research: Solid Earth
These plagioclase alignments and MPJs are similar to those observed in basaltic dykes in the Azores (Morei-
ra etal.,2015), Tenerife (Delcamp etal.,2012) and sills on the Isle of Mull (Holness & Humphreys,2003).
Delcamp etal.(2012) interpret that these alignments give basaltic dykes on Tenerife a preferential parting
direction, which they also relate to MPJs. Our observations corroborate this hypothesis.
Results from the ultrasonic pulse tests indicate that the dykes have anisotropic elastic properties (Figure8),
which we attribute to the plagioclase alignment and vesicle distribution, although other structures such
as microfractures could also play a role. P- and S-wave velocities were calculated in three perpendicular
directions by measuring the travel time of 50kHz ultrasonic pulses over distances of ∼20mm. Proper cou-
pling between the ultrasonic pulse transmitter and each sample was ensured using a coupling agent and by
applying a small force to the transmitter and receiver pads. As the flow direction of the dykes are unknown
THIELE ET AL.
10.1029/2020JB021477
7 of 17
Figure 5. Black, glassy chill margins (a, b, c) suggesting the exterior dyke was cooled prior to intrusion of the interior
one. Black arrows point toward glassy chill margins. The external dyke in (c) contains abundant vesicles, unlike the
interior one. External dyke margins are also glassy (d), and locally develop peperitic textures. The pen used for scale in
(c) and (d) is ∼14cm long.

Journal of Geophysical Research: Solid Earth
THIELE ET AL.
10.1029/2020JB021477
8 of 17
Figure 6. Cross-polarised (XP) photomicrographs of the two dykes sampled for geochronology, HV4 (a) and HV6 (b). The resulting heating spectra show that
emplacement (plateau) ages of the external (c) and internal (d) dykes are separated by ∼200 ka. Plateau steps are green, rejected steps are cyan. Inverse isochron
diagrams (e, f) are also shown for each dyke. Uncertainties are all 2σ.

Journal of Geophysical Research: Solid Earth
and likely variable, measurements in the dyke-parallel direction (which we refer to as L and X) represent an
arbitrary coordinate system within the flow-plane, while the dyke-perpendicular measurements (P-direc-
tion) is consistently perpendicular to the dyke margin (and flow direction).
The ultrasonic pressure- and shear-wave velocities VP and VS were used to estimate the Young's (E) and
shear (G) moduli in each of these directions, using Equations 1 and 2. Bulk-density (p) values for this
calculation were obtained by calculating the buoyancy of each sample in water, as described in detail by
Houghton and Wilson(1989).



22 2
22
34
SP S
PS
VV V
Ep
VV
(1)
2
S
G pV
(2)
Paired t-tests found no significant difference between the L- and X-directions, but comparisons with the
P-direction suggest significant differences in both E and, to a lesser extent, G (Table1). This indicates that
the dykes are an average of ∼8% stiffer under dyke-parallel strains than dyke-perpendicular ones, with
anisotropies of up to 25% measured on individual samples. Anisotropy of Poisson's ratio, which has been
shown to control stress rotation in anisotropic materials (Faulkner etal.,2006; Healy,2008), is also ∼8% on
average and up to 30% for individual samples.
The Brazilian tests also show significant differences between dyke-perpendicular and dyke-parallel direc-
tions (Table1), with lower tensile strength (unconfined tensile strength, UTS) and fracture toughness (K)
generally observed for failure parallel to the dyke margins (Figure8). Each pair of Brazilian discs were
loaded at 0.02mm/sec in an Instron 5982 100kN testing machine fitted with Brazilian frames such that one
sample failed parallel to the dyke margin (in the L–X plane) and the other perpendicular to it (X–P plane).
Three pairs of samples were discarded as invalid, as one or both samples did not fail along the disc axis but
instead underwent a mixture of shear and tensile fracturing. Peak stress σpeak was recorded and used to esti-
mate the ultimate tensile strength (UTS) in the direction of loading using Equation3 and measurements of
the samples' diameter D (45mm) and thickness t (±25mm).





peak
2
UTS
Dt
(3)
THIELE ET AL.
10.1029/2020JB021477
9 of 17
Figure 7. Plane-polarized (PPL) photomicrographs showing plagioclase alignments (a) and shear bands (b) observed due to the alignment of plagioclase laths
within dykes from various locations on La Palma. The open fractures in each of these (MPJs) are parallel to dyke strike, and are unfilled (except by epoxy during
sample preparation). These appear to have been influenced by the plagioclase alignment, which would give the dykes a preferential parting direction (Delcamp
etal.,2012), and sheet-like domains of high vesicularity (c). High resolution versions of these images can be downloaded from FigShare (Thiele,2020) for closer
inspection.

Journal of Geophysical Research: Solid Earth
Post-failure residual stress was also recorded and, using the methodology and calibration curves presented
by Guo etal.(1993), the magnitude of the stress-drop at failure of the Brazilian disc was used to estimate
K. While this method does not give as accurate a measure of K as more conventional tests (e.g., three-point
bend; Kuruppu etal.,2014), it is sufficient to provide a qualitative comparison of K in dyke-perpendicular
and dyke-parallel directions.
