Earth-Science Reviews 154 (2016) 14–28

Contents lists available at ScienceDirect

Earth-Science Reviews

journal homepage: www.elsevier.com/locate/earscirev

Fault linkage and relay structures in extensional settings—A review

Haakon Fossen a,b, Atle Rotevatn a
a
Department of Earth Science, University of Bergen, Allégaten 41, N-5007 Bergen, Norway
b
Museum of Natural History, University of Bergen, Allégaten 41, 7800, N-5020 Bergen, Norway

a r t i c l e i n f o a b s t r a c t

Article history: Normal fault relay structures form at all scales as faults interact or step out of their own plane during growth,
Received 31 July 2013 and their successive formation and destruction represent the most efficient way for faults to lengthen. Their
Received in revised form 21 November 2015 progressive evolution from the moment of fault overlap to a fully breached fault involves strain accumulation
Accepted 23 November 2015
in the fault overlap zone, initially through bending of layers and secondarily by the formation of fractures,
Available online 24 November 2015
deformation bands (porous sandstones) and subsidiary faults. These small-scale structures have more complex
Keywords:
orientation patterns than the typically strike-parallel orientations seen in ordinary damage zones away from
Relay structures sites of fault interaction. Breaching occurs at a given level of bending (dip or curvature) and at the achievement
Relay ramps of a critical level of fracture or deformation band density that again depends on local geometric and
Fault growth lithologic/mechanical conditions, but a ramp slope close to 10–15° at the onset of breaching seems to be
Fault interaction common. Relay zones are not only lateral communication paths for fluid flow across sealing faults, but
Fault stepovers their anomalously wide and well-developed damage zones make them conduits of vertical fluid flow in
petroleum, groundwater, CO2 sequestration and magma settings alike, and therefore also serve as sites of
ore deposits.
© 2015 Elsevier B.V. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2. Linked fault systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3. Fault interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.1. The role of boundary conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.2. Preexisting structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.3. The process of linkage: relay ramp formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.4. The breaching of relay ramps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.5. Fault linkage of curved fault segments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
4. Consequences for fluid flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
5. Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

1. Introduction deforming rocks where populations of structures evolve from small to

large sizes. In the most general sense this includes the formation and
Relay and transfer structures are locations of fault interaction where growth of fold populations as well as the development of fracture,
strain or displacement is transferred or relayed from one structure to vein and fault populations in any tectonic regime. However, modern
another (Figs. 1 and 2). They allow individual faults to have finite use is generally restricted to faults, and this paper will focus on normal
lengths and thus along-strike strain- or displacement variations, while fault relay zones, although most of the characteristics easily translate
the system as a whole (for example a rift) acts as a coherent system into other regimes and settings (e.g., Nicol et al., 2002).
that maintains a laterally constant amount of extensional strain as Dahlstrom (1969) realized that displacement on thrust faults in the
measured across the rift (e.g., Walsh and Watterson, 1991). Relays and Canadian Rocky Mountains varies along the faults, and found that as
transfer structures occur in many settings and scales in naturally displacement tapered out along one thrust fault it was commonly

http://dx.doi.org/10.1016/j.earscirev.2015.11.014
0012-8252/© 2015 Elsevier B.V. All rights reserved.
 H. Fossen, A. Rotevatn / Earth-Science Reviews 154 (2016) 14–28 15

zones are first-order relay structures in rift systems that also host sec-
ondary relay structures that are important both from a structural and
petroleum geology perspective. Hence, while transfer zones
technically are major relay zones that involve a number of faults and
smaller structures, the term relay structure is used about simpler
structures involving two (master)faults whose fault tips are interacting.
Consequently a transfer zone typically contains a number of relay
structures, which is evident from the examples portrayed in Fig. 3.
Even though the geologic term relay structure had been in
sporadic use for decades (e.g., Goguel, 1952), the term did not
catch on until it was explained and nicely illustrated by Larsen
(1988) based on his mapping of an extensional fault array in the
Permian of East Greenland. Since then, the term has for the most
Fig. 1. Normal fault relay structure with a ramp connecting the hanging wall and footwall, part been used about fault interaction structures in extensional
showing how relay ramps form when the faults are close and their two strain envelopes
settings (e.g., Peacock and Sanderson, 1991; Willemse, 1997; Crider
overlap (right). An isolated fault is shown in the left part of the figure.
and Pollard, 1998). In such settings, and in particular with respect
to petroleum geology and fluid flow, relay structures are particularly
being relayed to an adjacent and overlapping subparallel thrust. important as they occur in large numbers in rifts and faulted conti-
Dahlstrom (1969, p. 752) described this as “a kind of lap joint nental margins (e.g., Anders and Schlische, 1994; Young et al.,
wherein the fault whose displacement is diminishing is replaced by 2001; Jackson et al., 2002; Bense and Baalen, 2004; Elliott et al.,
an échelon fault whose displacement is increasing”. He referred to 2011). Relay structures are also important with respect to drainage
such thrust-fault relay structures as transfer zones, and realized patterns and facies variations along active faults that breach the
that they occurred along faults that were connected at depth by surface (Gawthorpe and Leeder, 2000; Athmer et al., 2010). In partic-
means of a basal décollement. ular, they focus sediment supply to local hanging-wall depocenters.
The terms transfer zone and accommodation zone were later used There has also been a recent interest in the damage (in most cases
about extensional structures by a number of authors (e.g. Bosworth, structures below seismic resolution) associated with relay structures,
1985, 1987; Rosendahl et al., 1986; Morley et al., 1990; Gawthorpe both in the context of fluid flow in petroleum and hydro reservoirs
and Hurst, 1993) for zones separating large (typically 100-km scale) (Fossen et al., 2005; Rotevatn et al., 2007), ore mineralization (Cox,
structural domains of different characteristic fault architecture or dip 2005; Xiao-shuang et al., 2005), hydrothermal systems (Faulds et al.,
directions. They transfer fault strain from one side of the rift to the 2013) and, for large relay structures, the control that strain-hardened
other, and classical models for transfer or accommodation zones exhibit relays may have on rupture propagation during earthquakes
arcuate faults bounding sub-basins (half-grabens) in rift systems, (Manighetti et al., 2009; Finzi and Langer, 2012).
primarily the East African rift system. They were commonly In this review we will look at relay structures in normal fault
portrayed together with low-angle detachments in the 1980s and populations. Although oppositely dipping fault interaction structures
90s (e.g., Bosworth, 1985; Faulds and Varga, 1998), in many cases are also regarded as relay structures (Morley et al., 1990), we only
more as a result of the strong focus on low-angle extensional detach- discuss such arrangements in terms of large transfer or accommodation
ments and listric faults at the time (e.g., Gibbs, 1984; Wernicke and zones. For simpler relay structures we limit our review to structures
Burchfiel, 1982) than from objective observations based on hard forming by interaction between (sub)parallel fault segments. We
geologic or geophysical data. However, in many cases transfer review important geometric and evolutionary aspects of fault
zones are not related to low-angle detachments, and they do not relay zones in extensional fault systems and discuss their implications
have to involve markedly curved master faults. Furthermore, transfer for fluid flow.

Fig. 2. Relay structure in Canyonlands National Park (Devils Lane) showing rapid fault displacement fallof (unusually high displacement gradient) toward the tip. The ramp itself contains
several faulted joints.
16 H. Fossen, A. Rotevatn / Earth-Science Reviews 154 (2016) 14–28

