AUTHORS
Richard M. Jones Woodside Energy,1 Adelaide Terrace, Perth, Western Australia,6000, Australia;[email protected]
Richard joined Woodside in September 2000and is currently working in the Trap and Pres-sure Team, New Ventures. He graduated witha joint B.Sc. degree (with honors) in geologyand economics from Keele University (1992)and a Ph.D. from Keele University (1996). Hehas worked extensively in the area of fault andtop seal evaluation and has been involved withseals research programs in Europe, the UnitedStates, and Australia. Current interests includestructural modeling, seal evaluation, wellborestability, and Liverpool FC. Richard is a memberof AAPG, PESA, and PESGB.
Richard R. Hillis National Centre forPetroleum Geology and Geophysics, AustralianPetroleum Cooperative Research Centre, Uni-versity of Adelaide, South Australia 5005,Australia; [email protected]
Richard holds the State of South Australia Chairin Petroleum Reservoir Properties/Petrophysicsat the National Centre for Petroleum Geologyand Geophysics (NCPGG), Adelaide University.He graduated with a B.Sc. degree (with honors)from Imperial College (London, 1985), anda Ph.D. from the University of Edinburgh(1989). After seven years at Adelaide Uni-versitys Department of Geology and Geophys-ics, Richard joined the NCPGG in 1999. Hismain research interests are in petroleum geo-mechanics and sedimentary basin tectonics.He has published approximately 50 papers andhas consulted to many Australian and inter-national oil companies in these topics. Richardis a member of AAPG, American GeophysicalUnion, Australian Society of Exploration Geo-physicists, European Association of Geoscien-tists and Engineers, Geological Society ofAmerica, Geological Society (London), Petro-leum Exploration Society of Australia, Societyof Exploration Geophysicists, and Society ofPetroleum Engineers.
An integrated, quantitativeapproach to assessingfault-seal riskRichard M. Jones and Richard R. Hillis
ABSTRACT
Fault sealing is one of the key factors controlling hydrocarbon accu-
mulations and trap volumetrics and can be a significant influence on
reservoir performance during production. Fault seal is, therefore, a
major exploration and production uncertainty. We introduce a sys-
tematic framework in which the geologic risk of faults trapping hy-
drocarbons may be assessed.
A fault may seal if deformation processes have created a mem-
brane seal, or if it juxtaposes sealing rocks against reservoir rocks,
and the fault has not been reactivated subsequent to hydrocarbons
charging the trap. It follows from this statement that the integrated
probability of fault seal can be expressed as {1 [(1 a)(1 b)]}(1 c), where a, b, and c are the probabilities of deformationprocess sealing, juxtaposition sealing, and of the fault being reacti-
vated subsequent to charge, respectively. This relationship provides
an assessment of fault-seal risk that integrates results from the crit-
ical parameters of fault-seal analysis that can be incorporated into
standard exploration procedures for estimating the probability for
geologic success. The integrated probability of fault seal for each
prospect can be visualized using the fault-seal risk web, which allows
rapid comparison of success and failure cases through construction of
prospect risk web profiles.
The impact of uncertainty (U ) and the value of information(VOI) for each aspect of fault sealing on the overall fault-seal risk
may be determined via the construction of data webs and the rela-
tion U = [1 {(P
nw) / n}] 100, where nw is the value givento each data web parameter and n is the number of data web com-ponents. For example, the data web components required to assess
fault reactivation risk are the orientation and magnitude of the in-situ
principal stresses, pore pressure, fault architecture, and the geo-
mechanical properties of the fault.
Risking of the Apollo prospect, Dampier subbasin, North West
shelf, Australia is presented as a worked example. Fault-seal risking
for the Apollo prospect has been conducted on 10- and 100-ft oil
Copyright #2003. The American Association of Petroleum Geologists. All rights reserved.
Manuscript received July 31, 2002; provisional acceptance August 27, 2002; revised manuscript receivedOctober 8, 2002; final acceptance October 10, 2002.
DOI:10.1306/10100201135
AAPG Bulletin, v. 87, no. 3 (March 2003), pp. 507524 507
508 An Integrated, Quantitative Approach to Assessing Fault-Seal Risk
columns to allow integration with volumetric probabilistic state-
ments. The critical parameter for fault-seal risking at the Apollo
prospect is the ability of disaggregation zone faults (low shale gouge
ratio fault gouge) to support increasingly large hydrocarbon col-
umns. Evaluation of the individual components for Apollo fault
sealing indicates a = 0.3 (10-ft column) and a = 0.1 (100-ft column),b = 0.2, and c = 0.1. The overall probability of the Apollo trap-bounding fault sealing a 10-ft oil column is 0.4 or 40% (seal con-
dition moderately unlikely). The likelihood that the fault seals oil
columns greater than 100 ft is 0.3 (seal condition unlikely). Data web
error margins for the Apollo prospect are 20% (juxtaposition un-
certainty), 26% (fault-rock process uncertainty), and 27% (fault re-
activation uncertainty). Recalculating each parameter by its un-
certainty, for a 10-ft oil column, the upper value of integrated fault-
seal risk is 0.5 (seal condition intermediate), and the lower value is
0.3 (seal condition unlikely). The upper value of integrated fault-
seal risk for a 100-ft oil column is 0.3 (seal condition unlikely), and
the lower value is 0.2 (seal condition very unlikely). The variation in
the Apollo final risk calculation reflects the lack of prospect-specific
data. The greatest VOI benefit for Apollo fault-seal prospectivity is
sedimentary architecture data.
