ELSEVIER
Tectonophysics 312 (1999) 283–301
www.elsevier.com/locate/tecto
Modelling the seismic properties of fast-spreading ridge crustal
Low-Velocity Zones: insights from Oman gabbro textures
Gwenaëlle Lamoureux a , Benoı̂t Ildefonse a,Ł , David Mainprice a
a
Laboratoire de Tectonophysique, CNRS UMR5568, ISTEEM, Université Montpellier II, 34095 Montpellier cedex 05, France
Received 21 August 1998; accepted 31 May 1999
Abstract
Although considerable progress has been made in the study of fast-spreading, mid-ocean ridge magma chambers over
the past fifteen years, the fraction of melt present in the chamber remains poorly constrained and controversial. We
present new constraints obtained by modelling the seismic properties of partially molten gabbros at the ridge axis. P-wave
velocities at low frequencies are calculated in the foliation=lineation reference frame using a differential effective medium
technique. The model takes into account the lattice preferred orientation of the crystalline phase and the average shape of
the melt phase. The structural parameters are obtained from the Oman ophiolite. The structural reference frame is given
by the general trend of the gabbro foliation and the melt fraction and shape are estimated using the textures of nine upper
gabbro samples. The estimated melt fraction and shape depend on the assumptions regarding which part of the observed
textures represent the melt in the gabbroic mush of the magma chamber. However, we can put limits on the reasonable
values for the melt fraction and shape. Our results are consistent with a melt fraction of the order of 10 to 20% in the
Low-Velocity Zone (i.e. the magma chamber), which is anisotropically distributed with the melt pockets preferentially
aligned parallel to the foliation and approximated by oblate ellipsoids with approximate dimensions of 4 : 4 : 1. These
results are also consistent with the seismic structure of the East Pacific rise at 9º300 . The anisotropic melt distribution can,
at least partially, explain the vertical velocity gradient described in the LVZ.  1999 Elsevier Science B.V. All rights
reserved.
Keywords: mid-ocean ridges; magma chambers; gabbro; anisotropy; seismic properties; Oman ophiolite
1. Introduction
During the last fifteen years, our understanding of
fast-spreading oceanic ridges has improved considerably, due to (1) the accumulation of geophysical
data at the fast-spreading East Pacific Rise (EPR)
(e.g. Hale et al., 1982; Vera et al., 1990; Solomon
and Toomey, 1992; Caress et al., 1992; Sinton and
Ł Tel.:
C33 467 14 36 03; Fax: C33 467 14 36 03; E-mail:
[email protected]
Detrick, 1992; Evans et al., 1994; Crawford et al.,
1999), and (2) the progress of the work in ophiolites
inferred to have been derived from a fast-spreading
environment such as the Oman ophiolite (e.g. Nicolas et al., 1988; Nicolas and Boudier, 1995; Boudier
et al., 1997). However, these two approaches lead
to some discrepancies in the ridge=magma chamber
models. One of the differences relates to the amount
of melt present in the crystallising mush at the ridge
axis, and consequently to the size of the magma
chamber. We provide herein some new constraints,
0040-1951/99/$ – see front matter  1999 Elsevier Science B.V. All rights reserved.
PII: S 0 0 4 0 - 1 9 5 1 ( 9 9 ) 0 0 1 8 3 - 3
284
G. Lamoureux et al. / Tectonophysics 312 (1999) 283–301
obtained by modelling the seismic properties of the
partially molten gabbros at the ridge axis using constraints derived from Oman gabbro microstructures,
and by comparing our results to the seismic data
at the EPR. The microstructure of upper gabbros is
used to estimate the shape distribution and amount
of the liquid phase in the magma chamber, and to
constrain how these parameters influence seismic
properties. From these data, we calculate the seismic
velocities, using the tensorial method of Mainprice
(1997).
2. What is the magma chamber?
Seismic data and tomography at the EPR image
a magma lens at the base of the sheeted dike complex (approximately 1–4 km wide, 50–200 m thick),
overlying a triangular Low-Velocity Zone (LVZ) approximately 6–8 km wide and 2–4 km thick (Vera
et al., 1990; Hale et al., 1982; Detrick et al., 1987;
Solomon and Toomey, 1992; Wilcock et al., 1992;
Barth and Mutter, 1996). From the seismic point
of view, the magma chamber is usually restricted
to the magma lens (e.g. Caress et al., 1992) or to
the magma lens and a narrow mush zone immediately below (e.g. Wilcock et al., 1992). The LVZ
is interpreted as a zone that contains a hot, mostly
solid material, with at most 5% of melt (e.g. Caress et al., 1992; Wilcock et al., 1992, 1995). This
estimate is based on the comparison between the
EPR seismic data and published experimental data
on ultra-sonic (high-frequency) velocities in partially
molten basalts (e.g. Murase and McBirney, 1973;
Manghani et al., 1986) and peridotites (e.g. Sato
et al., 1989), or seismic (low-frequency) attenuation
in gabbros at sub-solidus temperatures (Kampfmann
and Berckhemer, 1985).
On the other hand, the Oman ophiolite gabbros,
which are inferred to be produced at an ancient fastspreading centre (e.g. Nicolas et al., 1994; Nicolas
and Boudier, 1995; Boudier et al., 1997), display
dominantly magmatic microstructures (e.g. Nicolas
et al., 1988), suggesting that the amount of melt
present in the system (i.e., the LVZ) was high enough
to allow the deformation of the crystallising gabbros
to be controlled by the presence of melt (Nicolas
and Ildefonse, 1996; Ildefonse et al. in prep.). The
threshold from suspension flow to solid-state flow
in a magmatic mush is defined classically around
35–40% of melt (e.g. Arzi, 1978; Van der Molen
and Paterson, 1979; Fernandez and Gasquet, 1994).
On the basis of simple analogue experiments with
suspensions of paraffin wax particles in a mixture
of oil and alcohol, it has been suggested that this
threshold may be as low as 20% if the particles are
aligned (shape preferred orientation) by the flow of
the suspension (Nicolas, 1992; Nicolas et al., 1993).
This is the lowest published estimate for the transition from suspension flow to solid-state flow and
possibly the minimum estimate for the amount of
melt present in the crystallising mush at the ridge
axis. Thus, there is a large apparent discrepancy in
the amount of melt (and consequently in the size
of the magma chamber) estimated from the seismic
constraints at the EPR (<5%) and from the field
constraints in the Oman ophiolite. However, the latter estimate is not well constrained since it has been
shown that deformation in the Oman gabbros occurs
by grain boundary sliding accommodated by meltassisted pressure solution (Nicolas and Ildefonse,
1996). For such deformation, there is no clearly defined lower bound on the amount of melt required.
In this paper, we aim to put some new constraints
on the melt fraction and its seismic signature in partially molten gabbros, taking into account the spatial
distribution of melt estimated from textures of Oman
gabbros (we use the term ‘texture’ in the classical
geological sense, i.e., to describe the geometry of the
mineralogical aggregate at the microstructural scale,
characterised by the shape, size and distribution of
the different crystal phases).
3. Structural reference frame
In order to compare the results of our modelling
with the seismic velocity profiles of the EPR (e.g.