2. Linked fault systems

The evolution of large fault structures occurs by one or a combination

of the following mechanisms: (1) simple tip propagation and coalescence
or linkage of initially isolated, smaller fault segments (‘segment growth
and linkage’, sensu Trudgill and Cartwright, 1994; or ‘isolated fault
growth’, sensu Walsh et al., 2003); or (2) rapid establishment of the full
length of the fault and subsequent displacement accrual without signifi-
cant tip propagation (‘coherent fault growth’, sensu Walsh et al., 2003).
Although the two may be seen as competing models to explain fault
evolution, it is our view that these represent end-member models for
fault growth that are equally applicable in nature, probably even within
the same region or rift.
Isolated fault growth occurs with the development of long faults from
a population of more or less randomly distributed embryonic faults
nucleating from distributed natural heterogeneities in a lithologic unit
(e.g., Cowie et al., 2000; Soliva and Schultz, 2008; Fig. 4a). Experimentally
this setting is produced in a sandbox experiment where sand is overlying
a relatively homogeneously deforming material such as silica gel or basal
rubber membrane (e.g., McClay and Ellis, 1987; Vendeville et al., 1987;
Vendeville and Cobbold, 1988; Wu et al., 2015). As natural examples we
could envisage a system comprising a competent unit (sandstone,
limestone, basalt layer) overlying a softer or viscous unit (shale or salt)
(e.g., Trudgill and Cartwright, 1994) or clastic sediments sliding on a
low-angle décollement of evaporites or overpressured shale on a passive
margin (Duval et al., 1992; Fort et al., 2004; Rouby et al., 2011). Numerical
models of such settings have shown that fault linkage occurs as the em-
bryonic fault structures grow and interact (Cowie et al., 2000; Allken
et al., 2013). Hence the isolated fault growth model requires boundary
conditions that distribute strain homogeneously enough for individual
and isolated faults to nucleate in a relatively wide area.
The other end-member model, represented by the coherent fault
model, is the development of a long fault structure above a buried
fault undergoing reactivation (e.g., Giba et al., 2012) (Fig. 4b). Slip on
this fault imposes a strongly non-uniform extension with strain localiz-
ing in the cover above the fault, and the overall fault propagation is
upwards from the reactivated fault. In general the overall upward prop-
agation of basement faults through sedimentary cover tends to generate
tip line bifurcation and thus segmented fault systems that link up as
strain accumulates (e.g., fig. 8a in Childs et al., 2009). In detail, mechan-
ical stratification may complicate the growth history. For instance, faults
may initiate in strong layers above the basement fault and develop into
en-échelon faults that then propagate and interact laterally as well as
vertically (e.g., Jackson and Rotevatn, 2013).
Varieties of this model occur when the direction of least compressive
stress (σ3) is oblique to the strike of the underlying basement faults,
which produces shear displacements along these structures. Depending
on the angular relations the fault array in the cover will tend to consist
of individual segments that align with respect to the current stress field
(perpendicular to σ3) while the trend of the zone follows that of the base-
ment fault (Fig. 4c). Excellent examples of such developments are found
in active volcanic areas such as Hawaii and Iceland (e.g., Acocella et al.,
2000; Tentler and Mazzoli, 2005; Podolsky and Roberts, 2008) as well
as in rifts and continental margins (Giba et al., 2012; Jackson and
Rotevatn, 2013). Whether or not reactivation happens depends on the
orientation of preexisting faults relative to the new stress field and on
the strength of those faults (Bott, 1959), hence the contribution of each
of these processes will depend on local factors. In both cases, linked
fault systems form where individual segments interact and link up by Fig. 3. a) Structural map of the Suez rift, Egypt. Based on Bosworth (2015). b) 3D model
relay formation and breaching. The result of such a linkage process is to of the Cretaceous Tucano Rift, showing the trend of the Sergipano orogenic belt in the
form a longer fault from several individual small faults. Proterozoic basement and its connection to a transfer zone in the basin. Modified from
Milani and Davison (1988).
3. Fault interaction
such as fault density, fault distribution and spatial arrangement of faults
The likelihood that faults in a given fault population will interact in the fault population, the number of faults that grow relative to those
depends not solely on strain, but also on a number of other factors, that become inactive, and the size of the elastic strain field or stress
 H. Fossen, A. Rotevatn / Earth-Science Reviews 154 (2016) 14–28 17

perturbation around the faults (Cowie et al., 2000; Walsh et al., 2003;
Soliva et al., 2006). However, it also depends on whether the available
faults are more or less randomly distributed throughout the volume
of deforming rocks (e.g. a magmatic rift zone) or whether fault segmen-
tation is a consequence of complex vertical propagation of deeper
structures (Soliva and Schultz, 2008), similar to the twisted geometry
of dike fringes described repeatedly through the literature on dike
intrusion geometry (e.g., Anderson, 1951; Delaney and Pollard, 1981)
(Fig. 4c). The boundary conditions for the area or volume of rock
undergoing faulting are important factors in this respect.

3.1. The role of boundary conditions

The overall distribution or arrangement of faults in the popula-

tion is controlled by kinematic boundary conditions as well as
rheology and mechanical layer properties. By kinematic boundary
conditions we mean the external causes and controls on the type
(plane versus three-dimensional strain, strain rate) and magnitude
of strain in the deforming volume of sediments or rocks. Boundary
conditions are easier to visualize for physical models such as sand
box experiments, namely the base (flat, irregular, stretching rubber
sheet, etc.) and walls (fixed or moving). In nature, the base of a
deformed sedimentary sequence or basin can be a detachment or
salt layer, or heterogeneous with mechanically weak preexisting
faults and fabrics whose location, orientation and arrangement will
potentially influence the structures and their distribution in the
overburden (e.g., Fig. 4b–c).

3.2. Preexisting structures

Preexisting faults and fabrics in the basement of regions undergoing

extension may or may not influence the formation and localization of
transfer structures or relays. In general, it is the largest structures that
are influenced by basement anisotropy, and although transfer zones
can develop in the absence of basement structures (Schlische and
Withjack, 2009), a close correlation between basement structures and
transfer zones has been reported for a number of rifts worldwide. For
instance, the East African Rift is influenced both by a steep metamorphic
basement fabric and ductile/brittle basement shear zones that were
reactivated during rifting, locally with the formation of pseudotachylite
(Hetzel and Strecker, 1994; Smith and Mosley, 1993; Kinabo et al.,
2007). Similarly, the geometry of the Tucano Rift in NE Brazil is
influenced by the oblique fabrics and faults of the Sergipano orogenic
belt (Fig. 3b), which guided the establishment of an oblique transfer
zone and an associated change in fault polarity (Milani and Davison,
1988; Destro et al., 2003). The Suez-Red Sea Rift (Fig. 3a) is yet another
example where the locations and orientations of transfer or accommo-
dation zones relate to basement structures (Younes and McClay,
2002). In the last two examples many of the masterfaults are, on
average, fairly straight, as opposed to the curved geometries presented
schematically in early works on accommodation and transfer zones
(e.g., Rosendahl et al., 1986).
The strong influence of preexisting basement structures on first-
order transfer zones where strain is relayed from one side of the
rift to the other makes these structures different from smaller-scale
relay structures that may have little or no inheritance. Many, if not
most rift transfer zones are directly dictated by reactivating oblique
basement structures, whereas smaller relay structures form as the
Fig. 4. a) Uniform extension generating a population of incipient faults of which some
result of local stress interaction between overlapping fault tips.
grow, interact and link up to longer faults. b) Basement-controlled extension generating However, the location of such interacting faults may well be influ-
en-échelon faults in the cover that link up to a non-planar large fault. c) Variation of enced by preexisting structures, in which case there is an indirect
b) where the preexisting fault is inclined with respect to the extension direction. influence by preexisting structures. In general, the study of both
relay structures and transfer zones should involve a structural
study of the basement itself if possible.
18 H. Fossen, A. Rotevatn / Earth-Science Reviews 154 (2016) 14–28

Fig. 5. Zones of stress drop/increase around a normal fault (Fault 1). The growth rate of an overlapping fault tip (Fault 2) would be retarded as the tip enters the stress drop region of the
adjacent fault (Fault 1).

