Relationships between volcano gravitational spreading and magma intrusion

Abstract

Volcano spreading, with its characteristic sector grabens, is caused by outward flow of weak substrata due to gravitational loading. This process is now known to affect many present-day edifices. A volcano intrusive complex can form an important component of an edifice and may induce deformation while it develops. Such intrusions are clearly observed in ancient eroded volcanoes, like the Scottish Palaeocene centres, or in geophysical studies such as in La Réunion, or inferred from large calderas, such as in Hawaii, the Canaries or Galapagos volcanoes. Volcano gravitational spreading and intrusive complex emplacement may act simultaneously within an edifice. We explore the coupling and interactions between these two processes. We use scaled analogue models, where an intrusive complex made of Golden syrup is emplaced within a granular model volcano based on a substratum of a ductile silicone layer overlain by a brittle granular layer. We model specifically the large intrusive complex growth and do not model small-scale and shortlived events, such as dyke intrusion, that develop above the intrusive complex. The models show that the intrusive complex develops in continual competition between upward bulging and lateral gravity spreading. The brittle substratum strongly controls the deformation style, the intrusion shape and also controls the balance between intrusive complex spreading and ductile layer-related gravitational spreading. In the models, intrusive complex emplacement and spreading produce similar structures to those formed during volcano gravitational spreading alone (i.e. grabens, folds, en échelon fractures). Therefore, simple analysis of fault geometry and fault kinetic indicators is not sufficient to distinguish gravitational from intrusive complex spreading, except when the intrusive complex is eccentric from the volcano centre. However, the displacement fields obtained for (1) a solely gravitational spreading volcano and for (2) a gravitational spreading volcano with a growing and spreading intrusive complex are very different. Consequently, deformation fields (like those obtained from geodetic monitoring) can give a strong indication of the presence of a spreading intrusive complex. We compare the models with field observations and geophysical evidence on active volcanoes such as La Réunion Island (Indian Ocean), Ometepe Island (Nicaragua) and eroded volcanic remnants such as Ardnamurchan (Scotland) and suggest that a combination between gravitational and intrusive complex spreading has been active.

Full text

RESEARCH ARTICLE
Relationships between volcano gravitational spreading
and magma intrusion
Audray Delcamp &Benjamin van Wyk de Vries &
Mike R. James &L. S. Gailler &E. Lebas
Received: 15 July 2008 / Accepted: 13 October 2011 / Published online: 25 November 2011
#Springer-Verlag 2011
Abstract Volcano spreading, with its characteristic sector
grabens, is caused by outward flow of weak substrata due to
gravitational loading. This process is now known to affect
many present-day edifices. A volcano intrusive complex can
form an important component of an edifice and may induce
deformation while it develops. Such intrusions are clearly
observed in ancient eroded volcanoes, like the Scottish
Palaeocene centres, or in geophysical studies such as in La
Réunion, or inferred from large calderas, such as in Hawaii,
the Canaries or Galapagos volcanoes. Volcano gravitational
spreading and intrusive complex emplacement may act
simultaneously within an edifice. We explore the coupling
and interactions between these two processes. We use scaled
analogue models, where an intrusive complex made of Golden
syrup is emplaced within a granular model volcano based on a
substratum of a ductile silicone layer overlain by a brittle
granular layer. We model specifically the large intrusive
complex growth and do not model small-scale and short-
lived events, such as dyke intrusion, that develop above the
intrusive complex. The models show that the intrusive
complex develops in continual competition between upward
bulging and lateral gravity spreading. The brittle substratum
strongly controls the deformation style, the intrusion shape
and also controls the balance between intrusive complex
spreading and ductile layer-related gravitational spreading. In
the models, intrusive complex emplacement and spreading
produce similar structures to those formed during volcano
gravitational spreading alone (i.e. grabens, folds, en échelon
fractures). Therefore, simple analysis of fault geometry and
fault kinetic indicators is not sufficient to distinguish
gravitational from intrusive complex spreading, except when
the intrusive complex is eccentric from the volcano centre.
However, the displacement fields obtained for (1) a solely
gravitational spreading volcano and for (2) a gravitational
spreading volcano with a growing and spreading intrusive
complex are very different. Consequently, deformation fields
(like those obtained from geodetic monitoring) can give a
strong indication of the presence of a spreading intrusive
complex. We compare the models with field observations
and geophysical evidence on active volcanoes such as La
Réunion Island (Indian Ocean), Ometepe Island (Nicaragua)
and eroded volcanic remnants such as Ardnamurchan
(Scotland) and suggest that a combination between gravita-
tional and intrusive complex spreading has been active.
Keywords Volcano .Intrusive complex .Spreading .
Gravitational spreading .Rift zone .Analogue models
Introduction
Intrusions such as dykes, sills, cone sheets and intrusive
complexes can constitute an important part of volcano
Editorial responsibility: J.C. Phillips
A. Delcamp :B. van Wyk de Vries :L. S. Gailler :E. Lebas
Laboratoire Magmas et Volcans CNRS-UMR 6524,
Observatoire du Physique du Globe de Clermont,
Université Blaise Pascal,
Clermont-Ferrand, France
B. van Wyk de Vries
e-mail: b.vanwyk@opgc.univ-bpclermont.fr
A. Delcamp (*)
Department of Geography, Earth System Sciences,
Vrije Universiteit Brussel,
Brussel, Belgium
e-mail: delcampa@tcd.ie
M. R. James
Lancaster Environment Centre, Lancaster University,
Lancaster, UK
e-mail: m.james@lancs.ac.uk
Bull Volcanol (2012) 74:743–765
DOI 10.1007/s00445-011-0558-9

architecture and growth (e.g. Francis et al. 1993; Annen et
al. 2001; Mathieu and van Wyk de Vries 2009). While
dykes and sills are important near-surface intrusions and are
regularly detected by geodetic monitoring, large-scale
intrusive complex growth provides long-term, large-scale
deformation that can be seen on large volcanoes, such as
Hawaii and Etna (e.g. Delaney et al. 1990; Clague and
Denlinger 1994; Chiocci et al. 2011). Large intrusive
complexes clearly develop inside volcanoes, as seen in
eroded volcanoes (such as the Scottish Palaeocene centres
of Arndamurchan, Sky, Mull, Arran and Rum). They are
also detected by geophysical evidence, such as at La
Réunion Island (Gailler et al. 2009; Gailler and Lénat
2010), and from the presence of large calderas, as observed
on most large volcanoes (Roche et al. 2001).
Intrusion emplacement can induce surface deformation
on a volcano, in the extreme like the bulge observed at
Mount St. Helens prior to the May 1980 eruption, or during
inflation and dyke intrusion episodes at volcanoes such as
Hawaii or Piton de la Fournaise (Swanson et al. 1976;
Dieterich 1988; Cayol et al. 2000; Brooks et al. 2008).
Intrusive complex emplacement induces also deformation
in the host rock. For example, in the Ardnamurchan
Peninsula, Scotland, the surrounding sediments, schists
and volcanic rocks are folded around the intrusive complex
(Bailey et al. 1924; Tyrell 1928).
Volcano spreading is the lateral deformation of the
volcano edifice along a basal ductile layer that deforms
under the volcano weight. This mechanism has been widely
studied on the field (van Bemmelen 1949; Borgia et al.
1992,2000; Borgia and van Wyk de Vries 2003) and using
analogue and numerical modelling (Merle and Borgia 1996;
van Wyk and Matela 1998; Walter et al. 2006; Münn et al.
2006; Wooller et al. 2004; Morgan and McGovern 2005;
Delcamp et al. 2008). Grabens, en échelon faults, folds and
thrusts are the characteristic spreading-related structures.
Intrusive complex emplacement phenomena have been
studied separately from volcano gravitational spreading
(e.g. Merle and Venderville 1992). A link between
gravitational spreading and intrusion emplacement should
exist, as suggested by van Bemmelen (1949), Borgia (1994)
or Borgia and van Wyk de Vries (2003). According to
Borgia (1994), the evolution of a volcanic edifice is divided
into five stages that can be repeated, combined or absent:
(1) a building phase, (2) is followed by a compressive
phase due to an increase of volcanic load, (3) then a basal
thrust occurred along a decollement, (4) followed by
emplacement of an intrusive system and (5) finally, the
intrusive complex spreads. To Borgia et al. (2000), dyke
intrusion might contribute to the gravity spreading of the
edifice by creating an additional forceful outward push.
According to Merle and Vendeville (1992), vertical intru-
sion of a magma body generates gravity sliding along the
flank of the dome created by intrusion emplacement. In
spreading, the intrusive complex induces extension within
overlying brittle layer. Observed structures, such as those at
Concepción, Nicaragua (Borgia and van Wyk de Vries
2003), Poas, Costa Rica (Borgia et al. 1990) or Etna
(Borgia et al. 1992) should be a combination of both (1)
volcano gravitational spreading and (2) intrusive complex
emplacement and spreading.
The deformation related to small-scale intrusions has
been also studied using Mogi, Okada and similar elastic
models (e.g. Mogi 1958; Okubo and Watanabe 1989;
Okada 1992; Fukushima et al. 2005). Such models are
applicable for very small-scale elastic deformation relating
to, for example, dyke emplacement, but are not applicable
for large-scale nonelastic deformation, as considered here,
where faulting and viscous flow occur. At the scale of our
models, elastic deformation would only be in the order of
less than a fraction of a millimetre. Because of this
difference, we do not attempt any comparison with such
small-magnitude elastic models, and any comparison would
be meaningless.
Using scaled analogue models, we study the coupling of
the two processes to assess the influence of intrusive complex
emplacement in a spreading volcano. To emphasise the
influence, we compared spreading sandbox models with and
without intrusive complex emplacement. No such combined
models have been done previously, and previous studies have
either concentrated only on one process or the other. Finally,
we compare the model results with various eroded and recent
volcanic centres in both oceanic and continental contexts, and
we especially focused in on La Réunion Island where the
effect of intrusive complex emplacement and spreading
interacts with gravitational spreading of the volcano.
Modelling
Experimental setup
We study two types of models, one with a basal ductile
substratum diameter smaller than the volcanic edifice
(type I) and the other with a diameter of ductile basal
layer greater than the volcano (type II; Fig. 1). The first
situation simulates a volcano with a basal ductile layer
included in the edifice. Such a situation can be analogous
to an oceanic volcano like La Réunion, Guadeloupe or
Hawaii with within-edifice pelagic and detrital low
strength layers (Oehler et al. 2005), or any volcano that
has grown with weak inner layers (e.g. heavily weathered
rocks). The second type of model represents a volcano
built directly onto a ductile substratum layer, as it is
generally the case for subaerial volcanoes or oceanic
volcanoes on thick pelagic sediments.
744 Bull Volcanol (2012) 74:743–765