Specimens with MPJs were not tested due to the practical difficulty of preparing samples containing frac-
tures, however other studies have shown that fracturing induces significant elastic anisotropy (e.g., Heap
THIELE ET AL.
10.1029/2020JB021477
10 of 17
Figure 8. Geomechanical tests conducted using ultrasonic pulse velocity measurements (a) and Brazilian disc tests (b).
Young's modulus (a), tensile strength (b), shear modulus (c) and fracture toughness (d) were measured in directions
perpendicular and parallel to dyke orientation and associated flow fabrics. The results show that the dykes have larger
tensile strengths and fracture toughness in perpendicular directions and higher elastic moduli in parallel directions.
These differences make it easier to propagate fractures parallel to the dyke margins and increase the energy required for
cross-cutting fractures. Outliers (gray) were excluded when calculating the trend (dashed blue line). Equal values for
the parallel and perpendicular directions fall on the solid gray line.

Journal of Geophysical Research: Solid Earth
etal.,2009; Heap etal.,2010), and it seems reasonable to assume that a similar or even more pronounced
effect may be expected for fracture toughness and tensile strength.
4. Discussion
Based on our observations at Hoyo Verde, we hypothesize that (1) solidified dykes form important mechani-
cal discontinuities in basaltic volcanoes, and (2) the mechanical contrast between these dykes and the rocks
they intrude can explain the formation of the MPJs and multi-dykes. Mechanical layering and anisotropy
in volcanic rocks has previously been recognised as a significant control on dyke propagation and volcano
deformation (Gudmundsson, 2005b, 2012; Gudmundsson & Brenner, 2001; Kavanagh et al.,2006). Giv-
en this context, we use our observations to evaluate hypotheses (1) and (2) in the following sections, and
present some preliminary modeling results that explore their physical plausibility. Finally, we suggest that
multi-dykes represent reactivated magma pathways, and explore implications of this reactivation for the
organization of volcanic plumbing systems and the distribution of volcanic risks.
4.1. Stress-Concentration, Rotation, and the Formation of MPJs
The dykes at Hoyo Verde are much stiffer than the generally compliant breccia and pyroclastic tuff they
cross-cut. These volcanogenic rocks were not directly tested, but similar tuffs and scoria from Gran Canaria
and Tenerife have Young's moduli of <1–∼20GPa and shear moduli of <1–8GPa (de Vallejo etal.,2008;
Rodríguez-Losada etal.,2009), making them ∼3–15 times more compliant than the dykes (E≈35–70GPa,
G≈15–25GPa; Section3.2). This elastic contrast will cause stress concentration within the dykes during
volcanic inflation-deflation cycles and progressive gravitational loading, and, due to their oblique orienta-
tion, rotation of the principal compressive stress (σ1) toward parallelism with the dyke margins (Figure9).
The tendency for the surrounding, poorly cemented, granular rocks such as tuff and volcanic breccia to un-
dergo inelastic deformation (de Vallejo etal.,2008; Heap etal.,2014; J. S. Lee etal.,2012) would exaggerate
this effect, as would the dykes' elastic anisotropy (Section3.2). The distribution of stress within a stiff, inter-
connected network of solidified dykes embedded in 3-D within more compliant host rock is thus expected
to result in a complex and heterogeneous distribution of stress, akin to engineered composite materials.
As previously suggested by Porreca etal.(2006), we hypothesize that the stress concentration is sufficient
to initiate and drive joint propagation parallel to the dyke margins, facilitated by internal margin-parallel
structures such as plagioclase alignments and vesicle sheets. The formation of these initial fractures would
have enhanced the dyke's elastic anisotropy, further rotating the maximum compressive stress (Faulkner
etal.,2006; Heap etal.,2009,2010) toward parallelism with the dyke margins.
Interestingly, dykes with well-developed cooling joints were rarely observed to have MPJs. This suggests
that the presence of cooling joints inhibits the rotation of the principal compressive stress by reducing the
THIELE ET AL.
10.1029/2020JB021477
11 of 17
Mean difference Mean anisotropy Standard deviation p-value
Young's modulus (E)
X 4.6GPa 9% 4.9GPa 2.3×10−4
L 4.4GPa 8% 3.75GPa 1.4×10−5
Shear modulus (G)
X 1.6GPa 9% 2.2GPa 0.002
L 1.35GPa 7% 1.69GPa 0.003
Tensile strength (UTS) 3.5MPa 26% 2.4MPa 6.6×10−5
Fracture toughness (K) 0.35MPa.m0.5 17% 0.44MPa.m0.5 0.02
Note. Differences are all significant at a 0.05 level based on p-values produced from the paired t-tests.
Table 1
Paired t-Tests Comparing Dyke-Parallel Versus Dyke-Perpendicular Elastic Moduli, Tensile Strength, and Fracture
Toughness

Journal of Geophysical Research: Solid Earth
bulk stiffness of the dyke, while blunting of MPJ fracture tips along margin-perpendicular cooling joints
would further suppress their growth.
