93

FAULTS
Fractures are planar discontinuities, i.e. interruption of the rock physical continuity, due to stresses.
The geological fractures occur at every scale, so that no large volume of rock has any. These
discontinuities are related to sudden elastic relaxation of elastic energy stored in the crust.
The geological fractures have their economic importance. The loss of continuity in intact rocks
provides the necessary permeability for migration and accumulation of fluids such as groundwater
and petrol. Fractured reservoirs and aquifers are typically anisotropic since their transmissivity is
controlled by the conductive properties of fractures, which the local stress field partially controls.
Geological fractures may be partially or wholly healed by introduction of secondary minerals, often
giving rise to ore deposits, or by recrystallization of the original minerals.
Planar discontinuities along which rocks lose their cohesion during their brittle behaviour are
called:
- joints if there is no component of displacement parallel to the plane (there may be some very
small orthogonal parting; joints are extension fractures).
- faults if rocks on both sides of the plane have moved relative to each other, parallel to the plane
(faults are shear fractures).
- veins if the fractures are filled with secondary crystallisation.
Joints and faults divide the rocks in blocks that must be taken into consideration for engineering,
quarrying, mining and geomorphology.

Fault terminology
Definitions
Faults are defined when two adjacent blocks of rock have moved past each other in response to
induced stresses. The notion of localized movement leads to two genetically different classes of
faults reflecting the two basic responses of rocks to stress: brittle and ductile.
Brittle fault
A fault is a discrete fracture between blocks of rock that have been displaced relative to each other,
in a direction parallel to the fracture plane. A fault zone is a region containing several parallel or
anastomosing (i.e. branching and reconnecting) faults. Any fault-bounded sliver in a fault zone is a
horse. Fault and fault zones are identified either when an earthquake occurs or by geological
mapping showing that motion across a discontinuity has occurred in the past. On geologic maps,
only faults that affect the outcrop pattern are usually shown.

Faults

jpb, 2014

94
Ductile fault
Shear zones are the analogues in a ductile material of faults in a brittle material. Shear zones are
regions of localised but continuous ductile displacement, formed under conditions of elevated
temperature and/or confining pressure, in contrast to fault zones that are regions of localised brittle
deformation. Shear zones are thus ductile faults, by contrast to the brittle faults.

Geometrical classification
Fault plane
Faults dipping more than 45 are called high angle faults; Faults dipping less than 45 are called
low angle faults. In general, fault surfaces are curved. Undulation of fault-surfaces is commonly
seen in 3D seismic data. The fault corrugations thereby identified are attributed to the linkage of
fault-segments through time. A listric fault is a curved, concave upward fault, that is, it gradually
flattens with depth.
Where low-angle faults affect a set of nearly horizontal bedded rocks, they generally follow a
staircase path made up of alternating ramps and flats. The flats are where the overlying rocks slide
along a relatively weak bedding plane also called a dcollement plane, which refers to a surface
across which there is a discontinuity in displacement, strain or fold style. The ramps are fault
sections climbing through the stratigraphic sequence, typically at around 30 to the horizontal,
across stiff, competent layers. Ramps do not necessarily strike perpendicular to the movement
direction (frontal ramp) but are also found oblique or parallel to the transport direction (lateral
ramp or tear fault).

Faults

jpb, 2014

95

Most long faults are segmented, each segment having its individual history; fault segments are
usually not coplanar. The fault that intersected the ground surface while it was active is an
emergent fault, by opposition to blind faults that did not break the surface. Emergent faults
produce a topographic step, the fault scarp.
Attention: An emergent fault is not an exhumed fault. Make the difference!
Erosion reaching a fault makes a fault trace on the Earths surface.

Fault blocks
The rock immediately above and below a non-vertical fault or shear zone is referred to as the
hanging-wall and the footwall of the fault, respectively. Rocks that have been translated great
distances away from their original site are allochthonous. They come to rest on autochthonous
rocks, which have retained their original location. Parautochthonous refers to locally transported
rocks.
Growth faults
Growth faults are faults that operate during sedimentation and therefore displace an active surface
of sedimentation. They are dominantly listric. They form characteristically, but not exclusively, in
unconsolidated sediments freshly deposited in basins that are actively growing in breadth and depth.
Faults

jpb, 2014

96

As a clear result, sediments or specific layers are thicker on the hanging-wall than sediments and
layers of the same age on the footwall.

Kinematic classification
Slip
Slip is taken as the direction of movement of the hanging wall relative to the footwall. The
displacement vector connecting originally coincident points (the piercing points) on opposite sides
of the fault plane is called net slip. Its length provides the amount of displacement on the fault,
which generally is the addition of several movements.

Faults

jpb, 2014

97
The components of the net slip parallel to the strike and dip of the fault are the strike slip and the
dip slip. The rake is the angle measured within the fault plane down from the strike direction to the
line of slip. The plunge is the angle measured in the vertical plane that contains the slip line
between the horizontal in this plane and the slip line.
The offset shown by a planar feature in a vertical cross section perpendicular to the fault is called
the dip separation. The vertical component of the dip separation is the throw and the horizontal
component (perpendicular to the fault strike) is the heave. Notice that the dip separation is not
equivalent to the dip slip, the former depending on the orientation of the offset surface as well as on
the nature of the fault displacement.
Note: A bedding surface alone can never be used to determine slip
Faults are classified according to the direction of the relative movement between fault blocks,
which is related to the type of stress causing the fault.
Fault classes
Normal fault
A normal fault is a high angle, dip slip fault on which the hanging-wall has moved down relative
to the footwall. A normal fault brings younger rocks over older ones. Because of the separation of
geological horizons that results from normal faulting, such faults are also termed extension faults.

Extensional ramps cut down section in the direction of transport and are termed detachments
although a typical detachment has no roots and follows a stratigraphic horizon. A normal fault with
a dip less than 45 is sometimes called a lag or a denudation fault.
Reverse fault
A reverse fault is a dip slip fault on which the hanging-wall has moved up and over the footwall.
Consequently, old rocks are brought over younger ones. Such faults exhibit a repetition or overlap
of a geological horizon, and are accordingly also termed compression fault.