Vera et al., 1990), we need to define a common reference frame. At the sample scale which is the scale
of our modelling, the reference frame is the magmatic foliation and lineation of the gabbros. P-wave
velocities will be calculated in three directions (X,
Y , and Z , with XY the foliation plane and X the
lineation). In a typical lower crustal section from
Oman, the gabbro foliation in the lower two thirds
G. Lamoureux et al. / Tectonophysics 312 (1999) 283–301
285
2.2
3
EXTRUSIVE
BRECCIA
TED
4
2.5
S
DIKES
5.5
6.0
6.5
3.0
5
depth (km)
SHEETED DIKES
4.0
7.0 km/s
5.0
6
GABBROS
5.5
7
6.0
8
seismic ray paths
9
gabbro foliation
INTERLAYERED MAFICS/ULTRAMAFICS
10
MANTLE ULTRAMAFICS
-2
-1
0
1
2
3
4
5
6
7
8
9
10
distance (km)
Fig. 1. Schematic cross-section of the magma chamber. The seismic profile is after Vera et al. (1990), the foliation pattern is after
Boudier et al. (1996), and the seismic ray paths are after Wilcock et al. (1993).
is nearly parallel to the Moho, while upsection the
foliation steepens rapidly and progressively becomes
subparallel to the overlying sheeted dikes (Nicolas
et al., 1996). The lower gabbros are also layered,
but the layering tends to disappear upsection where
the gabbros are generally homogeneously foliated,
or locally nearly isotropic (e.g. Boudier et al., 1996).
Since the magmatic foliation develops in the magma
chamber, we assume that the structural pattern observed in the Oman gabbros reflects the structural
pattern in the present-day LVZs, an assumption that
is also supported by numerical modelling (Chenevez et al., 1998). The direction of the ray paths
used to obtain tomographic images of the lower crust
beneath the EPR axis is approximately horizontal below the magma lens (Wilcock et al., 1993; Wilcock
et al., 1995). They are either nearly perpendicular
to the upper gabbro foliation (i.e. parallel to Z ) or
nearly parallel to the lower gabbro foliation and lineation (i.e. parallel to X). The relationships of these
various elements are shown in Fig. 1, superimposed
to the Vp profile of the EPR at 9ºN (Vera et al.,
1990).
4. Calculation of the seismic properties from
petrofabrics and average shape of the melt phase
The method we use to model the seismic properties of the gabbroic mush is described in detail by
Mainprice (1997). It is a tensorial model based on the
standard ‘Differential Effective Medium’ technique
(e.g. Bruner, 1976) modified using the poro-elastic
method of Gassmann (1951) which was extended to
elastically anisotropic media (Brown and Korringa,
1975). Seismic velocities are calculated in the chosen
reference frame, at the low frequency limit (i.e., the
pore fluid is considered to be everywhere connected
and the pressure uniform) which corresponds to seismic frequencies, by taking into account the intrinsic
elastic constants, density, and lattice preferred orientation of all constitutive crystalline phases, and the
compressibility, density, and spatial distribution of
the liquid phase. The volume averaged shape of the
melt is represented by a triaxial ellipsoid. To a first
approximation, the attenuation is considered to result
only from the squirt-flow mechanism (e.g. Mavko,
1980; Schmeling, 1985; see discussion in Mainprice,
286
G. Lamoureux et al. / Tectonophysics 312 (1999) 283–301
Fig. 2. Typical textures in the Oman gabbros. (a) Subdoleritic texture (upper gabbros). (b) Tabular texture (upper gabbros). (c) Partly re-equilibrated lower gabbro texture
with lobate grain boundaries. (d) Strongly re-equilibrated mosaic texture in an anorthitic layer. The scale bar is 1 mm.
G. Lamoureux et al. / Tectonophysics 312 (1999) 283–301
1997). Calculations made with liquid represented by
oblate ‘pancake’ ellipsoids of various aspect ratio
demonstrate the strong influence of the distribution
of the liquid on the seismic response of a melt mush
(Mainprice, 1997). For this reason, we attempt in the
following section to constrain the shape of the liquid
phase from observations of Oman gabbro textures.
287
Microscopically, the textures in the upper gabbros
are dominantly subdoleritic or tabular, with euhedral
to subhedral plagioclase laths (Fig. 2a,b). Downsection, the plagioclase tends to become anhedral,
with either lobate or curviplanar grain boundaries,
and mosaic-like textures are frequently observed in
plagioclase-rich layers (e.g. Boudier et al., 1996)
(Fig. 2c,d). The subdoleritic textures are typically
encountered at the top of the igneous section, close
to the root of the sheeted dike complex; they reflect
faster cooling of the mush (i.e., there is not enough
time for the crystal grain boundaries to migrate toward a minimum surface energy pattern, the late
minerals crystallise in the interstices). Faster cooling of the upper gabbros is also supported by the
frequent compositional zoning of the euhedral plagioclases, which is never seen in the layered gabbros
(e.g. Pallister and Hopson, 1981). Tabular textures
(e.g. Boudier et al., 1996) are also present in the
upper gabbros; they are defined by tabular, nearly
euhedral plagioclases with average dimensions of
1:0 ð 0:2 mm which are often strongly aligned, and
clinopyroxene and olivine appearing either as large
oikocrysts or small interstitial crystals.
to determine the occlusion of porosity in crystallising granites, and by Jousselin and Mainprice
(1999) to estimate the shape of the melt phase
in impregnated peridotites of the Moho transition
zone in the Oman ophiolite. These interstitial minerals, crystallised from trapped melt, are dominantly
clinopyroxene and olivine when present, more or less
replaced by secondary minerals (amphibole, serpentine). Thus, the first step in our analysis required us
to separate the plagioclases from the other mineralogical phases. We obtained a binary image, with
the plagioclase in white, and the EBP (‘everything
but plagioclase’) phase in black using three different
techniques:
(1) Automatic image analysis of a sample section
after oxidation by heating in a furnace at 900ºC. Oxidation eventually changes the colours of the minerals
(olivine red, clinopyroxene green, plagioclase white,
serpentinised olivine yellow, and amphibole brown),
which can then be selected easily by a classical image analysis device, if the grains are big enough. We
used this technique only once in this work, since its
resolution was not sufficient.
(2) Automatic image analysis of a thin section.
The natural light image is acquired using a slide
scanner, a video camera, or a photo enlarger. This
technique is fast and allows easy separation of the
plagioclase (uncoloured) from the other phases.
(3) Drawing the thin section using a microfilm
reader. This technique is the most accurate, but also
the most time-consuming. Every grain and grain
boundary can be recognised with no ambiguity and
the grains manually assigned a colour.
Once we have a binary image, the average shape
of the selected phase is evaluated using an autocorrelation technique (Panozzo Heilbronner, 1992), which
gives an oriented average ellipse as a function of
size.
5.2. Image analysis, separation of mineral phases
5.3. Estimates of the melt fraction
We assume that the interstitial phases in subdoleritic or tabular textures, when present, represent to
a first approximation the melt phase at a given instant; and is thus the crystallised melt phase present
in the steady-state magma chamber (the LVZ), at
the structural level considered (upper gabbros). A
similar approach was used by Bryon et al. (1996)
The distribution of the mineral phases was determined on three orthogonal sections from each of nine
upper gabbro samples. The volume fraction of each
phase was then estimated by averaging the fraction
measured on the three orthogonal sections (Table 1).
