You are on page 1of 17

Tectonophysics 765 (2019) 129–145

Contents lists available at ScienceDirect

Tectonophysics
journal homepage: www.elsevier.com/locate/tecto

Modulation of fault strength during the seismic cycle by grain-size evolution T


around contact junctions
Sylvain Barbot
University of Southern California, United States of America

ARTICLE INFO ABSTRACT

Keywords: Earthquake prediction is fundamentally hampered by insufficient knowledge of the constitutive behavior of
Rate-state friction faults. Laboratory and theoretical work by Dieterich (1978) and Ruina (1983) formed the basis of rate-and-state
Rheology friction, a constitutive framework that captures the healing and weakening of frictional contacts. The rate-and-
Fault dynamics state law successfully explains the frictional behavior of a wide range of rocks and many phenomena associated
Fault physics
with fault slip. The physics underlying the rate-and-state friction law is not fully understood, but it is generally
accepted that several micro-mechanisms operate depending on the texture, maturity, fluid content, and other
physical and kinematic conditions of the fault. Here, I formulate a constitutive framework that unifies a number
of laboratory experiments on silicate rocks, and the gouges derived from these rocks, that describe the effect of
grain-size, temperature, real area of contact, gouge thickness, and fault roughness. Fundamental to the model is
that fault strength is controlled by the area of the interfacial contact junctions that support the shear and normal
loads. The curvature of asperities weakly controls the area of their contact junction and the resulting fault
strength. As a result, the dynamics of grain-size evolution around micro-asperities enables the seismic cycle. The
model quantitatively explains the correlation between the characteristic weakening distance and the gouge
thickness (Marone and Kilgore, 1993), the correlation of the static friction coefficient with the rate-dependence
parameters (Ikari et al., 2011), the dependence of frictional resistance to sub-solidus temperatures (Chester,
1994), and predicts a range of grain sizes that is compatible with observations at exposed fault zones. However,
the dynamics of grain-size evolution is still challenging to verify experimentally. The model features a reg-
ularization at vanishing slip speeds that implies distinct dynamics at low strength. For friction coefficients lower
than 0.1, seismic cycle simulations predict the emergence of frictional instabilities with complex source time-
functions. This is a significant departure from the predictions of rate-and-state friction that is most relevant to
the dynamics of décollements, serpentinized shear zones, and other low-strength faults.

1. Introduction explains many phenomena associated with fault slip (Barbot et al.,
2012; Duan et al., 2017; Hori et al., 2004; Horowitz and Ruina, 1989;
Friction intervenes in virtually every aspect of our lives. It is a topic Kaneko et al., 2016; Lapusta and Barbot, 2012; Lapusta and Rice, 2003;
of great practical importance for many engineering applications and it Liu and Rice, 2005; Mele Veedu and Barbot, 2016; Qiu et al., 2016; Rice
controls the dynamics of earthquakes (Brace and Byerlee, 1966). The and Ben-Zion, 1996; Rice and Ruina, 1983; Tse and Rice, 1986; Wei
static force that applies at solids in stationary contact has been de- et al., 2013), including in the quasi-static regime (Agata et al., 2019;
scribed experimentally first by Da Vinci, Amontons (1699) and Barbot et al., 2009; Bürgmann et al., 2002; Hsu et al., 2006; Marone
Coulomb (1785). Dieterich (1978, 1979) recognized how the velocity of et al., 1991; Masuti et al., 2016; Perfettini and Avouac, 2004; Perfettini
sliding and contact aging affect the frictional resistance and, together and Avouac, 2007; Rousset et al., 2012; Salman et al., 2017; Wei et al.,
with Ruina (1983), formulated the rate-and-state friction law, a con- 2015).
stitutive framework that successfully describes the behavior of many The experiments of Logan and Teufel (1986) and Teufel and Logan
materials such as rocks, paper, metals, and glass for a wide range of (1978) with thermodyes and those of Dieterich and Kilgore (1994,
experimental conditions (Baumberger et al., 1997; Berthoud et al., 1996) on transparent materials, shortly supported by several others
1999; Blanpied et al., 1991; Dieterich, 1972; Heslot et al., 1994; Kilgore (Ben-David et al., 2010; Goldsby et al., 2004; Persson, 2016; Rubinstein
et al., 1993; Leeman et al., 2016; Scholz, 1998; Scuderi et al., 2017) and et al., 2004), revealed that the dynamics of frictional sliding is governed

E-mail address: sbarbot@usc.edu.

https://doi.org/10.1016/j.tecto.2019.05.004
Received 25 January 2019; Received in revised form 9 April 2019; Accepted 9 May 2019
Available online 30 May 2019
0040-1951/ © 2019 The Author. Published by Elsevier B.V. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/BY/4.0/).
S. Barbot Tectonophysics 765 (2019) 129–145

by the physics of the inter-granular contacts that support the load. A Gouge model
These micro-asperities emerge due to the roughness of the material
surface at the micro- or even nano-scale (Bennett et al., 2017; Candela Gouge
Sliding
thickness W
and Brodsky, 2016; Candela et al., 2012; Harbord et al., 2017; Maegawa velocity V
et al., 2015; Pei et al., 2005; Persson et al., 2004; Persson, 2001), such
that only a minute fraction of the macroscopic interfacial area is in
Gouge
contact with the opposing wall. The evolution of the real area of contact
during fault slip also explains the modulation of electrical conductivity,
coda waves, and acoustic transmissivity across fault contacts
(Fukuyama et al., 2019; Kocharyan et al., 2018; Nagata et al., 2008; Country Strain-rate
Rouet-Leduc et al., 2018; Yamashita et al., 2014). These observations rock
explain why few surfaces are adhesive, i.e., strong in tension, despite
B Bare surface model
the potential importance of attraction forces, such as chemical bonding
(Dieterich and Conrad, 1984; Li et al., 2011) or van der Waals forces
(Pastewka and Robbins, 2014), the so-called adhesion paradox Surface
Sliding
roughness W
(Ciavarella et al., 2019; Lorenz et al., 2013). velocity V
The real area of contact is controlled at first order by the macro-
scopic normal stress (Dieterich and Kilgore, 1994; Dieterich and
Kilgore, 1996; Krick et al., 2012; Logan and Teufel, 1986; Lorenz et al.,
2013; Maegawa et al., 2015; Pastewka and Robbins, 2014; Persson, Bare surface

2016; Teufel and Logan, 1978; Yastrebov et al., 2015), but the area of
contact of the microasperities grows logarithmically during stationary Strain-rate
contact under confining pressure and decreases under shear (Dieterich
and Kilgore, 1994; Dieterich and Kilgore, 1996; Sahli et al., 2018). C Contact model
Stationary healing of the micro-asperities is also observed in nano-in-
dentation experiments that mimic a single inter-granular contact
(Goldsby et al., 2004; Thom et al., 2018). The variations of strength at Grain-size Grain
(local radius
the contact level provide a micro-mechanism for the modulation of of curvature)
frictional resistance during the seismic cycle.
One of the remaining challenges is to combine all these laboratory
observations into a physics-based constitutive framework at the con- Contact junction
tinuum level. Recent experimental work, however, has shed light on the
limitations of the rate-and-state friction framework, as processes acti-
Real area of contact density
scales with normal stress
vated by shear heating substantially reduce the frictional resistance of and grain size
the fault at high slip velocity (Beeler et al., 2008; Ferri et al., 2010;
Goldsby and Tullis, 2011; Han et al., 2007; Kitajima et al., 2011; Noda, Fig. 1. Schematic for the relationship between grain-size and real area of
2008; Rowe et al., 2019; Sulem and Famin, 2009; Toro et al., 2004; contact in a shear zone or for bare surfaces in direct frictional contact. A) In a
Toro et al., 2006). Other observations, particularly on carbonate gouge, the frictional strength is supported by closed grains in direct contact.
gouges, exhibit a unique sensitivity to temperature that is not readily The strain-rate is approximated by the ratio of the sliding velocity to the gouge
explained by rate-and-state friction, presumably due to the important thickness W. B) For bare surfaces in frictional contact, the strength is supported
role of pressure-solution enhanced compaction and dilatation (Chen by micro-asperities that form a small fraction of the nominal surface. The grain
is defined as the material immediately surrounding the contact junction. The
and Spiers, 2016; Chen et al., 2015; Fagereng and den Hartog, 2016;
macroscopic strain-rate is taken as the ratio of the sliding velocity to a wave-
Niemeijer and Spiers, 2007) or because localization is prevented
length characteristic of the fault roughness W. C) Grain-size is defined as the
(Nakatani and Scholz, 2004a; Nakatani and Scholz, 2004b). Other local radius of curvature at the contact junction (brown lines). In the proposed
colleagues model the rheology of fault gouge outside the framework of contact model, normal stress and grain-size control the real area of contact.
rate-and-state friction (Daub and Carlson, 2010; den Hartog and Spiers, Grain-size affects fault strength by modulating the real area of contact. Grain
2014; Lieou et al., 2014; Lieou et al., 2017; Niemeijer and Spiers, 2007). growth corresponds to time-dependent processes that flatten the surface.
In this paper, I focus on explaining the framework of rate-and-state Reduction in real area of contact accompanies comminution. (For interpretation
friction with the aging law within its applicable range of conditions. of the references to color in this figure legend, the reader is referred to the web
Naturally, the resulting model does not explain all these other ob- version of this article.)
servations, which will require further attention.
Many colleagues have already proposed a physical model for rate- stability, this approach has been useful to explain the slow-slip phe-
and-state friction based on the shared principle that the true area of nomenon for a wide range of physical conditions (Segall and Rice,
contact along frictionally slipping surfaces is considerably less than the 1995; Segall et al., 2010) and the low strength of continental strike-slip
apparent area, but most of this work has focused on the direct effect or faults (Sleep and Blanpied, 1992). In this paper, I present a friction
steady-state, modeled as the result of a thermally activated rheology model based on the physics of inter-granular contacts that applies to
with an activation energy and an activation volume (Aharonov and both bare contacts and fault gouges.
Scholz, 2018; Baumberger and Caroli, 2006; Baumberger et al., 1997; In the following sections, I start with the physical model description,
Heslot et al., 1994; Nakatani, 2001; Putelat et al., 2011). A unified show its mathematical equivalence with the classic formulation of
model for rate-and-state friction based on grain-level physics was pro- Dieterich (1979) and Ruina (1983) with the aging law in isothermal
posed by Sleep (1997) based on the gouge model of Sleep and Blanpied conditions, and with the formulation of Chester (1994) for non-iso-
(1992), where they associate the state variable with porosity and the thermal conditions. I then constrain the constitutive properties with
aging law describes compaction and dilatation of the gouge layer filling laboratory data, and finally discuss some implications on fault dy-
the fault core. As variations in pore space affect fluid pressure and fault namics.

130
S. Barbot Tectonophysics 765 (2019) 129–145

A Hertzian contact (Hertz, 1890) B Rough contact (Archard, 1957)

d2
Micro-asperity d1
d3
Micro-asperity
Real area of contact A

d1

Area of contact

d
C Rough surface (Greenwood & Williamson, 1966)
Micro-asperities
d
d

d d d

Fig. 2. Fault strength controlled by the asperities that support the load at the microscopic level. A) Hertz's theory describes the real area of contact for a single
2
spherical asperity under normal stress, predicting the real area of contact A~d ¯ 3 , where is the effective normal stress. B) The rough contact model assumes a fractal
distribution of asperities of decreasing size (Archard, 1957). C) The rough surface model assumes a random distribution of asperity heights (Greenwood and
Williamson, 1966). Both the rough fault and the rough contact models predict a linear dependence of fault strength on normal stress and a weak dependence on
asperity size.

