FAST: A New Technique for

Geomechanical Assessment of the

Risk of Reactivation-related Breach of
Fault Seals

ABSTRACT
Postcharge fault reactivation may cause fault seal breach. We present a new
methodology for assessment of the risk of reactivation-related seal breach: fault
analysis seal technology (FAST). The methodology is based on the brittle failure
theory and, unlike other geomechanical methods, recognizes that faults may
show significant cohesive strength. The likelihood of fault reactivation, which is
expressed by the increase in pore pressure (DP) necessary for fault to reactivate,
can be determined given the knowledge of the in-situ stress field, fault rock failure
envelope, pore pressure, and fault geometry. The FAST methodology was
applied to the fault-bound Zema structure in the Otway Basin, South Australia.
Analysis of juxtaposition and fault deformation processes indicated that the fault
was likely to be sealing, but the structure was found to contain a residual
hydrocarbon column. The FAST analysis indicates that segments of the fault are
optimally oriented for reactivation in the in-situ stress field. Microstructural
evidence of open fractures in a fault zone in the subsurface in an offset well and
an SP (self-potential) anomaly associated with a subseismic fault cutting the
regional seal in the Zema-1 well support the interpretation that seal breach is
related to fracturing.

INTRODUCTION
Fault sealing caused by juxtaposition and deformation processes has received
considerable attention,and techniques for the analysis of such, e.g., Allan
diagrams, juxtaposition diagrams, and shale smear algorithms, are widely applied
(Allan, 1989; Knipe, 1997; Bretan et al., 2003). Well-constrained lithological and
juxtaposition data, tied to seismically observable fault zones, can locate
potentially leaking sand-on-sand contacts across faults and predict whether such
sand-onsand contacts are likely to be sealed because of deformation processes
such as cataclasis or shale smearing ( Jev et al., 1993; Hippler, 1997; Fisher and
Knipe, 1998).However, whereas such analyses can define the sealing potential
of faults that have been inactive since hydrocarbon charge, they do not
incorporate the potential for seal breach because of fault reactivation subsequent
to charge.

Abundant evidence shows that faults and fractures provide high-permeability

conduits for fluid Flow during deformation in the brittle crust (e.g., Sibson, 1994;
Barton et al., 1995; Dewhurst et al., 1999). Juxtaposition or deformation process
seals maybe breached if the fault is reactivated subsequent to hydrocarbons
charging the trap (Jones and Hillis, 2003). Seal breach caused by fault
reactivation has been recognized as a critical risk in the Australian context. For
example, in the Timor Sea region, Neogene fault reactivation related to collision
between the Australian and Southeast Asian plates has breached many Jurassic
or older paleotraps (O’Brien andWoods, 1995; Hillis, 1998; Shuster et al., 1998).
The relative likelihood of fault reactivation can be assessed given the knowledge
of the prevailing stress field, fault orientation, pore pressure, and the failure
envelope for the fault rocks. Morris et al. (1996) defined slip tendency based on
the ratio of shear stress to normal stress acting on a fault surface. Ferrill et al.
(1999) defined dilation tendency based on the normal stress acting on a fault
plane normalized to the differential stress. Calculations of slip and dilation
tendency were used to assess the likelihood of fault reactivation at the proposed
high-level radioactive waste repository site at Yucca Mountain, Nevada. Wiprut
and Zoback (2000) determined the increase in pore pressure required to induce
reactivation of a normal fault in the Visund field, northern North Sea, assuming
cohesionless frictional failure. Finkbeiner et al. (2000) assessed the height of
hydrocarbon columns that could be sustained without inducing shear or tensile
failure on trapbounding faults in the Gulf of Mexico. These and related studies of
the relationship between stresses and fault reactivation and permeability (e.g.,
Barton et al., 1995; Hickman et al., 1997) assume that the failure envelopes for
fault rocks are described by a cohesionless friction law of the Byerlee (1978) type.
However, frictional sliding experiments on cohesionless joints or saw-cuts
through rocks of the type summarized by Byerlee (1978) do not describe the
failure envelopes for cemented fault rocks that may exhibit significant cohesion
(Dewhurst and Jones, 2002; Jones et al., 2002). An alternative geomechanical
parameter for assessing the risk of fault reactivation and associated seal breach
is presented herein that can incorporate the cohesive strength of faults and, in a
single parameter, express the risk of shear or tensile failure. The methodology is
applied to the Zema structure in the Otway Basin, South Australia, which contains
a fault-bound residual column. The technique can be readily modified to consider
areas where shear or tensile failure of intact cap rock presents the key
geomechanical risk, either because fault rocks are stronger than intact cap seal
rocks or because preexisting faults are misoriented for reactivation. Prior to
introducing the new approach, this chapter outlines the background concept of
structural permeability and summarizes previously used geomechanical
parameters for assessing fracture-related seal breach.