4.2. Anisotropy and the Formation of Multi-Dykes
Regardless of their mechanism of formation, the MPJs reduce the dyke-parallel fracture toughness while
encouraging the arrest of dyke-cross-cutting fractures. Similar anisotropy of fracture toughness has been
well documented in shales, and is known to divert fractures along the fabric (Chandler etal.,2016; Forbes
Inskip etal.,2018; H. P. Lee etal.,2015). Microstructural layering has also been shown to increase the frac-
ture toughness and strength of mollusc shells by several orders of magnitude, due to the arrest or deflection
of cross-cutting fractures (Kamat etal.,2000).
This anisotropy, combined with the rotation and increase of σ1 described in Section4.1, can explain the
observed capture and re-orientation of new dykes along older ones. Where dyke margins were weak, in-
creased stress within the older dyke would have encouraged the arrest of the cross-cutting dyke followed
by propagation along the older dyke's margin (e.g., the left-hand dyke in Figure2c). Where dyke margins
were well-bonded, the favored propagation path may instead have followed MPJs within the dyke, to form a
multi-dyke with internal chill-margins (e.g., Figure5), perhaps encouraged by delamination and opening of
the MPJs (Cook & Gordon,1964). The abundance of multi-dykes on La Palma (and elsewhere in the Canary
Islands) suggest that intrusions can remain captured over long distances (102–103m).
4.3. Numerical Modeling
As a preliminary investigation of these hypotheses, we have used Irazu (Lisjak etal.,2018) to construct 2-D
plane-strain finite element (FE) models that evaluate the stresses within and around 2 m thick solidified
dykes within a 100×50m domain. The dykes dip at angles of 30–75° (Figure9) and were given a Young's
modulus of 25GPa, 2.5 times stiffer than the surrounding host rocks (10GPa). A Poisson's ratio of 0.25 was
used for both the dyke and host rock. Boundary conditions correspond to the lithostatic stress at ∼1km
depth, producing lateral confinement and 25MPa overburden pressure). The corresponding stress field
under the given boundary conditions is obtained through the FE models.
Next, the propagation of a new fracture through this initial stress field was simulated using the bounda-
ry-element linear-elastic fracture mechanics code developed by Zhang etal.(2014) and initial stress fields
obtained by FE models. New magma was injected at 0.01m3/sec per unit of dyke thickness along an initially
10m long vertical fracture at the base of the model. The distribution of fluid pressure within the new dyke
and corresponding elastic stress induced in the surrounding host rock was calculated at each timestep, and
THIELE ET AL.
10.1029/2020JB021477
12 of 17
Figure 9. Finite element (Irazu) model of the initial-stress within and around a 2m thick dyke dipping at 30–75° within compliant host rock (a–d) under an
applied overburden stress of 25MPa and lateral confinement, equivalent to ∼1km depth. Each model extends 100m horizontally and 50m vertically, although
(a–d) only show a region of 20×20m to clearly show the rotation (thin black lines show the orientation of σ1) and increase in stress-magnitude caused
by stress-concentration within the solidified dyke. The stress concentration is maximized when the dyke is vertical (parallel to the loading direction), and
minimized when it is horizontal (perpendicular to the loading direction). This initial stress rotation causes the deflection of a subsequent intrusion (thick black
line) modeled using the linear-elastic fracture mechanics code described by Zhang etal.(2014). Note that these models do not include the effects of anisotropic
elastic moduli or fracture toughness, which would increase the amount of deflection (see text for details). Colors show the magnitude of the maximum
principal compressive stress σ1 (compressive stresses are positive).

Journal of Geophysical Research: Solid Earth
fracture growth triggered when the maximum stress intensity factor in one orientation reaches a critical
value equal to the mode I fracture toughness, which we set at 10MPa·m0.5 for both the dyke and the host
rock. Due to numerical limitations, we use relatively low elastic contrasts (2.5 times) and magma viscosity
(1Pa·sec). These models also do not include the previously described anisotropic elastic properties or frac-
ture toughness, so give minimum estimates for the amount of deflection that might be expected. In addi-
tion, actual magma injection conditions are likely to vary in rate and overpressure with time, but a constant
injection rate is used in this study.
Despite these simplifications, the models clearly show that (1) stress is concentrated and rotated in the
solidified dyke, and (2) this causes deflection of the cross-cutting fracture (Figure9). The amount of deflec-
tion varies from ∼1 to >10m depending on the dip of the solidified dyke, with smaller intersection angles
causing longer offsets. Browning and Gudmundsson(2015) developed a similar mechanical model to show
that stiff dykes within compliant fault damage zones can also capture intrusions by this mechanism.
The modeled geometries generally match those observed in the field, although with an order of magnitude
less offset (Figure2). We attribute this difference to the influence of (1) larger elastic contrasts, (2) elastic
anisotropy enhancing rotation of the principal stress, and (3) blunting and deflection of the cross-cutting
dyke tip along MPJs. The 2-D plane-strain geometry used in our model is also a significant simplification,
as most of the dykes we observed showed evidence of lateral propagation (Thiele etal.,2020). Nevertheless,
the results suggest that a mechanical explanation for the multi-dyke formation (Section4.2) is reasonable.
4.4. Implications of Reactivation for Magma Transport and Associated Hazards
By influencing the propagation path of intrusions and other fractures, solidified dykes give volcanic edifices
a “structural memory” of past stress states. Deflection of active dykes along older ones (i.e., reactivation)
will cause them to become misoriented with respect to regional or topographic far-field stress, and hence
lead to unexpected propagation paths. This reactivation will also encourage the self-organization of the
magma plumbing system, promoting re-use of older vents and directing dykes along rift zones or toward
shallow magma reservoirs.