INTRODUCTION
Hydrocarbon exploration and production strategies all involve an
element of risk. As with any investment strategy, the goal of the
venture capitalist is to minimize this risk. Geologic risk minimiza-
tion should begin with a focused evaluation of the chance of success.
That is determining the likelihood that all elements of the pe-
troleum system required for economically viable volumes of hy-
drocarbons to be developed and trapped have been satisfied. In
an economic environment where prospects are now considered
and drilled in a global context, it is essential that all prospects be
ranked via comparable risking criteria. The presence of a sealed
trap is one of the key factors in the evaluation of geologic risk for
exploration prospects (Rose, 1992; Otis and Schneidermann, 1997;
Watson, 1998). This paper presents a framework for quantifying
the risk associated with the development of an intact seal for fault-
bound prospects. This procedure
enables geologic fault-seal risk to be incorporated into standardrisking procedures,
allows data from different scales and pertaining to all fault-sealfailure mechanisms to be integrated,
facilitates rapid comparison of known sealing/leaking fault boundtraps to identify key field or basin-scale trap integrity controls, and
allows value of information (VOI) for critical data to be assessed.
Previous techniques for risking the likelihood of fault sealing
have tended to focus on one particular aspect of sealing such as
ACKNOWLEDGEMENTS
The authors are extremely grateful to colleaguesat the National Centre for Petroleum Geologyand Geophysics, Adelaide, Woodside Energy,and Shell for manuscript discussions. RussellDavies, Rob Knipe, Gavin Lewis, Frank Krieger,and James MacKay are thanked for providingconstructive and focused reviews that improvedearlier versions of the manuscript. The fault-sealrisk web as presented herein has evolved froma risk-web concept presented by former col-leagues of the first author at Rock DeformationResearch, University of Leeds. The integratedfault-seal risking procedure detailed in this paperforms part of the proprietary APCRC SealsConsortium. The consortium members Wood-side Energy, BHP Billiton, JNOC, Origin Energy,ChevronTexaco, Exxon-Mobil, Globex Energy,Santos, Anadarko Petroleum Corporation, andOMV are thanked for their permission topublish.
cross-fault reservoir-seal juxtaposition or membrane
seals created by deformation, shale gouge ratio (SGR)
or fault reactivation. There is no doubt that techniques
such as SGR mapping (Yielding et al., 1997) have proved
critical in reducing exploration and production fault-seal
uncertainty. However, to fully risk fault sealing, an as-
sessment of all possible mechanisms of seal failure (Fig-
ure 1) should be undertaken and the results integrated
into an overall fault-seal risk.
This paper presents a methodology for combining
the risks associated with juxtaposition seals, fault-rock
membrane seals, and fault reactivation into a single in-
tegrated, overall fault-seal risk. The fault-seal risk web
as presented herein has evolved from a risk web concept
presented by Knipe et al. (1995). Furthermore, because
reactivation is a critical risk in the Australian context,
and because assessing the risk of fault-seal breach caused
by reactivation has received less attention than the other
mechanisms, we summarize our methodology for assess-
ing the risk of reactivation-related fault-seal breach.
The risks associated with juxtaposition sealing, fault-
rock membrane sealing, and fault reactivation are each
assessed for the Apollo prospect in the Australian North
West shelf. This case study is used to demonstrate how
these risks are combined into an integrated, overall
fault-seal risk that considers uncertainty and assists in
determining data value of information.
MECHANISMS OF FAULT SEALING
Faults may seal if they juxtapose reservoir rocks against
sealing rocks (Allan, 1989; Freeman et al., 1990; Yielding
et al., 1997) or if the faulting process has generated a
membrane seal, for example, by cataclasis (Antonellini
and Aydin, 1994), cementation/diagenesis, framework
grain-claymixing (Knipe,1992), or clay smearing (Bouvier
et al., 1989; Gibson, 1994). It is not necessary for both
juxtaposition and deformation process seals to be de-
veloped in order for a fault to be sealing. If throw on the
fault juxtaposes sealing rocks against reservoir rocks, no
deformation process seal is required. Conversely, faults
can seal where there is sand-sand juxtaposition and cata-
clastic processes have reduced framework grain pore-
throat apertures such that the fault zone itself acts
as a membrane seal. For further details on cross-fault
reservoir-seal juxtaposition seals or membrane seals
created by deformation, the reader is referred to other
papers in this issue and the references herein.
Abundant evidence that active faults and fractures
provide high-permeability conduits for fluid flow dur-
ing deformation exists (e.g., Sibson, 1994; Barton et al.,
1995). Juxtaposition or deformation process seals may
be breached if the fault is reactivated subsequent to
hydrocarbons charging the trap. 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 reactivation related to collision between
the Australian and Southeast Asian plates has breached
many Jurassic or older paleotraps (Shuster et al., 1998).
Jones et al. (2000) has also presented microstructural
evidence for in-situ related fault reactivation in the
Otway basin, South Australia. In considering a pop-
ulation of faults and fractures, those that are critically
stressed (subject to a state of stress conducive to re-
activation) are prone to act as conduits for fluid flow
(Barton et al., 1995).