A thrust fault is a low-angle reverse fault along which the hanging wall forms thrust-sheets
(nappes) of allochthonous rocks emplaced over the autochthonous or parautochthonous footwall.
Most common, thrust faults ramp up section towards the surface in the direction of tectonic
transport.
Faults

jpb, 2014

98
Strike slip fault
Strike slip faults usually have very steep or vertical dips and the relative movement between the
adjacent blocks is horizontal, parallel to the strike of the fault plane. Large strike slip faults are also
referred to as transcurrent faults and wrench faults.

The sense of the strike-slip displacement on a fault is described by the terms sinistral and dextral.
A fault is sinistral if, to an observer standing on one block and facing the other, the opposite block
appears to have been displaced to his left. Conversely, if the movement is to the right the fault is
dextral.
A transfer fault is a strike slip fault that transfers displacement between two similarly oriented
fault segments (e.g. two normal faults). Transfer faults are usually confined to hanging walls of
detached systems (i.e. not affecting the basement) and terminate where they connect the linked
faults. Transfer faults and zones are lateral ramps that may accommodate differential displacement
and/or strain in adjacent blocks (different amounts of shortening or extension on both sides of the
fault).

Assuming that thrusts and normal faults strike at a high angle to the slip direction, transfer faults
linking two thrusts or normal faults are therefore nearly parallel to the movement direction.
Accordingly, transfer faults usually have strike slip components that vary along strike if
displacement changes across the transfer zone. Transfer zones (faults) usually terminate where they
connect to and terminate other faults or structures.
A tear fault is a strike slip fault that runs across the strike of a contractional or extensional belt and
accommodates differential displacement between two segments of the belt. Such faults are usually
confined to hanging walls of detached systems (i.e. not affecting their basement).

Faults

jpb, 2014

99

Transform fault
A transform fault is a strike slip fault at plate boundaries. There are three types:
Ridge-Ridge transforms link two segments of a constructive plate boundary.
Trench-Trench transforms link two segments of a destructive plate boundary.
Ridge-Trench transforms link a constructive plate boundary to a destructive one.

Ridge-Ridge transform faults are most common. They are fracture zones striking at nearly right
angles to the mid-oceanic ridges and seemingly offsetting the ridges. However, they differ from
transcurrent faults in that the direction of horizontal movements is in the opposite direction to that
required if the faults were strike slip faults responsible for offsetting the ridges after the latter were
formed.
Transform faults are active between the ridges and dead beyond the offsets, and they are parallel to
small circles centred at the poles of rotation of the plates. Displacement across them is much greater
than the length of the active segment.
Faults

jpb, 2014

100
Hybrid fault
The terms normal fault and reverse fault, while strictly defined for faults with zero strike slip
displacement, can also be used for faults with small strike slip displacements accompanying much
larger dip slip displacements. Where the strike slip and dip slip displacements are similar in
magnitude, the fault can be called an oblique slip fault.
Scissors fault
One fault block can rotate around an axis perpendicular to the plane of scissors faults.

Topographic effect
Recent vertical components of fault movement produce linear topographic steps, or scarps. Fault
scarps may be dissected by erosion to yield triangular faceted spurs.

Length/throw ratio
The general elliptical form of single-event movement planes suggests some relationship between
the maximum length of the fault plane (ellipse long axis) and its maximum down-dip height
(ellipse short axis). Slip distribution is further considered to be symmetrical about a central point of
maximum slip. This geometrical vision has led to mechanical models based on the assumption that
Faults

jpb, 2014

101

rocks are homogeneous elastic materials. These models relate the maximum displacement ( D max )
at the fault midpoint to the length ( L ) of the fault. Their general expression is:

2
D max 2 1
=
( d C y )
L
E
in which d is the shear driving stress (i.e. the shear stress leading to Coulomb frictional sliding),
y is the yield strength of the intact rock at the fault tip and C is a variable or a function that

specifies how the theoretical stress singularity is removed at the fault tip ( C= 1 in linear
displacement models). E and are the Youngs ratio and the Poissons ratio of the rock,
respectively.
The general equation given above shows that the maximum displacement/length ratio of faults
reflects three properties of the host rock: its shear modulus, its elastic strain limit and the shear
driving stress. Typical ratios for isolated, small normal faults in sedimentary rocks range from 0.002
to 0.04, with an average of 0.01. Ratios of 0.4 to 0.004 and an average of 0.02 were reported for
strike-slip faults in turbidites. Lithological variations may account for the range of values, with
weak rocks allowing higher strain gradients at the fault tips, hence higher D max L ratios than
stronger rocks.
3D seismic data indicate that the D max L ratio varies over a limited range around a nearly constant
value that depends on the tectonic setting:
*
D/L= 12-40: Several fault sets, faults frequently abut against each other
*
D/L= 25-75: One dominant fault set. Normal or reverse faults with a limited strike-slip
component
*
D/L= 50-150: Large component of strike-slip and also for shallow levels in deltaic (growth
fault) settings.
Systematically smaller maximum displacements for normal and thrust faults on Mars (by a factor of
ca. 5) and Mercury are related to the reduced gravity on these planets relative to the Earth.

Fault activity
Although every fault has moved and can be reactivated, geologists have developed a qualitative
three-term classification to appreciate seismic risks.
- An active fault has moved during the past 10 000 years.
- A potentially active fault has moved during the Quaternary
- An inactive fault has had movement older than the Quaternary.
However, it is difficult to prove that a fault is active without historical record of earthquakes on the
fault. Any fault is a weakness "capable" of reactivation.

Fault associations
Faults are rarely isolated. Subsidiary faults of lesser extent often accompany the major, master
fault. They are usually found in groups of the same type, often parallel and dividing the area in
blocks. The faults have the same sense of displacement.

Conjugate faults
Major blocks may be bounded by sets of conjugate faults, which means that faults of the same type
and formed during the same deformation episode occur in two symmetric sets with parallel strikes,
opposite dips and opposite or reciprocal sense of movement to each other.
Triaxial experiments (the three principal stresses have non-zero magnitudes) show that MohrCoulomb shear fractures (i.e. faults) are oriented systematically with respect to stress directions.

Faults

jpb, 2014

102

Conjugate faults intersect in a line parallel to the intermediate principal stress axis 2 .
The greatest principal stress 1 bisects the acute angle between the conjugate faults.
Striation orientations on a given fault are verified to be movement directions defined by the
intersection of the fault surface with the ( 1, 3 ) plane.