3.3. The process of linkage: relay ramp formation overlapping fault, but at some critical stress drop value, propagation
stops. The result is a (soft-linked) relay ramp or, if strain continues to
Once two subparallel fault segments get close enough they will be accommodated by the fault system, a breached relay (hard-link;
start to interact (Fig. 1). This is manifested by the retardation or tem- see next section) (Walsh and Watterson, 1991) (Fig. 6). If the two tips
poral arrest of the fault tips, the curving of fault tips in the overlap propagate simultaneously, both tip gradients steepen as the tips are
zone, the development of a complex zone of subsidiary structures influenced by each other's stress drop zone.
(faults, fractures, deformation bands) and the formation of a ramp. The critical spacing of faults, above which they do not interact, is
The ramp is a result of the steep displacement gradients that develop related to fault length and fault displacement; large faults have a
in the fault overlap region (Peacock and Sanderson, 1991; Nicol et al., wider stress perturbation zone than small faults and thus interact
1996), implying that the interaction of fault tips slows down the with faults that are farther away. This leads to a well-defined relation-
propagation rate of the tips involved (e.g., Maerten et al., 1999). ship between relay width and relay length (fault overlap) that is con-
Furthermore, a high displacement gradient in the fault tip regions stant over a wide range of fault sizes (Fig. 7). Soliva et al. (2006)
accentuates the ramp that forms between the two overlapping demonstrate that relay width can be related to mechanical layer
segments, whose dip direction is controlled by the arrangement thickness. In particular, observations indicate a characteristic spacing
and kinematics of the faults. of ~0.5 times the layer thickness over a wide range of scales (Fig. 15 in
The critical nearness or spacing at which two fault tips interact is of Soliva et al. (2006). This ratio is influenced by site-specific factors such
importance in understanding fault tip interaction during the growth as local fault geometry, fault weakness, rock properties, and preexisting
history of fault populations. Mechanically, this critical spacing has weak structures, which together contribute to the scatter in the data of
been related to the zone of stress reduction that occurs around faults about two orders of magnitude (Fig. 7). In the crust, mechanical layering
(e.g., Ackermann and Schlische, 1997; Cowie and Roberts, 2001; Soliva occurs at a range of scales from bed or lamina thickness to the thickness
et al., 2006) (Fig. 5). The effect of such a stress drop region has been ex- of the brittle crust. For a thickness of the brittle crust of 10–12 km,
plored by Willemse et al. (1996) and further by Gupta and Scholz a spacing of major extension faults in rift could be expected to be around
(2000), whose modeling confirmed that tip propagation is retarded as 5 km, which could produce first-order relay structures with a character-
a fault grows into the stress drop region of an overlapping fault. The istic width of ~ 5–6 km. Others use the elastic thickness of the litho-
consequence of this reduction in propagation rate is the aforementioned sphere to model the spacing (e.g., Spadini and Podladchikov, 1996).
development of a skewed displacement profile, with an average gradi- Morellato et al. (2003) found that many rifts show a characteristic
ent toward the overlap region that is up to 2.5 times that of isolated major fault spacing of 4–6 km, while other rifts show larger characteris-
fault tips (e.g., Soliva and Benedicto, 2004). The steepening of the dis- tic spacing (up to 30 km). Hence, individual relay ramps more than
placement gradient profile elevates the stress concentration at the tip, 30 km wide are observed in rifts, and in extreme cases the width can
which drives the fault to propagate into the stress drop region of the get close to 100 km (Peacock et al., 2000).

Fig. 6. Seismic data (variance timeslices) showing two overlapping faults forming a relay ramp (a) that becomes breached at depth. At 1500 ms (two-way time) the ramp is completely
breached (hard-linked). Modified from Giba et al. (2012).
 H. Fossen, A. Rotevatn / Earth-Science Reviews 154 (2016) 14–28 19

Fig. 7. Relationship between relay width and relay length (fault overlap). Data from Long and Imber's (2011) compilation, various other sources and own data. Regardless of size, ramp
length is on average 3–3.5 times their width.

3.4. The breaching of relay ramps in the Grabens area of Canyonlands National Park. We expect that
ramps in such stiff layers would breach at lower dip angles at burial
As faults overlap and the relay structure is established, the ramp will depths of N 1 km.
deform internally and breaching will eventually occur if strain keeps Giba et al. (2012) presented illustrative seismic data from a relay
accumulating. The deformation structures that form within evolving ramp offshore New Zealand (Fig. 6) that developed by reactivation of
relay structures depend greatly on the mechanical rock properties at an older fault similar to the situation portrayed in Fig. 4c. Because
the time of deformation. Soft clastic sediments at shallow depths will sedimentation occurred durig fault growth, we can see a progressively
most likely deform by non-cataclastic granular flow that over time more mature ramp as we move stratigraphically downward through
may localize into shear bands that may or may not involve strain hard- the synrift sequence, and hence an evolution from an unbreached to a
ening (e.g., Antonellini et al., 1994; Kristensen et al., 2013). Sandstones breached relay structure, as illustrated in Fig. 6. Giba et al. (2012) also
(as opposed to unlithified sand) are more likely to develop cataclastic demonstrate that the ramp becomes steeper downwards as strain
deformation bands that lead to strain hardening in the ramp and even- increases, and becomes breached as the maximum dip of the ramp
tually to faulting (Aydin et al., 2006; Fossen et al., 2007). For lithologies reaches around 13–14°. Breaching (or yield) criteria, such as the critical
that develop fractures rather than deformation bands, such as well- amount of dip or curvature of layers in the relay ramp, is potentially
indurated siliciclastic sediments (like the Canyonlands example; useful for predicting subseismic breaching faults from seismic interpre-
Fig. 2) and most limestones, effective strain weakening occurs and the tations. The limited amount of ramp dip data shown in Fig. 8 supports
breaching fault may establish itself along fracture zones. the idea that unbreached ramps have lower dips than barely breached
Several factors dictate the onset of breaching. One is the mechanical ramps, and that well-breached ramps display the largest dips. The
properties of the material within the relay zone. Soft materials such as latter may indicate that some ramps keep steepening also during the
poorly consolidated siliciclastic sedimentary layers would be expected breaching process, although relaxation may also be expected, creating
to accommodate more flexing prior to breaching than stiffer rocks a particularly wide range of dips for well-breached ramps. Clearly, the
such as well-indurated sandstone, limestone, basalt or basement rocks. critical dip depends on the material properties of the layers in the
A familiar exception is the probably still active Canyonlands National ramp (notably layer stiffness), and also on the local state of stress in
Park example, where relays are forming at the surface in quite stiff the perturbed stress field between the overlapping fault tips (Fig. 5),
Permian sandstones. In this case the fractures open in tension and both of which are related to burial depth, fault geometry, lithology
then shear (slipped fractures) so that the evolving ramp consists of and diagenetic history.
a number of extension fractures running across and along the ramp There is always a component of lengthening of the layers within
(Fig. 2). Hence some ramps reach maximum dips as high as 26° the ramp as they bend, as controlled by the ramp geometry (Ferrill
(Fossen et al., 2010) or perhaps 30° (Trudgill and Cartwright, 1994) and Morris, 2001). Additionally, there is a component of twisting of
20 H. Fossen, A. Rotevatn / Earth-Science Reviews 154 (2016) 14–28

Fig. 9. The twisting of a relay ramp as the two overlapping fault tips propagate and the
relay lengthens prior to breaching.

where both faults propagate and connect, the result becomes a lenticu-
lar relay zone (Fig. 10c), and displacement may be evenly or unevenly
distributed between the two faults. During continued accumulation of
displacement, this process results in a fault lens in map view and a
fault splay in cross-section. The Delicate Arch ramp (Fig. 11a) is an ex-
ample of an unbreached ramp where the orientation and distribution
of deformation bands in the ramp suggest that an upper-ramp breach
(Fig. 10b) was about to be established (Rotevatn et al., 2007). The
Peter Creek Ramp in Oregon, described by Crider and Pollard (1998)
(their fig. 4) is similar example where a zone of elevated fracture densi-
ty connects one of the fault tips with the other fault. In contrast, the
Canyonlands fault array seems to favor mid-ramp breaching
(Fig. 10d). The reason for this difference is probably the brittle nature
of the Canyonlands strata and the pre-existing fractures that get
reactivated. Soliva and Benedicto (2004) show small-scale examples
of this from limestones in the Pyrenees, and also some hybrid examples
of types 2 and 3. A large-scale example of a mid-ramp breach from
silisiclastic sedimentary strata in the North Sea is shown in Fig. 11b. In
Fig. 8. Distribution of maximum dip of relay ramps that are a) unbreached, b) barely general, the way that a ramp breaches is likely to be a consequence of
breached, and c) well breached. Outcrop-based data from Soliva and Benedicto (2004); geometric irregularities of the overlapping faults and lateral varia-
Huggins et al. (1995); Xu et al. (2011); Rotevatn and Bastesen (2012); Giba et al. tions in fault strength.
(2012), and Bastesen and Rotevatn (2012). Dip is relative to the general (regional) Crider and Pollard (1998) modelled ramp evolution numerically and
layer orientation.
found a concentration of Coulomb stress that bridges the two fault tips
in the relay zone. A breaching fault is expected to form in this region of
high stress values, and their idealized model predicts breaching in
the middle to upper part of the relay ramp (the range between
the layers in the ramp that develops as the fault overlap increases Figs. 10b and d). Crider and Pollard's (1998) model also suggest an ir-
(Fig. 9)—an effect that is most prominent for steep tip displacement regular or zigzagged composite bridging fault. The Grabens fault array
gradients. The stress and strain development within a ramp are also in Canyonlands National Park also does (Cartwright and Mansfield,
influenced by displacement profiles on the two faults, any non- 1998), but may not be representative for other and more common
parallelism or non-planarity of fault tips, and relative growth rates, fault populations for reasons mentioned above. In general, more de-
which makes it difficult to predict detailed stress, strain and fracture tailed observations are needed for a statistical evaluation of these char-
patterns within ramps. acteristics. The general picture is that within each fault array or
There are, in principal, three end-member classes of breached ramp faulted region a wide range of geometries are observed, typically
geometries that can be observed from field observations, seismic data with representation of all cases shown in Fig. 10 (see examples in
interpretation, physical experiments and numerical models (Fig. 10), Soliva and Benedicto, 2004).
forming by 1) single-tip breaching, 2) double-tip breaching, and It is important to realize that the transition from intact rock to a
3) mid-ramp breaching (Fig. 10). Single-tip breaching can be explained breached ramp is a gradual one that usually involves the accumulation
by one fault tip being arrested or retarded while the other tip is bending of small-scale deformation feature such as fractures, deformation
and eventually connecting with the other fault. This results in a curved bands and minor faults prior to wholesale breaching and fault
shape of the connecting fault in map view. If the upper (hanging-wall) coalescence. Hence, rocks constituting a ramp will undergo mechan-
fault tip is the one that connects (upper ramp breach), as shown in ical and petrophysical property changes throughout the evolution
Fig. 10b, the ramp is preserved in the footwall, which in a petroleum toward a fully breached ramp. As ramp-internal structures evolve,
setting could mean a trap if the ramp is large enough. In the case the scale-dependent concept of ductility becomes important. For
 H. Fossen, A. Rotevatn / Earth-Science Reviews 154 (2016) 14–28 21