Six samples display subdoleritic textures, and three
samples display tabular textures. For each of these
5. Estimate of the liquid phase geometry from
upper gabbro textures
5.1. Oman gabbro textures
12.3
39.2
49.4
3
47.7
41.3
0.6
10.3
46.7
6.5
41.2
32
9.3
1.09 : 1.07 : 1
1.5 : 1.1 : 1
1.6 : 1.5 : 1
2.1 : 1.5 : 1
9.8
46.4
3.2
41.9
22.2
19.5
14.1
25.5
35.2
17.3
6
3
45.1
41.7
88OA6a
88OA6b
subdoleritic subdoleritic
1.7 : 1.5 : 1
2.1 : 1.2 : 1
2.7 : 1.3 : 1
2.3 : 1.1 : 1
3
40.5
52.8
0.4
6.1
56.7
7.1
33.4
40.9
11.9
2.7
48.4
54.5
9.8
14.2
1.2 : 1.1 : 1
1.1 : 1.1 : 1
1.45 : 1.35 : 1
2 : 1.2 : 1
17.9
17.9
21
15.9
2.9
15.9
38.9
2
61.1
38.9
18.8
1
81.2
18.8
92OA83a
95OA129
95OA133
subdoleritic subdoleritic subdoleritic
1.45 : 1.14 : 1
1.8 : 1.2 : 1
20
20
9.4
29.4
2
70.6
29.4
94OA36
subdoleritic
1.6 : 1.4 : 1
1.85 : 1.55 : 1
2.5 : 1.75 : 1
15.6
12.6
8
15.9
21.7
8
3
64.3
28.2
0.2
5.9
32.5
92OA39a
tabular
1.9 : 1.8 : 1
3:2:1
2.31 : 1.9 : 1
2.05 : 1.4 : 1
6.2
37.7
0.1
60.6
24.8
8
14.1
24.9
31.1
14.2
1.6
3
60.7
33.7
92OA39b
tabular
2.75 : 3.25 : 1
3.1 : 3.9 : 1
4 : 3.7 : 1
18.3
18.3
13.9
32.2
3
67.8
32.2
94OB28
tabular
1.64 : 1.58 : 1
2.06 : 1.29 : 1
2.62 : 2.03 : 1
2.15 : 1.33 : 1
59.9
35.2
0.4
7.7
37.7
4.2
44.3
23.1
12.1
9.7
25.1
38.4
12.3
8.5
average
Modes are given for the different phases considered (average values from three perpendicular sections. EBP D ‘Everything But Plagioclase’), together with estimates of melt fraction and shape. See text for
further details.
Average shape of E (X : Y : Z)
Average shape of K (X : Y : Z)
Average shape of L (X : Y : Z)
Average shape of F (X : Y : Z)
Image acquisition technique (see text for number)
A. plagioclase (%)
B. clinopyroxene C olivine (%)
C. opaque (%)
D. grain boundaries (%)
E. EBP (B C C C DCpxCOl ) (%)
F. rims of zoned plagioclases (%)
G. cores of zoned plagioclases (%)
H. rims of Cpx C Ol (arbitrary) (%)
I. Cores of Cpx C Ol (arbitrary) (%)
J. small interstitial Cpx C Ol (%)
K. empirical melt fraction 1 (%) C C F C H
L. empirical melt fraction 2 (%) C C D C F C H
M. empirical melt fraction 3 (%) F C J
N. calculated minimum melt fraction (%)
Texture:
Table 1
Results of texture analysis for the nine upper gabbro samples
288
G. Lamoureux et al. / Tectonophysics 312 (1999) 283–301
G. Lamoureux et al. / Tectonophysics 312 (1999) 283–301
a
b
c
d
e
f
Fig. 3. Simplified sketches of an upper gabbro subdoleritic texture showing the different approximations to the melt fraction
(in black). (a) Initial texture. Plagioclase is shown in white with
zoned areas in light grey and clinopyroxene in grey including a
few plagioclase chadacrysts. (b) EBP (‘Everything But Plagioclase’) phase in black, plagioclase in grey. (c) Empirical melt
fraction 1. (d) Empirical melt fraction 2. (e) Rims of zoned plagioclases. (f) Empirical melt fraction 3. See text and Table 1 for
further explanations.
samples, we obtained different approximations of the
steady-state melt volume fraction, depending on what
we assume to be liquid in the gabbroic mush of the
LVZ (Figs. 3–5). First we measured the amount of the
EBP phase in the nine samples (Fig. 3b, Fig. 4b). It
ranges from 18 to 60% (Table 1, Fig. 5) and averages
38%. However, the EBP phase is not a valid approximation of the melt fraction, because the plagioclase is
not the only phase to crystallise early, and late-stage
melts can also crystallise plagioclase (Fig. 6).
Some of the samples studied (92OA39b,
92OA83a, 88OA6b, 88OA6a) display plagioclase
zoning, easily visible in a polarising optical microscope or on back-scattered images from a scanning
electron microscope (Fig. 7). Zoning of the plagioclases results from relatively rapid crystallisation in a
close system; the composition of the cores is around
An80 and the rims average An60 (Pallister and Hopson, 1981). It also demonstrates that the last melts
crystallise plagioclase. Thus, when present in the
samples (Table 1), we incorporate the plagioclaserim phase into the melt phase (Fig. 3c–f, Fig. 4c–f).
289
On the other hand, clinopyroxene (or olivine in
samples 92OA39b and 95OA129) starts to crystallise
above the eutectic temperature (Fig. 6). It forms large
crystals, often enclosing small plagioclases, which
can be distinguished easily from smaller interstitial
pyroxenes. The cores of these early clinopyroxenes
(and olivines) must be part of the crystallised phase
in the steady-state mush. Unfortunately, we have no
information on either the fraction or the geometry of
the parts of these crystals which were removed from
the liquid phase. Chemical zoning of clinopyroxenes
from the upper gabbros is very weak and does not
appear to be consistent (Pallister and Hopson, 1981);
we observed no zoning on SEM back-scattered images (Fig. 7). Moreover, clinopyroxene and olivine
in the upper gabbros are frequently hydrothermally
altered and replaced, at least partially, by secondary
amphibole and serpentine, which makes recognition
of subtle primary zoning impossible. Consequently,
the area of the core region that must be subtracted
from the liquid phase is unconstrained (Fig. 3c,d,
Fig. 4c,d). The last part of the texture which must
be considered when estimating the melt phase is the
grain-boundary phase, which ranges from 5 to 10%
on the drawings (Table 1). In Fig. 3c and Fig. 4c
(empirical estimate 1), the system is closed and melt
pockets are isolated; in Fig. 3d and Fig. 4d (empirical estimate 2), all the grain-boundaries are wetted
and the melt phase is connected. The empirical melt
fraction 1 ranges from 16 to 48% and averages 25%;
the empirical melt fraction 2 ranges from 22 to 55%
and averages 38% (Table 1, Fig. 5). These values
probably overestimate the steady-state melt fraction
in the gabbroic mush of the LVZ, but they are clearly
incompatible with values of a few percent suggested
by geophysicists for the LVZ at the EPR. Indeed,
they are of the same order of magnitude as the melt
fraction inferred to be present in the upper melt lens
by Hussenoeder et al. (1996) who compare their seismic data with the experimental values of Murase and
McBirney (1973). We shall consider these estimates
as upper bounds.