Assessing the Risk of Reactivation-Related Fault-SealBreach
Fault sealing caused by juxtaposition and deformation
processes has received considerable attention, and tech-
niques for the analysis of such (e.g., Allan diagrams, jux-
taposition diagrams and shale smear algorithms) are
widely applied. However, the risk of seal breach caused
by reactivation, although recognized (e.g., Nybakken,
1991; Grauls and Baleix, 1994), has received some-
what less attention. Our methodology for assessing the
risk of seal breach caused by fault reactivation was de-
veloped in response to the recognition that reactivation
is a critical risk in the Australian context.
We combine knowledge of the in-situ stress field
with that of fault geometry to assess the likelihood of
reactivation of mapped faults and associated seal breach
in the in-situ stress field. Other techniques have sim-
ilarly used information on in-situ stress and fault geom-
etry to investigate dynamic fault-sealing properties
(e.g., Morris et al., 1996; Ferrill et al., 1999; Wiprut and
Zoback, 2000; Finkbeiner et al., 2001). However, all
of these techniques assume cohesionless frictional fail-
ure of the fault rock. Dewhurst and Jones (2002) have
demonstrated that preexisting faults may not be cohe-
sionless. Thus, knowledge of the fault-rock failure en-
velope should be incorporated into predictions of fault
reactivation.
In-situ stress is determined from wellbore geome-
chanical data such as borehole breakouts as described,
for example, by Bell (1996). Fault orientations (dip and
dip direction) are determined from seismic interpreta-
tion, based on the offset between reflector terminations
Jones and Hillis 509
at a fault. Knowledge of the fault-rock failure envelope
can be determined from laboratory testing of intact fault
rocks (see Dewhurst and Jones, 2002). Given the above
information, there are three critical stages to assessing re-
activation risk.
1. A three-dimensional (3-D) Mohr diagram represent-
ing the state-of-stress and failure envelope for the
fault is constructed. The risk of reactivation of a plane
of any orientation can be expressed by the increase in
pore pressure (P) required to reduce the effectivestresses such that failure occurs on that plane, i.e.,
horizontal distance on a 3-D Mohr diagram between
a plane and the failure envelope.
2. The reactivation risk (P) is plotted on a polar plotof normals to all planes.
3. Fault geometry is transposed onto the reactivation
risk plot and the reactivation risk is determined for
the fault. The P value should be determined atvarious points along the fault, especially where signif-
icant changes in orientation and/or dip are observed
as risk of seal breach can vary significantly along a single
plane.
In addition to permitting the use of realistic fault-
rock failure envelopes derived from laboratory tests, this
methodology allows the likelihood of reactivation by
all modes of failure to be assessed in a single calculation,
as opposed to the separate slip and dilation tendency
analyses of Ferrill et al. (1999). The P technique issummarized in Figure 2 and described in greater detail
in Mildren et al. (2002).
QUANTIFYING THE RISK OF FAULT-SEALBREACH
The fundamental tenet of our approach follows from
the above discussion in that a fault may seal by mem-
brane seal if deformation processes have created an ef-
fective seal, or if displacement juxtaposes sealing rocks
against reservoir rocks, and the fault has not been re-
activated subsequent to hydrocarbons charging the
trap.
Expressed alternatively, a fault is sealing if both
cross- and up-fault flow are inhibited. Juxtaposition or
deformation process seals inhibit the former, and the
latter occurs if the fault is reactivated. This approach
forms the basis of BHP Petroleums logic tree for as-
sessing fault seal (Watson, 1998); however, the above
has not been previously converted into a quantitative
relationship describing integrated fault-seal risk.
It can be demonstrated that the integrated prob-
ability of a fault sealing (FS) can be expressed as
FS f1 1 a1 bg 1 c 1
where a, b, and c are the probabilities of deformationprocess sealing, juxtaposition sealing, and of the fault
being reactivated subsequent to the trap being charged
with hydrocarbons, respectively. A value of zero is as-
signed where there is no chance of a parameter providing
a seal, and one is assigned where a parameter definitely
provides a seal. For reactivation parameter c, a valueof 1 indicates the fault is definitely reactivated. The
three factors influencing fault sealing are assumed to be
510 An Integrated, Quantitative Approach to Assessing Fault-Seal Risk
Figure 1. Classification and critical risk factors of hydrocarbon seals.
independent (the probability of independent events
occurring is the product of their individual probabil-
ities). To combine the probabilities of juxtaposition
and deformation process sealing, the probability of
neither providing a seal should be considered. The
probability of neither juxtaposition nor deformation
process sealing being developed is (1 a)(1 b).The probability of juxtaposition and/or deformation
processes providing a seal is the complement of this,
i.e., {1 [(1 a)(1 b)]}. The product of the
Jones and Hillis 511
Figure 2. Summary work flow of risking faultreactivation and seal breach through nucleationof structural permeability networks.
probability of juxtaposition and/or deformation pro-
cesses providing a seal and the probability that
reactivation has not led to the seal being breached
gives the overall probability of a fault sealing (FS),
i.e., {1 [(1 a)(1 b)]} (1 c).The fault-seal risk web (FSRWeb) illustrates the
multiparameter approach required for integrated fault-
seal risking and presents an example of probability
and seal condition (Figure 3). The seal condition scale
is analogous to industry standard probability scales,
yet with the recognition that a seal condition value
of 0.5 does not reflect an equivocal probability of
sealing. Under this scheme, a risk value of 0.5 is in-
terpreted to reflect an intermediate chance of sealing
hydrocarbons. The Sherman-Kent scale may be used
for eliciting and quantifying confidence judgements
(Table 1).