Fig. 10. Patterns of relay ramp breaching. a) Unbreached relay ramp. b) Single-tip (upper-ramp) breach. c) Double breach. d) Mid-ramp breach.

example, what appears continuous and unbreached at the seismic The dip and curvature of ramps are directly reflected by the displace-
scale may reveal itself as highly deformed at outcrop scale. Another ment profiles of the two fault segments. In general, observations suggest
situation is where the overlapping fault tips are too close to be that most faults involved in relay structures (fault tip interaction) show
resolved on seismic data, resulting in the interpretation of a map-view convex displacement profiles where the gradient increases toward the
kink in the fault interpretation. tip (Fig. 12). This observation implies that the bending of layers is
Soliva and Benedicto (2004) suggested a criterion (c*) for most “forceful” in the tip region. However, some cases show a reduction
breaching based on the relationship between relay displacement in displacement gradient close to the tip (in the overlap region; see
(D, the sum of fault displacements across the relay) and relay Soliva and Benedicto (2004) for examples), implying that the curvature
width (fault separation, S) of the form D = c*S, where c* varies of the ramp is higher in the central part of the ramp. Assuming a positive
from 1 (D = S) to 0.27 (D = 0.27S) for relays that show evidence correlation between subseismic fracture and curvature, this geometry
of incipient breaching. Hence c* is a threshold value that holds favors a central breach (class 3, Fig. 10d). Furthermore, displacement
for centimeter- to kilometer-scale relays alike, suggesting that profile types vary within fault populations, indicating that they
relay-forming and -breaching processes are similar over a wide depend on local geometric or mechanical complications that are
range of scales. difficult to predict.
These geometric characteristics and their variability may have signif-
icant influence on depositional patterns where fault linkage occurs near
the surface. They may control drainage patterns and the hydrologic
conditions in a developing rift basin (Bergner et al., 2009), which
again affect the distribution of reservoir-quality deposits and thus are
of significant interest during exploration in rifts and continental
margins. In simple terms, steeper displacement variations associated
with ramp structures enhance their influence on depositional patterns,
and more work is needed to investigate the factors influencing such
structuring during fault linkage.

3.5. Fault linkage of curved fault segments

Above we considered relay structures defined by overlapping

parallel faults. However, even though large faults can be fairly straight
(Fig. 3a), particularly where guided by preexisting faults, several fault
systems are composed of fault segments that are curved in map view,
where the curved segment are interpreted as individual faults that
have linked up to a much longer fault system. This pattern is particularly
pronounced in large-scale normal fault systems, such as the Wasatch
Fault system in Utah (Fig. 13a), and rift systems such as the East
African rift system and the North Sea rift system (e.g., Scott and
Rosendahl, 1989; Fig. 13b), but occurs on all scales down to cm-scale
structures in physical models.
The Wasatch fault (Fig. 13a) is a 370 km long and composite normal
fault system composed of ten hard-linked curved segments (Machette
et al., 1991), and its characteristic curved geometry is particularly well
developed in the Provo-Salt Lake City area (Provo and Salt Lake City
segments). Curved geometries seem to be repeated at several scales
along this fault complex, but we will here confine ourselves to the
first-order structures, defined by ca. 40–50 km long segments with
several kilometers of throw. Several of the segment boundaries define
salients (“turtle backs”) that plunge westward into the Basin and
Fig. 11. Relay ramp in (a) Arches National Park (from Rotevatn et al., 2009b) and (b) the
North Sea (Murchison Field area, based on Young et al., 2001). The latter example is
Range extensional province.
based on seismic interpretation, hence small (subseismic) structures are not displayed Strikingly similar fault geometries of comparable size are seen in the
(but can be inferred from a). Jurassic sedimentary sequence of the northern North Sea rift system, for
22 H. Fossen, A. Rotevatn / Earth-Science Reviews 154 (2016) 14–28

Fig. 12. Displacement variations for a) non-interacting and b) interacting faults, measured from the point of maximum displacement (Dmax) to the fault tip. While non-interacting faults
show a close to linear displacement gradient, the more convex shape of the interacting fault data reflect the effect of growth restriction at sites of fault interaction (relays). Data from Nicol
et al. (1996) and Soliva et al. (2006).

Fig. 13. a) The Wasatch Fault in the Salt Lake area, Utah, portraying a strongly curved fault trace in map view. b) Strikingly similar fault patterns in the northern North Sea (base Cretaceous
unconformity). First-order faults with km-scale displacements are indicated (simplified).
 H. Fossen, A. Rotevatn / Earth-Science Reviews 154 (2016) 14–28 23

first-order faults with km-scale offsets (Fig. 13b) (Fossen et al., 2000). do not know the reason for these variations along this particular fault
Also in this area the segment boundaries define salients or cusps that system, but a curved geometry seems to be a common feature of
point toward the down-faulted hanging wall side, generally toward many large-scale extensional faults.
the rift axis. As a first-order approach to the evolution of both the North Sea
There are certain models and conditions that can explain curved and Wasatch Fault examples, we suggest that the faults initiated as
fault patterns in map view, although such fault patterns are not always individual segments whose fault tips at some point interacted at the
easy to understand. One factor that may be particularly important for locations of the salient. This model predicts maximum structural com-
large faults, such as the ones mentioned above, is the influence of plication and minimum displacement in the areas of fault interaction
preexisting heterogeneities, including the reactivation of older faults (i.e., in the salients), and depocenters near the middle of the curved
or fault arrays (Sevier thrusts in the case of the Wasatch Fault) that fault segments. Seismically this is consistent with the characteristic
formed in different stress regime(s). Reactivation of non-planar thrusts earthquake model (Schwartz and Coppersmith, 1984), which for seg-
is a well-known factor in the Basin and Range province (Coney and mented fault systems predicts that large earthquakes with similar
Harms, 1984) and reactivation of Permo-Triassic normal faults or (characteristic) size are repeated on individual segments. These charac-
Devonian extensional structures was important in the development of teristics are present in the Salt Lake and Gullfaks areas (Machette et al.,
the mid-late Jurassic North Sea rift system (Færseth, 1996; Fossen 1991; Fossen et al., 2000), and the similarity in both overall fault geom-
et al., 2000). Preexisting faults can dictate the location of younger faults etry and presence and location of subsidiary faults within the salients is
as well as their orientation and geometry. Interestingly the Wasatch striking (Fig. 14). For faults with km-scale offsets, gravitational destabi-
Fault system developed straighter fault segments in some areas, such lization of elevated footwall salients may add to the tectonically-driven
as the Weber segment north of the Salt Lake segment. At present we extensional faulting. Hence gravity-influenced crestal collapse may be

Fig. 14. Structural maps and cross-sections of the Salt Lake (a–b) and Gullfaks (c–d) salients. Geometrically the two areas are almost mirror images of each other, and we propose that both
formed as a result of fault linkage.
24 H. Fossen, A. Rotevatn / Earth-Science Reviews 154 (2016) 14–28

Fig. 15. The formation of curved faults (in map view) according to the model suggested by Wu and Bruhn (1994). The shear component in the tip regions of active faults causes the fault to
step out of its own plane in the sense shown in the figure, thereby causing a curved fault trace as the segments link up to a continuous fault surface.

