53

BRITTLE FAULTING
Because most rocks are brittle at low temperature and low confining (lithostatic) pressure, virtually
every rock at or near the Earth’s surface exhibits evidence of brittle failure, i.e. deformation-induced
loss of cohesion.
Brittle failure results from the irreversible and very rapid propagation and connection of cracks, a
process called fracturing. Cracks are grain-scale planes that pre-exist or nucleate under stress on
some microscopic defect of the material. Fracturing initiates at stress levels near the yield strength
(or maximal differential stress σ1 - σ3 ), marking the elastic limit of the material. Tensile stress cannot
act across the resulting physical discontinuity that has no cohesion.
By reference to laboratory observations, a broad mechanical classification of fractures recognizes two
fracture modes:
- Extension fractures, resulting from the initial separation of two formerly contiguous surfaces;
displacement is parallel to the minimum principal stress (i.e. maximum tension), which is
orthogonal to the fracture plane.
- Shear fractures resulting from initial displacement along the fracture plane; fractures and
displacement are oblique to the maximum principal stress (maximum compression).
This lecture discusses qualitatively the relationships of the earthquake source to faulting.

Brittle faulting is a process producing localized offset along a shear fracture. This lecture discusses
how and qualitatively comments the relationships of the earthquake source to brittle faulting.

Fault orientation relative to principal stress axes

Theoretical considerations
A force F that acts on P can be resolved into components normal (FN) and parallel (FS) to the plane P.
The components have magnitudes:
FN Fcos θ
= and FS Fsin θ
= (1)

with θ the angle between the direction of the force and the normal to the plane (see lecture on
mechanical aspects of deformation; forces and stresses).

Faulting jpb, 2020

54

Drawing cross-sections of a cube with the force of magnitude F acting normally to one cube face of
area A, the stress is by definition the concentration of force per unit area, which can be visualized as
the intensity of force if the cube is such that its faces = unit area = 1. Thus, the magnitudes of the
normal and shear components of stress across a plane P are:

σ N = FN A P = ( F A ) cos 2 θ = σ cos 2 θ
and (2)
σ
σS FS=
= AP ( F A ) sin θ cos
= θ sin 2θ
2
Typically, any rock is under a triaxial state of stress, and σ1 , σ 2 and σ 3 are the principal stresses with
σ1 ≥ σ2 ≥ σ3.

Remember! The convention in geology takes all positive stresses as compressive. In the
non-geological literature, extensional stresses are positive!

For practical purposes, one considers an arbitrary plane P within the body, parallel to σ 2 and whose
normal line makes an angle θ with σ1 ( θ is also the angle between the plane P and σ 3 ). One assumes
that, for an elementary treatment of the state of stress and of the quantitative relationship between the
normal and shear stresses, one may neglect σ 2 and only consider the two-dimensional principal plane
( σ1, σ3 ) . Then, lines in that plane represent traces of planes perpendicular to it and parallel to σ 2 .
All considered planes will make a line in this plane.
The stress tensor can be represented by its two principal components σ1 , and σ 3 . Where the principal
stresses are σ1 and σ 3 the equations for the normal and shear stresses across a plane whose normal is
inclined at θ to σ1 are:

=
σN
( σ1 + σ3 ) + ( σ1 − σ3 ) cos 2θ
2 2 (3)
( σ − σ3 ) sin 2θ
σS = 1
2
The principal stresses σ1 and σ3 can be assumed to represent the tectonic stress field (for instance,
compression σ1 is horizontal and σ 3 vertical). With this assumption, the equations (3) are important
because they can be applied to relate regional tectonic stresses to the normal and shear stresses on
local fault planes.
Equations (3) demonstrate that the value of σS is maximum when sin 2θ =1 i.e. 2θ= 90° . Thus, the
planes of maximum shear stress make a theoretical angle of 45° with σ1 and σ3 . The maximum
shear stress has the value ( σ1 - σ3 ) 2 .
In all cases where σ1 ≥ σ2 ≥ σ3 there are two planes of maximum shear stress. The paired faults,
called conjugate faults, develop more or less synchronously in both of the equally favored
orientations; conjugate faults intersect in a line parallel to the intermediate principal stress axis σ 2 .
In the special case where σ2 = σ3 or σ1 = σ2 , there is an infinite number of planes inclined at 45° to
σ1 or σ3 . All possible orientations are tangents to a cone.
Equations (1) and (3) imply that compressive normal stresses tend to inhibit sliding while shear
stresses tend to promote sliding on any plane. This lecture will now discuss how these conclusions
apply to brittle deformation, keeping in mind that the demonstration is valid for the onset of faulting
only.

Faulting jpb, 2020

 55

Shear fracture criteria

Shear failure occurs when loading creates shear stresses that exceed the shear strength of the rock. A
yield criterion is a hypothesis concerning the limit of elasticity under any combination of stresses.
Three main failure criteria are applied to rocks:
- The Coulomb criterion;
- The Mohr envelope;
- The Griffith crack theory;
- None of the basic “friction laws” discussed in the following paragraphs is based on mechanical
processes. They are only phenomenological, in-equation descriptions of experimental faulting.
This limitation is due to the complex and constant evolution of surface contacts during fault
displacement (roughness changes with asperities being sheared, plowing, stepping up and down,
the interlocking of asperities, and strength of the rock to be sheared and chemical reactions
producing wears).

Frictional sliding
Three coefficients refer to friction: (1) internal friction, to create a sliding surface;
(2) static friction to initiate movement on the sliding surface;
(3) dynamic friction to maintain sliding on the surface.
First investigations considered simple experiments with objects sliding down an inclined plane under
the action of gravity.

Only two forces were considered: the vertical weight of the object W and the horizontal force F ,
which at rest are balanced by opposite reactions Wr and Fr , respectively. Fr arises from frictional
forces that resist the motion that F would impel. At some magnitude Fr breaks and the object moves.
According to the weights of different objects, some proportionality between the two forces was
established:
If F = Fr < μ s W the object does not move, it is static.
If F = Fr > μ s W sliding initiates, the object is accelerated.
This is the Amonton law, in which the static friction µs is independent of the area of contact.
In reality, the shear component of weight WS acting parallel to the inclined plane initiates sliding
while the normal component WN is the resisting force. The static friction µs becomes the ratio of the
shear stress to the normal stress:

Faulting jpb, 2020

56

WS sinθ
μs = = = tanθ
WN cosθ
where θ is the inclination of the plane termed angle of friction.
Such observations lead to the concept that two bodies with a plane surface of contact are pressed
together by the normal stress σ N . The shear stress σS necessary to initiate sliding is related to σ N ,
by an equation of the form:
σS = f ( σ N ) (4)

which, reformulating the Amonton law, becomes:

σS = σ N tan φ =µs σ N (5)
φ is the angle of internal friction. It is the angle between the vector normal to the rupture plane and
the stress acting on this plane. The term internal friction is not strictly friction. It describes a material
property of slip resistance along the fracture and the state of the surfaces in contact (e.g. smooth versus
rough, fresh versus altered or coated, etc.). It also involves the coalescence of microcracks.
Experiments have shown that this linear relationship is valid for materials with no cohesion strength,
such as soils.

Coulomb criterion
Charles Augustin de Coulomb found that there is both a stress-dependent and stress-independent
component of shear strength. He proposed in 1776 that shear fracture occurs when the shear stress
along a potential fault plane overcomes two forces: (1) the cohesive strength of the material before
failure on that plane and (2) the resistance along that plane once it had formed. He expressed shear
resistance S as:
1
S c.a + N
=
n
where c is the cohesion per unit area, a is the area of the shear plane, N is the normal force on the
shear plane and 1 n is the coefficient of internal friction.
In modern terms, this equation is written:
σS = c + µ σ N (6)
where c is the cohesion, a material constant;
µ is the coefficient of internal friction, another material constant equivalent to tan φ , the
coefficient of internal friction seen for cohesionless soils (equation 5).
Both parameters are not inherent properties of the tested material but depend on the test conditions.
Equation (6) is often referred to as the Coulomb criterion.
Physical interpretation
Equation (6) assumes that shear fracture in solids involves two factors together:
- Breaking cohesive bonds between particles of intact rock (the c term); cohesion is a measure of
this internal bonding.
- Frictional sliding (the term µ , proportional to the normal compressive stress σ N acting across the
potential fracture plane); internal friction is caused by contact between particles.
This physical interpretation provides an acceptably good fit with many experimental data, which yield
cohesive strength of 10-20 MPa for most sedimentary rocks and about 50 MPa for crystalline rocks.
The average angle of internal friction is 30° for all.

Faulting jpb, 2020

 57

Experimental fractures
In triaxial experiments ( σ1 , σ 2 and σ3 have non zero magnitudes) the actual shear fractures form
dihedral angles smaller than 45° to the maximum compressive stress axis σ1 . A good average for
rocks is approximately 30°. The acute bisector of conjugate faults is parallel to σ1 . This geometrical
relationship between the stress axes and the shear fractures is widely utilized to define in the field the
orientation of the stress axes from measurements of faults. Understanding faulting implies
understanding the difference between this ideal angle and the fault direction. The factors that
contribute to this angular difference are included in the concept of “angle of internal friction”.
Three-dimensional deformation involves the activation of polymodal sets of fractures, usually
comprising variously oriented sets of conjugate (bimodal) faults. For instance, a pair of conjugate sets
forms a quadrimodal fault set. These quadrimodal faults intersect to form rhombohedral traces on
outcrop surfaces with σ1 and σ 2 bisecting the acute angles between the fault planes.

The notion of simultaneous conjugate faults is geologically valid, provided "simultaneous" loosely
means alternating over a short amount of time. Space problems ensue from strict simultaneity; they
can be solved only by rotation and alternating differential slip on each of the conjugate faults.

