Integrated seal assessment and geologic risk with …...Integrated seal assessment and geologic risk with application to the Skua Field, Timor Sea, Australia Scott D. Mildren 1,2,3,

Integrated seal assessment and geologic risk with application to the Skua Field,Timor Sea, Australia

Scott D. Mildren1,2,3, Richard R. Hillis1, 2, Tom Kivior1, 2, 4 and John G. Kaldi1, 2

Keywords: cap rock, fault, seal, risk, uncertainty, trap integrity, Skua Field, Timor Sea, Vulcan Sub-basin

Abstract

Trap integrity is one of four components comprisinggeological success, when risking prospects and generatingestimates of hydrocarbon reserves. The three remainingelements are source, reservoir and dynamics (timing/migration). We present an integrated work flow thatcombines cap rock and fault seal analyses to predict trapintegrity, so that it can be incorporated into prospectevaluation procedures. Sealing mechanisms of cap rocksand faults are represented by six parameters that inputinto the integrated work flow: Cap Rock Capacity, CapRock Geometry, Cap Rock Integrity, Fault PlaneCapacity, Juxtaposition Lithology Capacity and Post-Charge Reactivation. Quantitative expressions are usedto generate sealing probabilities that correspond to eachparameter, and subjective descriptions of data quality andquantity are used to modify them according to theiruncertainty. Sealing probabilities are combined in a wayconsistent with geological concepts to evaluate theprobability of a sealing cap rock (Pcap), the probabilitythat the bounding fault is sealing (Pfault) and the combinedprobability that the trap is sealing. The integrated workflow is iteratively evaluated for varying trap volumes togenerate probability distributions.

The Skua Field is located in an area with variable caprock sealing properties and susceptibility to fault-relatedleakage. The integrated work flow is applied to the SkuaField to demonstrate the evaluation process. For all trapvolumes, the sealing probability of the cap rock is lowerthan the sealing probability of the Skua Fault. Thissuggests the possibility that hydrocarbon leakage at Skuais associated with a cap rock-related failure mechanism,either in conjunction or in place of fault-related leakage,as previously thought. The probability that the trap issealing indicates that for hydrocarbon volumes of lessthan 7%, the trap is ‘more likely to be leaking than not’and ‘highly likely to be leaking’ when considering greaterhydrocarbon volumes. There is a 28% chance that thetrap can fill to 7% capacity and a 10% chance the trapcan fill to maximum capacity.

Introduction

Hydrocarbon exploration is a risky business. Petroleumcompanies today employ risk and uncertainty analyses inorder to maximise exploration performance and optimiseportfolio management. Otis and Schneidermann (1997)

illustrated the Chevron Overseas Petroleum Inc. approach,which is based on the play concept; source, reservoir, trapand dynamics (timing/migration). Risk assessment assignsa probability of success between 0 (failure) and 1 (success)to each of these four elements and the multiplicative resultyields the probability of geologic success (MacKay, 1996;Otis and Schneidermann, 1997). This paper concentrates onthe input parameters that comprise the geologic risk of thetrap within the play concept.

Previously, hydrocarbon traps have been judged byconsidering a checklist of the critical aspects of geologic risk(Goldstein, 1994; Otis and Schneidermann, 1997; Rose,2001). A subjective description of risk is associated with eachitem on the list and the ‘weakest’ element determines theprobable geologic success associated with the trap. This paperproposes a relationship between six elements of geologic risk,based on the sealing mechanisms of faults and cap rocksrather than using subjective estimates.

The parameters of geologic risk associated with cap rocksand faults have previously been individually discussed withinthe literature (Murris, 1980; Downey, 1984; Sluijk and Parker,1986; Allan, 1989; Bouvier et al., 1989; Freeman et al., 1990;Knipe, 1992; Antonellini and Aydin, 1994; Gibson, 1994;Yielding et al., 1997; Jones et al., 2000; Mildren et al. 2002a;Mildren et al. 2002b). Integrated studies of more than twoparameters are less frequent. Kaldi and Atkinson (1997)identified the three parameters that affect cap rock sealpotential to be seal capacity (the calculated hydrocarboncolumn height a lithology can support), seal geometry (thestructural position, thickness and areal extent of the lithology)and seal integrity (propensity for fracturing). Similarly, Joneset al. (2002) identified juxtaposition (sealing lithologiesjuxtaposed against non-sealing lithologies), deformationprocesses (sealing properties of the fault rock material) andfault reactivation (fracture networks associated with the fault)to be the elements associated with fault seal. These parametersfollow-on from the Watts (1987) mechanistic sealclassification, and each can compromise the integrity of ahydrocarbon trap.

Jones and Hillis (2003) went further to describe aquantitative relationship between the three parameters forfault-seal failure. However, in order to fully assess thegeologic risk of a hydrocarbon trap, the critical parametersfor the cap rock must be evaluated and combined withthose of the bounding faults. This paper builds on theapproach by Jones and Hillis (2003) to provide aframework for quantifying the risks associated with caprocks and incorporating it with a modified Jones et al.(2002) fault-seal relationship, to give a single quantitativerelationship for evaluating the geologic risk associatedwith a hydrocarbon trap. The six risk parameters arepresented in the context of fundamental tenets that describecap rock seal, fault seal and the overall hydrocarbonintegrity of the trap. Methods are suggested to enable the

1 Australian Petroleum Cooperative Research Centre2 Australian School of Petroleum, University of Adelaide3 Currently: JRS Petroleum Research, 45 Woodforde Road, Magill

SA 5072. Email: [email protected] Currently: Schlumberger Oilfield Service

276 Integrated seal assessment and geologic risk, Skua Field

conversion of quantitative parameter analyses intocorresponding probabilities of sealing.

A useful by-product of defining a single quantitativerelationship for trap risk is a holistic work flow, by which atrap can be comprehensively assessed. The advantage of thework flow is the assurance that all failure mechanisms areconsidered, independent of the evaluation methodology usedin each case. The focus of any integrated seal integrity studyis on the parameters deemed critical in the context of thattrap or basin. For example, in a proven exploration area,where fault reactivation is not considered an issue and faultthrow is minimal, efforts should concentrate on evaluatingthe cap seal. Probability estimates for fault seal are inherentlyincorporated, although they don’t need to be the result ofdetailed analyses. Ignoring the fault risk parameters in theabove situation implicitly assigns them a probability of 1,such that they do not decrease the risk associated with thetrap. However, the uncertainty of the value used for eachparameter must be considered. Based on the approach byNakanishi and Lang (2001), this paper also incorporatesprobability modifiers with respect to the uncertaintyassociated with the evaluation of each parameter.