expected where the salient parts of large fault blocks are tilted and water. Relay zones are therefore the loci of a range of fluid–rock in-
where footwall uplift is prominent. Gravitational sliding near the teractive processes, many of which are of economic significance.
surface will be the most obvious effect, but if the salient and fault One such example is provided in a study of a segmented rift system
block are large enough it may also have an effect on the deeper parts in New Zealand by Rowland and Sibson (2004), where a concentra-
of the system. A characteristic feature of both the Salt Lake and the tion of geothermal fields in fault stepovers were identified, and
Gullfaks salients is the formation of subsidiary (“shortcut”) faults that linked to enhanced vertical permeability caused by high structural
transect the salients, transferring some of the displacement more complexity in these zones. Evidence for increased geothermal fluid
directly between the main fault segments (Fig. 14). activity linked to fault linkage zones was also reported by Curewitz
A different model for curved fault traces employs the fact that lateral and Karson (1997), and by Dockrill and Shipton (2010) who showed
shear (strike–slip) components occur in the lateral tip regions of a that CO2 springs and seeps were co-located with structurally complex
normal fault because hanging-wall subsidence greatly exceeds footwall zones in a fault array. Also in volcanic systems, magma emplacement
uplift (Wu and Bruhn, 1994; Roberts, 1996). It has been suggested that has been correlated with dilational jogs or transverse fault relays
this asymmetry may generate en-échelon faults ahead of the fault tips in (Vigneresse and Bouchez, 1997) and extensional transfer zones (Dini
map view. Since, in accordance with the sense of shear in the tip regions, et al., 2008). Also in association with magmatism, where hydrothermal
the fault will step into the footwall, the result may be a curved fault as activity is high, economically important hydrothermal mineral deposits
these fault segments link up, as shown in Fig. 15 (Wu and Bruhn, may be related to fault relays or intersections. As an example, copper
1994). This model implies that the point of maximum displacement mineralization along the Lisbon Valley Fault and the Dolores Zone
occurs in the salients or convex parts of the curved faults (Fig. 15), of faults in Utah and Colorado (Breit and Meunier, 1990) are localized
which makes for a simple test. While this model does not seem to at relay zones. Furthermore, fault jogs, which commonly represent
apply to the Wasatch and Gullfaks faults (Fig. 14), Wu and Bruhn the locations of previous fault linkage zones, have been shown to be
(1994) suggest that the geometric pattern of the South Oquirrh associated with hydrothermal gold deposits in strike-slip systems
Mountains normal fault zone fits this model. (Micklethwaite and Cox, 2004, 2006).
Due to the ability of the fault relay zones to control fluid transport,
4. Consequences for fluid flow and therefore the loci of fluid–rock interaction, fault relays may also ex-
ercise strong control on diagenesis. For example, Eichhubl et al. (2009)
Relay ramps, breached or not, typically represent potential pathways showed that fluid migration and cementation along the Moab Fault
for vertical migration of fluids (Fig. 16). The reason for this is the (Utah) was focused at areas of fault linkage or intersection. Another
increased structural complexity found at fault relays, with increased example is presented by Sharp et al. (2010), where dolomitization of
numbers of faults and fractures and a wider range of orientations than Cretaceous carbonates in the Zagros Mountains of Iran have been
that of single, isolated faults (e.g. Sibson, 1996; Peacock and Parfitt, subject to strong structural controls; particularly, the authors note that
2002; Kim et al., 2004; Fossen et al., 2005). Relay zones therefore alteration is most extensive at faults/joint intersections.
represent an important control on fluid transport in the crust, affecting Fault relays and intersections are also important for the accumula-
all kinds of fluids, including hydrocarbons, CO2 and other volatiles, tion and retainment of hydrocarbons. They may act as vertical conduits
hydrothermal solutions, metamorphic fluids, magma, and ground for fluid migration into traps but may also have negative effect on seal
 H. Fossen, A. Rotevatn / Earth-Science Reviews 154 (2016) 14–28 25

Fig. 16. Fluid flow-pattern through a relay structure (layer A). A sealing fault may conduct fluids in the vertical direction (1), but much more efficiently so through a heavily fractured relay
structure (3). In a reservoir setting the relay structure typically provide bed-parallel hanging-wall–footwall communication (2). Layer B does not show a ramp geometry and is not
associated with flow of types 2 and 3.

integrity. In a study of petroleum fields in the Timor Sea, Australia, orientation variability in the linking damage zone leads to an increased
Gartrell et al. (2004) identified fault intersections as critical hydrocar- fracture:matrix ratio and connectivity, both of which lead to a higher
bon leakage zones. This is also supported by findings by Kristiansen overall effective permeability (e.g. Berkowitz, 1995). Adding to this is
(2011), who found that for fault-controlled structural hydrocarbon the fact that, when there is a great range in orientation, the chance of
traps in the Barents Sea (Norway), traps were generally underfilled or some of the fractures being optimally oriented for opening is great
dry where controlled by two or more interacting faults, whereas all under most stress conditions. Rotation of the local stress field, which
discoveries were associated with fault traps controlled by a single fault. is common in relay zones (Kattenhorn et al., 2000) may lead to principal
In addition to these aspects relevant for hydrocarbon exploration, stress orientations which, if oriented optimally with respect to fractures,
fault relay zones may also affect fluid flow between structural reservoir may promote greater fracture apertures that increase permeability
compartments on a production time scale. Soft-linked relay zones may (e.g. Tamagawa and Pollard, 2008). Relevant to this, drilling of active
offer cross-fault reservoir connectivity through folded but unbreached fault systems demonstrated that, although a small percentage of fault-
relay beds (Bense and Baalen, 2004; Manzocchi et al., 2008; Rotevatn related fractures only may be optimally oriented for opening under
et al., 2009a). However, this effect may be reduced if the relay zone is the regional stress regime, these maintain the largest hydraulic conduc-
associated with a linking damage zone of low-permeable deforma- tivities at any given time (Barton et al., 1995; Davatzes and Hickman,
tion bands in porous sandstone host rocks (Rotevatn et al., 2009b, 2005). Dilatant stress conditions at vertical and lateral fault jogs or
Rotevatn and Fossen, 2011). Another way that relay zones may aid intersections have also previously been suggested (e.g. Ferrill and
cross-fault fluid flow is the juxtaposition of multiple reservoir units Morris, 2003, Gartrell et al., 2004), and have important implications
at different stratigraphic levels due to folded beds and steep for the permeability structure at such locations.
displacement gradients (Manzocchi et al., 2010, their fig. 12). Relay-enhanced fault permeability should also be considered in a
In non-porous rocks, where fractures comprise most of the perme- temporal perspective. Whereas intact relays may provide cross-fault
ability, bed continuity in itself through a relay zone may have limited connectivity in porous sandstones, a breached relay in such rocks will
effect on cross-fault flow. However, greater density and variety in orien- represent a zone of enhanced low-permeable damage in the form of
tation of the permeability-controlling fractures may aid also cross-fault deformation bands, in addition to the breaching fault itself. In such
flow. Along isolated fault segments, high-permeablity fracture systems rocks, cross-fault permeability may therefore be increased in the
and fault slip surfaces generally have a positive effect on fault-parallel presence of a soft-linked relay (Rotevatn et al., 2009b), but reduced in
permeability, but the effect on cross-fault permeability is limited the presence of a breached relay. This contrasts findings in fractured
(Jourde et al., 2002) as most fractures are oriented subparallel to carbonate reservoir rocks (Bastesen and Rotevatn, 2012, Rotevatn
the fault. In relay zones, however, increased fracture intensity and and Bastesen, 2012), where it was shown that breached relays in
26 H. Fossen, A. Rotevatn / Earth-Science Reviews 154 (2016) 14–28

Fig. 17. Two fault maps from the North Sea, showing the fault pattern of the Beatrice Field (Inner Moray Firth basin) and that of the structurally more mature Gullfaks Field. Note how
faults are longer and better connected in the Gullfaks Field, where multiple previous relays can be inferred from kinks and jogs in the fault traces. The highly compartmentalized
Gullfaks Field requires many wells for efficient production while fewer wells are needed to produce the Beatrice Field with its well-connected fault blocks. Based on Husmo et al.
(2002) and Fossen (2010).