Lower bounds for the melt fraction can be estimated by limiting the melt phase to parts of the
texture that can be identified as late trapped liquid.
When present, the plagioclase zoning is considered
to be part of this trapped melt phase (Table 1, Fig. 3e,
Fig. 4e). The other obvious candidates to belong to
290
G. Lamoureux et al. / Tectonophysics 312 (1999) 283–301
G. Lamoureux et al. / Tectonophysics 312 (1999) 283–301
291
Fig. 5. Comparison of the EBP phase and different melt fraction estimates for the nine upper gabbro samples studied. See text for further
discussion.
the minimum melt fraction are the interstitial, small
clinopyroxenes (and olivines when present). We can
empirically estimate this contribution by simply removing all pyroxenes (and olivines) which are bigger
than an arbitrary chosen size (Fig. 3f, Fig. 4f). This
threshold size depends on the sample and ranges
from 0.5 to 1 mm. The corresponding empirical
melt fraction 3 (plagioclase zoning C small Cpx=Ol)
ranges from 8 to 17% and averages 12% (Table 1,
Fig. 5). Another way to estimate the contribution
of the Cpx=Ol phase is to calculate from the appropriate phase diagram (An67–Di–Fo, Biggar and
Humphries, 1981) the fraction of these minerals
which is consistent with the rims of zoned plagioclase. The respective phase proportions (Pl 54%–
Cpx 40.5%–Ol 5.5%, or Pl 53.4%–Cpx 46.6%) are
estimated from the position of the corresponding
eutectic points in the An67–Di–Fo diagram. The
minimum melt fraction calculated this way ranges
from 1.6 to 14.2% and averages 8.5% (Table 1,
Fig. 5).
5.4. Distribution of the melt fraction (average shape)
From the ellipses obtained by the autocorrelation
method, we construct ellipsoids representing the average shape of the selected phase (Fig. 8). The shape
anisotropy (aspect ratios of the ellipsoid; X : Y : Z )
of the EBP phase is generally very weak. It is slightly
stronger in tabular samples, especially in 94OB28,
reflecting the stronger shape fabrics in these samples
(compare Fig. 2b with Fig. 2a).
The average shapes of the empirical melt fractions 1 and 2 in the tabular samples also show
Fig. 4. Drawings of one of the studied textures (sample 88OA6b, section X Z ) showing different approximations for the melt fraction (in
black). (a–f) Same as Fig. 3. See text and Table 1 for further explanations.
292
G. Lamoureux et al. / Tectonophysics 312 (1999) 283–301
15
00
An
SP
E
IN
0
140
L
0
140
+L
L+
ITE
RTH
ANO
O
1300
1270
E
00
13
E+
00
00
15
DI
OP
SID
E+
L
L
T
RI
14
TE
RS
FO
00
16
00
17
0
180
Di
Fo
Fig. 6. Crystallisation paths for the nine studied samples in the An–Fo–Di phase diagram (after Morse (1980).
stronger anisotropy. The ellipsoids are slightly more
elongated and flatter when the grain boundaries are
incorporated into the melt phase (compare empirical
melt fraction 2 and 1 in Table 1 and Fig. 8). Similar
aspect ratios are also obtained by considering only
the rims of the zoned plagioclase. Thus, the average
shape of the estimated melt fraction is insensitive to
uncertainties regarding which phase represent melt.
The shape is generally prolate, intermediate between
a plane ellipsoid and a pancake shape, the anisotropy
is always weak (averaging about 2.5 : 1.5 : 1), except
for the strongly aligned tabular texture of 94OB28
(Fig. 2b) for which the ellipsoid aspect ratio is about
4 : 4 : 1 (Table 1, Fig. 8). The mean shape ellipsoid is
always parallel to the foliation of the sample, indicating that the distribution of the melt is directly related
to the fabric of the solid component of the partially molten rock. Similarly, the longest axis of the
ellipsoids is parallel to the lineation in the sample.
The anisotropy of melt distribution is weaker in
the subdoleritic gabbros, which were sampled in the
uppermost part of the gabbro section, near the root of
the sheeted dike complex. In this work, we favoured
this type of texture because their geometry is easier
to interpret in terms of liquid distribution. However,
they are not typical of the upper gabbro section,
which is usually more foliated. Therefore, we believe
that the mean measured distribution (2.5 : 1.5 : 1) is
a lower bound; the general melt distribution is more
likely to be similar to the one obtained in sample
94OB28 (i.e., of the order of 4 : 4 : 1).
5.5. Lower gabbros: a lower bound from the
distribution of opaque minerals
In the layered gabbro section (about two lower
thirds of the igneous crust), the textures tend to
equilibrate, with significant grain-boundary migration after primary crystallisation of the constituent
minerals, and the approximations we use in the upper gabbros are no longer valid. We did attempt
to perform the same type of analysis on one layered gabbro sample (95OA183, Fig. 9a), by using
the exceptionally abundant, late-crystallising opaque
(oxides and sulphides) phases as an approximation
to the trapped liquid phase. The opaque phase represents 4.5% of the sample volume, and its mean shape
is 4 : 4 : 1 (Fig. 9b). This shape is remarkably similar
to the most elongated ellipsoid obtained above for
the upper gabbros.
G. Lamoureux et al. / Tectonophysics 312 (1999) 283–301
293
Fig. 7. Scanning electron microscope back-scattered image showing the details of texture in sample 88OA6b. Zoned plagioclase is grey
and clinopyroxene is white. The scale bar is 0.1 mm.
6. Calculated seismic velocities
P-wave velocities were calculated using the estimates of the melt distribution described above, and
the preferred crystal orientations of the crystalline
phase. As the crystalline phase is dominantly plagioclase, we approximated its seismic properties by
using plagioclase fabrics characteristics of upper and
lower gabbros (Fig. 10). The fabric of the lower
gabbro (sample 90OF11, Fig. 10b) is stronger than
the upper gabbro one (sample 95OA129, Fig. 10a).
We use the elastic constants for An57 at ambient
conditions (Aleksandrov et al., 1974), and the bulk
elastic modulus for basalt (14.91 GPa) is calculated
from the P-wave velocity measured at 1200ºC and
ambient pressure; the effect of pressure on the basalt
elastic modulus can be neglected, as can the effects
of pressure and temperature on the plagioclase elastic constants (Mainprice, 1997; the maximum Vp
overestimate for dry plagioclase is 0.6 km s 1 ). The
basalt and An80 densities are 2.7 g cm 3 (Mainprice, 1997) and 2.72 g cm 3 (Smith and Brown,
1988, p. 290), respectively. Fig. 11 shows the evolution of the compressional velocities along the three
principal axes as a function of the melt fraction for
the upper gabbros. The initial anisotropy (prior to
partial melting) is very weak, reflecting the weak
plagioclase crystallographic fabric (Fig. 10a). Note
that the maximum velocity for a foliated aggregate
of plagioclase is perpendicular to the foliation and
not parallel as is the case for olivine and pyroxene
(Mainprice, 1997). Thus, the anisotropy of the crystalline aggregate is overestimated by considering just
the plagioclase fabric. Because of the compensating
effect of olivine and pyroxene, the anisotropy of a
typical gabbro (i.e., about 60% Pl and 40% Cpx=Ol)
is negligible.