The work flow for assessing geologic risk associated witha trap is implemented for the Skua Oil Field in the TimorSea. One of the main exploration challenges in the TimorSea is trap integrity. The area is considered to be reactivatedwithin the present-day stress environment and empiricalevidence strongly links hydrocarbon migration with faults(O’Brien and Woods, 1995; Lisk et al., 1998; O’Brien et al.,1998; Gartrell et al., 2002). In addition, the potential for the

regional sealing units to preserve hydrocarbon accumulationsvaries significantly across the basin (Kivior et al., 2002). Thisenvironment calls for an integrated approach to trapassessment and, therefore, the Skua Field is an appropriatelocation to illustrate the integrated framework approach forrisking and evaluating traps.

Parameters of seal risk and uncertainty

The qualitative relationships between the parameters usedto evaluate geologic risk of a hydrocarbon trap define aholistic work flow for seal evaluation (Fig. 1). The work flowis constructed around the fundamental statement describinga sealing trap, such that:

“a structurally bound trap is considered to be sealing if boththe cap rock and bounding faults are sealing”.

Where no seismically resolvable faults are present, a trapis solely dependent on the cap rock to be sealing. In thesetwo scenarios, either the faults and the cap rock need to beevaluated or only the cap rock needs to be evaluated.

Jones and Hillis (2003) proposed that a fault is sealing if:

“deformation processes have created a membrane seal or ifit juxtaposes sealing rocks against reservoir rocks, and thefault has not been reactivated post-charge”.

It follows that the three parameters that representmechanisms of fault-seal failure can be defined as: fault-plane

fault trap requires cap seal

fault traprequires fault seal

Trap TypeIs the trap cut

by a seismically-resolved fault?

NO YES

A fault trap is sealing if both thecap rock and the fault are sealing

A fault seals if deformation processes have created a membrane seal or if it juxtaposes

sealing rocks against reservoir rocks, and the fault has not been reactivated post-charge.

P = {1-[(1-a)*(1-b)]}*(c)fault

Probabilities (0-1)a:fault plane capacity

b: juxtaposition lithology capacityc: post-charge reactivation

Fault Seal

Fault Seal

A cap rock seals if it has membrane sealingproperties, continuously covers the trap

and is not cut by open fractures.

P = i*j*kcap

Probabilities (0-1)i: cap rock capacityj: cap rock geometryk: cap rock integrity

Cap Seal

ij

k

Cap Seal

if no fault P = 1fault

= (i*j*k) * {1-[(1-a)*(1-b)]}*(c)Ptrap

ab

c

Figure 1. Work flow for assessing the probability that a trap is sealing, based on six sealing parameters (see text). Venn diagrams illustratingthe relationships between key cap parameters for the probability of cap rock seal and the probability of fault seal are also included.Probability domains are shaded in blue.

Mildren et al. 277

capacity, juxtaposition lithology capacity and post-chargereactivation. These parameters are subsequently defined asthe likelihood that:

• deformation processes have created a membrane seal;• the fault juxtaposes sealing rocks against reservoir rocks;• the fault has not been reactivated subsequent to

hydrocarbon charge.

The corresponding probabilities that represent thelikelihood of the fault sealing are denoted (a), (b) and (c),respectively. Originally, Jones and Hillis (2003) describedthe parameter, (c), as the risk of reactivation such that a valueof c = 1 implies a fault is reactivated and leaking. For thesake of consistency with the other five parameters, this paperconsiders (c) to be the probability that the fault has not beencompromised and is sealing. Therefore c = 0 whenreactivation is likely and c = 1 when reactivation is notlikely (i.e. c = 1-c).

In a similar fashion it is proposed that a cap rock issealing if:

“it has membrane sealing properties, continuously coversthe trap and is not cut by open fractures”.

Fractures here are taken to be those below seismicresolution. If fractures (faults) cutting the seal are recognisedon seismic data, the trap is considered a fault trap. Thisstatement also describes the required components for asealing cap rock (Kaldi and Atkinson, 1997): cap rockcapacity, cap rock geometry and cap rock integrity.Alternatively, these parameters can be described as:

• the maximum hydrocarbon column height the cap rockcan sustain;

• the likelihood that the cap rock continuously covers the trap;• the likelihood that the cap rock is cut by open fractures.

The corresponding probabilities that describe thelikelihood of sealing by each of these parameters are denoted(i), (j) and (k), respectively.

Jones et al. (2002) defined a qualitative probabilityrelation for fault seal using the three sealing probabilities (a),(b) and (c), based on their fundamental statement. Usingprobability theory, an ‘and’ relationship between parametersis indicative of ‘intersection’ probability space. The ‘and’and ‘or’ relationship is indicative of ‘union’ probability space.The ‘and’ and ‘or’ relationships inherent in the statementsfor cap rock, fault and trap seal can therefore be representedas probability spaces comprised of the six risk parameters.An intuitive illustration of these relationships can also bemade using Venn diagrams (Fig. 1). Venn diagrams are aschematic representation of collections of sets that depictthe relationship between parameters using logic theory(Cundy and Rollett, 1989).

The fault seal statement can be represented as the intersectionof (c) with the union of (a) and (b) to give the relationship:

Pfault = {1-[(1-a)*(1-b)]}*c (Equation 1),

previously defined by Jones and Hillis (2003). Pfault is definedas the probability that a fault is sealing and can be illustrated

by Venn diagram (Fig. 1). Similarly, the cap seal statementcan be represented as the intersection of (i), (j) and (k) andrepresented by the relationship:

Pcap = i*j*k (Equation 2),

where Pcap is defined as the probability that the cap rock issealing (Fig. 1).

Each parameter is evaluated as a probability that the caprock/fault is sealing with values ranging between 0 and 1. Avalue of 0 implies that breach is assured and a value of 1implies that the seal will not be compromised. A value of 0.5implies there is significant uncertainty and that it is a “tossup” between sealing and leaking (Fig. 2).

Whereas risk implies the threat of loss, uncertainty refersto the range of probabilities that some conditions may exist

1.0

0.0

0.6

0.8 0.7

0.5 0.4 0.3 0.2 0.1

0.9 Definitely sealing Highly likely to be sealing

More likely to be sealing than not }

} More likely to be leaking than not

Possibly leaking, possibly sealing }

Highly likely to be leaking Definitely leaking

Significant Uncertainty

Certainty

Prob

abili

ty

Certainty

Figure 2. Subjective expressions of uncertainty corresponding withsealing probabilities (after Rose, 2001; Nakanishi and Lang, 2001).

or occur (Rose, 2001). Subjective expressions of confidence foreach critical parameter can be used to determine risk modifiers.Given that a probability of 0.5 implies there is significantuncertainty, an uncertainty modifier will alter the probabilitycloser to 0.5 relative to the size of the confidence range. Theuncertainty of the dataset used to calculate each parametercan be described on a subjective scale (i.e. plentiful, enough,moderate, poor and very poor (Nakanishi and Lang, 2001).The subjective scale permits numerical modification ofprobabilities according to data quality, quantity, resolutionand assumptions. An uncertainty-probability modifier matrixis used to alter parameter probabilities with respect to thesubjective description of uncertainty. For example, a poorlyprocessed 2D seismic dataset with no well ties could be used todefine the cap rock seal geometry parameter. The interpretationof that data might suggest that the sealing horizon entirely coversthe trap and is of considerably thickness, therefore beingassigned a probability of 1.0. However, the ‘poor’ quality ofthe dataset requires the probability to be modifiedaccordingly. Using the uncertainty-probability matrix(Table 1), the probability of 1.0 is modified to become 0.625for ‘poor’ quality data, reducing the likelihood of sealing to be‘more likely to be sealing than not’, rather than ‘definitely sealing’(Fig. 2).