STRUCTURAL PERMEABILITY
Our approach to the geomechanical risking of reactivation-related fault seal
breach is based on the concept of structural permeability (Sibson, 1996).
Structural permeability is the permeability created by the interaction of various
brittle structures (tensile fractures, shear fractures, and hybrid fractures; Figure
1). Such structures are generally created by the pressure of the infiltrating fluids
and can be represented by failure criteria expressed in terms of pore pressure
(Sibson, 1996) (Table 1). The theory of hybrid fracture generation is relatively
contentious. Hybrid fractures can be considered to be multiple jointing events
instead of tensile fractures with a shear component (Engelder, 1999). However,
the Coulomb–Mohr envelope is still considered to be a valid predictor of failure in
the tensile region as shown by Brace’s experiments (Brace, 1960; Engelder,
1999). The method presented herein is removed from the theory of hybrid fracture
generation and is used to predict brittle failure instead of the form by which it is
manifested.

FIGURE 1. The effective normal (sn 0 ) and shear (t) stresses leading to shear
and tensile fracturing assuming a cohesionless Coulomb failure envelope (t =
mssn 0 ) for shear reactivation of a preexisting fracture and a composite Griffith
(t2 4Tsn 0 4T2 = 0)– Coulomb (t = C + misn 0 ) failure envelope for intact rock.
The upper diagrams are schematic illustrations of the orientations of tensile and
shear fractures in a rock sample. Shear fracturing occurs where differential stress
(s1 s3, the diameter of Mohr circle) is relatively large compared to the tensile or
cohesive strengths and tensile fracturing at relatively lower differential stress.
Symbols are as in Table 1; us is the static friction coefficient along an existing
plane of weakness; and mi is the internal coefficient of rock friction for intact rock.
Fracture-related seal failure has widely been considered to occur solely because
of tensile fracturing (also termed natural hydraulic fracturing), whereby increasing
pore fluid pressure reduces the minimum effective stress to below the tensile
strength of the rock (e.g., Palciauskas and Domenico, 1980; Ozkaya, 1984; Bell,
1990; Engelder and Lacazette, 1990; Miller, 1995). Hence, Watts’ (1987)
explanation of seals that fail by fracturing is hydraulic seals. Such tensile or
natural hydraulic fracturing has, for example, been invoked to consider cap rock
leakage in the North Sea (Caillet, 1993; Gaarenstroom et al., 1993; Caillet et al.,
1997; Grauls, 1997). As discussed below, tensile fracturing can only occur with
increasing pore-fluid pressure if differential stress is relatively low (Figure 1; Table
1). Barton et al. (1995) combined in-situ stress measurements with information
on the orientations of hydraulically conductive fractures and faults in three Wells
in the southeastern United States and demonstrated that fractures and faults
optimally oriented for shear reactivation are the most important permeability
conduits. Wiprut and Zoback (2002) analyzed four fields in the northern North
Sea, concluding that faults that are critically stressed in the current stress field
(i.e., capable of slipping) tend to leak, whereas those that are not critically
stressed are more likely to be sealing. Given the evidence suggesting that both
tensile and shear fractures rupture to cause seal breaching, it is critical that any
methodology for assessing the risk of seal breach because of reactivation
incorporate the influence of both these elements of structural permeability.
Following Sibson (1996), we assume a composite Griffith–Coulomb failure
envelope (Figure 1). Hence, tensile failure is predicted where differential stress
is relatively low (Sibson, 1996). If 4T < (s1 s3) < 6T, hybrid tensile-shear failure
is predicted, and if (s1 s3) > 6T, shear failure is predicted. If the cohesive strength
of a reactivated fault zone is zero, the failure envelope passes through the origin
of the normal and shear stress plot, and reactivation in shear is the only posible
mode of failure, irrespective of the differential stress.
Following experimental data by Handin (1969), Sibson (1996, 1998) recognized
the influence of shear reactivation and noted that tensile fractures can only form
and provide conduits for fluid flow, where

