Download Wrench faults down to the asthenosphere

Vauchez, A., and Tommasi, A., 2003, Wrench faults down to the asthenosphere: Geological and geophysical evidence
and thermo-mechanical effects, in Storti, F., Holdsworth, R.E., and Salvini, F., eds., Intraplate Strike-Slip
Deformation Belts: London, Geological Society of London, Special Publication 210, p. 15-34.
Wrench faults down to the asthenosphere: Geological and
geophysical evidence and thermo-mechanical effects
Alain Vauchez and AndréaTommasi
Laboratoire de Tectonophysique, Université de Montpellier II et CNRS
Pl. Eugène Bataillon, F-34095 Montpellier cedex 5, France
[email protected]; [email protected]
Abstract:
We review a set of geological and geophysical observations that strongly support a coherent
deformation of the entire lithosphere in major intracontinental wrench faults. Tectonic studies of wrench
faults eroded down to the middle to lower crust show that, even in cases in which the lower to middle crust is
partially melted, strain remains localized (although less efficiently) in transcurrent shear zones. Seismic
profiling as well as seismic tomography and magnetotelluric soundings provide strong argument in favor of
major wrench faults crosscutting the Moho and deforming the upper mantle. Pn velocity anisotropy, shearwave splitting and electric conductivity anisotropy measurements over major wrench faults and in
transpressional domains support that a wrench fault fabric exists over most or even the entire lithosphere
thickness. These seismic and electrical anisotropies are generated by a crystallographic preferred
orientation of olivine and pyroxenes developed in the mantle during the fault activity, which is frozen in the
lithospheric mantle when the deformation stops. The preservation of such a "wrench fault-type" fabric within
the upper mantle may have major effects on the subsequent tectono-thermal behavior of continents, because
olivine is mechanically and thermally anisotropic. Indeed, the association of numerical models and
laboratory data on textured mantle rocks strongly suggests that the orogenic continental lithosphere is an
anisotropic medium with regards to its stiffness and to heat diffusion. This anisotropy may explain the
frequent reactivation, at the continents scale, of ancient lithospheric-scale wrench faults and transpressional
belts during subsequent tectonic events.
Introduction:
Horizontal displacements in transcurrent
faults represent one of the fundamental modes
of accommodation of deformation in the
crust. It is quite obvious that transcurrent
faults generated at transform plate boundaries,
like the San Andreas Fault in California or the
Alpine fault in New Zealand crosscut the
entire lithosphere. It is however less clear
whether intracontinental strike-slip fault
systems generated in active margins or in
collisional domains are only crustal structures
or are rooted in the upper mantle. The
penetration of a "wrench-fault type" tectonic
fabric (i.e., a vertical flow plane associated
with a horizontal flow direction) deep into the
upper mantle may have major geodynamic
implications, since it would generate an
anisotropy of the mechanical and thermal
properties of the lithospheric mantle and,
hence, modify the large-scale rheological
behavior of continental plates during
subsequent tectonic events (Tommasi et al.,
2001; Tommasi & Vauchez, 2001).
Assuming that major, i.e., continentalscale, strike-slip faults observed today at the
surface continue down to the base of the
lithosphere implies a strong mechanical
coupling between the various rheological
layers of the lithosphere. This raises the
question of the mechanical properties of the
hot middle to lower crust. Strain localization
should remain efficient enough to allow the
development of strike-slip faults zones at this
level. In addition, rheological contrasts
between the lower crust and the upper mantle
should remain moderate; otherwise the lower
crust would behave as a horizontal decoupling
level in which upper crustal wrench faults
would root. These issues have been already
addressed in a large number of studies on the
rheological stratification of the continental
lithosphere (e.g., Meissner et al., 2002;
Molnar, 1988; Ranalli & Murphy, 1987;
Vauchez et al., 1997), but experimental data
on the rheology of lower crustal materials is
so limited that these studies are not
conclusive.
In this paper, in order to evaluate how
deep a coherent "transcurrent fabric" may
penetrate, we analyze direct observations
from surface geology, which are of course
restricted to the crust, and indirect
information from geophysics and
geochemistry that gives a hint on the
crust/mantle coupling. We consider evidence
from active and fossil tectonic domains and
discuss observations from both individual
shear zones and broad transpressive domains.
The review of this broad dataset suggests that
major wrench faults do crosscut the entire
lithosphere. This leads us to discuss the effect
of these lithospheric-scale wrench faults on
the thermo-mechanical evolution of
continental plates.
Transcurrent shear zones and strain
localization in a hot middle to lower crust
If major transcurrent faults were rooted into
the crust, the wrench deformation in the upper
crust must be decoupled from the mantle flow.
Decoupling between crustal and mantle
deformations is supposed to be favored in the
middle to lower crust (especially in regions
displaying high geothermal gradients) by the low
stiffness of crustal material at high homologous
temperature (T/Tm, with Tm!=!melting
temperature). It would be marked by rooting of
the strike-slip faults into this low stiffness layer,
and therefore by a listric shape of the fault in
order to accommodate the transition from a
vertical to a horizontal flow plane.
In this section, we examine a set of
continental-scale transcurrent faults eroded to
increasingly deeper levels from the middle to the
lower crust. In all these cases, during transcurrent
deformation, the crustal levels exposed today
were submitted to high temperatures and even
partial melting. These levels represent former low
viscosity layers into which crustal-scale strike-slip
faults might have rooted.
.
Figure 1: The high-temperature Borborema shear zone system of Northeastern Brazil (Vauchez et al., 1995). (a) Sketch map
showing the complex pattern of transcurrent faults formed during the Neoproterozoic orogeny: (1) Neoproterozoic granitoids, (2)
Mid- and Late Proterozoic sedimentary basins, (3) Mesozoic sedimentary basins, (4) Neoproterozoic high-temperature shear zones,
and (5) Neoproterozoic low-temperature shear zones. (b) and (c) are two Landsat images showing segments of two major hightemperature wrench faults: the Patos and the West Pernambuco shear zones, respectively. Gray lines in (b) mark the shear zone
limits
2
Figure 2: High-temperature vertical foliation (S) and
horizontal mineral stretching lineation (L) in a mylonite
from the Borborema shear zone system. Deformation in this
felsic mylonite occurred at T>600°C. Location on Fig.1a.
Scale bar is 0.5m.
In northeastern Brazil, the Neoproterozoic
province of Borborema displays a complex
network of wrench faults (Figure 1) that are
several hundred kilometers long and up to 30 km
wide (e.g., Vauchez et al., 1995). Satellite images
highlight a clear textural contrast between the
shear zones and the country rock. This contrast is
mostly due to the transition from a predominant
low-angle metamorphic foliation outside the shear
zones to a steeply dipping mylonitic foliation
within the shear zones. At the satellite image
scale, the boundaries of the fault zones appear
usually rather sharp, although in the field a
continuous transition from the external flat-lying
foliation to the internal steeply dipping foliation
(half "flower-structure") is observed where no
subsequent reactivation concealed the original
relationships. Mylonites outcropping in the shear
zones were formed at depths of 16-18 km
(P!≤!500 MPa) and at high temperature (>650°C).
Figure 3: Migmatitic mylonite from the West
Pernambuco shear zone (see location on Figure 1a).
Downward view. Intense shearing occurred along a
subvertical foliation in a partially melted crust. White
layers are leucocratic neosome. Arrows indicate
dextral shear.
Under these conditions, the protoliths of the
mylonites
(metasediments, pre-kinematic
intrusives, felsic gneisses from the basement)
were partially melted and the resulting rock is
indeed a migmatitic mylonite. At these
temperature conditions, felsic rocks are expected
to display low viscosity, which will be further
decreased by partial melting. Nevertheless, even
when the degree of melting is rather high, the
foliation in the shear zones remains consistently
steeply dipping and bears a shallow-dipping
stretching lineation (Figure 2). Shear-sense
indicators developed in the partially melted
mylonites consistently support dextral wrenching
(Figure 3). Evidence of downward decrease of the
foliation dip, suggesting rooting of the faults, has
never been reported. On the contrary, a large
volume of mantle-derived magmas, especially
diorites, was emplaced as syn- or late-kinematic
dikes (Figure 4) and/or elongated plutons within
the shear zones (Neves et al., 2000; Vauchez et
al., 1995); this strongly suggests that the faults
were connected to a partially melted upper mantle.
The Neoproterozoic Mozambique belt in
Madagascar and East Africa is also characterized
by the development of a large network of wrench
faults (Figure 5) at ca. 530-500 Ma (Martelat et
al., 2000). The present-day level of exposure
shows rocks that were 20 to 30 km deep during
the deformation [0.5 to 1.1 GPa; \Martelat, 2000
#4157; Pili, 1997 #3960]. At these depths,
deformation took place at temperatures >750°C.