6.1. Calculations of seismic velocities for upper
gabbros
The results of the calculation for upper gabbros
are shown in Fig. 11 for the two melt distributions
estimated above (2.5 : 1.5 : 1 and 4 : 4 : 1). Since the
plagioclase fabric used for the crystalline phase is
weak (Fig. 10a), the initial anisotropy is also very
weak; Vp is slightly faster perpendicular to the foliation (VpZ). Because melt is preferentially aligned
with the foliation, VpZ decreases most rapidly with
the melt fraction. Consequently, the sense of the
294
G. Lamoureux et al. / Tectonophysics 312 (1999) 283–301
G. Lamoureux et al. / Tectonophysics 312 (1999) 283–301
295
Fig. 9. Analysis of a lower gabbro sample (95OA183). (a) Drawing of the texture. (b) Average shape of the opaque phase. See text for
further discussion.
anisotropy reverses, with the fastest velocities parallelling the foliation. If the melt is preferentially
aligned in one direction, this direction is the fastest
(Fig. 11a). For the estimated range of melt fractions
(8.5% to 38%), the melt shape estimate of 2.5 : 1.5 : 1
yields VpZ ranging from 4.15 to 6 km s 1 (Fig. 11a),
and the melt shape estimate of 4 : 4 : 1 yields VpZ
ranging from 3.7 to 5.7 km s 1 (Fig. 11b). These
velocities are generally consistent with the upper
half of the LVZ, considering the uncertainties in our
estimate of the melt fraction and the velocities of the
tomographic models (Fig. 1).
6.2. Calculations of seismic velocities for lower
gabbros
The only difference between our calculation of
lower gabbro velocities and the calculation shown in
Fig. 8. Ellipsoids representing the average shape of EBP, empirical melt fractions 1 and 2, and rims of zoned plagioclases. See Table 1
for the axis dimensions.
296
G. Lamoureux et al. / Tectonophysics 312 (1999) 283–301
a
[UVW] =100
[UVW] =010
Contours (%)
Max.Density = 2.83
Upper Hemisphere
pfJ = 1.37
[UVW] =001
Contours (%)
Contours (%)
1.00
2.00
1.00
2.00
3.00
1.00
2.00
3.00
N = 425
N = 425
N = 425
Min.Density =
Non-Polar data
0.05
Max.Density = 3.41
Upper Hemisphere
pfJ = 1.45
Min.Density =
Non-Polar data
0.06
Max.Density = 3.87
Upper Hemisphere
pfJ = 1.28
Min.Density =
Non-Polar data
0.05
b
[UVW] =100
[UVW] =010
Contours (%)
Max.Density = 6.42
Upper Hemisphere
pfJ = 2.26
[UVW] =001
Contours (%)
Contours (%)
1.00
2.00
3.00
4.00
5.00
6.00
1.00
2.00
3.00
4.00
5.00
1.00
2.00
3.00
N = 189
N = 189
N = 189
Min.Density =
Non-Polar data
0.00
Max.Density = 5.76
Upper Hemisphere
pfJ = 2.09
Min.Density =
Non-Polar data
0.00
Max.Density = 3.42
Upper Hemisphere
pfJ = 1.32
Min.Density =
Non-Polar data
0.11
Fig. 10. Lattice fabrics of plagioclase for (a) upper gabbro (sample 95OA129, subdoleritic texture), and (b) lower gabbro (sample
90OF11, tabular texture).
Fig. 11b is the stronger plagioclase fabric (Fig. 10b)
used for the crystalline phase. This results in a
stronger anisotropy at 0% melt (Fig. 12) but this
difference is eliminated with a few percent of melt
present. As explained above, information on the
geometry of the primary igneous texture at the
sample=thin section scale in the lower gabbros is
lost because of re-equilibration of the texture. It is
not possible to estimate the melt distribution from
the textures at the thin section scale, and a safer
approach is to estimate the melt fraction from the
calculated VpX assuming a given melt shape. The
velocities constrained by the seismic tomography are
approximately 5.5 to 7 km s 1 in this region, yielding a melt fraction ranging from 0 to 22% for an
average melt shape of 4 : 4 : 1 (Fig. 12).
7. Discussion
The analogy between the Oman ophiolite and
fast-spreading oceanic ridges has been presented
elsewhere (e.g. Nicolas et al., 1994; Nicolas and
Boudier, 1995; Boudier et al., 1997), and we will
not discuss it again. The main limitation of our approach is that the present-day microstructures and
textures of the Oman gabbros may not correspond
to the steady-state partially crystallised mush for two
reasons. First, it is not necessarily straightforward to
identify any primary igneous texture with the geometry of a trapped liquid (e.g. Means and Park, 1994;
McBirney and Hunter, 1995). Second, the textures
tend to equilibrate very easily under hypersolidus
conditions. In this study, this problem is minimized
by choosing samples from the uppermost part of the
G. Lamoureux et al. / Tectonophysics 312 (1999) 283–301
7
a
6.5
7
0.1
melt shape
2.5:1.5:1
6.5
Q-1 Z
0.1
melt shape
4:4:1
Q-1 Z
6
Q-1Y
0.01
5.5
Q-1X
VpX
4.5
0.001
VpY
4
Q-1Y
5
Q-1X
VpY
4.5
0.001
VpX
4
VpZ
3.5
0.01
5.5
Q- 1
Q- 1
5
Vp (km s-1)
6
Vp (km s-1)
b
297
3.5
VpZ
3
0.0001
0
10
20
30
40
50
3
0.0001
0
melt fraction (%)
10
20
30
40
50
melt fraction (%)
Fig. 11. Calculated P-wave velocity and attenuation in three orthogonal directions for upper gabbros, as a function of melt fraction for
average melt shapes of (a) 2.5 : 1.5 : 1, and (b) 4 : 4 : 1. For upper gabbros, the curves of interest are VpZ and Q 1 Z (bold lines). See text
for further explanations.
igneous section, which crystallised fast enough to
preserve apparent primary textures, with euhedral,
poikilitic and interstitial phases.