The model is constructed on a flat rigid base in
which a hole has been drilled. This aperture is
connected by a pipe to a reservoir containing Golden
syrup as a magma analogue (Mathieu et al. 2008).
Emplacement of the analogue intrusive complex is
generated by overpressure, as the reservoir can be raised
above the experimental table. After the experiments,
models are frozen at −18°C to preserve the form of the
intrusive complex shape. Photographs of the models were
taken at even time steps, allowing 2D horizontal defor-
mation to be tracked and structure development to be
followed. On some models, three cameras were used for
full 3D photogrammetry.
Scaling and dimensionless analysis
Scaling assures the necessary conditions for maintaining the
same geometric, time and force ratios between models and
natural cases (Hubbert 1937; Ramberg 1981; Middleton
and Wilcock 1994). Viscosity, density, cohesion and some
other parameters are well scaled from previous similar
experiments without intrusion (Merle and Borgia 1996;
Type II ‘continental’ volcano (rs>rc)
Type I ‘oceanic’ volcano (rs < rc)
Intrusive system (type I & II volcanoes)
Model
rs
rc
h
d
µ
τ0
a
volcano
edifice
ductile layer
Model
rs
rc
h
d
µ
τ0
a
volcano
edifice
ductile layer
rs
rvh
d µ
a
s
ddf
Golden syrup
tank
stable table
volcanic edifice brittle layer
(present or not)
ductile layer
pipe
f
sand edge
(absent or not)
sea level
shield volcano
volcanic edifice
(brittle behaviour)
low strength sediments,
unstable eruptive products,
weathered layers
(ductile behaviour)
resistant
substratum
Natural case
Natural case
pyroclastic products,
intrusive complex,
altered layers
(ductile behaviour)
a
b
c
Fig. 1 Experimental setup. Two types of models have been tested: a
type I ‘oceanic’model, where the radius of the edifice is greater than
the radius of ductile layer and btype II ‘continental’model, where the
radius of the volcano is smaller than the substratum; cSketch of the
experimental setup for intrusive models
Bull Volcanol (2012) 74:743–765 745

Oehler et al. 2005; Walter et al. 2006). The scaling here
follows closely to such previous studies (Table 1).
Length scale (l*) is 5.8×10
−5
, density scale ρ*) is 0.57
and gravity scale (g*) is 1. The result of the following
equation provides the stress and cohesion scale (σ*) that is
thus about 10
−5
:
l»r»g»¼s»ð1Þ
Using this, our model sand–plaster mix, which has a
cohesion of about 100 Pa (it varies between 66 and
336 Pa), is the equivalent to a rock with a cohesion of
about 10
7
Pa. Time scale (t*) is obtained from velocity
information. We prefer to use velocity (u) rather than the
previously used ‘time of deformation’(Merle and Borgia
1996) as this latter parameter is hard to determine in
nature and in the experiments. Characteristic flank
displacement rates in the models are about 10
−6
ms
−1
,
and we equate these to natural rates of about 10
−10
ms
−1
(equivalent to 1 cm per year). Thus, the velocity scale is
10
4
. Using the length scale, the time scale can be extracted
from the velocity scale:
l»=u»¼t»ð2Þ
Thus, the time scale (t*) is about 2× 10
−9
. The viscosity
scale is obtained from the stress and time scales
m»¼s»t»ð3Þ
The viscosity scale is thus 7×10
−14
, which is equivalent
to a substratum of 10
17–18
Pa s and an intrusive complex
viscosity of 10
13–14
Pa s The substrata values are coherent
with those used by other studies (Merle and Borgia 1996;
Cecchi 2003), and the intrusive values are coherent with a
viscosity expected for a semi-solidified magma body (e.g.
Holohan et al. 2008). Consequently, the intrusive complex,
composed of crystal-poor magma, crustal mush, cumulates
and partially solidified magma, is considered as one
mechanically coherent unit (Hill and Zucca 1987; Okubo
et al. 1997; Kauahikaua et al. 2000). Note that additional
low-viscosity intrusions that may rise from the intrusive
complex (i.e. dykes) are not and cannot be modelled here.
Viscosity of strata and magmatic bodies estimates may
vary by several orders of magnitude (e.g. Murase and
McBirney 1973;Marsh1981; Rosenberg 2001). Both
viscous intrusion and substrata behave as ductile material
and are thus likely to deform in the same manner, but with
different rates, over a large range of viscosity values.
Consequently, the unknown natural values will unlikely
affect the structural geometry observed, but will affect the
rates of deformation.
Dimensionless numbers
Once the basic scaling is completed, the parameters can be
cast into dimensionless numbers to study quantifiable
elements of the model such as geometric parameters or
fault density (Hubbert 1937; Ramberg 1981; Middleton and
Wilcock 1994).
There are a large number of parameters, but as many of
these are held constant, or do not change, they can be
removed from the dimensionless analysis. For example,
intrusion viscosity and diameter of feeding pipe stay
constant. Previous experiments (Cecchi 2003; Delcamp et
al. 2008) showed that substrata viscosity variation influen-
ces velocity, but not the type or geometry of the observed
deformation. This parameter is therefore not included in the
dimensional analysis. However, a scaled intrusion velocity
can still be extracted and can be used for comparisons. The
number of constant parameters allows a reduction of the
number of variables to six, expressed with the dimensions
of length and time (Table 2).
Table 1 Basic representative scaling parameters used, values for models and nature with their ratios
Parameters Units Model Nature Ratio (*)
Length (l) m 0.07 1,200 5.8× 10
−5
Density (ρ)kgm
−3
1,500 2,600 0.57
Gravity (g)ms
−2
9.81 9.81 1
Stress (σ)=l×ρ×gkg m
−1
s
−2
100 3× 10
7
3.4× 10
−6
Velocity (u)ms
−1
8×10
−6
3×10
−10
2.7× 10
4
Time (t)=l/us 1 min 900 years 2× 10
−9
Viscosity substrata (μ)kgm
−1
s
−1
10,000 1.4× 10
17
7×10
−14
Viscosity intrusion (μ)kgm
−1
s
−1
34×10
14
7×10
−14
Intrusion flux (f)m
3
s
−1
2×10
−8
2.2× 10
−4
9×10
−5
Dim. equivalence μ/σt=1 l/ut=1
Note that the basic three parameters, length, density and gravity, provide the values for scaling for the three dimensions length, mass and time.
Stress (and cohesion) are defined by the product, l×ρ×g. Velocity is estimated from models and nature, and time scaling is obtained from velocity
and length. The scaling is verified by the dimensionless equivalence, μ/σt=1/ut
746 Bull Volcanol (2012) 74:743–765

Using the Π-Buckingham theorem (e.g. Middleton and
Wilcock 1994), four dimensionless numbers can be
established. Π1 is the ratio of volcano height to its radius.
The volcano height for type I models includes the thickness
of ductile substratum that constitutes a part of the
edifice (i.e. ductile layer is included within the volcano
edifice). The tangent of this number corresponds to the
volcano slope.
Π1¼volcano height
volcano radius ¼h
rv
ð6Þ
Π2 is the ratio of the volcano height to the substrata
thickness,
Π2¼volcano height
thickness of substrata ðbrittle þductileÞ¼h
dfþdd
ð7Þ
Π3 is the ratio between the brittle layer and the ductile
layer thicknesses. For type I models, the brittle layer is
absent and Π3 can thus not be defined. For type II models,
Π3 can vary from zero to infinity.
Π3¼thickness of brittle substrata
thickness of ductile substrata ¼df
dd
ð8Þ
Π4 aims to constrain the influence of both the intrusive
flux and the displacement velocity on the deformation.
Π4¼spreading force
intrusive flux ¼uhðddþdfÞ
fð9Þ
The first three numbers are predefined by the experi-
mental setup, while Π4 is a consequence of the input
parameters and the system reaction as u(deformation
velocity) is not fixed or constrained during the experiment.
All values may vary with time according to the deformation
of the model. Other variables that have no dimensions can
be added, such as the number or density of faults and
number of grabens.
Materials
Brittle layer
Pure sand has been used often for models of brittle volcanic
rocks (e.g. Merle and Borgia 1996). In this study, sand is
mixed with 10% plaster to increase the cohesion to a value
scaled with natural rocks (Donnadieu 2000). Furthermore,
this also allows much more structural detail to appear at the
surface.
Ductile layer
Silicone is the ductile rock analogue material, referred to as
low strength layers (LSLs in Oehler et al. 2005) that can
constitute part of the volcano and/or the volcanic basement.
Viscosities of these natural layers are still poorly known,
but are estimated at 10
18
Pa s by Carena et al. (2000) and at
10
14–15
Pa s by Arnaud (2005). The range of viscosities will
influence the displacement velocities, but not the style of
deformation (Merle and Borgia 1996). We chose thus an
intermediary value of 10
17
Pa s. The silicone viscosity here
varies between 10
4
(clean silicone: SGM 36, Dow Corning)
and 8.10
4
Pa s (pure silicone mixed with a small quantity
of sand).
Magma
Magma and intrusive complexes are generally much less
viscous than the surrounding ductile rocks. Thus, we
Table 2 Significant model variables, given with the value ranges and their equivalents in nature. Π−Number description with model and
equivalent natural ranges, and their ratios
Parameters Definition Models Nature Dimensions
hVolcano height 0.01–0.07 170–1,200 m
r
v
Volcano radius 0.12–0.17 2,000–2,900 m
d
d
Thickness of ductile layer 0.006–0.032 100–500 m
d
f
Thickness of brittle layer 0–0.032 0–500 m
uDisplacement velocity 1.4×10
−6
–8.7× 10
−6
5×10
−11
–3×10
−10
ms
−1
fIntrusion flux 3×10
−9
–2×10
−8
3.3× 10
−5
–2.2× 10
−4
m
3
s
−1
μ
I
Intrusion complex bulk viscosity 3–610
14
Pa s
Ratio M/N
Π
1
Height/radius of volcano 0.07–0.52 0.05–0.6 ∼1
Π
2
Volcano height/substratum thickness 1.31–8.33 1.17–12 ∼1
Π
3
Brittle layer/ductile layer thickness 0–4.33 0–5∼1
Π
4
Spreading force/intrusion flux 0.08–1.14 0.04–10.9 ∼1
Bull Volcanol (2012) 74:743–765 747