The major shear zones in this domain are typically
several hundred kilometers long and up to 40 km
wide. Numerous minor ductile wrench faults
formed under similar P-T conditions are also
documented. The tectonic fabric in the shear
Figure 4: Diorite dikes injected in a porphyritic
granodiorite emplaced in the Pernambuco shear zone
(location on figure 1a). Dikes were emplaced within the
transcurrent shear zone and deformed before complete
solidification. No evidence of solid-state deformation has
been observed
3
zones is typical of ductile strike-slip faults: the
foliation is steeply dipping, the mineral-stretching
lineation is sub-horizontal and consistent shearsense criteria are observed. Outside the shear
zones, the granulites that form the country-rock
display a low-angle foliation and the fabric is
metamorphic-migmatitic rather than mylonitic.
According to Martelat et al. (2000), the
deformation regime in the southern Mozambique
belt was transpressional and the deformation was
partitioned; transcurrent shearing was localized
within the vertical shear zones and large-scale
folding accommodated transverse shortening.
Through a regional scale investigation of the Cand O-isotopes compositions of carbonates from
marbles and metabasites, Pili et al. (1997b) have
shown that CO2 in the major wrench faults of the
network has a mantle origin. This suggests that
these major faults were connected to the mantle.
On the other hand, in minor shear zones and in
metamorphic rocks outside the shear zones, CO2
has a crustal isotopic signature. In the same
region, Pili et al. (1997a) documented a
systematic association of a short wavelength
positive gravity anomaly to major strike-slip shear
zones that also supports a deep rooting of the
major wrench faults of the Mozambique belt. This
anomaly was interpreted as due to a shallower
crust-mantle boundary beneath the faults. Such an
upward deflection of the Moho might result from
thinning of the crust in response to the intense
stretching associated with simple shear in the fault
zones (Pili et al., 1997a).
In NE-Brazil as well as in the Madagascar
Neoproterozoic belts, strain localization in
transcurrent shear zones is observed even at
crustal levels where synkinematic temperatures
were high enough to induce partial melting. The
width of the fault zones is extremely large (several
tens of kilometers) compared to typical widths of
shear zones developed under lower temperature
conditions (cm to hundred m). This points out
that, at these high temperatures, strain localization
was less efficient and strain was distributed over a
larger volume of rocks than it is usually observed
in upper crustal shear zones. Rocks within the
shear zones display a high-temperature mylonitic
fabric largely due to dislocation creep assisted by
very effective diffusional processes (in particular
grain boundary migration), and consistent shear
criteria. In addition, petrological and geochemical
observations strongly suggest that fluids
percolated from the mantle into the crust along
these major shear zones, and therefore that the
faults were continuous through the upper mantle.
"Moho" faults vs. lithospheric faults
The observations presented above
strongly support that major transcurrent faults do
not root in some intracrustal decoupling level, but
rather crosscut the entire crust and are, in some
way, connected to the upper mantle. These
observations are however not sufficient to
evaluate whether those faults are rooted at the
crust-mantle interface or penetrate deeply into the
upper mantle. Clear evidence supporting that
major wrench faults crosscut the Moho and
penetrate deeply into the upper mantle is
nevertheless obtained by combining various
4
techniques of geophysical exploration of the
lithosphere. Evidence may be subdivided in two
groups. Seismic profiling, magnetotelluric
soundings, and seismic tomography have imaged
"Moho faults" (Diaconescu et al., 1997), i.e.,
discontinuities crosscutting the Moho beneath
several wrench faults observed at the surface. On
the other hand, electric conductivity anisotropy
evidenced in magnetotelluric soundings,
azimuthal anisotropy of Pn velocities, and Swaves splitting are directly related to the tectonic
fabric of the upper mantle and support that the
lithospheric mantle was deformed in major
wrench faults.
Electric conductivity anisotropy in the upper
mantle is interpreted as due to a preferred
orientation of graphite films elongated along the
foliation (Mareschal et al., 1995) or to an
anisotropic electrical conductivity in a "wet"
mantle due to the anisotropy of H+ diffusion in the
olivine crystal (Mackwell & Kohlstedt, 1990;
Simpson, 2001). In both cases, a "wrench-fault
type fabric" (i.e., a steeply-dipping flow plane, or
foliation, containing a subhorizontal flow
direction, or lineation) within the mantle would
generate a higher conductivity parallel to the trace
of the wrench fault observed at the surface.
Seismic anisotropy in the upper mantle, which
Figure 6: Cartoon illustrating the concept of lithospheric
fault, in which crustal fault zones broaden downward and tend
to coalesce forming a broad shear zone that cuts across the
entire lithospheric mantle. It displays the tectonic fabric
associated with the fault within the crust and the mantle, the
crystallographic fabric of olivine expected to develop in the
mantle section of such a fault zone (oriented in the structural
framework of the fault: X=lineation and Z=normal to the
foliation), and the splitting of a polarized incoming shear wave
that propagates across a lithospheric mantle displaying a
"wrench fault type" fabric. A seismic station locate above such
a lithospheric shear zone will record a fast shear wave
polarized parallel to the shear zone trend (X direction) and
strong delay times (>1s).
may be characterized by measurement of an
azimuthal anisotropy of Pn velocities or by the
splitting of teleseismic S-waves, results from the
lattice preferred orientation (LPO) of rockforming minerals during high-temperature
deformation by dislocation creep. Wrenchfaulting within the lithospheric mantle would
generate a LPO of olivine, the dominant mineral
phase in mantle peridotites, characterized by a
concentration of [100] axes close to the lineation
(i.e., subhorizontal) and of [010] axes normal to
the foliation plane [Figure 6; \Tommasi, 1999
#3215]. Olivine is elastically anisotropic. Thus if
deformation produces coherent olivine LPOs at
the scale of tens of km in the upper mantle, it also
results in anisotropic seismic properties
(Mainprice & Silver, 1993; Nicolas &
Christensen, 1987; Silver et al., 1999). P-waves
that propagate either parallel to the maximum of
[100] or [010] axes of olivine in the mantle are
respectively the fastest and the slowest. On the
other hand, S-waves propagating through a
deformed upper mantle split into two quasi-S
waves polarized in orthogonal planes; the fastest
one is polarized in a plane containing both the
maximum concentration of olivine [100] axis and
the propagation direction. The delay time between
the arrivals of the two split waves is proportional
to both the length of wave propagation path within
the deformed layer and the propagation direction
relative to the structural fabric; the largest Swaves splitting is observed for waves that
propagate at low angles to the maximum of [001]
axis. A wrench-fault fabric in the mantle would
therefore be evidenced (Figure 6) by a fast
propagation of P-waves (in particular, horizontally
propagating Pn waves) parallel to the fault and a
polarization of the fast split S-wave in a plane
containing both the direction of propagation of the
wave and the lineation, i.e., parallel to the fault
direction for waves having an almost vertical
incidence (such as SKS, SKKS, PKS...). It is also
in this case that the birefringence will be the
largest, leading to relatively large time lags
between the arrivals of the fast and slow split Swaves. Indeed, SKS splitting data above transform
boundaries, such as the Caribbean or the Alpine
fault in New Zealand, systematically display fast
shear waves polarized parallel to the transform
direction and delay times significantly larger than
1s, which imply that the entire lithosphere
deformed in a strike-slip regime (Klosko et al.,
1999; Russo et al., 1996).
These techniques "probe" the upper
mantle fabric with different spatial resolutions
and depth sensitivities. Magnetotelluric
5
soundings using a large spectrum of
measurement frequencies allow an evaluation
of the electrical conductivity anisotropy from
the crust to the asthenospheric mantle.
However, MT data depend on both anisotropy
and heterogeneity of electrical conductivity,
and reliable anisotropy determinations may
only be obtained when high-quality, longperiod MT transfer functions are available and
lateral conductivity gradients are small
(Simpson, 2001). Pn waves sample the
uppermost mantle (3-5 km beneath the
Moho), but the measured velocities depend on
both the anisotropy and the heterogeneity (in
temperature and composition) along the wave
path. Teleseismic S-waves splitting provides
reliable evidence of seismic anisotropy with a
very good spatial resolution (ca. 50 km), but
these measurements integrate all anisotropic
contributions along the wave path (which is
roughly vertical from the core-mantle
boundary to the surface for the most
commonly used SKS waves). Their
association should therefore allow us to better
constrain the structural fabric of the upper
mantle. Indeed, comparison of electric
conductivity anisotropy determined by
magnetotelluric soundings and S-waves
splitting measurements shows that the
direction of largest conductivity and the fast
split S-wave polarization plane are often
almost parallel (Barruol et al., 1997b;
Simpson, 2001; Wannamaker et al., 1996) or
make a slight, but consistent angle (Mareschal
et al., 1995). Ji et al. (1996) interpreted this
slight obliquity as representing the obliquity
between the foliation and the shear plane in
shear zones.
To investigate how deep a "wrench
fault fabric" may penetrate into the upper
mantle, we analyze geophysical data for
several ancient or active wrench faults and
transpressional belts. In each case,
transcurrent displacement, either in a single
fault or in a broader domain of transpressional
deformation, is supported by surface geology.
