Numerical Modeling of Fault Zone

PUBLICATIONS
Geochemistry, Geophysics, Geosystems

RESEARCH ARTICLE Development of branching brittle and ductile shear zones:
10.1002/2016GC006793
A numerical study
Key Points: Sven Erik Meyer1,2 , Boris J. P. Kaus1,2 , and Cees Passchier1,2
Strain weakening in ductile and
1
brittle shear localization affects shear Institute of Earth Sciences, Johannes Gutenberg University Mainz, Mainz, Germany, 2Center for Computational Sciences,
zone interactions and geometries University Mainz, Mainz, Germany
Temporal evolution of shear zone
branches changes the stress field of
the surrounding rocks
2-D viscoelastic-plastic numerical
Abstract Continental collision zones are usually associated with large-scale strike-slip shear zones. In
simulations reproduce observed most cases, these shear zones are complex and consist of multiple strands, varying in width, length, and
anastomosing strike-slip shear zone total displacement. Here we present 2-D numerical models to simulate the formation of such shear zones at
geometries
different depth levels within the crust, under either brittle (frictional/plastic) or ductile conditions. Localiza-
tion of shear zones is initiated by a material contrast (heterogeneity) of the material parameters. We system-
Supporting Information:
Supporting Information S1 atically test the rate of strain weakening in brittle and in ductile regimes to understand its influence on the
Table S1 development of shear zone networks. Our simulations suggest that the development of antithetic faults in a
Figure S1
brittle shear zone system is closely linked to a decrease in the angle of friction during deformation. In gen-
Figure S2
Figure S3 eral, variation of the strain weakening also has a significant influence on ductile shear zones. Numerical
Figure S4 results show that the geometry and thickness of the localized high strain zone are especially affected by
Figure S5
weakening mechanisms during deformation. Furthermore, the interconnection and interaction of the shear
Figure S6
strands lead to a more complex kinematic pattern, which lead to a local change in the maximum principal
stress axis. These interaction of shear strands may explain the occurrence of shear-related structures (e.g.,
Correspondence to:
S. E. Meyer, folds) or differing characteristics of shear zones, such as the thickness of shear zones or the orientation of
meyersv@uni-mainz.de the faults to the stress field, which are consistent with field observations.
Citation:
Meyer, S. E., B. J. P. Kaus, and
C. Passchier (2017), Development of
branching brittle and ductile shear
zones: A numerical study, Geochem. 1. Introduction
Geophys. Geosyst., 18, 2054–2075,
doi:10.1002/2016GC006793. Strike-slip shear zones, such as the San Andreas, North-Anatolian, and Dead Sea Faults, are an important
subject of study in the Earth sciences since they play an essential role in accommodating the Earth’s plate
Received 28 DEC 2016 tectonic displacement pattern and are locations of major earthquakes. Such crustal-scale shear zones con-
Accepted 28 MAR 2017 sist of segments with dominant brittle or ductile deformation, mainly depending on the depth within the
Accepted article online 22 MAY 2017
crust at which segments operate [e.g., Holdsworth, 2004]. Field studies on crustal-scale structures show that
Published online 14 JUN 2017
the brittle or cataclastic network of shear zones in the upper crust is connected to ductile mylonitic shear
zones in deeper parts of the crust [Sibson, 1977, 1987]. The surface trace of the San Andreas Fault (SAF) in
the upper crust is an example of shear zones developed in the brittle regime with a complex network of
right-lateral and left-lateral strike-slip shear strands [Sylvester, 1988], while the corresponding shear zone in
the lower crust is characterized by ductile deformation, as shown by seismic anisotropy [Bokelmann and Ber-
oza, 2000]. The brittle segments are mostly restricted to the upper crust and can be studied in active strike-
slip shear zone systems. Deeper equivalents of shear zone systems can be studied where they have been
uplifted along strands of active fault systems, or more commonly as fossil crustal-scale shear zones, known
as mobile belts (e.g., Arabian shield) [Stern and Johnson, 2010], which are particularly common in Precam-
brian terrains.
In addition to the vertical transitions from brittle to ductile deformation, shear zones have a complex 3-D
geometry. Detailed and quantitative studies have usually been limited to simple, straight planar shear zone
segments [Passchier, 1990a, 1990b, 1998; Grujic and Mancktelow, 1998; Piazolo et al., 2001; Schrank et al.,
2008], whereas field work shows that in most cases the straight segments form part of a complex systems
of kilometers-wide anastomosing and branching (The term ‘‘branch’’ is used for an active shear zone, which
is connected through joining or splitting with an additional localized shear zone.) shear zones [Passchier
C 2017. American Geophysical Union.
V et al., 1997; Faure et al., 2005; Archanjo, 2008; Carreras et al., 2010; Culshaw et al., 2011; Passchier and Platt,
All Rights Reserved. 2016]. Such anastomosing shear zone patterns can develop in different ways. Relatively subparallel shear
MEYER ET AL. DEVELOPMENT OF BRANCHING SHEAR ZONES 2054

Geochemistry, Geophysics, Geosystems 10.1002/2016GC006793
strands may connect to each other by small variations in orientation determined by a pre-existing structure
[Carreras, 2001; Mancktelow and Pennacchioni, 2005; Ponce et al., 2013]. An anastomosing or branching char-
acter of shear zones can also be caused by heterogeneities of the parent rock due to pre-existing structures
(e.g., layering, foliation, fractures) or variations in the strain intensity [e.g., Bell, 1986; Fusseis et al., 2006; Pen-
nacchioni and Mancktelow, 2007]. Alternatively, different shear sets may develop successively, which start to
join and form a more complex anastomosing pattern [e.g., Marquer et al., 1996; Arbaret et al., 2000; da Sil-
veira et al., 2009].
Experimental studies of shear zone networks have focused mainly on kinematics and strain distribution
[e.g., Cobbold et al., 1971; Harris and Cobbold, 1985; Dennis and Secor, 1987; Williams and Price, 1990; Her-
wegh and Handy, 1996; Huddleston, 1999] and on lithospheric-scale extensional or transform faults or on the
behavior of single ductile and brittle shear zones [Huismans and Beaumont; 2003; Hieronymus, 2004; Man-
cktelow, 2006; Choi et al., 2008; Gerya, 2010a, 2012, 2013a, 2013b]. However, little is known about how the
shear zones interact in a network. Individual shear strands can either be active at the same time as other
shear strands, or can be active after each other, whereas each deformation event will result in complex flow
patterns which will leave distinct traces in the rock. It is likely that the effective strength of rocks is reduced
with increasing strain, either by increased fluid pressures, mineral exchange and the formation of weak
layers (filled with clay minerals or mica), crystallographic preferred orientation or grain-size reduction [Sib-
son, 1990; Morrow et al., 2000; Connolly and Podladchikov, 2000; Bos and Spiers, 2002; Platt and Behr, 2011a,
2011b; Thielmann et al., 2015]. Strain weakening of either frictional plastic or viscous material properties was
shown to have a significant effect on the localization of shear zones [e.g., Frederiksen and Braun, 2001; Man-
cktelow, 2002, 2006; Huismans and Beaumont, 2007; Vitale and Mazzoli, 2008]. On a crustal scale, symmetric
extensional rifts form for small to moderate amounts of strain weakening, whereas asymmetric rift systems
form for fast and significant weakening [Huismans and Beaumont, 2003; Buiter et al., 2006].
Fully 3-D models of strike-slip faults show similarities in the development of simple network shear zones to
the field example of the San Andreas Fault, and results have confirmed the important role of strain weaken-
ing as a process during the development of a shear zone [Popov et al., 2012]. Nevertheless, those models do
not yield the complex networks of shear zones found in the field, possibly due to the limited numerical res-
olution of the 3-D models. No work has been carried out to date on the interaction of individual shear zone
strands in a strike-slip regime, how these form a larger branching shearing system and what the effect of
strain weakening is on the geometry of branching shear zones.
In this paper, we present 2-D forward numerical models to systematically study the influence of different
rock rheological regimes on the development of interconnecting strike-slip shear zones at different crustal
depths. We use large-scale models to compare the geometries of shear zone networks to classical field
examples. Our numerical modeling simulates shear zones at different crustal levels, which will help to
understand the effects of changes in material rheology on the development of individual shear zones and
their behavior during interconnection of shear strands.
2. Modeling Approach (Methodology)

The numerical model is based on a visco-elasto-plastic code named MILAMIN_VEP [Kaus, 2010; Thielmann
and Kaus, 2012], which solves the incompressible Stokes equations for visco-elasto-plastic rocks in combina-
tion with the energy conservation equation (see Thielmann and Kaus [2012] for a more extensive descrip-
tion). For the upper crust, we employ a Drucker-Prager yield criterion, which is equivalent to nonassociated
Mohr-Coulomb plasticity in 2-D for incompressible plane strain deformation, with the general equation for
shear stress (s):
s5C1 l rn
l5tan ð/Þ (1)
where C is the cohesion, l is the coefficient of internal friction, rn is the normal stress, and / is the angle of
internal friction [e.g., Vermeer and De Borst, 1984]. The second invariant of the deviatoric stress tensor (rII ) is
defined as:

rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi
2
rII 5 0:5 rxx 1ryy 12rxy 2 2 (2)
where rxx ; ryy are the normal stress in x and y directions and rxy is the shear stress. Shear zones are inter-
preted as a weaker zone in comparison to the surrounding undeformed rock [Chery et al., 2001; Provost and
Houston, 2003; Carpenter et al., 2011]. The strain weakening in a brittle-plastic regime is controlled by fric-
tional processes in the Mohr-Coulomb plasticity [Huismans and Beaumont, 2007]. We use the friction angle
and the cohesion to control the weakening of the material during ongoing deformation in our numerical
models [Poliakov and Buck, 1998; Lavier et al., 2000]. At yielding, the stress state is such that it touches the
yield criteria which is:

ry 520:5 rxx 1ryy sin ð/Þ1C cos ð/Þ (3)
Brittle failure occurs where the plastic yield stress (ry ) is reached by the second invariant of the deviatoric
stress tensor (rII ):
rII 5 ry (4)
F5rII 2ry (5)
where F is the yield function with the term F50 for brittle deformation. The initial rheological properties of
the matrix are / 5 308 and C 5 35 MPa. We apply strain weakening in which both parameters are linearly
reduced with increasing strain (c) until a given critical strain (ccrit Þ, after which full weakening is obtained
(see Figure 1b for sketch), according to
8 c
< /5/0 2ð/0 2/min Þ if c < ccrit
ccrit (6)
:
/5/min if c ccrit
The critical strain is poorly constrained, which is why we vary it on our models between 0.01 (rapid weaken-
ing) and 2 (slow weakening).
Our numerical code assumes a Maxwell visco-elasto-plastic rheology:
1 1 Dsij _ @Q
e_ ij 5_e viscous
ij 1_e elastic
ij 1_e plastic
ij 5 sij 1 1k (7)
2geff 2G Dt @rij
where e_ ij is the deviatoric strain rate, G is the elastic shear modulus, sij is the deviatoric stress, k_ is the plastic
Ds
multiplier, Q is the plastic flow potential, Dtij denotes the Jaumann rotation rate of the stress tensor [e.g.,
Thielmann et al., 2015], and geff is the effective viscosity and can be described by power-law (dislocation)
creep:
ðn121Þ
e_ II 1 1
geff 5g0 exp k 2 (8)
e_ 0 T T0
where g0 denotes the specified material viscosity at the initial stage for T0 (supporting information Table
S1), n is the stress exponent, e_ II is the second invariant of the strain rate tensor (defined in a similar way to
equation (2)), e_ 0 is a reference strain rate, k is the Arrhenius type temperature dependency of viscosity, T is
the temperature, and T0 is a reference temperature (in Kelvin).
Natural ductile shear zones are characterized by a fine-grained matrix, due to crystal-plastic processes and
dynamic recrystallization at high strain and elevated temperatures. These processes causing grain size
reduction result in a weakening of the material within the shear zone in comparison to the surrounding
host rock [e.g., Thielmann et al., 2015]. We implement this weakening effect in simulations occurring in the
ductile regime by adding a prefactor A to the viscosity formulation
gdef 5A g0 (9)
which is weakened in a linear fashion as a function of strain, similar to how the plastic material properties
are weakened (Figure 1c). We use a reference strain weakening (see supporting information Table S1) for
the numerical simulations presented in section 4.2, whereas section 4.2.3 discusses the effect of varying the