rocks are intact and devoid of faults

existing faults become severely misoriented for

shear reactivation

existing, favorably oriented, faults have regained cohesive strength because of

cementation

We add to this last point, ‘‘or where the faulting process has resulted in a zone
with significant cohesive strength.’’ Faulting may involve processes such as
cataclasis with quartz cementation that result in significant cohesion and,
therefore, tensile strength (as seen, for example, in the case study presented
below from the Otway Basin).
TABLE 1. Brittle failure criteria expressed in terms of pore pressure (P) and
necessary differential stress conditions. Criteria assume a Composite Griffith
Coulomb failure criterion, whereby C (cohesion) 2T
(tensile strength). The criteria apply to intact rock or reactivation of preexisting
fractures, provided that the appropriate values of T, C, and m (coefficient of
friction) are used. After Sibson (1996).

GEOMECHANICAL PARAMETERS FOR ASSESSING THE RISK OF

REACTIVATION RELATED SEAL BREACH

The first introduced geomechanical parameters for assessing the risk

of fracture-related seal breach assumed that such was caused by
tensile failure of the cap rock (Watts, 1987; Caillet, 1993;
Gaarenstroom et al., 1993), and thus, they were based on the criterion
for tensile failure (Table 1). For example, Gaarenstroom et al. (1993)
introduced the concept of retention capacity, which is given by the
difference between the minimum horizontal stress and pore pressure
(Figure 2). Retention capacity is thus the effective minimum horizontal
stress (sh 0 ). A positive retention capacity (or sh 0 ) reflects the
additional pore pressure (or hydrocarbon column height) that can be
developed prior to tensile failure by natural hydraulic fracturing. If
retention capacity is zero, then tensile fractures would develop if the
rock had no tensile strength. Retention capacity only considers the
risk of tensile (and not shear) failure and does not incorporate (tensile)
rock strength. In recognition of the fact that critically stressed shear
fractures present conduits for fluid flow, Morris et al. (1996) introduced
the concept of slip tendency, which is the ratio of shear stress to
effective normal stress acting on a fault and expresses the likelihood
of slip on a cohesionless fault (Figure 2). To assess risk caused by
both shear and tensile fractures, Ferrill et al. (1999) used slip
tendency and dilation tendency. The latter risks the likelihood of
dilation (tensile reactivation) of a fault on a linear scale from zero (if
s1 is normal to the fault) to one (if s3 is normal to the fault; Figure 2).
Slip and dilation tendency can be used together to geomechanically
risk the likelihood of fault seal breach caused by shear and tensile
fracturing. However, two separate parameters, neither of which
incorporates rock strength, must be assessed.
The Coulomb failure function was used by Castillo et al. (2000) to risk
fault seal breach caused by reactivation in the Australian Timor Sea.
The Coulomb failure function is the difference between the shear
stress acting on a fault and that required to cause failure on a
cohesionless fault (Figure 2). A negative Coulomb failure function
thus implied a stable fault, whereas a positive Coulomb failure
function was associated with low fault seal integrity. Wiprut and
Zoback (2002) used the critical pressure perturbation to risk fault seal
breach caused by reactivation in the northern North Sea. The critical
pressure perturbation is the increase in pore pressure required to
reduce the effective normal stress to the value that would cause slip
on a cohesionless fault (Figure 2). The Coulomb failure function and
the critical pressure perturbation both incorporate the coefficient of
sliding friction on a fault in assessing its risk of shear reactivation.
However, they do not allow for any cohesive strength on preexisting
faults nor for the development of tensile fractures.