The importance of dykes as structural discontinuities that favor strain localization onto volcanic rift zones
has been proposed by several authors. While discussing volcanic rift zones in the Canary Islands, Carrace-
do(1994) suggested that solidified dykes in the core of rift zones force active ones into parallelism with the
rift. Similarly, Walker(1992) suggested that dyke margins form structural weaknesses that provide prefer-
ential propagation pathways in the densely packed dyke-swarms beneath rift zones in the Hawaiian islands.
While the mechanics of our model for multi-dyke formation are somewhat different, treating dykes as
mechanically competent units rather than structural weaknesses, the overall effect will be the same: dykes
will tend to be re-directed along earlier volcanic rift zone dykes.
Similarly, multi-dykes formed by the mechanisms proposed above could focus dykes into shallow magma
chambers. Due to their greater stiffness, solidified dykes will carry stresses induced by a pressurized magma
chamber to a greater distance than more-compliant host rock, guiding dykes toward the chamber (Karl-
strom etal.,2009). Reactivation of older dykes to form multi-dykes would further enhance these processes,
assuming the older dykes also intersect the chamber. By increasing the probability of magma recharge in
this manner, a volcano might support smaller shallow magma chambers than otherwise expected. Similarly,
the chance of a batch of magma interacting with or assimilating older magma would also increase, allowing
for greater geochemical diversity of erupted products.
We speculate that it is also possible that dykes reactivating older intrusions to form multi-dykes are less
likely to be arrested at bedding interfaces, because the stiff pre-existing intrusion (and local stress field
within it) reduces the stress change experienced by the younger dyke as it passes between stratigraphic units
with different mechanical properties. A partially arrested dyke was observed at one location (Figure 10),
corroborating this hypothesis. Many studies have demonstrated that an anisotropy oriented at a high an-
gle to the dyke propagation direction (e.g., bedding) increases the chance of dyke arrest (e.g., Gudmunds-
son,2005a,2005b). However, to our knowledge, the effect of stiff mechanical discontinuities with similar
orientation to the propagation direction (e.g., older, solidified dykes) has received much less attention. If
the presence of these older intrusions reduces the chance of dyke arrest, then eruption (rather than arrest)
THIELE ET AL.
10.1029/2020JB021477
13 of 17

Journal of Geophysical Research: Solid Earth
becomes more likely in areas that have already been extensively intruded, such as along rift zones, promot-
ing localization of eruptive activity. Conversely, if the accommodation of previous intrusions increases the
stress contrast between stiff and compliant stratigraphic units, then dyke arrest becomes likely and eruption
from previously intruded regions is inhibited due to the formation of a “stress plug” (cf., Thiele etal.,2020).
We suggest that the interactions between concordant (bedding parallel) and discordant (highly oblique)
mechanical discontinuities, and their influence on dyke propagation, is a fertile avenue for future research.
Finally, an intrusion propagating along an older, solidified feeder-dyke might be expected to erupt in close
proximity to old vents. Vents and fissures formed during the Holuhraun eruption (Iceland) sometime be-
tween 1794 and 1864 were re-used by the 2014 Bárðarbunga eruption (Sigmundsson etal.,2014), which
Ruch et al. (2016) argue is evidence that the dyke that fed the eruption intersected and was captured by
a pre-existing fissure. While indistinguishable based on the available evidence, these observations could
equally be explained by the formation of a multi-dyke. Similarly, Carracedo etal.(1996) describe the 1677
Fuencaliente eruption, during which the main Strombolian vent formed on the much older (>3.3 ka) San
Antonio scoria cone. Although plausibly a coincidence, the location of this vent could also be elegantly
explained by the presence of a multi-dyke.
5. Conclusions
We propose that solidified dykes form highly discordant mechanical discontinuities in volcanic edifices.
Where dykes are stiffer than the material they intrude, local stress concentration and re-orientation can
favor the formation of dyke-parallel fractures, including MPJs and multi-dykes. Plagioclase alignments and
other internal structures in the solidified dykes, including MPJs that form during volcanic inflation/defla-
tion cycles, result in significant mechanical anisotropy that further encourages reactivation to form mul-
ti-dykes. Field observations of dykes on La Palma suggest that these processes can result in the capture and
re-orientation of dykes by up to 60°. Multi-dykes formed in this way are thus a type of mechanical memory,
as dykes intruded during previous volcanic activity and potentially different stress fields can influence fu-
ture dyke-propagation paths. Over longer timescales these processes influence the emergent structure of
shallow volcanic plumbing systems by capturing and redirecting dykes along rift zones, toward shallow
magma chambers, and possibly even old vents.
THIELE ET AL.
10.1029/2020JB021477
14 of 17
Figure 10. Photograph (a) and interpretation (b) of a partially arrested multi-dyke. The first dyke was arrested after
propagating through compliant bedded scoria into a stiff lava flow. A younger intrusion that propagated along the older
dyke (to form a multi-dyke) was not arrested, possibly because the older dyke reduced the stiffness contrast as it crossed
the interface. For a 3-D view of this outcrop the reviewer is referred to the digital outcrop models described by Thiele
etal.(2020), and available for download on Figshare (Thiele, Cruden, & Micklethwaite,2019).