Transcurrent shear zones
Recently, Pollitz et al. (Pollitz et al.,
2000; Pollitz et al., 2001) using a combination
of GPS and synthetic aperture radar (InSAR)
data, have shown that the deformation in the
years following the 1992-Landers and 1999Hector Mine major earthquakes in the Mojave
desert (California-USA) was about 3 times
greater than before the earthquakes. This
interseismic velocity field supports a right
lateral displacement parallel to the San
Andreas transform fault system (Figure 7).
According to these authors, the visco-elastic
relaxation of the lower crust and upper mantle
was the dominant postseismic process; this
requires that the lower crust acted as a
coherent stress guide coupling the upper crust
with the upper mantle (Pollitz et al., 2001).
Figure 7: (a) Structural sketch displaying the active faults
in California. (b) Shear-wave splitting in western California from
Hartog et al. (2001). Anisotropy beneath the westernmost
stations, i.e., those above the San Andreas Fault system, results
from the superposition of two anisotropic layers. The upper layer,
which corresponds to the lithospheric mantle, is characterized by
a polarization of the fast shear wave (black bars) in a plane
parallel to the San Andreas Fault system and a delay time close to
or even higher than1s. The easternmost stations display a simpler
anisotropy pattern (gray bars) that may be accounted for by a
single anisotropic layer with a roughly E-W flow direction. A
similar flow direction is inferred for the lower anisotropic layer
(gray bars) in the westernmost California. (c) Horizontal velocity
field showing the contemporary interseismic deformation across
southern California (relative to a group of GPS and VLBI stations
on the stable North American Plate). Geodetic data include
Global Positioning System (GPS), Very Long Baseline
Interferometry (VLBI), and Electro-optical Distance
Measurement (EDM) obtained by the Crustal Deformation
Working Group of the Southern California Earthquake Center
during the past three decades. Error ellipses are regions of 95%
confidence. Release 2, 1998, available at:
http://www.scecdc.scec.org:3128/group_e/release.v2
6
These conclusions are consistent with those
drawn from the analysis of the seismic anisotropy
measured across the San Andreas Fault system
slightly north of the Landers and Hector Mine
earthquakes area (Hartog & Schwartz, 2001;
Ozalaybey & Savage, 1995; Silver & Savage,
1994). The shear waves splitting parameters
retrieved from a large number of records
consistently suggest two layers of anisotropy
within the upper mantle (Figure 7). The upper
layer, which corresponds to the lithospheric
mantle, is characterized by a polarization of the
fast split shear wave parallel to the San Andreas
Fault system. This suggests that the lithospheric
mantle has a tectonic fabric consistent with the
crustal fabric, i.e., a steeply dipping foliation
bearing a subhorizontal lineation. Both geodetic
and seismologic observations therefore converge
toward a coherent deformation of the entire
lithosphere.
The Himalayan orogen provides some
of the best examples of active wrench faults
in an intracontinental setting. These faults
have accommodated large lateral
displacements associated with the India-Asia
collision (e.g., Tapponnier et al., 1986). The
main faults of the system have been mapped
over hundreds of kilometers and are
commonly several kilometers wide. The Red
River fault, for instance, was recognized over
1000!km from Tibet to the Gulf of Tonkin.
Pham et al. (1995) have performed a 70!km
long magnetotelluric profile across the Red
River fault system in North Vietnam (Yen Bai
region). In this area, the Red River system is
formed by three parallel transcurrent faults, a
few tens of kilometers apart (Tapponnier et
al., 1990). This MT survey (Figure 8) shows
that: 1) each fault is characterized by a high
conductivity zone down to the uppermost
mantle, 2) the Sông Hông fault, the main
branch of the Red River fault system,
separates two lithospheric domains presenting
contrasted electrical properties, and 3) a large
conductivity anisotropy is observed in both
the crust and the uppermost mantle; the
direction of highest conductivity is
consistently parallel to the strike of the faults.
This anisotropy is consistent with a steeply
dipping foliation within the uppermost mantle
as well as in the entire crust.
Figure 8: Magnetotelluric soundings from Pham et al.
(1995) across the Red River Fault system. (a) MT
geoelectrical section obtained by 2D numerical modeling
showing marked resistivity contrasts between domains
separated by the faults. Each bloc is characterized by its
longitudinal (i.e., parallel to the strike of the faults) and
transverse (in brackets) resistivities (in Wm). Lowresistivity domains beneath each branch of the fault are
displayed in light gray. Conductive zones in the lower
crust and uppermost mantle are displayed in medium and
dark gray, respectively. (b) MT sounding curves showing
a pronounced variation in apparent resistivity between
the transverse (normal to the strike of the faults) and
longitudinal direction in both the crust and the uppermost
mantle. The highest conductivity is parallel to the strike of
the faults, a result in good agreement with a "wrench fault
type" fabric in the uppermost mantle
In Tibet, seismic anisotropy
measurements have been performed above
and in the vicinity of two other well-know
major wrench faults, the Altyn Tagh and the
Kunlun faults. These faults, several hundreds
of kilometers long (1800 km for the Altyn
Tagh fault), have accommodated several
hundreds km of lateral escape during the
India-Eurasia collision (e.g., Tapponnier et
al., 1986). Wittlinger et al (1998) have
performed a seismic tomography study of an
area where the Altyn Tagh fault juxtaposes
Precambrian basement with the Qailam
sedimentary basin. This tomography shows a
southeastern domain characterized by lowvelocity perturbations in contrast with a
northwestern domain where high-velocity
7
perturbations dominate. The limit between
these domains is marked by a low-velocity
anomaly located just beneath the Altyn Tagh
fault (Figure 9a). From these results
Wittlinger et al. (1998) have suggested that
the Altyn Tagh fault in the mantle is ca. 40
km wide and is continuous down to a depth of
140 km at least. In addition, shear-wave
splitting measurements above the Altyn Tagh
fault (Herquel et al., 1999) show fast split
shear waves polarized in a plane parallel to
the trend of the fault and delay times between
the fast and slow S-waves arrivals of ca. 1s.
Such delay times require a thickness of
anisotropic mantle of ca. 100 km, in
agreement with the values of fault penetration
inferred from seismic tomography (Figure
9b). Shear-wave splitting measurement above
and across the Kunlun fault (Herquel et al.,
1999; McNamara et al., 1994) have reached
Figure 9: Mantle structure beneath the Altyn Tagh and
Kunlun active faults in Tibet. (A) Cross section displaying the
main geological structures and the P-wave velocity structure
across the Altyn Tagh fault system (Wittlinger et al., 1996).
Light gray and dark gray colors correspond to the crust and
mantle, respectively. Lighter shades in both layers indicate
domains of lower P-wave velocities. (B) Compilation of shearwave splitting measurements across the Kunlun and Altyn
Tagh faults from Herquel et al. (1999). Both faults are
characterized by a fast split shear wave polarized parallel to
the trend of the fault, contrasting significantly with the
anisotropy pattern away from the faults.
similar results. Approaching the Kunlun fault
zone the orientation of the fast S-wave
polarization plane progressively rotates into
parallelism with the trend of the fault,
suggesting a shear strain gradient and an
upper mantle fabric similar to that in the crust.
The 2s of delay time measured above the
Kunlun fault requires a thickness of
anisotropic material >200km, assuming a
steeply dipping flow plane and a
subhorizontal flow direction, thus larger than
the lithosphere thickness. This suggests that
the asthenosphere fabric also contributes to
the recorded anisotropy and deforms
somewhat coherently with the lithosphere.
Similar observations also characterize ancient
wrench faults whose fabric was frozen into
the lithospheric mantle at the end of the
orogenic evolution. The Great Glen – Walls
Boundary fault (GGWBF) is a major wrench
fault that belongs to a more complex fault
array developed in the northern segment of
the Caledonian belt between 428 and 390 Ma
(e.g., Stewart et al., 1999). Two segments of
the initial fault are exposed: the Great Glen
fault in Scotland and the Walls Boundary
fault in the Shetland Islands. Paleomagnetic
reconstructions suggest that several hundred
kilometers of sinistral strike-slip displacement
Figure 10: Shear-wave splitting in northern UK from
Helffrich (1995). Initials (e.g., MCD, LRW…) represent the
name of the stations. APM is the Absolute Plate Motion in
the hot-spot framework calculated using Morgan and
Morgan's model (see, Barruol et al., 1997a). Thick gray
line north of the Shetland Islands marks the location of the
UNST deep seismic reflection profile displayed in Fig. 11.
8
have been accommodated along this fault.
Shear-wave splitting has been measured
(Helffrich, 1995) at stations close to the
GGWBF in Scotland (Figure 10; station
MCD) and in the Shetland Islands (Figure 10;
station LRW). In both stations, the fast split
shear wave is polarized in a plane parallel to
the trace of the fault and a delay time of 0.94
and 0.53 s is observed between the arrivals of
the two SKS waves for MCD and LRW,
respectively. The fast S-wave polarization
direction close to the fault is significantly
oblique to the fast polarization direction
measured at other stations in the British
Caledonides (Barruol et al., 1997a).