Figure 1. (a) Setup of the large-scale models at different crustal levels, simulated by a bulk simple shear, (b) plastic (brittle) weakening: the
friction angle and cohesion decrease linearly with progressive strain until a critical strain c starting from an initial friction angle of 308.
Cohesion decreases from 35 to 1 MPa, (c) in viscous (ductile) weakening simulations, the viscosity decreases with progressive strain by a
prefactor A starting at 1.
prefactor and critical strain on the development of the shear zone networks. In our simulations, strain weak-
ening is implemented either in the brittle or ductile regimes, depending on the crustal depth at which the
simulations are performed.
2.1. Model Setup

The models presented in this study represent large-scale horizontal sections of the crust, 300 km long and
150 km wide at different crustal levels, simulating brittle faults in the upper crust (5–12 km depth) and duc-
tile shear zones in the lower crust (20–25 km depth) (Figure 1).
We implement two different rock properties in the models: an upper crust with Mohr Coulomb plasticity to
simulate brittle faults and a lower crust with viscous power-law rheology to simulate ductile shear zones
[Burg and Schmalholz, 2008; supporting information Table S1]. To initiate the shear zones, we introduce
round heterogeneities located in the center of the model, which correspond to experimental data for gra-
nitic compositions [Tirel et al., 2008; supporting information Table S1]. The effect of the shape, the size, and
the number of initial heterogeneities on the results will be discussed in section 4.1.1.
The velocity boundary conditions for the models are as follows (Figure 1a): The side boundaries are periodic,
the bottom is fixed, and the top boundary has a prescribed horizontal velocity such that the overall strain
rate (_e BG ) is constant throughout the simulation. This generates a dextral shear flow as seen from above
(simple shear), where the maximum compressive stress is oriented at an angle of 458 to the imposed shear
direction (Figure 1a) and parallel to the plane of the model. The model represents a horizontal segment of
the crust with a uniform reference temperature (upper crust 250–3008C, lower crust 500–7508C), and a

reference lithostatic pressure, based on an average crustal density of 2700 kg m23, of 265 MPa for 10 km
crustal depth and 662 MPa for 25 km crustal depth, which is added to the dynamic pressure of the model.
The boundary condition for the temperature consists of heat flow through the sides (periodic boundary),
while the top and bottom are isothermal. The heat source is given by the initial temperature and during the
simulation by shear heating (Hs Þ, which transforms the viscous and plastic strain into thermal energy,

Hs 5v sij e_ viscous
ij 1_e plastic
ij (10)
where sij is the deviatoric stress, v is the efficiency of shear heating (51 here), and e_ ij is the deviatoric strain
rate. Shear heating is present in all brittle and ductile shear zone simulations as an additional weakening
parameter [Regenauer-Lieb et al., 2001; Thielmann and Kaus, 2012]. As in other geodynamic codes, we
employ a lower and upper viscosity cutoff of 1018 and 1026 Pa s, to avoid rounding errors in the direct solv-
ers used to solve the corresponding system of linear equations.
The numerical Lagrangian mesh consists of about 1.4 million quadrilateral finite elements with linear shape
functions for velocity, temperature, and discontinuous linear function for pressure (Q1P0 elements). During
the simulation, remeshing is applied if elements are too distorted and we use particles to track material
properties and stresses.
In the model runs presented in this paper, we change the rheological parameters in equations (4–9) to
understand how these affect the development of brittle or ductile shear zones, and to determine if the
parameters have an influence on the geometry of complex branching strike-slip shear zones.
3. Results
3.1. Brittle Shear Zones—Reference Model
For sufficiently large stresses, the upper crust is brittle, which can be represented by Mohr-Coulomb plastic-
ity. In the first simulation (reference model), we use standard rock properties (supporting information Table
S1) for the upper crust to illustrate the temporal development of brittle shear zones. We employ an overall
strain rate of 10214 s21 with a background lithostatic pressure of 264:9 MPa and temperature of 3008C,
which represent the pressure-temperature conditions for a horizontal model in the upper crust at 10 km
depth. Under these conditions, the effective viscosity of the matrix is initially 1.3 3 1021 Pa s for the matrix
and 7.5 3 1020 Pa s for the inclusions. Stresses are initially zero but build up visco-elastically until they
reach the yield stress, after which plastic (brittle) failure occurs, initially around the two imposed weaker het-
erogeneities. Strain weakening of the friction angle and cohesion is activated until a critical strain of 0.6 is
reached, after which full weakening is obtained (Figures 1b and 1c). Throughout this paper, orientation of
shear zones is given with respect to the long axis of the model, which is also the imposed shear direction.
The top of the model, where the maximum compressive stress is oriented at an angle of 458, moves to the
right in a dextral shear sense, Figure 1a. In all simulations (we performed over 100 in total), the first brittle
shear zones grow at an angle of 5–88 with a dextral shear sense (Figure 2a): such shear zones are described
in the literature as synthetic Riedel fractures (R). These so-called main faults (R-faults, Figure 2b) spread to
the sides of the model and join at the center at 0.11 Ma, leaving the tips of the shear zones unaffected. This
leads to a complex geometry of strands in the form of ‘‘bridge structures’’ (Figure 2b). In a later stage, two
new shear zone types develop that have a different shear sense, namely one set of synthetic faults with an
orientation of 166–1808 (P-faults, Figure 2b) and a second set of antithetic faults at 20–758 (R0 -faults, Figure
2c). The shear zones in our simulation have orientations that are typical for Riedel structures as described
by Cloos [1928], Riedel [1929], and Ahlgren [2001], except for the R0 -faults which have a larger variation in ori-
entation. In the early model stages, the R0 -faults grew at an angle between 208 and 508. After 13% strain,
new R0 -faults developed at a higher angle (between 70–758) (Figure 2c). The numerical models show that
the shear strands stay active, since they display continuous slip motion after the formation of the branched
structure.
The interconnection of brittle shear zones into a network has a significant influence on the total pressure
distribution within the model, where total pressure is a sum of the lithostatic (264:9 MPa) and tectonic pres-
sure. A plot of the total pressure (Figure 2a) shows that during early stages, the shear zones show a
decrease of pressure relative to the surrounding matrix, which is a common feature of shear bands in a

Figure 2. Model development for a case with reference parameters in the brittle regime (supporting information Table S1) and two anomalies in the center with an overall strain rate of
e_ BG 510214 s21 , showing (a) the development of the main faults in the strain rate field; black rounded spots represent the initial position of the anomaly, (b) the interaction of the main
faults and weak growth of synthetic faults, (c) the development of a complex shear zone system with antithetic/synthetic faults; Figures 2a–2c show in addition the evolution of the total
pressure (in MPa) at the same time steps. In Figures 2a–2c, vertical white and red lines are originally vertical passive strain markers that show deformation in the model.
pressure-dependent rheology (Mohr-Coulomb plasticity) [Mancktelow, 2006]. The center of the model is
dominated by a low-pressure zone (100 MPa) located in the bridge between the tips (Figure 2b). With
increasing strain the total pressure distribution changes, whereby the evolution of new shear strands within
the fault network play a significant role. The gentle bending trend of the newly developed synthetic and
antithetic faults shows high-pressure zones, which exceed the low-pressure zones by 400 MPa (Figure 2c).
The low-pressure zones are located primarily along the straight part of the shear strands. An additional
observation is that the shear zone branches that connect to the opposite side of the imposed shearing (Fig-
ure 2c, I and II), exhibit a zone of high pressure (3802500 MPa) between the shear strands (junction zone).
In summary, the results show that left and right-lateral brittle faults (R and R0 ) interconnect to form an inter-
connected network of shear zones, where shear strands in the branched system exhibit a continuous activ-
ity (local strain rate), i.e., they remain active. The interacting of the newly nucleated faults (e.g., R0 -fault) with
the existing shear zones lead to simultaneous active branched shear structures, which subsequently affects
the total pressure distribution of rocks in the model.
3.1.1. Influence of Heterogeneities and Material Properties on Evolution of Shear Zones
In a purely homogeneous model with identical rock parameters and rheology, no brittle shear zones form,
which shows that material heterogeneities are required to induce localized deformation. An interesting
question in this respect is, how strong the deviation of the rheological parameters between the anomaly
and the matrix should be to localize brittle shear zones. To address this, we performed several numerical
models that employ different specified material viscosities, friction angles, and cohesion of the anomaly.
Results suggest that parameters that cause a small difference of >0.1% in the effective viscosity between