- The material shortens parallel to 1 and expands parallel to 3 .

These observations are the basis for a dynamic interpretation of fault systems. In addition,
Anderson emphasised that the earth surface is a free surface with a fluid, the air, and fluids are
unable to support any shear stress (it is a physical definition of fluids). Therefore, the earth surface
is a principal plane of stress (remember that a principal stress is per definition orthogonal to a noshear plane). Assuming a bulk horizontal attitude of the surface of the earth (which is nearly true in
low relief regions), one of the three principal stresses is close to the vertical. The type of conjugate
fault that develops near the surface depends on which of the three principal stresses is sub-vertical:
1 vertical: Normal faults dipping 60.

2 vertical: Vertical strike-slip faults.

3 vertical: Thrusts dipping 30.
This interpretation involves that the vertical stress is the lithostatic pressure and that regional stress
variations are due to changes in magnitude of the horizontal stresses relative to the vertical
gravitational load. There are three possible ways:
- Both horizontal principal stresses decrease by different amounts in magnitude.
- Both horizontal principal stresses increase by different amounts in magnitude.
- One horizontal principal stress increases while the other horizontal principal stress decreases.

Faults

jpb, 2014

103

Exercise
Assuming that one of the principal stresses is the gravitational load, draw three Mohr
diagrams with the same Coulomb criterion, the same vertical stress and
corresponding to the three standard states of stress postulated by Anderson.
Comment in terms of differential stresses involved in normal, thrust and strike-slip
faulting.
This formulation explains many fault systems but low-angle normal faults and high-angle thrusts
are identified natural cases that do not abide by the Andersons rules. Explanations can be the role
of anisotropies or pre-existing fractures in natural rocks, which affect fault orientation, and possible
strain along the 2 -direction. Other explanations involve rotation of fault planes towards nonconventional attitudes.

Synthetic and antithetic faults

Synthetic faults are parallel and have the same relative movement as the master fault. The
subsidiary yet genetically related set of conjugate faults, dipping in the opposite direction to the
master fault, is called antithetic.

This is orange

Normal faults
A down-dropped block bounded by conjugate normal faults dipping toward each other is a graben
and a relatively elevated block bounded by outward-dipping normal faults is a horst. Rifts are
major grabens that extend for long distances. A graben bounded by a single set of normal faults on
one side of a tilted fault block has a triangular profile and is called a half-graben.

In ideal graben and horst systems, the growth rates of the faults are equal so that the fault blocks do
not rotate and the grabens remain symmetric throughout the extension event. In natural fault
systems, however, faults grow at different rates and therefore give rise to asymmetric grabens and
Faults

jpb, 2014

104
block rotation. In fact, faulting is usually associated with rotation. There are two types of normalfault along which rotation is important: Planar and listric faults.
Planar faults
Planar, rotational normal faults occur above a basal detachment or a brittle-ductile transition. They
separate juxtaposed and tilted blocks without internal deformation. Both the faults and fault-blocks
rotate simultaneously about an axis roughly parallel to the strike of the faults (rigid body rotation
resulting in domino or bookshelf faulting).

Each fault block has its own half graben. Each fault must have the same amount of displacement
and tilting or there are space problems at the bottom of the system (opening of voids). Planar,
rotational faults and blocks generally abut against transfer, scissors faults.
Listric faults
Normal faults, in particular the master faults, are in general listric (concave upward). They may
look steep on the surface outcrops, although they are basically horizontal at depth, which gives rise
to space problems if the adjacent blocks are rigid: when the opposing blocks are displaced, they
cannot remain uniformly in contact and a gap must develop between the hanging-wall and the
footwall.

The voids may be filled by broken rocks of the fault walls, or may provide sites at which minerals
are subsequently deposited from circulating fluids. In order to maintain geometric compatibility,
beds in the hanging wall have to rotate and dip towards the fault. Collapse and rotation of the
hanging-wall toward the fault produces block rotations, fills this gap and flexure is eventually
accommodated by a number of antithetic and/or synthetic normal faults that sole into the low-angle
master fault. Note that the triangular shape of the half-graben over a rollover defines the dip of the
associated listric fault.

Faults

jpb, 2014

105

The listric fault geometry is important because it can accommodate a much larger amount of
extension than planar faults for the same amount of slip. The steep part of the detachment is
sometimes called the break-away fault.
Curved and ramping fault planes
In general, fault surfaces are curved. A listric fault is a curved, concave upward fault. Fault planes
gradually flattening with depth in a consistent stress field meet several explanations.
One is inherent to the shape of the Mohr envelope (see chapter on Faulting). Its progressively
decreasing slope with increasing confining pressure implies that the fault orientation flattens with
depth.
Another one concerns the material properties. Different rocks have different friction angles. The
failure (Coulomb) criterion of a rock with high friction angle and likely high cohesion (e.g.
sandstone) differs from that of a rock with lower friction angle and cohesion (e.g. clay). These
conditions plotted in a Mohr-Coulomb diagram show that a fault should change orientation with
respect to 1 across the boundary if these rocks are interlayered. Both rocks have to be at the
rupture state. This implies that the two rocks are under different states of stress. Assuming the same
pore pressure in both, the normal and shear components must have the same values on the layer
boundary. This is true only at the point where the two Mohr circles at failure intercept. The N and

S axes coincide with the x and z axes of real (physical space), respectively. Therefore, one
should be able to represent stress and physical spaces together on the same diagram. Consider
normal faulting with 1 normal to horizontal bedding. A horizontal line through the ( N , S )
stress point is the trace of the horizontal bedding plane on which the considered N and S stress
components are acting. The intersections of this line with the Mohr circles are called poles. The
chords connecting poles to stress point 1 define the real orientation of the plane orthogonal to the
local 1 , hence the local direction of the maximum stress in the corresponding rock. Knowing from
2 the angle of the fault plane to this local 1 , one can construct the fault in the two adjacent rocks
and readily see how variations in shear strength deviate fault planes.
The same game can be played for thrusting conditions. Fault plane deviation is even more
pronounced if layer-parallel shearing is involved, as it can happen when beds are inclined to
regional maximum compressive stress after folding or tilting, or when a viscous layer (salt, or
molten middle crust on a larger scale) laterally flows below a brittle cover. Then the fault tends to
follow the ductile layer
Faults