Interestingly, several seismic profiles
performed across the GGWBF, in Scotland as
well as in the Shetland Islands (e.g.,
Klemperer & Hobbs, 1991; Klemperer et al.,
1991; McGeary, 1989), show topography and
a change in the seismic expression of the
Moho tightly associated with the trace of the
GGWBF at the surface (Figure 11). These
features have been interpreted as due to the
fault crosscutting the Moho and bounding two
initially remote domains that show contrasted
seismic responses. This interpretation is in
good agreement with shear-wave splitting
measurements. Altogether these results
strongly suggest that the GGWBF, rather than
being rooted in some crustal decoupling level
(McBride, 1995), is a lithospheric fault that
crosscuts the Moho and penetrates deeply into
the upper mantle.
The well-known South Armorican Shear
Zone (SASZ) in Brittany, France, is a major
intracontinental transcurrent fault formed
during the Hercynian orogeny. Surface
Figure 11: Deep seismic reflection profile across the
Shetland platform (McGeary, 1989). M1, M2, M3 indicate
Moho reflectors. D refers to diffraction hyperbolae.
Figure 12: Deep lithospheric structure beneath the South
and North Armorican shear zones (SASZ and NASZ,
respectively) in Brittany, West of France. (a) P-velocity
model from Judenherc et al. (in press) showing that the
SASZ separates a northern domain characterized by high
seismic velocities from a southern domain, where low
velocities dominate. High P-wave velocities below the
lithosphere (below 90 km) are interpreted as representing a
fossil slab. (b) Shear-wave splitting measurements.
Approaching the SASZ, the fast split shear wave
polarization turns parallel to the trend of the fault,
suggesting a coherent tectonic fabric in both the crust and
the mantle. In contrast, shear-wave splitting measurements
above the NASZ do not show fast shear waves polarized
parallel to the fault trend, suggesting that this latter is a
crustal fault.
geology evidence of strain localization and
strike-slip displacement has been reported in a
large number of papers (e.g., Berthé et al.,
1979; Jegouzo, 1980). The fault is located
north of the high-pressure domain that marks
the trace of the suture between two collided
continents. A seismic velocity model of the
structure of the lithosphere down to 200km
beneath Brittany has been obtained through a
recent passive seismology experiment (Granet
9
et al., 2000; Judenherc, 2000). P-wave
velocity perturbation models show a marked
contrast between two domains (Figure 12):
the northeastern domain is characterized by a
positive velocity anomaly, whereas the
southwestern domain displays negative
anomalies. The limit between these two
domains coincides with the trace of the SASZ
and is observed down to the base of the
lithosphere. In addition, the direction of fast
propagation of Pn waves and the direction of
the polarization plane of the fast split shear
wave are consistently parallel to the trend of
the SASZ. The delay time between the fast
and slow split shear waves at stations close to
the SASZ is consistently larger than 1s, also
suggesting that the entire lithosphere displays
a "wrench-fault type" fabric (Judenherc,
2000). These results are very consistent and
altogether suggest that the South Armorican
Shear Zone crosscuts the entire lithosphere.
Combined magnetotelluric (MT) and
seismic anisotropy measurements (Figure 13)
have been recently performed in the vicinity
of the Proterozoic Great Slave Lake shear
zone (GSLSZ; e.g., Hanmer et al. 1992), in
northwestern Canada (Wu et al., 2002). This
NE-SW trending dextral wrench fault is 25
km wide and its magnetic expression can be
correlated over 1300 km. This study provided
interesting insights on the lithospheric
structures associated with this major wrench
fault: 1) The fault is associated with a crustalscale resistive zone which is coincident with a
magnetic low, 2) The resistivity structure in
the lower crust to upper mantle is
approximately 2D with a geoelectric strike
N60°E parallel to the large-scale trend of the
GSLSZ, and 3) There is a close parallelism
between the orientation the fast split shear
wave polarization plane and the geoelectric
strike retrieved from long-period MT
measurements. This similarity of seismic and
electric conductivity anisotropies suggests
that they both have an origin related to the
wrench fault fabric of the lithospheric mantle
beneath the GSLSZ.
Transpressional orogenic domains
Often, orogenic domains as a whole
have been submitted to a transpressional
deformation characterized by the association
of thrusting normal to the belt and lateral
escape accommodated by transcurrent faulting
parallel to the belt. Recently, Meissner et al.
(2002) using Pn anisotropy measurements
have shown that in such domains the
uppermost (sub-Moho) mantle is
characterized by a fast propagation of Pwaves parallel to the trend of the belt,
pointing to a flow fabric in the uppermost
mantle dominated by the lateral escape of
lithospheric blocks. Shear-wave splitting
measurements in active and fossil orogenic
areas also record orogen-parallel flow
directions in the upper mantle (e.g., Savage,
1999; Silver et al., 1999; Vauchez & Nicolas,
1991). Fast shear waves are polarized parallel
to the trend of the transpressional belts, even
in domains where crustal deformation is
essentially accommodated by thrusting, and
delay times frequently ≥ 1s indicate that this
"wrench fault type" flow fabric affects the
entire lithospheric mantle.
Figure 13: Comparison of magnetic field data, MT highconductivity strikes for the period band of 20-500s, and SKS
fast directions for the Great Slave shear zone (Wu et al.
2002). H and L refer to magnetic highs and lows,
respectively. Delay times between the arrivals of the two
split SKS waves are of 1.1-1.5 s.
10
Taiwan is currently deforming in
response to the oblique convergence between
the Philippines and the Eurasian plates. As a
result, the crust displays evidence of a
transpressive deformation and strain is
partitioned between thrusting and wrench
faulting normal and parallel to the belt,
respectively (e.g., Huang et al., 2000;
Lallemand et al., 2001). Shear-wave splitting
measurements by Rau et al. (2000) display
nevertheless a coherent pattern over the entire
Taiwan Island (Figure 14). S-waves generated
in the Benioff zone by local earthquakes that
probe the mantle above the subduction zone,
are split. The fast shear wave is polarized
parallel to the tectonic grain and delay times
are up to 2s. These observations suggest that
the upper mantle beneath Taiwan has a
homogeneous transcurrent /transpressional
fabric due to northward tectonic escape, i.e., a
transport direction parallel to the active
orogen.
The Neoproterozoic Ribeira orogenic
belt of Southeastern Brazil formed during the
final amalgamation of Gondwana between
580 and 540 Ma (Egydio-Silva et al., 2002).
The southern and central domains of the belt
were subjected to an oblique convergence
between the South American and African
protocontinents (Figure 15a). This resulted in
development of numerous dextral wrench
faults, hundreds of kilometers long and up to
10 kilometers wide, oriented parallel or
slightly oblique to the belt. In the central
domain, the current level of erosion (17-20
km) shows mylonites that formed at high
temperature (T>800°C) and continued to
deform during a slow cooling down to ca.
740°C. Southward, the erosion level is more
superficial and the shear zones are marked by
mylonites formed under amphibolite facies
metamorphic conditions (Vauchez et al.,
1994). The wrench faults reworked a slightly
older low-angle foliation due to thrusting
toward the South American protocontinent.
During the late orogenic stages, both orogennormal thrusting and orogen-parallel wrench
faulting occurred. As a whole, the southern-
Figure 14: Deep structure beneath the active Taiwan
orogen. A: simplified map showing the geodynamic
situation of the Taiwan orogen (from Lallemand et al.,
2001). B : Shear-wave splitting measurements (Rau et al.,
2000) using S waves from local earthquakes and
teleseismic ScS.
central Ribeira belt represents a
transpressional orogenic segment about
100km wide and almost 1000 km long
(Trompette, 1994). Shear-wave splitting
measurements performed over the southern
branch of the Ribeira belt (Heintz et al., 2000)
have yielded a coherent pattern characterized
by a polarization of the fast S-wave in a
direction parallel to the orogenic grain (Figure
15b), suggesting that the bulk volume of
lithosphere in the transpressional domain has
a "wrench fault-type fabric". Larger delay
times between the fast and slow shear waves
arrivals (up to 2.5s) have usually been
retrieved from data recorded above or close to
the main shear zones, suggesting that strain
was not homogeneously accommodated but
was somewhat localized in the main shear
zones.
11
The Pyrenees in Western Europe
(Figure 16) formed during the Mesozoic due
to displacement of Iberia relative to Eurasia.