the anomaly and matrix are sufficient to trigger localization of brittle shear zones. Models in which the
anomaly was not circular or in which the anomaly had a different size show that its shape does not play a
significant role in the initiation of brittle shear zones, provided that the anomaly is numerically resolved. We
therefore use a standard circular shape with 100 m radius. In models with only a single or more than two
anomalies, a network of R, R0 and P-faults formed similar to the reference model. The reasons for the similar
geometry of the shear zone network are (1) the periodic boundary conditions in case of a single propagated
shear zone and (2) the fact that the localized shear zones affect the surrounding rocks and therefore the
propagation of a nearby fault. If the anomalies are sufficiently close, one shear zone develops and the
neighboring fault remains inactive (see supporting information Figure S1). In summary, we conclude that
the number of anomalies does not play a significant role in the development of a different geometry of
shear zone networks. We therefore use two anomalies in all subsequent numerical simulations.
In addition, we performed several simulations in which we changed the creep-law parameters such that a
different effective viscosity of the matrix was obtained (see supporting information). In these simulations,
brittle faults occurred under standard conditions (supporting information Table S1) for specific material vis-
cosities between 1021 21022 Pa s. Furthermore, a reduction of 18 in the friction angle between the anomaly
and the matrix is sufficient to initiate heterogeneous behavior and shear zone nucleation.
3.1.2. Effect of Crustal Depth on Shear Zone Networks
Shear zones can form at different crustal depths and, with increasing depth, the rheology of the material
changes due to increasing temperature and pressure. Major shear zones can form crustal-scale structures
that pass through the ductile as well as the brittle regime of the crust. In our 2-D simulations, we simulate
the development of a strike-slip fault network at different crustal depths within the brittle regime to under-
stand the effect of the environment parameters.
We performed a total of 25 numerical models with reference properties (supporting information Table S1)
but with variations in the overall strain rate (_e BG ), lithostatic pressure (P), and temperature (T). The simula-
tions refer to the localization of brittle shear zones, and therefore involve a strain weakening by a reduction
of the friction angle and cohesion, whereas the mechanism of weakening in a ductile regime is not
involved. We varied the depth of the models from 2 to 14 km in 2 km steps under a constant overall strain
rate of e_ BG 510214 s21 (supporting information Table S1). The temperature change with depth is based on a
geotherm of T5273 K1ð25zÞ, where z is the depth in km. The results show that variations in lithostatic
pressure (with constant temperature) do not have a significant effect on the geometry of the shear zones,
but brittle failure occurs in the numerical models only down to 20 km depth for a lower thermal gradient.
However, temperatures over 3308C, which corresponds to a depth of ca. 13 km for our standard thermal
gradient (258C per km), result in an overall viscous behavior of the model and no brittle failure occurs. The
numerical observation can be explained by the following equation:
rxy 52 leff e_ BG (11)
In a system of constant overall strain rate, the shear stress is proportional to the effective viscosity of the
rocks (equations (8) and (11)). An increase in temperature reduces the effective viscosity and the shear
stress (rxy ) until it is less than the yield stress (ry Þ, which occurs when the temperature exceeds 3308C in
our model for a background strain rate of e_ BG 5 10214 s21 . This dependency can be explained with a
‘‘Christmas tree’’ strength profile, which shows that under low temperature and at shallow depth brittle fail-
ure is reached, whereas at higher temperature and depths, below the brittle-ductile transition, viscous flow
dominates. To summarize, the models show that variation of the lithostatic pressure and the temperature
has an indirect effect on localization of brittle shear zones by influencing rock viscosity. This leads to an
overall viscous deformation of the model instead of the localization of fault zones.
3.1.3. Effect of Strain Weakening on Brittle Shear Zone Networks
With ongoing deformation, grain-size reduction through damage results in a weaker behavior of cataclastic
zones. Previous studies have indicated the importance of strain weakening on plate tectonics [Gueydan
et al., 2014], which includes cohesion loss [Buck, 1993], frictional strength reduction [Bos and Spiers, 2002],
and fluid pressure variations [Sibson, 1990; Chester et al., 1993; Rice, 1992; Faulkner and Rutter, 2001; Carpen-
ter et al., 2011]. Strain weakening can be simulated by a strain rate [Behn et al., 2002] or strain dependent
weakening of material properties [Huismans and Beaumont, 2003]: while the former represents the fault
weakness on a seismic timescale, the latter does so on a geological timescale. Previous studies on

Figure 3. Effect of strain weakening on shear zone geometry. (a and b) Effect of a decrease in the cohesion. (c and d) Effect of decreasing the friction angle by various amounts with
ongoing critical strain.
lithospheric normal and transform faults have shown that strain weakening has a strong effect on the devel-
opment of shear zones [e.g., Lavier et al., 2000; Huismans and Beaumont, 2003; Hieronymus, 2004; Choi et al.,
2008; Gerya, 2010, 2012, 2013a, 2013b; Allken et al., 2011, 2012]. In contrast to the previous studies, our
numerical models represent a horizontal profile within the crust. Here we test the effects of a decrease of
the friction angle and cohesion in relation to variation of the critical strain (ccrit ) on the geometry of branch-
ing shear zones. We focused the first models on the friction angle and the cohesion with a constant critical
strain of 0.6 using the reference model setup (supporting information Table S1). Decreasing the cohesion
from an initial value of 35 MPa results in three types of behavior: (1) localization of parallel main faults for a
decrease in the cohesion to 20 MPa, (2) joining of the main faults for a value of 4 < C < 20 MPa, and (3)
localization of weak synthetic faults together with the main faults for a cohesion 3 MPa (Figures 3a and
3b). Simulations in which cohesion was kept constant but the friction angle / was decreased from 308 to
108 over a critical strain of 0.6 show that shear localization of the main faults is accompanied by synthetic
faults for a friction angle of 12 < / < 30 , but that antithetic faults developed only if the friction angle is
reduced to 12 (Figures 3c and 3d).
In further simulations, we varied the critical strain after which full weakening is achieved in models with a
simultaneous reduction of the friction angle (/) from 30 to 10 and cohesion (C) from 35 MPa to 1MPa. In
the first set of simulations, we used a critical strain of 0.01, which mimics a very rapid weakening of the
matrix. The weak rheological behavior of the material leads to a large number of brittle shear zones in a
short amount of time, which formed a complex network of main, synthetic and antithetic faults. Further-
more, the synthetic faults show a stronger variation in displacement along the individual strands and a
more curved geometry (Figure 4a). By increasing the critical strain to 0.6, the shear zone network geometry
changed and showed a stronger concentration of strain on a smaller number of shear strands. The main
faults are characterized by a thinner zone of deformation compared to the softer models (Figure 4b). The
connection of shear strands forms by an accumulation of simultaneously active antithetic and synthetic
faults over a longer period. Next, we employed a critical strain 2, which resulted in main faults that were
active for a longer period of time and in synthetic faults with weak antithetic shear zones (Figure 4c).
To clarify the difference in shear zone geometry, we computed the proportion of the activity (strain rate) on
synthetic and antithetic faults in the models. In order to quantify this in a systematic manner, we defined as

Figure 4. Weakening controlled by the reduction of friction angle and cohesion with ongoing strain. In all models, the friction angle and the cohesion are decreased from 308 to 108 and
35 to 1 MPa for different critical strains (a) rapid strain weakening with c50:01, (b) intermediate conditions with c50:6, (c) slow weakening with c52, (d) activity of synthetic and anti-
thetic faults under different critical strains.
a ‘‘shear zone’’ any zone where the strain rate is at least twice the overall (background) strain rate, while the
vorticity characterizes the shear sense of the shear zone. The results show that after 0.35 Ma the largest pro-
portion of antithetic faults experience a critical strain between 0.3 and 0.6, and that 22% of the active shear
zones in the model box have a sinistral shear sense (antithetic faults). A higher and lower values of the criti-
cal strain show a reduction of the proportion of antithetic faults (Figure 4d). The analysis shows that most
faults are synthetic, but also that the proportions of synthetic/antithetic faults are time dependent. The
activity of the antithetic faults is relatively large for small values of critical strain, with the greatest propor-
tion of antithetic faults developing for a critical strain of 0.3–0.6 and the proportion decreasing for larger
critical strains (Figure 4d).
In summary, strain weakening and the rate at which the frictional parameters decrease, has a significant
effect on the development of branched brittle faults in the models. Furthermore, the complexity of brittle
branching shear zones patterns is more dependent on the friction angle than on the cohesion.
3.2. Ductile Localized Shear Zones

Ductile shear zones were simulated in our models by using matrix properties for a lower crust [Burg and
Schmalholz, 2008] with a specified material viscosity of g0 53:051021 Pa s, a stress exponent of n53 and an
anomaly that mimics a granite rock type with a specified material viscosity of g0 57:491020 Pa s [Tirel
et al., 2008] and a stress exponent of n53:2. The models were performed at a depth of 20 km in the lower
crust with a lithostatic pressure of 530 MPa, a temperature of 5008C and an overall strain rate (_e BG ) of
10214 s21 , which results in stresses that are below the brittle yield stress at this depth. Strain weakening
was incorporated by linearly reducing the specified material viscosity by a prefactor 0.01, between strain 0
and 0.6 (Figure 1c), with identical anomaly locations as in the reference model (Figures 1a and 2). The first
ductile shear zone develops at a strain of 34.4% after 1.09 Ma along the two anomalies in the center of the
model. We defined as a ‘‘shear zone’’ any zone where the strain rate is at least three times the overall (back-
ground) strain rate. The shear zones are first oriented parallel to the imposed shear direction in a narrow
zone and fan out toward their termination (Figure 5a). After 1.29 and 1.54 Ma, the tip of the main shear
zones change its orientation by 258 to join the upper and lower shear strand (Figures 5b–5d) by forming an
anastomosing structure. In the interaction zone, the shear zones widen considerably. In general, the

Figure 5. (a–d) Temporal evolution of the standard model with ductile shear zones.
individual shear strands vary in thickness and the main faults show a dextral shear sense with some minor
sinistral (antithetic) shear zones (antithetic) at 1158 to the imposed shear direction. The shear strands differ
in their displacement and the interaction of the lower strand with the upper shear zones leads to a decrease
in the activity of the lower shear strand. Every strand remains active over the duration of the simulation.
3.2.1. Effect of Heterogeneous Material (Number of Anomalies)
The presence of heterogeneities is important in the development of localized ductile shear zones, and as in
the brittle simulations, a small variation in viscosity or in the weakening parameters is sufficient to produce
localized shear zones along the boundary of the anomaly and matrix. The results showed that the geometry
of the ductile shear zone strands are like the results of simulations with two anomalies, but differ only in the
number of shear strands (see supporting information). For consistency with the earlier brittle simulations,
we use two anomalies in the following simulations.
3.2.2. Internal and External Influence of Material Behavior
Our models of ductile shear zones are based on a power-law rheology, and as a consequence the shear
zones are zones with reduced effective viscosity. In general, rock types differ in their specific material viscos-
ity (g0 ) and stress exponent (n). To understand the effect of the stress exponent and the specified material
viscosity on the development and on the interaction of the shear zones, we performed simulations in which
the powerlaw exponent n was varied between 1 (linear viscous) and 6. The results show that for n < 3 the
shear zones are wider (9–20 km) and more regular in shape, whereas for n > 3, strain is concentrated in a
narrower zone and the main shear strand accumulates more displacement (see supporting information).
Similar results are described in Mancktelow [2002, 2006]. Simulations in which the viscosity prefactor g0 was
varied show that lower viscosities (1021 Pa s) lead to a weaker behavior, wider ductile shear strands and
wider connection zones in comparison to models with a higher initial viscosity (1022 Pa s) (see supporting
information).
3.2.3. Effect of Strain Weakening
Field observations and modeling studies suggest that shear zones are weaker in comparison to the sur-
rounding host rock. To study the effect of viscous strain weakening, we focused our simulations on under-
standing the relation of the rate and amount of effective viscosity reduction on the characteristics of ductile
shear zone networks. We performed a total of six models with a constant overall strain rate of 10214 s21 , in
which we varied both the maximum weakening prefactor between 1 and 0.1, and the critical strain after
which full weakening is achieved from 0.1 to 1.5. Simulations with a weakening prefactor larger than 0.55