The construction of the fault-seal risk profile allows
a rapid visual assessment of the prospect risks and their
relative importance. Comparison of adjacent known
sealing/leaking traps and their fault-seal risk profile
may aid identification of key regional fault-seal issues.
INCORPORATING UNCERTAINTY
A key factor in any risking procedure is the influence of
uncertainty. Indeed, the more uncertainty that attends
a given prospect, the more a systematic expression of
subjective probability is needed (Rose, 1992). Each
component of fault-seal risk, as expressed by equation
1, has a component of uncertainty. This uncertainty
reflects a less than complete data set and technical
limitations of geologic data (e.g., seismic resolution).
Juxtaposition, deformation processes, and reactivation
fault-seal uncertainties are calculated via construction
of data webs (Figures 46). Each data web provides a
framework for evaluating the uncertainty for each
aspect of fault-seal risk and its associated impact on
overall fault-seal risk.
The uncertainty (U ) associated with each aspectof fault-seal risk is calculated by the summation of the
data web values given to each critical parameter (nw),divided by the total number of parameters (n). Criticalparameters to be included in each data web were chosen
with reference to key publications (e.g., Allan, 1989;
Knipe, 1992; Yielding et al., 1997, Hillis, 1998). U ex-pressed as a percentage can be written as
U 1 X
nw=n 100 2
The juxtaposition uncertainty illustrated in Figure
4 is expressed as
U 1 f0:9 0:6 0:9=3g 100 20% 3
This value, in conjunction with deformation process
and reactivation aspects, may be used to identify the
upper and lower integrated fault-seal risk values by
positively and negatively factoring each parameter of
equation 1 by their respective uncertainty value. These
data can be used to assess VOI and determine whether
it is economically viable to generate additional data to
reduce prospect uncertainty. Traps that are risked as
marginal (borderline geologic success), yet having a
high uncertainty may be worth additional investment,
especially when calculated returns are believed to be
high. The VOI for the Apollo prospect is discussed in
the following worked example.
A WORKING EXAMPLE: APOLLO PROSPECT,NORTH WEST SHELF, AUSTRALIA
Geologic Setting
The Apollo prospect is a rollover structure on the down-
thrown side of the northeast-southwesttrending Egret
fault system, Dampier basin, permit WA-10-R, North
West shelf, Australia (Figure 7). The reservoir target is
the Tithonian synrift Angel Formation having the Ber-
riasian Forestier Claystone forming the cap seal (Figure
8). The main risk for Apollo prospectivity is fault seal-
ing caused by the presence of large potential leak win-
dows generated by sand-sand juxtaposition. A detailed
review of the fault-seal workflow and interpretation are
beyond the scope of this publication. Therefore, only a
technical summary of prospect fault-seal evaluation
is given so as to illustrate application of the risking
methodology.
Juxtaposition Analysis
Seismic mapping suggests Egret fault displacement
maintains Angel sand-sand communication through-
out the Apollo prospect. Construction of detailed Allan
diagrams confirms Angel sand-sand juxtaposition over
closure (see Figure 9). Juxtaposition fault sealing is
therefore assigned a high-risk seal condition of 0.2 (seal
condition very unlikely). In this example, no distinc-
tion in juxtaposition seal risk is made between 10- and
100-ft oil columns as mercury injection capillary pressure
512 An Integrated, Quantitative Approach to Assessing Fault-Seal Risk
Jones and Hillis 513
Figu
re3
.Fa
ult-
seal
risk
web
and
seal
prob
abili
tyco
nditi
on(s
eete
xtfo
rde
tails
).
data from offset wells indicates the membrane seal
capacity of equivalent nonreservoir strata is capable of
supporting greater buoyancy pressures than both col-
umns would exert. In this example, maintaining a con-
stant juxtaposition parameter value between 10- and
100-ft column scenarios simply reflects the risk that
such a seal mechanism exists. The Apollo juxtaposition
data web is illustrated in Figure 4. Three-dimensional
seismic over the prospect has allowed fault-zone archi-
tecture to be accurately mapped; hence, a confidence
514 An Integrated, Quantitative Approach to Assessing Fault-Seal Risk
Table 1. Sherman-Kent Scale for Eliciting and Quantifying Confidence Judgements*
Proven; definitely true 98100%
Virtually certain; convinced 9098%
Highly probable; strongly believe; highly likely 7590%
Likely; probably true; about twice as likely to be true as untrue; chances are good 6075%
Chances are about even, or slightly better than even or slightly less than even 4060%
Could be true but more probably not; unlikely; chances are fairly poor;
two or three times more likely to be untrue than true 2040%
Possible but very doubtful; only a slight chance; very unlikely indeed; very improbable 220%
Proven untrue; impossible 02%
*Meyer and Brooker, 1991; Watson, 1998.