This motion, generated by the opening of the
Atlantic ocean between North America and
Iberia, was mainly accommodated along the
North Pyrenean fault (e.g., Choukroune,
1992). At first, the deformation regime was
transtensive and several pull-apart basins
formed. Then, during the final stages of the
evolution it became transpressive and finally
compressive. Indeed, the North Pyrenean
fault, i.e., the rupture between Iberia and
Eurasia, reactivated an older, pervasive
transpressive fabric formed during the late
stages of the Hercynian orogeny (e.g.,
Bouchez & Gleizes, 1995; Vauchez &
Barruol, 1996). Shear-wave splitting
measurements performed across the Pyrenees
and adjacent areas revealed a very consistent
pattern of anisotropy (Barruol et al., 1998).
The fast shear wave polarization plane is
usually oriented parallel to the belt, and the
delay between the fast and slow S-wave
Figure 16: Shear-wave splitting in the Pyrenees and adjacent areas. (a) Sketch map of the main hercynian structural directions
in the Pyrenees and adjacent regions. NPF is for the North Pyrenean Fault and SASZ for the South Armorican Shear Zone (see
Figure 12). The relative position of Iberia relative to Europa is the current position. (b) Shear-wave splitting measurements in
the Pyrenees (Barruol et al., 1998). At each location, the size of the circle is proportional to the delay time that is usually >1s
and the line indicates the polarization of the fast split shear wave.
12
arrivals is larger than 1s, even beyond the
Mesozoic Pyrenees belt. Pn anisotropy
measurements (Judenherc et al., 1999) are in
good agreement with S-wave splitting
measurements; the fast propagation direction
o f Pn is also parallel to the
Hercynian/Pyrenean tectonic fabric,
suggesting that the entire lithosphere beneath
the probed area has a coherent "wrench fault
type" fabric.
The analysis of the seismic anisotropy
data for the active orogen of Taiwan, the
Neoproterozoic Ribeira belt and the
Hercynian/Alpine Pyrenean belt leads to
similar conclusions. S-waves splitting results
are consistent with seismic anisotropy models
in which the lithospheric mantle deforms by
homogeneous transpression, instead of the
partitioned mode displayed by the crust
(Tommasi et al., 1999). However, the tectonic
fabric of the mantle does not correspond to
the classical transpression as defined by
Sanderson and Marchini (1984, i.e., with a
vertical stretching), but rather to lengtheningthinning shear (i.e., plane transpression,
Tikoff & Fossen, 1999; Tommasi et al.,
1999). This deformation regime involves
simultaneous shortening normal and
stretching parallel to the trend of the belt and
results in a lateral escape of the lithospheric
mantle. This may explain why observation of
a seismic anisotropy coherent with orogennormal thrusting at the scale of the lithosphere
is so scarce (e.g., Silver, 1996).
Lithospheric wrench faults: Thermomechanical effects
The various examples presented above
converge toward a model of major wrench
faults deeply rooted into the upper mantle.
Especially seismic tomography and shearwave splitting observations support that the
fault fabric affects the entire lithosphere
thickness. The width of the domain presenting
a "wrench fault-type" fabric likely ranges
between several tens of kilometers for a single
fault to several hundreds of kilometers for a
transpressional domain involving various
transcurrent and thrust faults. Moreover,
seismic anisotropy observations using longperiod data such as SKS waves imply that the
olivine lattice preferred orientation associated
with this "wrench fault-type" fabric,
characterized by horizontal [100] axes and
vertical (010) planes, both parallel to the fault
trace, is coherent at scales ≥ 50 km.
On the other hand, the olivine crystal
does not only display an anisotropic elasticity,
which leads to the observed seismic
anisotropy. The plastic deformation and
thermal diffusivities of olivine also are highly
anisotropic (Bai et al., 1991; Chai et al., 1996;
Durham & Goetze, 1977; Kobayashi, 1974).
Thus if major wrench faults are characterized
by a coherent olivine lattice preferred
orientation that affects the entire lithosphere
over domains several hundreds (or thousands
in the case of a transpressional belt) of
kilometers long and tens (or hundreds) of
kilometers wide, these domains might also be
the source of a large-scale mechanical and
thermal anisotropy within the continental
lithosphere that may influence the thermomechanical behavior of the plate during
subsequent tectonic events.
Strain-induced mechanical anisotropy of the
continental lithosphere
Experimental deformation of olivine single
crystals under different orientations relative to
its crystallographic lattice shows that olivine
has only three independent slip systems and
that these systems display significantly
different strength or critical resolved shear
stress (CRSS) values (Bai et al., 1991;
Durham & Goetze, 1977). Under hightemperature conditions, the (010)[100] slip
system displays the lowest critical resolved
shear stress; this means that, compared to the
other possible slip systems, for a given stress
it is able to accommodate the largest slip rate,
or, conversely, that it requires the lowest
built-up resolved shear stress to accommodate
a given strain rate. In other words, for
deformation in the dislocation creep regime,
which is expected to prevail in the
lithospheric mantle in active areas, olivine
displays an anisotropic viscosity.
In a lithospheric wrench fault, the
weakest (010)[100] slip system is oriented
parallel to the fault, i.e. the olivine crystals are
preferentially oriented with the (010) plane
13
subvertical and the [100] axis horizontal,
parallel to the shear direction. The question is
whether the anisotropic mechanical behavior
of the olivine single crystal combined with
such a LPO coherent over large scales in the
lithospheric mantle may result, at the scale of
the lithospheric mantle, in an anisotropy of
viscosity large enough to influence the
deformation of the lithosphere during
subsequent tectonic solicitations.
Tommasi and Vauchez (2001) used a
polycrystal plasticity model to investigate the
effect of a pervasive "wrench fault-type"
fabric frozen in the lithospheric mantle on the
continental breakup process. In this work, the
deformation of an anisotropic continental
lithosphere in response to an axi-symmetric
tensional stress field produced by an
upwelling mantle plume was evaluated by
calculating the deformation of textured
olivine polycrystals representative of the
lithospheric mantle at different positions
above a plume head (Figure 17). These
models show that an LPO-induced
mechanical anisotropy of the lithospheric
mantle may result in directional softening,
leading to heterogeneous deformation.
Reactivation of the inherited crystallographic
fabric, which is favored by tensional stresses
oblique to its trend, is characterized by higher
strain rates than other deformation regimes.
The reactivation of the pre-existing
fabric also results in higher strain rates than
those accommodated by an isotropic mantle
in similar conditions. During continental
rifting, this mechanical anisotropy may thus
induce strain localization in domains where
extensional stress is oblique (30-60°) to the
pre-existing mantle fabric. The directional
softening associated with olivine LPO frozen
in the lithospheric mantle may also guide the
propagation of the initial instability that will
follow the pre-existing structural trend. The
inherited mantle fabric also controls the
deformation regime, imposing a strong strikeslip shear component to the deformation. An
LPO-induced mechanical anisotropy may
therefore explain both the systematic
reactivation of ancient collisional belts during
rifting (structural inheritance) and the onset of
transtension within continental rifts.
Figure 17. Predicted deformation of a lithosphere
displaying a wrench fault-type fabric above a mantle plume
(Tommasi & Vauchez, 2001). (a) Strain rate (Von Mises
equivalent strain rate, normalized relative to the isotropic
behavior) as a function of the orientation of the radial
tensional stress relative to the [100] axis maximum of the
preexisting LPO for points above the plume head periphery
for three models with different initial LPOs. (b) Normal
and shear components of the strain rate tensor (normalized
by the Von Mises equivalent strain rate displayed by an
isotropic polycrystal) for the model in which the initial
LPO is the model aggregate. The reference frame is
defined relative to the preexisting mantle fabric: X is
parallel to the [100] axis maximum, i.e., parallel to the
preexisting structural trend, Y is normal to the preexisting
shear plane, and Z is vertical. Positive normal strain rates
denote extension and negative ones, shortening. Gray
region marks orientations that may trigger strain
localization
14
Figure 18 . Compressional deformation of a lithosphere
displaying a wrench fault-type fabric. Calculated strain rates
(Von Mises equivalent strain rate, normalized relative to the
isotropic behavior) are displayed as a function of the orientation
of the imposed shortening relative to the (010) plane maximum of
the preexisting LPO.
These results, obtained for a specific
geodynamic case, can be extended to a more
general situation. In major strike-slip faults
and transcurrent/transpressional orogenic
domains, the inherited fabric of the
lithospheric mantle should induce a
directional softening, with the consequence
that this fabric should be preferentially
reactivated. Development of new structures
oblique to the preexisting shear zones should
only be observed when the new tectonic
solicitations (either distensive or compressive,
Figure 18) are normal or parallel to the
inherited foliation, i.e., when no shear stresses
are applied parallel to the inherited fabric. In
most cases, reactivation will occur through
transtension or transpression, and the relative
proportion of simple and pure shear depends
on the obliquity of the stress axes relatively to
the inherited fabric.