Geologic risk differs from economic risk in that it isindependent of volume predictions and project economics(Watson, 1998). However, decisions must ultimately beguided by commercial risk and therefore prospectivity shouldbe incorporated in the form of a probabilistic volumestatement. Each of the parameters has a dependency onelements relating to the considered trap volume, either


directly, such as trap height, or indirectly, such as the location ofhigh risk areas relative to the trap closure. Therefore volumetricrisk assessment can also be incorporated when evaluating trapprospectivity using this integrated work flow. For example, caprock capacity evaluated at the base of a structure is less criticalthan at the crest. A low probability of sealing (high risk) evaluatedat the base of the structure will not affect the geological riskassociated with the trap (Ptrap), when considering hydrocarbonvolumes less than 100% of the trap closure. It is recommendedthat the integrated work flow presented herein be iterativelyemployed for a range of fill volumes to define a volumetricrisk distribution.

Cap Rock Capacity

Kaldi and Atkinson (1997) defined Cap Rock Capacity asthe maximum hydrocarbon column height that a cap rockcan hold before the buoyancy pressure of the column exceedsthe capillary resistance of the seal. The probability that thecap rock is sealing is intrinsically linked to the volume ofhydrocarbons being considered (volumetric trap height).Therefore, Cap Rock Capacity can be quantitativelyevaluated using the following relationship:

If the maximum hydrocarbon column height that the caprock can maintain exceeds the volumetric height of the trap,the Cap Rock Capacity exceeds 1 and the probability ofsealing i = 1. If the maximum column height is less than thecapacity of the trap, then the cap rock capacity is less thanone and defines the probability of sealing (i.e. i = Cap RockCapacity).

Trap height can be evaluated using standard seismicinterpretation techniques and can be varied according to thehydrocarbon volume being considered. The maximumcolumn height of the cap rock is commonly determined byMercury Injection Capillary Pressure (MICP) analysis. MICPanalysis is based on the capillary law governing liquidpenetration into small pores. Capillary forces in the seal area function of surface and interfacial liquid tensions, pore-throat size and shape, and the wetting properties of the rock(Schowalter, 1979; Vavra et al., 1992). The threshold pressureis the pressure at which mercury enters the pore system as acontinuous filament as it is injected into the sample. Injectionpressures from the mercury/air system can be converted tothe hydrocarbon/brine system. The maximum column heightis then determined from the calculated hydrocarbon–brinethreshold pressure, in situ fluid densities, and the pressuregradient of pure water.

Cap Rock Geometry

Kaldi and Atkinson (1997) defined Cap Rock Geometry asthe likelihood that the cap rock continuously covers the trap.In order to continuously cover the trap, the cap rock unitmust have a greater areal extent than the trap and must not betectonically denuded from the structure due to sub-seismicfaulting (the influence of seismically-resolvable faults on trapintegrity is assessed under fault seal). A quantitativerelationship describing Cap Rock Geometry should consistof two components representing coverage by the cap rockand thickness of the cap rock relative to the seismic resolvablefault throw, such that:

Cap Rock Coverage can be quantitatively defined as theareal extent of the cap rock relative to the areal extent of thetrap and expressed by the following relationship:

The areal extent of the trap is dependent on the consideredhydrocarbon volume.

Cap Rock Thickness is quantitatively defined as the along-fault thickness of the cap rock relative to the seismically-resolvable fault throw and expressed as:

Cap Rock Thickness evaluates the likelihood that the caprock thickness is greater than the throw of any cross-cuttingfaults.

To convert Cap Rock Geometry to a probability (j),Cap Rock Coverage and Cap Rock Thickness are restrictedto values between 0 and 1. When Cap Rock Coverageexceeds a value of 1, the cap rock is considered to coverthe areal extent of the volumetric trap and assigned aprobability of 1. Similarly, when Cap Rock Thicknessexceeds 1, the thickness of the cap rock exceeds the throwof any cross-cutting faults and is assigned a probabilityof 1. When both of these criteria are met, Cap RockGeometry is deemed to be sealing and j = 1. For all othervalues of Cap Rock Coverage and Cap Rock Thickness,the probability of sealing with respect to the Cap RockGeometry is evaluated using Equation 4 (i.e. j = Cap RockGeometry).

All variables associated with Cap Rock Coverage andCap Rock Thickness can be estimated from seismicinterpretation and standard mapping techniques.

Table 1.Uncertainty-probability matrix used to modify sealing probabilities, using subjective descriptions of data quality and quantity(after Rose, 2001; Nakanishi and Lang, 2001).

(maximum column height)Cap RockCapacity =

(volumetric trap height)(Equation 3)

Cap Rock Geometry = Cap Rock

Coveragex Cap Rock

Thickness(Equation 4)

(areal extent of cap rock) Cap Rock Coverage =

(areal extent of trap)(Equation 5)

(along-fault cap seal thickness) Cap Rock Thickness =

(seismically-resolvable fault throw) (Equation 6)

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Plentiful 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Enough 0.125 0.2 0.275 0.35 0.425 0.5 0.575 0.65 0.725 0.8 0.875

Moderate 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75

Poor 0.375 0.4 0.425 0.45 0.475 0.5 0.525 0.55 0.575 0.6 0.625

Dat

a Q

uant

ity

and

Qua

lity

Very Poor 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5

Risk associated with critical parameter

Mildren et al. 279

Cap Rock Integrity

Cap Rock Integrity is defined as the likelihood that the caprock is not cut by sub-seismic, open, natural fractures. Thisparameter is governed by the rheology of the cap rock andthe in situ stress environment. It is the relationship betweenthese two elements that dictates the likelihood of brittle failureoccurring within the cap rock (Mildren et al., 2002b).Measuring the likelihood of brittle failure is best illustratedusing a 3D Mohr diagram, incorporating a Griffith-Coulombfailure envelope representing the strength of the cap rock(Fig. 3a). The likelihood of brittle failure is measured as theminimum horizontal ‘distance’ between the failure envelopeand the circumference of the Mohr Circle. This measure isthe pressure change required to initiate brittle failure (∆P).The lower the ∆P value, the closer the Mohr circle is to thefailure envelope and the higher the risk of fracturing the caprock. If the cap rock has been subjected to a deformationhistory, pre-existing fractures may be ‘activated’ under thepresent-day stress conditions. The likelihood that thesefractures are hydraulically conductive may also be evaluatedusing this technique, albeit with a failure envelope indicativeof the pre-existing fracture strength (Mildren et al., 2002a).