Figure 6. Comparison of the models with a different prefactor and critical strain under constant overall strain rate of 10214 s21 for a time step 1.8–2 Ma; (a and b) varying of the prefactor
by constant critical strain 0.6, dispersed shear zones developed for a prefactor 1–0.08 where constricted shear zones occurred for a prefactor of 0.01; (c) critical strain ccrit 50:1 with a pre-
factor 1–0.01, development of a wide shear zone (dispersed) with undeformed blocks; (d) increase of the critical strain to ccrit 50:3 leads to narrower shear zones with similar geometries
of syn-(marked as i) or antithethic shear strands to the previous setup.
do not exhibit any ductile shear zones [Mancktelow, 2002, 2006], whereas a single shear zone develops for a
small period of time if the prefactor is between 0.55 and 0.25. A larger amount of weakening (0.25–0.12)
leads to a more stable activity of the shear strands with the result that the shear zones form a weak connec-
tion of the shear strands. The shear zones are more intensely localized close to the anomaly and have a
more disperse strain to the junction zone. With a critical value of 0:12, the shear zone strands show a con-
tinuous activity during the entire simulations. Characteristic of these shear zones is that at a distance from
the anomaly the strain fans out and affects a broader zone, whereas the width of the shear zone strands
and the junction zone varies between 3 and 30 km (Figure 6a). A further decrease of the prefactor to 0.03–
0.01 leads to shear strands with high strain with the width of the shear zone strands and junction zone vary-
ing between 3–10 km (Figure 6b), which we term ‘‘constricted’’ shear zones. Strain weakening of the mate-
rial has a significant influence on the characteristics of the development of shear zones as well as on the
width of the shear strands. A weak behavior of the material with a reduction of more than 1/100 of the spe-
cific material viscosity with ongoing strain leads to narrow high strain zones.
Next, we employed different critical strain values, using a reduction prefactor of 0.01 for a time-span of 1.8–
2 Ma (55–63% strain, Figures 6c and 6d). The results show that shear zones develop only for a critical strain
between 0.1 and 2, whereas the maximum limit can be extended for values of ccrit 2 at later model stages
(strain larger than 94%). The results show that the localized shear zones for a reduction prefactor of 0.01
and a critical strain field of 0.1–2 differ in the thickness of the strands, in the geometry, and in the intensity
of strain rate (Figures 6b–6d). The ductile shear zone for ccrit 50:120:3 is characterized by a wide junction
zone with a width of over 30 km and by development of internal synthetic faults between the main faults.
The new synthetic faults with an angle of 1608 to the imposed shear direction forms an anastomosing pat-
tern of shear zones (Figures 6c and 6d, i), whereas the material between the high strain zones stays unde-
formed (unsheared zones). An increase of the critical strain to 0:3 with the same prefactor led to a
stronger shear nucleation in a thinner zone and a thinner junction zone, with a width of 5–10 km (Figures
6b–6d). Furthermore, a lower number of shear zones nucleated in the model resulting in a simpler geome-
try through the absence of oblique trending synthetic high strain zones, whereas the shear motion is

Figure 7. Influence of strain weakening on ductile shear zones by decreasing the viscosity by a prefactor linked to a critical strain (ccrit )
under constant overall strain rate 10214 s21 , after a time step of 1.8–2 Ma. We define an inactive shear zone as a structure with strain rate
close to the overall strain rate of 10214 s21. The numbers showed next to the symbols on the plots indicate the end value of the
prefactor (A).
stronger along the constricted shear zones. An overview of the effect of strain weakening on shear zone
nucleation indicates that shear zones occur only with a higher prefactor >0.2 under a limited range if critical
strains of 0.6–0.8 (Figure 7). In summary, constricted shear zones occur in a limited range only and require a
critical strain value 0:6 and a reduction prefactor of 0:01.
Simulations that are performed for a longer time (3.18 Ma, with a strain of more than 100%), results in simi-
lar geometries, but differ in the width and the number of shear zones. The numerical runs for a critical strain
higher than 1.3 only show an active phase in the beginning of the simulations, as indicated by the slip motion
of the markers (Figure 8a). At a later stage, the shear zone strands are intermittently inactive, which means
that the strain rate along the propagating shear strand has nearly the same value as the overall strain rate
10214 s21. During advanced stages of the constricted models (10 km wide shear zones; critical strain 0.3–
0.8), new antithetic and synthetic shear zones develop with an angle of 1158–1408 and 458 between the
straight trending shear strands (Figure 8b), whereas the dispersed models (>11 km wide shear zones; critical
strain 0.1–0.3) show a similar trend during earlier stages. The dispersed localization of the shear strands lead
to a wide deformed zone, where the newly developed synthetic and antithetic high strain zones form an
interacting anastomosing structure (Figures 8c and 8d) within the wide shear zone. The junction zone has
both undeformed blocks (called ‘‘lozenges’’) and verging overturned folds, mainly between the lozenge struc-
tures. In summary, the long-term numerical models with a critical strain 0.1–0.3 (dispersed shear zones) indi-
cate well-developed lozenges objects and a higher number of large-scale folds in the junction zone. The
models simulating shear zones of constricted strain develop smaller-scale fold structures (Figure 8b) in the
thinly propagated high-strain areas, after a relatively large amount of strain (>150%).
4. Discussion
4.1. Effect of Numerical Parameters on Shear Zone Patterns
In all our simulations, shear heating was active and thus provides a weakening mechanism in addition to
the prescribed strain weakening. Results for a typical viscous localization simulations show that the increase

Figure 8. (a) Inactive shear zones where unsheared zones are surrounded by high-strain zones; (b) constricted shear zones with internal shear-related folds and initiation of synthetic
faults; (c and d) evolution of undeformed blocks and shear-related folds. Example of a dispersed shear zone with a different movement of undeformed blocks (A, B, C, and D). Black lines
represent the high strain zones within the shear zone to illustrate the anastomosing trend.
of temperature within the shear zone is only a few degrees, whereas the brittle localization has a shear zone
that is up to 908C warmer, which may act as an additional source of weakening [Brun and Cobbold, 1980;
Kameyama et al., 1999; Schmalholz and Duretz, 2015]. However, additional simulations without shear heating
give very similar results (supporting information Figure S2). We also tested the effect of numerical resolution
on model behavior and found that higher resolution simulations result in quite similar overall shear zone
geometries, although the shear zones themselves are somewhat narrower and local differences occur (sup-
porting information Figure S3). These results suggests that the main conclusions of our work are robust.
4.2. Evolution and Interaction of Faults (Upper Crust)

Most crustal-scale strike-slip shear zones exhibit a complex network of shear zone strands in the brittle (e.g.,
San Andreas Fault) [Wallace, 1990] and the ductile (e.g., Nadj shear zone) [Stern, 1994] regime. The rheologi-
cal behavior of the material is a key to better understanding the variations in geometry of such shear zones.
Examples of large-scale shear zones show a straight, as well as a curved trace of the fault at the Earth’s sur-
face. One well-exposed classical example of a branched network of brittle faults is the San Andreas Fault
(SAF) [Wallace, 1990]. The SAF represents a complex connected structure of fault strands, indicative of a var-
iation in the orientation of the shear planes relative to the stress field. It is still debated how such fault
strands interact and how they influence the propagation—or even initiation of—neighboring fault zones.
In our numerical 2-D models, we focused primarily on the geometry of brittle shear zones and the interac-
tion of the shear strands as a result of a decrease of either viscous or frictional strength with strain. The
results show that the development of antithetic faults (R0 ) depends on a decrease of the friction angle coef-
ficient and is not related to a loss of cohesion. Antithetic faults occur in our simulations due to a decrease of
the friction angle from an initial value of 308 to a value lower than 128 (friction coefficient <0.21). Further-
more, the R0 -faults appeared at a later geological time in models with decreasing friction angle (Figure 2). In
addition, we observed that the rate of decrease of frictional strength plays an important role in the activity
(strain rate) of antithetic faults (Figure 4d). The models exhibit the highest amount of shear activity on anti-
thetic faults for a critical strain of 0.3–0.6. In comparison to natural observations, the geometry of the final
mature fault system with right-lateral and left-lateral faults shows similarities to the SAF structures as

Figure 9. (a) Network of anastomosing ductile shear zones—Qazaz, Ajjaj, Hanabiq, and Hamadat shear zones, Najd shear zone system in Saudi Arabia; (b) Riedel shear zones in foliated
cataclasite, G€
uney Detachment, Aydin, Turkey; (c) 10–20 km wide shear zones with F3 folds in quartzo-feldspathic granulites, southern Madagascar [Martelat et al., 1999; de Wit et al.,
2001]; (d) shear-related folds of quartz veins in a graphitic marble, Brandberg-West-Fm, Goantagab, Namibia.
described by Wallace [1990] or to the Anatolian Fault system [Barka and Kadinsky-Cade, 1988]. Previous
work on the SAF and on foliated cataclasite confirmed the crucial role of weakening of fault zones, with
measurements of a low friction coefficient of 0.1–0.21 due to the presence of clay minerals or high fluid
pressure [Bird and Kong, 1994; Zoback, 2000; Collettini et al., 2009; Faulkner et al., 2010; Carpenter et al.,
2011], in the numerical model. Our results show that a low friction coefficient <0.21 is required to form
branched networks of fault zones, similar to those observed in the SAF system. They also highlight the
importance of strain weakening. In the simple linear weakening function, we employ in our simulations, a
critical strain of 0.3–0.6 gives results that are most similar to natural observations. The use of a linear strain
weakening function is likely an oversimplification of natural weakening processes, which may occur as the
result of many different processes, e.g., grain size reduction, thermal increase, each of which most likely has
its own dependence on strain.
In general, field studies [Davis et al., 1999; Katz et al., 2004], numerical models [McKinnon and Garrido de la
Barra, 1998], and analogue experiments [Naylor et al., 1986; Schreurs, 1994; Marques and Coelho, 2001;
Coelho et al., 2006] are consistent with Coulomb theory, where R and R0 are oriented at 458 6 //2 to the
maximum compressive stress (Figure 10d). A number of field studies on small-scale Riedel structures indi-
cate a variation in the orientation of R and R0 -faults, for example in the Jurassic Navajo Sandstone (Capitol
Reef National Park, Utah) [Katz et al., 2004] or in the foliated cataclasite, G€
uney Detachment (NW of Aydin,
Turkey) [Gessner et al., 2001] (Figure 9b). In our numerical model, the antithetic faults display a similar trend
of variation with angles between 208 and 758 to the imposed shear direction (Figure 10c). Our simulations
suggest that the development of R-faults in the early stages leads to a change in the maximum compressive
stress direction between the shear faults (junction zone), from an initial value of 458 to a final angle of 108
(Figures 10a and 10b). The low angle of the maximum compressive stress forms antithetic faults (R0 ) within
a 408 range (Figure 10d). The angle of the local maximum compressive stress direction is therefore not a
stable value. For complex branched shear zones, the angle varied locally within a range of 5–458, which
explains the curvilinear trend of the variably oriented antithetic faults between 208 and 758 (Figures 2c