2. Constitutive model that the latter are themselves made of even smaller asperities and for
1 26
this configuration finds A d 9 27 . The power exponents change de-
I start with the assumption that the normal load and shear stress pending on the geometrical assumptions, but his results show that in-
across two solids in contact is supported by a small fraction of the creasing contact roughness leads to an asymptotic linear relationship
macroscopic surface consisting of a finite number of micro-asperities between the area of contact and normal stress and an asymptotically
(Bowden and Tabor, 1964). I refer to a micro-asperity, or a grain, as the vanishing dependence on grain-size. Greenwood and Williamson
topographic high of local radius of curvature d surrounding and in- (1966) simply assume a random distribution of asperity heights
cluding a contact junction (Fig. 1). The real area of contact is the area of 1
(Fig. 2C) and find A d 2 for the statistically expected value. These
the contact junction. In a fault gouge, a grain may form a closed vo- models suggest that the local radius of curvature of an asperity exerts a
lume. If the grain is spherical, d represents the radius of the grain before weak control on the area of contact that can be captured with a power-
deformation. For contacts along two bare surfaces, the grain is loosely law relationship.
defined as the region of sharp topography around the contact junction Considering the density of the real area of contact A , i.e., the ratio
that is most affected by the local stress concentration. The fate of other of the real area of contact to the macroscopic or nominal contact area, I
grains in the rock matrix that are not directly participating in the propose the following dependence on cohesion, normal stress, and grain
contact is irrelevant and ignored. size
Fault strength is controlled by the asperities that support the load at m
the microscopic level. Assuming a spherical geometry of the sur- c + µ0 d n
A= ,
rounding asperity, Hertz's theory (Timoshenko and Goodier, 1951) d0 (1)
should describe the real area of contact at these junctions (Fig. 2A).
where 0 A 1 is the contact density, c is the cohesion, μ0 is the
However, Hertzian contacts fail to predict the linear relationship be-
coefficient of friction, χ is the indentation hardness of the material
tween the real area of contact, frictional resistance, and normal stress
(Brace, 1960; Dieterich and Kilgore, 1996; Goldsby et al., 2004), d is the
that is explicit in Amonton's law. This paradox is resolved by fault
grain size, d0 is a reference grain size, and m/n is a power exponent with
roughness in two end-member models that assume either a fractal
∣m ∣ ≪ n. Typical values of m are in the range from 1 to 4, but negative
distribution of micro-asperities (Archard, 1957) or a random distribu-
values are not excluded theoretically. The power exponent n is of the
tion of asperity heights (Greenwood and Williamson, 1966). In a
2 order of 60. The linear form (c + µ 0 )/ is compatible with the for-
Hertzian contact, the area of contact scales with A d 3 , where is
mulation of Bowden and Tabor (1964) and captures Amonton's law for
the effective normal stress that includes the effect of pore pressure. Not
a moderate range of normal stress (Byerlee, 1978; Lorenz et al., 2013;
only does this model violate Amonton's law, but it also predicts a strong
Stesky, 1978; Yastrebov et al., 2015). A more general relationship that
dependence of fault strength on grain-size that is incompatible with
captures the full range of the area of contact, including near saturation
experimental findings (Biegel et al., 1989; Dieterich, 1979; Logan and
(i.e., as A approaches 1) and at vanishing normal stress, is outside the
Teufel, 1986; Marone and Kilgore, 1993; Teufel and Logan, 1978).
scope of the study. In particular, the range of normal stress leading to
Archard (1957) assumes that the grain surface is made of a succession
1 8 saturation of the real area of contact corresponds to depths in the Earth
of smaller asperities (Fig. 2B) and finds A d 3 9 . He then assumes where plastic flow dominates. The multiplicative factor (d/d0)m/n

131
S. Barbot Tectonophysics 765 (2019) 129–145

represents the effect of the curvature of the grain that surrounds the where d is the size of the asperities involved in inter-granular contacts,
contact junction (Archard, 1957; Greenwood and Williamson, 1966). p is a power exponent in the range from 1 to 3 depending on the
Although it is second-order, this effect is important for the seismic cycle dominant mechanism (Atkinson, 1988), G0 a characteristic rate of
(Rice, 1983). growth, H is the activation enthalpy, and 1/λ is a critical strain for
With the accumulation of fine-grain material associated with com- microstructural evolution that controls the rate of grain-size reduction
minution, wear, and the abrasion of the fault walls (Biegel and Sammis, and associated weakening. For later convenience, I also define
2004; Chester and Logan, 1986; Giger et al., 2008), the macroscopic
contact may be represented by simple shear in a narrow zone delineated G = G0 exp
H
,
by two more cohesive surfaces (Fig. 1). The deformation within the RT (5)
gouge layer (cataclastic rock or breccia, wear, and their alteration
products) can be treated as the shearing of a granular material at the which represents an effective rate of growth. The first term on the right-
continuum mechanics level (Barbot and Fialko, 2010; Byerlee and hand side of Eq. (4) assumes that the rate of grain growth depends on
Savage, 1992; Chester, 1994; Paterson and Wong, 2005; Scott et al., the current grain curvature (Atkinson, 1988). This term assumes a
1994; Sleep, 1997). For bare contacts, we can define the strain-rate as single, dominant healing mechanism. More realistic models could in-
the instantaneous velocity of sliding divided by a characteristic wave- corporate more mechanisms, each with its own activation energy. The
length defined by the fault roughness (Dieterich, 1979; Harbord et al., second term in Eq. (4) represents comminution and is proportional to
2017; Power et al., 1987; Sammis and Biegel, 1989; Scholz, 1988). The the plastic strain-rate.
slip behavior for both slip on clear, machine-finished surfaces and slip The evolution law is parameterized with a single dynamic variable
across surfaces separated by a gouge layer (Dieterich, 1979; Marone to represent the grain-size distribution. This may be an over-simplifi-
and Kilgore, 1993; Marone et al., 1990; Marone et al., 2009) can then cation, but as such, it implies either that the surface is self-similar
be described with the same physical model based on inter-granular (Candela et al., 2012; Power et al., 1987; Sammis et al., 1987; Sammis
contacts (Fig. 1). and Biegel, 1989; Sammis and King, 2007) or that this parameter re-
The area of contact controls fault strength and strain-rate, through presents the maximum grain-size, that is, the grain-size above which
the proposed power-law stress-strain-rate relationship the power-law breaks down. Describing the continuous distribution of
grain-size brings great complications and is an area of improvement
n
Q (e.g., Bush et al., 1975; Persson, 2001; Rozel et al., 2011), but the ex-
= 0 exp ,
A RT (2) perimental work by Badt et al. (2016) indicates that the power-law
scaling of fractal surfaces is stable under shear, hence the distribution
where is the plastic strain-rate within the shear zone, τ is the norm of can be represented by a single scaling factor that increases with normal
the shear traction vector parallel to the fault, A is the reference stress, stress and decreases with shear (Badt et al., 2016).
and an Arrhenius law is assumed for the thermal activation with the Even though this mechanism is not specifically invoked, Eq. (4) is
activation energy Q, the absolute temperature T, and the universal gas the same as for dynamic recrystallization (Chester, 1989; De Bresser
constant R, as was proposed by Chester (1994, 1995). The constant 0 is et al., 1998; De Bresser et al., 2001; Hall and Parmentier, 2003; Montési
a reference strain-rate. The power exponent n ≫ 1 is large, typically of and Hirth, 2003; Nishihara et al., 2006), which is important in the
the order of 60, to represent the fact that sliding is catastrophic when ductile substrate at lower-crustal and mantle depths, because the phy-
the shear traction exceeds a fraction of the material strength defined by sics describing grain-size evolution, for example, surface tension or
the real contact area density. However, no macroscopic shear de- minimization of surface energy, can take the same functional form for a
formation occurs in the absence of shear traction. Combining Eqs. (1) solid-solid phase intra-granular interface and for the solid-fluid phase
and (2), we obtain the following flow law in multiplicative form interface at the grain boundary (Atkinson, 1988).
n m The combination of Eqs. (3) and (4) represents a closed-form con-
d Q
= 0 exp , stitutive model for the evolution of velocity, stress, and the state of the
c + µ0 d0 RT (3) contact junctions bearing the load. The proposed constitutive frame-
where the size of the asperities bearing the contact junctions has been work incorporates physical quantities such as particle size or the
made explicit. Because the area of contact depends inversely on the thickness of the shear layer, that are measurable in the laboratory or in
indentation hardness (A 1/ ) and the macroscopic yield strength the field, at least in principle (Fagereng and Sibson, 2010; Sone et al.,
depends linearly on indentation hardness ( A ), this material 2012). The model represents an alternative to the one proposed by
property is absent in the final flow law (Eq. (3)). Sleep (1997) and Sleep and Blanpied (1992) based on the evolution of
The remaining constituent of the model is to allow grain-size, or porosity in the gouge layer. However, the two models may not be
equivalently, the local radius of curvature, to vary dynamically. mutually exclusive as bulging and flattening of the surface surrounding
Microfracturing leads to grain-size reduction within the fault zone the contact junctions may not occur without significant changes in
(Sammis et al., 1987; Sammis and Biegel, 1989). But this is counter- porosity or micro-crack length.
acted by other, time-dependent processes acting at longer times scales,
such as surface migration, lattice diffusion, surface diffusion (Atkinson,
1988), intra-crystalline plasticity (Hulikal et al., 2015; Mitchell et al., 3. Correspondence with rate-and-state friction
2013; Perfettini and Molinari, 2017), static recrystallization (Chester,
1989), chemical bounding (Li et al., 2011; Liu and Szlufarska, 2012), or The constitutive framework based on Eqs. (3) and (4) can be shown
diffusion of the elastic energy stored in the grains, that lead to time- to closely match the rate-and-state friction formulation of Dieterich
dependent grain growth, or equivalently, flattening. Viscoelastic effects (1979) and Ruina (1983) with the aging law under isothermal condi-
are possible in the brittle field at the micro-scale level because the loads tions. The correspondence with the logarithmic form of rate-and-state
endured by the micro-asperities are an order of magnitude higher than friction is immediate if we notice that power-laws of the form xy with
the macroscopic stress and close to the yield limit (e.g., Teufel and y ≪ 1 can be written xy = exp (y ln (x)) and then approximated by the
Logan, 1978). A constitutive framework that captures this behavior is linear terms of the Taylor series expansion of exp (x), i.e.,
x y = 1 + y ln (x ) + O (ln2 (x )) .
the thermally activated grain-size evolution law
First, to connect the grain-size with the state variable of rate-and-
G0 H state friction in isothermal conditions, I define a state variable for the
d= exp d,
p dp 1 RT (4) age of the grain, or equivalently, the duration of grain growth, as

132
S. Barbot Tectonophysics 765 (2019) 129–145

=
1 p
d . halite can be approximated by a power-law and a similar result for
G (6) shear zones was discussed by Montési and Hirth (2003) using different
The instantaneous and reference strain-rates are then associated with physical assumptions. It is likely that a and b depend on temperature, as
the velocity across the shear zone has been shown for amorphous materials (Ronsin and Coeyrehourcq,
2001) and carbonate rocks (Chen and Spiers, 2016; Chen et al., 2015).
V = 2W (7)
Healing may also be mediated by the growth and welding of contact
and the reference velocity junctions mediated by pressure solution creep (Chester and Higgs,
1992; Yasuhara et al., 2005) but we do not include these effects in this
Q
V0 = 2W 0 exp , study.
RT (8)
The classic formulation of rate-and-state friction appears to be an
where W is the characteristic wavelength of the system (Fig. 1). Com- approximation of a power-law rheology under isothermal conditions.
bining Eqs. (1) and (6), linearizing Eq. (1), neglecting cohesion, and Allowing temperature to vary, defining the age of the grains, or the age
defining the reference grain-size adequately, we obtain the following of contact, as follows
relationship for the real area of contact (see Appendix A)
1 H
= exp d p,
µ0 m V0 G0 RT0 (15)
A= 1+ ln ,
pn L (9) with the reference velocity
which was proposed by Ronsin and Coeyrehourcq (2001). This first Q
result indicates that the age of the grain (Eq. (6)) coincides with the age V0 = 2W 0 exp ,
RT0 (16)
of contact defined by Dieterich (1979). We can recast Eq. (3) as an
expression for strength as a function of velocity and state and otherwise following the same steps as above, we find the following
1 m
constitutive behavior (see Appendix B)
V n
V0 pn
= µ0 , V Q 1 1 V0
V0 L (10) µ = µ 0 + a ln + + b ln ,
V0 R T T0 L (17)
where the terms in parenthesis with their exponent are of order one,
compatible with the approximation µ 0 used as a simple failure with the evolution law
criterion. Indeed, the effects of the state variable and sliding velocity on H 1 1 V
strength are small, of the order of a few percents, but sufficient to en- = exp ,
R T T0 L (18)
able frictional instabilities and the seismic cycle (Rice, 1983; Ruina,
1983). For sufficiently large power exponent n, one can reduce Eq. (10) which corresponds to the thermally-activated rate-state formulation
to the linear terms of its power series. The additive, logarithmic form proposed by Chester and Higgs (1992) and Chester (1994) based on
directly emerges (see Appendix A) extensive experimental data. By extension, the proposed multiplicative,
thermally-activated form of rate-and-state friction explains the relevant
V V0
= µ 0 + a ln + b ln . observations of these studies and provides a specific micro-physical
V0 L (11) framework to justify the thermal activation.
Finally, combining Eqs. (4) and (6), we obtain the aging law of
Dieterich (1979) exactly, 4. Laboratory constraints on physical parameters