Brittle failure estimates using this methodology havepreviously been calibrated using multiple traps in the TimorSea (Mildren et al., 2002a). This analysis suggests that for:

• ∆P ≤ 10 MPa, the trap is more likely to be leaking thannot;

• 10 ≤ ∆P ≤ 15 MPa, trap could be leaking or sealing;• ∆P ≥ 15 MPa, trap is more likely to be sealing than not.

This calibration can be used to define the probabilitythat the cap rock is not cut by sub-seismic natural fractures(k), such that:

Fault Plane Capacity

Fault Plane Capacity is defined as the likelihood that themembrane of the fault rock is sufficient to oppose thebuoyancy pressure of the hydrocarbon column. Fault rockprocesses that can occur during deformation affect thesealing capacity. These processes include: cataclastic grainreduction; clay smear development along the fault plane;and phyllosilicate/framework generated from impuresandstones and cement precipitation (Watts, 1987; Knipe,1997). Each process results in the reduction of pore-throatsize within the fault rock, and in turn increases the capillarythreshold pressure (Pc) and ultimately the Fault PlaneCapacity. Several techniques have been proposed forassessing membrane seals such as shale smear algorithmsor direct measurement of hydrocarbon capacities usingmercury injection capillary pressures (MICP). Themaximum column height of hydrocarbons that the faultrock material can sustain is used to qualitatively evaluateFault Plane Capacity relative to the considered trapvolume, such that:

Where the maximum column height of the fault rockexceeds the volumetric trap height being considered, FaultPlane Capacity is greater than 1 and considered to be sealing(i.e. a = 1). Where the Fault Plane Capacity is less than thevolumetric trap height, the probability of sealing is thefraction of the trap height that will be retained (i.e. a = FaultPlane Capacity).

Juxtaposition Lithology Capacity

Juxtaposition Lithology Capacity is defined as thelikelihood that lithologies juxtaposed against reservoir unitsacross trap-bounding faults have seal capacities sufficientto oppose the buoyancy pressure of the hydrocarboncolumn. Fault plane juxtaposition diagrams can be used toidentify potential leak windows where low capillarypressure (Pc) units, such as clean sandstones, juxtapose thereservoir units (Allan, 1989). In a similar manner to faultplane capacity, juxtaposition lithology capacity isquantitatively evaluated by measuring the maximumhydrocarbon column height that the juxtaposed lithologycan support relative to the volumetric trap height beingconsidered, such that:

∆P

τ

Cap rock failure envelope

σn'σ3' σ2' σ1'

∆P

τ

Fault rock failure envelope

σn'σ3' σ2' σ1'

Resolved stresses for particular fault

segment

Figure 3. Graphical presentation of FAST methodology for (a), initiating brittle failure in a cap rock; and (b), reactivating a pre-existingfault geometry. All fault/fracture orientations plot within the shaded area of the 3D Mohr circle, when incident stresses are resolved intoshear (τ) and effective normal stress (σn’). Delta P (∆P) is the pressure required for the fault/fracture of interest (point within the Mohrcircle) to intersect with the failure envelope.

1.00.9 } ∆P > 25 MPa

0.80.7 } 15 < ∆P ≤ 25 MPa

0.60.50.4

} 10 < ∆P ≤ 15 MPa

0.30.2 } 0 < ∆P ≤ 10 MPa

0.1

k =

Prob

abili

ty

0.0 } ∆P ≤ 0

(Equation 7)

(max. column height fault rock)Fault PlaneCapacity =

(volumetric trap height)(Equation 8)


Where the maximum column height exceeds theconsidered height of the trap, it is considered to be sealingand therefore the corresponding probability b = 1. Wherethe maximum column height is less than the consideredtrap height, the probability of sealing is the retained fraction(i.e. b = Juxtaposition Lithology Capacity).

Post-charge reactivation

Active faults and fractures can provide high permeabilityconduits for fluid flow during deformation. Previouslyformed fault juxtaposition or deformation processmembrane seals may be breached if the fault is reactivatedpost-charge. Similarly to Cap Rock Integrity, the risk ofpost-charge reactivation is dictated by the mechanicalproperties of the fault rock, the contemporary stressenvironment and the structural geometry of the fault itself.Using the FAST technique (Mildren et al., 2002a), thelikelihood of reactivation is estimated as the pressurechange (∆P) required to initiate brittle failure of the fault.

All fault and fracture orientations plot within the shadedarea of the 3D Mohr diagrams presented in Figure 3. Insteadof evaluating the closest point between the failure envelopeand the Mohr circle, as is done for cap rock integrity, therisk of reactivation is measured from the point within theMohr circle representing the particular fault geometry tothe failure envelope (Fig. 3b). The probability of sealing(no reactivation) is again based on the previously publishedTimor Sea ∆P calibration study (Mildren et al., 2002a) and∆P values are converted to sealing probabilities using thesame scheme devised for cap integrity such that:

Probability of trap integrity

The overall probability of a sealing trap is evaluated bycombining the six uncertainty-modified seal probabilitiesdetermined from the quantitative relationships. Combinationof these probabilities is based on the fundamental statementthat describes the requirements for a sealing trap (a fault-bounded trap is sealing if both the cap rock and the fault aresealing). Just as probability theory was used to establishquantitative relationships between the parameters for caprocks and faults, so too can these probabilities be combinedto evaluate the trap. The ‘and’ relationship requires both thecap rock and the faults to be sealing, therefore implying amultiplicative relationship, such that the probability that atrap seals hydrocarbons can be expressed as:

Ptrap = Pcap * Pfault (Equation 11)

By substituting Equations 1 and 2, Equation 11 becomes:

Ptrap = (i*j*k)*{1-[(1-a)*(1-b)]}*c. (Equation 12)

If a trap is not bounded by any faults, probabilities a, band c become 1.0, and the relationship becomes identical tothe probability for a sealing cap rock (Equation 2).

Multiple calculations of Ptrap are evaluated for various trapvolumes, to generate a volumetric distribution of sealingprobabilities to establish the geologic risk. Ptrap can then becombined with assessment of other elements of the playconcept, to determine the likelihood of economic andcommercial success.