Figure 10. Distribution of the maximum compressive stress orientation on localized brittle shear, ccrit 50:6. (a–b) The interaction of the shear zone strands leads to local variations in the
compressive stress direction between 108 and 458, time step 0.17–0.42 Ma. (c) Overview of shear localization, R0 -faults show a curvilinear trend and different angles to the imposed shear
direction. (d) Geometry of synthetic and antithetic faults in dextral simple shear, SZ is the imposed shear direction.
and 10c). A change of the maximum compressive stress in the evolution of the Riedel—shears may be a
suitable explanation for the variation of the antithetic faults in the small-scale field example of the Jurassic
Navajo Sandstone described by Katz et al. [2004] or for the structures of the G€ uney Detachment [Gessner
et al., 2001] (Figure 9c). In addition, the synthetic faults in the numerical model are affected in a similar man-
ner, with a few faults occurring as secondary structures parallel to the imposed shear direction (Figure 10c).
Furthermore, in our simulations, the fault zones that formed first and exhibit a 5–88 angle to the overall
shear direction are very similar to what is observed in nature [e.g., Mancktelow and Pennacchioni, 2005, Fig-
ure 5a].
Our modeling observations are similar to several natural field examples of strike-slip fault zones. The net-
work of the San Andreas, Cucamonga, Glen Helen, and San Jacinto Faults exhibits a similar orientation of
the constituent fault zones as in our simulations. The different angles of the faults are described as splay
faults [Scholz et al., 2010], with the primary fault rotated into a parallel position relative to the plate motion
vector. This led to a misaligned orientation of the fault plane, and therefore to new faults (secondary struc-
tures) developed at an ideal angle to the principal stress. We propose a further possible explanation, namely
that the evolution of primary fault structures and the interconnecting bridge structures can change the local
maximum principal stress direction, which lead to new synthetic faults with an orientation even parallel to
the bulk imposed shear direction (plate slip motion). The amount of slip displacement of the numerical
models differs from the natural examples, and one reason can be a different strain rate in comparison to
our model.
Our simulations suggest that the curvilinear trend of synthetic and antithetic faults also affects the distribu-
tion of total pressure. The numerical models show that during early stages, total pressure is typically
reduced along the shear zones compared to that of the host rock, as described by Mancktelow [2006, 2007].
With ongoing strain the distribution of pressure changes because of newly formed synthetic and antithetic
faults. The cross-linking shear faults are separated by a high-pressure zone, which is related to bending of
the shear zones, whereas the straighter trend of the faults hosts a low-pressure zone (Figure 2c). Fluid trav-
eling from the high-pressure to the low-pressure zone is likely to locally change the shear strength of the
shear zone by way of a weakening effect by mineral reactions and affect the pore pressure [Sibson, 1990;
Rice, 1992; Ridley, 1993; Connolly and Podladchikov, 2000; Jeffries et al., 2006; Schmalholz et al., 2014]. This
has potentially important implications for fluid flow and the development of ore deposits. A concentration

of fluids along fault planes can (1) increase fluid pressures and lead to further brittle fracturing in the sur-
rounded rocks (e.g., veins) and (2) cause hydrothermal fluids to react with the wall-rocks to produce hydro-
thermal ore deposits. Our simulations show that significant variations in pressure occur in fault zone
networks, which may be useful in identifying hydrothermal ore deposits in natural shear structures, provid-
ing that they formed under comparable conditions.
4.3. Numerical Simulations in Ductile Regimes (Lower Crust)

Many publications have established that strain weakening plays a significant role in localizing ductile shear
zones [e.g., White et al., 1980; Mancktelow, 2002, 2006; Huismans and Beaumont, 2003; Buiter et al., 2006;
Kaus and Podladchikov, 2006; Thielmann and Kaus, 2012; Gerya, 2013a, 2013b; Montesi, 2013]. Our numerical
simulations demonstrate that the rate of weakening within the shear zone has a significant influence on the
characteristics, the geometry, and on the evolution of shear-related structures of the shear zones.
4.3.1. Characteristics of Shear Zones
The numerical models show that the width of a propagating shear zone is controlled by certain values of
critical strain and reduction prefactors of viscosity. We have found that a sharp threshold between thin and
wide shear zones exists at specific values of prefactor 100 and a critical strain of 0.5 (Figures 6a and 6b).
In this study, we classify a narrow zone as less than 5 km wide and a wide zone as more than 6 km wide.
Field observations on the Neoproterozoic East African Orogen display a wide variation in the thickness of
individual shear strands within the shear zone network, e.g., the Ampahiny/Vorokafotra shear zone in Mada-
gascar with a thickness of 10–20 km [de Wit et al., 2001] and the localized thinner shear zones on the Najd
Fault system in Saudi Arabia [Stern and Johnson, 2010], which have a width between 0.6 and 5 km. As dem-
onstrated by our numerical models, the rate of weakening is one possible explanation for such distinct var-
iations in the thickness of shear zones. A further example of a branched network of variable-width ductile
shear zones is shown at the Northern Cap de Creus in Spain [Carreras, 2001]. In our simulations, the rate of
weakening is linear: in the case of a local heterogeneity of the material, a deviation in weakening from the
critical strain (0.5) by a prefactor of 100 can lead to a local change in width of the shear zone from a nar-
row to a wide shear zone. A strength drop of 80–99% along the shear zone to the protolith rock is seen in
the numerical models and should also be observed in the natural field example. Experimental studies on
quartzofeldspathic gneiss have shown a strength drop of 30–60% to the protolith rocks within the shear
zone [Holyoke and Tullis, 2006], and as much as 70% for mica-bearing weak layers [Park et al., 2006]. The dif-
ferent weakening values in our models compared to those of the experimental results in the literature may
be due to a difference in the geological timescale and the P-T conditions of our models compared to those
of the natural setting. The degree of dynamic recrystallization, which is related to temperature and strain
rate, is one factor involved in strain weakening. Field estimates suggest values of 77% or even higher weak-
ening within the crustal-scale shear zone of the Parry Sound protolith domain in the Greenville orogeny
[Gerbi et al., 2010]. Estimations of B€
urgmann and Dresen [2008] and Mehl and Hirth [2008] indicate strength
drops approximately equal to our numerical models, which shows that our values are at least consistent
with those determined from (small) experimental samples.
Microstructural observations of mylonitic shear zones are characterized by a grain-size reduction due to
dynamic recrystallization, a change in mineral content and the development of a preferred crystallographic ori-
entation. Experimental and natural field examples of deformed rocks (shear zones) have shown that several
dynamic recrystallization mechanisms are accompanied by dislocation creep [e.g., Tullis et al., 1973; White, 1976;
Drury and Urai, 1990; de Bresser et al., 2001]. For the higher temperature simulations (5008C), glide-controlled
and climb-controlled dislocation creep are active mechanisms in the minerals of quartz and feldspar. The stress
exponent has typical values between 3 and 5. Therefore, we used a stress exponent of n 5 3 for the matrix, in
accord with natural deformation by dislocation creep. Increasing the stress exponent in our numerical models
to a value of 6 promotes the concentration of high strain rate along the shear zone and a change in the number
of shear strands (supporting information Figure S4). This shows that a few parameters in the power law-creep
function also have influence on the characteristics of such branched networks of ductile shear zones.
4.3.2. Geometry of Ductile Shear Zones
As well as a change in the thickness of shear zones, the shear zone geometry is also affected by strain weak-
ening. As demonstrated in Figure 6, the simulated constricted shear zones in the models have a simple
structure of two parallel trending shear strands and the connection to a single shear zone (junction zone).
This geometry can be described as a dextral freeway triple junctions [Passchier and Platt, 2016], where both

Figure 11. Sketch summarizing our results on anastomosing shear zones formed in brittle and ductile rheologies; (a and b) development
of Riedel shear zones to a complex network of brittle faults, antithetic and synthetic faults indicating different directions of propagation;
(c and d) showing the interconnection of shear strands and the development of an anastomosing pattern of synthetic and antithetic faults
within the dispersed shear zones, which bounds undeformed blocks.
shear strands are simultaneously active in a dextral slip motion. Structures like this model are common in
orogenic collision zones—the Najd shear zone system in the Arabian Shield is one example of a complex
linkage of left-lateral and right-lateral crustal-scale strike-slip networks. A few of the shear zone geometries
from the field example in Figure 9a show similarities to the numerical models of the junction zone and the
connection of the shear strands. Therefore, the numerical models give us two important pieces of informa-
tion about such natural structures: (1) strain weakening during the deformation may attain similar values as
in our numerical models and (2) simultaneous activity of all shear strands in the shear zone network is possi-
ble. The theoretical model of the Qazaz shear zone [Meyer et al., 2014] assumes a simultaneous activity of
the shear strands, whereas it is not always clear in the field if the shear strands are impaired by their inter-
connection. The similarities in the geometry of our numerical models to such natural structures suggests
that the shear strands operated simultaneously in the natural shear zone networks. The detailed observa-
tion of the whole network system of the Najd shear zone (Figure 9a) brings to light a more complex geome-
try compared to that of the models, with a dominant NW-SE trend and a higher number of obliquely
oriented WNW-ESE and N-S trending shear zones. The N-S trending Hanabiq shear zone even indicates an
antithetic slip motion (dextral) and forms a branch with a synthetic shear zone (Hamadat). A similar struc-
ture is observed in numerical models that simulate the transition to dispersed shear zones. At a critical
strain between 0.3 and 0.1, the simulations show junction zones of antithetic and synthetic shear strands.
The structure is mirror-inverted compared to the field example (Figures 6c, 6d, 11c, and 11d). In general,
such dispersed shear zones differ not only in the thickness of their shear strands, but also display a higher
number of high-strain strands (Figure 6c). These are conspicuous in the Amphahiny shear zone in Madagas-
car (Figure 9c). The shear zone there bears similarities to the numerical model in terms of the thickness of
shear branches (10–20 km wide) and in the diagonally trending synthetic shear zones between the main
shear strands (Figures 6d and 9c). Local variations in the rates of strain weakening along the collision zone
induced by rheological heterogeneity of the material could be one explanation for the different geometry
of the network of the Najd-shear zone system and the Amphahiny/Vorokafotra shear zones in Madagascar.
In addition to the dispersed shear zones for critical strain between 0.1 and 0.3 and a prefactor of 0.01, the
newly interconnected shear strands also form an anastomosing network (Figures 6c and 11d). Studies with
bimineralic materials, reported by Jammes et al. [2015], show that anastomosing fabrics in shear zones are a
consequence of a high proportion of weak minerals in the rock. Our numerical models demonstrate that a
specific range in the rate of strain weakening can also lead to the formation of a wide single shear zone
with an anastomosing pattern (Figure 11d). In general, the change in the rock viscosity from the initial stage
to the final stage influences the physical characteristics of ductile shear zones.
4.3.3. Shear Related Structures
Our numerical models of dispersed shear zones also include zones of undeformed blocks (lozenges) and
shear-related fold structures, as shown in Figure 8c. The evolution of such structures is strongly related to
the disperse propagation of the ductile shear zones. The development of sigmoidal and rhomboidal