In order to approximate the distribution of the
melt phase, we have made a series of hypotheses,
leading to various estimates of both the melt content
and the melt distribution. Rather than choosing a
7
0.1
melt shape
4:4:1
6.5
Q-1 Z
0.01
5.5
Q-1Y
5
Q- 1
Vp (km s-1)
6
Q-1X
VpY
4.5
0.001
VpX
4
3.5
VpZ
3
0.0001
0
10
20
30
40
50
melt fraction (%)
Fig. 12. Calculated P-wave velocity and attenuation for lower
gabbros, as a function of melt fraction for an average melt shape
4 : 4 : 1. For lower gabbros, the curves of interest are VpX and
Q 1 X (bold lines). See text for further explanations.
single value, we favour determining bounds on the
melt fraction (see above). The upper limit ranges
from 25% to 38%, depending on our assumptions
about the distribution of the melt along the grain
boundaries. The best estimate for the grain-boundary
contribution is somewhere between these two endmembers. Experimental studies in various types of
rocks show that the distribution of melt at the aggregate scale is anisotropic (e.g. Longhi and Jurewicz,
1995; Laporte et al., 1997; Cmiral et al., 1998).
Melt preferentially wets planar crystalline faces of
low solid–melt interfacial energy, and is thus related
to crystalline anisotropy. However, there are a few
studies on the melt topology in partially molten,
deforming gabbroic assemblages, and the results
are somewhat contradictory. Longhi and Jurewicz
(1995) report a distribution of wetting angles (45º on
average) reflecting the crystalline anisotropy in an
anorthite aggregate, whereas Dimanov et al. (1998)
observed melt pools that are probably unconnected
in experiments on partially molten labradorite, with
no wetting of the grain boundaries. The relationships
between crystal orientations and the distribution of
melt are still not well established and cannot be
included in this work.
It is also important to consider the scale at which
the melt is distributed. We can infer that the melt
is distributed homogeneously in the upper gabbros,
298
G. Lamoureux et al. / Tectonophysics 312 (1999) 283–301
since they generally have no layering and display a
homogeneous, scale-independent structure. The melt
distribution inferred at the sample=thin section scale
should be representative of the melt-related seismic
signature of the rock at the scale of the seismic
wavelength. For the lower gabbros, the situation is
more complicated since they are heterogeneous by
nature. It has been proposed recently that the lower
part of the magma chamber is fed by the episodic
intrusion of sills (Boudier et al., 1996; Kelemen
et al., 1997). Seismic models in the lower crust have
a resolution of 1 km at best, which probably exceeds
the vertical dimension of these sills. However, there
are no constraints on the timing and the scale of these
intrusions. Consequently, their effect on the average
melt shape used in the model, and thus on the seismic
signature of the lower part of the LVZ is unclear. We
also know from the geochemistry that lower gabbros
are cumulates (Pallister and Hopson, 1981; Kelemen
et al., 1997). A significant amount of melt has to
escape from the system (at least 50% according to
Kelemen et al., 1997). This implies that a large
amount of melt migrates through the gabbroic mush.
However, the mechanism for melt migration remains
unknown. Short-scale chemical variations apparently
preclude pervasive porous flow, and favour a channel
model (Korenaga and Kelemen, 1997). However,
they are presently no field observations to support
7
this mechanism. We have no good constraints on the
melt distribution at a given time in the lower part of
the LVZ.
In the upper gabbro, the lower limit on our estimate of the melt content is of order 10% (12.3% for
the interstitial minerals only, 8.5% for the calculation
using the plagioclase zoning). If we accept 10% as
a reasonable estimate for the minimum melt content,
the corresponding VpZ is 5.7 km s 1 for an average melt shape of 4 : 4 : 1 (Fig. 11b). To explain the
lowest velocities in the uppermost part of the LVZ
(4 to 3 km s 1 , Fig. 1), one needs a melt content
increasing upward toward the melt lens, and=or an
increase in the anisotropy of the melt distribution.
Fig. 13 shows the decrease of VpZ and VpX with
the increasing ellipticity of the average melt shape.
The largest effect is observed perpendicular to the
foliation when the average melt shape changes from
1 : 1 : 1 to 10 : 10 : 1 (Fig. 13a). For all directions, Vp
is relatively insensitive to melt shape once the ellipticities exceed 10 (Fig. 13), and VpZ will unlikely be
lower than 5 km s 1 (Fig. 13a) with a melt fraction
of 10%. Thus, the anisotropy of the melt distribution
alone is not sufficient to explain the lowest velocities
in the uppermost part of the LVZ.
In the lower gabbros, a VpX 5.5–6 km s 1 , a
range typical of the lower part of the LVZ (Fig. 1),
corresponds to a melt fraction ranging from 9.5 to
6.5
a
VpZ
b
VpX
6.5
6
Vp (km s-1)
Vp (km s-1)
6
5.5
1:1:1
5
2:2:1
100:100:1
4.5
4:4:1
10:10:1
100:100:1
10:10:1
5.5
1:1:1
2:2:1
4:4:1
5
4
3.5
4.5
0
5
10
15
20
melt fraction (%)
25
30
0
5
10
15
20
25
30
melt fraction (%)
Fig. 13. Variation in P-wave velocity as a function of melt fraction for various melt shapes. (a) VpZ (perpendicular to the foliation). (b)
VpX (parallel to the lineation).
G. Lamoureux et al. / Tectonophysics 312 (1999) 283–301
22% for a 4 : 4 : 1 melt ellipsoid. Again, we have
a lower bound of order 10%. The velocity along
the lineation (VpX), that is parallel to the maximum
elongation of the representative melt ellipsoid, is not
sensitive to the average shape of the melt fraction
(Fig. 13b). Increases in the velocity at the bottom
of the LVZ (from 6 to 7.5 km s 1 in Fig. 1) are
likely to be related to a progressive increase with
depth of the abundance of ultramafic, olivine-rich
layers in the basal layered gabbros, as is commonly
observed in Oman. Increases in Vp would then be
the result of the higher Vp of olivine and to the
Vp anisotropy of olivine which counteracts that in
plagioclase, since the fast velocity is parallel to the
lineation. A decrease in the melt fraction downward
is not required, although it may contribute to the
trend.
The accuracy of the velocity models (Fig. 1) is
weak, especially in the lower half of the crust (Dunn
et al., 1999). Thus our estimate of the melt fraction
for the lower gabbros from Vp is not well constrained. Nevertheless, our results are consistent with
melt fraction ranging from 10% to 30% in the LVZ.
The upper bound is not well constrained since estimates are highly variable (see Table 1). We can infer
that the melt fraction is similar throughout the LVZ,
as implied by the constant viscosity in the numerical
models of Chenevez et al. (1998). The upper bound
can then be lowered to 22%, which is the upper limit
for the lower gabbros. With a melt fraction of the
order of 10 to 20%, the seismic structure of the LVZ,
and in particular its vertical velocity gradient, can be
explained by the combined structural effects of the
melt being preferentially aligned in the developing
foliation, and of the foliation being variably oriented
with respect to the seismic ray paths. This melt content is still at least twice as large as independent
estimates previously obtained for the LVZ on the basis of seismic (e.g. Caress et al., 1992; Wilcock et al.,
1992, 1995) or electromagnetic (Evans et al., 1994)
experiments on the EPR. As Wilcock et al. (1995)
point out, no laboratory studies completely simulate
the conditions of the seismic experiments. We note
that our estimate of the melt fraction is consistent
with (1) the field-derived evidence (Nicolas and Ildefonse, 1996), (2) compliance measurements on the
EPR which yield an estimate of 2–18% for the melt
fraction in the LVZ (Crawford et al., 1999), and (3)
299
a recent three-dimensional tomographic model of the
EPR at 9º300 N (Dunn et al., 1999) which yields an
estimate of 5–22% at 4 km depth and 2–11% at 6
km depth for the melt fraction.