introduce a viscosity contrast between intrusive complex
and sediment layers. Golden syrup is a good analogue for
magma (Mathieu et al. 2008;2009). Viscosity varies
between 3.5 Pa s at 25°C and 0.084 Pa s at 60°C. Room
temperature varied between 20°C and 27°C during experi-
ments corresponding to a variation of Golden syrup
viscosity from about 5.9 to 3 Pa s.
Magma viscosities vary from 10 to 10
18
Pa s depending
on crystal content and melt composition. According to the
scaling, Golden syrup represents here an intermediate value
which can correspond to a crystal-rich magma, or can
represent a bulk value for an entire intrusive complex (e.g.
Murase and McBirney 1973;Marsh1981; Rosenberg
2001).
Magma fluxes in nature are estimated from volumes
emitted during eruptions, from estimated ages and edifice
volumes, or from deformation and degassing data
(Wadge 1982; McNight and Williams 1997; Hasenaka
and Carmichael 1985; Beauducel et al. 2000; Wicks et al.
2002). Approximate values vary from 0.01 to 0.5 m
3
s
−1
.
Fluxes in models (without eruption) are 10
−9
–10
−8
m
3
s
−1
that scale to 10
−4
–10
−3
m
3
s
−1
in nature. Thus, models
correspond here to 1/100 of the estimated natural fluxes.
These values can be coherent if magma storage does not
correspond to the total amount of magma flux in the
system, some being lost by eruption (a situation that
occurs when the intrusive flux is increased in the models)
or by crystallisation. Furthermore, note that we study here
the growth of an intrusive complex over a long period,
when flux is lower than during short eruptive episodes, i.e.
when most flux estimates have been calculated.
Results
Analogue model description
The parameters we changed and considered were the
volcano slope, the thickness of brittle and ductile layers,
and the intrusive flux. Three brackets of slope values were
tested: ∼10°, ∼20° and ∼30°. It is difficult to construct a
cone with an exact slope, and the angle of each cone was
measured with a clinometer to obtain the true slope angle.
The variation of Π1 values reflects the variation of the
slope value. The thickness of the brittle and ductile layers is
fixed at the start of the experiment, but as the slope, they
vary with time. The intrusive flux was kept low to avoid the
intrusive body appearing at the surface. The flux varies also
with time.
In all the models, characteristic spreading structures were
formed, i.e. grabens, folds (Borgia 1994; Merle and Borgia
1996; Borgia et al. 2000; Walter et al. 2006) and en échelon
strike-slip faults (Delcamp et al. 2008). To describe the
models, we used cardinal points where the north corre-
sponds to the top of the picture.
Type I internal ductile layer volcano
For type I model, steep cones of about 30° were
characterised by intense faulting and fracturing. The
structures formed a dense nearly radial network with certain
sectors slumping out more rapidly than others.
For intermediate slope models (∼20°) fewer structures
formed. A single transverse graben was formed first and
passed through the edifice centre and the intrusive complex
location (Fig. 2a). Subsequent minor grabens were formed,
as well as slumps at the model edge.
The deformation of the low-slope models (∼10°) was
concentrated around the intrusive complex, where a highly
fractured uplift was observed. Small slumps and collapses
were formed at the foot of the edifice as the edge spread out
(Fig. 2b). Few or no grabens were formed in this case.
Type II volcano with substrata ductile layer
These models all evolved in a predictable pattern, with the
early formation of flank grabens and summit fracturing and
flattening (Fig. 2c, d). In models with thick brittle layers,
strike-slip faults developed at the cone foot and were
connected to graben extremities. These faults were propa-
gating in the brittle layer. In the early stages, a subtle
surface bulge with minor fracturing provided evidence of
the intrusive complex growth within the spreading volcano.
This bulge was rapidly transformed into a densely fractured
depression. The bulging phase was not observed for
experiments with a high spreading displacement velocity,
for example for those with thin brittle layers.
In models without a brittle layer, many grabens formed
within the cone. However, once a brittle layer was
introduced, the number of grabens decreased. In fact, an
increase of brittle layer thickness limits the formation of
grabens (see “Dimensionless numbers”section). This effect
was previously observed on gravitational spreading models
(Delcamp et al. 2008). For models with a thick brittle layer,
only a single graben was formed and traversed the edifice
through its centre. Smaller grabens forming discrete
spreading sectors developed from this central graben. An
example of a typical model is J3 (Fig. 2c) that spreads
initially towards the east from a single NNW–SSE
transversal graben. Strike-slip displacements along the
faults have been observed and were associated with the
formation of en échelon fractures (zoom of the graben on
Fig. 2c). The preferential spreading direction (eastward for
the model J3 for example) was found to be due to model
asymmetry, where the thickness of brittle layer varies
slightly (Cecchi 2003).
748 Bull Volcanol (2012) 74:743–765

.
.
.
.
.
.
.
..
.
Fig. 2 a L2 model (Π1= 0.27).
After 35 min; we observed
the formation of an arcuate
transversal graben going
through the intrusion point; bL3
model (Π1= 0.12) after 1 h and
5 min, significant horizontal
deformation above the intrusion
is seen by surface cracks and
exhumation of underlying
layering; cJ3 experiment
after 50 min (Π1=0.15 and
Π3= 4.33); dJ1 model
(Π1= 0.17 and Π3 = 0) after
25 min; two centres of
deformation are clearly visible,
one organised around the
intrusion point (in grey) and
the other (in black lines)
organised around the summit
Bull Volcanol (2012) 74:743–765 749

Thrusts and folds occasionally formed along the model
border when the experiment was limited by a sand
boundary. This even occurred when the boundary was set
far from the model volcano (more than 50 cm). At such
distances in nonintrusion experiments, no thrusting was
observed. Thrusts and folds also appeared at the base of
individual spreading sectors.
Dimensionless numbers analysis
We note that the number of graben increased with slope
(Π1) and the number of graben decreased with increasing
thickness of brittle layer (represented by Π3; Fig. 3a). For
low Π3 values, multiple graben formed, while a Π3 >1.25
allowed the formation of a single graben that traversed the
edifice and passed above the intrusive complex. For larger
Π3 values (>3; Merle and Borgia 1996), spreading should
be inhibited and no structures should form. Even above this
limit, the experiments with intrusive complexes spread,
leading to graben formation (experiments I5, J3 and K3,
Tab le 3). However, for much higher Π3values,the
intrusion propagated as a sill under the edifice and formed
a bulge at the cone foot. A poorly defined graben was
formed at the summit. During the experiments, it was
possible to follow the sill propagation as a graben
progressively formed above the sill.
In thin brittle layer models (Π3 <0.52), the volcano foot
was surrounded by folds. For Π3 >1, no folds were
observed and strike-slip faults radial to the cone were
rather formed.
The Π4 number, and thus displacement velocity and
intrusive flux, did not play any role on the density of
structures observed over the range studied. However, there
is a linear relationship between Π4 and Π1 (Fig. 3b). The
length of the intrusion is linked to Π3 (Fig. 3c).
Morphology of the intrusions
Intrusive complexes were intruded from the feeder pipe into
the ductile basement and then rose into the spreading cone.
In all cases, they intruded and developed within the ductile
level, attested by the thin silicone layer that enveloped the
intrusive complex. Intrusive complex shapes could be
classified into three principal morphologies:
1. The first type was characterised by a cylindrical base of
variable height and diameter (Fig. 4). The summit
could be flat or slightly bulged. The border of the
intrusion was undulating and ridged. These small dyke-
like, pointed intrusion features were linked to substrata
silicone ridges, which formed at the intersection
between the conjugate faults that formed individual
graben. This first type of intrusive complex shape
occurred when the basal ductile layer was covered by a
very thin layer of brittle material (0<Π3<1). The
y = 11.789x-0.499
area of transversal graben
formation
area of fold
formation
a
b
c
number of grabens vs. Π3
number of grabens
0
5
10
15
20
25
0
0.2
0.4
0.6
0.8
1.0
1.2
0 1 2 3 4 5
Π3
Π1
Π3
Π4
Π4 vs Π1
0 0.1 0.2 0.3 0.4 0.5
0
0.05
0.1
0.15
0.2
0.25
0 1 2 3 4
H3 (Π3=0.3)
H5 (Π3=1.5)
I2-2 (Π3=0)
I4 (1h;Π3=2.8)
I5 (1h;Π3=3.6)
J1 (Π3=0)
J3 (50minutes;Π3=4.33)
K4 (1h;Π3=2)
evolution of horizontal displacement velocity
0
0.000002
0.000004
0.000006
0.000008
0.00001
0 2 4 6
slope of 30°
slope of 20°
slope of 10°
mean of displacement velocity (m/s) vs Π3
e
d
Π3
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
20 0 40 60 80
maximal intrusion length
(m)
mean of displacement velocity
(mm/min)
displacement velocity
(mm/min)
time (minutes)
Fig. 3 Graphs of parameters varying in the models: anumber of grabens
vs. Π3 (brittle/ductile thickness ratio); bΠ4 (spreading force/intrusive
flux) vs. Π1 (volcano slope); cmaximum length of the intrusion (in
metres) vs. Π3; daverage of displacement velocity (m s
−1
)vs.Π3; e
evolution of the horizontal displacement velocity (mm min
−1
) vs. time
750 Bull Volcanol (2012) 74:743–765