FAST: A NEW GEOMECHANICAL PARAMETER FOR ASSESSING

THE RISK OF REACTIVATION-RELATED SEAL BEACH

We propose a new geomechanical parameter for assessing the risk

of reactivation-related seal breach, the fault analysis seal
technology(FAST),which allows for the input of a failure envelope with
cohesion. Jones et al. (2002) and Dewhurst and Jones (2002) have
demonstrated that fault rocks may show significant postdeformation
lithification caused by cementation that results in the regaining of
cohesive and, therefore, tensile strength. Thus, knowledge of the fault
rock failure envelope should be incorporated into predictions of fault
reactivation. The likelihood of fault reactivation and associated seal
breach can be assessed by the FAST method given the knowledge
of the stress field, pore pressure, fault orientation, and the failure
envelope for the fault rocks.

The in-situ stress field can be determined by a variety of wellbore

geomechanical techniques. Density and check-shot velocity data
yield the vertical stress, borehole breakouts and drilling-induced
tensile fractures yield the orientation of the horizontal stresses, leak-
off and extended leak-off tests yield the minimum horizontal stress,
and the maximum horizontal stress can be determined by the
occurrence or nonoccurrence of breakouts and drilling-induced
tensile fractures and knowledge of rock strength (see, e.g., Bell, 1996;
Moos and Zoback, 1990, for detailed discussion of these techniques).
Knowledge of the fault failure envelope can be determined from
laboratory testing of intact fault rocks (Handin and Jaeger, 1957;
Handin, 1969; Jaeger and Cook, 1976; Dewhurst and Jones, 2002).
To compensate for fault plane heterogeneity or when no strength data
are available, sensitivity analysis of fault failure envelopes should also
be undertaken. Fault orientation (dip and strike) is determined from
depth-converted seismic interpretation. For three-dimensional (3-D)
seismic data, the risk of reactivation can be mapped onto the
interpreted fault plane geometry in 3-D. For two-dimensional (2-D)
seismic data, the dip and dip azimuth of the fault are determined from
the offset between reflector terminations. A centerline point in the
mapped fault polygon can then be assigned the fault dip and strike
and, in turn, the risk of reactivation. Given the requisite information,
three stages to assessing reactivation risk thus exist.

1) A 3-D Mohr diagram representing the state-of-stress and failure

envelope for the fault is constructed (Figure 3). The risk of reactivation
of a plane of any orientation is expressed by the increase in pore
pressure (DP) required to cause its reactivation, i.e., horizontal
distance on a 3-D Mohr diagram between a fault plane and the failure
envelope. 2) The reactivation risk (DP) for all planes is plotted on a
polar diagram of normals to planes (Figure 3). 3) The appropriate
reactivation risk (DP) is mapped either onto the fault plane (3-Ddata;
Figure 4)or fault polygon centerline points (2-D data).