Graphical representation
The general form of equation (6) also predicts that failure points in a σS σ N diagram should lie on a
straight line with slope µ and intersecting the σS ordinate at the shear strength c = τ0 . The line is
characteristic of many rocks tested at moderate confining pressures and describes their shear failure
to a good approximation.
The graphical representation shows that:
- Any stress condition below the Coulomb criterion line is safe, but shear failure occurs as soon as the
stress condition touches the line. Then slip relax stress, thereby pre-empting stress conditions that
would exceed the failure line (i.e. Mohr circles growing beyond the criterion line).

Faulting jpb, 2020

58

- The shear fractures form at less than 45° to σ1 since the shearing resistance line has a positive slope.
This slope represents the competing effects of σ N and σS on shear fractures. Both the minimum
normal stress and maximum shear stress promote together new shear fractures. The acute angle
between conjugate faults in rocks (60 to 70°) is an optimization of these two conditions.

- Since slip depends on the magnitude and not the sign of shear stress, there are two lines symmetrical
with respect to the horizontal axis of normal stresses. These lines have slope angles ± φ . Accordingly,
the acute angle between the conjugate faults is bisected by the greatest principal stress σ1 .
- The criterion is linear but since rocks cannot sustain large tensile stresses a tension cut-off (a vertical
line on the negative side of normal stresses) is often introduced.
The graphic shows that the angle of internal friction φ , which is a material property, is related to the
φ
angle θ of the fault plane: θ= + 45 . The linear failure criterion imposes that new shear fractures
2
make with σ1 a well-defined angle of ± ( 45° − φ 2 ) .

Faulting jpb, 2020

 59

Linear relationship between σ1 and σ3

The graphical representation of a Mohr circle shows that the radius is:
σ1 − σ3  σ1 + σ3 c 
= +  sin φ
2  2 tan φ 
Multiplying both sides by 2 and transposing terms, this becomes:
σ1 (1 − sin φ ) =
σ3 (1 + sin φ ) + 2c.cos φ
Expressing σ1 :
2c.cos φ  1 + sin φ 
=σ1 + σ3   (7)
1 − sin φ  1 − sin φ 

Playing with the classical sine and cosine identities sin α = ± 1 − sin 2 α and cos α = ± 1 − sin 2 α :

=
cos φ
=
1 + sin 2 φ (1 − sin φ=)(1 + sin φ ) 1 + sin φ
1 − sin φ (1 − sin φ )2 (1 − sin φ )(1 − sin φ ) 1 − sin φ

Substituting this result into equation (7) gives:

1 + sin φ  1 + sin φ 
=σ1 2c + σ3  
1 − sin φ  1 − sin φ 
Since c and φ are material constants, this alternate expression of the Coulomb criterion is of the
linear form:
σ1 = a + bσ3
1+ sin φ
a = 2c b und b=
1− sin φ
This relation shows that σ1 and σ 3 are linearly related when a fracture occurs.

Faulting jpb, 2020

60

Real-world: Byerlee law

Compilation of experimental data on samples with pre-cut fault surfaces indicates that frictional
sliding (slip of pre-existing fracture planes) is independent of rock type, except for many clay-rich
rocks. Resistance to shearing depends on confining pressure σ N . Two best-fitting, general and
empirical equations expressing this relation are known as Byerlee’s law:
For σ N < 200MPa σS= 0.85 σ N (8)
For 200MPa < σ N < 2000MPa σS = 0.6 σ N + 50 MPa (9)
These laws were established at room temperature but emphasize that friction is proportional to the
normal stress, which presses fault blocks together.

Mohr conditions
Christian Otto Mohr proposed in 1900 that σS and σ N are related in general by a non-linear function
specific to the material considered. The function represents the shape of the envelope to a series of
Mohr circles at failure for a given material: the Mohr envelope, which is an empirical curve that
delimits the field of failure of the material.
Form of the Mohr envelope; graphical representation
The Mohr envelope is slightly concave towards the σ-axis and is symmetric with respect to this axis.
Its shape and its position vary for each material and are empirically obtained as follows:
- A cylinder of rock is axially compressed under constant confining pressure. The axial load is
gradually increased until the rock fails. The Mohr circle at failure contains a point P, which
represents the σS and σ N stresses on the failure surface and indicates the orientation of the
planes along which the rock has failed, at an angle θ to σ 3 (here the confining pressure).
- 2θ is the trigonometric angle between the radius normal to the tangent to the Mohr circle at P and
the σ -axis.
- Several experiments are performed on identical samples of one rock type at different confining
pressures. They yield slightly different Mohr circles. The tangent to these circles, which
passes through all P points, represents the Mohr envelope or failure envelope for the
particular rock.

A series equation expresses this envelope:

Faulting jpb, 2020

 61

2c cos φ + σ3 (1 + sin φ )
σ1 =
1 − sin φ
Drawing the symmetrical failure criterion below the axis of normal stresses sets conjugate shear
fractures at angles −θ to σ3 .

The shape of the Mohr envelope also illustrates that:

- The curve cuts the normal stress axis only at one point, which means that it is impossible to cause
shear fractures with hydrostatic pressure.
- The shear stress required to produce failure increases with the confining pressure.
- The Mohr circles at failure become progressively larger with the size of the confining pressure.
- At high confining pressures, the envelope delineates a symmetrical pair of parallel and horizontal
lines at a critical shear strength (Von Mises criterion), which means that the material becomes
perfectly plastic, while the ductile flow is pressure-insensitive.
- The progressively decreasing slope with increasing pressure to horizontal at high pressure
represents the progressive transition from brittle to ductile behavior.

At intermediate confining pressures, the fracture strength usually increases linearly with increasing
confining pressure. The angle between this line and the horizontal axis is the angle of internal friction
φ and the slope of the envelope is the Coulomb coefficient µ with, as above
µ tan φ
=
In soil mechanics, the curved envelope is considered as a line.
Physical concept and interpretation
The curvature of the Mohr envelope is attributed to the increasing proportion of cracked compared to
intact areas on the incipient rupture surface. The strength of this incipient plane is then a combination

Faulting jpb, 2020

62

of the frictional resistance across the cracked areas and some measure of the bulk cohesion strength of
the material in the intact areas.
Once a Mohr envelope is established on a given rock type, this curve can be used to predict both the
ultimate strength and the fault angle in tests at other confining pressures. The shearing resistance can
be calculated with the equation:
( σ − σ ) sin 2θ
σS = 1 3
2
2θ is the trigonometric angle between the radius through P (normal to the tangent to the circle) and the
σ-axis.

A simple geometrical construction shows that the angle 2θ in a Mohr construction must exceed a right
angle by a value of φ , which expresses the angle of internal friction. The consistent orientation of new
shear surfaces corresponding to Mohr-Coulomb failure is:
2θ = ± ( 90° + φ )
The sign ± implies that there are two symmetrically oriented (conjugate) fault directions.

The two-dimensional Mohr-Coulomb failure theory assumes that failure is only a function of the
differential stress ( σ1 − σ3 ) , i.e. the diameter of the Mohr circle. σ2 has no influence.
The law is based on the following concept:
- If a Mohr circle representing a particular combination of σ1 and σ 3 does not intersect the
envelope, the material will not fracture and remains elastic.
- If the Mohr circle touches or intersects the Mohr envelope, the material will fracture. The contact
point defines the orientation of the fracture plane. Note that tangency of the circles to the Mohr
envelope determines the inclination of the failure plane that will form: only one angle θ is possible
at failure.
Faulting jpb, 2020
 63

- No part of any Mohr circle can go above the envelope, into the failure field, because the critical
stresses are exceeded.
In practice, only a limited portion of the envelope is available; the hydrostatic tension (negative
stresses) has not been attained experimentally.

Griffith criterion
Physical background
Alan Arnold Griffith noticed that strength values measured on glass in tension are smaller than those
predicted by solid-state theory.
Breaking inter-atomic bonds
The solid-state theory considered the atomic scale, with a crack developing through breaking at once
all inter-atomic bonds across the whole crack surface. Knowing the strength of single bonds, the
calculated strength to break all the bonds is 10 to 1000 times bigger than the experimentally measured
tensile strength. The discrepancy between theoretical and empirical strength is explained with
microscopic flaws or cracks found either on the surface or within the material, and these cracks
produce local stress magnification at their tips. The fracturing of material has to start somewhere
within the material. It usually starts at a location of stress concentration.
Stress magnification
The theory of elasticity had shown that defects or holes in infinite isotropic plates of linear elastic
materials amplify stresses at their boundaries. They are stress raisers. The amount of stress
magnification depends primarily on the shape, location, and orientation of the defect. In two
dimensions, the two extremities of an elliptical hole lying orthogonal to the direction of the remote
(e.g. regional) and uniaxial tensile stress σ are head-points where the boundary is parallel to stress.
The amount of stress magnification at these two tips is related to both the pore geometry, i.e. the long
(2L) and short ( 2 ) axes of the elliptical hole, and to the far-field stress. The maximum stress σmax at
the tips of the ellipse is given by:
 2L 
σmax ≈ σ 1 +
  

where ( 1 + 2L  ) is the stress concentration factor. Thus, the larger is the axial ratio, the greater the
stress concentration. The ellipse flattens to a crack if L >>>  . For circular holes, L =  and the stress
at the two “extremity” points is three times larger than the applied stress ( σmax =
σ3 ), independently
of the hole’s size.

Faulting jpb, 2020

64

The theory expresses stress concentration as a function of r, the radius of curvature of the ellipse at
its ends, which is related to its length and width:
r = 2 L
Solving this for  and substituting into the L/l ratio in the σmax equation, the tip stress attains the
form:
σmax = (
σ 1+ 2 L r )
and if L >>> r :
2σ L r
σmax =

This criterion suffers from one major drawback: if r → 0 (the case for a longitudinal crack), then σmax
→ ∞. This is not realistic because no material can withstand infinite tensile stress and there is no
obvious tendency for cracks to lengthen spontaneously.