thicker the brittle layer, the fewer ridges formed,
corresponding to the decrease in the number of graben.
The ridges were also larger. Type I volcanoes (internal
ductile layer) with a slope of 20° and 30° also
developed such intrusive complex morphology (L1
and L2 experiments; Table 3).
2. The second type of intrusive complex morphology
occurred for models covered with a thin brittle layer
(1<Π3<1.56). The intrusion base was more spread
and flattened, and the top was bulged (Fig. 5a).
Crests are visible on the outer parts and are linked to
the traces of fault-related silicone ridges and to the
graben in which the intrusive complex was emplaced.
3. The third type formed when the brittle layer was thick
(2.22<Π3<4.33): intrusive complexes were principally
developed in two opposite directions and formed fat,
short intrusions connected to silicone ridges (Fig. 5b).
For thicker brittle layers, the intrusion spread as a sill
only and eventually rose on a radial dyke-like intrusion
to erupt at the cone foot.
The three main morphologies represent end-members
and general cases. In detail, each intrusive body displayed
slight variations, and occasionally, morphology could be
very different. For example, J1 model (Π1= 0.17 and Π3=0)
allowed the formation of an almost perfect cylinder.
Table 3 Description of the experiments
Angle h(m) d
d
(m) d
f
(m) r
v
(m) r
s
(m) Other information Number of structures
H1 30° 0.057 0.008 0 0.12 0.51 i.f the most rapid/c.s 28
H2 30° 0.0572 0.0078 0 0.15 0.3 c.s 28
H3 30° 0.0497 0.0078 0.0025 0.12 0.28 b.d.s 21
H4 30° 0.0572 0.0078 0 0.1325 0.27 e.s/b.d.s 28
H5 30° 0.05 0.0064 0.01 0.14 0.27 e.s/b.d.s 13
H3sp 30° 0.0565 0.006 0.0025 0.135 0.27 e.s/c.s
H5sp 30° 0.07 0.007 0.01 0.14 0.25 e.s/c.s
I1-1 20° 0.038 0.009 0 0.15 0.28 e.s/edifice without cohesion/b.d.s 24
I1-2 20° 0.0536 0.0104 0 0.145 0.28 e.s/edifice without cohesion/b.d.s 25
I2-1 20° 0.0498 0.0092 0 0.1325 0.28 e.s/cohesive edifice/b.d.s 28
I2-2 20° 0.0542 0.0088 0 0.145 0.28 e.s/cohesive edifice/b.d.s 21
Iannexe2 20° 0.0395 0.0105 0 0.135 0.29 e.s/b.d.s 19
Iannexe3 20° 0.0304 0.0096 0.005 0.13 0.28 e.s/b.d.s 16
Iannexe4 20° 0.0318 0.0102 0.011 0.145 0.29 e.s/b.d.s 14
I3 20° 0.037 0.008 0.01 0.13 0.27 e.s/b.d.s 8
I4 20° 0.0395 0.008 0.0225 0.14 0.29 e.s/b.d.s 5
I5 20° 0.035 0.0075 0.0275 0.125 0.25 e.s/b.d.s 7
J1 10° 0.023 0.012 0 0.1375 0.3 e.s/b.d.s 21
J2 10° 0.015 0.01 0.01 0.13 0.3 e.s/b.d.s 13
J3 10° 0.023 0.0075 0.0325 0.155 0.27 e.s/b.d.s 7
K1 20° 0.04525 0.00775 0 0.125 0.28 e.s/i.f decreased/ b.d.s 18
K2 20° 0.036 0.009 0.02 0.125 0.27 e.s/i.f<i.f K1/ b.d.s 8
K3 10° 0.01 0.0075 0.025 0.1375 0.29 e.s/i.f= i.f K2/ b.d.s 7
K4 20° 0.03 0.01 0.02 0.135 0.28 e.s/i.f < i.f K2/b.d.s 7
L1 30° 0.069 0.011 0 0.155 0.125 b.d.s 22
L2 20° 0.043 0.007 0 0.155 0.125 b.d.s 16
L3 10° 0.02 0.01 0 0.165 0.125 b.d.s 3
L4 10° 0.0225 0.0095 0 0.16 0.125 b.d.s 9
M20° 0.016 0.024 0 0.15 0.6275 e.s/i.f= i.f K4/ b.d.s 33
M2 20° 0.032 0.032 0 0.17 0.6315 e.s/ b.d.s 14
I 30° 0.075 0 0.018 0.125 0.435 No brittle substratum, no volcano
In bold, type I volcano; in italic, type II volcano
Angle model slope, hmodel height, d
d
thickness of ductile layer, d
f
thickness of brittle layer, r
v
volcano radius, r
s
basal layer radius. c.s clean
silicone, i.f intrusion flux, b.d.s silicone mixed with sand, e.s sand edge border
Bull Volcanol (2012) 74:743–765 751

All the intrusive complexes mimic the structures
observed at the surface. Ridges, dykes and undulations
corresponded to the graben formed at the surface. For
example, for the L3 model, the intrusive body reproduced
the horseshoe-shaped graben (Fig. 2b).
Displacement
Different intrusive rates were applied to models of same
geometry. For example, the models H1 and H2 displayed
the same geometry, but H1 had a higher flux (Table 3). In
this case, the displacement rate was higher for H1. In H1,
unlike H2, no clear structural pattern occurred and the
model was rather characterised by disorganised fracturing.
Moreover, the intrusive complex pierced the surface.
A correlation between Π3 and the displacement velocity
exists, where higher Π3 induces lower displacement
velocity (Fig. 3d). Models with Π3> 2 show an acceleration
phase that becomes more pronounced with greater brittle
layer thickness (Fig. 3e). After this acceleration, velocities
generally decreased. We note that I4 model shows
acceleration, but velocity does not diminish afterwards.
2D displacement velocity fields
Using pictures, horizontal displacement fields were calcu-
lated over sequential time steps (Fig. 6). For models with
no intrusion, greater Π1 and Π2 are always associated with
faster spreading, and the spreading rate decreases as the
volcano height is reduced. The first result to note is that
displacement fields obtained for gravitational spreading
alone and for models associated with intrusive complex
emplacement clearly differ. The horizontal displacement
fields were more variable for models with an intrusive
complex (Fig. 6a–c) than without (i.e. classic spreading
cone, Fig. 6d).
In addition, displacement occurred further out from the
centre. Above the intrusive body, the models displayed a
Fig. 4 Results from model I1-2
(Π1= 0.37 and Π3 = 0): apicture
and sketch of exhumed silicone
substrata layer showing central
intrusion and ridges associated
with grabens. The close
relationship between the
intrusion fingers and the ridges
is clearly visible; bthe intrusion
extracted after freezing, in two
views; ccross-sectional sketch
of the model setup
752 Bull Volcanol (2012) 74:743–765

greater horizontal component than for models without
intrusion, and the maximum displacement on the volcano
flanks extended further out.
A model with an intrusive complex that is offset from
the volcano summit produces two centres of deformation.
For example, visual inspection of the J1 model shows two
areas of deformation around which are organised two
systems of grabens (Fig. 2d). These two centres are also
visible on horizontal displacement field maps (Fig. 6a).
When spreading is restricted to one sector, such as in model
J3 (Fig. 6b), the horizontal displacement field is clearly
concentrated on this spreading sector (Fig. 2c), in which the
intrusive complex develops.
3D displacement velocity fields
We used stereophotogrammetry to obtain horizontal and
vertical velocity components. Three cameras and the
10cm
N
N
silicone ridge
dyke
N
crest
dyke
cross section
intrusion (top view)
N
dd=0.8cm
df=2.2cm
20°
3.9cm
brittle layer
ductile layer
sand edge
basal silicone
intrusion without silicone envelope
location of transversal graben
transversal graben
location of transversal graben
intrusion
dyke/ridge connnection
I4. T=1h55
flattened
basement
bulged
summit
crest
silicone
silicone
top view
graben
marks
1 cm East
drop
shape
NorthSouth
b
a
Fig. 5 Details of intrusion shapes and relationships with substrata: a
model I3 (Π1= 0.28 and Π3 = 1.25); intrusion shape after freezing,
note the upper bulge and the flatter base with ridges; bI4 model (Π1=
0.28 and Π3=2.81); picture at the top is the model after 1 h and
55 min. Photo in the middle and associated sketch correspond to
exhumed frozen model. On the bottom photo, the thin silicone layer
that covers the frozen intrusion has been removed to show clearly the
intrusion shape. Dykes were formed and propagated in the silicone
ridges below the grabens
Bull Volcanol (2012) 74:743–765 753

geometry were calibrated with Visual Measurement System
(VMS; Robson and Shortis 2004,http://www.geomsoft.
com). Points were manually chosen in VMS to be sure that
there was low initial error. Point positions were tracked on
each image. Then, the coordinates (x,y,z) were obtained
using a script written in Matlab. The results are synthesised
in Fig. 7.
For one model, we obtain three figures. The X-axis
corresponds to time from the start of the experiment. The
Y-axis represents radial distance from the centre of the
model, so zero is the cone summit and the greatest number
is the model edge. Associated colours give displacement
magnitude for each distance and time. Dark blue areas are
those with no data. The first figure is the variation of
velocity magnitude (horizontal and vertical component)
with time (X-axis). The second figure is the variation of
the horizontal velocity component with time, and the third
figure is the variation of the vertical velocity component
with time.
Figure 7a is a model with an initial slope of 30° and no
brittle substrata. Velocity reaches a maximum and is
concentrated at the centre, but the value varies with time.
The horizontal component is highest between 6 and 10 cm.
We note an increase of the velocity at 4 cm from the centre
between 14 and 20 min and a maximum at the end of the
experiment corresponding to the intrusive complex rising
close to the surface. Vertical velocity reaches a maximum at
the centre and decreases outwards. However, we observe an
increase of velocity towards the end, which again corre-
sponds to the approach of Golden syrup towards the
surface.
Figure 7b corresponds to a model with the same
geometry to that of the model in Fig. 7a, with an initial
slope of 30°, but without any intrusive complex, i.e. where
200 400 600 800 1000 1200 1400
200 400 600 800 1000 1200 1400
1600 200 400 600 800 1000 1200 1400 1600
200
400
600
800
1000
1200
1400
200
400
600
800
1000
1200
1400
200
400
600
800
1000
1200
1400
0.12
0.16
0.2
0.24
0.28
0.32
0.36
0.4
0.44
0.48
0.52
0.56
0.6
0.64
0
0.04
0.08
0.12
0.16
0.2
0.24
0.28
0.32
0.36
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
velocity over 1h-J3
a
c
b
velocity over 50 minutes-E1
1800 2000
1600
1400
1400
1200
1000
1000
800
600
600
400
200
0.4
0.3
0.2
0.1
0
d
velocity over 30 minutes-J1
velocity over 30 minutes-H5
Fig. 6 Horizontal velocity displacement field maps: amap of J1
model (Π1= 0.17 and Π3 = 0) determined over 30 min (model in
Fig. 2d); bmap for J3 model (Π1=0.15 and Π3 = 4.33) over 1 h; cH5
model (Π1= 0.36 and Π3 = 1.56) over 30 min (model in Fig. 2c); d
map of spreading model with no intrusion for comparison
754 Bull Volcanol (2012) 74:743–765