V The proposed framework (Eqs. (3)–(4)) satisfies a number of see-


=1 .
L (12) mingly unrelated constraints from laboratory observations. The con-
The relationships between the parameters of the multiplicative form stitutive parameters can be uniquely estimated combining insight from
Eqs. (3) and (4) with the classic parameters of rate-and-state friction in several key experiments. In this section, I start by considering the
additive form Eqs. (11) and (12) are found by identification findings of Dieterich and Kilgore (1994) and Dieterich and Kilgore
(1996) that document the effect of surface preparation on fault strength
µ0
a= , in stationary contact and during sliding. I then explain the correlation
n
m between the kinematic and the static friction coefficients found by Ikari
b = a. et al. (2011) and Ikari et al. (2016). I discuss the dependence of friction
p (13)
on temperature that was previously analyzed by Chester and Higgs
The characteristic weakening distance (1992) and Chester (1994). Finally, I compare the predictions of grain-
2W size of the physical model based on numerical simulations with those
L= found in fault zones (e.g., Chester and Logan, 1986; Frost et al., 2009).
p (14)
The laboratory experiments of Dieterich and Kilgore (1994) and
is proportional to the characteristic wavelength of the system (see Dieterich and Kilgore (1996) using transparent materials indicate a
Methods). The gouge model of Sleep (1995) also predicts a linear cor- dependence between the real area of contact and macroscopic normal
relation between the characteristic weakening distance and the gouge stress, a “logarithmic” increase of surface area with hold time during
thickness, which emerges naturally when the weakening term is pro- stationary contact, and its correlation with the state variable of the
portional to the anelastic strain-rate. aging law during the seismic cycle.
This concise derivation proves the mathematical equivalence be- The experimental data of Dieterich and Kilgore (1996) on the static
tween the constitutive framework (Eqs. (3)–(4)) and the classic rate- strength of bare contacts between transparent materials indicate a first-
and-state formulation (Eqs. (11)–(12)) upon the change of variable (Eq. order, linear dependence of the real area of contact on normal stress
(6)) and a truncated power-series expansion. From a practical stand- and a second-order, but systematic dependence on grain-size. Different
point, a power-law form is more adequate because the logarithmic form surfaces are prepared using grits of increasing number, which largely
is ill-posed for vanishing velocities (Bizzarri, 2011; Rice and Ben-Zion, eliminate the asperities above a fixed size. To model these observations
1996), and therefore not applicable for the full parameter range. with Eq. (1), I simply assume that grain-size is inversely proportional to
Chester (1989) showed that the velocity-strengthening behavior of grit number. The data of Dieterich and Kilgore (1996) is explained

133
S. Barbot Tectonophysics 765 (2019) 129–145

3.5
Data (D&K96) Grit # Model Grit #
60 60
100 100
3.0 240 240
Acrylic
Calcite
Glass
Quartz

2.5
Area of contact (%)

Acrylic
2.0

1.5

Calcite
1.0

Glass
0.5

Quartz
0
0 5 10 15 20 25 30
Normal stress (MPa)

Fig. 3. Effect of loading and grain-size on the real area of contact in stationary
contact. The experimental data of Dieterich and Kilgore (1996) on acrylic,
calcite, glass, and quartz (open and closed symbols) are explained by Eq. (1).
The profiles are for grit #60 (solid), #100 (long dashed), and #240 (short
dashed).
Fig. 4. Change of the area of contact during the seismic cycle. A) Effect of the
slip velocity for acrylic at 2.5 MPa normal stress from the data of Dieterich and
Table 1
Kilgore (1994) (black profile) and model Eq. (1) (red profile). B) Evolution of
Material properties controlling the dependence of the real area of contact of
friction (black profile) during the experiment shown in A). C) Area of contact
bare surfaces on confining pressure and grain size based on the experimental
during stationary hold for glass at 10 MPa (Dieterich and Kilgore, 1994) (black
findings of Dieterich and Kilgore (1996) and the physical model Eq. (1). A
profile) and the model Eq. (1). The area is relative to a reference value after 1 s
friction coefficient of μ0 = 0.6 is assumed for all experiments. †The indentation
of hold. (For interpretation of the references to color in this figure legend, the
hardness for these materials is measured independently on different samples
reader is referred to the web version of this article.)
(e.g., Brace, 1963; Dieterich and Kilgore, 1994; Dieterich and Kilgore, 1996;
Ferguson et al., 1987; Marsh, 1964).
Material Cohesion c Reference stress χ Exponent Indentation hardness
(MPa) (MPa) m/n (MPa)† state-dependence of friction and the logarithmic hardening with time at
static contact under constant normal load. The proposed formulation
Acrylic 1 460 0.14 400 explains the experimental data well (Fig. 5). In addition, the numerical
Calcite 4 1500 0.007 1800
simulation described in Appendix C indicates that the power-law terms
Glass 15 7300 0.04 5500
Quartz 5 10,000 0.04 12,000 and their logarithmic equivalent differ by less than 0.17% throughout
the seismic cycle.
The correlation between the characteristic weakening distance and
the gouge thickness found by Marone and Kilgore (1993) is predicted
(Fig. 3) assuming μ0 = 0.6 for all materials and the material properties based on Eq. (14) with λp/2 ≈ 100 to match their experimental data.
shown in Table 1. In all these cases, the parameter χ is found close to The characteristic weakening strain is therefore of the order of 0.5%.
the indentation hardness of these materials (Brace, 1963; Dieterich, The model of Eqs. (3)–(4) also predicts the independence of the char-
1979; Dieterich and Kilgore, 1994; Dieterich and Kilgore, 1996; acteristic weakening distance with grain-size and the evolution of grain-
Ferguson et al., 1987; Marsh, 1964) and the grain-size dependence is size during sliding, compatible with experiments (Marone and Kilgore,
weak, confirming the hypotheses of the model. 1993). Small grains do not make more unstable faults. Instead, the
The main features of the experimental data for velocity jumps for linear scaling (Eq. (14)) between the gouge thickness and the critical
acrylic at = 2.5 MPa normal stress (Dieterich and Kilgore, 1994) can weakening distance, which controls stability (Gu et al., 1984; Rice and
be explained with χ = 460 MPa, c = 0, μ0 = 0.76, and m/pn = 0.10 Ruina, 1983; Ruina, 1983), explains the stabilization of fault slip with
(Fig. 4A), except for some long-wavelength variations that are not gouge accumulation (Beeler et al., 1996; Engelder et al., 1975; Marone
captured by the model. The increase of the real area of contact with et al., 1990; Scholz et al., 1972). For bare surfaces, the increase in the
time for glass at 10 MPa during stationary hold (Dieterich and Kilgore, weakening distance with large slip over multiple slip cycles is compa-
1994) can be explained for μ0m/pn = 0.02 (Fig. 4C) (see Methods). tible with the removal of small-wavelength striations on the fault sur-
The mathematical equivalence between Eqs. (3)–(4) and (11)–(12) face (Voisin et al., 2007). As gouge is a product of wear and may always
is only valid upon a truncated power-series expansion, and one may be present, the two length scales emerging from surface roughness and
question the validity of this approximation. To address this concern, I gouge thickness may coexist and the characteristic weakening distance
compare the prediction of the multiplicative form Eq. (3) for steady- may scale with whichever is the largest, compatible with larger slip
state friction (see Appendix B) with slide-hold-slide experiments on weakening distance in fault gouge.
sandstone (Dieterich, 1972). These results were key to establish the The correlation between the static friction coefficient μ0 and the

134
S. Barbot Tectonophysics 765 (2019) 129–145

0.86 1.2

Data for sandstone from Dieterich (1972) Data Carpenter et al. (2016)
0.84 4.1 MPa 5.4 MPa 9.5 MPa and Ikari et al. (2016)
18.7 MPa 48.0 MPa San Gregorio
1.0 Fault
Constitutive model Multiplicative model with n=70
Alpine Fault
0.82 Calaveras
Fault

Rate-dependence parameter a (x10-2)


Westerly
0.80 0.8
Granite
Zuccale Fault
Coefficient of friction

Kodiak
0.78
Ghost Rocks
Biotite

0
0.6

7
Andesine

n=
0.76
Kaolinite Costa Rica
Decollement

0.74 0.4 Montmorillonite


Rochester Nankai
Shale Decollement

0.72 Talc

0.2
SAFOD
Core
0.70

0
0.68 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
100 101 102 103 104 105 106 Static friction parameter 0
Hold time (seconds)
Fig. 6. Rate-dependence parameter a in rate-and-state friction Eq. (11) for
Fig. 5. Increase of the static coefficient of friction in stationary contact for
natural and synthetic gouges (Carpenter et al., 2016; Ikari et al., 2016). The
quartz sandstone in slide-hold-slide tests (Dieterich, 1972) from a range of
relationship μ0 = n a of Eq. (13) explains the data in the least-squares sense for
normal stress from 2 to 48 MPa (symbols) and predictions from the multi-
n = 70 ± 4 (solid profile with slope 1/n; uncertainty is for one standard de-
plicative form Eq. (10) of rate-and-state friction (solid lines). The hold time
viation).
corresponds to the duration of stick under constant normal stress.

change in slip velocity (Chester and Higgs, 1992). Different effective


rate-dependence parameter a found for a wide range of natural and activation energies correspond to several micro-mechanisms of de-
synthetic gouges (Ikari et al., 2011; Ikari et al., 2016) indicates a formation, presumably acting at different pressures, shear stress, and
somewhat uniform n = 70 ± 4 for the velocity exponent for all these even temperatures.
materials (Fig. 6), although these data show some scatter. The high It is not possible to fully describe the thermally activated sliding
power exponent in the multiplicative form of Eq. (3) is compatible with behavior without resolving the shear heating effect (Chen and Rempel,
the asperity creep model that requires n ≥ 40 to explain aging at sta- 2015; De Lorenzo and Loddo, 2010; Kitajima et al., 2011; Noda and
tionary contact by viscoelastic flow (Mitchell et al., 2013). For re- Lapusta, 2010; Rempel and Weaver, 2008; Segall and Bradley, 2012;
ference, typical laboratory value of μ0 and a give rise to n = μ0/a = 60. Segall and Rice, 2006; Sleep, 1995; Sleep, 2019), but the sign of ΔQ
Using Arrhenius laws for grain growth (or equivalently, flattening) controls the importance of shear heating. For positive ΔQ, sliding is
and for the flow law, the model correctly predicts the modest change of temperature-hardening and stable (negative feedback). For negative
frictional resistance with varying sub-solidus temperatures (Chester and ΔQ, there is a potential for thermal instabilities, which are conditioned
Higgs, 1992; Chester, 1994; Stesky, 1978). The temperature depen- on the details of heat transport. Chester (1994) found separate values
dence of strain-rate in Eq. (3) and of grain growth at stationary contact for Q and H in dry and wet quartz gouge of the order of 89 ± 23 kJ/mol
in Eq. (4) implies a temperature control of static friction at fixed hold with effective activation energy ΔQ ranging from −30 kJ/mol to 21 kJ/
times controlled by an effective activation energy (see Appendix B and mol. Therefore, the temperature dependence may have an important
the work of Chester and Higgs, 1992) influence on fault dynamics due to thermal weakening, notwith-
standing the additional strong weakening mechanisms activated at high
b
Q= H Q. slip speed.
a (19)
The constitutive framework (Eqs. (3)–(4)) also predicts a realistic
The available data for sandstone (Nakatani, 2001) and Westerly dynamic range of grain-size in fault zones. Constraining the evolution of
granite (Mitchell et al., 2013) require an activation energy of grain-size in laboratory experiments still remains a challenge, but grain-
ΔQ = 36.3 kJ/mol and ΔQ = 16.9 kJ/mol, respectively, to explain the size can be estimated in numerical simulations. To explore this prop-
variation of static friction from room temperature to 900 °C (Fig. 7 and erty, I simulate the evolution of stress, velocity, grain-size, and contact
Supplementary Materials). As the multiplicative form of rate-and-state area in a spring-slider model (see Appendix C) with the parameters
friction finds the same form as proposed by Chester and Higgs (1992) listed in Table 2 using a Runge-Kutta numerical solver (Press et al.,
and Chester (1994) upon a Taylor series expansion, the model naturally 1992). With p = 3, the numerical simulation indicates a range of about
explains the experimental results of these studies, including that a three orders of magnitude of grain-sizes throughout the nucleation and
sudden change in temperature gives rise to a sudden change in the propagation of frictional instabilities and relocking (Fig. 8). Grain-sizes
frictional resistance to sliding followed by a transient, similar to, but in from 1 μm to 1 mm emerge with the characteristic rate of growth of
the opposite sense, as the change in friction produced by a sudden G = 10 μm3/s, which is theoretically possible around room temperature

135
S. Barbot Tectonophysics 765 (2019) 129–145

A 0.90 with an activation energy of H=160 kJ/mol. These results qualitatively


Data for Westerly granite agree with estimates and measurements of grain-size in fault cores
from Mitchell et al. (2013) (Chester and Logan, 1986; Frost et al., 2009; Sammis and Ben-Zion,
Arrhenius law with Q=16.9 kJ/mol 2008; Sammis and King, 2007; Yund et al., 1990).
0.85 Grain-size evolution offers an alternative interpretation of the sta-
bility of fault slip where the evolution of grain-size within the gouge is
Friction parameter

of primary importance. The stability of a spring-slider system governed


by Eqs. (3–4) is controlled by the stress exponents m and p, i.e., con-
0.80 ditionally stable (Ruina, 1983) for m > p and linearly stable for m < p.
The ratio p/m also seems to control the style and size of nucleation
(Rubin and Ampuero, 2005; Viesca, 2016). This implies that fault sta-
bility and overall dynamics are controlled by the grain-size dependence
0.75 of the real area of contact and of grain-growth. These results are
broadly compatible with the interpretation of Sleep (2006), who found
that steady-state frictional resistance is controlled by the aspect ratio of
the asperities bearing the load.
0.70 The grain-size evolution law proposed in Eq. (4) corresponds to the
0 100 200 300 400 500 600
Temperature (ºC)
aging-law of rate-and-state friction, upon the change of variable Eq. (6).
B 0.80 However, the aging law has been shown to only poorly predict the
symmetry of decelerating and accelerating velocity jumps in laboratory
experiments (e.g., Beeler et al., 1994; Nakatani, 2001). The slip law
Data for Na-feldspar gouge
from Nakatani (2001) (hold time=3162 s)
predicts this behavior more accurately but fails to capture the time-
Arrhenius law with Q=36.3 kJ/mol
dependent healing at stationary contact. Some formulations have ad-
0.75

dressed this issue (e.g., Kato and Tullis, 2001; Nagata et al., 2012;
Sleep, 2012). In principle, any evolution law can be directly translated
Friction parameter

0.70 to a revised grain-size evolution law by applying the change of variable


of Eq. (6).