The Skua Field

BHP Petroleum discovered the Skua Field in 1985 with thecompletion of Skua-2 (Osborne, 1990). Skua-1 had earlieridentified sandstone reservoirs in the Eocene, LateCretaceous and Early Jurassic (Osborne, 1990; Table 1).However, with the exception of minor shows in the LateCretaceous (Puffin Formation) and the late Paleocene(Johnson Formation), the commercial hydrocarbonaccumulation was restricted to the Early Jurassic (PloverFormation). Skua-2 intersected a 6 m residual hydrocarboncolumn in the Puffin Sandstone and 9 m gross (3 m net)column of moveable oil in the Plover Formation across theSkua Fault. The find was confirmed by Skua-3, whichintersected a 46.5 m oil column. A total of six wells weredrilled subsequent to Skua-3 in order to delineate theaccumulation and develop the field for production.

Swift-1 and Rowan-1 were also drilled in the Skua area,southeast and northwest of the Skua tilted fault blockrespectively, to test adjacent tilted fault blocks. Both wellswere plugged and abandoned, although Swift-1 intersectedan 8.5 m gross (4.6 m net) oil column in the Late JurassicLower Vulcan Formation, and in Rowan-1, fluorescence wasidentified in the Paleocene Johnson Formation, Late JurassicLower Vulcan Formation and Early Jurassic PloverFormation sandstones.

Structural elements

The Skua area is located within the Bonaparte Basin onthe North West Shelf of Australia in the southeast portionof the Timor Sea (Fig. 4). Several phases of deformationhave contributed to the structurally complex architectureof the Timor Sea region. The northwest Palaeozoicstructural grain, exemplified by the Petrel Sub-basin, wasoverprinted by northeast-trending features during Middleto Late Jurassic rifting. Late Miocene collision of thenorthern Indo-Australian Plate margin with the southeastAsian Plate generated a fault system that is marginallydivergent from the older Jurassic fault trends. Reactivatedlate Palaeozoic basement faults trending north-northwestand north have also been observed at the base Cretaceouslevel (O’Brien et al., 1993; Cowley and O’Brien, 2000;Gartrell et al., 2002).

The Skua, Rowan and Swift traps are located on LateJurassic, northeast-trending tilted fault blocks that are part

(max. column height of leastsealing juxtaposed lithology)

JuxtapositionLithologyCapacity

=(volumetric trap height)

(Equation 9)

1.00.9 } ∆P > 25 MPa

0.80.7 } 15 < ∆P ≤ 25 MPa

0.60.50.4

} 10 < ∆P ≤ 15 MPa

0.30.2 } 0 < ∆P ≤ 10 MPa

0.1

c =

Pro

babi

lity

0.0 } ∆P ≤ 0

(Equation 10)

Mildren et al. 281

of a series of elongate northeast-trending horsts, grabens andhalf-grabens along the southeastern margin of the Vulcan Sub-basin (Patillo and Nicholls, 1990; Woods, 1994; Fig. 5). TheJurassic faults associated with each of these structures dipto the northwest approximately 50° within the CretaceousJamieson Formation, and flatten out with depth (Rowan Fault,Skua Fault, Spruce Fault and Swift Fault; Fig. 6). A series ofpost-rift faults are superimposed on these Jurassic structures

in the Skua area. Gartrell et al. (2002) observed a set of right-stepping en echelon normal faults above the Rowan Fault thatcross-cut the late Miocene reflector (RF1, RF2, RF3, RF4,RF5 and RF6; Fig. 6). A similar set of en echelon normalfaults were observed above the Skua Fault. However, onefault (SF1) terminates just above the early Miocene reflectorand two (SF2 and SF3) terminate just below and have beenattributed to an earlier late Eocene reactivation event (Gartrell

Rainbow 1

Tahbilk 1

Willeroo 1

Talbot 2

Yering 1

Snipe 1

Kimberley 1

Woodbine 1

Maret 1

AC/P20

Langhorne 1

Yarra 1

AC/P22

Puffin 3

Grebe 1

Champagny 1

Prion 1

Pascal 1

3

86

7/7a

2

5Birch 1

Rowan 1

AC/P27

AC/P26

Anson 1

Anderdon 1

Katers 1

Conway 1

AC/RL1

Brontosaurus 1

AC/RL3

49

AC/P21Allaru 1

Great Eastern 1

Caversham 1Rothbury 1

Eclipse 1

East Swan 2East Swan 1

Eclipse 2

Parry 1

Pituri 1

Lucas 1

Puffin 4

Longleat 1

Taltarni 1

Leeuwin 1

AC/P31

AC/P4Part 1(2)

AC/P4TL

Padthaway 1

Elasmosaurus 1

AC/P32

Vulcan 1,1a,1b

Hadrosaurus 1

Keeling 1

Swift 1

Skua 1

Bilyara 1

Talbot 1

Puffin 2

Puffin 6

Puffin 5

Puffin 1

Swan 3/3aSwan 1Swan 2

Paqualin 1

Tenacious 1/ST1Tenacious West 1/ST1

Elm 1

Anson North 1

AC03-2

Maple 1

Cash 1

CANNING

BASIN

BROWSE

BASIN

KIMBERLEY

BLOCK

BONAPARTE

BASIN

Timor

Sahul

Platform

VULC

AN S

UB-B

ASIN

Ashmore

Platform

Lond

onde

rry

Hig

h

PETREL

SUB-BASIN

Timor

Sea

SpruceProspect

SWAN GRABEN

Swift-1

SW

IFT

FAULT

BLO

CK

SKUA F

AULT

BLOCK

RO

WAN

FAU

LT

BLOCK

0 2

km

SKUA F

IELD

Rowan-1

A

A'

SW

AN

GRABEN

VULCAN

SUB-BASIN

SKUA

TRO

UG

H

Figure 4. Location of the Skua trap withinthe Bonaparte Basin in the Timor Sea.Petroleum exploration leases and adjacentwells are shown for reference. Section line A-A’is illustrated in Figure 5.

Figure 5. Structural cross-section A-A’ across the southeastern margin of the Swan Graben (after Fittal and Cowley, 1992). Section illustrates thetilted fault block constituting the Skua trap. Jurassic reservoir (Plover Formation) is highlighted in green and Cretaceous sealing package(Woolaston, Gibson, Fenelon formations and base Puffin claystones) are shown in blue. Refer to section location in Figure 4.

SWAN GRABEN

Puffin Formation

Intra-Val. D/C

Tithonian U/CU. Vulcan Fm.

L. Vulcan Fm.

Callovian U/C

Top Cretaceous

Top Palaeocene

Puffin Formation

L. VulcanFm.

Callovian U/C

0 2km5000

4000

3000

2000

1000

0 m

A'ARowan-1st Skua-1 Swift-1

SKUA FIELD

MONTARA TERRACE

Top Jamieson Formation

Top Jamieson Formation

Plover Formation


et al., 2002). The post-rift faults are marginally oblique to thestrike of the Jurassic faults and are characterised by steeperdip magnitudes.