Figure 4. Juxtaposition data web. Exam-ples of the position on each parameterspine with respect to data quality andquantity are given. In the example illus-trated, the combined juxtaposition uncer-tainty is 20% with the greatest uncertaintyrelated to sedimentary architecture (seetext for details).
value of 0.9 is ascribed to this parameter on the data
web spine. Sedimentary architecture is less certain with
recognized facies variability across the basin and gamma
log-defined strata from adjacent wells. A data web con-
fidence value of 0.6 is ascribed to sedimentary architec-
ture reflecting a lack of prospect-specific data. Principal
fault throw magnitude is ascribed a data web confidence
value of 0.9 reflecting 3-D seismic mapping of the fault
zone. Integrated uncertainty with respect to a juxtapo-
sition seal mechanism is calculated as 20% (equation 2
and Figure 4).
Fault-Rock Process Analysis
Prediction of fault-rock processes using Juxtaposition
software (Knipe, 1992) focused on the likelihood of
sealing through sand-sand deformation given the like-
lihood of reservoir-reservoir juxtaposition. Both grain
boundary (dissagregation zones) and cataclastic pro-
cesses occur through clean (< 15% clay content) sand
deformation yet under distinct effective stress mag-
nitudes (Fisher and Knipe, 1998). Timing of faulting
relative to burial history is, therefore, a crucial element
to consider when evaluating membrane seal formation.
Structural reconstruction of the prospect indicates depth
of burial at the time of faulting to be less than 2 km
at relatively low vertical effective stress and early in
the reservoir burial history. Faulting of clean Angel
Formation sands is modeled to generate disaggregation
zones (Figure 10a) that, without diagenetic enhance-
ment, will act as leak windows and allow cross-fault
communication. Shale gouge ratio values correspond-
ing to reservoir-reservoir deformation range from ap-
proximately 14 to 16 reflecting faulting of relatively
clean/low GR sands (Figure 10b). Microstructural and
petrophysical analysis of core scale faults from an ad-
jacent field confirms the presence of disaggregation
zone faults in the Angel Formation strata. No post-
deformation diagenetic processes were observed in
these faults, suggesting that diagenetic enhancement
of Apollo disaggregation zone seal capacity is unlikely.
Mercury injection capillary pressure evaluation of these
faults indicates an oil seal capacity of less than 1 ft.
Hydrocarbon trapping greater than the 10-ft oil column
through fault-rock process sealing was subsequently
risked as 0.3 (seal condition unlikely). The potential for
Jones and Hillis 515
Figure 5. Fault rock process data webfor the Apollo prospect. Integrated un-certainty associated with fault rock pro-cess sealing is 26% (see text for details).
the trap-bounding fault baffling an oil column greater
than 100 ft is risked as 0.1 (seal condition extremely un-
likely). The fault-rock processes seal uncertainty is cal-
culated as 26% (Figure 5).
Fault-Seal Reactivation Analysis
As discussed above, reactivation of faults nucleates
networks of structural permeability that can result in
across- and up-fault hydrocarbon migration. P ex-presses the likelihood of seal failure (see Figure 2). A
high P value represents a relatively low-risk fault asa significant pore pressure increase is required to in-
duce failure. Structural permeability modeling of the
Apollo prospect fault shows the majority of the fault
within closure to be low-reactivation risk having Pof approximately 38 MPa. However, a high-risk (P< 5 MPa) fault section is identified related to a change
in fault strike (see Figure 11). This high-risk, low Pregion is coincident with the southern margin of clo-
sure for the Egret field on the footwall side of the
Apollo structure. Egret is underfilled, and, as charge is
not believed to be limited, it is feasible that structural
permeability networks may be responsible for con-
trolling the volume of hydrocarbons in the footwall
trap. Given that P for the Apollo prospect fault isabout 38 MPa, a seal condition of 0.1 (extremely un-
likely that the fault is reactivated) is used for risking
the likelihood of seal breach. The fault reactivation
uncertainty is calculated as 27% (Figure 6).
INTEGRATED FAULT-SEAL RISKING OF THEAPOLLO PROSPECT
Quantitative fault-seal risking was conducted on the
likelihood of supporting 10- and 100-ft oil columns
to allow integration with prospectivity and probabi-
listic volumetric statements. The critical structural risk
parameter for Apollo is the ability of disaggregation zone
faults (low SGR fault gouge) to support increasingly large
hydrocarbon columns. Given the above assessments of
516 An Integrated, Quantitative Approach to Assessing Fault-Seal Risk
Figure 6. Reactivation data web for theApollo prospect. Integrated uncertaintyassociated with the likelihood that thefault seal is unreactivated is 27% (seetext for details).
Jones and Hillis 517
Figure 7. Three-dimensional two-way traveltime image of Apollo (hanging wall) and Egret (footwall) prospects. Fault plane istransparent to allow visualization of both footwall and hanging-wall prospects. Figure insert: location of North West shelf, petroleumprovince, and Dampier subbasin.
518 An Integrated, Quantitative Approach to Assessing Fault-Seal Risk
Figure 8. Generalized Triassic/Jurassic stratigraphy of Dampier subbasin.
Figure 9. Allan diagram ofApollo fault. Note the presenceof large areas of sand-sandjuxtaposition in prospect clo-sure.
Jones and Hillis 519
Figu
re1
0.
Faul
tro
ckty
pepr
edic
tion
for
Apo
llopr
ospe
ct.N
ote
the
deve
lopm
ent
ofdi
sagg
rega
tion
zone
sdu
eto
sand
-san
dde
form
atio
npr
oces
ses
atlo
wef
fect
ive
stre
ss.T
hese
zone
sar
em
odel
edto
form
leak
win
dow
s(s
eete
xtfo
rde
tails
).Sh
ale
goug
era
tioda
taar
epr
esen
ted
for
com
pari
son.