0.65 5. Implications for fault dynamics

Using a power-law for the velocity dependence of friction as in Eq.


0.60 (3) is more realistic for vanishing velocities than the alternative form of
Eq. (11), as the logarithm of velocity predicts negative shear stress for
sufficiently low slip speeds, and even infinite negative stress for truly
0.55 stationary contact. Several regularization forms have been proposed to
0 100 200 300 400 500 600 700 800 900 work around this issue (see the review by Bizzarri, 2011) to model
Temperature (ºC) deformation on weak faults, e.g., sub-horizontal décollements in ac-
Fig. 7. Effect of temperature. A) Friction coefficient μ0 for Westerly granite cretionary prisms or in thrust and fold belts (Hubbard et al., 2015; Ong
from slide-hold-slide experiments (Mitchell et al., 2013) and the prediction et al., 2019), or to incorporate strong weakening mechanisms (e.g.,
from multiplicative model with the effective activation energy ΔQ = 16.9 kJ/ Karato and Barbot, 2018; Noda and Lapusta, 2010; Noda et al., 2009).
mol (solid profile). B) Static friction coefficient μ0 for Na-feldspar gouge from The proposed constitutive framework does not require additional
similar experiments (Nakatani, 2001) and model with ΔQ = 36.3 kJ/mol regularization as the power-law relationship is well-behaved at van-
(Appendix B). ishing velocities. To explore the implications of this form of regular-
ization, I explore the dynamics of faults of decreasing strength in quasi-
dynamic models of the seismic cycle with the radiation-damping ap-
proximation (Barbot, 2018; Goswami and Barbot, 2018; Horowitz and
Table 2 Ruina, 1989; Kato, 2003; Lambert and Barbot, 2016; Li and Liu, 2016;
Physical parameters used to simulate seismic cycles with a spring-slider model Liu and Rice, 2005; Qiu et al., 2016; Rice, 1993; Tse and Rice, 1986; Wu
(Fig. 8). The model predicts effective grain-sizes between 1 μm and 1 mm and a and Chen, 2014). The model consists of a single vertical strike-slip fault
real area of contact in the range of 0.5–0.6% for a dynamic range of velocity embedded in an elastic half-space with the model parameters listed in
between 10−12 and 1 m/s. the caption of Fig. 9. I have checked numerical convergence and proper
Parameter Symbol Value Unit discretization, which requires >20 samples within the critical nuclea-
tion size. Although, the effect of static friction is typically ignored in
Rigidity G 30 GPa
studies of fault stability (e.g., Rice and Ruina, 1983; Rubin and
Velocity power exponent n 60
State variable exponent m 3 Ampuero, 2005; Ruina, 1983; Viesca, 2016), these numerical simula-
Activation energy Q 400 kJ/mol tions indicate that the seismic cycle becomes increasingly complex with
Temperature T 0 °C decreasing static fault strength (i.e., decreasing value of μ0), every other
Reference velocity V0 10−6 m/s
parameters being the same (Fig. 9). Considering frictional parameters
Reference length scale W 10−2 m
Reference friction coefficient μ0 0.6
that produce simple cycles of elasto-dynamic ruptures on strong faults
Effective confining pressure 100 MPa (Fig. 9A, B), partial ruptures of the seismogenic zone and multi-modal
Cohesion c 0 MPa source-time functions emerge spontaneously on weak faults
Activation enthalpy H 160 kJ/mol (Fig. 9C–H). This is reminiscent of the complex dynamics found in the
Grain-size evolution exponent p 3
accretionary prism of subduction zones (Carlos Villegas-Lanza et al.,
Growth parameter G0 10 μm3/s
Reference strain 1/λ 10−2 2016; Dixon et al., 2014; Jiang et al., 2012; Wallace et al., 2016;
Wallace and Beavan, 2006; Wallace et al., 2017), where shortening is
accommodated by a low-dipping, low-strength décollement (Hubbard

136
S. Barbot Tectonophysics 765 (2019) 129–145

A 60 B

55
Stress (MPa)
50

45

40
0 10 20 30 40 50 60 70 80 90 100 0.45 0.5 0.55 0.6 0.65
Contact area (%)
C D
100
Velocity (m/s)

10-3
10-6
loading
10-9
rate
10-12

0 10 20 30 40 50 60 70 80 90 100 0.45 0.5 0.55 0.6 0.65


Contact area (%)
E F
2
1 mm
Grain-size (mm)

100
h
wit
10-1 th /3
1 ow e1

gr
10-2 tim
1 micron
10-3

10-4 0
0 10 20 30 40 50 60 70 80 90 100 20 25 30 35 40
G Time (yr)
H
Density of contact area (%)

0.7
logarithmic
0.6 growth

0.5

0.4
0 10 20 30 40 50 60 70 80 90 100 20 25 30 35 40
Time (yr) Time (yr)

Fig. 8. Seismic-cycle simulation in a spring-slider block with dynamic recrystallization assuming μ0 = 0.6, = 100 MPa, a = 10−2, a − b = − 4 × 10−3, L = 1 cm,
G = 10 μm3/s, and λ = 100. A) Shear stress asa function of time. The rapid stress drops are due to frictional instabilities. B) Stress as a function of contact area. The
arrow indicates the direction of time evolution. C) Slip velocity as a function of time. The peaks are due to frictional instabilities. The dashed line indicates the
loading rate of 10−9 m/s. Velocities significantly below this rate correspond to fault locking. D) Phase diagram of slip velocity as a function of contact area. The arrow
indicates the time evolution. E) Grain-size as a function of time (log-linear plot). F) Grain-size as a function of time (linear plot) showing growth with the cubic-root of
time in the inter-seismic period. G) Real area of contact density as a function of time (log-linear plot). H) Real area of contact density (linear plot) indicating
logarithmic increase during the inter-seismic period.

et al., 2015; Ong et al., 2019). Further investigation of fault dynamics 6. Conclusions
on weak faults is warranted.
The regularization that involves replacing the exp (x) terms with Consideration of the relationship between fault strength and the real
sinh (x) terms in the friction law (Rice and Ben-Zion, 1996) assumes an area of contact, and of the weak modulation of the area of contact
exponentially decaying probability of backward slip at contact junc- junctions by grain-size evolution gives rise to the following multi-
tions, giving rise to a linearization of the stress-velocity relationship at plicative, thermally activated form of rate-and-state friction
small velocities. The complex sequence of partial ruptures (Fig. 9C–H) µ0
b
obtained with the multiplicative form Eq. (2) does not appear with the
a
V0 a Q 1 1
V = V0 exp
regularization form of Rice and Ben-Zion (1996), indicating that the c + µ0 L R T T0 (20)
form of regularization may play a more important role in fault dy-
namics than previously thought. Further experimental work is needed with the evolution law
to validate these candidate expressions. Relaxation tests (e.g., Chester H 1 1 V
and Higgs, 1992; Higgs, 1981; Rutter and Mainprice, 1978), although = exp ,
R T T0 L (21)
less popular, seem most adequate for this task because they can achieve
low deformation rates. that captures the temperature dependence of fault strength and reg-
ularizes the constitutive behavior near vanishing slip velocity.

137
S. Barbot Tectonophysics 765 (2019) 129–145

Slip velocity (m/s)

10-12 10-9 10-6 10-1

Moment rate (x1017 Nm/s)


A 0 B 20
0
= 0.6 (strong fault) 0
= 0.6 (strong fault)
5
Depth (km)

nucleation single-peaked
10 10 source-time
propagation function
15
termination
20 0
0 5 10 15 20 25 30 0 5 10 15 20 25 30
Time (s) Time (s)

Moment rate (x1017 Nm/s)


C 0 D 10
0
= 0.2 0
= 0.2
5 8 single-peaked
Depth (km)

nucleation source-time
6
10 function
propagation 4 of increasing
15 complexity
termination 2
20 0
0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35
Time (s) Moment rate (x1017 Nm/s) Time (s)
E 0 F 1
0
= 0.1 (weak fault) 0
= 0.1 (weak fault)
5
Depth (km)

secondary
moment multi-modal
10
release source-time
function
15

20 0
0 50 100 150 200 250 300 0 50 100 150 200 250 300
Time (s) Time (s)
Moment rate (x1017 Nm/s)

G 0 H 2
0
= 0.05 (very weak fault) 0
= 0.05 (very weak fault)
5
Depth (km)

partial
10 rupture 1 long-tailed
source-time
15 function

20 0
0 5 10 15 20 25 30 0 5 10 15 20 25 30
Time (s) Time (s)
Fig. 9. Effect of fault strength on the dynamics of fault slip. The strike-slip fault is assumed infinite in the strike direction with a = 10−2, L = 1 cm, = 100 MPa,
V0 = 10−6 m, and Vs = 3 km/s. In the velocity-weakening region, between 5 and 15 km depth, we have b − a = 4 × 10−3; in the velocity-strengthening region
b − a = − 4 × 10−3. The sampling size is 35 m. The fault is loaded at the background rate VL=10−9 m/s. The following snapshots are provided for the fifth event in
the seismic cycle. A) Evolution of slip velocity on a strong fault (μ0 = 0.6) during the nucleation, propagation, and arrest of a seismic rupture. B) Corresponding
source-time function, assuming a length of 10 km. C) Evolution of slip velocity for μ0 = 0.2. D) Corresponding source-time function with increasing complexity. E)
Evolution of slip velocity for μ0 = 0.1. F) Multi-modal source-time function. G) Slip velocity for a partial rupture breaking the top of the seismogenic zone of a very
weak fault with μ0 = 0.05. A previous, slow-slip event ruptures the down-dip section. H) Corresponding long-tailed source-time function.

The physical model proposed here facilitates the upscaling of la- rheology and friction and may simplify the treatment of the brittle-
boratory data by providing an explicit relationship for the frictional ductile transition (e.g., Chester, 1989; Noda and Shimamoto, 2010;
parameters a, b, and L with micro-mechanical properties. The possible Reber et al., 2015; Shimamoto and Noda, 2014).
discrepancy in the characteristic weakening distance measured in the Grain-size evolution around contact junctions offers new insight
laboratory and inferred seismologically (Cocco et al., 2009; Marone, into the mechanics of frictional slip and provides a physical basis to a
1998) could be elucidated by the difference between the gouge thick- constitutive framework with already extensive experimental support.
ness and fault roughness in the laboratory versus in nature (Fagereng However, it does not address well-known shortcomings, including the
and Sibson, 2010; Sibson, 2003). The increasing thickness of fault zones neglect of off-fault deformation (Bhat et al., 2007; Griffith et al., 2010;
with depth below the seismogenic zone may enable slow-slip events by Sammis et al., 2009) and thermally activated processes that take
stable weakening near the brittle-ductile transition zone (Goswami and dominance at high slip-velocity (Ferri et al., 2010; Han et al., 2007;
Barbot, 2018; Sone et al., 2012; Ujiie et al., 2018). Toro et al., 2004), and uncertainties regarding the use of one or more
Writing the constitutive behavior of faults as a nonlinear viscoe- state variables. The constitutive framework should also be refined to
lastic process draws a strong connection between the worlds of afford a consistent treatment of near-solidus temperatures (Blanpied

138
S. Barbot Tectonophysics 765 (2019) 129–145

et al., 1995; Brown and Fialko, 2012; Mitchell et al., 2013; Stesky et al., Acknowledgments
1974), dilatancy (Marone et al., 1990; van den Ende et al., 2018), and
anisotropy (Power et al., 1987; Sleep, 1998). I would like to thank Dr. Kelin Wang and Dr. Eiichi Fukuyama for
With further integration of laboratory, in situ, and remote-sensing their editorial work on the special issue. The manuscript benefitted
observations into physical models of rock deformation, we may even- from the insightful reviews of Dr. Norman Sleep and an anonymous
tually elucidate the earthquake phenomenon. reviewer, and from discussions with Dr. Emily Brodsky.