4
8
12
16
20
24
28
4
8
12
16
20
24
28
4
8
12
16
20
24
28
4
8
12
16
20
24
28
4
8
12
16
20
24
28
4
8
12
16
20
24
28
distance in cm from edifice centre
distance in cm from edifice centre
distance in cm from edifice centre
distance in cm from edifice centre
distance in cm from edifice centre distance in cm from edifice centre
middle flank
exterior of the edifice exterior of the edifice
middle flank
7.5.10-6
5.10-6
2.5.10-6
0
0
5.10-6
1.10-6
-2.5.10-6
-5.10-6
2.5.10-6
1.5.10-5
1.10-5
5.10-6
0
7.5.10-6
5.10-6
2.5.10-6
0
0
-5.10-6
-1.10-5
-1.5.10-5
time (minutes) time (minutes)
m.s-1
time (minutes)
time (minutes) time (minutes)
ab
0
Velocity magnitude
Horizontal velocity component
Vertical velocity component
Velocity magnitude
Horizontal velocity component
Vertical velocity component
5 10 15 20 25 30 35 40
510 152025303540
5 10 15 20 25 30 35 40
100 20 30 40 50 60 70 80
1002030405060708090
time (minutes)
1002030405060708090
m.s-1
m.s-1
m.s-1
m.s-1 m.s-1
Fig. 7 3D evolution of displacement velocities for spreading only and
spreading with intrusion models: awith an intrusion and an initial
slope of 30°; bwithout an intrusion and an initial slope of 30°. Y-axis
is the radial distance from the centre of the model. X-axis is the time,
shading gives the velocity magnitude. At the top, evolution of total
velocity magnitude in time; in the middle, evolution of the horizontal
component of velocity during time; at the bottom, evolution of the
vertical component of velocity during time
Bull Volcanol (2012) 74:743–765 755

only gravity spreading occurs. In this case, velocity
magnitude decreases from the centre outwards and we
observe a peak at 6 cm, over the mid-flank of the edifice.
Areas with lower displacement rates reach a constant
velocity sooner than areas with initial high displacement
rates. Maximum of velocity does not occur at the same
location for the xand ycomponents. Vertical velocity
reaches a maximum (downwards) at the centre and
decreases outwards, while horizontal velocity reaches a
maximum (outwards) at the flank (12 cm).
Interpretations of the analogue model results
Role of the brittle substrata
The thickness of substrata brittle layer was found to be the
most important factor that influences the number of
structures formed, the intrusive complex morphology and
the displacement velocity. In general, for spreading volcano
models, increasing the brittle substrata layer (increasing
Π3) restructures the stress field and diminishes the number
of grabens (Merle and Borgia 1996; Delcamp et al. 2008).
Previously, we noted a relationship between the length of
the intrusion and Π3 (Fig. 3c). Intrusive complexes were
emplaced and developed within the grabens. Consequently,
in models where few grabens were formed, a larger relative
volume of Golden syrup was intruded below each graben.
This allowed the intrusive complex to develop fewer but
longer branches. In addition, fewer numbers of grabens
mean that fewer structures can accommodate the deforma-
tion. The displacements along these structures are thus
higher allowing asymmetry to develop. This produces
sector spreading. The spreading sectors formed in the part
of the model where the brittle substratum is thinner.
On the J3 model, gravitational spreading alone was
totally inhibited by a thick brittle substrata (Π3= 4.33);
however, grabens and volcanic rift zones were still formed
(Fig. 2c). This model had structures organised around a
slight thickness variation in the brittle substrata. Conse-
quently, a slight asymmetry in the model favours an
asymmetric development of the intrusive complex that
may have concentrated stresses enough to generate the
limited intrusion-related spreading.
Role of the intrusive complex
The intrusive complex emplacement does not create new
types of structures compared to those formed in a spreading
volcano. At low Π3 values when spreading is already
occurring, intrusion and substrata contribute together to the
formation of the structures, and the two deformation fields
are merged.
When the spreading was supposedly inhibited by the
presence of a thick brittle layer of constant thickness (i.e.
Π3>3), a poorly defined graben or a set of fractures formed
at the summit, i.e. above the intrusive body. The rest of the
deformation was characterised by the propagation of the
complex as a sill below the edifice. This shows that
intrusive complexes can favour a moderate localised
spreading if the volcano is initially in a stability domain.
Obviously, for a very thick brittle cover, spreading would
not be triggered at all. In these cases, the intrusion formed a
sill-like body in the ductile layer. Thus, the intrusive
complex spread, but not the overlying volcano. No precise
limit was found between spreading induced by intrusive
complex emplacement and sill formation, and the change-
over is probably very sensitive to small variabilities in
model construction.
In intrusion models, even when the model boundary was
set up far from the edifice itself, the formation of thrusts
and folds along the edge occurred. At such distances, in no-
intrusion gravitational spreading experiments, no thrusting
was observed. Thus, with intrusion the deformation field
extends further out. This effect is also seen in the
deformation field on Fig. 7,wherein7a there is
deformation registered up to the model edge, but not in
7b, the nonintrusion case.
Two different systems might be responsible for gravity
spreading, the first due to a basal ductile layer under the
volcano and the second due to the growth and the spreading
of the intrusive complex as suggested by e.g. van
Bemmelen (1949), Borgia (1994), Borgia et al. (2000),
Borgia and van Wyk de Vries (2003).
In a few experiments, such as model J1 (Fig. 2d), we
observe two distinguishable systems of graben. Thus, two
deformation complexes may exist: one due to spreading of
silicone layer and the second due to intrusive complex
emplacement.
Displacement velocity and intrusive flux
We previously observed a linear relationship between Π4
and Π1(Fig. 3b)that can be due to a combined effect: the
steeper the cone, the faster the system spreads on the ductile
layer, and consequently, the steeper the cone, the faster the
intrusion may spread. This suggests that the cone may slide
along the intrusion. A similar effect was found for deep
intrusions into volcanoes with decollements by Mathieu
and van Wyk de Vries (2009), where the intrusion domes
the substrate, causing outwards sliding.
For the same geometry, a higher intrusive flux
induces a disorganised structural pattern, and there is
higher displacement velocity (see for example models
H1 and H2). Consequently, intrusive flux influenced
deformation style.
756 Bull Volcanol (2012) 74:743–765

In comparing Fig. 7a with 7b, we see clearly that adding
the intrusion disturbs significantly the displacement field,
which is concentrated around the intrusion. Furthermore,
deformation extends further from the edifice and is more
prolonged as also observed in the 2D horizontal displace-
ment measurements.
Shape and evolution of the intrusion during spreading
The intrusive complex develops vertically and horizontally
depending on the surrounding substrata and edifice. The
extreme case is a model with no volcano or brittle layer:
experiment I (see Table 3) consists of an intrusion in a
silicone layer without sand cover. Deformation is charac-
terised by a bulge stage rapidly followed by a spreading
phase. The intrusion does not pierce the silicone, and its
summit flattens as its diameter increases. In the main set of
models, intrusive complex morphologies also show these
bulging and spreading stages (Figs. 4and 5) with the
presence of summit bump and crests (bulging phase), as
well as long intrusive branches and lateral bulge
formation (spreading phase). The intrusive complex is
thus affected by vertical and horizontal expansion. The
bulging stage is followed by the spreading stage, though
there may be a continuous competition between gravity
spreading and Golden syrup push, as suggested by the
velocity fluctuations.
Rapid outward horizontal deformation is observed above
the rising intrusive complex; at this location, the fine plaster
model surface is highly fragmented allowing the exposure
of the inner cone layers. Displacement field patterns show
this rapid spreading, for example for the J1 model (Fig. 6a).
The above results are in accord with the hypothesis of
Borgia (1994) where the intrusive complex can spread and
also highlights the important link between gravity spread-
ing and the intrusive complex. The fifth stage described by
Borgia (1994), i.e. spreading of intrusive complex, occurred
from the very beginning of the experiments. Consequently,
some of the five stages described by Borgia (1994) could
combine and operate at the same time.
Comparisons with natural cases
In this section, we concentrate on comparing the structures
observed in the topography, in the field and by geophysics
on La Réunion Island, as there is a clear link between major
intrusive complexes and the structure of the island (Gailler
and Lénat 2010; Michon and Saint-Ange 2008). Other
interesting cases are presented elsewhere; for example, the
Mull Central Complex in Scotland has been interpreted as a
gravity sliding and uplifting intrusion that has deformed the
surrounding substrata and volcanic rocks (Mathieu et al.
2008). The examples set out in Fig. 8can also be
interpreted in the light of the modelling. For example, the
Ardnamurchan intrusion (Scotland) has small protuberan-
ces, linked to faults, and these features may be similar to
the ridges seen in some of the analogue intrusions; a close
look at the structure of the intrusion would be interesting to
see if there is evidence of the intrusion growth and
spreading (Fig. 8a). For the Concepción and Maderas
volcanoes (Borgia and van Wyk de Vries 2003), where
there is a clear spreading signature, the postulated intrusive
complexes lie under the edifice; thus, on geological
evidence alone, there is no clear way of distinguishing
gravitational spreading along from intrusion-related spread-
ing (Fig. 8b). However, there are two lines of evidence that
suggest that interplay has been operating: (1) The graben
system on Maderas is slightly offset with respect to the
central crater. This may be related to an offset intrusion. (2)
On Concepción, Borgia and van Wyk de Vries (2003)
report mudstones outcropping at up to 200 m above lake
level, on the edifice flank. This may not be related solely to
the folding related to the spreading of Concepcion, but may
also be related to uplift due to the intrusive complex. For
both of these cases, dedicated deformation monitoring and
gravity modelling might provide confirmation of the
interaction between the intrusive complex and the edifice
spreading.
Piton des Neiges
La Réunion Island is composed of an active edifice, Piton
de la Fournaise, by a dormant edifice, Piton des Neiges, and
a mainly buried volcano, Les Alizés (Lénat et al. 2001). In
the topography of the island, best viewed on a shaded relief
digital elevation model, there is a clear star or flower-
shaped pattern of escarpments and valleys (locally called
Cirques) that look similar to the grabens observed during
our experiments (Fig. 9a, b). This pattern has been already
suggested by Borgia et al. (2000) and van Wyk de Vries et
al. (2001) to be a spreading related set of sector grabens.
While the pattern appears clear in the topography, finding
clear evidence in the field has been more difficult (e.g. van
Wyk de Vries et al. 2001). Here, we report evidence from
several field surveys in Cilaos Cirque and elsewhere. In the
field, few faults are found, even with excellent outcrops and
huge cliffs. However, the rocks forming the central part of
Piton des Neiges contain severely fractured rock, where
lavas are transformed to breccias (Fig. 10). The original
fresh lavas and their associated breccias were already
extensively fractured by emplacement and cooling mecha-
nisms, and such densely fractured units do not favour the
localisation of discrete faults as the preexisting fractures
themselves take up significant amounts of strain (Lebas
2009). Thus, on La Réunion, diffuse shear zones made up
Bull Volcanol (2012) 74:743–765 757