jpb, 2014

106

More in Mohr: Physical space in Mohr diagram

Displaying angles twice their actual, physical value is a visual inconvenience of Mohr diagrams.
Representing the real orientation of the planes on which stresses act is achieved by defining the
pole on the Mohrs circle. The easiest way to locate this pole is to draw a horizontal line through the
( N , S ) point under consideration. This line represents a plane parallel to the ( N ) axis, parallel
to the 1 direction and containing 2 . In terms of stress and space axes, the line is actually drawn
is accordingly parallel to the normal of a plane on which a stress, 1 , is known. Drawing a vertical
line parallel to the normal of a plane on which 3 is acting would yield the actual a diametrically
opposite intersection pole, which would be in the lower semi-circle. Discarding sign conventions
and being interested with orientations only, the upper semi-circle is sufficient.
An elementary theorem of Euclidian geometry states that an inscribed angle extended from any
point on a circle is one half of the angle subtended at the center from the same arc. A special
situation is that an angle inscribed in a semi-circle is a right angle. Therefore, the two lines drawn
from the antipole to the 1 and 3 points are perpendicular and one of these two lines defines
with the horizontal line through the ( N , S ) point an angle = . These two lines, from the
antipole point to the principal stresses, thus physically represent the orientation of the two
perpendicular planes on which N and S are normal and shear stresses, respectively.
Faults

jpb, 2014

107

Fault plane deviation can be even more pronounced if layer-parallel shearing is involved, as it
happens when beds are inclined to regional maximum compressive stress after folding or tilting, or
when a viscous layer (salt, or molten middle crust on a larger scale) laterally flows below a brittle
cover.
Then the fault tends to follow the ductile layer.

Faults

jpb, 2014

108

Thrusts
Two styles are commonly invoked to describe thrust tectonics: thin-skinned and thick-skinned
tectonics. They refer to the degree of basement involvement in the considered thrust system.

Thin-skinned tectonics
In many foreland fold-and-thrust belts, the sedimentary cover is detached from the basement
typically along fault planes with ramp-flat geometries. The sole thrusts remain above the strong
crystalline basement left undeformed. This style of deformation is known as thin-skinned tectonics.
Then, bedding plays a controlling factor in generating staircase, flatramp systems.

Thin-skinned tectonics implies large horizontal displacements whereby the stratigraphic sequence
above the floor dcollement can be piled up several times, thrust sheet upon thrust sheet. Thrust
sheet are generally thin compared to their lateral extent. Thrust faults may develop in sequence
either forward (which is termed prograding) or backward from the first thrust. Where the later
thrust develops in the footwall of the original thrust, the earlier, higher thrust sheets are carried
forward by the later, lower ones, which earned the name of piggyback thrusting. Conversely, if the
thrust development migrates backwards, an overstep sequence develops.
Thrust sequences often result in the stacking up of a series of thrust sheets separated by subparallel
thrust faults making up an imbricate zone or schuppen structure. When an imbricate zone is
bounded at the top and bottom by master thrusts or dcollement surfaces, the whole package is
called a duplex. Individual imbricate sheets within the duplex are called horses; they typically are
lense shaped in cross section. The typical duplex structure therefore consists of a flat-lying roof
thrust and a floor thrust (also called sole thrust) enclosing a stacked-up pile of horses.
Backthrusts are subsidiary thrusts with a displacement opposite to that of the main thrust. The
uplifted hanging-wall block between a thrust and a backthrust forms a pop-up. If the backthrust
truncates an earlier thrust, a triangle zone is formed.
Faults

jpb, 2014

109
Thick-skinned tectonics
In metamorphic regions thrusting is commonly associated with intense and distributed ductile
deformation. The staircase, flat and ramp geometry is not respected. Major sole thrusts extend
steeply down to the basement. Although thrust zones tend to follow surfaces of rheological contrast,
they involve the basement. This style is termed thick-skinned tectonics.

Eroded thrusts
A window (or fenster) is an erosion exposure of the rocks below a thrust fault that is completely
surrounded by rocks above the thrust. A klippe is an isolated, erosion remnant of a thrust sheet
completely surrounded by rocks of the footwall.

Gravity-driven thrusts
Slip-sheets are formed when a coherent part of a series has slipped away, as gravity collapse, from
an anticline crest and rests on an eroded surface within one of the adjacent synclines.

Strike-slip faults
Strike-slip faults are in general vertical and develop at ca. 30 to the horizontal compression
direction. Major strike-slip faults can be traced over several hundred kilometres, in which case they
are not simple planar movement planes. They often develop a system of right-stepping and leftstepping faults. Where right stepping faults generate an extensional zone, left-stepping faults
generate a compressive zone, and vice versa, according to the sense of displacement on the master
fault.

Faults

jpb, 2014

110

Subsidiary faults of curviplanar master faults

Usually subsidiary faults belong to the same class as the host master fault. This relationship is not
respected where the master fault is curviplanar, hence imposing a complex strain of the hanging
wall through the development of accommodation faults.
- Rotation of the hanging wall of a listric normal fault may change a former antithetic normal
fault into an apparent reverse one.

Reverse faults may form in the hanging wall of a convex-upwards normal fault.

Reverse faults may form in tilted layers to accommodate layer-parallel stretching due to larger
scale normal faulting.

Faults

jpb, 2014

111
-

Normal faults may form in the hanging wall of a convex-upwards thrust.

Subsidiary fractures: Riedel shears

Subsidiary Riedel shear fractures propagate a short distance out of the main fault. Riedel shears
form a network commonly developed during the embryonic stages of faulting. They form a
systematic array that seems to be self-similar for a wide variety of materials over a wide scalerange. The basic geometry consists of conjugate R and R fractures so that the acute bisector is the
direction of maximum compressive stress.
R Riedel shears develop at a small angle (typically 10-20) to the main fault, often in an en
chelon array, and are synthetic to the main fault. En-chelon describes the aligned pattern of a
series of parallel, short fractures that are arranged like rungs of a ladder seen in perspective. In
simple shear, the principal stress 1 is at 45 to the main slip plane. The Mohr-Coulomb failure

criterion predicts that conjugate failure surfaces are inclined at ( 45 2 ) to 1 , where is the

angle of internal friction. The acute angle of R Riedel shears with the main fault is equal to 2 .
This angle points in the direction of the relative sense of movement on the master fault.
R shears are conjugate, antithetic to the R(iedel) shears (i.e. with offsets opposite to the
bulk movement) oriented at a high angle [ 90 + ( 2 ) ] to the main fault plane. They preferentially
occur between two parallel R shears. R and R shears intersect in an acute angle = 90 .