The crustal fabric in lithospheric-scale
shear zones also contributes to this
mechanical anisotropy. Indeed, localized
deformation in the middle and lower crust
gives rise to strong LPOs. Crustal minerals,
in particular micas that are important phases
in mylonites, display a still stronger
mechanical anisotropy than olivine; their
layered structure results in plastic
deformation accommodated by glide on the
(001) plane only. In addition, strength
variations in polymineralic crustal rocks
often gives rise to a mm- to cm-scale
compositional layering parallel to the shear
zone that, at a larger scale, also contributes to
a directional weakening and reactivation of
the shear zone. Finally, grain-size reduction
associated with shearing in the upper/middle
crust may result in an isotropic strainsoftening within the shear zone; at these
depths, the shear zone will thus act as a
planar weak heterogeneity localizing the
subsequent deformation.
Repeated reactivations of major
transcurrent shear zones or domains during
long periods of time and the necessity for the
cause of this persistence to be in the
lithospheric mantle have been recognized
long ago (e.g., Watterson, 1975). Many
examples of such reactivation in various
geodynamic environments are available in the
literature. Tommasi and Vauchez (2001) have
already discussed those related to the
reactivation of lithospheric-scale shear zone
or transpressional belts during continental
rifting. So we will focus on one of the best
illustrations of the reactivation of a collisional
wrench fault as a transform boundary: the
development of the Newfoundland-AzoresGibraltar transform plate boundary at the
northern edge of the central Atlantic Ocean
during the Early Mesozoic (Figure 19). The
Newfoundland-Azores-Gibraltar fault zone
formed a major
15
Figure 19: Fit of the Central and North Atlantic ocean
showing that the initial rift in the central domain
propagated parallel to the Hercynian orogen and that the
New Foundland-Açores-Gibraltar hercynian wrench fault
was reactivated in the Mesozoic as a transform fault
transferring extension from the Central Atlantic basin to
the Thetys basin.
Hercynian dextral strike-slip fault
zone that offsets the Appalachians orogenic
front in Newfoundland (Keppie, 1989).
During the final stages of the AppalachianVariscan convergence, this fault
accommodated the relative displacement
between the Iberian and North African blocks.
This fault subsequently played a major role on
the Central Atlantic initial rifting, limiting one
of the promontories of the North American
stable margin. Indeed, the opening of the
central Atlantic Ocean took place almost
simultaneously from Florida to the
Newfoundland-Azores-Gibraltar transform
(the first Central Atlantic magnetic anomaly,
M25 is identified along this entire segment
(Owen, 1983)), but further northward
propagation of the Central Atlantic leading to
separation between Eurasia and North
America did not occur until Late Cretaceous
time. From mid-Jurassic to Late Cretaceous
time, the Newfoundland-Azores-Gibraltar
transform connected the Central Atlantic and
the Thetys oceanic basins, accommodating
the differential motion between Africa and
Europe.
Thermal conductivity anisotropy
Heat transfer is a key process
controlling the Earth's dynamics, since
temperature is a major parameter controlling
the rheological behavior of both crustal and
mantle rocks. Thermal conductivity in both
mantle and crust is usually assumed to be
isotropic. Yet, experimental data shows that,
at ambient conditions, the dominant mineral
phases in the crust and upper mantle display a
large anisotropy of thermal diffusivity. In
olivine, for instance, heat conduction parallel
to the [100] crystallographic axis is 1.5 times
faster than parallel to the [010] axis (Chai et
al., 1996). Quartz and micas, the main
constituents of crustal mylonites also display
a strongly anisotropic thermal conductivity,
with the highest and lowest conductivities
parallel to the [001] axis and within the (001)
plane, respectively (Clauser & Huenges,
1995).
This thermal anisotropy is also observed
at the rock scale. Recent studies combining
petrophysical modeling and thermal
diffusivity measurements on upper mantle
rocks (Tommasi et al., 2001) show that a
deformation-induced olivine LPO may result
in a significant thermal diffusivity anisotropy
in the uppermost mantle: heat transport
parallel to the olivine [100] axes
concentration (flow direction) is up to 30%
faster than normal to the flow plane ([010]
concentration). Moreover, in the studied
temperature range (300 to 1250 °K), the
thermal diffusivity anisotropy does not
depend on temperature, suggesting it might be
preserved even at higher temperatures
corresponding to asthenospheric conditions.
Seismic anisotropy data, like those presented
in the previous sections, indicate that major
wrench faults are characterized by a coherent
olivine lattice preferred orientation that
affects the entire lithosphere over domains
several hundred (or thousand in the case of a
transpressional belt) of kilometers long and
tens (or hundreds) of kilometers wide. This
"wrench fault type" fabric should therefore
induce a large-scale thermal diffusivity
anisotropy in the lithospheric mantle,
characterized by faster heat conduction within
16
the shear zone parallel to the shear direction
and slower conduction normal to the shear
zone.
A similar thermal anisotropy should be
present in the crustal section of a lithospheric
shear zone. Laboratory measurements of
thermal conductivity of gneisses drilled in the
KTB borehole show up to 40% of anisotropy
(Buntebarth, 1991). In these samples, which
display mineralogical compositions (quartz,
micas, and feldspars) and microstructures
similar to those of high-temperature mylonites
in the Borborema, Ribeira, and Madagascar
shear zones, heat conduction parallel to the
foliation plane is on average 1.2 times faster
than normal to it. A weaker anisotropy is
observed within the foliation plane, with the
highest conductivity measured parallel to the
lineation. Comparison between measured
thermal conductivities and those predicted by
petrophysical modeling suggests that,
similarly to the mechanical anisotropy, the
major contributions to the gneisses thermal
conductivity anisotropy stems from the strong
LPO of micas and quartz (Siegesmund, 1994).
Existence of a large-scale, straininduced thermal anisotropy in the upper
mantle implies that the temperature
distribution, rheology, and, hence, the upper
mantle dynamics depend on its deformation
history. Olivine orientations frozen in the
continental lithosphere may modify plumelithosphere interactions for instance.
Enhanced thermal diffusivity along
lithospheric-scale wrench zones, i.e., parallel
to the olivine [100] preferred orientation, may
lead to anisotropic heating of the lithosphere
above a mantle plume, favoring the
reactivation of these structures during
continental break-up (Tommasi & Vauchez,
2001; Vauchez et al., 1999; Vauchez et al.,
1998). Such a control of the pre-existing
lithospheric structure on the propagation of a
thermal anomaly may be inferred, for
instance, from tomographic images of the
East African rift in Kenya (Achauer & krispgroup, 1994). In these images, the lowvelocity seismic anomalies display two main
trends: a N-S trend, parallel to the surface
expression of the East African rift, and a NWSE trend following Neoproterozoic structures
that were reactivated during the Mesozoic to
give rise to the Anza rift.
Conclusion
Geological
and
geophysical
observations in active and fossil orogenic
belts converge to support that major wrench
faults are rooted into the upper mantle. Huge
transcurrent shear zones (several hundreds of
kilometers long and a few tens of kilometers
wide) in Brazil and Madagascar have been
eroded down to levels where deformation was
accommodated under high temperature
conditions (650°C to >800°C) in partially
melted rocks. It is remarkable that under these
high-temperature and, hence, low-viscosity
conditions, which were highly favorable to
development of a decoupling level, no
evidence of rooting of these shear zones has
been observed; on contrary, strain was still
localized in wide transcurrent shear zones.
Seismic profiling, seismic tomography, Pn
azimuthal anisotropy and magnetotelluric
soundings also support that several major
wrench faults crosscut the Moho discontinuity
and penetrate the uppermost mantle. In
addition, shear-wave splitting measurements
and electric conductivity anisotropy above
major strike-slip faults are in agreement with
a "wrench fault type" mantle fabric coherent
across most or even the totality of the
lithosphere thickness. Indeed, transform fault
boundaries as the San Andreas Fault, for
which a connection with the mantle is
required, display geophysical characteristics
similar to those of the main intracontinental
faults, either active or fossil. A similar
conclusion is reached for transpressional
orogenic domains deforming in response to
oblique convergence/collision.
The existence of a "wrench fault-type"
fabric into the continental mantle, besides
inducing anisotropic elastic and electrical
properties, may result in the development of a
directional softening and an anisotropic
conduction of heat in the continental mantle.
These anisotropic properties probably
influence the large-scale tectonic behavior of
the continents. Reactivation of the inherited
mantle fabric represents in most cases the
most economic behavior in terms of energy.
17
Only in very specific situations (solicitation
orthogonal or parallel to the ancient fabric),
the pre-existing fabric of the lithospheric
mantle will not be reactivated. Preferential
propagation of continental break-up parallel
to ancient orogenic belts as well as the
systematic reactivation of major wrench faults
likely result from both a directional softening
and an anisotropic heat transfer due to
wrench-type olivine-preferred orientations
frozen in the continental mantle.
Finally, the work by Pollitz et al.
(Pollitz et al., 2000; Pollitz et al., 2001) that
suggests that the mantle beneath active
wrench faults deform coherently with the
crust and, in some way, determines the
interseismic characteristics of the fault raises
the question of the effect of the mechanical
anisotropy of the lithospheric mantle on the
dynamics of active faults. Characteristics of
the fault like the slip rate, the stress building
rate and therefore the magnitude and the
recurrence of earthquakes could be affected
by a lower stiffness of the mantle in a specific
direction.