FW=
foot
wal
l;H
W=
hang
ing
wal
l.
520 An Integrated, Quantitative Approach to Assessing Fault-Seal Risk
Figu
re1
1.
Apo
llofa
ult
reac
tivat
ion
pred
ictio
n.Fi
gure
left,
stru
ctur
alpe
rmea
bilit
ypo
les-
to-p
lane
ster
eone
t.H
igh-
risk
faul
tse
ctio
nsst
rike
betw
een
070
and
160
and
dip
grea
ter
than
60j.
Figu
reri
ght,
map
ping
ofre
activ
atio
nri
skon
tofa
ult
trac
e(s
eete
xtfo
rde
tails
).
Jones and Hillis 521
Figu
re1
2.A
pollo
faul
t-se
alri
skw
eb,r
isk
prof
iles,
and
inte
grat
edpr
ospe
ctfa
ult-
seal
risk
stat
emen
tsfo
r10
-an
d10
0-ft
oilc
olum
ns.N
ote
the
incr
ease
dfa
ult-
zone
proc
esse
sse
albr
each
risk
asso
ciat
edw
ithth
ela
rger
oil
colu
mn.
the individual components of fault sealing [a = 0.3 (10-ftcolumn) and a = 0.1 (100-ft column); b = 0.2; c = 0.1]and utilizing equation 1, the probability that the Apollo
trap-bounding fault seals a 10-ft oil column is 0.4 (seal
condition moderately unlikely), see Figure 12. The like-
lihood that the fault seals oil columns greater than 100 ft
is 0.3 (seal condition unlikely).
Data web errors for the Apollo prospect are 20%
(juxtaposition uncertainty), 26% (fault-rock process
uncertainty), and 27% (fault reactivation). Recalculat-
ing each parameter by its uncertainty, the upper
(positive) value of integrated fault-seal risk for a 10-ft
oil column is 0.5 (seal condition intermediate) The
lower integrated risk value for a 10-ft oil column is 0.3
(seal condition unlikely). The upper value of integrated
fault-seal risk for a 100-ft oil column is 0.3 (seal con-
dition unlikely). The lower integrated risk value for a
100-ft oil column is 0.2 (seal condition very unlikely).
The variation in the final Apollo risk calculation re-
flects the lack of prospect-specific data and the critical
importance of accurately determining the probability of
fault-seal failure through reactivation. Data employed
in the final integrated fault-seal risk calculations are
summarized in Table 2.
These data indicate the greatest uncertainty and,
therefore, most influencing parameters on Apollo fault-
seal prospectivity are sedimentary architecture/clay con-
tent and fault failure envelope. Given that the average
fault P within closure is approximately 38 MPa, theApollo fault is unlikely to be reactivated under present-
day in-situ stress conditions (hence, c = 0.1). Therefore,there appears little business driver or technical justi-
fication to collect further information on the failure
envelope of the fault. Following this rationale, the
greatest VOI for Apollo is in obtaining additional data
on the sedimentary architecture and clay content var-
iability. Value of information analysis is increasingly
being used as a discriminator to justify data acquisition.
Value of information can be thought of as value of a
project executed with certain information minus value
of that project executed without that information.
Hence, VOI components include benefit of informa-
tion and cost of acquiring information. The cost of ac-
quiring the analysis is relatively simple to ascertain in
that most data have a monetary value attached. The
degree of uncertainty and time range in project net
present value (NPV) influence the VOI benefit. For
example, early in project development, the range in
potential NPV may be large as many factors such as
field compartmentalization and migration pathway may
carry significant uncertainty. At this stage in project ap-
praisal, key data would carry a high VOI as it is likely
to significantly influence the decision-making process.
As projects mature, the impact of additional data on
reducing the NPV range is likely to have relatively minor
impact as overall project uncertainty will have been
reduced and time to revenue will be shorter. Hence,
data will carry a relatively reduced VOI benefit. The
cost of acquiring additional data needs to be evaluated
with reference to impact on overall prospectivity. Given
the impact of sedimentary architecture and clay content
data on Apollo fault-seal prospectivity range and poten-
tial time to revenue, VOI benefit in this case is likely to
outweigh the cost of acquisition or reduce uncertainty
range through modeling. VOI justification of this state-
ment is that additional data may sufficiently reduce
522 An Integrated, Quantitative Approach to Assessing Fault-Seal Risk
Table 2. Apollo Fault Risking and Seal Parameter Uncertainty*
Apollo Risking Parameters
Risk Web Values Juxtaposition Fault-rock Processes Reactivation
10-ft oil column 0.2 0.3 0.1
100-ft oil column 0.2 0.1 0.1
Data Web Components and
Confidence Parameters Fault-zone architecture: 0.9 Depth of burial: 0.9 Stress regime: 0.7
Sedimentary architecture: 0.6 Sedimentary architecture: 0.6 Failure envelope: 0.6
Fault throw: 0.9 Fault throw: 0.9 Fault architecture: 0.9
Hydrocarbon column height: 0.8
Stratigraphic clay content: 0.6
Data Web Uncertainty 20% 26% 27%
*See text for details.
stratigraphic uncertainty to the extent that a drill-or-
not-to-drill decision can be made. In the negative case,
a decision to divert well costs could be made, thus sav-
ing considerable dollar value. Whether to always re-
duce key uncertainties as identified by the fault-seal
risking process should be made in conjunction with VOI
studies.