The Zema-1 well in the Otway Basin, South Australia, intersected a

69-m (226-ft) paleogas leg and a 15-m (49-ft) paleo-oil leg in the
Lower Cretaceous Pretty Hill Formation (Lyon et al., 2005). The Zema
structure has fault-dependent closure in the footwall of the Zema fault.
The strike of the Zema fault varies between east– west and
northwest–southeast, and it dips approximately 708 to the north (Lyon
et al., 2005). Detailed analysis suggested that juxtaposition and fault
deformation processes together were likely to provide an adequate
seal for the observed paleocolumn (Jones et al., 2000; Lyon et al.,
2005).Hence, the existing methodologies for assessing fault seal for
inactive faults did not consider the observed paleocolumn, and the
propensity for reactivation-related breach of the fault seal at Zema-1
was investigated.
FIGURE 2. Geomechanical risking parameters. See text for full
discussion and references. Retention capacity risks the likelihood of
tensile failure of intact cap rock. Other techniques risk the reactivation
of a preexisting fault, the orientation of which is specified by the dot in
the 3-D Mohr circles. The position of the fault in the gray-shaded área
in the 3-D Mohr circles is defined by the relative orientation of the fault
and the principal stresses. CFF = Coulomb failure function.
FIGURE 3. The FAST methodology.(A) Three-dimensionalMohr circle
for Otway Basinstate-of-stress and failure envelopefor laboratory-
testedcataclasite. (B) Polar diagramof normals to planes coloredby
increase in pore pressure(DP) required for reactivation.The crosses
represent the orientationsof individua lelementsof the Zema fault.
The in-situ stress tensor in the area was constrained, as described in
more detail by Jones et al. (2000), using density and check-shot log
data (for vertical stress), leakoff tests and one extended leak-off test
(minimumhorizontal stress), and the occurrence of drilling-induced
tensile fractures (máximum horizontal stress).Maximum horizontal
stress orientation of 1568N was inferred from breakouts in the nearby
Katnook- 3 well. Pore pressures are hydrostatic. In the Depth range
of interest of 2500– 3000m(8200–10,000 ft), the following stress
gradients apply:
Minimum horizontal stress (sh) = 16.1 MPa/km Overburden stress
(sv) = 22.4 MPa/km
Maximum horizontal stress (sH) = 28.7 MPa/km
Pore pressure (P) = 9.8 MPa/km
Maximum horizontal stress (sH) orientation = 1568N

Unusually, core is available through a fault zone in the reservoir

(Pretty Hill Formation) in the offset Banyula-1 well. The core
intersected cataclasites that were tested in a standard triaxial cell with
deformation features oriented at 308 to s1. The failure envelope
derived from testing of the cataclasites, as described in more detail
by Dewhurst and Jones (2002) is
t = 5:40 + 0.78 o’n

where t is the shear stress at failure, sn 0 is the effective normal stress

(i.e., sn P), and pressures are in megapascals. We have used the
failure envelope for this core in our analysis, although we recognize
that the core is through the reservoir and not the seal, which may have
a different rheology (see discussions in Boult et al., 2003; Dewhurst
and Jones, 2003). Such potential variation in fault rock failure
envelopes is a key driver for the sensitivity analysis (in very weak and
strong fault rocks) described in the following section. The entire region
is covered by 2-D seismic data and a 3-D survey covers
approximately half of the Zema structure. The geometry of the fault
plane was determined by depth conversion and subsequent
interpretation of the 2-D and 3-D seismic data over the Zema
structure. Applying the FAST methodology to the above data
demonstrates that significant elements of the Zema fault are optimally
oriented for strike-slip reactivation in the in-situ stress field (Figure 3).
The most at-risk segments are those striking approximately
northwest–southeast (Figure 4). Given that juxtaposition and fault
deformation processes are likely to have created an adequate seal,
and that significant sections of the trap-bounding fault are optimally
oriented for reactivation, we interpret that fault reactivation is a likely
candidate for seal breach and, thus, the presence of the paleocolumn
in the Zema-1 well.
Two pieces of additional evidence are consistent with the
interpretation that fault reactivation is responsable for seal breach in
the Zema-1 well. First, microstructural analysis of cataclasite samples
from the fault zone in the offset Banyula-1 well indicates the presence
of open fractures (Jones et al., 2000). Illites span the fractures that
must be open in the subsurface. Second, a minor fault (not apparent
on seismic data) is interpreted in the Laira Formation regional seal
from dipmeter data in the Zema-1 well. This minor fault is associated
with a significant SP anomaly, indicating that fault zones do indeed
provide a permeable zone in the regional seal (Figure 5). We are
unable to determine whether this subseismic fault intersects the Zema
reservoir. The main Zema fault is interpreted to cut the Zema-1 well
near the base of the overlying Eumeralla Formation, and its SP effect
is harder to ascertain because of the more frequent permeable sands
in the Eumeralla Formation.