Faults

jpb, 2014

112
P shears are synthetic minor faults symmetrically oriented to the R shears with respect to
the fault plane, at + 2 from the fault plane. P shears generally link R shears and tend to occur for
large displacements.
As for R Riedel shears, there may be P shears conjugate with P shears (at 90 2 from
the fault plane) but these have relative minor importance.
Y shears are synthetic microfaults sub parallel to, and slipping coherently with the main
fault.

Fault population and networks

The analysis of fault populations has shown that faults exhibit many characteristics of fractals; a
fractal is an entity that has geometrical properties (e.g. shape) either independent of its size (selfsimilar) or exhibiting a simple relation with its size (self-affine). This characteristic means that a
pattern of faults viewed from satellite pictures looks the same as the pattern of faults seen in an
outcrop. It implies that some fault properties (e.g. the length/throw ratio) are relatively independent
of fault size. However, analysis of faults on 3D data sets has shown that, in log-log space, there is a
simple linear relationship between fault frequency and fault size. This relationship between fault
size (length or maximum throw) and the number of faults with a certain size (few large faults, many
small faults) can be used to predict the density of small, sub-seismic (i.e. below the limit of seismic
resolution) faults.
There are a few important terms describing fault arrays in map view.
En chelon pattern
En chelon means "in step-like arrangement" and describes a consistently overstepping and
overlapping alignment of subparallel, closely-spaced structures that are oblique to the planar zone
in which they occur. Such patterns are commonly related to potential faulting.

Faults

jpb, 2014

113
Merging faults
In brittle regions and damage zones, faults frequently splay into complex arrangements of smaller
faults that curve away from the direction of the master fault. The line of connection where a fault
splits into two fault surfaces of the same type is a branch line. Beyond this line, branching splay
faults form an imbricate fan that spreads the displacement over a volume of rock. A splay is a
small, often inactive, fault segment or branch created during fault coalescence (hard linking) or
propagation (branching).

Anastomosing pattern
Riedel shears of any scale may merge with one another to form an anastomosing network of
fractures in a narrow fault zone whose bulk borders are parallel to the main fault. Anastomosing
refers to a branching and rejoining network of irregular surfaces or lines interlaced like braided
streams or veins.

Scaling of fault distribution

A power law distribution is frequently advocated to describe fault distributions. In practice, the
significance of a single power law exponent might reveal hierarchical patterns in a complex fracture
system that includes all sort of fracture with different scales.

Fault anatomy
Faults have not infinite extent. They consist of the slip surface (fault core for thick faults made of
anastomosing shear fractures) and an enveloping damage zone that spreads over some width. These
two elements either simply die out along their strike or terminate against other faults. In the latter
case, faults either merge with or are truncated by the other faults. As a result of fault growth and
coalescence, faults develop into a fault network.

Tip line and tip zones

Single, isolated faults are approximately elliptical surfaces along which most of the slip has taken
place; this elliptical shape is broken if the fault intersects the earth surface. The aspect ratios
between length and width of the elliptical normal-fault planes tend to be >1 and <5.
Faults

jpb, 2014

114

The relative displacement must fade out outward. It drops to zero at the tip line, which encloses this
movement plane. In other words, the tip line separates slipped from non-slipped rocks. Beyond this
limiting line, the fault displacement is accommodated and dies out across a tip zone in various
ways, depending on the ratio of fault length to fault displacement.

If displacement is very small relative to the dimensions of the active segment, space and
continuity problems are accommodated by gradual reduction in displacement toward the fault
extremities and suitably distributed, non-brittle deformation (penetrative strain and/or folding)
through the surrounding solid rock along the fault tip., the ductile bead,
Where displacement is bigger relative to fault dimensions, strain around the fault can be
accommodated within a damage zone by clusters of smaller faults and additional fractures,
particularly in hanging-walls. Then a ductile bead envelopes the brittle damage zone.

Faults

jpb, 2014

115

Damage zones
Damage zones are arrays of entangled minor faults and fractures along larger fault planes. The
density of faults and fractures usually decays exponentially away from the master fault. Damage
zones occur because of stress concentration, particularly at fault tips and in linking zones; they also
occur to accommodate displacement variations into or along faults. Initiation, propagation and
interaction classify damage zones into tip-, wall- and linking-damage zones based on position
within and around a fault zone.
- A tip damage zone develops in response to stress concentration at a fault termination.
- Wall damage zones can be distributed along the whole trace of a fault. They may represent tip
damage zones abandoned in the wall rocks as faults propagated through the rock. They may also
represent wall-rock deformation associated with the build-up of slip on faults.

Linking damage zones are caused by the interaction and linkage of fault segments in a relatively
small region. They are complicated due to cumulative displacement and interaction of the tip
and wall damage zones of two neighbouring faults. Consequently, linking damage zones can
develop a wide range of fracture patterns depending on the interaction between the two fault
segments.

Fault termination
Mesoscopic tip damage zones in front of tip lines are categorized into four subdivisions according
to the nature and orientation of faults and fractures developed: wing cracks, horsetail splays,
synthetic and antithetic splay faults. Tip-damage zones are easy to recognize, even at a large scale.
Faults

jpb, 2014

116
Wing cracks
Wing cracks occur where there is a rapid decrease in slip at the fault tip, at fault plane irregularities
such as bends, steps, or relay zones and at points of variable frictional properties along the fault
surface. They abut fault planes and are extension fractures that tend to curve towards parallelism
with the local maximum principal stress direction, in dilational quadrants of the fault front.