of dense fracture networks are expected for any graben
faults rather than localised fault planes (Figs. 10 and 11).
In outcrops that are located at about 2 km below the
original surface, seen in deep valleys, the lavas and breccias
are affected by strong chloritisation and zeolitisation, and
the fractures are sealed with secondary minerals (Rançon
1985). On outcrops at this depth, mainly low-angle zones of
breccia and low-angle faults are found, accompanied by
sills and outwardly inclined dykes (Figs. 10,11). Some of
the intrusions have highly lobate margins, indicating slow
intrusion into ductile country rock, and some of the faults
have ductile features, such as schistosity development and
foliation (e.g. Famin and Michon 2010). The interface
between brittle and ductile structural features is diffuse, and
brittle features found alongside ductile features, such as the
inclined straight-sided dykes and lobate dykes (Fig. 10),
indicate a strain rate dependence of the deformation. The
inclined dykes also indicate a probable outwards sliding of
the upper part of the edifice that could be interpreted as
sector collapse (Famin and Michon 2010) or spreading-
related displacement, or a combination. The brittle–ductile
interface is also marked in the Salazie Cirque by the
outcrop of the top of gabbroic bodies, with highly deformed
upper contracts, beautifully described by Famin and
Michon (2010).
Geophysical data can be used to determine the extent of
the intrusive bodies on Piton des Neiges. Figure 9c
combines the shaded relief image with gravimetric data.
There is a strong positive gravity anomaly at the centre of
the edifice that can be clearly linked a dense intrusive complex
(Fig. 9d; Gailler and Lénat 2010). Three-dimensional gravity
and magnetic modelling shows a flattened laccolith shape
with a central peak. The modelled 3D and plan view shapes
of this intrusive complex are similar to the type 1 intrusions
observed in the models (Fig. 4b). Furthermore, the gravity
modelling-derived intrusive complex shape has branches
that coincide with the Cirques and escarpments observed
in the topography, suggesting an intimate link between
the intrusion shape and the inferred structural pattern, a
feature also seen in the models.
sediments intrusive complex
volcanic schists
bedding
monocline
schistosity
30°
40°
25°
35°
5 km N
excrescence
a
56°
40N
6°10W 6°5
laminated gabbro inward-directed tilting of
magmatic layer planes
present level of exposure
volcanoclastic sediments
dolerite intrusion 5 km
central
sagging
Lake
Nicaragua
Concepción
Maderas
N
5 km
b
WE
1 km
85°40
11°
33
85°34
edifice
mudstone
tertiary flysch
intrusive complex
detrital sediments
ductile layer
cross section of Ardnamurchan intrusive complex
Fig. 8 a Tertiary intrusive complex of Ardnamurchan (Scotland),
some thick finger-like protuberances from the main body are
connected to strike-slip faults (cross section modified from O’Driscoll
et al. 2006)bShaded relief image and structural sketch of Concepción
and Maderas volcanoes (Nicaragua). These two volcanoes have
spread, or are spreading, and have probably well-developed intrusive
complexes (Borgia and van Wyk de Vries 2003). Note that on
Maderas, some grabens are significantly offset to the east of the actual
summit crater, suggesting two centres of deformation, one gravity-
controlled, and the other possibly intrusion-generated
758 Bull Volcanol (2012) 74:743–765

Piton des Neiges
Piton de la Fournaise
a
c
10 km
N
N
Indian Ocean
PdF
PdN
Faults and escarpments
submarine
edifice and deposits
emerged
island
Indian Ocean
b
coastal slump
d
location
of fig.10
and 11
location
of fig.10
and 11
Fig. 9 a Digital Elevation
Model of La Reunion Island and
submarine flanks. This image
shows the star-shaped central
part of Piton des Neiges in the
context of the broad submarine
flanks that make up the greatest
part of the edifice. Note that the
star graben-like topography
stops at sea level, roughly at
about the same altitude as
brittle–ductile transition level
observed in the eroded edifice.
The outer submarine edifice is
composed of less coherent
sediments, debris avalanche
deposits and hyaloclastites
(e.g. Oehler et al. 2005) that
may accommodate the outward
spreading either on strike-slip
faults or by diffuse folding. The
recent sediments and clastic
deposits would also tend to
mask such structures. bImage
of the subaerial part of the island
with the major escarpments
indicated. Note the star-shaped
pattern over Piton des Neiges
(PdeN) and the horseshoe shape
of Piton de la Fournaise (PdeF).
cTopography of La Réunion
island with the draped Bouger
gravity anomaly. This shows the
close association between the
gravity anomaly and the
star-shaped pattern of
escarpments on Piton des
Neiges. d3D view of the
modelled Piton des Neige
intrusion from Gailler and
Lénat (2010). Coloured scale in
metres. This shows that the
shape of the intrusion has a
similar upper shape to the type 1
intrusions created in the models
Bull Volcanol (2012) 74:743–765 759

The morphology of the intrusive complexes that
have been formed during our experiments thus repli-
cates well the intrusion shape observed in Piton des
Neiges. In addition, the relationships between the
intrusion shape and the structures are similar. The
position of the intrusion is also within the ductile
altered and brecciated volcanic layers. There is extreme
fracturing in the central part of both the models and in
the natural example. We propose that, similar to our
model, the graben structures and fracturing observed
around the Piton des Neiges are linked to the spreading
of the edifice and to the growth and spreading of the
intrusive complex (Fig. 9c, d). The spreading of Piton
des Neiges would have occurred principally during the
formation of the intrusive complex, and the displacement
would have been accommodated at or below sea level by
the outward movement of the volcano flanks (Fig. 12).
The many of the coastal and submarine debris avalanches
mapped around Piton des Neiges (e.g. Oehler et al. 2005)
could have been triggered by this outward displacement
that would steepen the submarine flanks. These steep and
avalanching structures are seen also in the models where
they could be considered as the equivalent of La Réunion
coastal scarps. The debris avalanches may have covered
spreading-related thrust faults, such as those seen below
the Hilina slump in Hawaii (Morgan et al. 2003). Or, as
the flanks of Piton des Neiges are quite shallowly
dipping, thrusts may not have formed, but the graben
may have rather transferred deformation to strike-slip
faults that would leave little topographic expression
(Delcamp et al. 2008).
Piton de la Fournaise
The structures interpreted from outcrop and topography on
Piton de La Fournaise (e.g. Upton and Wadsworth 1965;
Merle et al. 2010) do not have a radial pattern like on Piton
des Neiges, but rather have a series of horseshoe-shaped
faults, superimposed on some caldera-like depressions
surrounded by flatter benches (Fig. 9b). The calderas have
been interpreted as magmatic calderas (Bachèlery 1981), as
hydrothermal calderas (Merle and Lénat 2003) or were
related to a combination of basement fault reactivation and
hydrothermally altered layers (Michon et al. 2007). While
no large gravity high is associated with the main cone of
Piton de la Fournaise, a major high is sited at the east coast
(called the Grand Brulé anomaly; Malengreau et al. 1999;
Gailler 2010) that is known from exploratory drill cores to
Fig. 10 Field images and
sketches of the structure in Piton
des Neiges. aExample of the
intensely fractured interior of
Piton des Neiges. This outcrop
shows a lave sequence, intruded
by dykes, which has been
intensively brecciated. Note the
brecciated dyke above scale
person’s head and dipping
brecciated lava flow core in
the centre of the image. b
Brecciated thrust zone near to
the outcrop in a, showing
steeply dipping lavas, cut by
low-angle thrust, that is intruded
above by a low-angle intrusion.
Similar geometries are also
reported by Famin and Michon
(2010). cLow-angle sheared
dyke (in pink) intruded into lava
sequence. The shear zone
probably predated the intrusion,
but the intrusion has also been
deformed within the shear zone
760 Bull Volcanol (2012) 74:743–765