DISCUSSION
If a fault is considered to be definitely reactivated sub-
sequent to hydrocarbon charge, the integrated fault-
seal risk determined using the methodology herein is
zero, i.e., no chance of fault seal. We believe that the
assumption that fault reactivation postcharge leads to
seal breaching, and loss of hydrocarbons from the trap
provides a reasonable and practical basis for the as-
sessment of fault-seal exploration risk. However, the
reactivation risk is a critical parameter that must be
assessed with caution.
We follow Sibsons (1992) fault-valve model that
proposes that fault rupture leads to the creation of
fracture permeability and fluid discharge. Only during
fault movement, or at stresses close to criticality, does
the fault act as a conduit for fluid flow. Hence, reacti-
vation must occur postcharge for the seal to be breached.
The fault may again seal after reactivation. For example,
fluid flow along the fault may promote cementation, or
fault movement may assist the development of deforma-
tion process seals, and the fault may again seal hydro-
carbons once quiescent, if the reservoir is recharged.
Further reactivation would again cause the fault to leak,
hence, the process may be cyclic (Sibson, 1992; 1994).
Increasing pore pressure while the fault is quiescent
may play a critical role in initiating the next phase of
reactivation.
The assumption that fault reactivation postcharge
leads to seal breach is valid in the Australian context
where numerous fault traps, the bounding faults of
which have been reactivated, are associated with re-
sidual columns witnessed by geochemical analysis of
fluid inclusions in the reservoir (e.g., Lisk et al., 1998).
The reactivated trap-bounding faults are commonly
associated with anomalous hydrocarbon-related diage-
netic zones and in some cases present-day seepage of
hydrocarbons into the water column (OBrien and
Woods, 1995). These faults have clearly been reacti-
vated postcharge and acted as conduits for the escape
of hydrocarbons. In other basins, reactivation may be
relatively minor and may not always lead to loss of an
entire accumulation. The risk posed by reactivation can
therefore be thought of as area specific and should be
calibrated by the type of observations that confirms its
significance in the Australian context. Furthermore,
the P methodology for assessing the risk of reactiva-tion and associated seal breach uses knowledge of the
current in-situ stress field and is only appropriate if re-
activation is occurring at the present day, or in a stress
field similar to that of the present day. If reactivation
occurred under a paleostress regime, then reactivation-
related risk should be assessed with reference to that
paleostress regime.
The location of high-risk reactivation points rela-
tive to trap geometry/spill point also needs to be con-
sidered when producing a final integrated fault-seal
risk statement. It is feasible that reactivation may occur
downdip of the crest of the structure and may not breach
the entire accumulation. Hence, one may wish to gen-
erate a series of risk statements that link to volumetric
statements that consider several scenarios, i.e., reac-
tivation at the crest of the trap and/or reactivation of a
section of the fault downdip from the crest but above
the trap spill point. Regardless of the approach taken,
it is critical that other members of the team are fully
aware of the exploration/production implications as-
sociated with the reactivation risk.
CONCLUSIONS
A quantitative assessment of fault-seal risk that in-
tegrates parameters from different aspects of fault-seal
analysis in a consistent framework may be determined
if the risks associated with juxtaposition sealing, de-
formation process sealing, and reactivation are known.
The impact of uncertainty and VOI for each aspect of
fault sealing may be determined via the construction
of data webs and modification of equation 1. The fault-
seal risk web profile provides a powerful tool for vi-
sualizing each parameter probability of fault sealing and
allows rapid comparison with proven success/failure
prospect cases.
REFERENCES CITED
Allan, U. S., 1989, Model for hydrocarbon migration and entrapmentwithin faulted structures: AAPG Bulletin, v. 73, p. 803811.
Antonellini, M., and A. Aydin, 1994, Effect of faulting on fluid flowin porous sandstones: petrophysical properties: AAPG Bulletin,v. 78, p. 355377.
Barton, C. A., M. D. Zoback, and D. Moos, 1995, Fluid flow along
Jones and Hillis 523
potentially active faults in crystalline rock: Geology, v. 23,p. 683686.
Bell, J. S., 1996, In situ stresses in sedimentary rocks (part 1): measure-ment techniques: Geoscience Canada, v. 23, p. 85100.
Bouvier, J. D., K. Sijpesteijn, D. F. Kleusner, C. C. Onyejekwe, andR. C. van der Pal, 1989, Three dimensional seismic interpreta-tion and fault sealing investigations, Nun River, Nigeria: AAPGBulletin, v. 73, p. 13971414.
Dewhurst, D. N., and R. M. Jones, 2002, Geomechanical,microstructural and petrophysical evolution in experimentallyreactivated cataclasites: application to fault seal prediction:AAPG Bulletin, v. 86, no. 8, p. 13831405.
Ferrill, D. A., J. Winterle, G. Wittmeyer, D. Sims, S. Colton, A. Ams-trong, and A. P. Morris, 1999, Stressed rock strains ground-water at Yucca Mountain, Nevada: GSA Today, v. 9, p. 18.
Finkbeiner, T., M. Zoback, P. Flemings, and B. Stump, 2001, Stress,pore pressure, and dynamically constrained hydrocarboncolumns in the South Eugene Island 330 field, northern Gulfof Mexico: AAPG Bulletin, v. 85, p. 10071031.