Exercise
Numerically play with the shape of holes (from circle to flat ellipses) to visualize effects
of porosity shape on stress concentration.
Postulate
Griffith postulated in 1920 that numerous sub-microscopic flaws such as bubbles, fine pores or simply
grain boundaries, now called Griffith cracks, are naturally spread throughout apparently
homogeneous and elastic materials. Griffith cracks in rocks may be original or induced imperfections
along grain boundaries or within grains.
Concept
Such minute openings have the shape of extremely elongated and flattened ellipsoids across which
atomic bonds are originally broken. A relatively low applied stress is variably amplified, depending
on the orientation of the flattened ellipsoid with respect to the applied stress, at the sharply curved
ends of these pre-existing cracks. Amplification produces stresses much higher than the mean stress
in the material, to the point where the amplified local stress reaches the strength required to break the
atomic bonds. Then the cracks spread spontaneously from their tips, where only a few chemical bonds
have to be broken, at applied stress smaller than the theoretical tensile strength of the material.
Progressive rupture and connection of the network of microcracks ultimately lead to brittle failure.

Faulting jpb, 2020

 65

Necessary condition for crack growth in two dimensions

Griffith’s analysis integrated the elastic strain energy required to create the new surface area of an
expanding single crack in a thin plate. His approach abides by the first law of thermodynamics: “When
a system goes from non-equilibrium to equilibrium state there is a net decrease in energy”. In a very
small material volume, brittle fracture occurs when the decrease in the strain energy during an
incremental crack growth is equal or exceeds the energy absorbed to create the new surfaces of the
crack. The applied external load creates the elastic energy stored in the cracked plate. The
thermodynamic argument is complex and can be found in textbooks dedicated to fracture mechanics.
The demonstration shows that the tip tensile stress σT at either end of the crack equals the atomic
bonds when:
2AE
σT = (10)
πλ
where A is the surface energy per unit area of the crack (i.e. the energy required to create new surface),
E is the effective Young’s (elasticity) modulus of the rock and λ =L 2 , i.e. the elliptical crack half-
length. Assumptions imply that the crack extends in its plane. Equation (10) shows that short cracks
require larger far-field stresses to grow than long cracks. In other words, the greater is the length of
the cracks, the lower the macroscopic tensile strength of the material.

Experiment:
Take a piece of paper and pull strong on both extremities. How difficult is it to tear
apart? Make a cut in the middle? Is it easier? Make on a similar piece of paper two
cuts of different lengths. Result?
Critical stress intensity
Equation (10) is a result expressed in terms of far-field load. It can also be expressed in terms of the
stress at the crack tips. The necessary thermodynamic criterion for crack propagation through opening
only (i.e. motion orthogonal to L, with no lengthwise shear displacement) is:
K I = σT πλ
where K I is the critical stress intensity factor also called fracture toughness at fracture propagation.
It is a material property expressed in MPa.m1/2. It defines fracture propagation in the considered
material. K I depends on parameters such as temperature, confining pressure and chemical/fluid
environment. At room temperature, it varies from about 0.1 MPa m1/2 for coal to 3.5 MPa m1/2 for
granite and dunite.
Theory of failure
Griffith extended the concept to materials containing a large number of elliptical cracks oriented at
random. He assumed that these cracks propagate when the tensile stress reaches the critical stress
intensity. In materials with cracks of different axial ratios, cracks with the largest axial ratios most
likely grow first unstably and lengthwise into intact material. They interact and link under the influence
of applied tensile or compressive stress to reach the failure of the rock. This theory leads to Griffith’s
criterion for failure expressed by a curved line in a Mohr diagram:
)2
( σ1 − σ3= 8T0 ( σ1 + σ3 ) if σ1 + 3σ3 ≥ 0
and
σ3 =−T0 if σ1 + 3σ3 ≤ 0
where T0 is the uniaxial tensile strength of the material (the lowest intercept between the failure
envelope and the horizontal, normal stress abscissa).
Note also that:

Faulting jpb, 2020

66

σ3 =0 then σ1 =8T0
which suggests that uniaxial compressive stress at crack expansion is always eight times the uniaxial
tensile strength. This is not consistent with experiments as the uniaxial compressive strength of most
rocks is 10 to 50 times the uniaxial tensile strength. Typically, T0 ≈ 40MPa for intact rock.
Graphical representation
The Griffith criterion can be expressed in function of normal stress σ N and shear stress σS acting on
the plane containing the major axis of the crack:
σ S2 4T0 ( T0 − σ N )
= (11)
This equation gives a parabolic envelope for failure points in the tension side of the plot.

On the compression side, equation (11) becomes:

σS 2T0 + µσ N
=
which is a modified version of the Coulomb criterion (equation 6). When σ N = 0 , σS = 2T0 , which
represents cohesion c . This relationship ( c = 2T0 ) fits well experimentally derived curves for
shear failure.

Faulting jpb, 2020

 67

Brittle failure criterion

In experiments, initial failure occurs at the peak stress that the rock can sustain, known as static
friction. The complete criterion for brittle failure is obtained by linking the two criteria (Griffith to
Mohr-Coulomb; equations 6 and 10) at the point where they meet. For most rocks, the coefficient of
sliding friction introduced to account for postulated closure of Griffith cracks under compression is
0.5 < µ < 1 (0.75 is a representative value). Linking the parabolic Griffith criterion to the straight
Coulomb criterion explains a complete transition, with increasing mean stress (therefore depth), from
extension fractures that are parallel to σ1 (with negative σ values, left of the σS ordinate axis where
σ N =)0 to faults inclined at ca. 30° to σ1 through shears with a dilatational component inclined at
lower angles to σ1 . Experiments and field evidence corroborate the occurrence of shear fractures at
very low angles to σ1 .
The type of fracture that occurs within intact rocks depends on the ratio of differential stress ( σ1 − σ3 )
to tensile strength T0 . For µ =0.75 :
compressional shear failure occurs when ( σ1 − σ3 ) > 5.66T0 ,
extensional shear requires 5.66T0 > ( σ1 − σ3 ) > 4T0
extension fracturing requires ( σ1 − σ3 ) < 4T0 .
Effects of environmental and material factors
Compressive strength is the capacity of a material to withstand axially directed compression. The
compressive strength of rock is usually defined by the ultimate stress, the maximum stress the rock
can withstand. Well-controlled compression tests have generated complete stress-strain curves of
various rock specimens, so deciphering the role of various material and physical/chemical parameters.

Effects of fluids
Pores in rocks are primary features like vesicles in volcanic rocks and intergranular spaces in any rock.
Secondary pore spaces are deformation-induced microcracks, joints, and faults. Pores of natural rocks
contain fluids (e.g. water, oil, gaseous phase, and melt in deep levels of the Earth) which affect the
failure of rocks in two ways: (1) the mechanical effect of fluid pressure, which accelerates the
propagation of microfractures and (2) the chemical interactions between the rock and the fluid; this
effect is attributed to the weakening of the crystalline framework at the highly stressed tips of cracks
by a stress corrosion process involving rapid hydrolysis of silicon-oxygen bonds.
Fundamentally, fluid pressure counteracts and therefore decreases the confining pressure. When stress
is applied to wet rocks, volume changes first translate into pore pressure changes. Excess pore pressure
may or may not escape depending on the permeability of the rock at the time available.

Remember few definitions

Porosity describes how densely the material is packed. It is the ratio of the non-solid volume to the
total volume of material. Porosity, therefore, is a dimensionless fraction between 0 and 1. The value
ranges from <0.01 for granite, up to 0.5 for porous sandstone. It may also be represented in percent
terms by multiplying the fraction by 100%. Porosity provides the void for fluids to flow through rock
material. High porosity therefore naturally leads to high permeability.
Water content is a measure indicating the amount of water the rock material contains. It is the
dimensionless ratio of the volume of water to the bulk volume of the rock material.
Permeability is a measure of the ability of a material to transmit fluids. Rocks generally have very
low permeability. Permeability of the rock mass presents limited interest because the fluid flow is
concentrated in fractures. The physical unit of permeability is m2.

Faulting jpb, 2020

68

Pore pressure
If the pores are interconnected and communicate with the Earth's surface, which is common in the
uppermost crust, the hydrostatic pressure in pore water at any depth is equal in all directions, outward
from the pore space, to the weight of the water column from this depth to the surface.
Therefore, the pore fluid pressure within a rock increases as the rock is buried. To quantify this
consideration, the hydrostatic pore pressure at any depth is about 0.25 - 0.3 times the lithostatic or
rock pressure, if one takes the mean density of the pore fluid to be 1.0 g cm-3 and the mean density of
the rock column to be 2.5 to 3.0 g cm-3. However, the pore pressure can vary temporarily and spatially
in the lithosphere in a wide range of values.
The pore-fluid factor λ is the ratio between the fluid pressure and the lithostatic pressure:
λ =Pfluid Prock
which defines the fluid-pressure level at different depths.

For dry rocks λ = 0.