lozenges in the model indicate that they are not stationary objects within the shear zones. Moreover, most
of the undeformed blocks exhibit no change in shape or in volume during the deformation. Detailed analy-
sis of the internal kinematic flow shows that internal synthetic and antithetic high-strain zones form an
interacting network within the wide deformed zone (Figure 11d). These high strain zones, parallel, and
trending 1158 to 1408 to the imposed shear direction lead to a lateral variation in flow and to a more com-
plex kinematic pattern in the wide shear zone. Through the inhomogeneous flow, the undeformed blocks
(Figures 8c and 8d) move within the wide shear zones, surrounded by high-strain zones. The evolution of
the lozenges in our numerical models show strong visual similarities to the structures described by Pollard
and Aydin [1984], Pennacchioni [2005], and Mann [2007]. Natural field examples of lozenges in southern
Madagascar, southern Switzerland, and in north-eastern Spain (Figure 9c) [Simpson, 1982; Martelat et al.,
1999; Carreras, 2001; Ponce et al., 2013] exhibit a similar trend in the development of the undeformed blocks
in relation to the interaction of shear zones.
The heterogeneous main flow in the numerical models, described earlier, produces shear-related folds (Fig-
ures 8c and 8d) as a result of the flow perturbation. During the evolution of the folds an increase in geomet-
ric complexity occurs as deformation proceeds, whereby the structures that formed during early stages are
initially characterized as open folds with an axial surface at 908 to the shear direction. In later stages, how-
ever, the open folds start to tighten and their axial surfaces rotate parallel to the imposed shear direction
(Figures 8c and 8d). Local zones of high strain, mostly between the undeformed blocks, display the complex
geometry of verging overturned folds and several types of shear-related folds. The evolution described
above may explain the large-scale as well as small-scale folds observed in shear zones (Figures 9c and 9d), if
internal low-angle to high-angle high strain zones are present within the shear zones. Differences in wave-
length, thickness, and amplitude of the folds are influenced by changes in the internal flow perturbation,
whereas viscosity contrast plays no role in the simulations. Similar structures of shear-related late folds
[Howard, 1968; Bell, 1978; Platt, 1983; Mandal et al., 2004] were described by Carreras et al. [2005] in a classic
example from NE Spain. The development of these fold structures seems to be caused by local or general
perturbations of flow, related to rigid objects or rheological effects of different material properties between
layers [Cobbold and Quinquis, 1980; Lister and Williams, 1983; Platt, 1983; Bons and Jessell, 1998]. Studies on
shear-related late folds in Cadaques, NE Spain, have described folds in relation to a change in kinematics
[Carreras et al., 2005]. Our numerical models demonstrate that the evolution of internal time-dependent
synthetic and antithetic high-strain zones in combination with the main shear flow within the shear zone
can generate shear-related folds through a more complex kinematic pattern. In our models, these fold struc-
tures form passively and are not induced by a material contrast. If material contrasts had also been taken
into account, folding would most likely have been even more pronounced.
5. Conclusions
We performed 2-D numerical simulations to understand how branching shear zones form under simple
shear for either brittle or ductile rheologies. Our results can be summarized as follows:
1. The initiation and propagation of simulated brittle and ductile shear zones are strongly related to the
rate of strain weakening with ongoing deformation. A minor change in this parameter results in distinctly
different geometries or characteristics of the shear strands (e.g., thickness, presence of undeformed
blocks, etc.)
2. A low friction coefficient (<0.21) that is weakened within a critical strain of 0.6–0.8 during brittle defor-
mation is required to form antithetic faults in a shear zone network.
3. The time-dependent development of brittle shear strands changes the internal stress field of the sur-
rounding rocks, which promotes the formation of new shear strands with orientations that differ from
their predecessors.
4. The linkage and the simultaneous activity of brittle shear strands lead to an irregular distribution of the
total pressure, with zones of overpressure and underpressure developed along the shear zones and in
the wall rock between the strands.
5. In most simulations, ductile shear zones are relatively wide (‘‘dispersed’’). Yet narrow (or ‘‘constricted’’)
shear zones are observed if a large amount of weakening (prefactor of 0.01) is applied together with a
specific range of weakening rates (for critical strains of 0.6–0.8).

6. Dispersed shear zones exhibit a higher number of localized shear zones in comparison to narrow shear
zones. Furthermore, the interconnection and interaction of variously oriented synthetic and antithetic
shear strands led to an anastomosed fabric of the shear zone network.
7. The development of lozenges and shear-related fold structures in dispersed shear zones in the numerical
models results from simultaneously interaction of high-strain zones in the heterogeneous flow perturba-
tion. The interaction of high-strain zones with different local strain rates led to local differences in the
effective viscosity of the homogeneous material, which initiate the fold structures.
Acknowledgments References
We thank Laurent Montesi, Stefan
Ahlgren, S. G. (2001), The nucleation and evolution of Riedel shear zones as deformation bands in porous sandstone, J. Struct. Geol., 23(8),
Schmalholz, an anonymous reviewer,
1203–1214.
and particularly Neil Mancktelow for
Allken, V., R. S. Huismans, and C. Thieulot (2011), Three-dimensional numerical modeling of upper crustal extensional systems, J. Geophys.
their constructive comments. The
simulation data and the code to Res., 116, B10409, doi:10.1029/2011JB008319.
reproduce the results presented in this Allken, V., R. S. Huismans, and C. Thieulot (2012), Factors controlling the mode of rift interaction in brittle-ductile coupled systems: A 3D
study will be provided on request. This numerical study, Geochem. Geophys. Geosyst., 13, Q05010, doi:10.1029/2012GC004077.
study was supported by the Center for Arbaret, L., J. P. Burg, G. Zeilinger, N. Chaudhry, S. Hussain, and H. Dawood (2000), Pre-collisional anastomosing shear zones in the Kohistan
Computational Sciences (SRFN) at the arc, NW Pakistan, Geol. Soc. Spec. Pub., 170(1), 295–311.
University of Mainz and by the German Archanjo, C. J., M. H. B. M. Hollanda, S. W. O. Rodrigues, B. B. B. Neves, and R. Armstrong (2008), Fabrics of pre- and syntectonic granite plu-
Research Foundation (DFG), PA 578/ tons and chronology of shear zones in the Eastern Borborema Province, NE Brazil, J. Struct. Geol., 30, 310–326.
16–1. Barka, A. A., and K. Kadinsky-Cade (1988), Strike-slip fault geometry in Turkey and its influence on earthquake activity, Tectonics, 7, 663–684.
Behn, M. D., J. Lin, and M. T. Zuber (2002), A continuum mechanics model for normal faulting using a strain-rate softening rheology: Impli-
cations for thermal and rheological controls on continental and oceanic rifting, Earth Planet. Sci. Lett., 202, 725–740.
Bell, T. H. (1978), Progressive deformation and reorientation of fold axes in a ductile mylonite zone: The Woodroffe thrust, Tectonophysics,
44(1–4), 285–298, 301–320.
Bell, T. H. (1986), Foliation development and refraction in metamorphic rocks: Reactivation of earlier foliations and decrenulation due to
shifting patterns of deformation partitioning, J. Metamorph. Geol., 4, 421–444.
Bird, P., and X. Kong (1994), Computer simulations of California tectonics confirm very low strength of major faults, Geol. Soc. Am. Bull., 106,
159–174.
Bokelmann, G. H., and G. C. Beroza (2000), Depth-dependent earthquake focal mechanism orientation: Evidence for a weak zone in the
lower crust, J. Geophys. Res.: Solid Earth, 105(B9), 21683–21695.
Bons, P. D., and M. W. Jessell (1998), Folding in experimental mylonites, in Fault-Related Rocks—A Photographic Atlas, edited by A. W. Snoke,
pp. 366–367, Princeton University Press, N. J.
Bos, B., and C. J. Spiers (2002), Frictional-viscous flow of phyllosilicate-bearing fault rock: Microphysical model and implications for crustal
strength profiles, J. Geophys. Res., 107(B2), 2028, doi:10.1029/2001JB000301.
Brun, J. P., and P. R. Cobbold (1980), Strain heating and thermal softening in continental shear zones: a review, J. Struct. Geol., 2(1–2),
149–158.
Buck, W. R. (1993), Effect of lithospheric thickness on the formation of high- and low-angle normal faults, Geology, 21, 933–936.
Buiter, S. J., A. Y. Babeyko, S. Ellis, T. V. Gerya, B. J. Kaus, A. Kellner, G. Schreurs and Y. Yamada (2006), The numerical sandbox: comparison
of model results for a shortening and an extension experiment, Geol. Soc. Spec. Pub., 253, 29.
Burg, J., and S. Schmalholz (2008), Viscous heating allows thrusting to overcome crustal-scale buckling: Numerical investigation with appli-
cation to the Himalayan syntaxes, Earth Planet. Sci. Lett., 274(1), 189–203.
B€
urgmann, R., and G. Dresen (2008), Rheology of the lower crust and upper mantle: Evidence from rock mechanics, geodesy, and field
observations, Annu. Rev. Earth Planet. Sci., 36, 531–567, doi:10.1146/annurev.earth.36.031207.124326.
Carpenter, B. M., C. Marone, and D. M. Saffer (2011), Weakness of the San Andreas Fault revealed by samples from the active fault zone,
Nat. Geosci., 4, 251–254, doi:10.1038/ngeo1089.
Carreras, J. (2001), Zooming on Northern Cap de Creus shear zones, J. Struct. Geol., 23, 1457–1486.
Carreras, J., E. Druguet, and A. Griera (2005), Shear zone-related folds, J. Struct. Geol., 27, 1229–1251.
Carreras, J., D. M. Czeck, E. Druguet, and P. J. Hudleston (2010), Structure and development of an anastomosing network of ductile shear
zones, J. Struct. Geol., 32(5), 656–666.
Chery, J., M. D. Zoback, and R. Hassani (2001), An integrated mechanical model of the San Andreas fault in central and northern California,
J. Geophys. Res., 106, 22,051–22,066.
Chester, F. M., J. P. Evans, and R. L. Biegel (1993), Internal structure and weakening mechanism of the San Andreas Fault, J. Geophys. Res.,
98, 771–786.
Choi, E., L. Lavier, and M. Gurnis (2008), Thermomechanics of mid-ocean ridge segmentation, Phys. Earth Planet. Inter., 171, 374–386.
Cloos, H. (1928), Experimenten zur inneren Tektonic, Centralblatt Mineral. Paleontol. B, 609.
Cobbold, P. R., J. W. Cosgrove, and J. M. Summers (1971), Development of internal structures in deformed anisotropic rocks, Tectonophysics,
12(1), 23–53.
Cobbold, P. R., and H. Quinquis (1980), Development of sheath folds in shear regimes, J. Struct. Geol., 2, 119–126.
Coelho, S., C. Passchier, and F. Marques (2006), Riedel shear control on the development of pennant veins: Field example and analogue
modelling, J. Struct. Geol., 28, 1658–1669.
Collettini, C., A. Niemeijer, C. Viti, and C. Marone (2009), Fault zone fabric and fault weakness, Nature, 462, 907–910, doi:10.1038/nature08585.
Connolly, J. A. D., and Y. Y. Podladchikov (2000), Temperature-dependent viscoelastic compaction and compartmentalization in sedimen-
tary basins, Tectonophysics, 324, 137–168.
Culshaw, N. G., C. Gerbi, J. H. Marsh, and L. Plug (2011), Heterogeneous amphibolite facies deformation of a granulite facies layered proto-
lith: Matches Island Shear System, Parry Sound Domain, Grenville Province, Ontario, Canada, J. Struct. Geol., 33, 875–890.
da Silveira, A. B., J. Cabral, H. Perea, and A. Ribeiro (2009), Evidence for coupled reverse and normal active faulting in W Iberia: The Vidi-
gueira–Moura and Alqueva faults (SE Portugal), Tectonophysics, 474(1), 184–199.