Fisher, Q. J., and R. J. Knipe, 1998, Fault sealing processes insiliciclastic sediments, in G. Jones, Q. J. Fisher, and R. J.Knipe, eds., Faulting, fault sealing and fluid flow in hydro-carbon reservoirs: Geological Society (London) Special Pub-lication 147, p. 117134.
Freeman, B., G. Yielding, and M. Badley, 1990, Fault correlationduring seismic interpretation: First Break, v. 8, no. 3., p. 8789.
Gibson, R. G., 1994, Fault zone seals in siliciclastic strata of theColumbus Basin, offshore Trinidad: AAPG Bulletin, v. 78,p. 13721385.
Grauls, D. J., and J. M. Baleix, 1994, Role of overpressures and insitu stresses in fault-controlled hydrocarbon migration: a casestudy: Marine and Petroleum Geology, v. 11, p. 734742.
Hillis, R. R., 1998, Mechanisms of dynamic seal failure in the TimorSea and Central North Sea, in P. G. Purcell and R. R. Purcell,eds., The sedimentary basins of Western Australia 2: Pro-ceedings of Petroleum Exploration Society of Australia Sym-posium, Perth, Western Australia, Bath, Geological Society,p. 313324.
Jones, G., Q. J. Fisher, and R. J. Knipe, 1998, Faulting, fault sealingand fluid flow in hydrocarbon reservoirs: Geological Society(London) Special Publication 147, 319 p.
Jones, R. M., P. Boult, R. R. Hillis, S. D. Mildren, and J. Kaldi,2000, Integrated hydrocarbon seal evaluation in the PenolaTrough, Otway basin: Australian Petroleum Production andExploration Association Journal, v. 40, no. 1, p. 194211.
Knipe, R. J., 1992, Faulting process and fault seal, in R. M. Larson,H. Breke, B. T. Larsen, and E. Talleraas, eds., Structural andtectonic modelling and its application to petroleum geology:Norwegian Petroleum Society Special Publication 1, Stavan-ger, p. 325342.
Knipe, R. J., G. Jones, Q. J. Fisher, B. Clennell, E. White, D.
Brown, E. McAllister, B. Kidd, A. Harrison, A. Farmer, N.Porter, and Z. Shipton, 1995, Fault and fault sealing in N. Seareservoirs: Amalgamated Data Report, Pp Internal RDRReport, p. 250.
Lisk, M., M. P. Brincat, P. J. Eadington, and G. W. OBrien,1998, Hydrocarbon charge in the Vulcan sub-basin, in P. G.Purcell and R. R. Purcell, eds., The sedimentary basins ofWestern Australia 2: Proceedings of Petroleum ExplorationSociety of Australia Symposium, Perth, Western Australia:Bath, Geological Society, p. 288302.
Meyer, M., and J. Brooker, 1991, Eliciting and analysing expertjudgment: A practical guide: London, Academic Press.
Mildren, S. D., R. R. Hillis, and J. Kaldi, 2002, Calibratingpredictions of fault seal reactivation in the Timor Sea:Australian Petroleum Production and Exploration AssociationJournal, v. 42, p. 187202.
Morris, A., D. A. Ferrill, and D. B. Henderson, 1996, Slip-tendencyanalysis and fault reactivation: Geology, v. 24, p. 275278.
Nybakken, S., 1991, Sealing fault traps an exploration concept ina mature petroleum province: Tampen Spur, northern NorthSea: First Break, v. 9, p. 209222.
OBrien, G. W., and E. P. Woods, 1995, Hydrocarbon-relateddiagenetic zones (HRDZs) in the Vulcan sub-basin, Timor Sea:recognition and exploration implications: Australian Petro-leum Production and Exploration Association Journal, v. 35,p. 220252.
Otis, R. M., and N. Schneidermann, 1997, A process for evaluatingexploration prospects: AAPG Bulletin, v. 81, p. 10871109.
Rose, P. R., 1992, Chance of success and its use in petroleumexploration, in R. Steinmetz, ed., The business of petroleumexploration: AAPG Treatise of Petroleum Geology, p. 7186.
Shuster, M. W., S. Eaton, L. L. Wakefield, and H. J. Kloosterman,1998, Neogene tectonics, greater Timor Sea area, offshoreAustralia: implications for trap risk: Australian PetroleumProduction and Exploration Association Journal, v. 38,p. 351379.
Sibson, R. H., 1992, Implications of fault-valve behaviour forrupture nucleation and occurrence: Tectonophysics, v. 211,p. 283293.
Sibson, R. H., 1994, Crustal stress, faulting and fluid flow, in J.Parnell, ed., Geofluids: origin, migration and evolution of fluidsin sedimentary basins: Geological Society (London) SpecialPublication 78, p. 6984.
Watson, P., 1998, A process for estimating geological risk of petro-leum exploration prospects: Australian Petroleum Productionand Exploration Association Journal, v. 38, p. 577583.
Wiprut, D., and M. D. Zoback, 2000, Fault reactivation and fluidflow along a previously dormant normal fault in the northernNorth Sea: Geology, v. 28, p. 595598.
Yielding, G., B. Freeman, and D.T. Needham, 1997, Quantitativefault seal prediction: AAPG Bulletin, v. 81, p. 897917.
524 An Integrated, Quantitative Approach to Assessing Fault-Seal Risk