Ratios of pore pressure to lithostatic pressure above 0.8 have been measured in oil fields. Ratios
approaching 1.0 are considered for water-saturated, unconsolidated fresh sediments.
For hydrostatically pressured sediments λ = 0.4
If pore fluids are not connected to the surface, several mechanisms can generate fluid overpressure
(pore pressures greater than hydrostatic, i.e. λ > 0.4). Important mechanisms are:
- Seismic shocks can rapidly amplify the fluid pressure for a short time, which may even fluidize
water-saturated soil or unconsolidated sediments.
- Reduction of pore space (compaction) in fluid-rich sediments by burial or tectonic deformation
and rapid fluid release by dehydration of mineral assemblages during diagenesis. Rates of loading
or dehydration may exceed the escape rate of pore fluids in low permeability rocks. This is
particularly true if pores do not form an open system connected with the surface.
- Deeper in the crust fluids internally expand at elevated temperatures and new fluids are added by
fluid-releasing metamorphic reactions.
Faulting jpb, 2020
 69

- Magmatic intrusions and melt segregation in melting rocks generate fluids faster than they can
escape.
Under these circumstances, the presence of fluids can considerably reduce the effective brittle strength
of rocks. A rock overpressured by fluids in great depths may break as if it were near the Earth's surface.
Fluid-triggered crack propagation likely assists emplacement of magmatic intrusions, in that of dikes
and sills.
Effective normal stress
The stress state within the pores is hydrostatic and the fluid pressure acts so as to oppose the lithostatic
stress caused by the overburden: Pore fluids support some of the load that would otherwise be
supported by the rock matrix and, therefore, the pore pressure Pf is a component of the total stress.
Pf acts uniformly against the normal stress σN on the rock according to the equation:
σeff =σ N − Pf

The pore pressure reduces all the lithostatic stresses by an amount Pf to give the effective normal
stress σeff . Thus the principal stresses become:

σ1eff =σ1 − Pf
σeff
2 =σ2 − Pf
σ3eff =σ3 − Pf
The pore pressure produces no shear stress, and hence no shear deformation. The net result of pore
fluid pressure is to allow the rock to behave as if the confining pressures were lowered by an amount
equal to Pf .
Graphical representation
A Mohr diagram can illustrate how pore pressure promotes faulting.
- A circle representing the state of stress in a rock with a zero pore pressure is drawn. The effective
normal stress σieff
i equals the total normal stress σii . The rock is stable under these conditions if
the circle is below the failure envelope.
- If the pore pressure is increased gradually (for example during burial), σ1eff and σ3eff are smaller
than σ1 and σ 3 , respectively, but the differential stress does not change:

( σ1eff − σ3eff ) = ( σ1 − σ3 )
The ‘dry’ circle representing the regional stress keeps the same size but inexorably migrates along
the normal stress abscissa from its original position towards lower compressive stresses, closer to
the failure envelope. The amount of translation of the stress circle is determined by the magnitude
Pf of the fluid pressure.
- If Pf is large enough, the circle will hit the Mohr envelope and faulting will occur. The pressure
of the pore fluid thus allows faulting even though the shear stresses are too small for faulting in
the dry rock, or wet rock at lower pore pressures. This effect is verified in triaxial tests. Internal
fluid pressure reduces markedly both the fracture strength and the ductility of rocks, which both
are functions of the effective confining pressure. This effect also accounts for the increased
occurrence of landslides in the aftermath of heavy rainfall.

Faulting jpb, 2020

70

Hydraulic fracturing
By simply increasing the fluid pressure, the outward push of the fluid creates tensile stress sufficient
to cause crack propagation at the pore and crack tips. This process is called hydraulic fracturing. In
this way, an originally compressional stress regime can be changed so that one or more of the principal
stresses becomes effectively tensile and the conditions for tensile failure can be satisfied.

A fluid opens a fracture if the fluid pressure Pf equals or exceeds the normal stress σN acting on the
fracture. A vertical line locating the fluid pressure divides the Mohr circle into two fields:
- A domain with Pf ≥ σ N (left side of the fluid pressure line) in which fractures can dilate.
- A domain with Pf ≤ σ N (right side of the fluid pressure line) in which fractures remain closed.
The condition for fracture opening is expressed from equation (3) as:
σ + σ3 σ1 − σ3
Pf ≥ 1 + cos 2θ
2 2
Faulting jpb, 2020
 71

The driving stress ratio DSR defines the ranges of orientations of fractures that can dilate under a
given fluid pressure:
σ + σ3
Pf − 1
= DSR 2 ≥ cos 2θ
σ1 − σ3
2
which is expressed more simply in terms of mean stress σ and maximum shear stress σS max defined
in the lecture on stresses:
P −σ
DSR = f ≥ cos 2θ
σS max
A Mohr diagram represents the range of fracture orientations that will open under this bounding
condition. The pore pressure, the intersection of the Mohr circle with the vertical line passing through
the pore pressure, and the center of the Mohr circle (the mean stress) define a triangle whose horizontal
side, lying on the normal stress axis, is Pf − σ = σSmax cos 2θ . Fracturing represented by the
intersection point on the Mohr circle is thus written:
Pf − σ ≥ σSmax cos 2θ

Faulting jpb, 2020

72

This relationship is an expression of DSR , which uses the relative magnitudes of stresses and fluid
pressure to predict the range of fracture orientations able to dilate.
From the Mohr construction one readily sees that:
If DSR < −1 , Pf < σ3 , no fractures open.
If −1 < DSR < 1, σ3 < Pf < σ1 , a limited range of fracture orientations can open.
If DSR > 1, Pf > σ1 , fractures of any orientation can open (brecciation of the rock).
The same reasoning and the same constructions can be used in three dimensions.
When Pf < σ2 , the poles to fractures able to dilate define a cluster distribution around σ3. A narrow
cluster region indicates a low driving fluid pressure ( Pf ≈ σ3 ).
When Pf > σ2 , poles to fractures able to dilate form a girdle distribution perpendicular to σ1 . A wide
pole distribution indicates a high driving fluid pressure ( Pf ≈ σ1 ).

Accordingly, the structural expression of hydraulic fracturing can vary from randomly oriented
extensional fractures (resulting in a breccia when Pf > σ1 ) through aligned extensional fractures to
shear fractures.

Note that this mechanism is independent of the depth at which it may occur. This process is routinely
applied in the petroleum industry to create fractures in low-permeability rocks. Besides, crack
propagation increases space for fluid whose pressure diminishes unless there is additional fluid
entering the system. Since crack propagation takes place only at some yield strength, hydraulic
fracturing typically occurs in pulses, when pore pressure reaches the necessary value.
Note also that tensile failure can occur in rocks without the aid of a high internal fluid pressure: for
example, during the contraction of a layer as a result of desiccation of sediment or the cooling of an
igneous body.
Faulting jpb, 2020
 73

Fluid pressure in shearing resistance

The shearing resistance of the rock is the shear stress on a potential fault plane that is just sufficient
to initiate fault movement. The conditions for sliding along a plane in fluid-saturated materials are
given by a simple modification of the Mohr-Coulomb criterion for failure (equation 6),
σS = c + µ ( σ N − Pf ) (12)

This equation has several inferences.

- A rock will exhibit essentially the same shearing resistance when σ N = 1 kbar, Pf = 0 and when
σ N = 2 kbar, Pf = 1 kbar , because the effective normal stress is 1 kbar in both cases. The effect
discussed here is one in which increased fluid pressure changes the state of stress throughout the
rock, which somehow reduces the shear stress required for faulting.
- While the normal stress σ N tends to strengthen the fault plane by pushing together the opposing
rock blocks, hence increasing friction, the fluid pressure acts to weaken the fault by pushing the
opposing rock blocks apart. In other words, increased pore pressure dampens down the effect of
confining pressure, which strengthens a fault as it increases with depth.
In both cases, pore pressure can be thought of as having a lubricating effect on faults in the sense that
it reduces frictional, shearing resistance to movement.

Faults with favorable orientations for slip or dilation present potential fluid flow pathways. Therefore,
recent research on earthquake control is much concerned with pore pressure effects because it is the
one critical variable that can be manipulated to some extent by man. The basic idea is that a suitable
local increase in pore pressure, brought about by pumping fluids into drill holes, could lower the
shearing resistance of rocks sufficiently to trigger local faulting and small earthquakes. Many such
small earthquakes may gradually release the stored energy that would otherwise accumulate for a
single large and catastrophic earthquake. A correlation between the rate of fluid injection and

Faulting jpb, 2020

74

earthquake frequency has been observed. These studies tend to confirm that small earthquakes can be
triggered or suppressed by judicious control of pore pressure, but it is not yet clear what effect such
control of small earthquakes will have on the occurrence of large ones.
High pore pressures are considered to play a vital role in some examples of low-angle thrust faulting.
Thrust sheets 30-100 km wide pose serious mechanical problems if they moved dry because friction
along their bases would seem to require either impossibly high shearing stresses in the sheets (if they
are pushed from behind) or unreasonably steep slopes (if they slide downhill under the influence of
gravity). Pore pressures higher than normal could allow thrust sheets to be pushed more easily or to
slide down slopes with dips as low as a degree or two. However, an alternative explanation for large
thrust sheets is that the basal shear zone, or detachment horizon, dominantly behaved as a viscous
material and not as a frictional material.
Equation (12), like the Coulomb equation for dry rocks (6), fits many experimental data. However,
this equation relates the shearing resistance entirely to the pressure of the pore fluid when chemical
properties of the fluid may also be important, for example, in controlling the rate of stress corrosion
at the tips of cracks.
Mode of deformation
A series of triaxial compression tests with a constant confining pressure but with various pore
pressures illustrates the influence of pore-water pressure on the behavior of porous rock. There is a
transition from ductile to brittle behavior as pore pressure is increased from 0 to higher values. Fluid
pressure lowers the elastic limit, thus increases the ductility field of rocks (i.e. their capacity for change
of shape without gross fracturing) in experiments. In that way, pore pressure influences the mode of
deformation, for example from ductile under moderate confining and fluid pressures to brittle under
similar confining pressure but high pore pressure. The current hypothesis is that when the frictional
resistance is higher than the shearing strength of rock, the rock shows ductile behavior. However, this
brittle/ductile transition is still incompletely understood.
Dry rock
If the pore pressure can dissipate, it tends to zero. The rock is most likely work-hardened.