Davis, G. H., A. P. Bump, P. E. Garcia, and S. G. Ahlgren (1999), Conjugate Riedel deformation band shear zones, J. Struct. Geol., 22, 169–190.
de Bresser, J. H. P., J. H. ter Heege, and C. J. Spiers (2001), Grain size reduction by dynamic recrystallization: Can it result in major rheological
weakening?, Int. J. Earth Sci., 90, 28–45.
Dennis, A. J., and D. T. Secor (1987), A model for the development of crenulations in shear zones with applications from the Southern
Appalachian Piedmont, J. Struct. Geol., 9(7), 809–817.
de Wit, M. J., S. A. Bowring, L. D. Ashwal, L. G. Randianasolo, V. P. I. Morel, and R. A. Rambeloson (2001), Age and tectonic evolution of Neo-
proterozoic ductile shear zones in southwestern Madagascar, with implications for Gondwana studies, Tectonics, 20, 1–45.
Drury, M. R., and J. I. Urai (1990), Deformation-related recrystallization processes, Tectonophysics, 172, 235–253.
Faulkner, D. R., and J. H. Rutter (2001), Can the maintenance of overpressured fluids in large strike-slip fault zones explain their apparent
weakness?, Geology, 29, 503–506.
Faulkner, D. R., C. A. L. Jackson, R. J. Lunn, R. W. Schlische, Z. K. Shipton, C. A. J. Wibberley, and M. O. Withjack (2010), A review of recent
developments concerning the structure, mechanics and fluid flow properties of fault zones, J. Struct. Geol., 32, 1557–1575, doi:10.1016/
j.jsg.2010.06.009.
Faure, M., E. Be Mezeme, M. Duguet, C. Cartier, and J.-Y. Talbot (2005), Paleozoic tectonic evolution of medio-Europa from the example of
the French Massif Central and Massif Armoricain, in The Southern Variscan Belt, edited by R. Carosi et al., J. Virtual Explorer, 19, 5.
Frederiksen, S., and J. Braun (2001), Numerical modelling of strain localisation during extension of the continental lithosphere, Earth Planet.
Sci. Lett., 188, 241–251.
Fusseis, F., M. R. Handy, and C. Schrank, (2006), Networking of shear zones at the brittle-to-viscous transition (Cap de Creus, NE Spain),
J. Struct. Geol., 28(7), 1228–1243.
Gerbi, C., N. Culshaw, and J. Marsh (2010), Magnitude of weakening during crustal-scale shear zone development, J. Struct. Geol., 32(1),
107–117.
Gerya, T. (2010), Dynamical instability produces transform faults at mid-ocean ridges, Science, 329, 1047–1050.
Gerya, T. (2012), Origin and models of oceanic transform faults, Tectonophysics, 522–523, 34–56.
Gerya, T. (2013a), Three dimensional thermomechanical modeling of oceanic spreading initiation and evolution, Phys. Earth Planet. Inter.,
214, 35–52.
Gerya, T. (2013b), Initiation of transform faults at rifted continental margins: 3D petrological-thermomechanical modeling and comparison
to the Woodlark basin, Petrology, 21(6), 550–560.
Gessner, K., S. Piazolo, T. G€ ung€or, U. Ring, and C. W. Passchier (2001), Tectonic significance of deformation patterns in granitoid rocks of
the Menderes Massif, southwest Turkey, Int. J. Earth Sci., 89, 766–780.
Gueydan, F., J. Pr ecigout, and L. G. J. Montesi (2014), Strain weakening enables continental plate tectonics, Tectonophysics, 631, 189–196.
Grujic, D., and N. S. Mancktelow (1998), Melt-bearing shear zones: Analogue experiments and comparison with examples from southern
Madagascar, J. Struct. Geol., 20, 673–680.
Harris, L. B., and P. R. Cobbold (1985), Development of conjugate shear bands during bulk simple shearing, J. Struct. Geol., 7(1), 37–44.
Herwegh, M., and M. R. Handy (1996), The evolution of high temperature mylonitic microfabrics: Evidence from simple shearing of a quartz
analogue (norcamphor), J. Struct. Geol., 18, 689–710.
Hieronymus, C. F. (2004), Control on seafloor spreading geometries by stress-and strain-induced lithospheric weakening, Earth Planet. Sci.
Lett., 222(1), 177–189.
Holdsworth, R. E. (2004), Weak faults-rotten cores, Science, 303(5655), 181–182.
Holyoke, C.W., III, and J. Tullis (2006), Mechanisms of weak phase interconnection and the effects of phase strength contrast on fabric
development, J. Struct. Geol., 28, 621–640.
Howard, K. A. (1968), Flow direction in triclinic foliated rocks, Am. J. Sci., 266, 758–765.
Huddleston, P. (1999), Strain compatibility and shear zones: Is there a problem?, J. Struct. Geol., 21, 923–932.
Huismans, R. S., and C. Beaumont (2003), Symmetric and asymmetric lithospheric extension: Relative effects of frictional-plastic and viscous
strain softening, J. Geophys. Res., 108(B10), 2496, doi:10.1029/2002JB002026.
Huismans, R. S., and C. Beaumont (2007), Roles of lithospheric strain softening and heterogeneity in determining the geometry of rifts and
continental margins, Geol. Soc. Spec. Pub., 282(1), 111–138.
Jammes, S., L. L. Lavier, and J. E. Reber (2015), Localization and delocalization of deformation in a bimineralic material, J. Geophys. Res. Solid
Earth, 120, 3649–3663, doi:10.1002/2015JB011890.
Jefferies, S. P., R. E. Holdsworth, C. A. J. Wibberley, T. Shimamoto, C. J. Spiers, A. R. Niemeijer, and G. E. Lloyd (2006), The nature and impor-
tance of phyllonite development in crustal-scale fault cores: an example from the Median Tectonic Line, Japan, J. Struct. Geol., 28(2),
220–235.
Kameyama, M., D. A. Yuen, and S. I. Karato (1999), Thermal-mechanical effects of low-temperature plasticity (the Peierls mechanism) on the
deformation of a viscoelastic shear zone, Earth Planet. Sci. Lett., 168(1), 159–172.
Katz, Y., R. Weinberger, and A. Aydin (2004), Geometry and kinematic evolution of Riedel shear structures, Capitol Reef National Park, Utah,
J. Struct. Geol., 26, 491–501.
Kaus, B., and Y. Podladchikov (2006), Initiation of localized shear zones in viscoelastoplastic rocks, J. Geophys. Res., 111, B04412, doi:
10.1029/2005JB003652.
Kaus, B. (2010), Factors that control the angle of shear bands in geodynamic numerical models of brittle deformation, Tectonophysics, 484,
36–47.
Lavier, L. L., W. R. Buck, and A. N. B. Poliakov (2000), Factors controlling normal fault offset in an ideal brittle layer, J. Geophys. Res., 105,
23,431–23,442.
Lister, G. S., and P. F. Williams (1983), The partitioning of deformation in flowing rock masses, Tectonophysics, 92, 1–33.
Lubliner, J. (2008), Plasticity Theory, Dover Publications, New York.
Marquer, D., N. Challandes, and T. Baudin (1996), Shear zone patterns and strain distribution at the scale of a Penninic nappe: the Suretta
nappe (Eastern Swiss Alps), J. Struct. Geol., 18(6), 753–764.
Marques, F. O., and S. Coelho (2001), Rotation of rigid elliptical cylinders in viscous simple shear flow: Analogue experiments, J. Struct.
Geol., 23, 609–617.
Martelat, J. E., K. Schulmann, J. M. Lardeaux, C. Nicollet, and H. Cardon (1999), Granulite microfabrics and deformation mechanisms in
southern Madagascar, J. Struct. Geol., 21, 671–687.
Mancktelow, N. (2002), Finite-element modelling of shear zone development in viscoelastic materials and its implications for localisation
of partial melting, J. Struct. Geol., 24, 1045–1053.
Mancktelow, N. S. (2006), How ductile are ductile shear zones?, Geology, 34(5), 345–348.