such rocks may provide better cross-fault connectivity compared to of the structural elements of which they are constituted. While their
soft-linked relays. The reason for this is that in such rocks, fracture tendency to reduce lateral fluid flow within many sandstone reser-
system development and permeability enhancement was found to voirs is well known, their possible role as vertical conduits is less
progress during continual relay growth and linkage, and to be greatest explored. Particular attention should be paid to this aspect during
at breached relays. planning and monitoring of CO2 sequestration and hydrocarbon
A consequence of fault growth through linkage is the development exploration/production alike.
of long and continuous faults at advanced stages. As relay ramps form More detailed work is needed to understand the variability of ramp
and breach, these long faults compartmentalize reservoirs and may pro- evolution and destruction. Detailed geometric analysis of ramps at
duce long and isolated fault blocks that act as individual fluid compart- various scales and settings are needed to relate geometry (dip,
ments. The linkage and breaching of relays leads to more connected curvature) to small-scale damage and to location of breaching. The
fault systems, and a relationship between strain and fault population influence of three-dimensional fault geometry on ramp development
maturity exists. The Gullfaks Field in the northern North Sea (Fossen is also needed: ramps are commonly considered at a single stratigraphic
and Hesthammer, 1998) may serve as an example of a structurally ma- level only, while true fault geometry may have important bearings on
ture fault system where high degree of fault connectivity and structural their development.
complexity has required a high number of production and injection
wells (Fig. 17). The strain in the part of the Gullfaks Field shown in
Acknowledgments
Fig. 17b is more than 40%, approximately twice that of the structurally
less mature fault array of the Beatrice field shown in Fig. 17a.
We are grateful for careful and constructive reviews by George Hilley
and editor Manfred Strecker.
5. Concluding remarks

The formation of relay structures during the development of normal References

fault populations is extremely common, from the cm-scale to tens of Ackermann, R.V., Schlische, R.W., 1997. Anticlustering of small normal faults around
kilometers in width. They portray geometric properties that appear to larger faults. Geology 25, 1127–1130.
be scale independent, including their shape (3–3.5 times longer than Acocella, V., Gudmundsson, A., Funiciello, R., 2000. Interaction and linkage of extension
fractures and normal faults: examples from the rift zone of Iceland. J. Struct. Geol.
their width) and the displacement variations along overlapping fault
22, 1233–1246.
tips and therefore ramp geometry. Typically the steepness of ramps at Allken, V., Huismans, R.S., Fossen, H., Thieulot, C., 2013. 3D numerical modelling of
the point of breaching is around 10–15°, and the breaching process graben interaction and linkage: a case study of the Canyonlands grabens, Utah.
itself results in a wide and complicated zone of damage that is largely Basin Res. 25, 14.
Anders, M.H., Schlische, R.W., 1994. Overlapping faults, intrabasin highs, and the growth
inactivated after breaching. These damage zones may have different of normal faults. J. Geol. 102, 165–180.
functionality with respect to fluid flow, depending on the properties Anderson, E.M., 1951. The Dynamics of Faulting. Oliver & Boyd, Edinburgh.
 H. Fossen, A. Rotevatn / Earth-Science Reviews 154 (2016) 14–28 27