Stratigraphy

The Early Jurassic Plover Formation is recognised as thereservoir for the commercial hydrocarbon accumulation atSkua. Reservoir quality varies from very good to poor, withpatchy occurrences of low porosity and permeability zonescomprised of large quantities of detrital silt and clay (BHPPetroleum, 1992). Extensive erosion during the Callovianand intra-Kimmeridgian tectonic events exposed thesediments of the Early Jurassic. On the Skua high, the intra-Valanginian and intra-Kimmeridgian unconformities arecoincident and merge with the Callovian unconformity(Osborne, 1990; Fig. 5). The ‘Callovian et al.’ unconformitymarks the top of the reservoir over the Skua structure.

The Skua cap rock is a package of Cretaceoussediments predominantly consisting of the Woolaston,Gibson and Fenelon formations (WGF). They form a thick,hemipelagic slope depositional sequence of marls andcacilutites (MacDaniel, 1988; Pattillo and Nicholls, 1990).Disconformably underlying the WGF are calcareousclaystones of the Jamieson Formation, which is generallyless than 10 m thick and which is absent over the Skuatilted fault block itself. The WGF is overlain by calcareousclaystones that comprise the base of the Puffin Formation.The WGF and the claystone packages above and beloware considered the cap rock package at the Skua trap andthese sediments as a whole will be referred to as the WGFin this paper (Fig. 5).

Empirical evidence for leakage

A significant amount of empirical evidence for hydrocarbonleakage has been documented across the Timor Sea regionover the last 10 years (Martin and Cawley, 1991; O’Brienet al., 1992; O’Brien and Woods, 1995; O’Brien et al.,1998). Present-day hydrocarbon leakage has been identifiedusing various remote survey techniques: Airborne LaserFluorescence (ALF), geochemical sniffer and Synthetic

Aperture Radar (SAR). Historical hydrocarbon leakage hasbeen identified by mapping acoustic anomalies termedHydrocarbon-Related Diagenetic Zones (HRDZs; O’Brienand Woods, 1995; Cowley and O’Brien, 2000) and fluidinclusion analyses (Lisk and Eadington, 1994; Lisk et al.,1998).

ALF anomalies are observed in association with the SkuaFault and there also exists a relatively large anomaly northof Swift-1, associated with the Spruce prospect (Fig. 7).Sniffer surveys indicate the presence of methane and ethaneconcentrations adjacent to the Skua Fault and in surroundingareas. SAR reveals several ocean surface slicks towards thesouthern limit of Skua and adjacent to Spruce. All evidencesuggests active hydrocarbon leakage associated with theSkua and Spruce structures (O’Brien and Woods, 1995).Cowley and O’Brien (2000) observed three HRDZsassociated with the Skua structure that appear to formpreferentially at the intersection of the Palaeozoic and

Figure 6. East-looking section of Skua areausing interpretation of the Skua HV11 3Dseismic survey (Gartrell et al., 2002).Jurassic rift faults are coloured purple, lateMiocene post-rift faults are yellow, earlyMiocene post-rift faults are red and lateEocene post-rift faults are green. Comparesection to Figure 5 for reference.

Figure 7. Empirical evidence for hydrocarbon leakage in the Skuaarea (O’Brien et al., 2001). HRDZs are delineated by faster (red)two-way times (TWT).

Mildren et al. 283

Jurassic fault systems adjacent to Skua-4 and Skua-7A. TheSpruce Fault is also associated with HRDZs and the SwiftFault is associated with the most intense and largest in thearea. No HRDZs are evident adjacent to Rowan-1ST.

GOI (grains containing oil-filled inclusions) fluidinclusion analyses are available for six wells in the Skua area:Skua-3, Skua-4, Skua-6, Skua-8, Skua-9 and Swift-1 (Lisket al., 1998; Gartrell et al., 2002). These data indicate asystematic change in residual hydrocarbon column height,progressively increasing in wells towards the southwestmargin of the Skua Field (Gartrell et al., 2002), with amaximum residual column height of approx 10 m. Analysisof Swift-1 indicates the presence of a 17 m palaeo-hydrocarbon column (Lisk et al., 1998). Gartrell et al.

(2002) interpreted the systematic change across the Skuatrap as evidence for a tilted palaeo-oil–water contact, asoriginally suggested by Lisk et al. (1998). Thisinterpretation considerably reduces the apparent leakagefrom the Skua trap and suggests that little leakage hasoccurred since the trap was charged. However, as shownby Gartrell et al. (2002), conventional oil shows incuttings, core and sidewall cores are much moreextensive beneath the GOI anomalies. Although Gartrellet al. (2002) dismissed the conventional shows as beingless reliable than the GOI, the conventional shows couldbe indicative of a more recent palaeo-oil column andindicative of much more oil leakage from Skua thanpredicted by GOI.

Skua stress tensor

The Skua area in situ stress tensor has been determined inorder to evaluate two of the six key parameters affecting trapintegrity: cap rock integrity and post-charge reactivation. Allstress magnitude data (including pore pressure) werecalculated to provide stress gradients accurate at trap level(ca. 2,200 m).

Stress orientation

Seven Stratigraphic High-Resolution Dipmeter Tool(SHDT) logs were interpreted for borehole breakouts todetermine the in situ stress orientation across the Skuaarea (Table 2, Figs 8 and 9). A total of 24 boreholebreakouts were observed over 3,543.7 m of caliper logwith a cumulative length of 304.4 m. Mean maximumhorizontal stress orientations (σHmax) were rankedaccording to the World Stress Map ranking scheme(Zoback, 1992). On a scale of A to E, where A is thehighest quality, only three wells exhibited a mean with aranking of D or greater (Rowan-1, Skua-5 and Swift-1).These wells were used to determine an unweighted σHmax

mean orientation of 058°N to be used in the faultreactivation calculations. No borehole breakouts wereobserved within Skua-6 and Skua-9ST.

(a) (b)

Unweighted Mean = 058 N Standard Deviation = 18 Breakouts

Figure 8. Summary of breakout-derived σHmax orientations forwells in the Skua area with aquality ranking D or greater(i.e., Rowan-1, Skua-5 andSwift-1); (a) Rose strikediagram of breakout derivedσHmax orientations. Maximumpetal length is 4 breakouts; (b)Depth-azimuth plot showingorientation of σHmax with depthand breakout lengths.

Rowan-1

Skua-5

Swift-1

Skua-6

Skua-1

Skua-9

Skua-7ASkua-4

Skua-3Skua-8

Skua-2

σHmax orientation (breakout)

Quality:ABCD

LEGEND

no stress indicators observed

no data

Skua Fault

N0 500 1000

Metres

▲

Figure 9. Map of breakout-derived σHmax orientations for wells in theSkua area with a quality ranking D or greater. Wells where no breakoutswere interpreted are also included. Long axis of stress indicators isproportional to World Stress Map quality ranking (Zoback, 1992).