Anticracks
Anticracks are solution surfaces (stylolites) symmetrical to wing cracks with respect to the main
shear surface. They are orthogonal to the local 1 direction in compressional quadrants of the fault
front.
Pinnate fractures
Feather (or pinnate) fractures are extension fractures that form en chelon arrays along slip
surfaces. Tension gashes are pinnate fractures filled with mineral crystallization.
Splay shear fractures
Synthetic splay faults
Synthetic splay faults are geometrically and mechanically similar to wing cracks but are finer and
more closely spaced with relatively low angles to the master fault. They have the same sense of slip
as the main fault, and may link with a neighbouring fault segment. Horsetail fractures splay
asymmetrically out, often on one side of the main fault in a fan-shaped network. They tend to
develop where slip dies out more gradually towards the fault tip than for wing cracks.

Faults

jpb, 2014

117
Antithetic splay faults
Antithetic splay faults have a sense of slip opposite to that of the main fault and tend to develop at
high angles to the master fault. They are isolated fractures separated from the master fault and they
often increase their length and spacing away from the fault tip.

Relay zones; linking damage zones

A relay structure transfers displacement from one fault segment to another segment. The overstep
is the discontinuous interval between two sub-parallel faults (strictly speaking between similar
structures). If the normal to the tip of one overstepping fault intersects the other faults, there is
overlap. The alternative is underlapping. The distance between overlapping, parallel fault
segments is the separation. Separation / overlap ratios provide a crude measure of fault interaction.

The degree of interaction is better determined when separation and overlap values are normalised to
the fault length of one and any of the two interacting faults. Overlaps are described as right- or leftstepping, depending on the sense the jump goes from one structure to the next.
A relay ramp is an area of bent bedding that transfers displacement between two overstepping
faults with the same dip direction.

Relay zones may lead to large structures. Whatever their scale, they are transient features, evolving
during fault propagation until they are replaced by breaching faults that connect the interacting
fault segments (evolution from soft- to hard-linkage) to make a single, through-going fault surface.
Single- and double-tip linkages are obvious patterns that will form fault bends and jogs, where both
the dip and strike of a strike-slip fault change.

Faults

jpb, 2014

118

The change in attitude of the fault plane produces compressional or extensional stepping zones,
according to the shape of the step with respect to the movement on the master fault. For example, if
a dextral fault steps to the right, the overlap zone or the transfer zone where the fault segments run
parallel faces the direction of shear: it is in extension. Transverse normal faulting forming voids
filled with vein material or low topography areas with basin sediments (rhomb-shaped basin)
commonly accommodate extension in these zones. Conversely, solution structures, anticlines, or
some high topography mark a compressional overlap, which faces opposite to the direction of shear.
Overlap structures above strike-slip faults are called flower-structures if several splay faults root
into the main strike slip zone. Fault throw gradients tend to be highest in relay zones.

Faults

jpb, 2014

119

Fault rocks
Faulted rocks form bounding walls on either side of the fault zone, which is often filled with fault
rock. The accepted classification of fault rocks, i.e. rock types created by fault generation, uses
cohesion at the time of fault movement and the presence of a planar fabric. Two main types of fault
rocks are identified:
- Incohesive fault rocks in brittle fault zones. These rocks have a random fabric.
- Cohesive fault rocks in ductile shear zones. These rocks have a foliated fabric.

Incohesive, non-foliated fault rocks

Faulting in the brittle field is characterised by comminution (grinding) of rocks. The resulting
incohesive, non-foliated fault rocks are called cataclasites.
Description - Definition
Cataclasites are randomly oriented aggregates of angular, broken fragments of the rocks that
compose the fault walls. The fragments range in size and may be held together by some cementing
material generally formed by precipitation between fragments of infilling minerals from circulating
fluids. According to the size of the elements, one distinguishes:
* Fault breccia when visible, angular fragments constitute more than 30% of the rock
volume;
* microbreccia if the fragments are microscopic;
* gouge when more than 70% of the material consists essentially of very fine-grained,
clayey and often dark powder containing small angular fragments. Clay minerals result
from weathering and/or hydrothermal alteration of crushed fault-wall rocks. Gouges and
equivalent fine-grained fault wears are rarely consolidated.
Late movements may impart a distinct planar fabric in these rocks, which, however, fundamentally
remain incohesive, i.e. unconsolidated.
Setting
Major faults do not exhibit a discrete slip surface but a planar core up to several meters in
thickness, essentially formed of wear detritus derived from the fault walls. Cataclastic and cracked
rocks also constitute the damage zone, the broad volume of deformed wall rocks around the fault
core. Many rich veins of metalliferous ores occur in this setting, with hydrothermal minerals
cementing the rock clasts. In that case the cohesionless rock at the time of faulting has acquired a
secondary cohesion.
Cataclasis
The incohesive fault rocks are essentially formed by cataclasis, the deformation process involving
fracturing of grains and grain boundaries along with dilatancy allowing rigid-body rotation
between granular elements. Cataclasis results from the initiation, propagation, interaction and buildup of slip along the fault. It can be described as the mechanical granulation, crushing and/or milling
down to powder any coherent rock. The process is common in the upper crust where strain rates are
fast and confining pressures and temperatures relatively low (< 500 MPa, 200-300C).
The size frequency of crushed particles is a measure of the energy used for cataclasis. Fault gouges
show size frequencies with fractal (i.e. scale invariant) dimensions > 1.6.

Cohesive, foliated fault rocks

Cohesive, foliated, commonly lineated fault rocks belong to the mylonite series. They are
characterised by a foliated or streaky structure, in thin section, and are typical of ductile shear
zones. Grains of the parent rock have been reduced in size without the loss of primary cohesion.
The fine-grain size and distinctive microstructure are due entirely to ductile deformation (viscous
creep) accompanied by recrystallisation. They often contain larger fragments or relict minerals from
the parent rock; these fragments are called porphyroclasts. Mylonitization is a gradual process of
Faults

jpb, 2014

120
grain size reduction in which three types of rocks are distinguished on the relative proportion of
porphyroclasts to fine-grain matrix:
* A protomylonite is a rock in the early stages of mylonitisation, containing more than 50%
porphyroclasts.
* A true mylonite contains 10-50% porphyroclasts.
* Extreme grain size reduction and dynamic recrystallisation may produce a hard, flint-like,
dark fault filling of ultramicroscopic grains containing less than 10% of tiny porphyroclasts. These
rocks are known as ultramylonites.
* The term blastomylonite can be used for such rocks that have been extensively
recrystallized and strain grains annealed after mylonitisation.
* Phyllonite is a mica-rich mylonite that has the mesoscopic appearance of schist.
Fault rocks - Terminology
metamorphic
conditions