Journal of Geophysical Research: Solid Earth
Data Availability Statement
The geomechanical and geochronology datasets presented in this work are available in the Supplemen-
tary Material. UAV surveys and the derived digital outcrop models can be downloaded from https://doi.
org/10.26180/5d688c17f2ed2. Geomechanical and geochronology data can be downloaded from https://doi.
org/10.6084/m9.figshare.13332815.v3.
References
Bemis, S. P., Micklethwaite, S., Turner, D., James, M. R., Akciz, S., Thiele, S. T., & Bangash, H. A. (2014). Ground-based and UAV-Based
photogrammetry: A multi-scale, high-resolution mapping tool for structural geology and paleoseismology. Journal of Structural Geolo-
gy, 69, 163–178. https://doi.org/10.1016/j.jsg.2014.10.007
Bonneville, A., & Capolsini, P. (1999). THERMIC: A 2-D finite-element tool to solve conductive and advective heat transfer problems in
Earth sciences. Computers & Geosciences, 25(10), 1137–1148. https://doi.org/10.1016/S0098-3004(99)00075-8
Browning, J., & Gudmundsson, A. (2015). Caldera faults capture and deflect inclined sheets: An alternative mechanism of ring dike for-
mation. Bulletin of Volcanology, 77(1), 4. https://doi.org/10.1007/s00445-014-0889-4
Bubeck, A., Walker, R. J., Healy, D., Dobbs, M., & Holwell, D. A. (2017). Pore geometry as a control on rock strength. Earth and Planetary
Science Letters, 457, 38–48. https://doi.org/10.1016/j.epsl.2016.09.050
Budkewitsch, P., & Robin, P.-Y. (1994). Modeling the evolution of columnar joints. Journal of Volcanology and Geothermal Research, 59(3),
219–239. https://doi.org/10.1016/0377-0273(94)90092-2
Carracedo, J. C. (1994). The Canary Islands: An example of structural control on the growth of large oceanic-island volcanoes. Journal of
Volcanology and Geothermal Research, 60(3–4), 225–241. https://doi.org/10.1016/0377-0273(94)90053-1
Carracedo, J. C., Day, S., Guillou, H., & Badiola, E. R. (1996). The 1677 eruption of La Palma, Canary Islands. Estudios Geológicos, 52(3–4),
103–114. https://doi.org/10.3989/egeol.96523-4258
Carracedo, J. C., Day, S. J., Guillou, H., & Gravestock, P. (1999). Later stages of volcanic evolution of La Palma, Canary Islands: Rift evolu-
tion, giant landslides, and the genesis of the Caldera de Taburiente. Geological Society of America Bulletin, 111(5), 755–768. https://doi.
org/10.1130/0016-7606(1999)111<0755:lsoveo>2.3.co;2
Chandler, M. R., Meredith, P. G., Brantut, N., & Crawford, B. R. (2016). Fracture toughness anisotropy in shale. Journal of Geophysical
Research: Solid Earth, 121(3), 1706–1729. https://doi.org/10.1002/2015JB012756
Cook, J., & Gordon, J. E. (1964). A mechanism for the control of crack propagation in all-brittle systems. Proceedings of the Royal Society of
London Series A: Mathematical and Physical Sciences, 282(1391), 508–520.
de Vallejo, L. I. G. L., Hijazo, T., & Ferrer, M. (2008). Engineering geological properties of the volcanic rocks and soils of the Canary Islands.
Soils and Rocks, 31, 3–13.
Delcamp, A., Troll, V. R., van Wyk de Vries, B., Carracedo, J. C., Petronis, M. S., Pérez-Torrado, F. J., & Deegan, F. M. (2012). Dykes and
structures of the NE rift of Tenerife, Canary Islands: A record of stabilization and destabilization of ocean island rift zones. Bulletin of
Volcanology, 74(5), 963–980. https://doi.org/10.1007/s00445-012-0577-1
Dering, G. M., Micklethwaite, S., Thiele, S. T., Vollgger, S. A., & Cruden, A. R. (2019). Review of drones, photogrammetry and emerging
sensor technology for the study of dykes: Best practises and future potential. Journal of Volcanology and Geothermal Research, 373,
148–166. https://doi.org/10.1016/j.jvolgeores.2019.01.018
Ehlers, C., & Ehlers, M. (1977). Shearing and multiple intrusion in the diabases of Åland archipelago. Geologinen tutkimuslaitos.