Appendix A. Physical model for rate-and-state friction

In this Appendix, I provide a more detailed derivation of the equivalence between Eqs. (3)–(4) and (11)–(12). We start with
n m
d Q
= 0 exp ,
c + µ0 d0 RT (A1)
and
G0 H
d= exp d,
pd p 1 RT (A2)
or equivalently, using Eq. (5),
G
d= d,
pd p 1
(A3)
where the parameters are described in the main text. At stationary contact, the grain growth follows
d (t ) p = d1 p + G t , (A4)
where t is time and d1 is a starting grain-size. This serves as a motivation to define
dp
=
G (A5)
as a state variable representing the age of the grain, or, in other words, the approximate duration of grain growth. Introducing the change of variable
Eq. (A5) into Eq. (A2), we obtain
1 dp
p d dp 1 =1 p ,
G G (A6)
or, simply,
=1 p . (A7)
Assuming a uniform strain-rate in the fault normal direction leads to
V
= ,
2W (A8)
where V is the slip velocity across the shear zone. Combining Eqs. (A7) and (A8), we obtain
p
=1 V .
2W (A9)
Writing the characteristic weakening distance as
2W
L= ,
p (A10)
we finally obtain
V
=1 ,
L (A11)
i.e., the aging law of Dieterich (1979). Eq. (A11) indicates that the age of the grain in Eq. (A5) and the age of contact defined by Dieterich (1979) are
one and the same. I now consider the rate-and-state formulation of Eq. (A1), recast as
1 m
V n
d n
= (c + µ 0 ) ,
V0 d0 (A12)
where
V Q
= exp .
V0 0 RT (A13)
y
Expanding the powers following x = exp (y ln x), we obtain

1 V m d
= (c + µ 0 ) exp ln + ln .
n V0 n d0 (A14)
The logarithmic dependence on velocity and grain-size combined with the small value of the pre-factors, i.e., n ≫ 1 and n ≫ ∣ m∣, allows us to consider

139
S. Barbot Tectonophysics 765 (2019) 129–145

a power series expansion of Eq. (A14). Using the Taylor series expansion exp (x) = 1 + x + O(x2), ignoring the cohesion for simplicity, we obtain

1 V m d
µ0 1+ ln + ln .
n V0 n d0 (A15)
Using Eqs. (A5), (A15) and
GL 2GW
d0 p = = ,
V0 p V0 (A16)
I obtain a familiar expression of frictional resistance with the static, rate-dependent, and state-dependent terms

1 V m V0
µ0 1+ ln + ln .
n V0 pn L (A17)
−14 3 −6
For reference, with p = 3, G = 10 m /s, V0 = 10 m/s, and L = 0.1 μm, I obtain d0 = 100 μm, which is representative grain-size for fault cores
(Chester and Logan, 1986; Frost et al., 2009). By identification with the rate-and-state friction formulation

V V0
= µ 0 + a ln + b ln ,
V0 L (A18)
an expression for the rate-and-state parameters can be obtained
1
a = µ0 ,
n
m m
b = µ0 = a,
pn p
2W
L= .
p (A19)
Following the same approach, the real area of contact
m
c + µ0 d n
A= ,
d0 (A20)
can be approximated, neglecting cohesion, by
µ0 m d
A= 1+ ln .
n d0 (A21)
The ratio m/n and the material properties χ for acrylic, glass, calcite, and quartz are constrained by the experiments of Dieterich and Kilgore (1996)
(see Table 1). Using Eqs. (A5), (A16), and (A19), we find

µ0 b V0
A= 1+ ln ,
µ0 L (A22)
compatible with the expression proposed by Ronsin and Coeyrehourcq (2001).
To maintain the use of parameters that have been classically used and discussed in multiple studies, we can formulate rate-and-state friction in
multiplicative form as follows
µ0
a
b
V0 a
V = V0 ,
c + µ0 L (A23)
or, equivalently,
a b
V µ0
V0 µ0
= (c + µ 0 ) ,
V0 L (A24)
coupled with the evolution law
V
=1 .
L (A25)
This formulation features a regularization at vanishing slip velocity without further modification.

Appendix B. Temperature dependence of static friction

I look at the effect of temperature on grain-growth to explain the temperature dependence of static friction found in laboratory experiments
(Chester and Higgs, 1992; Chester, 1994; Mitchell et al., 2013; Nakatani, 2001). Let us start with the temperature-dependent constitutive law
n m
d Q
= 0 exp ,
µ0 d0 RT (B1)
and the grain-size evolution law

140
S. Barbot Tectonophysics 765 (2019) 129–145

G0 H
d= exp d.
pd p 1 RT (B2)
The Arrhenius law in the first term on the right-hand side of Eq. (B2) captures the thermal activation of grain-growth. To go from strain-rate to
velocity, I define
V = 2W , (B3)
and

Q
V0 = 2W 0 exp ,
RT0 (B4)
where T0 is the temperature at which the friction is μ0 for the stable velocity of sliding V0. The multiplicative form of the friction law becomes
1 m
V n
d n
Q 1 1
µ = µ0 exp .
V0 d0 nR T T0 (B5)
With the previously defined constants a = μ0/n, b = μ0m/pn, and d 0 = G0 L / V0 , the truncated power-series expansion of Eq. (B5) becomes
p

V Q 1 1 V0
µ = µ 0 + a ln + + b ln ,
V0 R T T0 L (B6)
where I have assumed 0 < Q/nRT ≪ 1.
To make the dependence of temperature explicit in the state evolution law, I consider the change of variable
dp
= .
G0 exp ( ) H
RT0 (B7)
Instead of the aging law for the state variable, I get the modified evolution law

H 1 1 V
= exp ,
R T T0 L (B8)
with the constant L = 2T/pλ. Temperature affects the rate-and-state constitutive relationship Eq. (B6) and the state evolution law Eq. (B8). The pair
Eqs. (B6) and (B8) corresponds to the temperature-dependent constitutive framework of Chester and Higgs (1992) and Chester (1994).
I consider the temperature dependence of friction at steady-state first and then during stationary contact. At steady-state, we have the state
variable

L H 1 1
ss = exp
V R T T0 (B9)
and the steady-state friction reads

Q Hb/a 1 1 V
µss = µ0 + a + (a b) ln .
R T T0 V0 (B10)
If the steady-state friction is evaluated at V = V0, it appears temperature-dependent with the apparent activation energy
b
Q= H Q.
a (B11)
For slide-hold-slide experiments, the weakening terms in Eq. (B2) can be neglected during stationary contact, leading to
G0 H
d= exp .
pd p 1 RT (B12)
Upon integration, we find
H
d (t ) p = d1 p + G0 exp t,
RT (B13)
where d1 is a reference value at t = 0, subsequently ignored. Looking asymptotically at large time t, the friction evolution becomes
1 m
pn
Q m H V n
G0 t
µ (t ) = µ 0 exp .
nRT pn RT V0 d 02 (B14)
Using the truncated power series expansion with b = μ0m/pn, I get
µ µeffective (T ) + b ln(t ), (B15)
showing that the rate of healing does not change with temperature and that the static friction appears temperature dependent, following
Q H b/a
µeffective (T ) = µ0 + a ,
RT (B16)
b
where some constants independent of temperature have been neglected for clarity. The effective activation energy is also Q = a
Q . Assuming
H
a = 0.01, the laboratory data of Mitchell et al. (2013) for the temperature dependence of Westerly granite can be explained for ΔQ = 16.9 kJ/mol;

141
S. Barbot Tectonophysics 765 (2019) 129–145

the laboratory data of Nakatani (2001) of Na-feldspar for ΔQ = 36.3 kJ/mol (Fig. 7).
In simulations of the seismic cycle with the additive form of Eq. (11) of rate-and-state friction where the static friction coefficient is constant, its
value does not affect the dynamics, i.e., the values taken by the state variable and the velocity. The static fiction controls the absolute level of shear
stress, but not its dynamic range. If the temperature is kept constant, it has no effect on the dynamics. However, if the temperature is allowed to vary,
there will be a negative feedback between shear heating and fault hardening by temperature increase if ΔQ is positive, and a positive feedback
between shear heating and fault weakening by increasing temperature if ΔQ is negative.

Appendix C. Block modeling with a spring-slider

Simple models of frictional instabilities in a spring-slider block at constant temperature can be obtained by coupling the multiplicative form of
rate-and-state friction
a b
V µ0
V0 µ0
= µ0 ,
V0 L (C1)
the evolution law
V
=1 ,
L (C2)
and the equation for the elastic interaction between the spring and the slider
= k (V VL ) V (C3)
where k is the spring stiffness, VL is the imposed loading rate, and α is a damping coefficient. The acceleration can be obtained with
µ
V k (V VL) µ0
b /
= µ .
V a + V
µ0 (C4)
Time-series of the state vector y = (τ, θ, V) can then be obtained numerically using the Runge-Kutta method (Press et al., 1992). The grain-size and
b
contact area can be obtained using d = (Gθ)1/p and A = µ 0 ( V0/L) µ0 / , respectively. An example is shown in Fig. 8.
The numerical simulation in Fig. 8 with the parameters listed in Table 2 can be used to investigate the range of values taken by the arguments of
the exponential function in Eq. (A14) to justify the approximation exp (x) = 1 + x + O(x2). The values taken by the arguments of the exponential
function vary in the range of ±6%. As a result, the exponential function and its Taylor expansion differ by less than 0.17%.

References frictional strength. Nature 463 (7277), 76.