8. Conclusions
We have analysed the texture of nine gabbro
samples from the Oman ophiolite, in order to estimate the melt fraction and its average shape in
the crystallising mush present in the steady-state
magma chamber (LVZ) at fast spreading ridges. We
have chosen upper gabbro samples because they display primary textures with interstitial phases that
are assumed to represent the approximate geometry
of melt in the magmatic mush of the steady-state
magma chamber. Our approach does not allow a precise estimate of the melt fraction in the upper half of
the LVZ. However, it places bounds on the possible
melt fraction and on the spatial distribution of melt at
the aggregate scale. The average melt fraction ranges
from approximately 10% to 30%. It is important to
note that the lower bound (¾10%) is significantly
higher than previous estimates obtained from most
geophysical studies on the EPR, and is in better
agreement with estimates derived from the Oman
ophiolite and recent geophysical studies on the EPR
(Crawford et al., 1999; Dunn et al., 1999). The melt
is preferentially aligned with the mineral foliation,
and has an average shape which can be approximated
by a 2.5 : 1.5 : 1 ellipsoid. We think that this shape
probably underestimates ellipticity, and should be
taken as a minimum, since most of the samples used
for our analysis display weaker foliation than the
typical upper gabbros. The 4 : 4 : 1 ellipsoid obtained
for the sample with a well foliated tabular texture
more likely represents the average melt shape. Direct
estimates of the melt fraction and distribution are not
possible in the lower gabbros. However, an estimate
can be obtained indirectly from our modelling. In
the lower part of the LVZ, the influence of the melt
shape is weak because the melt lenses are nearly
aligned with the seismic ray paths; the melt fraction is estimated from velocities in the tomographic
image (Fig. 1) and is about 10 to 20%.
Previous interpretations of the seismic structure of
the East Pacific Rise at 9º300 (Fig. 1) were based on
300
G. Lamoureux et al. / Tectonophysics 312 (1999) 283–301
the assumption that the melt was distributed isotropically and, on the basis of the available experimental
data, concluded that the melt fraction was small. We
suggest that the same seismic structure is consistent with a larger amount of melt (of the order of
10–20%) which is anisotropically distributed. The
anisotropy does not need to be very large (of the
order of 4 : 4 : 1 with the long axes parallel to the
foliation plane) to account for the amount of melt
estimated from the textures of Oman upper gabbros,
and the vertical velocity gradient reported in the LVZ
of the EPR. The latter observation can be explained
by the preferred orientation of the melt, and by the
way it is oriented with respect to the seismic ray
paths. Lower velocities in the upper half of the LVZ
are not exclusively a result of a decrease with depth
of the melt content and=or temperature. They may
also result from the steepening of the flow pattern
in the upper part of the magma chamber beneath
the melt lens, as reflected by the steepening of the
magmatic foliation in the Oman gabbros.
Acknowledgements
This work benefited from high-quality thin sections done by Christophe Nevado, and from discussions with various persons, especially Françoise
Boudier, Adolphe Nicolas, and David Jousselin. We
thank William Wilcock and an anonymous reviewer
for very careful and helpful reviews.
References
Aleksandrov, K.S., Alchikov, U.V., Belikov, B.P., Zaslavskii, B.I.,
Krupnyi, A.I., 1974. Velocities of elastic waves in minerals
at atmospheric pressure and increasing precision of elastic
constants by means of EVM (in Russian). Izv. Acad. Sci.
USSR, Geol. Ser. 10, 15–24.
Arzi, A.A., 1978. Critical phenomena in the rheology of partially
melted rocks. Tectonophysics 44, 173–184.
Barth, G.A., Mutter, J.C., 1996. Variability in oceanic crustal
thickness and structure: multichannel seismic reflection results
from the northern East Pacific Rise. J. Geophys. Res. 101,
17951–17975.
Biggar, G.M., Humphries, D.J., 1981. The plagioclase, forsterite,
diopside, liquid equilibrium in the system CaO–Na2 O–MgO–
Al2 O3 –SiO2 . Mineral. Mag. 44, 309–314.
Boudier, F., Nicolas, A., Ildefonse, B., 1996. Magma chambers
in the Oman ophiolite: fed from the top or from the bottom?
Earth Planet. Sci. Lett. 144, 239–250.
Boudier, F., Nicolas, A., Ildefonse, B., Jousselin, D., 1997. EPR
microplates, a model for the Oman ophiolite. Terra Nova 9,
79–82.
Brown, R.J.S., Korringa, J., 1975. On the dependence of the
elastic properties of a porous rock on the compressibility of
the pore fluid. Geophysics 40, 608–616.
Bruner, W.M., 1976. Comment on ‘Seismic velocities in dry and
saturated cracked solids’. J. Geophys. Res. 81, 2573–2576.
Bryon, D.N., Atherton, M.P., Cheadle, M.J., Hunter, R.H., 1996.
Melt movement and the occlusion of porosity in crystallizing
granitic systems. Mineral. Mag. 60, 163–171.
Caress, D.W., Burnett, M.S., Orcutt, J.A., 1992. Tomographic
image of the axial low-velocity zone at 12º500 N on the East
Pacific Rise. J. Geophys. Res. 97, 9243–9263.
Chenevez, J., Machetel, P., Nicolas, A., 1998. Numerical models
of magma chambers in the Oman ophiolite. J. Geophys. Res.
103, 15443–15455,
Cmiral, M., Gerald, J.D.F., Faul, U.H., Green, D.H., 1998. A
close look at dihedral angles and melt geometry in olivine–
basalt aggregates: a TEM study. Contrib. Mineral. Petrol. 130,
336–345.
Crawford, W.C., Webb, S.C., Hildebrand, J.A., 1999. Constraints
on melt in the lower crust and Moho at the East Pacific Rise,
9 degrees 480 N, using seafloor compliance measurements. J.
Geophys. Res. 104, 2923–2939.
Detrick, R.S., Buhl, P., Vera, E., Mutter, J., Orcutt, J., Madsen, J.,
Brocher, T., 1987. Multi-channel seismic imaging of a crustal
magma chamber along the East Pacific Rise. Nature 326, 35–
41.
Dimanov, A., Dresen, G., Wirth, R., 1998. High-temperature
creep of partially molten plagioclase aggregates. J. Geophys.
Res. 103, 9651–9664.
Dunn, R.A., Toomey, D.R., Solomon, S.C., 1999. Three-dimensional seismic structure and physical properties of the crust
and shallow mantle beneath the East Pacific Rise at 9º300 N. J.
Geophys. Res. (submitted).
Evans, R.L., Sinha, M.C., Constable, S.C., Unsworth, M.J., 1994.
On the electrical nature of the axial melt zone at 13ºN on the
East Pacific Rise. J. Geophys. Res. 99, 577–588.
Fernandez, A.N., Gasquet, D.R., 1994. Relative rheological evolution of chemically contrasted coeval magmas: example of the
Tichka plutonic complex (Morocco). Contrib. Mineral. Petrol.