Effects of confining pressure

Triaxial compression tests at various confining pressures show that with increasing confining pressure:
- the peak strength increases;
- there is a transition from typically brittle to fully ductile behavior; the confining pressure that
causes the post-peak reduction in strength is called the brittle-ductile transition pressure. The
brittle-ductile transition pressure (ca 50 MPa for marble) varies with rock type;
- the region incorporating the peak of the axial stress-axial strain curve flattens and widens;
- the post-peak drop in stress to the residual strength reduces and disappears at high confining stress.
- The ductility of most rocks (i.e. how much strain is accommodated before reaching ultimate
strength and failure) increases with increasing confining pressure. In the ductile regime, two lines
parallel to the σ N -axis in a Mohr diagram represent by the Von Mises criterion.

Faulting jpb, 2020

 75

The dependence of shearing resistance on normal stress is shown by a differential stress / confining
pressure diagram on which yield stress (arbitrarily taken as the differential stress at 2% strain) and
ultimate strength of rocks is plotted. For most rocks both the yield stress and the ultimate strength
increase almost linearly with increased confining pressure. In other words, rocks are stronger in
confined compression and an increase in pressure suppresses the formation of new fractures.

Faulting jpb, 2020

76

Effect of temperature
Temperature affects the mechanical properties of rocks much less than confining and pore pressures.
From above 200 to 500°C, rocks are ductile. The yield stress and ultimate strength are reduced with
increased temperature, but this reduction varies from one rock to another, depending on their
mineralogical composition. Heating usually increases ductility.

Effect of strain rate

Temperature affects the mechanical properties of rocks much less than confining and pore pressures,
and effects are rather limited from above 200 to 500°C, according to their mineralogical composition,
rocks are ductile. The yield stress and ultimate strength are reduced with increased temperature, but
this reduction varies from one rock to another. Heating usually increases ductility.

Effect of planar anisotropy

Rocks may contain well-developed planar anisotropy such as bedding and foliation. The intact rock
material may be stronger in the direction of maximum shear stress than along the anisotropy. In this
case, shear fractures are created along the planar anisotropy, therefore at angles that differ from those
estimated from the σN σS Mohr envelope. New shear fractures are forced by the orientation of the
plane of anisotropy whose lower shear resistance reduces the shear strength of the bulk material. This
textural weakening is a shear softening mechanism.
Peak strength
The peak strengths vary with the orientation of the plane of anisotropy (of weakness) with respect to
the principal stress directions. Experimental curves of differential stress at failure versus inclination
Faulting jpb, 2020
 77

of anisotropy are close to concave upward parabolas. The curves are lifted upward with increasing
confining pressure (increase in breaking strength) while shear fractures tend to develop at small angles
to the compression direction.

The analytical solution shows that the differential stress can be given by the equation:
2 ( c w + σ3 tan φw )
σ1 − σ3 =
(1 − tan φw cot β ) sin 2β
where c w = cohesion of the plane of anisotropy
φw = angle of friction of the plane of anisotropy
β = inclination of the anisotropy plane to the compression direction
The minimum strength occurs when:
tan 2β = − cot φw or β= 45° + ( φw 2 )
This shows that when the rock containing an existing weakness plane that is about to become a failure
plane, the rock has the lowest strength.
Orientation of shear fractures
In anisotropic rocks, the angle of faulting varies considerably, depending on the orientation of the
anisotropy with respect to the principal stress direction.
- Rocks compressed parallel to the anisotropy sustain the greatest differential stress and shear
fractures occur at less than 30° to the compression direction.
- Rocks compressed at 45 – 30 ° to anisotropy show the least strength.
- Shear fractures develop parallel to the anisotropy plane when the latter is inclined 15 to 45° to the
compression direction.

Faulting jpb, 2020

78

- Rocks compressed orthogonal to the anisotropy sustain the greatest differential stress and shear
fractures occur at 30° to the compression direction.

Slip tendency
Slip is likely to occur on a fracture when the resolved shear stress σ S on that plane equals or exceeds
the frictional resistance to sliding. Frictional resistance is proportional to the effective normal stress
σeN acting across that surface. The slip tendency Ts of a surface is the ratio of maximum resolved
shear stress to normal stress acting on that surface:
Ts = σS σ N
Ts is a measure of the relative likelihood that a fault or fracture will undergo slip. It is sensitive to both
the form of the stress tensor and the orientation of the studied surface. The maximum value of Ts is
limited by the slope of the sliding envelope. Whether or not a surface slips depends upon details of
local conditions such as rock or fault cohesive strength, if any, the coefficient of internal friction and
the orientation of the fault or fracture surface. Ts that causes slip on a cohesionless surface is often
referred to as the fault strength in earthquake focal mechanism analysis. Under most crustal conditions,
faults with Ts ≥ 0.6 are ideally oriented for slip (Byerlee law).

Dilation tendency
Dilation of fractures is largely controlled by the resolved shear stress, which is a function of lithostatic
and tectonic stresses and fluid pressure. The normal stress on a fracture depends on the magnitude and
direction of the principal stresses relative to the fracture plane. The ability of a fracture to dilate and
transmit fluid is directly related to its aperture, which in turn is a function of the effective normal stress
acting upon it. The normal stress can be computed for surfaces of all orientations within a known or
hypothesized stress field. This normal stress can be normalized by comparison with differential stress.
The resulting dilation tendency Td for a surface is then defined as:

Td = ( σ1 - σ N ) ( σ1 - σ3 )

Maximum depth for faulting in dry rocks - listric faults

The dependence of shearing resistance on normal stress leads to the prediction that the resistance of
dry rock to faulting should be greater at greater depths. This leads to the maximum depth for faulting
if the shear strength of a given rock type becomes so high that deformation can occur by other
mechanisms at lower stresses. In effect, the stress conditions required for flow by crystal slip are a
limiting factor verified experimentally.

Note also that curved faults may also result from a uniform stress orientation due to the effect of
downward increase in confining pressure. This change is shown again by the change in slope of the
Mohr failure envelope. Ductile flow dominates at 10-20km depth, depending on rock composition and
regional thermal gradients.

Faulting jpb, 2020

 79

Fault orientation relative to principal strain axes

Conjugate faults, whether they are thrust, and normal or strike-slip faults, occur under plane strain
conditions and intersect in a line parallel to the intermediate principal strain axis λ 2 . The conjugate
angle is normally an acute angle except where modified by internal rotations. The minimum principal
strain axis λ3 bisects the conjugate angle and occupies, along with the maximum principal strain axis
λ1 , a plane perpendicular to λ 2 . Minor structures like gashes and striations can be used to help to
define the principal strain axes. Tension gashes are oriented perpendicular to the axis of maximum
extension λ1 . Striations are parallel to the line of intersection of each fault with the ( λ 2 ; λ3 ) plane.

Faulting jpb, 2020

80

Fault orientation and anisotropic porosity

Anisotropic permeability in fractured aquifers arises from the abundance and distribution of
subparallel faults and fractures and permeability of associated damaged zones (fault breccias).

Faults and fractures orthogonal to the maximum principal stress tend to close, thereby reducing
permeability perpendicular to the regional compression direction. Conversely, faults and fractures
striking perpendicular to the minimum principal stress direction (i.e. parallel to the regional
compression) tend to open, which relatively enhances their permeability. Regional stress combined
with existing fractures thus influence the regional permeability and fractures favorably oriented for
slip or dilation in the ambient stress field tend to be the most active groundwater flow pathways. This
effect produces a transmissivity anisotropy.

Faulting History
Microfracturing and Dilatancy
When a rock is compressed and shortened at a constant rate in the laboratory, brittle deformation
begins before any fault forms across the sample.
Up to a stress difference of roughly half the fracture strength, the specimen shortens by a fraction of 1
% and its volume decreases by an even smaller amount, governed by the compressibility of constituent
minerals. Through this stage, the original dimensions of the specimen are recoverable and the behavior
is elastic.
At stress differences greater than about half the fracture strength, inelastic effects begin to be
noticeable. The volume of the specimen no longer decreases with continuing longitudinal shortening
but instead increases slightly; the specimen is now dilatant. Acoustic emissions monitored during
experiments and swarms of microseisms signal the opening of cracks and microfractures (the Griffith
cracks) with limited propagation on a granular scale. The intensity of microfracturing activity
increases while approaching the fracture strength. Microfracturing events become more frequent and
spatially more concentrated near the eventual fault plane. Such observations suggest that

Faulting jpb, 2020

 81

microfracturing and dilatancy may also precede fault displacements in nature. Shallow depth, dilatant
faults thus produce open spaces in which secondary deposits may crystallize.

According to Griffith’s theory, a crack properly oriented with respect to the directions of the principal
stresses propagates in the direction normal to the maximum tensile stress. The rate at which
microfractures grow is an important factor determining the rate or stress difference at which brittle
yielding will occur on larger scales. How propagating microfractures link up to give a macroscopic
fault is still poorly understood, even in rocks deformed under laboratory conditions. In some tests, for
example, extension microfractures form early and link up through suitably oriented grain boundary
cracks to form a fault. In other tests, however, through-going faults are thought to predate the
associated extension microfractures.
Fractures propagate from initial crack tips in three, or in a combination of the three mutually
orthogonal ways. These three basic modes of displacement and stresses are:
- mode 1 = opening mode: Tensile opening normal to the crack plane and propagation as
extension fracture along the plane of the original crack;
- mode 2 = sliding mode = down-dip slip: Displacement parallel to the crack plane and normal
to the crack tip-line with propagation as in-plane shear fracture along the plane of the
original crack;
- mode 3 = tearing mode = strike-slip: Displacement parallel to both the crack plane and its tip-
line with propagation as shear fractures.

Faulting jpb, 2020

82

Propagation - Interaction - Coalescence

Fault systems evolve by the growth and linkage of smaller, individual fault segments. Nucleation
describes the appearance of independent faults. Propagation describes the incremental growth of the
fault dimensions and/or displacement. Coalescence describes the process by which initially isolated
faults connect to form new and larger faults. Before faults are physically attached (hard linkage) to
become structurally and mechanically coherent, they propagate towards one another. Their stress
fields may already interact through an unfaulted relay zone of overlap (soft linkage) where elastic and
ductile strain is concentrated.