Bennett, A.I., Harris, K.L., Schulze, K.D., Urueña, J.M., McGhee, A.J., Pitenis, A.A., Müser,
M.H., Angelini, T.E., Sawyer, W.G., 2017. Contact measurements of randomly rough
Agata, R., Barbot, S.D., Fujita, K., Hyodo, M., Iinuma, T., Nakata, R., Ichimura, T., Hori, surfaces. Tribol. Lett. 65 (4), 134.
T., 2019. Rapid mantle flow with power-law creep explains deformation after the Berthoud, P., Baumberger, T., G'sell, C., Hiver, J.-M., 1999. Physical analysis of the state-
2011 Tohoku mega-quake. Nat. Commun. 10 (1), 1385. and rate-dependent friction law: static friction. Phys. Rev. B 59 (22), 14,313.
Aharonov, E., Scholz, C.H., 2018. A physics-based rock friction constitutive law: steady Bhat, H.S., Olives, M., Dmowska, R., Rice, J.R., 2007. Role of fault branches in earthquake
state friction. J. Geophys. Res. 123 (2), 1591–1614. rupture dynamics. J. Geophys. Res. 112 (B11309), 16. https://doi.org/10.1029/
Amontons, G., 1699. De la resistance causée dans les machines, tant par les frottemens des 2007JB005027.
parties qui les composent, que par roideur des cordes qu'on y employe, & la maniere Biegel, R.L., Sammis, C.G., 2004. Relating fault mechanics to fault zone structure. Adv.
de calculer l'un & l'autre. In: Mémoires de l'Académie Royale Des Science, pp. Geophys. 47, 65–111.
206–222. Biegel, R.L., Sammis, C.G., Dieterich, J.H., 1989. The frictional properties of a simulated
Archard, J.F., 1957. Elastic deformation and the laws of friction. Proc. R. Soc. Lond. A 243 gouge having a fractal particle distribution. J. Struct. Geol. 11 (7), 827–846.
(1233), 190–205. Bizzarri, A., 2011. On the deterministic description of earthquakes. Rev. Geophys. 49
Atkinson, H., 1988. Overview no. 65: theories of normal grain growth in pure single phase (RG3002). https://doi.org/10.1029/2011RG000356.
systems. Acta Metall. 36 (3), 469–491. Blanpied, M.L., Lockner, D.A., Byerlee, J.D., 1991. Fault stability inferred from granite
Badt, N., Hatzor, Y.H., Toussaint, R., Sagy, A., 2016. Geometrical evolution of interlocked sliding experiments at hydrothermal conditions. Geophys. Res. Lett. 18 (4), 609–612.
rough slip surfaces: the role of normal stress. Earth Planet. Sci. Lett. 443, 153–161. Blanpied, M.L., Lockner, D.A., Byerlee, J.D., 1995. Frictional slip of granite at hydro-
Barbot, S., 2018. Asthenosphere flow modulated by megathrust earthquake cycles. thermal conditions. J. Geophys. Res. 100 (B7), 13,045–13,064.
Geophys. Res. Lett. 45, 6018–6031. https://doi.org/10.1029/2018GL078197. Bowden, F.P., Tabor, D., 1964. The Friction and Lubrication of Solids, Part II. Clarendon
Barbot, S., Fialko, Y., 2010. A unified continuum representation of postseismic relaxation Press, Oxford.
mechanisms: semi-analytic models of afterslip, poroelastic rebound and viscoelastic Brace, W., 1960. Behavior of rock salt, limestone, and anhydrite during indentation. J.
flow. Geophys. J. Int. 182 (3), 1124–1140. https://doi.org/10.1111/j.1365-246X. Geophys. Res. 65 (6), 1773–1788.
2010.04678.x. Brace, W., 1963. Behavior of quartz during indentation. J. Geol. 71 (5), 581–595.
Barbot, S., Fialko, Y., Bock, Y., 2009. Postseismic deformation due to the Mw 6.0 2004 Brace, W., Byerlee, J., 1966. Stick-slip as a mechanism for earthquakes. Science 153
Parkfield earthquake: stress-driven creep on a fault with spatially variable rate-and- (3739), 990–992.
state friction parameters. J. Geophys. Res. 114 (B07405). https://doi.org/10.1029/ Brown, K.M., Fialko, Y., 2012. ‘Melt welt’ mechanism of extreme weakening of gabbro at
2008JB005748. seismic slip rates. Nature 488 (7413), 638.
Barbot, S., Lapusta, N., Avouac, J.P., 2012. Under the hood of the earthquake machine: Bürgmann, R., Ergintav, S., Segall, P., Hearn, E.H., McClusky, S., Reilinger, R.E., Woith,
towards predictive modeling of the seismic cycle. Science 336 (6082), 707–710. H., Zschau, J., 2002. Time-dependent distributed afterslip on and deep below the
Baumberger, T., Caroli, C., 2006. Solid friction from stick-slip down to pinning and aging. Izmit earthquake rupture. Bull. Seismol. Soc. Am. 92 (1), 126–137.
Adv. Phys. 55 (3–4), 279–348. Bush, A., Gibson, R., Thomas, T., 1975. The elastic contact of a rough surface. Wear 35
Baumberger, T., Berthoud, P., Caroli, C., 1997. Physical analysis of the state- and rate- (1), 87–111.
dependent friction law. II. Dynamic friction. Phys. Rev. B 60 (6), 3928–3939. Byerlee, J., 1978. Friction of rock. Pure Appl. Geophys. 116, 615–626.
Beeler, N.M., Tullis, T.E., Weeks, J.D., 1994. The roles of time and displacement in the Byerlee, J., Savage, J., 1992. Coulomb plasticity within the fault zone. Geophys. Res. Lett.
evolution effect in rock friction. Geophys. Res. Lett. 21, 1987–1990. 19 (23), 2341–2344.
Beeler, N.M., Tullis, T.E., Blanpied, M.L., Weeks, J.D., 1996. Frictional behavior of large Candela, T., Brodsky, E.E., 2016. The minimum scale of grooving on faults. Geology 44
displacement experimental faults. J. Geophys. Res. 101 (B4), 8697–8715. (8), 603–606.
Beeler, N.M., Tullis, T.E., Goldsby, D.L., 2008. Constitutive relationship and physical basis Candela, T., Renard, F., Klinger, Y., Mair, K., Schmittbuhl, J., Brodsky, E.E., 2012.
of fault strength due to flash heating. J. Geophys. Res. 113 (B01401). https://doi.org/ Roughness of fault surfaces over nine decades of length scales. J. Geophys. Res.
10.1029/2007JB004988. 117 (B8).
Ben-David, O., Rubinstein, S.M., Fineberg, J., 2010. Slip-stick and the evolution of Carlos Villegas-Lanza, J., Nocquet, J.-M., Rolandone, F., Vallée, M., Tavera, H., Bondoux,

142
S. Barbot Tectonophysics 765 (2019) 129–145

F., Tran, T., Martin, X., Chlieh, M., 2016. A mixed seismic–aseismic stress release with implications for rate-and state-variable friction laws relevant to earthquake
episode in the Andean subduction zone. Nat. Geosci. 9 (2), 150–154. mechanics. J. Mater. Res. 19 (1), 357–365.
Carpenter, B., Ikari, M., Marone, C., 2016. Laboratory observations of time-dependent Goswami, A., Barbot, S., 2018. Slow-slip events in semi-brittle serpentinite fault zones.
frictional strengthening and stress relaxation in natural and synthetic fault gouges. J. Sci. Rep. 8 (1), 6181.
Geophys. Res. 121 (2), 1183–1201. Greenwood, J.A., Williamson, J., 1966. Contact of nominally flat surfaces. Proc. R. Soc.
Chen, J., Rempel, A., 2015. Shear zone broadening controlled by thermal pressurization London, Ser. A 295 (1442), 300–319.
and poroelastic effects during model earthquakes. J. Geophys. Res. 120 (7), Griffith, W.A., Nielsen, S., Di Toro, G., Smith, S.A., 2010. Rough faults, distributed
5215–5237. weakening, and off-fault deformation. J. Geophys. Res. 115 (B8).
Chen, J., Spiers, C.J., 2016. Rate and state frictional and healing behavior of carbonate Gu, J., Rice, J.R., Ruina, A.L., Tse, S.T., 1984. Slip motion and stability of a single degree
fault gouge explained using microphysical model. J. Geophys. Res. 121 (12), of freedom elastic system rate and state dependent friction. J. Mech. Phys. Sol. 32,
8642–8665. 167–196.
Chen, J., Verberne, B.A., Spiers, C.J., 2015. Effects of healing on the seismogenic potential Hall, C.E., Parmentier, E., 2003. Influence of grain size evolution on convective in-
of carbonate fault rocks: experiments on samples from the Longmenshan Fault, stability. Geochem. Geophys. Geosyst. 4 (3).
Sichuan, China. J. Geophys. Res. Solid Earth 120 (8), 5479–5506. Han, R., Shimamoto, T., Hirose, T., Ree, J.-H., Ando, J., 2007. Ultralow friction of car-
Chester, F., Higgs, N., 1992. Multimechanism friction constitutive model for ultrafine bonate faults caused by thermal decomposition. Science 316. https://doi.org/10.
quartz gouge at hypocentral conditions. J. Geophys. Res. 97 (B2), 1859–1870. 1126/science.1139763.
Chester, F., Logan, J.M., 1986. Implications for mechanical properties of brittle faults Harbord, C.W., Nielsen, S.B., De Paola, N., Holdsworth, R.E., 2017. Earthquake nuclea-
from observations of the Punchbowl fault zone, California. Pure Appl. Geophys. 124 tion on rough faults. Geology 45 (10), 931–934.
(1–2), 79–106. Heslot, F., Baumberger, T., Perrin, B., Caroli, B., Caroli, C., 1994. Creep, stick-slip, and
Chester, F.M., 1989. Dynamic recrystallization in semi-brittle faults. J. Struct. Geol. 11 dry friction dynamics: experiments and a heuristic model. Phys. Rev. E 49 (6),
(7), 847–858. 4973–4988.
Chester, F.M., 1994. Effects of temperature on friction: constitutive equations and ex- Higgs, N.G., 1981. Mechanical Properties of Ultrafine Quartz, Chlorite and Bentonite in
periments with fault gouge. J. Geophys. Res. 99 (B4), 7247–7261. Environments Appropriate to Upper-Crustal Earthquakes. 267 pp. Texas A &M
Chester, F.M., 1995. A rheologic model for wet crust applied to strike-slip faults. J. Univ., College Station, TX.
Geophys. Res. 100 (B7), 13,033–13,044. Hori, T., Kato, N., Hirahara, K., Baba, T., Kaneda, Y., 2004. A numerical simulation of
Ciavarella, M., Joe, J., Papangelo, A., Barber, J., 2019. The role of adhesion in contact earthquake cycles along the Nankai trough in southwest Japan: lateral variation in
mechanics. J. R. Soc. Interface 16 (151) 20180,738. frictional property due to the slab geometry controls the nucleation position. Earth
Cocco, M., Tinti, E., Marone, C., Piatanesi, A., 2009. Scaling of slip weakening distance Planet. Sci. Lett. 228 (3–4), 215–226.
with final slip during dynamic earthquake rupture. Int. Geophys. 94, 163–186. Horowitz, F.G., Ruina, A., 1989. Slip patterns in a spatially homogeneous fault model. J.
Coulomb, C., 1785. The theory of simple machines. Mem. Math. Phys. Acad. Sci. 10 Geophys. Res. 94 (B8), 10,279–10,298.
(161–331), 4. Hsu, Y.-J., Simons, M., Avouac, J.-P., Galetzka, J., Sieh, K., Chlieh, M., Natawidjaja, D.,
Daub, E.G., Carlson, J.M., 2010. Friction, fracture, and earthquakes. Annu. Rev. Condens. Prawirodirdjo, L., Bock, Y., 2006. Friction afterslip following the 2005 Nias-Simeulue
Matter Phys. 1 (1), 397–418. earthquake, Sumatra. Science 312, 1921–1926.
De Bresser, J., Peach, C., Reijs, J., Spiers, C., 1998. On dynamic recrystallization during Hubbard, J., Barbot, S., Hill, E.M., Tapponnier, P., 2015. Coseismic slip on shallow
solid state flow: effects of stress and temperature. Geophys. Res. Lett. 25 (18), décollement megathrusts: implications for seismic and tsunami hazard. Earth Sci.
3457–3460. Rev. 141, 45–55.
De Bresser, J., Ter Heege, J., Spiers, C., 2001. Grain size reduction by dynamic re- Hulikal, S., Bhattacharya, K., Lapusta, N., 2015. Collective behavior of viscoelastic as-
crystallization: can it result in major rheological weakening? Int. J. Earth Sci. 90 (1), perities as a model for static and kinetic friction. J. Mech. Phys. Solids 76, 144–161.
28–45. Ikari, M.J., Marone, C., Saffer, D.M., 2011. On the relation between fault strength and
De Lorenzo, S., Loddo, M., 2010. Effect of frictional heating and thermal advection on pre- frictional stability. Geology 39 (1), 83–86.
seismic sliding: a numerical simulation using a rate-, state- and temperature-depen- Ikari, M.J., Carpenter, B.M., Marone, C., 2016. A microphysical interpretation of rate-and
dent friction law. J. Geodyn. 49 (1), 1–13. state-dependent friction for fault gouge. Geochem. Geophys. Geosyst. 17 (5),
den Hartog, S.A., Spiers, C.J., 2014. A microphysical model for fault gouge friction ap- 1660–1677.
plied to subduction megathrusts. J. Geophys. Res. 119 (2), 1510–1529. Jiang, Y., Wdowinski, S., Dixon, T.H., Hackl, M., Protti, M., Gonzalez, V., 2012. Slow slip
Dieterich, J.H., 1972. Time-dependent friction in rocks. J. Geophys. Res. 77, 3690–3697. events in Costa Rica detected by continuous GPS observations, 2002–2011. Geochem.
Dieterich, J.H., 1978. Time-dependent friction and the mechanics of stick-slip. Pure Appl. Geophys. Geosyst. 13 (4).
Geophys. 116 (4–5), 790–806. Kaneko, Y., Nielsen, S.B., Carpenter, B.M., 2016. The onset of laboratory earthquakes
Dieterich, J.H., 1979. Modeling of rock friction 1. Experimental results and constitutive explained by nucleating rupture on a rate-and-state fault. J. Geophys. Res. 121 (8),
equations. J. Geophys. Res. 84 (B5), 2161–2168. 6071–6091.
Dieterich, J.H., Conrad, G., 1984. Effect of humidity on time-and velocity-dependent Karato, S.-I., Barbot, S., 2018. Dynamics of fault motion and the origin of contrasting
friction in rocks. J. Geophys. Res. 89 (B6), 4196–4202. tectonic style between Earth and Venus. Sci. Rep. 8 (1), 11 884.
Dieterich, J.H., Kilgore, B.D., 1994. Direct observation of frictional contacts: new insights Kato, N., 2003. Repeating slip events at a circular asperity: Numerical simulation with a
for sliding memory effects. Pure Appl. Geophys. 143, 283–302. rate-and state-dependent friction law. Bull. Earthq. Res. Inst., Univ. Tokyo 78,
Dieterich, J.H., Kilgore, B.D., 1996. Imaging surface contacts: power law contact dis- 151–166.
tributions and contact stresses in quartz, calcite, glass and acrylic plastic. Kato, N., Tullis, T.E., 2001. A composite rate- and state-dependent law for rock friction.
Tectonophysics 256 (1–4), 219–239. Geophys. Res. Lett. 28, 1103–1106.
Dixon, T.H., Jiang, Y., Malservisi, R., McCaffrey, R., Voss, N., Protti, M., Gonzalez, V., Kilgore, B.D., Blanpied, M.L., Dieterich, J.H., 1993. Velocity dependent friction of granite
2014. Earthquake and tsunami forecasts: Relation of slow slip events to subsequent over a wide range of conditions. Geophys. Res. Lett. 20 (10), 903–906.
earthquake rupture. Proc. Natl. Acad. Sci. U. S. A. 111 (48), 17,039–17,044. Kitajima, H., Chester, F.M., Chester, J.S., 2011. Dynamic weakening of gouge layers in
Duan, B., Liu, D., Yin, A., 2017. Seismic shaking in the North China basin expected from high-speed shear experiments: assessment of temperature-dependent friction,
ruptures of a possible seismic gap. Geophys. Res. Lett. 44 (10), 4855–4862. thermal pressurization, and flash heating. J. Geophys. Res. 116 (B8).
Engelder, J.T., Logan, J.M., Handin, J., 1975. The sliding characteristics of sandstone on Kocharyan, G.G., Ostapchuk, A.A., Pavlov, D.V., 2018. Traces of laboratory earthquake
quartz fault-gouge. Pure Appl. Geophys. 113 (1), 69–86. nucleation in the spectrum of ambient noise. Sci. Rep. 8 (1), 10 764.
Fagereng, A., den Hartog, S.M., 2016. Subduction megathrust creep governed by pres- Krick, B.A., Vail, J.R., Persson, B.N., Sawyer, W.G., 2012. Optical in situ micro tribometer
sure-solution and frictional-viscous flow. Nat. Geosci. 10, 51–57. for analysis of real contact area for contact mechanics, adhesion, and sliding ex-
̊
Fagereng, Å., Sibson, R.H., 2010. Melange rheology and seismic style. Geology 38 (8), periments. Tribol. Lett. 45 (1), 185–194.
751–754. Lambert, V., Barbot, S., 2016. Contribution of viscoelastic flow in earthquake cycles
Ferguson, C.C., Lloyd, G.E., Knipe, R.J., 1987. Fracture mechanics and deformation within the lithosphere-asthenosphere system. Geophys. Res. Lett. 43 (19), 142–154.
processes in natural quartz: a combined Vickers indentation, SEM, and TEM study. Lapusta, N., Barbot, S., 2012. Models of earthquakes and aseismic slip based on labora-
Can. J. Earth Sci. 24 (3), 544–555. tory-derived rate and state friction laws. In: Bizzarri, A., Bhat, H.S. (Eds.), The
Ferri, F., Toro, G.D., Hirose, T., Shimamoto, T., 2010. Evidence of thermal pressurization Mechanics of Faulting: From Laboratory to Real Earthquakes. Research Signpost,
in high-velocity friction experiments on smectite-rich gouges. Terra Nova 22 (5), Trivandrum, Kerala, India, pp. 153–207.
347–353. Lapusta, N., Rice, J.R., 2003. Nucleation and early seismic propagation of small and large
Frost, E., Dolan, J., Sammis, C., Hacker, B., Cole, J., Ratschbacher, L., 2009. Progressive events in a crustal earthquake model. J. Geophys. Res. 108 (B4) 2205.
strain localization in a major strike-slip fault exhumed from midseismogenic depths: Leeman, J., Saffer, D., Scuderi, M., Marone, C., 2016. Laboratory observations of slow
structural observations from the Salzach-Ennstal-Mariazell-Puchberg fault system, earthquakes and the spectrum of tectonic fault slip modes. Nat. Commun. 7 (11), 104.
Austria. J. Geophys. Res. 114 (B4). Li, D., Liu, Y., 2016. Spatiotemporal evolution of slow slip events in a nonplanar fault
Fukuyama, E., Yamashita, F., Mizoguchi, K., 2019. Voids and rock friction at subseismic model for northern Cascadia subduction zone. J. Geophys. Res. 121 (9), 6828–6845.
slip velocity. In: Earthquakes and Multi-Hazards around the Pacific Rim, Vol. II. Li, Q., Tullis, T.E., Goldsby, D., Carpick, R.W., 2011. Frictional ageing from interfacial
Springer, pp. 87–107. bonding and the origins of rate and state friction. Nature 480 (7376), 233.
Giger, S.B., Cox, S.F., Tenthorey, E., 2008. Slip localization and fault weakening as a Lieou, C.K., Elbanna, A.E., Carlson, J.M., 2014. Grain fragmentation in sheared granular
consequence of fault gouge strengthening? Insights from laboratory experiments. flow: weakening effects, energy dissipation, and strain localization. Phys. Rev. E 89
Earth Planet. Sci. Lett. 276 (1–2), 73–84. (2), 022–203.
Goldsby, D., Tullis, T.E., 2011. Flash heating leads to low frictional strength of crustal Lieou, C.K., Daub, E.G., Guyer, R.A., Ecke, R.E., Marone, C., Johnson, P.A., 2017.
rocks at earthquake slip rates. Science 334 (6053), 216–218. Simulating stick-slip failure in a sheared granular layer using a physics-based con-
Goldsby, D.L., Rar, A., Pharr, G.M., Tullis, T.E., 2004. Nanoindentation creep of quartz, stitutive model. J. Geophys. Res. 122 (1), 295–307.