Esteve, S., Enrique, P., & Galán, G. (2014). The camptonites in the multiple intrusion of Platja Fonda (Girona, NE Spain): Mechanisms of
intrusion and geochemistry. Journal of Geosciences, 59(1), 23–40. https://doi.org/10.3190/jgeosci.161
Faulkner, D. R., Mitchell, T. M., Healy, D., & Heap, M. J. (2006). Slip on 'weak' faults by the rotation of regional stress in the fracture damage
zone. Nature, 444, 922–925. https://doi.org/10.1038/nature05353
Forbes Inskip, N. D., Meredith, P. G., Chandler, M. R., & Gudmundsson, A. (2018). Fracture properties of nash point shale as a function of
orientation to bedding. Journal of Geophysical Research: Solid Earth, 123(10), 8428–8444. https://doi.org/10.1029/2018JB015943
Gaffney, E. S., Damjanac, B., & Valentine, G. A. (2007). Localization of volcanic activity: 2. Effects of pre-existing structure. Earth and
Planetary Science Letters, 263(3), 323–338. https://doi.org/10.1016/j.epsl.2007.09.002
Gass, I. G. (1968). Is the Troodos Massif of Cyprus a fragment of mesozoic ocean floor? Nature, 220, 39–42. https://doi.org/10.1038/220039a0
Gibson, I. L., & Walker, G. P. L. (1963). Some composite rhyolite/basalt lavas and related composite dykes in eastern Iceland. Proceedings
of the Geologists' Association, 74(3), 301–IN3. https://doi.org/10.1016/s0016-7878(63)80025-5
Gudmundsson, A. (1984). Formation of dykes, feeder-dykes, and the intrusion of dykes from magma Chambers. Bulletin of Volcanology,
47(3), 537–550. https://doi.org/10.1007/BF01961225
Gudmundsson, A. (2005a). The effects of layering and local stresses in composite volcanoes on dyke emplacement and volcanic hazards.
Comptes Rendus Geoscience, 337(13), 1216–1222. https://doi.org/10.1016/j.crte.2005.07.001
Gudmundsson, A. (2005b). Effects of mechanical layering on the development of normal faults and dykes in Iceland. Geodinamica Acta,
18(1), 11–30. https://doi.org/10.3166/ga.18.11-30
Gudmundsson, A. (2011). Rock fractures in geological processes. Cambridge University Press. https://doi.org/10.1017/cbo9780511975684
Gudmundsson, A. (2012). Strengths and strain energies of volcanic edifices: Implications for eruptions, collapse calderas, and landslides.
Natural Hazards and Earth System Sciences, 12(7), 2241–2258. https://doi.org/10.5194/nhess-12-2241-2012
Gudmundsson, A., & Brenner, S. L. (2001). How hydrofractures become arrested. Terra Nova, 13(6), 456–462. https://doi.
org/10.1046/j.1365-3121.2001.00380.x
Guo, H., Aziz, N. I., & Schmidt, L. C. (1993). Rock fracture-toughness determination by the Brazilian test. Engineering Geology, 33(3),
177–188. https://doi.org/10.1016/0013-7952(93)90056-i
Guppy, E. M., & Hawkes, L. (1925). A composite Dyke from Eastern Iceland. Quarterly Journal of the Geological Society, 81(1–4), 325–340.
https://doi.org/10.1144/GSL.JGS.1925.081.01-04.13
Healy, D. (2008). Damage patterns, stress rotations and pore fluid pressures in strike-slip fault zones. Journal of Geophysical Research,
113(B12). https://doi.org/10.1029/2008jb005655
Heap, M. J., Baud, P., Meredith, P. G., Vinciguerra, S., & Reuschlé, T. (2014). The permeability and elastic moduli of tuff from Campi Fle-
grei, Italy: Implications for ground deformation modeling. Solid Earth, 5(1), 25–44. https://doi.org/10.5194/se-5-25-2014
THIELE ET AL.
10.1029/2020JB021477
15 of 17
Acknowledgments
The authors gratefully acknowledge
the staff at Parque Nacional Caldera de
Taburiente for their generous support
and hospitality during collection of
the field data presented in this study.
ST was supported by a Westpac Future
Leaders Scholarship and Australian
Postgraduate Award. ARC's research on
magma plumbing systems is supported
by Australian Research Council Discov-
ery Grant DP190102422. The University
of Melbourne Ar-Ar Laboratory receives
support under the AuScope program of
the National Collaborative Research In-
frastructure Strategy (NCRIS). Finally,
we would like to thank John Browning,
Dave Healy, Hans Jørgen Kjøll and
an anonymous reviewer for their con-
structive and helpful comments. Open
Access funding enabled and organized
by Projekt DEAL.

Journal of Geophysical Research: Solid Earth
Heap, M. J., Faulkner, D. R., Meredith, P. G., & Vinciguerra, S. (2010). Elastic moduli evolution and accompanying stress changes with
increasing crack damage: Implications for stress changes around fault zones and volcanoes during deformation. Geophysical Journal
International, 183(1), 225–236. https://doi.org/10.1111/j.1365-246X.2010.04726.x
Heap, M. J., Vinciguerra, S., & Meredith, P. G. (2009). The evolution of elastic moduli with increasing crack damage during cyclic stressing
of a basalt from Mt. Etna volcano. Tectonophysics, 471(1), 153–160. https://doi.org/10.1016/j.tecto.2008.10.004
Holness, M. B., & Humphreys, M. C. S. (2003). The Traigh Bhan na Sgurra Sill, Isle of Mull: Flow localization in a major magma conduit.