SENSITIVITY ANALYSIS
A sensitivity study of the results was also undertaken despite a failure
envelope being available for fault rocks in the Otway Basin. This is, in
part, to illustrate the application of sensitivity analysis to areas where
failure envelopes are not available. Furthermore, the cataclastic zone
tested in the Otway Basin is from an offset well and in the reservoir
and not the cap rock. The failure envelope may change significantly
along the fault plane because of changes in the rock types through
which the fault cuts or because of laterally variable fluid Flow and
diagenetic processes along the fault. Hence, the robustness of the
results was consideredwith respect to very weak and very strong fault
rock failure envelopes.
The maximum horizontal stress is generally the least well-constrained
component of the in-situ stress field. Hence, the sensitivity of the
results to a range in the magnitude of maximum horizontal stress was
also assessed. The most likely stress regime of the area (above) is a
strike-slip regime, whereby sH > sv > sh. The robustness of the results
were tested for a much lower maximum horizontal stress value equal
to the vertical stress, such that the stress regime would be transitional
between strike-slip and normal (sH sv > sh)
and also for the maximum value that maximum horizontal stress could
attain, i.e., the frictional limit (Zoback and Healy, 1984) in strong rocks
(coefficient of friction of 0.8).
The following end-member caseswere considered to assess the
robustness of the assessment of reactivation risk:

weak fault rocks (t = 0.3sn 0 ) and most likely stress regime above

strong fault rocks (t = 20 + sn 0 ) and most likely stress regime

above

lower sH (22.4 MPa/km, transitional strike-slip or normal fault stress

regime) and failure envelope as determined from lab testing

higher sH (37.1 MPa/km, frictional limit for coefficient of friction of

0.8) and failure envelope as determined from lab testing

Differences in the risk of reactivation in each of the above four

scenarios exist (Figure 6). However, major similarities are present in
the most at-risk fault orientations that allow important generalizations
to be made. A conservative approach to the sensitivity analysis is to
avoid fault planes that are prone to reactivation in any single scenario
and to focus on planes not prone to reactivation in any of the
scenarios. Planes suitably oriented for reactivation in any single
scenario strike between 100 and 2108N and dip greater than 408
(Figure 6). Planes not suitably oriented for reactivation in any of the
scenarios strike between 40 and 908N and dip greater than 608
(Figure 6).

The differences between the reactivation risk plots for each scenario
essentially reflect the different mode of brittle reactivation that
predominates in each case. Shear failure tends to predominate in the
high máximum horizontal stress (hence, high differential stress) case
and the weak fault rock case, because (s1 s3) > 6T. Hence, planes
oriented approximately 308 to s1 and containing the s2 direction are
prone to reactivation. Tensile failure is more significant in the low
maximum horizontal stress (hence, low differential stress) case and
the strong fault rock case, because (s1 s3) < 4T. Hence, planes
orthogonal to s3 are prone to reactivation.
These differences also highlight the advantage of the FAST
methodology over the Coulomb-based risking algorithms: critical
pressure perturbation (CPP) and Coulomb failure function
(CFF).Where differential stress is high or fault rock strength is weak,
the FAST method will produce almost identical results to the CPP
method. A small differential stress or strong failure envelope
increases the risk of fault orientations critically oriented for tensile
failure relative to those for shear failure. Therefore, CPP and CFF may
overestimate the risk associated with shear failure and ignore the risk
associated with fault orientations critical for tensile failure. Significant
differences exist between the DP values in the four scenarios. Major
increases in pore pressure are required to reactivate even optimally
oriented faults in the lower bound sH and strong fault cases.
Significant portions of the fault are at stresses beyond failure in the
weak fault case. Although extreme scenarios have been used, these
variations illustrate that the errors in this technique preclude it being
used for predicting hydrocarbon column heights. A 1-MPa variation in
DP is equivalent to the buoyancy pressure associated with 1 km (0.6
mi) of oil column, assuming the hydrostatic gradient is 9.8MPa/km
( 1 g cm3) and oil gradient 8.8 MPa/km
( 0.9 g cm3). As shown by Figure 6, the
potential errors in the methodology are greater than 1 MPa. This
applies to any geomechanical methodology because of the errors
inherent in estimating in-situ stresses and rock failure parameters. We
believe that the value of the technique lies in its application to areas
where reactivation-related breach is suspected to be an issue and in
the relative risking of fault-bound prospects in such an area. The DP
values need to be calibrated with reference to the occurrence of intact
and breached columns in specific basins. For example, in the Timor
Sea, DP values less than 10 MPa represent a significant risk of
reactivation-related seal breach (Mildren et al., 2002).
FIGURE 6. Sensitivity analysis of the likelihood of reactivation in the
Otway Basin. (A) Weak fault rocks; (B) strong fault rocks; (C) lower
limit sH; and (D) upper limit sH. Plots are polar diagrams of normals
to planes colored by DP values.