Hard linkage drastically changes the geometry of fault planes, which, conversely, do not significantly
change during soft linkage. The fundamental concept is that growth and connection of individual fault
segments take place as total displacement increases. The linkage and the resulting variations in
displacement along main faults affect the location and thickness of syn-faulting sediments at various
stages of fault development.

Faulting jpb, 2020

 83

Stress anomalies, in particular at fault tips, locally perturb both the magnitude and direction of the
regional stress field. The stress field in the overlap zone controls the growth of the overlapping faults,
hence the linkage geometry. There are four main types of fault interaction: intersection, interaction,
single-tip interaction, and double-tip interaction.

Faulting jpb, 2020

84

Displacement / Stress History

Rocks under stress undergo strain until they eventually break, forming a fault. Major faults show large
total displacements that have accumulated incrementally over a long period. Their history
encompasses interseismic periods of no-slip and seismic displacements.
Stable / unstable sliding
Fault creep is very slow slip, at typical average rates of a few cm/yr, without a perceptible earthquake.
This steady, “aseismic” displacement probably takes place under near-constant shear stress within
ductile shear zones and faults lubricated by clay minerals.

The observation that the slip of natural faults can occur with or without generating detectable seismic
waves is paralleled by the laboratory observation that frictional sliding between rock surfaces can
occur with or without detectable and sharp stress drops. Stable sliding occurs at a constant velocity
without jerks and stress drops. Unstable sliding, also called stick-slip behavior, repetitively occurs
with prominent jumps corresponding to episodic stress drops.
Stick-slip model
Seismic and interseismic faulting is compared to sliding of a weight resting on a flat and rough surface
(plastic deformation) and pulled laterally through a helical spring (elastic deformation). The pull
exerted by the spring must be larger than static friction to displace the weight.
The typical history is as follows (imagine also moving a piano):
If the surface is smooth, the weight glides at a more or less constant rate for a given stress, (an example
is pushing something on ice).
If the surface is rough, motion consists of two stages:
- The applied force first slowly increases but nothing happens to the weight. Shear stress increases
along the boundary between the weight and the slide surface as the system is loaded elastically, but
static friction temporarily keeps the weight from moving.
- When the tension in the spring reaches a critical value, the weight leaps ahead. The stored elastic
strain is released as the spring shortens and the basal shear stress drops. Back to memories of moving
a piano, it is well experienced that it takes more effort to start shifting an object than to keep it moving
once it is sliding. This is because the static friction sticking the weight to the surface is larger than the
dynamic friction opposing motion when it starts.
- Once basal stress has dropped to below dynamic friction the weight is stuck again.
- If loading is maintained, the cycle of alternate jerky sliding and stress release starts over.

Faulting jpb, 2020

 85

Coulomb law of sliding

Laboratory experiments on the sliding between two solids show that it follows, as a first
approximation, the Coulomb law of sliding (not to be confused with the Coulomb fracture criterion).
The mathematical expression is the same as equation (4):
σS = f ( σ N ) (4)

This is a macroscopic law. At the microscopic scale, the actual contacts only concern small fractions
of the surface, with a total area proportional to σ N . Microscopic processes will be different, depending
on whether the bumps are overcome through elastic, brittle or ductile deformation. The coefficient f
has a value at standstill (static friction coefficient µs ) larger than during sliding (kinetic coefficient
µ k ). Frictional sliding is such that:
(a) Either the shear stress equals or is lower than for static conditions, and there is no movement,
or
(b) it is equal and there is sliding with an undetermined speed.
This slip condition (see equation 5) is linear and passes through the origin of a Mohr diagram. The
Coulomb law of sliding accepts the relationship between shear stress and normal stress. Irregular
interseismic times can be replicated by varying the ratio of the friction force and the normal force (the
coefficient of static friction), which depends on some microscopic and often invisible properties
(asperities, strength, stability, etc.) of the fault surface. Reducing the normal stress reduces the static
friction to be overcome for sliding (unclamping effect), hence reduces the interseismic time.
Conversely, increasing the normal fault clamps the model fault and increases interseismic time.
Frequency of events
Stick-slip results from a familiar phenomenon: it is harder to start an object sliding than to keep it
sliding. This is because the static friction against sliding exceeds the dynamic friction that opposes
motion once sliding starts. The actual friction force that must be overcome depends on microscopic
details of how rough is the gliding surface. This means that many parameters cause the weight to be
stuck despite an applied force. All of these causes are metastable states because the stuck weight is in
a stable position, yet not the lowest energy state since friction induces elastic strain in both the floor
and the weight, and this strain corresponds to a certain amount of stored elastic energy. Among the
metastable states, the set of configurations the weight will visit while performing its jerky motion has
some particular importance. These states are marginally stable. A slight increase in the applied force
can lead to almost any response: the same amount of driving force yields sometimes a small forward
jump, sometimes a large one. The marginally stable states are believed to lack a typical time or length
scale, which leads to a correlation function that describes the frequency with which events occur. This
function is commonly a power law. The lack of a typical scale is similar to the configuration of a
thermodynamic system at a critical temperature. This notion has led to the recent concept of Self-
Organised Criticality.

Reminder: The critical behavior of thermodynamic systems is well understood. For all
temperatures, one can disturb the system locally and the effect of the perturbation will
influence only the local neighborhood. If a transition temperature is reached, the local
distortion will propagate throughout the entire system and something extraordinary
happens (e.g. change of phase, solid to liquid for water at 0°C). The system is critical in
the sense that all members of the system influence each other.
Stick-slip sliding on faults
It is now widely believed that the most important cause of earthquakes is large-scale stick-slip behavior
on faults. The seismic cycle is subdivided into (1) preseismic, (2) coseismic and (3) postseismic
phases:

Faulting jpb, 2020

86

1.) Stress and mainly elastic strain accumulate over a long preseismic period of tectonic loading
until
2.) Frictional resistance along the fault is overcome when sudden and violent coseismic failure of
the fault occurs until
3.) The stress drops to a value, the residual strength, at which friction prevents further slip.
Aftershock activity decays inversely with time.
4.) Postseismic creep and the formation and collapse of fault-related dilatancy drives transient
pulses of hydrothermal fluids.

As in the initial fracturing of rock specimens, there may be noticeable microfracturing activity
associated with and premonitory of each of the stress-drop episodes during frictional sliding
experiments. This possible 2.a) phase of preseismic anelastic deformation may involve foreshock
activity and accelerating precursory slip.
Stress increase remains low and the upper limit of the saw-tooth stress/displacement relationship is a
threshold known as the sliding friction. The usual magnitude of the stress drop associated with
earthquakes is in the range of 10 to 100 bars, with a logarithmic mean at about 30 bars (3 MPa).
Faulting and fluid displacement
Fault ‘valve’ models have been proposed, where increasing fluid pressures below sealed reverse faults
periodically trigger displacements with subsequently enhanced fluid flow and re-crystallization of the
fault zone (silicification or calcitisation, etc). Fault ‘pump’ models suggest that the coseismic collapse
of strain-induced micro-crack dilatancy within the host rock flushes fluid through the broken fault
zone rock. Post-seismic creep and compaction of newly formed cataclasite, in turn, lead to pore volume
reduction with fluid expulsion and vein emplacement in the adjacent wall rocks.
Variables influencing stick-slip behavior
Temperature, effective confining pressure, pore fluid chemistry, and rock type are among the
variables that determine whether stick-slip or stable sliding occurs.

Faulting jpb, 2020

 87

- High temperatures suppress stick-slip and this is one possible reason why earthquake foci are
restricted to shallow depths (less than 20 km). Stick-slip is nearly absent at low temperatures under
low effective confining pressure.
- At higher effective confining pressures, there is a positive effective normal stress on the fault plane
and stick-slip behavior appears, the magnitude of which is dictated by the frictional resistance to
sliding on the fault plane. The effects of pore fluid chemistry, as opposed to pore fluid pressure,
are not yet clear, but important effects may exist if stress corrosion at crack tips is a factor
governing the growth rate of microfractures.
Slow slip: slow earthquakes
Instrumental records, often combining a range of methods (seismic, GPS, long-term strain
measurements in boreholes), indicate that other processes than dynamic rupture (i.e. seismic rupture
lasting seconds to minutes) release the strain energy accumulated in the Earth’s crust. “Aseismic” slip
(i.e., rupture propagating over hours to several months) describes several types of usually
discontinuous events such as episodic, non-volcanic tremor or almost imperceptible, low-frequency
rumblings.

Slow slip is still poorly understood and its occurrence vs. faster seismic rupture is variably attributed
to the fault structure (e.g. shape and asperities), fault material (e.g. gouge composition, fluid pressure)
and depth (rather shallow in the crust). Slow slip involves deformation rates expected to develop
brittle-ductile structures in the fault zones, which geologically remain a discontinuity in the crust. The
process has been documented for all types of faults and tectonic regimes.

Fault segments
Friction forces vary at different points on a fault. The instantaneous rates of slip may, therefore,
strongly fluctuate, ranging from zero to rates comparable to the elastic wave velocity of the
surrounding rock (ca 7 km.s-1). Thus, an earthquake may emanate from the fault portion where the slip
rate is temporarily very high while, simultaneously, another portion of the same fault is locked or is

Faulting jpb, 2020

88

only exhibiting slow, aseismic creep. An analogy is provided by a series of blocks connected by
springs and pulled from one extremity. Each spring represents a segment. The complex interplay of
relative movements between blocks simulates the diachronous and segmented movement on the fault.