very low grade

low grade

medium

Cataclasite

Mylonite

incohesive

cohesive

Fragments %

Principal
types of fault
rocks

high grade

Matrix %

Breccia

Protomylonite

<50

Microbreccia

Mylonite

50<<90

Ultramylonite

>90

>30

<30

Gouge

Pseudotachylite (molten)

recrystallised
Blastomylonite

>85

Pseudotachylites
Pseudotachylites form thin dark veins of cohesive and non-foliated rock along some faults. They
look like volcanic glass and typically occur in a branching network of injection veins stemming
from the fault zone into the usually crystalline wall rock. These veins indicate that the
pseudotachylite was in a fluid state. The glassy matrix contains rock inclusions and microscopic
spherulites. Pseudotachylites formed by local friction melting of the wall rocks during rapid
movement along the fault plane. In most cases, later devitrification has removed the glassy texture.
Pseudotachylites are believed to form when a seismic movement and local decompression trigger
rapid melting followed by quenching and solidification of the molten material.

Fault rocks and depth

The different types of fault rocks tend to form at different depths:
Incohesive gouge and breccia:
Cohesive cataclasites and pseudotachalites:
Cohesive mylonites:

0-5 km
10(15) km
> 10(15) km

Determination of slip
Several methods may be employed, if possible in combination, to determine a fault displacement.

Offset of geological structures

A first indication for knowing a fault is based on the knowledge of the stratigraphic sequence. Then,
for example in a drill hole, a thrust is responsible for the repetition of strata, and a normal fault for
omission of strata. This simplification ignores however the possible strike-slip movement on strikeslip faults with an original dip.
Faults
jpb, 2014

121
The complete determination of displacement requires identifying the positions on either side of the
fault of two originally coincident points. Such points may be defined by the intersection of the fault
plane and a linear element such as a fold hinge line, which pre-existed and was displaced by the
fault. Where the offset linear marker is defined by the intersection of two planar structures, the
planar structures must both predate the faulting. Offset contact lines on geological maps may be
misleading because they are defined by the intersection of a geological plane that predates the fault
and a topographic surface that post-dates the fault. The same may be true of offsets shown in cross
sections, or observed in outcrop surfaces of any orientation. To distinguish the offset of planar
markers from offset of linear markers the former are referred to as separations.
A lower limit can be placed on the net slip of a thrust fault where thrusting has brought older rocks
over younger ones. For example, the distance between klippe and window places a minimum limit
on the displacement of the fault. Note, however, that the klippe-to-window method is lousy if a line
drawn from window to klippe is not at least approximately parallel to the direction of displacement.

Diagnostic movement structures within brittle fault planes

Rocks in contact at a fault plane often have smooth, shiny or polished surfaces of mineralised
material known as slickensides. Slickensides are due to abrasive action, may be featureless, but
sometimes feel smoother in the direction of slip. In fact, parallel striations are often found on fault
surfaces. The angle measured within the inclined fault plane between the horizontal line (the fault
strike) and the line marked by striations is called pitch or rake.

These cataclastic lineations (slickenlines) are elongated parallel to the slip vector on the fault.
Associated asymmetric surface features are kinematic indicators of the sense of slip.
The most common and reliable fault kinematic criteria are:
Asperity ploughing
Slickensides commonly display a linear striation or corrugation (thereby describing parallel ridges
and grooves that occur over a range of several scales) experimentally shown to be abrasion
scratches parallel to the direction of relative fault movement. Most striations (slickenlines) are
defined by mineral streaks in the fine-grained material along fault planes. Some striae may be
grooves or gutters deeply furrowed on one side of a fault by hard particles drawn over it by the
other side. Ridges and grooves can be long, linear, metre scale undulations of the fault plane. On a
microscopic scale, a dimensional preferred orientation of grains marks such lineations, in particular
in soft sediments. Such linear features indicate the slip direction but not its sense.

Faults

jpb, 2014

122

Tool marks, tracks and debris streaking

A striating, erosive object can be pinned in one wall of the fault plane, while matching depressions
or indentations are present on the opposing surface. A spoon-shaped depression around a hard clast
provides the sense of relative movement, with the hard object at the distal end of the pit it produced.
Conversely, debris can be deposited in the direction of slip behind a protruding asperity, erosion
sheltering creating a tail of lightly cemented gouge material accumulated behind hard asperities,
which is the movement direction.

Steps
Chatter marks are small, asymmetric steps facing in one direction and roughly perpendicular to
the striations. These steps were traditionally interpreted as indicators of the sense of displacement,
with the riser facing the relative displacement direction (congruous steps). However, it has been
shown experimentally that incongruous steps accompany frictional-wear striations or obliquestylolite columns and form so that the step risers are opposed to the movement vector. Therefore,
there is no general rule about the kinematics significance of the steps alone.
Riedel shears
Small striated fractures in Riedel-shear attitudes commonly truncate fault planes nearly orthogonal
to the slip direction. R and R-shears tend to be regularly spaced and impel a serrated profile to the
fault plane, with steps facing the movement direction.

Faults

jpb, 2014

123

Combinations of R and P-Riedel shears, intersecting nearly perpendicular to the slip direction,
result in alternating striated surfaces (P-shears facing in movement direction) and non-striated
surfaces (R-shears in lee side of asperities). The intensity of striation or non-striation depending on
the attitude of topographic irregularities on the fault plane is a common kinematic criterion.

Friction fractures
Friction fractures dip forward into the fault with respect to the movement direction (R Riedel
shears). They are concave so that their intersections have crescent-shapes with their long axis
transverse to the direction of movement on the fault surface. The two crescent tips indicate the
movement direction of the missing block.

Accretionary mineral steps

Preferred directional growth of minerals during faulting produces fibrous crystals with their long
axes parallel to the prevailing direction of slip into dilating cracks, on the lee side of congruous
steps and asperities. These slickenfibers (or accretionary growth fibers) are particularly valuable
Faults
jpb, 2014

124
because they indicate the sense of movement. Curved or superposed crystal fibres can preserve a
record of changes in the instantaneous direction of fault displacement. By contrast, ordinary
slickenside striations may often be erased and overprinted if changes occur in the direction of fault
displacement so that ordinary striations may record only the latest uniform displacement.