Mancktelow, N. S. (2007), Tectonic pressure: Theoretical concepts and modelled examples, Lithos, 103(1–2), 149–177.
Mancktelow, N. S., and G. Pennacchioni (2005), The control of precursor brittle fracture and fluid-rock interaction on the development of
single and paired ductile shear zones, J. Struct. Geol., 27, 645–661.
Mandal, N., S. K. Samanta, and C. Chakraborty (2004), Problem of folding in ductile shear zones: A theoretical and experimental investiga-
tion, J. Struct. Geol., 26, 475–489.
Mann, P. (2007), Global catalogue, classification and tectonic origins of restraining- and releasing bends on active and ancient strike-slip fault
systems, in Tectonics of Strike-Slip Restraining and Releasing Bends, edited by W. D. and P. Mann, Geol. Soc. Spec. Publ., 290, 367–385.
Mehl, L., and G. Hirth (2008), Plagioclase preferred orientation in layered mylonites: Evaluation of flow laws for the lower crust, J. Geophys.
Res., 113, B05202, doi:10.1029/2007JB005075.
Meyer, S. E., C. Passchier, T. Abu-Alam, and K. St€ uwe (2014), A strike-slip core complex from the Najd fault system, Arabian shield, Terra
Nova, 26, 387–394.
McKinnon, S. D. and I. Garrido de la Barra (1998), Fracture initiation, growth and effect on stress field: A numerical investigation, J. Struct.
Geol., 20, 1673–1689.
Mont esi, L. G. J. (2013), Fabric development as the key for forming ductile shear zones and enabling plate tectonics, J. Struct. Geol., 50,
254–266, doi:10.1016/j.jsg.2012.12.011.
Morrow, C. A., D. E. Moore, and D. A. Lockner (2000), The effect of mineral bond strength and absorbed water on fault gouge frictional
strength, Geophys. Res. Lett., 27, 815–818.
Naylor, M. A., G. Mandl, and C. H. K. Sijpesteijn (1986), Fault geometries in basement-induced wrench faulting under different initial stress
states, J. Struct. Geol., 8, 737–752.
Park, Y., S.-H. Yoo, and J.-H. Ree (2006), Weakening of deforming granitic rocks with layer development at middle crust, J. Struct. Geol., 28,
919–928, doi:10.1016/j.jsg.2006.02.005.
Passchier, C. W. (1990a), Reconstruction of deformation and flow parameters from deformed vein sets, Tectonophysics, 180, 185–199.
Passchier, C. W. (1990b), A Mohr circle construction to plot the stretch history of material lines, J. Struct. Geol., 12, 513–515.
Passchier, C. W. (1998), Monoclinic model shear zones, J. Struct. Geol., 20, 1121–1137.
Passchier, C. W., and J. P. Platt (2016), Shear zone junctions of zippers and freeways, J. Struct. Geol., 95, 188–202, doi:10.1016/j.jsg.2016.10.010.
Passchier, C. W., S. W. J. den Brok, J. A. M. van Gool, M. Marker, and G. Manatschal (1997), A laterally constricted shear zone system—The
Nordre Strømfjord steep belt, Nagssugtoqidian Orogen, W. Greenland, Terra Nova, 9, 199–202.
Pennacchioni, G. (2005), Control of the geometry of precursor brittle structures on the type of ductile shear zone in the Adamello tonalites,
Southern Alps (Italy), J. Struct. Geol., 27, 627–644.
Pennacchioni, G., and N. S. Mancktelow (2007), Nucleation and initial growth of a shear zone network within compositionally and structur-
ally heterogeneous granitoids under amphibolite facies conditions, J. Struct. Geol., 29, 1757–1780.
Platt, J. P. (1983), Progressive folding in ductile shear zones, J. Struct. Geol., 5, 619–622.
Platt, J. P., and W. M. Behr (2011a), Lithospheric shear zones as constant stress experiments, Geology, 39, 127–130, doi:10.1130/G31561.1.
Platt, J. P., and W. M. Behr (2011b), Grainsize evolution in ductile shear zones: Implications for strain localization and the strength of the
lithosphere, J. Struct. Geol., 33, 537–550, doi:10.1016/j.jsg.2011.01.018.
Piazolo, S., S. M. ten Grotenhuis, and C. W. Passchier (2001), New apparatus for controlled general flow modelling of analog materials, Geol.
Soc. Am. Mem., 193, 235–244.
Pollard, D. D., and A. Aydin (1984), Propagation linkage of Oceanic Ridge segments, J. Geophys. Res., 89, 10017–10028.
Poliakov, A. N. B., and W. R. Buck (1998), Mechanics of Stretching Elastic-Plastic-Viscous Layers: Applications to Slow-Spreading Mid-Ocean
Ridges, vol. 106, pp. 3, AGU, Washington, D. C., doi:10.1029/GM106p0305.
Popov, A. A., S. V. Sobolev, and M. D. Zoback (2012), Modeling evolution of the San Andreas Fault system in northern and central California,
Geochem. Geophys. Geosyst., 13, Q08016, doi:10.1029/2012GC004086.
Ponce, C., E. Druguet, and J. Carreras (2013), Development of shear zone-related lozenges in foliated rocks, J. Struct. Geol., 50, 176–186.
Provost, A.-S., and H. Houston (2003), Stress orientations in Northern and Central California: Evidence for the evolution of frictional strength
along the San Andreas plate boundary system, J. Geophys. Res., 108(B3), 2175, doi:10.1029/2001JB001123.
Regenauer-Lieb, K., D. Yuen, and J. Branlund (2001), The initiation of subduction: Criticality by addition of water?, Science, 294, 578–580.
Rice, J. R. (1992), Fault Mechanics and Transport Properties of Rocks, edited by B. Evans and T.-F. Wong, pp. 475–503, Academic Press,
London.
Ridley, J. (1993), The relations between mean rock stress and fluid flow in the crust: With reference to vein-and lode-style gold deposits,
Ore Geol. Rev., 8, 23–37.
Riedel, W. (1929), Zur Mechanik Geologischer Brucherscheinungen, Zentral-blatt Mineral. Geol. Paleontol. B, 354–368.
Schmalholz, S. M., T. Duretz, F. L. Schenker, and Y. Y. Podladchikov (2014), Kinematics and dynamics of tectonic nappes: 2-D numerical
modelling and implications for high and ultra-high pressure tectonism in the Western Alps, Tectonophysics, 631, 160–175.
Schmalholz, S. M., and T. Duretz (2015), Shear zone and nappe formation by thermal softening, related stress and temperature evolution,
and application to the Alps, J. Metamorphic Geol., 33(8), 887–908.
Scholz, C. H., R. Ando, and B. E. Shaw (2010), The mechanics of first order splay faulting: The strike-slip case, J. Struct. Geol., 32(1), 118–126.
Schrank, C. E., M. R. Handy, and F. Fusseis (2008), Multiscaling of shear zones and the evolution of the brittle-to-viscous transition in conti-
nental crust, J. Geophys. Res., 113, B01407, doi:10.1029/2006JB004833.
Schreurs, G. (1994), Experiments on strike-slip faulting and block rotation, Geology, 22, 567–570.
Sibson, R. H. (1977), Fault rocks and fault mechanisms, J. Geol. Soc. London, 133, 191–213.
Sibson, R. H. (1987), Earthquake rupturing as a mineralizing agent in hydrothermal systems, Geology, 15, 701–704, doi:10.1130/0091-7613.
Sibson, R. H. (1990), Conditions for fault-valve behaviour, in Deformation Mechanisms, Rheology and Tectonics, edited by R. J. Knipe and
E. H. Rutter, Geol. Soc. Spec. Publ., 54, 15–28.
Simpson, C. (1982), Strain and shape-fabric variations associated with ductile shear zones, J. Struct. Geol., 8(2), 111–115, 117–122.
Stern, R. J. (1994), Arc assembly and continental collision in the Neoproterozoic East African orogen: Implications for the consolidation of
Gondwana, Annu. Rev. Earth Planet. Sci., 22, 319–351.
Stern, R. J., and P. Johnson (2010), Continental lithosphere of the Arabian Plate: a geologic, petrologic, and geophysical synthesis, Earth-Sci-
ence Reviews, 101(1), 29–67.
Sylvester, A. G. (1988), Strike-slip faults, Geol. Soc. Amer. Bull., 100(11), 1666–1703.
Thielmann, M., and B. J. P. Kaus (2012), Shear heating induced lithospheric localization: Does it result in subduction?, Earth Planet. Sci. Lett.,
359–360, 1–13, doi:10.1016/j.epsl.2012.10.002.

Thielmann, M., A. Rozel, B. J. P. Kaus, and Y. Ricard (2015), Intermediate-depth earthquake generation and shear zone formation caused by
grain size reduction and shear heating, Geology, 43(9), 791–794.
Tirel, C., J. P. Brun, and E. Burov (2008), Dynamics and structural development of metamorphic core complexes, J. Geophys. Res., 113,
B04403, doi:10.1029/2005JB003694.
Tullis, J., J. M. Christie, and D. T. Griggs (1973), Microstructures and preferred orientations of experimentally deformed quartzites, Bull. Geol.
Soc. Am., 84, 297–314.
Vermeer, P., and R. De Borst (1984), Non-associated plasticity for soils, concrete and rock, Heron, 29(3), 1–64.
Vitale, S., and S. Mazzoli (2008), Heterogeneous shear zone evolution: The role of shear strain hardening/softening, J. Struct. Geol., 30,
1363–1395.
Wallace, R. (1990), The San Andreas fault system, California, U.S. Geol. Surv. Prof. Pap., 1515, 283.
White, S. (1976), The effects of strain on the microstructures, fabrics, and deformation mechanisms in quartzites, Philos. Trans. R. Soc. Lon-
don A, 283, 69–86.
White, S. H., S. E. Burrows, J. Carreras, and N. D. Shaw (1980), On mylonites in ductile shear zones, J. Struct. Geol., 2, 175–187, doi:10.1016/
0191-8141(80)90048-6.
Williams, P. F., and G. P. Price (1990), Origin of kink bands and shear-band cleavage in shear zones: An experimental study, J. Struct. Geol.,
12(2), 145–164.
Zoback, M. D. (2000), Strength of the San Andreas, Nature, 405, 31–32.

Numerical Modeling of Fault Zone

Uploaded by

Copyright:

Available Formats

You might also like

Numerical Modeling of Fault Zone

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Numerical Modeling of Fault Zone

Uploaded by

Copyright:

Available Formats

PUBLICATIONS

Geochemistry, Geophysics, Geosystems

MEYER ET AL. DEVELOPMENT OF BRANCHING SHEAR ZONES 2054

2. Modeling Approach (Methodology)

l5tan ð/Þ (1)

MEYER ET AL. DEVELOPMENT OF BRANCHING SHEAR ZONES 2055

F5rII 2ry (5)

MEYER ET AL. DEVELOPMENT OF BRANCHING SHEAR ZONES 2056

2.1. Model Setup

MEYER ET AL. DEVELOPMENT OF BRANCHING SHEAR ZONES 2057

MEYER ET AL. DEVELOPMENT OF BRANCHING SHEAR ZONES 2058

MEYER ET AL. DEVELOPMENT OF BRANCHING SHEAR ZONES 2059

MEYER ET AL. DEVELOPMENT OF BRANCHING SHEAR ZONES 2060

MEYER ET AL. DEVELOPMENT OF BRANCHING SHEAR ZONES 2061

3.2. Ductile Localized Shear Zones

MEYER ET AL. DEVELOPMENT OF BRANCHING SHEAR ZONES 2062

MEYER ET AL. DEVELOPMENT OF BRANCHING SHEAR ZONES 2063

MEYER ET AL. DEVELOPMENT OF BRANCHING SHEAR ZONES 2064

MEYER ET AL. DEVELOPMENT OF BRANCHING SHEAR ZONES 2065

4.2. Evolution and Interaction of Faults (Upper Crust)

MEYER ET AL. DEVELOPMENT OF BRANCHING SHEAR ZONES 2066

MEYER ET AL. DEVELOPMENT OF BRANCHING SHEAR ZONES 2067

MEYER ET AL. DEVELOPMENT OF BRANCHING SHEAR ZONES 2068

4.3. Numerical Simulations in Ductile Regimes (Lower Crust)

MEYER ET AL. DEVELOPMENT OF BRANCHING SHEAR ZONES 2069

MEYER ET AL. DEVELOPMENT OF BRANCHING SHEAR ZONES 2070

MEYER ET AL. DEVELOPMENT OF BRANCHING SHEAR ZONES 2071

MEYER ET AL. DEVELOPMENT OF BRANCHING SHEAR ZONES 2072

MEYER ET AL. DEVELOPMENT OF BRANCHING SHEAR ZONES 2073

MEYER ET AL. DEVELOPMENT OF BRANCHING SHEAR ZONES 2074

MEYER ET AL. DEVELOPMENT OF BRANCHING SHEAR ZONES 2075

You might also like