Antonellini, M.A., Aydin, A., Pollard, D.D., 1994. Microstructure of deformation bands in Fossen, H., Hesthammer, J., 1998. Structural geology of the Gullfaks Field, northern North
porous sandstones at Arches National Park, Utah. J. Struct. Geol. 16, 941–959. Sea. Geol. Soc. Lond. Spec. Publ. 127, 231–261.
Athmer, W., Groenenberg, R.M., Luthi, S.M., Donselaar, M.E., Sokoutis, D., Willingshofer, E., Fossen, H., Johansen, T.E.S., Hesthammer, J., Rotevatn, A., 2005. Fault interaction in porous
2010. Relay ramps as pathways for turbidity currents: a study combining analogue sandstone and implications for reservoir management; examples from Southern
sandbox experiments and numerical flow simulations. Sedimentology 57, 806–823. Utah. AAPG Bull. 89, 1593–1606.
Aydin, A., Borja, R.I., Eichhubl, P., 2006. Geological and mathematical framework for Fossen, H., Odinsen, T., Færseth, R.B., Gabrielsen, R.H., 2000. Detachments and low-angle
failure modes in granular rock. J. Struct. Geol. 28, 83–98. faults in the northern North Sea rift system. Geol. Soc. Lond. Spec. Publ. 167, 105–131.
Barton, C.A., Zoback, M.D., Moos, D., 1995. Fluid flow along potentially active faults in Fossen, H., Schultz, R.A., Rundhovde, E., Rotevatn, A., Buckley, S., 2010. Fault linkage
crystalline rock. Geology 23, 683–686. and graben stepovers in the Canyonlands (Utah) and the North Sea Viking
Bastesen, E., Rotevatn, A., 2012. Evolution and structural style of relay zones in layered Graben, with implications for hydrocarbon migration and accumulation. AAPG
limestone–shale sequences: insights from the Hammam Faraun Fault Block, Suez Bull. 94, 597–613.
Rift, Egypt. Geol. Soc. Lond. 169, 477–488. Fossen, H., Schultz, R.A., Shipton, Z.K., Mair, K., 2007. Deformation bands in sandstone—a
Bense, V.F., Baalen, R.V., 2004. The effect of fault relay and clay smearing on groundwater review. Geol. Soc. Lond. 164, 755–769.
flowpatterns in the Lower Rhine Embayment. Basin Res. 16, 397–411. Gartrell, A., Zhang, Y., Lisk, M., Dewhurst, D., 2004. Fault intersections as critical
Bergner, A.G.N., Strecker, M.R., Trauth, M.H., Deino, A.L., Gasse, F., Blisniuk, P., Dünforth, hydrocarbon leakage zones: integrated field study and numerical modelling of an
M., 2009. Tectonic and climatic control on evolution of rift lakes in the Central example from the Timor Sea, Australia. Mar. Pet. Geol. 21, 1165–1179.
Kenya Rift, East Africa. Quat. Sci. Rev. 28, 2804–2816. Gawthorpe, R.L., Hurst, J.M., 1993. Transfer zones in extensional basins: their structural
Berkowitz, B., 1995. Analysis of fracture network connectivity using percolation theory. style and influence on drainage development and stratigraphy. Geol. Soc. Lond.
Math. Geol. 27, 467–483. 150, 1137–1152.
Bosworth, W., 1985. Geometry of propagating continental rifts. Nature 316, 625–627. Gawthorpe, R., Leeder, M.R., 2000. Tectono-sedimentary evolution of active extensional
Bosworth, W., 1987. Off-axis volcanism in the Gregory rift, east Africa: implications for basins. Basin Res. 12, 195–218.
models of continental rifting. Geology 15, 397–400. Giba, M., Walsh, J.J., Nicol, A., 2012. Segmentation and growth of an obliquely reactivated
Bosworth, W., 2015. Geological Evolution of the Red Sea: Historical Background, Review, normal fault. J. Struct. Geol. 39, 253–267.
and Synthesis. pp. 45–78. Gibbs, A.D., 1984. Structural evolution of extensional basin margins. Geol. Soc. Lond. 141,
Bott, M.H.P., 1959. The mechanics of oblique slip faulting. Geol. Mag. 46, 110–117. 609–620.
Breit, G.N., Meunier, J., 1990. Fluid inclusion, d18O, and 87Sr/86Sr evidence for the origin Goguel, J., 1952. Traité de Tectonique. Masson, Paris.
of fault-controlled copper mineralization, Lisbon Valley, Utah, and Slick Rock District, Gupta, A., Scholz, C.H., 2000. A model of normal fault interaction based on observations
Colorado. Econ. Geol. 85, 884–891. and theory. J. Struct. Geol. 22, 865–879.
Cartwright, J.A., Mansfield, C.S., 1998. Lateral displacement variation and lateral Hetzel, R., Strecker, R., 1994. Late Mozambique Belt structures in western Kenya and their
tip geometry of normal faults in Canyonlands National Park, Utah. J. Struct. influence on the evolution of the Cenozoic Kenya Rift. J. Struct. Geol. 2, 189–201.
Geol. 20, 3–19. Huggins, P., Watterson, J., Walsh, J.J., et al., 1995. Relay zone geometry and displace-
Childs, C., Manzocchi, T., Walsh, J.J., Bonson, C.G., Nicol, A., Schöpfer, M.P.J., 2009. ment transfer between normal faults recorded in coal-mine plans. J. Struct.
A geometric model of fault zone and fault rock thickness variations. J. Struct. Geol. 17, 1741–1755.
Geol. 31, 117–127. Lower and middle Jurassic. In: Husmo, T., Hamar, G.P., Høiland, O., Johannesen, E.P.,
Coney, P.J., Harms, T.A., 1984. Cordilleran metamorphic core complexes: Cenozoic Rømuld, A., Spencer, A.M., Tritterton, R. (Eds.), The Millennium Atlas: Petroleum
extensional relics of Mesozoic compression. Geology 12, 550–555. Geology of the Central and Northern North Sea. Geol. Soc., London, pp. 129–155.
Cowie, P., Roberts, G.P., 2001. Constraining slip rates and spacings for active normal faults. Jackson, C.A.L., Rotevatn, A., 2013. 3D seismic analysis of the structure and evolution of a
J. Struct. Geol. 23, 1901–1915. salt-influenced normal fault zone: a test of competing fault growth models. J. Struct.
Cowie, P., Gupta, S., Dawers, N.H., 2000. Implications of fault array evolution for synrift Geol. 53. http://dx.doi.org/10.1016/j.jsg.2013.06.012.
depocentre development: insights from a numerical fault growth model. Basin Res. Jackson, C.A.L., Gawthorpe, R., Sharp, I.R., 2002. Growth and linkage of the East Tanka
12, 241–261. fault zone, Suez rift: structural style and syn-rift stratigraphic response. J. Geol. Soc.
Cox, S.F., 2005. Coupling between deformation, fluid pressures, and fluid flow in 159, 175–187.
ore-producing hydrothermal systems at depth in the crust. Econ. Geol. 100th Jourde, H., Flodin, E.A., Aydin, A., Durlofsky, et al., 2002. Computing permeabilities of
Anniversary, 39–76. fault zones in eolian sandstone from outcrop measurements. AAPG Bull. 86,
Crider, J.G., Pollard, D.D., 1998. Fault linkage: three-dimensional mechanical interaction 1187–1200.
between echelon normal faults. J. Geophys. Res. 103, 675–692. Kattenhorn, S.A., Aydin, A., Pollard, D.D., 2000. Joints at high angles to normal fault strike:
Curewitz, D., Karson, J.A., 1997. Structural settings of hydrothermal outflow: fracture an explanation using 3-D numerical models of fault-perturbed stress fields. J. Struct.
permeability maintained by fault propagation and interaction. J. Volcanol. Geotherm. Geol. 22, 1–23.
Res. 79, 149–168. Kim, Y.-S., Peacock, D.C.P., Sanderson, D.J., 2004. Fault damage zones. J. Struct. Geol. 26,
Dahlstrom, C.D.A., 1969. Balanced cross sections. Can. J. Earth Sci. 6, 743–757. 503–517.
Davatzes, N.C., Hickman, S.H., 2005. Controls on fault-hosted fluid flow; Preliminary Kinabo, B.D., Atekwana, E.A., Hogan, J.P., Modisi, M.P., Wheaton, D.D., Kampunzu, A.B.,
results from the Coso geothermal field. Geothermal Research Council Transactions, 2007. Early structural development of the Okavango rift zone, NW Botswana. J. Afr.
paper, CA, p. 144. Earth Sci. 48, 125–136.
Delaney, P.T., Pollard, D.D., 1981. Deformation of host rocks and flow of magma Kristensen, M.B., Childs, C., Olesen, N.Ø., Korstgård, J.A., 2013. The microstructure and
during growth of minette dikes and breccia-bearing intrusions near Ship Rock, internal architecture of shear bands in sand–clay sequences. J. Struct. Geol. 46,
New Mexico. USGS Professional Paper 1202 (61 pp.). 129–141.
Destro, N., Alkmim, F.F., Magnavita, L.P., Szatmari, P., 2003. The Jeremoabo transpressional Kristiansen, K., 2011. Vertical fault leakage in the western part of the Hammerfest Basin.
transfer fault, recôncavo–Tucano rift, NE Brazil. J. Struct. Geol. 25, 1263–1279. Unpublished MSc thesis, University of Bergen, (94 pp.)
Dini, D.S.W., Innocenti, F., Rocchi, S., 2008. Magma emplacement in a transfer zone: the Larsen, P.-H., 1988. Relay structures in a Lower Permian basement-involved extension
Miocene mafic Orano dyke swarm of Elba Island, Tuscany, Italy. Geol. Soc. Lond. system, East Greenland. J. Struct. Geol. 10, 3–8.
Spec. Publ. 302, 131–148. http://dx.doi.org/10.1144/SP302.10. Long, J.J., Imber, J., 2011. Geological controls on fault relay zone scaling. J. Struct. Geol. 33,
Dockrill, B., Shipton, Z.K., 2010. Structural controls on leakage from a natural CO2 geologic 1790–1800.
storage site: Central Utah, U.S.A. J. Struct. Geol. 32. Machette, M.N., Personius, S.F., Nelson, A.R., Schwartz, D.P., Lund, W.R., 1991. The Wasatch
Duval, B., Cramez, C., Jackson, M.P.A., 1992. Raft tectonics in the Kwanza Basin, Angola. fault zone, Utah—segmentation and history of Holocene earthquakes. J. Struct. Geol.
Mar. Pet. Geol. 9, 389–390. 13, 137–149.
Eichhubl, P., Davatzes, N.C., Becker, S.P., 2009. Structural and diagenetic control of fluid Maerten, L., Willemse, E.J.M., Pollard, D.D., et al., 1999. Slip distributions on intersecting
migration and cementation along the Moab fault, Utah. Am. AAPG Bull. 93, 653–681. normal faults. J. Struct. Geol. 21, 259–271.
Elliott, G.M., Wilson, P., Jackson, C.A.L., Gawthorpe, R.L., Michelsen, L., Sharp, I.R., 2011. The Manighetti, I., Zigone, D., Campillo, M., Cotton, F., 2009. Self-similarity of the largest-scale
linkage between fault throw and footwall scarp erosion patterns: an example from segmentation of the faults: implications for earthquake behavior. Earth Planet. Sci.
the Bremstein Fault Complex, offshore Mid-Norway. Basin Res. 23, 1–18. Lett. 288, 370–381.
Færseth, R.B., 1996. Interaction of Permo-Triassic and Jurassic extensional fault-blocks Manzocchi, T., Childs, C., Walsh, J.J., 2010. Faults and fault properties in hydrocarbon flow
during the development of the northern North Sea. J. Geol. Soc. 153, 931–944. models. Geofluids 10, 94–113.
Faulds, J.E., Varga, R.J., 1998. The role of accommodation zones and transfer zones in Manzocchi, T., Heath, A.E., Palananthakumar, B., Childs, C., Walsh, J.J., 2008. Faults in
the regional segmentation of extended terranes. Geol. Soc. Am. Spec. Paper 323 conventional flow simulation models: a consideration of representational
(45 pp.). assumptions and geological uncertainties. Pet. Geosci. 14, 91–110.
Faulds, J.E., Hinz, N.H., Dering, G.M., Siler, D.L., 2013. The hybrid model —the most accom- McClay, K.R., Ellis, P.G., 1987. Analogue models of extensional fault geometries. Geol. Soc.
modating structural setting for geothermal power generation in the Great Basin, Lond. Spec. Publ. 28, 109–125.
Western USA. GRC Trans. 37, 3–10. Micklethwaite, S., Cox, S.F., 2004. Fault-segment rupture, aftershock-zone fluid flow, and
Ferrill, D.A., Morris, A.P., 2001. Displacement gradient and deformation in normal fault mineralization. Geology 32, 813.
systems. J. Struct. Geol. 23, 619–638. Micklethwaite, S., Cox, S., 2006. Progressive fault triggering and fluid flow in aftershock
Ferrill, D.A., Morris, A.P., 2003. Dilational normal faults. J. Struct. Geol. 25, 183–196. domains: examples from mineralized Archaean fault systems. Earth Planet. Sci. Lett.
Finzi, Y., Langer, S., 2012. Damage in step-overs may enable large cascading earthquakes. 250, 318–330.
Geophys. Res. Lett. http://dx.doi.org/10.1029/2012GL052436 (5 pp.). Milani, E.J., Davison, I., 1988. Basement control and transfer tectonics in the Recôncavo–
Fort, X., Brun, J.P., Chauvel, F., 2004. Salt tectonics on the Angolan margin, synsedimentary Tucano–Jatobá rift, Northeast Brazil. Tectonophysics 154, 41–70.
deformation processes. AAPG Bull. 88, 1523–1544. Morellato, C., Redini, F., Doglioni, C., 2003. On the number and spacing of faults. Terra
Fossen, H., 2010. Structural Geology. Cambridge University Press (463 pp.). Nova 15, 315–321.
28 H. Fossen, A. Rotevatn / Earth-Science Reviews 154 (2016) 14–28