Acknowledgements: J. M. Lardeaux and
J. E. Martelat provided the map and images of
the Madagascar shear zones and M. Granet
the seismological results on the Armorican
massif. We thank C. Teyssier and L. Burlini
for constructive reviews.
References
Achauer, U. & krisp-group (1994) New ideas on the
Kenya rift based on the inversion of the combined
dataset of the 1985 and 1989/90 seismic
tomography experiments. Tectonophysics, 236,
305-329.
Bai, Q., Mackwell, S.J., & Kohlstedt, D.L. (1991)
High-temperature creep of olivine single crystals. 1.
Mechanical results for buffered samples. Journal of
Geophysical Research, 96, 2441-2463.
Barruol, G., Helffrich, G., Russo, R., & Vauchez, A.
(1997a) Shear wave splitting around the northern
Atlantic: frozen Pangean lithospheric anisotropy?
Tectonophysics, 279, 135-148.
Barruol, G., Silver, P.G., & Vauchez, A. (1997b)
Seismic anisotropy in the eastern United States:
Deep structure of a complex continental plate.
Journal of Geophysical Research - Solid Earth, 102,
8329-8348.
Barruol, G., Souriau, A., Vauchez, A., Diaz, J., Gallart,
J., Tubia, J., & Cuevas, J. (1998) Lithospheric
anisotropy beneath the Pyrenees from shear wave
splitting. Journal of Geophysical Research, 103,
30039-30054.
Berthé, D., Choukroune, P., & Jegouzo, P. (1979)
Orthogneiss, mylonite and non coaxial deformation
of granites: the example of the South Armorican
Shear Zone. Journal of Structural Geology, 1, 3142.
Bouchez, J.L. & Gleizes, G. (1995) Two-stage
deformation of the Mont-Louis-Andorra granite
pluton (Variscan Pyrenees) inferred from magnetic
susceptibility anisotropy. Journal of the Geological
Society of London, 152, 669-679.
Buntebarth, G. (1991) Thermal properties of KTBOberpfalz VB core samples at elevated temperature
and pressure. Scientific Drilling, 2, 73-80.
Chai, M., Brown, J.M., & Slutsky, L.J. (1996) Thermal
diffusivity of mantle minerals. Physics and
Chemistry of Minerals, 23, 470-475.
Choukroune, P. (1992) Tectonic evolution of the
Pyrenees. Annual Review of Earth and Planetary
Sciences, 20, 143-158.
Clauser, C. & Huenges, E. (1995). Thermal
conductivity of rocks and minerals. In Rock physics
& phase relations: A handbook of physical
constants (ed T. Ahrens), pp. 105-126. American
Geophysical Union, Washington.
Diaconescu, C., Knapp, J.H., Brown, L.D., & Steer,
D.N. (1997) Moho faults. In American Geophysical
Union Fall meeting (ed A.G. Union), Vol. 78, pp.
F724. EOS, Transactions, San Francisco.
Durham, W.B. & Goetze, G. (1977) Plastic flow of
oriented single crystals of olivine. 1. Mechanical
data. Journal of Geophysical Research, 82, 57375753.
Egydio-Silva, M., Vauchez, A., Bascou, J., & Hippert,
J. (in-press) High temperature deformation in the
neoproterozoic transpressional Ribeira belt,
Southeast Brazil). Tectonophysics.
Granet, M., Judenherc, S., & Souriau, A. (2000) Des
images du systeme lithosphere-asthenosphere sous
la France et leurs implications geodynamiques;
l'apport de la tomographie telesismique et de
l'anisotropie sismique. Bulletin de la Societe
Geologique de France, 171, 149-167.
Hanmer, S., Bowring, S.A., Breemen, O.v., and
Parrish, R.R., 1992, Great Slave Lake shear zone,
NW Canada; mylonitic record of early Proterozoic
continental convergence, collision and indentation:
Journal of Structural Geology, v. 14, p. 757-773.
Hartog, R. & Schwartz, S.Y. (2001) Depth-dependent
mantle anisotropy below the San Andreas fault
system: Apparent splitting parameters and
waveforms. Geophysical Research Letters, 106,
4155-4167.
Heintz, M., Debayle, E., Vauchez, A., & Assumpção,
M. (2000) Seismic anisotropy and surface wave
tomography of South America. EOS Transactions
of American Geophysical Union, 81.
18
Helffrich, G. (1995) Lithospheric deformation inferred
from teleseismic shear wave splitting observations
in the United Kingdom. Journal of Geophysical
Research, 100, 18195-18204.
Herquel, G., Tapponnier, P., Wittlinger, G., Mei, J., &
Danian, S. (1999) Teleseismic shear wave splitting
and lithospheric anisotropy beneath and across the
Altyn Tagh Fault. Geophysical Research Letters,
26, 3225-3228.
Huang, C.-Y., Yuan, P.B., Lin, C.-W., Wang, T.K., &
Chang, C.-P. (2000) Geodynamics processes of
Taiwan arc-continent collision and comparison with
analogs in Timor, Papua New Guinea, Urals and
Corsica. Tectonophysics, 325, 1-21.
Jegouzo (1980) The South Armorican shear zone.
Journal of Structural Geology, 2, 39-47.
Ji, S., Rondenay, S., Mareschal, M., & Senechal, G.
(1996) Obliquity between seismic and electrical
anisotropies as a potential indicator of movement
sense for ductile shear zones in the upper mantle.
Geology, 24, 1033-1036.
Judenherc, S. (2000) Etude et caractérisation des
structures hercyniennes à partir de données
sismologiques : le cas du Massif Armoricain. Thèse
d'Université, Université Louis Pasteur - Strasbourg
I, Strasbourg.
Judenherc, S., Granet, M., & Boumbar, N. (1999) Twodimensional anisotropic tomography of lithosphere
beneath France using regional arrival times. Journal
of Geophysical Research, 104, 13201-13215.
Judenherc, S., Granet, M., Brun, J.P., & Poupinet, G.
(in press) La collision hercynienne dans le Massif
armoricain: mise en évidence d'une différentiation
lithosphérique à partir de la tomographie et de
l'anisotropie sismique. Bulletin de la Société
géologique de France.
Keppie, J.D. (1989). Northern Appalachian terranes
and their accretionary history. In Terranes in the
Circum-Atlantic Paleozoic orogens (ed R.D.
Dallmeyer), Vol. Geological Society of America
Special Paper 230, pp. 159-192.
Klemperer, S.L. & Hobbs, R. (1991) The BIRPS Atlas.
Deep seismic reflection profiles around the British
Isles Cambridge University Press, New York.
Klemperer, S.L., Ryan, P.D., & Snyder, D.B. (1991) A
deep seismic reflection transect across the Irish
Caledonides. Journal of the Geological Society of
London, 148, 149-164.
Klosko, E., Wu, F., Anderson, H., Eberhart-Phillips,
D., McElvilly, T.V., Audoine, E., Savage, M.K., &
Gledhill, K.R. (1999) Upper mantle anisotropy in
the New Zealand region. Geophysical Research
Letters, 26, 1497-1500.
Kobayashi, Y. (1974) Anisotropy of thermal diffusivity
in olivine, pyroxene, and dunite. Journal of Physics
of the Earth, 22, 359-373.
Lallemand, S., Liu, C.-S., Angelier, J., & Tsai, Y.B.
(2001) Active subduction and collision in Southeast
Asia. Tectonophysics, 333, 1-7.
Mackwell, S.J. & Kohlstedt, D.L. (1990) Diffusion of
hydrogen in olivine; implications for water in the
mantle. Journal of Geophysical Research, 95, 50795088.
Mainprice, D. & Silver, P.G. (1993) Interpretation of
SKS-waves using samples from the subcontinental
lithosphere. Physics of the Earth and Planetary
Interiors, 78, 257-280.
Mareschal, M., Kellet, R.L., Kurtz, R.D., Ludden, J.N.,
Ji, S., & Bailey, R.C. (1995) Archean cratonic
roots, mantle shear zones and deep electrical
anisotropy. Nature, 375, 134-137.
Martelat, J.-E., Lardeaux, J.-M., Nicollet, C., &
Rakotondrazafy, R. (2000) Strain pattern and late
Precambrian deformation history in southern
Madagascar. Precambrian Research, 102, 1-20.
McBride, J.H. (1995) Does the Great Glen fault really
disrupt Moho and upper mantle structure?
Tectonics, 14, 422-434.
McGeary, S. (1989) Reflection seismic evidence for a
Moho offset beneath the Walls boundary strike-slip
fault. Journal of the Geological Society of London,
146, 261-269.
McNamara, D., Owens, T.J., Silver, P.G., & Wu, F.T.
(1994) Shear wave anisotropy beneath the Tibetan
Plateau. Journal of Geophysical Research, 99,
13655-13665.