Journal of Petrology, 44(11), 1961–1976. https://doi.org/10.1093/petrology/egg066
Houghton, B. F., & Wilson, C. J. N. (1989). A vesicularity index for pyroclastic deposits. Bulletin of Volcanology, 51(6), 451–462. https://doi.
org/10.1007/BF01078811
Kamat, S., Su, X., Ballarini, R., & Heuer, A. H. (2000). Structural basis for the fracture toughness of the shell of the conch Strombus gigas.
Nature, 405(6790), 1036–1040. https://doi.org/10.1038/35016535
Kanaris-Sotiriou, R., & Gibb, F. G. F. (1985). Hybridization and the petrogenesis of composite intrusions: The dyke at An Cumhann, Isle
of Arran, Scotland. Geological Magazine, 122(4), 361–372. https://doi.org/10.1017/S0016756800031800
Karlstrom, L., Dufek, J., & Manga, M. (2009). Organization of volcanic plumbing through magmatic lensing by magma chambers and
volcanic loads. Journal of Geophysical Research, 114(B10). https://doi.org/10.1029/2009JB006339
Karlstrom, L., Dufek, J., & Manga, M. (2010). Magma chamber stability in arc and continental crust. Journal of Volcanology and Geother-
mal Research, 190(3), 249–270. https://doi.org/10.1016/j.jvolgeores.2009.10.003
Kavanagh, J. L., Menand, T., & Sparks, R. S. J. (2006). An experimental investigation of sill formation and propagation in layered elastic
media. Earth and Planetary Science Letters, 245(3), 799–813. https://doi.org/10.1016/j.epsl.2006.03.025
Kidd, R. G. W., & Cann, J. R. (1974). Chilling statistics indicate an ocean-floor spreading origin for the Troodos complex, Cyprus. Earth and
Planetary Science Letters, 24(1), 151–155. https://doi.org/10.1016/0012-821x(74)90020-x
Kuruppu, M. D., Obara, Y., Ayatollahi, M. R., Chong, K. P., & Funatsu, T. (2014). ISRM-suggested method for determining the mode I static
fracture toughness using semi-circular bend specimen. Rock Mechanics and Rock Engineering, 47(1), 267–274. https://doi.org/10.1007/
s00603-013-0422-7
Lee, H. P., Olson, J. E., Holder, J., Gale, J. F. W., & Myers, R. D. (2015). The interaction of propagating opening mode fractures with pre-
existing discontinuities in shale. Journal of Geophysical Research: Solid Earth, 120(1), 169–181. https://doi.org/10.1002/2014JB011358
Lee, J. S., Maharidge, R. L., Franquet, J. A., & Dick, A. D. (2012). Stress-dependent brittle-to-ductile deformation of tuff in geothermal
wells. In Paper presented at the 46th U.S. Rock Mechanics/Geomechanics Symposium.
Lisjak, A., Mahabadi, O. K., He, L., Tatone, B. S. A., Kaifosh, P., Haque, S. A., & Grasselli, G. (2018). Acceleration of a 2D/3D finite-discrete
element code for geomechanical simulations using General Purpose GPU computing. Computers and Geotechnics, 100, 84–96. https://
doi.org/10.1016/j.compgeo.2018.04.011
Marinoni, L. B., & Gudmundsson, A. (2000). Dykes, faults and palaeostresses in the Teno and Anaga massifs of Tenerife (Canary Islands).
Journal of Volcanology and Geothermal Research, 103(1), 83–103. https://doi.org/10.1016/s0377-0273(00)00217-1
Matchan, E. L., & Phillips, D. (2014). High precision multi-collector 40Ar/39Ar dating of young basalts: Mount Rouse volcano (SE Austral-
ia) revisited. Quaternary Geochronology, 22, 57–64. https://doi.org/10.1016/j.quageo.2014.02.005
Moreira, M. A., Geoffroy, L., & Pozzi, J. P. (2015). Magma flow pattern in dykes of the Azores revealed by anisotropy of magnetic suscepti-
bility. Journal of Geophysical Research: Solid Earth, 120(2), 662–690. https://doi.org/10.1002/2014JB010982
Neal, C. A., Brantley, S. R., Antolik, L., Babb, J. L., Burgess, M., Calles, K., etal. (2018). The 2018 rift eruption and summit collapse of
Kīlauea Volcano. Science, 363, 367–374. https://doi.org/10.1126/science.aav7046
Platten, I. M. (2000). Incremental dilation of magma filled fractures: Evidence from dykes on the Isle of Skye, Scotland. Journal of Struc-
tural Geology, 22(8), 1153–1164. https://doi.org/10.1016/s0191-8141(00)00024-9
Porreca, M., Acocella, V., Massimi, E., Mattei, M., Funiciello, R., & Debenedetti, A. (2006). Geometric and kinematic features of the dike
complex at Mt. Somma, Vesuvio (Italy). Earth and Planetary Science Letters, 245(1), 389–407. https://doi.org/10.1016/j.epsl.2006.02.027
Rivalta, E., Taisne, B., Bunger, A. P., & Katz, R. F. (2015). A review of mechanical models of dike propagation: Schools of thought, results
and future directions. Tectonophysics, 638, 1–42. https://doi.org/10.1016/j.tecto.2014.10.003