DISCUSSION
The considerable evidence that fault and fracture reactivation leads
to fluid redistribution has been summarized by Sibson (1992, 1994),
Muir-Wood (1994), and Dewhurst et al. (1999). Sibson’s (1992) fault-
valve model provides a mechanism whereby the cycling of tectonic
shear stress and/or fluid pressure is linked to episodic fault instability
and, in turn, episodic fluid redistribution. Cementation or hydrothermal
precipitation may lead to faults resealing during periods of stability.
The observed episodic discharge of hydrocarbons from
overpressured compartments (e.g.,Hunt, 1990) is consistent with this
model. At the geological timescale, hydrocarbons clearly leak
episodically and not continually up trap-bounding faults, because if
leakage was continual at rates greater than charge, accumulations
would not develop. In the Australian context, the presence of
paleocolumns witnesses the fact that trapbounding faults do seal over
significant periods and, thus, thatmost trap-bounding faults are not
permanently open conduits for leakage (O’Brien andWoods, 1995).
The model followed herein, like Sibson’s (1992) faultvalvemodel,
assumes that the observed episodic breaching of faults and
associated fluid redistribution is associated with reactivation.
In some cases, fault rocks may be stronger than the surrounding
rocks (Dewhurst and Jones, 2002; Jones et al., 2002), and hence, the
risk of failure of intact cap rock must be considered as well as the risk
of fault reactivation (Boult et al., 2002). Indeed, the risk of failure of
intact cap rock may be greater than the risk of reactivating misoriented
faults where the intact cap rock is weaker (Streit, 1999; Hillis and
Nelson, 2005). To assess the risk of cap rock failure, it is simply
necessary to add an intact rock failure envelope to the analysis. The
risk of failure of intact cap rock is the increase in pore pressure that
can be sustained prior to failure by the point on the Mohr circle closest
to the failure envelope. In the case of cap rock analysis, a single value
is provided for the (crest of the) prospect (as is the case with retention
capacity; Figure 2), whereas the risk of fault reactivation varies with
the geometry of the fault. The strengthof fault rockmaterial obtained
at Banyula-1 suggests that sand-sand fault contacts are strong in the
Otway Basin. The sensitivity analysis reveals the risk of fracturing to
be very high when using a weak failure envelope. It is possible that
the strength of the cap rock is weaker than the fault, and the
generation of cap rock fractures at Zema is also a possibility. It is
necessary to consider the location of at-risk fault segments (low-DP
zones) with respect to the trap as a whole. If segments of the fault
with low DP are not coincident with the top of the structure, then
reactivation may not lead to breaching of the entire column.
Reactivation must postdate hydrocarbon charge for it to cause seal
breaching. If postcharge reactivation occurred in the geological past
and within a paleostress regime that differed significantly from that of
the present day, it should not be risked with reference to the insitu
stress field. The risk of reactivation may be assessed with reference
to the in-situ stress field if leakage is associated with present-day
geochemical anomalies such as those witnessed by marine
geochemical sniffers and airborne laser fluorescence (Bishop and
O’Brien, 1998) or if it is associated with faults that cut a Young seabed
and land surface. In the North Sea, the significance of the in-situ
stress field to the assessment of fault leakage is witnessed by Heffer
and Fox’s (1996) compilation indicating that nonsealing faults are
strongly preferentially oriented in the in-situ maximum horizontal
stress direction. Although in many cases, the in-situ stress field is
appropriate, if leakage occurred in a paleostress field different to that
of the present day, then that paleostress field must be applied to
consider the risk of reactivation (Gartrell and Lisk, 2005).
All geomechanical methodologies for risking reactivation-related seal
breach assign risk to seismically mapped faults. Hence, it is an implicit
assumption that leakage associated with reactivation occurs on the
seismically mapped structures or on structures with similar
orientations. Structures that are not seismically imaged may be prone
to reactivation and contribute to seal breach. Indeed, the subseismic
fault cutting the Laira Formation regional seal in the Zema example is
permeable. Hence, at the prospect scale, it is an assumption that
seal-breaching fractures are parallel or subparallel to the mapped
fault.