116, 316–326.
Gassmann, F., 1951. Über die Elastizität poroser Medien. Vierteljahrsschr. Naturforsch. Ges. 96, 1–23.
Hale, L.D., Morton, C.J., Sleep, N.H., 1982. Reinterpretation of
seismic reflection data over the East Pacific Rise. J. Geophys.
Res. 87, 7707–7717.
Hussenoeder, S.A., Collins, J.A., Kent, G.M., Detrick, R.S.,
Harding, A.J., Orcutt, J.A., Mutter, J.C., Buhl, P., 1996. Seismic analysis of the axial magma chamber reflector along the
southern East Pacific Rise from conventional reflection profiling. J. Geophys. Res. 101, 22087–22105.
Jousselin, D., Mainprice, D., 1999. Melt topology and seismic
G. Lamoureux et al. / Tectonophysics 312 (1999) 283–301
anisotropy in mantle peridotites of the Oman ophiolite. Earth
Planet. Sci. Lett. 164, 553–568.
Kampfmann, W., Berckhemer, H., 1985. High temperature experiments on the elastic and anelastic behavior of magmatic
rocks. Phys. Earth Planet. Inter. 40, 223–247.
Kelemen, P., Koga, K., Shimizu, N., 1997. Geochemistry of
gabbro sills in the crust–mantle transition zone of the Oman
ophiolite: implications for the origin of the oceanic lower
crust. Earth Planet. Sci. Lett. 146, 475–488.
Korenaga, J., Kelemen, P.B., 1997. Origin of gabbro sills on the
Moho transition zone of the Oman ophiolite: implications for
magma transport in the oceanic lower crust. J. Geophys. Res.
102, 27729–27749.
Laporte, D., Rapaille, C., Provost, A., 1997. Wetting angles,
equilibrium melt geometry, and the permeability threshold of
partially molten crustal protoliths. In: Bouchez, J.L., Hutton,
D., Stephens, W.E. (Eds.), Granite: From Segregation of Melt
to Emplacement Fabrics. Kluwer, Dordrecht, pp. 31–54.
Longhi, J., Jurewicz, S.R., 1995. Plagioclase-melt wetting angles
and textures: implications for anorthosites. Lunar Planet. Sci.
26, 859–860.
Mainprice, D., 1997. Modelling the anisotropic seismic properties of partially molten rocks found at mid-ocean ridges.
Tectonophysics 279, 161–179.
Manghani, M.H., Sato, H., Rai, C.S., 1986. Ultrasonic velocity
and attenuation measurements on basalt melts to 1500ºC: role
of composition and structure in the viscoelastic properties. J.
Geophys. Res. 91, 9333–9342.
Mavko, G.M., 1980. Velocity and attenuation in partially molten
rocks. J. Geophys. Res. 85, 5173–5189.
McBirney, A.R., Hunter, R.H., 1995. The cumulate paradigm
reconsidered. J. Geol. 103, 114–122.
Means, W.D., Park, Y., 1994. New experimental approach to
understanding igneous texture. Geology 22, 323–326.
Morse, S.A., 1980. Basalts and Phase Diagrams. Springer, Berlin,
477 pp.
Murase, T., McBirney, A.R., 1973. Properties of some common
igneous rocks and their melts at high temperatures. Geol. Soc.
Am. Bull. 84, 3563–3592.
Nicolas, A., 1992. Kinematics in magmatic rocks with special
reference to gabbros. J. Petrol. 33, 891–915.
Nicolas, A., Boudier, F., 1995. Mapping oceanic ridge segments
in Oman ophiolites. J. Geophys. Res. 100, 6179–6197.
Nicolas, A., Ildefonse, B., 1996. Flow mechanism and viscosity
in basaltic magma chambers. Geophys. Res. Lett. 23, 2013–
2016.
Nicolas, A., Reuber, I., Benn, K., 1988. A new magma chamber
model based on structural studies in the Oman ophiolite.
Tectonophysics 151, 87–105.
Nicolas, A., Freydier, C., Godard, M., Vauchez, A., 1993.
301
Magma chambers at oceanic ridges: how large? Geology 21,
53–56.
Nicolas, A., Boudier, F., Ildefonse, B., 1994. Evidence from
the Oman ophiolite for active mantle upwelling beneath a
fast-spreading ridge. Nature 370, 51–53.
Nicolas, A., Boudier, F., Ildefonse, B., 1996. Variable crustal
thickness in the Oman ophiolite: implication for oceanic crust.
J. Geophys. Res. 101, 17941–17950.
Pallister, J.S., Hopson, C.A., 1981. Semail ophiolite plutonic
suite; field relations, phase variation, cryptic variation and
layering, and a model of a spreading ridge magma chamber. J.
Geophys. Res. 86, 2593–2644.
Panozzo Heilbronner, R., 1992. The autocorrelation function: an
image processing tool for fabric analysis. Tectonophysics 212,
351–370.
Sato, H., Sacks, I.S., Murase, T., 1989. The use of laboratory
velocity data for estimating temperature and partial melt fraction in the low-velocity zone: comparison with heat flow and
electrical conductivity studies. J. Geophys. Res. 94, 5689–
5704.
Schmeling, H., 1985. Numerical models on the influence of
partial melt on elastic, anelastic and electric properties of
rocks, Part I. Elasticity and anelasticity. Phys. Earth Planet.
Inter. 41, 34–57.
Sinton, J.M., Detrick, R.S., 1992. Mid-ocean ridge magma chambers. J. Geophys. Res. 97, 197–216.
Smith, J.V., Brown, W.L., 1988. Feldspar Minerals, Volume
1. Crystal Structures, Physical, Chemical and Microtextural
Properties. Springer, Berlin, 2nd revised and extended edition.
Solomon, S.C., Toomey, D.R., 1992. The structure of mid-ocean
ridges. Annu. Rev. Earth Planet. Sci. 20, 329–364.
Van der Molen, I., Paterson, M.S., 1979. Experimental deformation of partially-melted granite. Contrib. Mineral. Petrol. 70,
299–318.
Vera, E., Mutter, J.C., Buhl, P., Orcutt, J.A., Harding, A.J., Kappus, M.E., Detrick, R.S., Brocher, T.M., 1990. The structure of
0- to 0.2-m.y.-old oceanic crust at 9ºN on the East Pacific Rise
from expanded spread profiles. J. Geophys. Res. 95, 15529–
15556.
Wilcock, W.S.D., Solomon, S.C., Purdy, G.M., Toomey, D.R.,
1992. The seismic attenuation structure of a fast-spreading
mid-ocean ridge. Science 258, 1470–1474.
Wilcock, W.S.D., Dougherty, M.E., Solomon, S.C., Purdy, G.M.,
Toomey, D.R., 1993. Seismic propagation across the East
Pacific Rise: finite difference experiments and implications for
seismic tomography. J. Geophys. Res. 98, 19913–19932.
Wilcock, W.S.D., Solomon, S.C., Purdy, G.M., Toomey, D.R.,
1995. Seismic attenuation structure of the East Pacific Rise
near 9 degrees 300 N. J. Geophys. Res. 100, 24147–24165.