Rodríguez-Losada, J. A., Hernández-Gutiérrez, L. E., Olalla, C., Perucho, A., Serrano, A., & Eff-Darwich, A. (2009). Geomechanical pa-
rameters of intact rocks and rock masses from the Canary Islands: Implications on their flank stability. Journal of Volcanology and
Geothermal Research, 182(1), 67–75. https://doi.org/10.1016/j.jvolgeores.2009.01.032
Ruch, J., Wang, T., Xu, W., Hensch, M., & Jónsson, S. (2016). Oblique rift opening revealed by reoccurring magma injection in central
Iceland. Nature Communications, 7, 12352. https://doi.org/10.1038/ncomms12352
Sigmundsson, F., Hooper, A., Hreinsdóttir, S., Vogfjörd, K. S., Ófeigsson, B. G., Heimisson, E. R., etal. (2014). Segmented lateral dyke
growth in a rifting event at Bárðarbunga volcanic system, Iceland. Nature, 517, 191–195. https://doi.org/10.1038/nature14111
Stewart, M. A., Klein, E. M., & Karson, J. A. (2002). Geochemistry of dikes and lavas from the north wall of the Hess Deep Rift: Insights into
the four-dimensional character of crustal construction at fast spreading mid-ocean ridges. Journal of Geophysical Research, 107(B10),
EPM41–EPM424. https://doi.org/10.1029/2001JB000545
Sun, M.-D., Chen, H.-L., Zhang, F.-Q., Wilde, S. A., Dong, C.-W., & Yang, S.-F. (2013). A 100Ma bimodal composite dyke complex in the
Jiamusi Block, NE China: An indication for lithospheric extension driven by Paleo-Pacific roll-back. Lithos, 162–163, 317–330. https://
doi.org/10.1016/j.lithos.2012.11.021
Thiele, S. T. (2020). La Palma multi-dyke data, Figshare. https://doi.org/10.6084/m9.figshare.13332815.v3
Thiele, S. T., Cruden, A. R. & Micklethwaite, S. (2019). Volcán taburiente digital outcrop models, Figshare. https://doi.
org/10.26180/5d688c17f2ed2
Thiele, S. T., Cruden, A. R., Micklethwaite, S., Bunger, A. P., & Köpping, J. (2020). Dyke apertures record stress accumulation during sus-
tained volcanism. Scientific Reports, 10(1), 1–9. https://doi.org/10.1038/s41598-020-74361-w
Thiele, S. T., Grose, L., Cui, T., Cruden, A. R., & Micklethwaite, S. (2019). Extraction of high-resolution structural orientations from digital
data: A Bayesian approach. Journal of Structural Geology, 122, 106–115. https://doi.org/10.1016/j.jsg.2019.03.001
Valentine, G., & Krogh, K. (2006). Emplacement of shallow dikes and sills beneath a small basaltic volcanic center - The role of pre-ex-
isting structure (Paiute Ridge, southern Nevada, USA). Earth and Planetary Science Letters, 246(3), 217–230. https://doi.org/10.1016/j.
epsl.2006.04.031
van den Hove, J., Grose, L., Betts, P. G., Ailleres, L., Van Otterloo, J., & Cas, R. A. F. (2017). Spatial analysis of an intra-plate basaltic vol-
canic field in a compressional tectonic setting: South-eastern Australia. Journal of Volcanology and Geothermal Research, 335, 35–53.
https://doi.org/10.1016/j.jvolgeores.2017.02.001
THIELE ET AL.
10.1029/2020JB021477
16 of 17

Journal of Geophysical Research: Solid Earth
Walker, G. P. L. (1992). "Coherent intrusion complexes" in large basaltic volcanoes - a new structural model. Journal of Volcanology and
Geothermal Research, 50(1), 41–54. https://doi.org/10.1016/0377-0273(92)90036-d
Zhang, X., Bunger, A. P., & Jeffrey, R. G. (2014). Mechanics of two interacting magma-driven fractures: A numerical study. Journal of
Geophysical Research: Solid Earth, 119(11), 8047–8063. https://doi.org/10.1002/2014JB011273
References From the Supporting Information
Lee, J.-Y., Marti, K., Severinghaus, J. P., Kawamura, K., Yoo, H.-S., Lee, J. B., & Kim, J. S. (2006). A redetermination of the isotopic abun-
dances of atmospheric Ar. Geochimica et Cosmochimica Acta, 70(17), 4507–4512. https://doi.org/10.1016/j.gca.2006.06.1563
Ludwig, K. R. (2012). User's manual for IsoPlot 3.75. A geological toolkit for Microsoft Excel. In Berkeley geochronology center special pub-
lication No. 5. https://doi.org/10.1145/2166966.2167008
Phillips, D., Matchan, E. L., Honda, M., & Kuiper, K. F. (2017). Astronomical calibration of 40Ar/39Ar reference minerals using high-pre-
cision, multi-collector (ARGUSVI) mass spectrometry. Geochimica et Cosmochimica Acta, 196, 351–369. https://doi.org/10.1016/j.
gca.2016.09.027
Steiger, R. H., & Jäger, E. (1977). Subcommission on geochronology: Convention on the use of decay constants in geo- and cosmochronol-
ogy. Earth and Planetary Science Letters, 36(3), 359–362. https://doi.org/10.1016/0012-821x(77)90060-7
THIELE ET AL.
10.1029/2020JB021477
17 of 17