CONCLUSIONS
Analysis of cross-fault lithological juxtaposition and of deformation
processes are routinely used to assess the likelihood of fault seal.
However, these techniques cannot incorporate the potential for seal
breach caused by fault reactivation subsequent to charge.Faults that
are suitably oriented to be reactivated in the prevailing stress field
provide conduits for fluid flow. The concepts of brittle failure and
structural permeability provide the basis of a technique to assess
whether seismically mapped faults are likely to be reactivated in the
in-situ stress field and thus associated with seal breaching.
Faults may show significant cohesive strength caused by
postdeformation cementation. The existing geomechanical
parameters used to assess the likelihood of reactivation-related seal
breach assume that preexisting faults have no cohesive strength.
Hence, we have introduced a new methodology (FAST) that allows
input of a laboratory-derived fault rock failure envelope or a range of
likely fault rock failure envelopes. The likelihood of reactivation is
expressed by the increase in pore pressure (DP) required to cause
the fault to reactivate given the orientation in the in-situ stress field.
This translates to an advantage over the Coulombderived
geomechanical methodologies because it incorporates tensile failure
in scenarioswhere either fault rock material is very strong or the
differential stress is small.
The technique can be readily modified to consider areaswhere failure
of intact cap rock presents the key geomechanical risk. Sensitivity
studies incorporating variable fault strength parameters are required
until fault strength can be mapped in detail across fault planes,
possibly by relating strength to damage processes using gouge
estimates. Once this has been achieved, the identification of across-
fault leakage vs. along-fault leakage can be determined.
The FAST methodology was applied to the faultbound Zema structure
in the Otway Basin, South Australia. Juxtaposition and fault
deformation processes indicated that the fault was likely to be sealing,
but the structure was found to contain a residual hydrocarbon column.
The FAST analysis indicates that segments of the fault are optimally
oriented for reactivation in the in-situ stress field.Microstructural
evidence of open fractures in a fault zone inthe subsurface in an offset
well and an SP anomaly associated with a subseismic fault cutting the
regional seal in the Zema-1 well support the interpretation that seal
breach is related to fracturing.
The FAST methodology provides a powerful tool for predrill
assessment of the risk of fault reactivationrelated seal breach in all
stress scenarios and fault strengths. However, limitations to such
geomechanical techniques exist. First, they assume that reactivation
occurs in the in-situ stress field as can be determined from wellbore
data. Second, predrill risk can only be assigned to seismicallymapped
faults (and not subseismic faults).
Finally, the errors associated with the techniques preclude
themfrombeing used to assess likely hydrocarbon column heights.
Nonetheless, these techniques provide a powerful method for the
relative ranking of faultbound prospects in an area where reactivation-
related breach presents an exploration risk.