The average slip rate of major faults over long periods is of special interest in tectonics. It may be
estimated geologically by using several methods to obtain the displacement and dividing by some
estimate of the elapsed time. Average slip rates may also be obtained on some faults (e.g., oceanic
transform faults) by applying the theory of sea-floor spreading. Two methods are available for making
relatively short-term determinations of slip rate. The first is to carry out repeated surveys of networks
of geodetic stations either side of a fault. The second is a simple seismological method in which the
surface wave magnitudes of all earthquakes occurring on a given fault in a given interval of time are
used to estimate the total slip on the fault during that time and, hence, the average slip rate. Calculated
slip rates are in approximate agreement with geodetic measurements. In some cases, the slip rates
indicated by this method are lower than those indicated geodetically, suggesting either that some
additional slip is occurring by creep or that strain is accumulating for an earthquake. Besides,
structures and topographic features observed along faults indicate that the pre-faulting strain is not
entirely recovered during a seismic event. Aftershocks also indicate that the total amount of released
strain/energy is not instantaneous.

Sliding along existing faults

In natural conditions, failure is likely to take the form of sliding along pre-existing fractures rather
than opening new ones. Sliding along pre-existing fractures does not have to overcome cohesion, only
sliding friction. In such conditions, the relationship between σS and σN is equation (5):
σS =σ N tan φ (5)
a line that goes through the origin in a Mohr diagram, and, therefore may intersect a Mohr circle of a
given state of stress in two points. Sliding will be possible along any surface oriented at angles with
σ 3 that corresponds to the intercepted arc of the Mohr circle. This relationship shows that increased
pore pressure, which shifts Mohr circles to the left, may reactivate existing fractures.

Exercise
Draw a Mohr circle representation with σ1 = 100 MPa, σ3 = 50 MPa, for a rock with
a cohesion strength of 20 bar and a failure surface whose slope is 25°. Draw the same
body with a pre-existing fracture.
Faulting jpb, 2020
 89

As the coefficient of friction is relatively well defined for most rocks, it can be shown that frictional
sliding will occur on pre-existing faults when σ1 σ 3 ≈ 3 . Faulting is controlled by the vertical
principal stress σ v , and one horizontal principal stress, σ 3 in extension and σ1 in compression. In
cases of hydrostatic pore pressure, these relationships show that in extensional areas σ 3 ≈ 0.6 σ v , in
compression areas σ1 ≈ 2.3 σ v , and in strike-slip faulting areas, when σ v ≈ ( σ1 + σ3 ) 2 , then
σ1 ≈ 2.2 σ 3 . In situ stress measurements to depths of about 2 km at intraplate sites have confirmed
these simple equations.

Cataclastic flow
Fluid overpressure enhances hydraulic fracturing and cataclastic flow. Cataclasis refers to grain
fracturing while friction between grains is sufficiently large to inhibit sliding on grain boundaries.
Cataclastic flow refers to deformation during which distributed microfractures at grain scale produce
clasts that frictionally slide past each other and rotate. Grain crushing happens at such a minute scale
that the macroscopic structure of the rock seems to results from flow in a ductile manner. However,
microfractures and grain rotation are brittle, cataclastic features. Consequently, fault movement takes
place across a commonly narrow, cataclastic fault zone. Owing to these numerous microcracks, hence
porosity and permeability, fault zones tend to channelize the fluid pathways. Channelized fluid flow
weakens the faults considerably, which accounts for the concentration of brittle deformation into
highly localized planar discontinuities.

Locking
Many cataclastic processes may increase the number of asperities, i.e. points and areas where fault
rocks are in contact. The total area of contact may cause high resistance to sliding, to the level where
adhesion prevents further faulting. Adhesion is mechanical, due to interpenetration and locking of
fault blocks, and chemical due to secondary crystallization bonding surfaces along the fault zone.

Faulting and earthquakes

Studying fault movements and faulting is important to get a better understanding of earthquakes
because earthquakes are thought to release strain and stress on pre-seismically locked faults.

Elasto-plastic model
Earthquake reconstruction uses a spring pulling a mass resting on a rough surface. The pull on the
spring must exceed the frictional resistance to sliding of the mass for the latter to move. Before any
slip, one can pull the spring back and forth; this is elastic behavior. But once the mass slips it does not
move back (the displacement is permanent) while the spring recovers some or all of its strain.

Model relevance
Plate movements entail relative movement between adjacent crustal blocks. Consequently, two
reference points taken away, on both sides of the bounding fault zone move with respect to each other.
These relative movements are continuous in the lower, ductile crust where fault creep is permanently
active. The lower crust carries the upper, brittle crust in which the shallower section of the fault does
not instantaneously follow the deeper, steady, ductile movement. The brittle fault section is locked
while elastic strain accumulates with a gradient centered on the fault plane (a straight line across the
fault, initially joining the two distant points, becomes sigmoidal). This portion of the seismic cycle,
during which strain and stresses increase with time along the fault, corresponds to the interseismic
interval. When the local stresses exceed the frictional strength of the fault zone, slip suddenly occurs
with an earthquake along a newly created or reactivated fracture. The upper crustal rocks
instantaneously release their accumulated elastic strain (the sigmoidal line snaps back to its original
straight shape but is broken by the fault). This coseismic strain reduction is the elastic rebound. A

Faulting jpb, 2020

90

new earthquake cycle begins and will end with the next seismic slip. The rapid coseismic slip
increases stress in the ductile crust. As these stresses are relieved after the earthquake, either through
aseismic fault slip or through more broadly distributed viscous flow at depth, the upper-crustal fault
is re-stressed and the fault region re-strained. This phase corresponds to the postseismic period, which
grades into the steady interseismic stress accumulation progressing to the next earthquake.

Accordingly, faults are seismic sources. The elastic behavior of rocks and the frictional properties of
faults govern brittle deformation. The earthquake cycle provides an important framework for
forecasting earthquakes and thereby mitigating their effects. In theory, one can predict the next
earthquake on a particular fault from the deformation that accompanied the previous earthquake. At
hand are the slow rate of permanent deformation determined from geologic observations (millimeters
to centimeters per year), the repeat times between earthquakes on the same fault segment (earthquake
recurrence, ranging from tens to several thousands of years) and the rate of interseismic strain
accumulation determined from geodetic observations. In practice, this prediction has proven quite
difficult because of the generally long time between earthquakes and the relatively short history of
geodetic measurements. Besides, the cyclic deformation is highly idealized. Because of the paucity of
appropriate observations, it is still not known whether interseismic strain accumulation or permanent
deformation can be characterized by uniform rates on the time scale of individual earthquakes. In
addition, fault properties and irregularities change during each seismic event, so that the total friction
resistance to movement changes and alters the interseismic interval. In any case, the identification of
seismic sources in an area is the first step in evaluating the earthquake risk. Because rocks have typical
strength and plate tectonic frameworks have rather steady rates, particular faults tend to generate
characteristic earthquakes with the same maximum magnitude.

First motion
Seismic waves are vibrations generated by the rupture and sudden movement of rock at the
hypocenter, or focus, of the earthquake. Two types of elastic, seismic waves are produced:
- Body waves radiate through the Earth’s interior in all spatial directions from the source region.
- Surface waves propagate along the Earth’s surface.

Faulting jpb, 2020

 91

P (primary) and S (secondary) body waves have different characteristics in different directions and
travel within the Earth faster than surface waves. Therefore, P waves are the first to arrive from a
distant earthquake. The first motion they impel to the ground (the impetus) is divided into two classes:
- Upward, thus away from the earthquake source, which means that the related fault movement has
pressed and pushed away all material points along the expansion direction from focus to the record
point.
- Downward, thus towards the earthquake source, which means that the related fault movement has
depressed and pulled in all material points in the contraction direction of propagation.
Hence, the direction of the first ground motion at a particular place depends on its position relative to
the hypocenter and the type of faulting that occurred, i.e. whether the fault is moving rocks toward or
away from the direction of the seismic station. In conclusion, the first-motion study of earthquakes
can be interpreted in terms of fault orientation and displacement direction.

S waves produce a sideways shearing motion at right angles to the direction of propagation. Surface
waves are the most damaging and consist of a complex horizontal and rolling (Rayleigh) motions.

Focal mechanism solutions

The direction of elastic forces released during an earthquake can be determined from the first motions
recorded at as many seismograph stations as possible around the earthquake.
Concept
Upward first motion releases compression at the recording point; conversely, downward first motion
expresses extension (dilation). Neither push (compression) nor pull (dilation) occurs along the fault
plane because slip is shearing motion only. This is also valid for the non-material auxiliary plane,
which is normal to both the fault plane and the slip direction. The fault plane and the auxiliary plane
(the nodal planes) divide the space around the focus into four alternate quadrants where the first
motion is a push or a pull. This quadrant distribution is characteristic of most natural earthquakes.
Implementation
The goal is to have the first compressional/dilational ground motion recorded by a sufficient number
of widely spaced seismograms in different directions from the earthquake. Seismic stations recording
very small or no first motion align along nodal lines. Generalized to three dimensions, these lines
construct the nodal planes.
The standard procedure is as follows:
- Plot on lower-hemisphere, equal-area projections the oriented station points (considering wave
trajectories in the spherical Earth) where the first P waves had push (downward first-motion of P
traces on a seismogram) or pull (upward first-motion of P traces on a seismogram) character.
Faulting jpb, 2020
92

- Identify the pair of orthogonal nodal planes that best divide the space around the focus into
compressive and dilatative quadrants.
- If faulting caused an earthquake, then one of the nodal planes is the fault plane and the other is the
auxiliary plane.