143
S. Barbot Tectonophysics 765 (2019) 129–145

Liu, Y., Rice, J.R., 2005. Aseismic slip transients emerge spontaneously in three-dimen- surface roughness with application to contact mechanics, sealing, rubber friction and
sional rate and state modeling of subduction earthquake sequences. J. Geophys. Res. adhesion. J. Phys. Condens. Matter 17 (1), R1.
110 (B08307). https://doi.org/10.1029/2004JB003424. Persson, B.N., 2001. Theory of rubber friction and contact mechanics. J. Chem. Phys. 115
Liu, Y., Szlufarska, I., 2012. Chemical origins of frictional aging. Phys. Rev. Lett. 109 (18) (8), 3840–3861.
186,102. Power, W., Tullis, T., Brown, S., Boitnott, G., Scholz, C., 1987. Roughness of natural fault
Logan, J.M., Teufel, L., 1986. The effect of normal stress on the real area of contact during surfaces. Geophys. Res. Lett. 14 (1), 29–32.
frictional sliding in rocks. Pure Appl. Geophys. 124 (3), 471–485. Press, W.H., Teukolsky, S.A., Vetterling, W.T., Flannery, B.P., 1992. Numerical Recipes in
Lorenz, B., Krick, B., Mulakaluri, N., Smolyakova, M., Dieluweit, S., Sawyer, W., Persson, C: The Art of Scientific Computing, 2nd ed. Cambridge Univ. Press, New York 994 pp.
B., 2013. Adhesion: role of bulk viscoelasticity and surface roughness. J. Phys. Putelat, T., Dawes, J.H., Willis, J.R., 2011. On the microphysical foundations of rate-and-
Condens. Matter 25 (22) 225,004. state friction. J. Mech. Phys. Solids 59 (5), 1062–1075.
Maegawa, S., Itoigawa, F., Nakamura, T., 2015. Effect of normal load on friction coeffi- Qiu, Q., et al., 2016. The mechanism of partial rupture of a locked megathrust: the role of
cient for sliding contact between rough rubber surface and rigid smooth plane. Tribol. fault morphology. Geology 44 (10), 875–878.
Int. 92, 335–343. Reber, J.E., Lavier, L.L., Hayman, N.W., 2015. Experimental demonstration of a semi-
Marone, C., Kilgore, B., 1993. Scaling of the critical slip distance for seismic faulting with brittle origin for crustal strain transients. Nat. Geosci. 8 (9), 712.
shear strain in fault zones. Nature 362, 618–620. Rempel, A., Weaver, S., 2008. A model for flash weakening by asperity melting during
Marone, C., Raleigh, B.C., Scholz, C.H., 1990. Frictional behavior and constitutive mod- high-speed earthquake slip. J. Geophys. Res. 113 (B11).
eling of simulated fault gouge. J. Geophys. Res. 95 (B5), 7007–7026. Rice, J.R., 1983. Constitutive relations for fault slip and earthquake instabilities. Pure
Marone, C., Scholz, C.H., Bilham, R., 1991. On the mechanics of earthquake afterslip. J. Appl. Geophys. 21, 443–475.
Geophys. Res. 96, 8441–8452. Rice, J.R., 1993. Spatio-temporal complexity of slip on a fault. J. Geophys. Res. 98 (B6),
Marone, C., Cocco, M., Richardson, E., Tinti, E., 2009. The critical slip distance for seismic 9885–9907.
and aseismic fault zones of finite width. Int. Geophys. 94, 135–162. Rice, J.R., Ben-Zion, Y., 1996. Slip complexity in earthquake fault models. Proc. Natl.
Marone, C.J., 1998. Laboratory-derived friction laws and their application to seismic Acad. Sci. 93 (9), 3811–3818.
faulting. Annu. Rev. Earth Planet. Sci. 26, 643–696. Rice, J.R., Ruina, A.L., 1983. Stability of steady frictional slipping. J. Appl. Mech. 50,
Marsh, D., 1964. Plastic flow in glass. Proc. R. Soc. Lond. A Math. Phys. Sci. 279 (1378), 343–349.
420–435. Ronsin, O., Coeyrehourcq, K.L., 2001. State, rate and temperature–dependent sliding
Masuti, S., Barbot, S., Karato, S., Feng, L., Banerjee, P., 2016. Upper mantle water stra- friction of elastomers. In: Proceedings of the Royal Society of London a:
tification inferred from the 2012 Mw 8.6 Indian Ocean earthquake. Nature 538, Mathematical, Physical and Engineering Sciences. vol. 457. The Royal Society, pp.
373–377. https://doi.org/10.1038/nature19783. 1277–1294.
Mele Veedu, M., Barbot, S., 2016. The Parkfield tremors reveal slow and fast ruptures on Rouet-Leduc, B., Hulbert, C., Bolton, D.C., Ren, C.X., Riviere, J., Marone, C., Guyer, R.A.,
the same asperity. Nature 532 (7599), 361–365. https://doi.org/10.1038/ Johnson, P.A., 2018. Estimating fault friction from seismic signals in the laboratory.
nature17190. Geophys. Res. Lett. 45 (3), 1321–1329.
Mitchell, E., Fialko, Y., Brown, K., 2013. Temperature dependence of frictional healing of Rousset, B., Barbot, S., Avouac, J.P., Hsu, Y.-J., 2012. Postseismic deformation following
westerly granite: experimental observations and numerical simulations. Geochem. the 1999 Chi-Chi earthquake, Taiwan: implication for lower-crust rheology. J.
Geophys. Geosyst. 14 (3), 567–582. Geophys. Res. 117 (B12405), 16.
Montési, L.G.J., Hirth, G., 2003. Grain size evolution and the rheology of ductile shear Rowe, C.D., Lamothe, K., Rempe, M., Andrews, M., Mitchell, T.M., Di Toro, G., White,
zones: from laboratory experiments to postseismic creep. Earth Planet. Sci. Lett. 211, J.C., Aretusini, S., 2019. Earthquake lubrication and healing explained by amorphous
97–110. nanosilica. Nat. Commun. 10 (1), 320.
Nagata, K., Nakatani, M., Yoshida, S., 2008. Monitoring frictional strength with acoustic Rozel, A., Ricard, Y., Bercovici, D., 2011. A thermodynamically self-consistent damage
wave transmission. Geophys. Res. Lett. 35 (6). equation for grain size evolution during dynamic recrystallization. Geophys. J. Int.
Nagata, K., Nakatani, M., Yoshida, S., 2012. A revised rate-and state-dependent friction 184 (2), 719–728.
law obtained by constraining constitutive and evolution laws separately with la- Rubin, A.M., Ampuero, J.-P., 2005. Earthquake nucleation on (aging) rate and state
boratory data. J. Geophys. Res. 117 (B2). faults. J. Geophys. Res. 110 (B11312) 24 PP.
Nakatani, M., 2001. Conceptual and physical clarification of rate and state friction: Rubinstein, S.M., Cohen, G., Fineberg, J., 2004. Detachment fronts and the onset of dy-
frictional sliding as a thermally activated rheology. J. Geophys. Res. 106 (B7), namic friction. Nature 430 (7003), 1005.
13,347–13,380. Ruina, A., 1983. Slip instability and state variable friction laws. J. Geophys. Res. 88 (10),
Nakatani, M., Scholz, C.H., 2004a. Frictional healing of quartz gouge under hydrothermal 359–10,370.
conditions: 1. Experimental evidence for solution transfer healing mechanism. J. Rutter, E., Mainprice, D., 1978. The effect of water on stress relaxation of faulted and
Geophys. Res. 109 (B7). unfaulted sandstone. In: Rock Friction and Earthquake Prediction. Springer, pp.
Nakatani, M., Scholz, C.H., 2004b. Frictional healing of quartz gouge under hydrothermal 634–654.
conditions: 2. Quantitative interpretation with a physical model. J. Geophys. Res. Sahli, R., Pallares, G., Ducottet, C., Ali, I.B., Al Akhrass, S., Guibert, M., Scheibert, J.,
109 (B7). 2018. Evolution of real contact area under shear and the value of static friction of soft
Niemeijer, A., Spiers, C., 2007. A microphysical model for strong velocity weakening in materials. Proc. Natl. Acad. Sci. U. S. A. 115 (3), 471–476.
phyllosilicate-bearing fault gouges. J. Geophys. Res. 112 (B10). Salman, R., Hill, E.M., Feng, L., Lindsey, E.O., Mele Veedu, D., Barbot, S., Banerjee, P.,
Nishihara, Y., Shinmei, T., Karato, S.-I., 2006. Grain-growth kinetics in wadsleyite: effects Hermawan, I., Natawidjaja, D.H., 2017. Piecemeal Rupture of the Mentawai Patch,
of chemical environment. Phys. Earth Planet. Inter. 154 (1), 30–43. Sumatra: The 2008 Mw 7.2 North Pagai Earthquake Sequence. J. Geophys. Res. Solid
Noda, H., 2008. Frictional constitutive law at intermediate slip rates accounting for flash Earth 122 (11), 9404–9419.
heating and thermally activated slip process. J. Geophys. Res. 113 (B09302), 12. Sammis, C., King, G., Biegel, R., 1987. The kinematics of gouge deformation. Pure Appl.
Noda, H., Lapusta, N., 2010. Three-dimensional earthquake sequence simulations with Geophys. 125 (5), 777–812.
evolving temperature and pore pressure due to shear heating: effect of heterogeneous Sammis, C.G., Ben-Zion, Y., 2008. Mechanics of grain-size reduction in fault zones. J.
hydraulic diffusivity. J. Geophys. Res. 115 (B12314) 24 PP.,. Geophys. Res. 113 (B2).
Noda, H., Shimamoto, T., 2010. A rate- and state-dependent ductile flow law of poly- Sammis, C.G., Biegel, R.L., 1989. Fractals, fault-gouge, and friction. Pure Appl. Geophys.
crystalline halite under large shear strain and implications for transition to brittle 131 (1–2), 255–271.
deformation. Geophys. Res. Lett. 37 (L09310). https://doi.org/10.1029/ Sammis, C.G., King, G.C., 2007. Mechanical origin of power law scaling in fault zone rock.
2010GL042512. Geophys. Res. Lett. 34 (4).
Noda, H., Dunham, E.M., Rice, J.R., 2009. Earthquake ruptures with thermal weakening Sammis, C.G., Rosakis, A.J., Bhat, H.S., 2009. Effects of off-fault damage on earthquake
and the operation of major faults at low overall stress levels. J. Geophys. Res. 114 rupture propagation: experimental studies. In: Mechanics, Structure and Evolution of
(B07302), 27. Fault Zones. Springer, pp. 1629–1648.
Ong, M., Qing, Su, Barbot, S., Hubbard, J., 2019. Physics-based scenario of earthquake Scholz, C.H., 1988. The critical slip distance for seismic faulting. Nature 336, 761–763.
cycles on the Ventura Thrust system, California: the effect of variable friction and Scholz, C.H., 1998. Earthquakes and friction laws. Nature 391, 37–42.
fault geometry. Pure Appl. Geophys. https://doi.org/10.1007/s00024-019-02111-9. Scholz, C.H., Molnar, P., Johnson, T., 1972. Detailed studies of frictional sliding of granite
Pastewka, L., Robbins, M.O., 2014. Contact between rough surfaces and a criterion for and implications for earthquake mechanism. J. Geophys. Res. 77 (32), 6392–6404.
macroscopic adhesion. Proc. Natl. Acad. Sci. 111 (9), 3298–3303. Scott, D.R., Marone, C.J., Sammis, C.G., 1994. The apparent friction of granular fault
Paterson, M.S., Wong, T.-f., 2005. Friction and sliding phenomena, Experimental rock gouge in sheared layers. J. Geophys. Res. 99 (B4), 7231–7246.
deformation? In: The Brittle Field, pp. 165–209. Scuderi, M.M., Collettini, C., Viti, C., Tinti, E., Marone, C., 2017. Evolution of shear fabric
Pei, L., Hyun, S., Molinari, J., Robbins, M.O., 2005. Finite element modeling of elasto- in granular fault gouge from stable sliding to stick slip and implications for fault slip
plastic contact between rough surfaces. J. Mech. Phys. Solids 53 (11), 2385–2409. mode. Geology 45 (8), 731–734.
Perfettini, H., Avouac, J.-P., 2004. Postseismic relaxation driven by brittle creep: a pos- Segall, P., Bradley, A.M., 2012. The role of thermal pressurization and dilatancy in
sible mechanism to reconcile geodetic measurements and the decay rate of after- controlling the rate of fault slip. J. Appl. Mech. 79 (3) 031,013.
shocks, application to the Chi-Chi earthquake, Taiwan. J. Geophys. Res. 109 Segall, P., Rice, J.R., 1995. Dilatancy, compaction, and slip instability of a fluid infiltrated
(B02304). fault. J. Geophys. Res. 100 (22), 155–22,171.
Perfettini, H., Avouac, J.-P., 2007. Modeling afterslip and aftershocks following the 1992 Segall, P., Rice, J.R., 2006. Does shear heating of pore fluid contribute to earthquake
Landers earthquake. J. Geophys. Res. 112 (B07409). nucleation? J. Geophys. Res. 111 (B09316), 17.
Perfettini, H., Molinari, A., 2017. A micromechanical model of rate and state friction: 1. Segall, P., Rubin, A.M., Bradley, A.M., Rice, J.R., 2010. Dilatant strengthening as a me-
Static and dynamic sliding. J. Geophys. Res. 122 (4), 2590–2637. chanism for slow slip events. J. Geophys. Res. 115 (B12305).
Persson, B., 2016. Silicone rubber adhesion and sliding friction. Tribol. Lett. 62 (2), 34. Shimamoto, T., Noda, H., 2014. A friction to flow constitutive law and its application to a
Persson, B., Albohr, O., Tartaglino, U., Volokitin, A., Tosatti, E., 2004. On the nature of 2-d modeling of earthquakes. J. Geophys. Res. 119 (11), 8089–8106.