Slickolites
Surface irregularities or asperities may show a striated or stylolitic surface facing movement
(compression) direction of the missing block and unstriated slopes towards the movement
(extension) direction. Slickolite defines dissolution surfaces facing the displacement direction with
micro-stylolitic peaks pointing in the upstream direction at a low angle to the fault surface.
Unstriated, mineralised fissures
Straight, forward-dipping straight and lunate fractures form at a low angle to the fault plane and
tend to rotate towards higher angle relationships during deformation of wall rocks due to friction
slippage. They open for secondary mineral crystallization during rotation. Straight fractures are
named comb fractures. The two horns of crescent fractures point in the direction of slip.

Diagnostic movement structures within fault zones

Movement-related structures in fault zones and observed on erosion surfaces orthogonal to the fault
plane and parallel to the slip direction are used in determining the fault kinematics.
Tension gashes
Tensile fractures or tension gashes (T fractures) that are parallel to the regional maximum
principal stress (compression) may appear at an angle typically less than 45 to the fault plane, near
the fault plane. Their intersection with the fault surface is nearly perpendicular to the cataclastic
lineation. Their angular relationship may be helpful to infer the sense of slip, as discussed for the
Faults

jpb, 2014

125
pinnate joints. In addition, they may take S or Z shapes depending on the leftward or rightward
sense of shear along the fault, respectively.

Exercise
Draw in two and three dimensions tension fractures related (1) to a normal fault, (2)
to a thrust.
Drag folds
Bedding adjacent to faults may curve in the direction of movement of the opposite block. These
local flexures are called drag folds. The appellation is interpretation that faulting is initiated first
and then frictional shear deformation along the fault causes folding as one block is dragged along
the other.

The resulting geometry would not be different if fault-parallel, ductile shearing precedes faulting
through a shear zone, whether this has a constant (simple shear) or a downward decreasing
thickness (trishear).

The use of drag folds can be misleading because curvature of opposite sense to the displacement,
termed reverse drag, is common. Reverse drag is clearly independent of true drag effects but hardly
distinguished from true drag when they appear separately. In addition, the orientation of such folds
is often not controlled by the movement direction but rather the intersection between bedding and
the fault plane. Drags should be used with extreme care.

Faults

jpb, 2014

126
Conjugate Riedel shears

Exercise
Discuss and sketch how conjugate Riedel-shears may help defining the sense of
relative movement of the main fault.

Experimental faulting in sand

Very early, students of the mechanical behaviour of rocks have noted that dry sand exhibit fault
structures similar to those recognised in rocks. In effect, it has been empirically verified that dry
sand satisfies a yield criterion of Coulomb type behaviour with an internal friction angle similar to
those of rocks (30-40, depending on close packing) and cohesion of the order of 100 Pa.
Therefore, dry sand is an excellent analogue for simulating brittle deformation of the upper crust in
the Earths field of gravity.
A famous experiment is due to Hubbert. He made a glass box divided into two compartments by a
rigid, movable partition that can be translated parallel to the long axis of the box. The box was filled
using loose sand with different colours (and in Hubberts case, plaster) so that layers and horizontal
markers can be seen. When the partition is moved to the right (with a screw), the left-hand side
compartment is lengthened while the right hand side compartment is shortened by the same amount.

Faults

jpb, 2014

127
The first thing that happens is the development in the lengthened compartment of a distinct normal
fault that dips 60-65. In the meantime nothing other than a slight bulging is observable in the
shortened compartment. As the partition is cranked farther, reverse faults that dip 25-30 occur in
the shortened side. These dip in the same direction and extend from the bottom of the box near the
foot of the partition to the surface of the sand where escarpments are produced.
The behaviour of the sand is obviously identical in both compartments. Thus, the double
experiment illustrates that
- 1) rupture is reached faster in extension than in compression
- 2) Normal faults are steeper than reverse faults.
- 3) Andersons classification of faults (fault plane orientation with respect to stress ellipsoid) is
applicable in near-surface conditions.

Summary
Faults are fractures along which macroscopically visible slip has occurred. Faults are not theoretical
planes of highest shear stress because a friction factor depending on each lithology has to be taken
into consideration. High pore pressures are able to promote faulting under conditions where it
would not occur in dry rocks.
Three basic types are recognised 1) normal faults, 2) thrust faults, and 3) strike-slip faults. Large
faults consist of a central fault core and an enveloping damage zone. Faults develop in the brittle
regime and represent sudden displacement events producing earthquakes. They initiate with the
nucleation of small shear fractures that propagate, interact and link to form large faults. Their total
movement generally results from the addition of many smaller slip events. Slickensides, striations
and grooves caused by surface roughness (asperities) are used to define the offset directions of the
walls.

Recommended literature
Gupta, A. & Scholz, C. H. 2000. A model of normal fault interaction based on observations and
theory. Journal of Structural Geology 22(7), 865-879.
Hancock, P. L. 1985. Brittle microtectonics: principles and practice. Journal of Structural Geology
7(3/4), 437-457.
Hancock, P. L. & Barka, A. A. 1987. Kinematic indicators on active normal faults in western
Turkey. Journal of Structural Geology 9(5-6), 573-584.
Hubbert, M. K. 1961. Mechanical basis for certain familiar geologic structures. Geological Society
of America Bulletin 62, 355-372.
Mandl, G. 1988. Mechanics of tectonic faulting. Elsevier, Amsterdam.
McClay, K. R. 1992. Thrust tectonics. Chapman & Hall, London.
Monzawa, N. & Otsuki, K. 2003. Comminution and fluidization of granular fault materials:
implications for fault slip behavior. Tectonophysics 367(1-2), 127-143.
Petit, J.-P. 1987. Criteria for the sense of movement on fault surfaces in brittle rocks. Journal of
Structural Geology 9(5-6), 597-608.
Ramsay, J. G. & Huber, M. I. 1987. The techniques of modern structural geology - Volume2 : Folds
and fractures. Academic Press, London.
Twiss, R. J. & Moores, E. M. 1992. Structural geology. W.H. Freeman & Company, New York.
Wernicke, B. & Burchfiel, B. C. 1982. Modes of extensional tectonics. Journal of Structural
Geology 4(2), 105-115.
Faults

jpb, 2014

128

Faults

jpb, 2014