The resulting plot is a focal-mechanism (or fault plane) solution. The direction of displacement lies
within the fault plane and is perpendicular to the auxiliary plane. It is the pole to the auxiliary plane.
Hence, the slip direction and type of movement along the fault can be determined easily from the
arrangement of compressional (conventionally shaded) and dilatational (shown blank) quadrants. Note
also that the radius of the focal sphere has no value.
Application
The direction of compression (P-axis, the bisector of dilational quadrants) and the direction of tension
(T-axis, the bisector of compressional quadrants) derived from the quadrant distribution are
information on the stress field responsible for the considered earthquake. Accepting that P and T axes
approximate the directions of maximum and minimum principal stresses, respectively, one can define
the nature of the fault.
One problem is that a solution for an earthquake yields two nodal planes that both are possible faults
since both have the same first motion. This ambiguity is an inherent characteristic of focal mechanism
solutions based solely on P-wave impetus. In principle, it can be overcome by the study of S-wave
characteristics or by other means. In practice, geological knowledge and/or observation, especially the
orientation of surface breaks associated with an earthquake, quickly help to distinguish the fault plane
from the auxiliary plane.
Another problem is that a pre-existing fault plane controls the seismic radiation pattern more than the
in situ stress field: the earthquake focal plane mechanism always has the P- and T-axes at 45° to the
fault plane and the B axis in the plane of the fault.
To use earthquake focal mechanisms to determine stress orientations, one considers average P-, B-,
and T-axes for earthquakes occurring on different faults within a limited region.

Faulting jpb, 2020

 93

Furthermore, the amplitude and frequency content of seismic waves only give information about the
magnitude of stress released in an earthquake (stress drop) and not the absolute stress levels. In
general, stress drops are about 1–10 MPa, a very small fraction of the shear stress that causes the
earthquake.
In conclusion, a fault-plane solution is a method for using the seismograms from an earthquake to
study the geometry and sense of motion on faults. This method permits interpretations of present-day
movements on deeply buried or otherwise concealed faults, especially in the oceans, and has proved
important by enabling the relative motions of lithospheric plates to be determined.

Attention: News media generally report the location of the epicenter, which is the point on the surface
of the Earth directly above the focus; geophysical information includes the location of the epicenter
and the depth to the focus. A focal mechanism solution is obtained by first determining the location
of the focus.

Magnitude and fault movement

Geophysical information also includes the magnitude, a measure of the strength of the earthquake.
The Richter magnitude (M) of an earthquake was first determined as the largest amplitude A of
seismic waves recorded on a seismogram normalized to a local reference amplitude A 0 all expressed
on a logarithmic scale and multiplied by an empirical factor of 3:
M = log ( A A 0 )
Since the Richter magnitude is a logarithmic scale, a whole number unit increase on the scale represents a
ten-fold increase in vibration amplitude.
The Gutenberg-Richter power law states that the average number n of earthquakes with magnitudes
≥ M recorded per year by a particular seismic network is a decreasing exponential function of M:
log n = a − bM
with a and b constants characteristic of the particular region. This law describes the frequency of
different sized earthquakes. In most regions, b is about 1 so that earthquakes generally become nearly
10 times more frequent for every unit decrease in magnitude (for instance, there are 10 times fewer
earthquakes of magnitude 5 than earthquakes of magnitude 4). The Gutenberg-Richter rule is
acceptably valid for earthquakes at epicentral distance < 600 km and magnitude M < 7.
Magnitude is a measure of the amount of energy released, which is proportional to the fault
displacement. For earthquakes of magnitude >7, the moment magnitude (Mw) is used. It accounts for
the average amount of slip on the fault that produced the earthquake, the rupture area and the shear
modulus of rocks that failed. The Richter and moment magnitude scales are comparable over the range
of magnitude 3 to 7.
The energy E released during an earthquake is given in Joules by:
E = 1011.8+1.5M
A magnitude increase of 2 is equivalent to an energy increase of approximately 1000 times.

Displacement and rupture length

The ratio between seismic slip us and the rupture length Ls is considered to be:
10−4 > u s Ls > 10−5
The ratio between total displacement u t and total fault length L t is estimated as:
u t L t ≈ 0.1
Case studies have shown that the rupture point propagates along the fault plane, generally and mostly
in one direction, unzipping locked portions of the fault at a rate of ca 3km/s.

Faulting jpb, 2020

94

Earthquake intensity - Material amplification

The intensity of an earthquake is related to local damage, duration and other effects on structures and
people. Earthquake intensity is scaled from I to XII on the qualitative Modified Mercalli scale.
Intensity varies with distance from the epicenter and from place to place within the shaken area.
Earth materials have different elastic properties and, therefore, respond differently to seismic waves.
The amplitude of shaking is increased in unconsolidated sediments because the energy of a seismic
wave remains practically constant when passing from one rock type to another. Since the seismic
velocity is lower in unconsolidated than in consolidated rock, keeping energy constant implies that the
displacement amplitude of the waves must be bigger. This material amplification is important for
vulnerability evaluation. It is often the reason why buildings on unconsolidated rock are destroyed
while buildings in the surrounding consolidated rock are not. Besides, shaking may produce an
increase in pore-water pressure of water-saturated weak sediments, which results in sediment
liquefaction and flow.

Slip rate and recurrence

Any fault has a long, discontinuous movement history. The slip rate is the ratio of the amount of
displacement to the time interval over which that displacement took place. The average recurrence
interval is the average time interval between earthquakes.
However, these may vary in time, which casts suspicion on average values. Frequently, earthquakes
are clustered events separated by relatively long periods of quiescence. The duration of the period
between two earthquakes along the same fault has been attributed to several physical parameters such
as the relative velocity between the two walls of the fault, the mineralogy, and anisotropy of the fault
walls and the fault rocks, the morphology of the fault plane, the thermal state, the fluid pore-pressure,
and others.

Reactivation
Recorded seismicity is largely due to the reactivation of existing fault planes. Conditions for
reactivation depend on frictional properties, which in turn depend on fluid pressure and the fault
orientation with respect to principal stresses. Two-dimensional analyses with σ 2 parallel to the strike
of the fault show that reactivation is relatively easy for an effective stress ratio ( σ1 -Pf ) ( σ3 -Pf ) ≈ 4
and low angles (ca 15 to 45°) between the fault plane and σ1 . Faults at a high angle are locked unless
the fluid pressure is sufficiently high to virtually eliminate friction. It is easier to create new slip planes
if existing faults have unfavorable conditions.

Faulting jpb, 2020

 95

Exercise
Combine equations (3) and (5) to express the stress ratio in function of the angle
between the fault plane and σ1 ; draw the curves for friction coefficient 0.6 and 0.85.

Conclusion
Experimental results can be summarised as follows:
- Fractures are generated when the effective stress exceeds the tensile strength of the rock.
- At a given temperature and strain rate, increased confining pressure increases both the yield stress
and ultimate strength;
- At a given confining pressure, increased temperature or decreased strain rate lowers both yield
stress and ultimate strength.
In short, brittle deformation is highly pressure-sensitive.
The simple and widely applicable Coulomb criterion predicts a linear strength-pressure relationship.
At higher confining pressures where ductile behavior begins, or at very low values of σ 3 where other
special effects enter, plots of σ N against σS at failure are characteristically concave toward the σ N
axis. It is the empirical Mohr failure envelope tangent to Mohr circles at different σS and σ N at
failure. This line separates stable and unstable stress states, the latter giving rise to brittle deformation.
The mechanistic Griffith failure criterion describes the tensile failure. It assumes that real materials
contain slit-like imperfections (Griffith cracks) that open, propagate and link, which leads to
macroscopic rupture.
The increase in fluid pressure reduces the effective stresses and brings the stress state closer to the
failure envelope.
Faulting occurs to relieve accumulated shear stress on faults. The sudden movement of strained crustal
blocks that break along new or pre-existing faults causes earthquakes. Earthquake displacement does
not occur over the whole length of a large fault: a small area of a fault displaces at one time with other

Faulting jpb, 2020

96

areas displacing at another time. However, earthquake displacement at one particular area along a fault
may occur at regular recurrence intervals.
The deformation of the upper crust is mostly accommodated by important displacements along fault
zones. Such large displacements result from seismic slip increments representing reactivation events
of faults over lengthy periods. Stress cycling accompanying episodic reactivation is likely to be the
norm in the vicinity of major faults whose seismic cycles involve long periods of elastic strain and
stress accumulation, driven by aseismic ductile deformation at depth, ultimately released by sudden
fault slip events. As a corollary, earthquake focal mechanisms are the most ubiquitous indicators of
stress in the lithosphere.

Recommended literature

Hubbert M.K. & Rubey W.W. - 1959. Role of fluid pressure in mechanics of overthrust faulting: I.
Mechanics of fluid-filled porous solids and its application to overthrust faulting. Geological
Society of America Bulletin. 70 (2), 115-166, 10.1130/0016-7606(1959)70[115:ROFPIM]2.0.CO;2
Jaeger J.C. - 1969. Elasticity, fracture and flow: with engineering and geological applications. third,
Methuen & Co LTD and Science Paperback, London. 268 p.
Jaeger J.C. & Cook N.G.W. - 1979. Fundamentals of rock mechanics. 3, Chapman and Hall, London.
593 p.
Jaeger J.C., Cook N.G.W. & Zimmerman R.W. - 2007. Fundamentals of rock mechanics. Fourth
edition. Blackwell Publishing, Oxford. 475 p.
Jolly R.J.H. & Sanderson D.J. - 1997. A Mohr circle construction for the opening of a pre-existing
fracture. Journal of Structural Geology. 19 (6), 887-892, 10.1016/S0191-8141(97)00014-X
Mandl G. - 1988. Mechanics of tectonic faulting. Elsevier, Amsterdam. 407 p.
Mandl G. - 1999. Faulting in brittle rocks. Springer, Berlin. 434 p.
Savage J.C., Byerlee J.D. & Lockner D.A. - 1996. Is internal friction friction? Geophysical Research
Letters. 23 (5), 487-490, 10.1029/96GL00241
Scholz C.H. - 1992. The mechanics of earthquakes and faulting. 2nd, Cambridge University Press,
Cambridge. 439 p.
Sibson R.H. - 1989. Earthquake faulting as a structural process. Journal of Structural Geology. 11 (1-
2), 1-14, 10.1016/0191-8141(89)90032-1

Faulting jpb, 2020