144
S. Barbot Tectonophysics 765 (2019) 129–145

Sibson, R.H., 2003. Thickness of the seismic slip zone. Bull. Seismol. Soc. Am. 93 (3), Ujiie, K., Saishu, H., Fagereng, Å., Nishiyama, N., Otsubo, M., Masuyama, H., Kagi, H.,
1169–1178. 2018. An explanation of episodic tremor and slow slip constrained by crack‐seal veins
Sleep, N.H., 1995. Frictional heating and the stability of rate and state dependent fric- and viscous shear in subduction mélange. Geophys. Res. Lett. 45 (11), 5371–5379.
tional sliding. Geophys. Res. Lett. 22 (20), 2785–2788. van den Ende, M., Chen, J., Ampuero, J.-P., Niemeijer, A., 2018. A comparison between
Sleep, N.H., 1997. Application of a unified rate and state dependent theory to the me- rate-and-state friction and microphysical models, based on numerical simulations of
chanics of fault zones with strain localization. J. Geophys. Res. 102, 2875–2895. fault slip. Tectonophysics 733, 273–295.
Sleep, N.H., 1998. Rake dependent rate and state friction. J. Geophys. Res. 103 (B4), Viesca, R.C., 2016. Self-similar slip instability on interfaces with rate-and state-dependent
7111–7119. friction. Proc. R. Soc. Lond. A 472 (2192), 20160,254.
Sleep, N.H., 2006. Real contacts and evolution laws for rate and state friction. Geochem. Voisin, C., Renard, F., Grasso, J.-R., 2007. Long term friction: from stick-slip to stable
Geophys. Geosyst. 7 (8). sliding. Geophys. Res. Lett. 34 (13).
Sleep, N.H., 2012. Microscopic elasticity and rate and state friction evolution laws. Wallace, L., et al., 2016. Near-field observations of an offshore Mw 6.0 earthquake from
Geochem. Geophys. Geosyst. 13 (12). an integrated seafloor and subseafloor monitoring network at the Nankai Trough,
Sleep, N.H., 2019. Thermal weakening of asperity tips on fault planes at high sliding southwest Japan. J. Geophys. Res. 121 (11), 8338–8351.
velocities. Geochem. Geophys. Geosyst. https://doi.org/10.1029/2018GC008062. Wallace, L.M., Beavan, J., 2006. A large slow slip event on the central Hikurangi sub-
Sleep, N.H., Blanpied, M.L., 1992. Creep, compaction and the weak rheology of major duction interface beneath the Manawatu region, North Island, New Zealand.
faults. Nature 359 (6397), 687. Geophys. Res. Lett. 33 (11).
Sone, H., Shimamoto, T., Moore, D.E., 2012. Frictional properties of saponite-rich gouge Wallace, L.M., Kaneko, Y., Hreinsdóttir, S., Hamling, I., Peng, Z., Bartlow, N., D’Anastasio,
from a serpentinite-bearing fault zone along the Gokasho-Arashima Tectonic Line, E., Fry, B., 2017. Large-scale dynamic triggering of shallow slow slip enhanced by
Central Japan. J. Struct. Geol. 38, 172–182. overlying sedimentary wedge. Nat. Geosci. 10 (10), 765.
Stesky, R., 1978. Rock friction-effect of confining pressure, temperature, and pore pres- Wei, M., Kaneko, Y., Liu, Y., McGuire, J.J., 2013. Episodic fault creep events in California
sure. In: Rock Friction and Earthquake Prediction. Springer, pp. 690–704. controlled by shallow frictional heterogeneity. Nat. Geosci. 6 (7), 566.
Stesky, R.M., Brace, W.F., Riley, D.K., Robin, P.Y., 1974. Friction in faulted rock at high Wei, S., Barbot, S., Graves, R., Lienkaemper, J.J., Wang, T., Hudnut, K., Fu, Y.,
temperature and pressure. Tectonophysics 23 (1-2), 177–203. Helmberger, D., 2015. The 2014 Mw 6.1 South Napa earthquake: a unilateral rupture
Sulem, J., Famin, V., 2009. Thermal decomposition of carbonates in fault zones: slip- with shallow asperity and rapid afterslip. Seismol. Res. Lett. 86 (2A), 344–354.
weakening and temperature-limiting effects. J. Geophys. Res. 114 (B03309), 14. Wu, Y., Chen, X., 2014. The scale-dependent slip pattern for a uniform fault model
Teufel, L., Logan, J., 1978. Effect of displacement rate on the real area of contact and obeying the rate-and state-dependent friction law. J. Geophys. Res. 119 (6),
temperatures generated during frictional sliding of Tennessee sandstone. Pure Appl. 4890–4906.
Geophys. 116 (4–5), 840–865. Yamashita, F., Fukuyama, E., Mizoguchi, K., 2014. Probing the slip-weakening me-
Thom, C.A., Carpick, R.W., Goldsby, D.L., 2018. Constraints on the physical mechanism of chanism of earthquakes with electrical conductivity: rapid transition from asperity
frictional aging from nanoindentation. Geophys. Res. Lett. 45 (24), 13–306. contact to gouge comminution. Geophys. Res. Lett. 41 (2), 341–347.
Timoshenko, S., Goodier, J., 1951. Theory of Elasticity, 2nd ed. McGraw-Hill, New York. Yastrebov, V.A., Anciaux, G., Molinari, J.-F., 2015. From infinitesimal to full contact
Toro, G.D., Goldsby, D.L., Tullis, T.E., 2004. Friction falls towards zero in quartz rock as between rough surfaces: evolution of the contact area. Int. J. Solids Struct. 52,
slip velocity approaches seismic rates. Nature 427, 436–439. 83–102.
Toro, G.D., Hirose, T., Nielsen, S., Pennacchioni, G., Shimamoto, T., 2006. Natural and Yasuhara, H., Marone, C., Elsworth, D., 2005. Fault zone restrengthening and frictional
experimental evidence of melt lubrication of faults during earthquakes. Science 311 healing: the role of pressure solution. J. Geophys. Res. 110 (B6).
(5761), 647–649. Yund, R., Blanpied, M., Tullis, T., Weeks, J., 1990. Amorphous material in high strain
Tse, S.T., Rice, J.R., 1986. Crustal earthquake instability in relation to the depth variation experimental fault gouges. J. Geophys. Res. 95 (B10), 15,589–15,602.
of frictional slip properties. J. Geophys. Res. 91 (B9), 9452–9472.

145

You might also like