N Azi SD Q ? L Azi SD Q Azi SD Q

Rowan-1 12°29'53.51 124°23'36.96 8 063 8.6 B 114.4 060 7.3 B 062 8.2 B

Skua-2 12°30'34.56 124°24'14.98 10 044 49.0 E 151.1 012 57.0 E 019 66.3 E

Skua-4 12°29'35.52 124°25'32.88 1 066 0.0 E 4.5 066 0.0 E 066 0.0 E

Skua-5 12°28'26.11 124°26'37.20 2 053 10.1 D 21.0 048 8.5 D 060 7.2 D

Skua-6 12°29'15.88 124°26'18.85 No Breakouts Interpreted

Skua-9ST 12°29'52.58 124°25'09.85 No Breakouts Interpreted

Swift-1 12°32'14.28 124°27'05.40 3 039 33.4 D 13.4 048 22.9 D 046 26.0 D

0

1

2

3

Dep

th (k

m)

20 40 60 80 1000Pressure (MPa)

Figure 10. Principal stress magnitude gradients [mudweight (—), porepressure (—),σhmin (—), σv (—) and σHmax (—) at 2,200 m] derivedfrom empirical data [LOPs (•), DST (x), Mud weight (+++++) and verticalstress (+++++)] at approximate depth of the reservoir-seal interface.

Table 2. Breakout derived mean σHmax orientations. Magnetic declination of 3° has been applied. N is the total number of breakouts and ΣL thetotal length (m) of breakouts in the well. Azi and SD are the mean strike (000–180°N) of breakout-derived σHmax orientations in the well and thestandard deviation in degrees as determined by directional statistics. Q is the quality rating of the measurement, determined using the WorldStress Map ranking system (Zoback, 1992). Ecc (eccentricity) is the maximum difference between orthogonal calliper pairs.

Well Name Latitude Longitude Un-weighted


Stress magnitudes

The weight of the overburden is assumed to be equal to thevertical stress magnitude (McGarr and Gay, 1978). It iscalculated by integrating density logs acquired fromhydrocarbon exploration wells according to the followingrelationship:

σv = ∫o

z ρ(z)g.dz

where:

ρ = density of the overlying rock column [gcm-3];z = depth [m];g = acceleration due to gravity [ms-2].

Density logs are not normally run from the surface;therefore, vertical stress cannot be determined by simplyintegrating the density log data from the surface to the depthof interest. However, sonic velocity and density are stronglyrelated. Check-shot velocity survey data are used to determineaverage sonic velocity from the surface to the top of thedensity log data, and the average velocities are then convertedto average densities. See Serra (1984) for details of theoperation of sonic and density logs, and Balch and Lee (1984)for the operation of check shot surveys. Vertical stress profileswere generated for Skua-4 and Swift-1 and are shown inFigure 10.

Derivation of the minimum horizontal stress magnitude(σhmin) is based on the assumption that lower-bound leak-off pressures are an estimate of fracture opening pressure(Breckels and Eeklen, 1982). The fracture opening pressurein turn estimates the instantaneous shut-in pressure andtherefore estimates σhmin. Four leak-off measurements wereavailable to constrain the σhmin gradient in the Skua area(Fig. 10).

The maximum horizontal stress magnitude is the mostdifficult component of the stress tensor to define. Boreholestresses can be modeled to replicate the occurrence, or non-occurrence of deformation observed at the borehole wall,such as borehole breakout and drilling-induced tensilefractures (DITF). A small number of breakouts wereidentified on SHDT logs in several of the Skua area wells.Mildren et al. (1994) identified DITF on FormationMicroScanner images from wells adjacent to the Skua area,implying that their presence at Skua is possible. Wellbore

stresses were modeled to determine σHmax magnitude basedon breakout occurrence and the likelihood of DITFoccurrence (Fig. 10).

Pore pressures are directly measured in situ by formationtests. Drill-stem test data were available from three wells inthe Skua area (Rowan-1, Skua-5 and Swift-1) each indicatinga consistent pressure gradient (Fig. 10).

The stress magnitude estimates indicate that the stressregime in the Skua area is strike-slip, such that σhmin < σv <σHmax. This result is consistent with regional estimates of thein situ stress field across the Bonaparte Basin (Hillis et al.,1997; Castillo et al., 1998; Mildren and Hillis, 2000; Mildren

0

1

2

3

Dep

th (k

m)

20 40 60 80 1000Pressure (MPa)

Figure 10. Principal stress magnitude gradients [mudweight (—), porepressure (—),σhmin (—), σv (—) and σHmax (—) at 2,200 m] derivedfrom empirical data [LOPs (•), DST (x), Mud weight (+++++) and verticalstress (+++++)] at approximate depth of the reservoir-seal interface.

total length (m) of breakouts in the well. Azi and SD are the mean strike (000–180°N) of breakout-derived σHmax orientations in the well and thestandard deviation in degrees as determined by directional statistics. Q is the quality rating of the measurement, determined using the WorldStress Map ranking system (Zoback, 1992). Ecc (eccentricity) is the maximum difference between orthogonal calliper pairs.

Well Name Latitude Longitude Un-weighted Length-weighted Ecc-weighted

Mildren et al. 285

et al., 2002a). A summary of all stress magnitude gradientsused to evaluate fault reactivation and cap rock integrity atSkua are listed in Table 3.

Skua trap risk

A probability distribution was generated for the Skua trap,by iterating the integrated work flow (Fig. 1) for a variety oftrap volumes. Sealing probabilities were generated for eachtrap volume (0–100% of structural closure) and modifiedaccording to uncertainty. However, although the probabilitydistribution for all trap volumes was calculated, calculationsare only shown for trap volumes of 100%, 10% and 1%.Where optimal data were unavailable to evaluate a parameter,an alternate method or logical assumption was used togenerate a probability, and the appropriate uncertaintymodifier was applied. This highlights the flexibility of theapproach in the real world, where data availability varies,and also illustrates the independence of the method to specificevaluation methodologies.

Cap rock seal parameters

The geometry of the Skua trap was evaluated from aninterpretation of the depth-converted HV11 3D seismicsurvey. The ‘Callovian et al’ unconformity was interpretedas the interface between the top of the Jurassic PloverFormation reservoir and the Cretaceous WGF cap rocksediment package. Maximum trap closure is defined by the2,300 m contour on the depth-converted, top Plover horizon,indicating 129 m from crest (2,171 m) to spill point (Fig. 11).Volumetric calculations indicate that 10% trap volume isequivalent to a column height of 67 m and 1% trap volume isequivalent to 29 m.