be a dense gabbroic body (Rançon et al. 1987; Rançon
1990). Also, there is a north–south elongated gravity high
to the west of the active cone under the Plaine des Sables
(Gailler et al. 2009).
The structural pattern seen could be consistent with
sector spreading, as originally suggested by Upton and
Wadsworth (1965), and the geometry looks similar to the
L3 model of this study (Fig. 2b). The upper Plaine des
Sables gravity anomaly fits the location and geometry that
an intrusive complex associated with the Plaine des Sables
fault would have, according to the models. The Grand
Brûlé anomaly is thought to be associated to an older
edifice, the Alizés volcano, and its relationship to the
present topography is not clear.
The present rift zone that extends to the north and
south of Dolomieu crater may be a manifestation of an
elongated intrusion that is developing at depth. While
there is no clear gravity evidence for this, the
deformation data show that there is a body at depth
(around sea level), and recent eruptions have produced
highly crystal-rich lavas that may represent the partial
destruction of the intrusive complex. Also, ground
deformation data (e.g. Froger et al. 2004) indicate
large-scale displacement of the seaward flank towards
the sea. While much of this displacement could be
attributed to shallow dyke emplacement, the underlying
cause could be sector spreading and intrusion develop-
ment that is partly accommodated by dyke intrusion. The
dykes may record the rapid intrusion events in the context
of generally slow seaward spreading, and they could be
the equivalent of the inclined straight dykes seen in Piton
des Neiges (Fig. 11).
pahoehoe
water
breccia
inclined dyke
a
outward-inclined dyke
brecciated shear-zone
low angle minor shears high angle fracture
zone minor shears at 45°
to main shear plane
b
cd
Fig. 11 a Example of a lobate sided dyke intruded into ductile
pahoehoe and breccia sequence with intense chloritisation and
zeolitisation. On the extreme top left of the image, a more recent
straight-sided low-angle dyke cuts the sequence. bExample of lobate
dykes intruded into altered pahoehoe lavas. cLow-angle fault zones
with higher angle straight-sided dykes. dDetail of the shear fracture in
a low-angle shear zone
Bull Volcanol (2012) 74:743–765 761

Conclusions
A volcano based on a ductile basal layer will slowly spread
leading to the formation of grabens and en echelon faults.
The analogue models performed in this study show that
adding an intrusion into a spreading volcano does not
significantly change the structural pattern: no distinctive
and diagnostic new structures are formed. A simple
structural analysis is thus not sufficient to differentiate the
spreading of a ductile substratum from the spreading of an
intrusive complex. Fortunately, deformation fields greatly
differ, and the role of intrusive complex emplacement
within a spreading volcano can be constrained in the field
with deformation or gravity measurements. In addition, the
spreading of an intrusive complex in a spreading volcano
can be observed if the complex is offset from the volcano
summit as observed in our models and possibly at Maderas
volcano.
Shallow and small intrusive complexes were modelled
here (0.1 to 0.2 km
3
). Their morphology depends essen-
tially on the thickness of the brittle layering. Without this
layer, intrusive bodies were circular with a bulged summit
Brecciated upper brittle part of edifice, with
primary lava fracturing, weathered horizons,
and uncompacted sediments: high angle
fractures and breccia zones.
Altered brittle-ductile zone with closed fractures
and cemented breccias. Mostly outward dipping
fractures and some low angle shear-zones
Highly altered and sheared ductile-brittle zone
with low angle shear zones, mostly dipping at
low angle outwards. General movement relates
to outwards sliding, but stretching and intrusive
complex inflation also generate inward dipping
shear zones. Outward dipping brittle dykes.
Ductile bounded dykes and outer altered, sheared
intrusive complex.
a
b
brecciated interior
of grabens
outward dipping dykes
and brittle-ductile
shear zones
coastal deposits
low angle faults
and shallow slumps main intrusive
complex highly altered and ductile
carapace of the intrusive
complex
volcano-delta
deposits
sea level
?
central conduit
zone
stretched zone above
intrusive complex
strike-slip faults and shear
zones relay intrusive complex
expansion and spreading to free
outer slope
poorly constrained lower
geometry of complex;
dense early volcano core
and cumilates
probable gravitational spreading on ductile
pelagic sediments and internal low strength layers
Fig. 12 a Sketches of low-angle deformation in Piton des Neiges. b
Summary cross section of Piton des Neiges with associated deforma-
tion. This shows the large intrusive complex, shaped like a laccolith
with a slightly dome top. It shows the brecciated and altered rock
around the intrusion and the graben structures, as well as the low-
angle slide planes that accommodate the outward intrusion push and
spreading. The submarine flanks of the volcano provide an open
boundary for the deformation to disperse, and flank steepening is
accompanied by slumping and landsliding, although many structures
are obscured by debris avalanches and sediments derived from the
rapid erosion of the island. Piton de la Fournaise would have a similar
structure, but is asymmetric
762 Bull Volcanol (2012) 74:743–765

and a cylindrical base. For a brittle layer thickness equal or
greater than the ductile layer, the complex propagated and
spread as elongated branches that followed a transversal
graben. A complex that is emplaced into a spreading
volcano will thus more likely propagate horizontally than
vertically. The horizontal propagation can favour rift zone
development, and can drive flanks sideways, to create
slumps and possible sector collapse.
In the analogue models, horizontal displacements due to
substratum spreading were increased when the intrusive
fluxes were higher. This observation indicates that spread-
ing of ductile substrata and intrusive complex are intimately
linked. These two processes can play together or operate at
different times, one triggering and/or favouring the other.
The experiments show that the three-armed rift zone, a
typical feature of oceanic islands, is associated with sector
spreading that are due to (1) a presence of a thick brittle
edifice layer and (2) the irregularly and variability of this
layer. A thicker brittle layer (1) reduces the number of
grabens to obtain the three-armed feature, whereas the
asymmetry (2) promotes the sector spreading.
A good relationship has been achieved between natural
cases and our models, using field and geophysical
evidence. We have interpreted the geology and structure
of La Réunion island in terms of the development of a
major intrusive complex for Piton des Neiges, with
associated intense brecciation and the formation of a star-
shaped set of graben. The lateral spreading of this volcano
may have been responsible for the large number of debris
avalanches seen on the submarine flanks. For Piton de la
Fournaise, there has probably been sector spreading, and
the interpreted intrusive bodies in the volcano fit this
hypothesis. Their present rift zone may be evidence of a
developing intrusive complex below the summit, but
elongated north–south, like the earlier Plaine des Sables
body. This study lays the ground to reinterpret volcanic and
intrusive structures in terms of coupled gravity and magma
intrusion tectonics and provides information that can be
used to distinguish such processes on active volcanoes.
Acknowledgements Tate and Lyle kindly provided us with all the
Golden syrup we wished for. The work was partially supported by
ANR 06-CATT-013-01 grant VOLKARISK. The authors acknowledge
the two anonymous reviewers for their constructive comments that
greatly help to improve the manuscript.
References
Annen C, Lénat JF, Provost A (2001) The long-term growth of
volcanic edifices: numerical modeling of the role of dyke
intrusion and lava-flow emplacement. J Volcanol Geotherm Res
105:263–289
Arnaud N (2005) Les Processus de démantèlement des volcans, cas
d’un volcan bouclier en milieu océanique : Le Piton des Neiges
(île de La Réunion). PhD Thesis, Université de La Réunion, Saint
Denis
Bachèlery P (1981) Le Piton de la Fournaise (Ile de la Réunion): étude
volcanologique, structurale et pétrologique. PhD Thesis, Uni-
versité Blaise Pascal, Clermont-Ferrand
Bailey EB, Clough CT, Wright WB, Richey JE, Wilson GV (1924)
Tertiary and post-tertiary geology of Mull Loch Aline and Oban.
British Geological Survey. Mem Geol Surv Great Brit 44:172–
184
Beauducel F, Cornet FH, Suhanto E, Duquesnoy T, Kasser M (2000)
Constraints on magma flux from displacements data at Merapi
volcano, Java. J Geophys Res 105:8193–8204
Borgia A (1994) Dynamic basis of volcano spreading. J Geophys Res
99:17791–17804
Borgia A, van Wyk de Vries B (2003) The volcano-tectonic evolution
of Concepción, Nicaragua. Bull Volcanol 65:248–266
Borgia A, Burr J, Montero W, Morales LD, Alvarado GE (1990)
Fault-propagation folds induced by gravitational failure and
slumping of the Central Costa Rica volcanic range: implications
for large terrestrial and Martian volcanic edifices. J Geophys Res
95:14,357–14,382
Borgia A, Ferrari L, Pasquarè G (1992) Importance of gravitational
spreading in the tectonic and volcanic evolution of Mount Etna.
Nature 357:231–235
Borgia A, Delaney PT, Denlinger RP (2000) Spreading volcanoes.
Annu Rev Earth Planet Sci 28:3409–3412
Brooks BA, Foster J, Sandwell D, Wolfe CJ, Okubo P, Poland M,
Myer D (2008) Magmatically triggered slow slip at Kilauea
Volcano, Hawaii. Science 321(5893):1177
Carena S, Borgia A, Pasquarè G, Battaglia A, Ferraris M, Martelli L,
De Nardo MT (2000) Gravity synclines. J Geophys Res
105:21,819–21,833
Cayol V, Dieterich JH, Okamura AT, Miklius A (2000) High magma
storage rates before the 1983 eruption of Kilauea, Hawaii.
Science 288:2343. doi:10.1126/science.288.5475.2343
Cecchi E (2003) Reconstruction 3D pour la volcanologie: apports
d’une méthode multi-vues par photogrammétrie numérique.
PhD Thesis, Université B.Pascal, Clermont-Ferrand, pp. 132–
134
Chiocci FL, Coltelli M, Bosman A, Cavallaro D (2011) Continental
margin large-scale instability controlling the flank sliding of Etna
volcano. Earth Planet Sci Lett 305:57–64
Clague DA, Denlinger RP (1994) Role of olivine cumulates in
destabilizing the flanks of Hawaiian volcanoes. Bull Volcanol
56:425–434
Delaney PT, Fiske RS, Miklius A, Okamura AT, Sako M (1990) Deep
magma body beneath the summit and rift zones of Kilauea
Volcano, Hawaii. Science 247:1311–1346
Delcamp A, van Wyk de Vries B, James MR (2008) The influence of
edifice slope and substrata on volcano spreading. J Volcanol
Geotherm Res 177:925–943
Dieterich JH (1988) Growth and persistence of Hawaiian volcanic rift
zones. J Geophys Res 93:4258–4270
Donnadieu F (2000) Déstabilisation des édifices volcaniques par les
cryptodômes: modélisation analogique et approche numérique.
PhD Thesis, Université B.Pascal, Clermont-Ferrand, p. 39
Famin V, Michon L (2010) Volcano destabilization by magma
injections in a detachment. Geology 38:219–222
Francis P, Oppenheimer C, Stevenson D (1993) Endogenous growth
of persistently active volcanoes. Nature 366:554–557
Froger JL, Fukushima Y, Briole P, Staudacher T, Souriot T, Villeneuve
N (2004) The deformation field of the August 2003 eruption at
Piton de la Fournaise, Reunion Island, mapped by ASAR
interferometry. Geophys Res Letters 31:L14601, 5 PP
Fukushima Y, Cayol V, Durand P (2005) Finding realistic dike models
from interferometric synthetic aperture radar data: the February
Bull Volcanol (2012) 74:743–765 763