Morley, C.K., Nelson, R.A., Patton, T.L., Munn, S.G., 1990. Transfer zones in the East African Sibson, R.H., 1996. Structural permeability of fluid-driven fault-fracture meshes. J. Struct.
rift system and their relevance to hydrocarbon exploration in rifts. AAPG Bull. 74, Geol. 18, 1031–1042.
1234–1253. Smith, M., Mosley, P., 1993. Crustal heterogeneity and basement influence on the
Nicol, A., Gillespie, P., Childs, C., Walsh, J.J., 2002. Relay zones between mesoscopic thrust development of the Kenya Rift, East Africa. Tectonics 12, 591–606.
faults in layered sedimentary sequences. J. Struct. Geol. 24, 709–727. Soliva, R., Benedicto, A., 2004. A linkage criterion for segmented normal faults. J. Struct.
Nicol, A., Watterson, J., Walsh, J.J., Gillespie, P.A., 1996. The shapes, major axis orientations Geol. 26, 2251–2267.
and displacement patterns of fault surfaces. J. Struct. Geol. 18, 235–248. Soliva, R., Schultz, R.A., 2008. Distributed and localized faulting in extensional settings:
Peacock, D.C.P., Parfitt, E.A., 2002. Active relay ramps and normal fault propagation on insight from the North Ethiopian Rift–Afar transition area. Tectonics 27. http://dx.
Kilauea Volcano, Hawaii. J. Struct. Geol. 24, 729–742. doi.org/10.1029/2007TC002148 (19 pp.).
Peacock, D.C.P., Sanderson, D.J., 1991. Displacements, segment linkage and relay ramps in Soliva, R., Benedicto, A., Maerten, L., 2006. Spacing and linkage of confined normal faults:
normal fault zones. J. Struct. Geol. 13, 721–733. importance of mechanical thickness. J. Geophys. Res. 111 (B1), 17p. http://dx.doi.org/
Peacock, D.C.P., Price, S.P., Pickles, C.S., 2000. The world's biggest relay ramp: hold with 10.1029/2004JB003507.
hope, NE Greenland. J. Struct. Geol. 22, 843–850. Spadini, G., Podladchikov, Y., 1996. Spacing of consecutive normal faulting in the
Podolsky, D.M.W., Roberts, G.P., 2008. Growth of the volcano-flank Koa'e fault system, lithosphere: a dynamic model for rift axis jumping (Tyrrhenian Sea). Earth Planet.
Hawaii. J. Struct. Geol. 30 (10), 1254–1263. Sci. Lett. 144, 21–34.
Roberts, G.P., 1996. Variation in fault-slip directions along active and segmented normal Tamagawa, T., Pollard, D.D., 2008. Fracture permeability created by perturbed stress fields
fault systems. J. Struct. Geol. 18, 835–845. around active faults in a fractured basement reservoir. AAPG Bull. 92, 743–764.
Rosendahl, B.R., Reynolds, D.J., Lorber, Burgess, C.F., McGill, J., Scott, D., Lambiase, J.J., Tentler, T., Mazzoli, S., 2005. Architecture of normal faults in the rift zone of central north
Derksen, S.J., 1986. Structural expressions of rifting: lessons from Lake Tanganykia, Iceland. J. Struct. Geol. 27, 1721–1739.
Africa. Geol. Soc. Lond. Spec. Publ. 25, 29–43. Trudgill, B., Cartwright, J., 1994. Relay-ramp forms and normal-fault linkages,
Rotevatn, A., Bastesen, E., 2012. Fault linkage and damage zone architecture in tight Canyonlands National Park, Utah. Geol. Soc. Am. Bull. 106, 1143–1157.
carbonate rocks in the Suez Rift (Egypt): implications for permeability structure Vendeville, B., Cobbold, P.R., 1988. How normal faulting and sedimentation interact to
along segmented normal faults. Geol. Soc. Lond. Spec. Publ. 374. http://dx.doi.org/ produce listric fault profiles and stratigraphic wedges. J. Struct. Geol. 10, 649–659.
10.1144/SP374.12. Vendeville, B., Cobbold, P.R., Davy, P., Brun, J.P., Choukroune, P., 1987. Physical models of
Rotevatn, A., Fossen, H., 2011. Simulating the effect of subseismic fault tails and process extensional tectonics at various scales. Geol. Soc. Lond. Spec. Publ. 28, 95–107.
zones in a siliciclastic reservoir analogue: implications for aquifer support and trap Vigneresse, J.L., Bouchez, J.L., 1997. Successive granitic magma batches during pluton
definition. Mar. Pet. Geol. 28, 1648–1662. emplacement: the Case of Cabeza de Araya (Spain). J. Petrol. 38, 1767–1776.
Rotevatn, A., Buckley, S.J., Howell, J.A., Fossen, H., 2009a. Overlapping faults and their Walsh, J.J., Watterson, J., 1991. Geometric and kinematic coherence and scale effects in
effect on fluid flow in different reservoir types: a LIDAR-based outcrop modeling normal fault systems. Geol. Soc. Lond. Spec. Publ. 56, 193–203.
and flow simulation study. AAPG Bull. 93, 407–427. Walsh, J.J., Bailey, W.R., Childs, et al., 2003. Formation of segmented normal faults: a 3-D
Rotevatn, A., Fossen, H., Hesthammer, J., 2007. Are relay ramps conduits for fluid flow? perspective. J. Struct. Geol. 25, 1251–1262.
Structural analysis of a relay ramp in Arches National Park, Utah. Geol. Soc. Lond. Wernicke, B., Burchfiel, B.C., 1982. Modes of extensional tectonics. J. Struct. Geol.
Spec. Publ. 270, 55–71. 4, 105–115.
Rotevatn, A., Tveranger, J., Howell, J.A., 2009b. Dynamic investigation of the effect of a Willemse, E.J.M., 1997. Segmented normal faults: correspondence between three-
relay ramp on simulated fluid flow: geocellular modelling of the Delicate Arch dimensional mechanical models and field data. J. Geophys. Res. 1902, 675–692.
Ramp, Utah. Pet. Geosci. 15, 45–58. Willemse, E.J.M., Pollard, D.D., Aydin, A., 1996. Three-dimensional analyses of slip
Rouby, D., Nalpas, T., Jermannaud, P., 2011. Gravity driven deformation controlled by the distributions on normal fault arrays with consequences for fault scaling. J. Struct.
migration of the delta front: the Plio-Pleistocene of the Eastern Niger Delta. Geol. 18, 295–309.
Tectonophysics 513, 54–67. Wu, D., Bruhn, R.L., 1994. Geometry and kinematics of active normal faults, South Oquirrh
Rowland, J.V., Sibson, R.H., 2004. Structural controls on hydrothermal flow in a segmented Mountains, Utah: implication for fault growth. J. Struct. Geol. 16, 1061–1075.
rift system, Taupo Volcanic Zone, New Zealand. Geofluids 4, 259–283. Wu, J.E., McClay, K., Frankowicz, E., 2015. Niger Delta gravity-driven deformation above
Schlische, R.W., Withjack, M.O., 2009. Origin of fault domains and fault-domain the relict Chain and Charcot oceanic fracture zones, Gulf of Guinea: insights from
boundaries (transfer zones and accommodation zones) in extensional provinces: analogue models. Mar. Pet. Geol. 65, 43–62.
result of random nucleation and self-organized fault growth. J. Struct. Geol. 31, Xiao-shuang, X., Jing-ru, T., Hua, K., Shao-xun, H., 2005. Control of relay structure on min-
910–925. eralization of sedimentary-exhalative ore deposit in growth faults of graben systems.
Schwartz, D.P., Coppersmith, K.J., 1984. Fault behavior and characteristic earthquakes: J. Cent. S. Univ. Technol. 12, 340–345.
examples from the Wasatch and San Andreas Fault Zones. J. Geophys. Res. 89 Xu, S., Nieto-Samaniego, A.F., Alaniz-Alvarez, S.A., et al., 2011. Structural analysis of a
(B7), 5681. relay ramp in the Querétaro graben, central Mexico: implications for relay ramp
Scott, D.L., Rosendahl, B.R., 1989. North Viking Graben: an east African perspective. AAPG development. Rev. Mex. Cienc. Geol. 28, 275–289.
Bull. 73, 155–165. Younes, A.I., McClay, K., 2002. Development of accommodation zones in the Gulf of
Sharp, I., Gillespie, P., Morsalnezhad, D., Taberner, C., Karpuz, R., Verge, S.,.J., Horbury, A., Suez–Red Sea rift, Egypt. AAPG Bull. 86, 1003–1026.
Pickard, N., Garland, J., Hunt, D., 2010. Stratigraphic architecture and fracture- Young, M.J., Gawthorpe, R.L., Hardy, S., 2001. Growth and linkage of a segmented normal
controlled dolomitization of the Cretaceous Khami and Bangestan groups: an outcrop fault zone; the Late Jurassic Murchison–Statfjord North Fault, northern North Sea.
case study, Zagros Mountains, Iran. Geol. Soc. Spec. Publ. 329, 343–396. J. Struct. Geol. 23, 1933–1952.