Meissner, R., Mooney, W.D., & Artemieva, I. (2002)
Seismic anisotropy and mantle creep in young
orogen. Geophysical Journal International, 149, 114.
Molnar, P. (1988) Continental tectonics in the
aftermath of plate tectonics. Nature, 335, 131-137.
Neves, S.P., Vauchez, A., & Feraud, G. (2000)
Tectono-thermal evolution, magma emplacement,
and shear zone development in the Caruaru area
(Borborema Province, NE Brazil). Precambrian
Research, 99, 1-32.
Nicolas, A. & Christensen, N.I. (1987). Formation of
anisotropy in upper mantle peridotites - A review.
In Composition, structure and dynamics of the
lithosphere-asthenosphere system (eds K. Fuchs &
C. Froidevaux), Vol. 16, pp. 111-123. Am.
Geophys. Un., Washington, D.C.
Owen, H.G. (1983) Atlas of continental displacement:
200 million years to the present Cambridge
University Press, Cambridge.
Ozalaybey, S. & Savage, M.K. (1995) Shear wave
splitting beneath the western United States in
relation to plate tectonics. Journal of Geophysical
Research, 100, 18135-18149.
Pham, V.N., Boyer, D., Nguyen, V., & Nguyen, T.K.T.
(1995) Propriétés électriques et structure profonde
de la zone de faille du Fleuve Rouge au Nord ViêtNam d'après les résultats de sondage magnétotellurique. Comptes Rendus de l'Académie des
Sciences de Paris, 320, 181-187.
Pili, E., Ricard, Y., Lardeaux, J.-M., & Sheppard, S.
(1997a) Lithospheric shear zones and mantle-crust
connections. Tectonophysics, 280, 15-29.
Pili, E., Sheppard, S., Lardeaux, J.-M., Martelat, J.E.,
& Nicollet, C. (1997b) Fluid flow versus scale of
shear zones in the lower crust and the granulite
paradox. Geology, 25, 15-18.
19
Pollitz, F.F., Peltzer, G., & Bürgmann, R. (2000)
Mobility of continental mantle: Evidence from
postseismic geodetic observations following the
1992 Landers earthquake. Journal of Geophysical
Research, 105, 8035-8054.
Pollitz, F.F., Wicks, C., & Thatcher, W. (2001) Mantle
Flow Beneath a Continental Strike-Slip Fault:
Postseismic Deformation After the 1999 Hector
Mine Earthquake. Science, 293, 1814-1818.
Ranalli, G. & Murphy, D.C. (1987) Rheological
stratification of the lithosphere. Tectonophysics,
132, 281-295.
Rau, R.-J., Liang, W.-T., Kao, H., & Huang, B.-S.
(2000) Shear wave anisotropy beneath the Taiwan
Orogen. Earth and Planetary Science Letters, 177,
177-192.
Russo, R.M., Silver, P.G., Franke, M., Ambeh, W.B.,
& James, D.E. (1996) Shear wave splitting in
northeast Venezuela, Trinidad, and the eastern
Caribbean. Physics of the Earth and Planetary
Interiors, 95, 251-275.
Sanderson, D. & Marchini, R.D. (1984) Transpression.
Journal of Structural Geology, 6, 449-458.
Savage, M.K. (1999) Seismic anisotropy and mantle
deformation: What have we learned from shear
wave splitting? Reviews of Geophysics, 37, 65-106.
Siegesmund, S. (1994) Modelling the thermal
conductivity observed in paragneisses of the KTB
pilot hole. Scientific Drilling, 4, 207-213.
Silver, P.G. (1996) Seismic anisotropy beneath the
continents: Probing the depths of geoloy. Annual
Review of Earth and Planetary Sciences, 24, 385432.
Silver, P.G., Mainprice, D., Ben Ismail, W., Tommasi,
A., & Barruol, G. (1999). Mantle Structural
Geology from Seismic Anisotropy. In Mantle
Petrology : Field observations and high pressure
experimentation: A tribute to Francis R. (Joe) Boyd
(eds Y. Fei, C.M. Bertka & B.O. Mysen), Vol.
Special Publication, 6, pp. 79-103. The
Geochemical Society, Washington.
Silver, P.G. & Savage, M.K. (1994) The interpretation
of shear wave splitting parameters in the presence
of two anisotropic layers. Geophysical Journal
International, 119, 949-963.
Simpson, F. (2001) Resistance to mantle flow inferred
form the electromagnetic strike of the Australian
upper mantle. Nature, 412, 632-634.
Stewart, M., Strachan, R.A., & Holdsworth, R.E.
(1999) Structure and early kinematic history of the
Great Glen Fault Zone, Scotland. Tectonics, 18,
326-342.
Tapponnier, P., Lacassin, R., Leloup, P.H., Schärer, U.,
Zhong, D., W., W.H., Liu, X.H., Ji, X.C., Zhang,
L.S., & Zhong, J.Y. (1990) The Ailao Shan/Red
River metamorphic belt: Tertiary left-lateral shear
between Indochina and South China. Nature, 343,
431-437.
Tapponnier, P., Peltzer, G., & Armijo, R. (1986). On
the mechanics of the collision between India and
Asia. In Collision Tectonics (eds M.P. Coward &
A.C. Ries), Vol. 19, pp. 115-157. Geological
Society of London.
Tikoff, B. & Fossen, H. (1999) Three-dimensional
reference deformations and strain facies. Journal of
Structural Geology, 21, 1497-1512.
Tommasi, A., Gibert, B., Seipold, U., & Mainprice, D.
(2001) Anisotropy of thermal diffusivity in the
upper mantle. Nature, in press.
Tommasi, A., Tikoff, B., & Vauchez, A. (1999) Upper
mantle tectonics: Three-dimensional deformation,
olivine crystallographic fabrics and seismic
properties. Earth and Planetary Science Letters,
168, 173-186.
Tommasi, A. & Vauchez, A. (2001) Continental rifting
parallel to ancient collisional belts: An effect of the
mechanical anisotropy of the lithospheric mantle.
Earth and Planetary Science Letters, 185, 199-210.
Trompette, R. (1994) Geology of western Gondwana
(2000 - 500 Ma), Rotterdam.
Vauchez, A. & Barruol, G. (1996) Shear waves
splitting in the Appalachians and the Pyrenees:
Importance of the inherited tectonic fabric of the
lithosphere. Physics of the Earth and Planetary
Interiors, 95, 127-138.
Vauchez, A., Barruol, G., & Nicolas, A. (1999)
Comment on "SKS splitting beneath rift zones".
Journal of Geophysical Research, 104, 1078710790.
Vauchez, A. & Nicolas, A. (1991) Mountain building:
Strike-parallel displacements and mantle
anisotropy. Tectonophysics, 185, 183-201.
Vauchez, A., Pacheco-Neves, S., Caby, R., Corsini, M.,
Egydio-Silva, M., Arthaud, M., & Amaro, V.
(1995) The Borborema shear zone system.
Journal of South American Earth Sciences, 8,
247-266.
Vauchez, A., Tommasi, A., & Barruol, G. (1997)
Rheological heterogeneity and mechanical
anisotropy of the continental lithosphere. In AGU
fall meeting, Vol. 78, pp. 680. American
Geophysical Union, San Francisco, Ca.
Vauchez, A., Tommasi, A., & Barruol, G. (1998)
Rheological heterogeneity, mechanical anisotropy,
and tectonics of the continental lithosphere.
Tectonophysics, 296, 61-86.
Vauchez, A., Tommasi, A., & Egydio-Silva, M. (1994)
Self-indentation of continental lithosphere.
Geology, 22, 967-970.
Wannamaker, P.E., Chave, A.D., Booker, J.R., Jones,
A.G., Filloux, J.H., Ogawa, Y., Unsworth, M.,
Tarits, P., & Evans, R. (1996) Magnetotelluric
experiment probes deep physical state of
southeastern U.S. EOS, 77, 329, 333.
Watterson, J. (1975) Mechanism for the persistence of
tectonic lineaments. Nature, 253, 520-522.
Wittlinger, G., Masson, F., Poupinet, G., Tapponnier,
P., Mei, J., Herquel, G., Guilbert, J., Achauer, U.,
Guanqi, X., Danian, S., & Team, L.K. (1996)
Seismic tomography of northern Tibet and Kunlun:
Evidence for crustal blocks and mantle velocity
contrasts. Earth and Planetary Science Letters, 139,
263-279.
20
Wittlinger, G., Tapponnier, P., Poupinet, G., Mei, J.,
Danian, S., Herquel, G., & Masson, F. (1998)
Tomographic evidence for localized lithospheric
shear along the Altyn Tagh Fault. Science, 282, 7476.
Wu, X., Ferguson, I.J., & Jones, A.G. (2002)
Magnetotelluric response and geoelectric structure
of the Great Slave Lake shear zone. Earth and
Planetary Science Letters, 196, 35-50.
21