2000 eruption of Piton de la Fournaise. J Geophys Res 110.
doi:10.1029/2004JB003268
Gailler LS (2010) Structure interne d’un système volcanique océanique
de type point chaud, La Réunion: Approches géophysiques. PhD
thesis, University B. Pascal, Clermont-Ferrand
Gailler LS, Lénat JF (2010) Three-dimensional structure of the
submarine flanks of La Réunion inferred from geophysical data.
J Geophys Res B: Solid Earth 115:B12105
Gailler LS, Lénat JF, Lambert M, Levieux G, Villeneuve N, Froger JL
(2009) Gravity structure of Piton de la Fournaise volcano and inferred
mass transfer during the 2007 crisis. J Volcanol Geotherm Res
184:31–48
Hasenaka T, Carmichael ISE (1985) The cinder cone of Michoacan-
Guanajauto, Central Mexico: their age, volume, distribution and
magma discharge rate. J Volcanol Geotherm Res 25:105–124
Hill DP, Zucca JJ (1987) Geophysical constraints on the structure of
Kilauea and Mauna Loa volcanoes and some implications for
seismomagmatic processes. U.S. Geol Surv Prof Pap 1350 pp
903–917
Holohan EP, van Wyk de Vries B, Troll VR (2008) Analogue models
of caldera collapse in strike-slip tectonic regimes. Bull Volcanol.
doi:10.1007/s00445-007-0166-x
Hubbert MK (1937) Theory of scale models as applied to the study of
geologic structures. Bull Geol Soc Am 48:1459–1520
Kauahikaua J, Hildenbrand TG, Webring M (2000) Deep magmatic
structures of Hawaiian volcanoes; imaged by three-dimensional
gravity models. Geology 28(10):883–886
Lebas E (2009) Genèse et nature de la fracturation des édifices
volcaniques: de l’échelle du volcan à l’échelle de la coulée. Ms
Thesis, University B. Pascal, Clermont-Ferrand, p. 51
Lénat JF, Gibert-Malengreau B, Galdeano A (2001) A new model for
the evolution of the volcanic island of Réunion (Indian Ocean). J
Geophys Res B106(5):8645–8663
Malengreau B, Lénat JF, Froger JL (1999) Structure of Réunion Island
(Indian Ocean) inferred from the interpretation of gravity
anomalies. J Volcanol Geotherm Res 88:131–146
Marsh BD (1981) On the crystallinity, probability of occurrence, and
rheology of lava and magma. Contrib Mineral Petrol 78:85–98
Mathieu L, van Wyk de Vries B (2009) Edifice and substrata
deformation induced by intrusive complexes and gravitational
loading in the Mull volcano (Scotland). Bull Volcanol 71
(10):1133–1148
Mathieu L, van Wyk de Vries B, Holohan EP, Troll VR (2008) Dykes,
cups, saucers and sills: analogue experiments on magma
intrusion into brittle rocks. Earth Planet Sci Lett 271(1–4):1–13
McNight SB, Williams SN (1997) Old cinder cone or young
composite volcano? The nature of Cerro Negro, Nicaragua.
Geology 25:339–342
Merle O, Borgia A (1996) Scaled experiments of volcanic spreading. J
Geophys Res 101:13,805–13,817
Merle O, Lénat JF (2003) Hybrid collapse mechanism at Piton de La
Fournaise volcano, Réunion Island, Indian Ocean. J Geophys Res
108(B3):2166
Merle O, Vendeville B (1992) Modélisation analogique de chevauchements
induits par des intrusions magmatiques. C R Acad Sci Paris 315(série
II):1541–1547
Merle O, Mairine P, Michon L, Bachèlery P, Smietana M (2010)
Calderas, landslides and paleo-canyons on Piton de la Fournaise
volcano (La Réunion Island, Indian Ocean). J Volcanol Geotherm
Res 189:131–142
Michon L, Saint-Ange F (2008) Morphology of Piton de la Fournaise
basaltic shield volcano (La Réunion Island): characterization and
implication in the volcano evolution. J Volcanol Geotherm Res
113(B03203):1–19
Michon L, Saint-Ange F, Bachèlery P, Villeneuve N, Staudacher T
(2007) Role of the structural inheritance of the oceanic
lithosphere in the magmato-tectonic evolution of Piton de la
Fournaise volcano (La Réunion Island). J Geophys Res 112:
B04205
Middleton GV, Wilcock PR (1994) Mechanics in the Earth and
environmental sciences. Cambridge University Press, Cambridge
Mogi (1958) Relation between the eruptions of various volcanoes and
deformations of the ground surfaces around them. Bull Earth Res
Inst 36:99–134
Morgan JK, McGovern PJ (2005) Discrete element simulations of
gravitational volcanic deformation: 1. Deformation structures and
geometries. J Geophy Res 110:B05402. doi:10.1029/2004JB003252
Morgan JK, Moore GF, Clague DA (2003) Slope failure and volcanic
spreading along the submarine south flank of Kilauea volcano,
Hawaii. J Geophy Res 108:24–15. doi:10.1029/2003JB002411
Münn S, Walter TR, Klügel A (2006) Gravitationnal spreading
controls rift zones and flank instabilities on El Hierro, Canary
Island. Geol Mag 143(3):257–268
Murase T, McBirney AR (1973) Properties of some common igneous
rocks and their melts at high temperatures. Bull Geol Soc Amer
84:3563–3592
O’Driscoll B, Troll VR, Reavy J, Turner P (2006) The Great Eucrite
intrusion of Ardnamurchan, Scotland: re-evaluation of the ring-
dyke concept. Geology 34:189–192
Oehler JF, van Wyk de Vries B, Labazuy P (2005) Landslides and
spreading of oceanic hot-spot and arc shield volcanoes on Low
Strength Layers (LSLs): an analogue modeling approach. J
Volcanol Geotherm Res 144(1–4):169–189
Okada Y (1992) Internal deformation due to shear and tensile faults in
a half-space. Bull Seism Soc Am 82:1018–1040
Okubo S, Watanabe H (1989) Gravity change caused by a fissure
eruption. Geophys Res Lett 16:445–448
Okubo PG, Benz HM, Chouet BA (1997) Imaging the crustal magma
sources beneath Mauna Loa and Kilauea volcanoes, Hawaii.
Geology 25:865–870
Ramberg H (1981) Gravity, deformation and the Earth’s crust in
theory, experiments and geologic application, vol 2. Academic,
London, p 452
Rançon JP (1985) Hydrothermal history of Piton des Neiges volcano
(Réunion Island, Indian Ocean). J Volcanol Geotherm Res
26:297–315
Rançon JP (1990) Lithostratigraphie du forage du Grand Brûlé.
Implications volcanologiques. In: Lénat JF (ed) Le volcanisme de
la Réunion. Monographie. Centre de Recherches Scientifiques,
Clermont-Ferrand, pp 187–200
Rançon JP, Lerebour P, Auge T (1987) Mise en évidence par forage
d’une chambre magmatique ancienne à l’aplomb de la zone
orientale du Piton de La Fournaise (Ile de La Réunion).
Implications volcanologiques. CR Acad Sc Paris 304:55–60
Robson S, Shortis MR (2004) Visual Measurement System. http://
www.geomsoft.com
Roche O, van Wyk de Vries B, Druitt TH (2001) Sub-surface
structures and collapse mechanisms of summit pit craters. J
Volcanol Geotherm Res 105:1–18
Rosenberg CL (2001) Deformation of partially molten granite: a
review and comparison of experimental and natural case studies.
Int J Earth Sciences 90:60–76
Swanson DA, Duffield WA, Fiske RS (1976) Displacement of the
south flank of Kilauea Volcano: the result of forceful
764 Bull Volcanol (2012) 74:743–765

intrusion of magma into the rift zone. US Geol Surv Prof
Pap 963:39
Tyrell GM (1928) The geology of Arran. PhD Thesis, Glasgow
University, British Geological Survey
Upton BGJ, Wadsworth WJ (1965) Geology of Réunion Island, Indian
Ocean. Nature 207:151–154
Van Bemmelen RW (1949) The geology of Indonesia: general geology
of Indonesia and adjacent archipelagos. Government Printing
Office, The Hague
van Wyk de Vries B, Matela R (1998) Volcano-induced deformation:
numerical models of substratum flexure, spreading and extrusion.
J Volcanol Geotherm Res 81:1–18
van Wyk de Vries B, Cecchi E, Robineau B, Merle O, Batchèlery P
(2001) Factors governing the volcano-tectonic evolution of La
Réunion Island: a morphological, structural and laboratory
modelling approach. Journal of Conference Abstracts 6:800
Wadge G (1982) Steady state of volcanism: evidence from eruptive
histories of polygenetic volcanoes. J Volcanol Geotherm Res
87:4035–4039
Walter TR, Klügel A, Münn S (2006) Rift zone formation on overlapping
volcanoes by gravitational spreading. Terra Nova 18:26–33
Wicks CW, Dzurisin D, Ingebritsen S, Thatcher W, Lu Z, Iverson J
(2002) Magmatic activity beneath the quiescent Three Sisters
volcanic center, central Oregon Cascade range, USA. Geophys
Res Let 29:10–1029
Wooller L, van Wyk de Vries B, Murray JB, Rymer H, Meyer S
(2004) Volcano spreading controlled by dipping substrata.
Geology 32(7):573–576
Bull Volcanol (2012) 74:743–765 765