Mercury injection capillary pressures (MICP) have beenmeasured from samples of the WGF at five well locationsacross the Skua structure. Seal capacities, calculated fromcapillary pressure measurements (maximum hydrocarboncolumn heights that can be retained), are dependant on manyparameters, such as hydrocarbon phase, oil and gas densities,sub-surface pressure and temperature conditions, interfacialtension and contact angle. Maximum column heights were

calculated for eight samples, using Skua reservoir conditionsand measured oil and gas densities (Table 4). Upper and lowerbound estimates were made for each sample, by varying theinterfacial tension and contact angle (Fig. 12). Althoughcalculations were made for oil and gas phases, seal capacityat Skua has been evaluated assuming oil phase only. Oil phaseseal capacities vary considerably across the Skua trap,ranging from 69 m at Skua-5 to 296 m at Skua-1. Cap RockCapacity was evaluated using Equation 3 for the 100%, 10%and 1% trap volumes and minimum mean oil phase column

Table 3. Summary of stress magnitudes calculated for 2,200 mdepth in the Skua area and corresponding stress magnitudegradients used to evaluate brittle failure.

Table 4. Reservoir conditions at Skua used toconvert capillary threshold pressures, derived fromMICP analyses, to maximum hydrocarbon columnheights displayed in Figure 12. Note that reservoirconditions at Skua-3 were used to evaluate thesample taken at Skua-1.

2250

2350

2500

SKUA-1SKUA-1

SKUA-2SKUA-2

SKUA-6SKUA-6

SKUA-5SKUA-5

SKUA-4SKUA-4

SKUA-3SKUA-3SKUA-8SKUA-8

SWIFT-1SWIFT-1

ROWAN-1ROWAN-1

652000

8620000

12 32 00 S

12 30 00 S

12 28 00 S

124 24 00 E 124 26 00 E 124 28 00 E

2400

2450

2450

2300

2375

654000 656000 658000

8618000

8622000

8616000

8614000

N▲

Figure 11. Depth contour map at the ‘Callovian et al’ unconformity,representing the top of the Jurassic reservoir at Skua. Dashed whiteline delineates trap closure.

115

296

96

69

125

174

108

139

79

203

6650

89

124

7799

0

50

100

150

200

250

300

350

400

450

500

Skua-1 Skua-1 Skua-3 Skua-5 Skua-8 Skua-8 Skua-9 Skua-9

Col

umn

Hei

ght (

m)

Figure 12. Maximum hydrocarbon column heights derived frommercury injection capillary pressures and Skua reservoir conditionsof WGF samples across the Skua Field. See Table 4 for reservoirconditions. Blue and red markers are column heights determinedfor oil and gas phases, respectively. Upper and lower boundestimates for each sample were generated by varying interfacialtension and contact angle.

Well Depth Reservoir Temp. Gas Gravity Salinity API

Stress Component Magnitude

(MPa)

Gradient

(MPa/m)

Pore Pressure (Pp) 22 0.010

Vertical Stress (σv) 34 0.016

Minimum Horizontal Stress (σhmin) 49 0.022

Maximum Horizontal Stress (σHmax) 68 0.031

(m) Pressure (MPa) (°C) (g/cm3) (ppm) gravity (°)

Skua-1 2,288.7 22.92 96.7 0.71 100,000 43Skua-3 2,288.7 22.92 96.7 0.71 100,000 43

Skua-5 2,339.2 23.43 88 0.775 110,000 39.9

Skua-8 2,286.5 22.90 100 0.74 110,000 42.2

Skua-9 2,286.5 22.90 95.8 0.71 110,000 43.7

Considered Max. Column Cap Rock Probability Quality & ModifiedVolume (%) Height Capacity (i) Quantity Probability (i)

100 69 0.53 0.535 Enough 0.526

10 108 1.61 1.00 Moderate 0.750

1 108 3.72 1.00 Moderate 0.750

Considered Cap Rock Cap Rock Probability Quality & ModifiedVolume (%) Coverage Thickness (j) Quantity Probability (j)

100 1 0.9 0.9 Plentiful 0.910 1 0.9 0.9 Plentiful 0.9

1 1 0.9 0.9 Plentiful 0.9


heights relevant to the trap closure being considered(Table 5). The two samples obtained from Skua-9 were usedto define the maximum column height for trap volumesbetween 0–17% (2,250 m contour) and all MICP sampleswere used to evaluate maximum column height for trapvolumes between 17–100% (2,300 m contour).

Although MICP analyses were performed on a relativelyhigh number of samples, the uncertainty associated with thecolumn height calculations restricts the data quality andquality description to ‘enough’. The uncertainty modifierslisted in Table 1 were used to modify the sealing probabilitiesaccordingly (Table 5).

The WGF package was mapped across the Skua area, toassess its areal extent and thickness distribution (Figs 11and 13). The minimum thickness of the WGF is 90 m withinthe 100% closure of the Skua trap, and therefore, Cap RockCoverage (Equation 5) exceeds 1.0 for all trap volumes.Along the northwest fault-bounded margin of the trap, thethickness of the cap rock is 90 m or greater. Fault throw atthe level of the Cretaceous sediments lies between 50 and100 m, and therefore, using a maximum fault throw of100 m indicates that Cap Rock Thickness (Equation 6) isless than 1 (Table 1). The data uncertainty is minimal andthe quality and quantity is considered to be ‘plentiful’(Table 6).

The in situ stress tensor at the depth of the cap rock(2,200 m) is used to evaluate the pressure required to initiatetop seal failure using a modified FAST technique (Table 3;Mildren et al., 2002a; Mildren et al., 2002b). Although thestress tensor is well defined by local data, there is littleinformation regarding the strength of the seal. In the absenceof any physical strength information, upper and lower boundcalcareous shale and marl strengths were used to assess thelikelihood of brittle failure (fracture generation) within thecap rock (coefficient of friction, µ = 0.3–0.5; cohesivestrength, Co = 2–10). Figure 14 illustrates the ∆P valuesevaluated for all fracture orientations using the end memberstrength scenarios. In the lower bound case, a large rangeof fracture orientations correspond to ∆P values of 0 Mpa,or less. Only steeply-dipping fracture orientations (greaterthan 70°) correspond with ∆P values greater than 15 MPa.Intuitively, the upper bound strength scenario is less inclinedto failure with ∆P values greater than 12 MPa.

The relationship presented in Equation 10 is used toconvert the minimum (highest risk) ∆P to a probability ofsealing. The minimum ∆P observed within the lower boundstrength scenario corresponds to a sealing probability of 0.15and a probability of 0.5 for the upper bound scenario. Theprobability of sealing corresponding to Cap Rock Integrityis taken midway between these two end members (0.325)

Figure 13. Isopach map generated for the WGF sediment packagecomprising the seal of the Skua trap. Dashed white line delineatestrap closure.

50

60

70

80

90

100

110

120

130

140

150

160

170

180

190

200

210

220

230

240

250

260

Thic

kness (

m)

▲N

Table 5. Parameters used to evaluate Cap RockCapacity, the corresponding sealingprobability (i), the subjective description ofdata quality and quantity, and the modifiedsealing probability for trap volumes of 100%,10% and 1%, respectively.

Cap Rock Capacity

652000

8620000

12 32 00 S

12 30 00 S

12 28 00 S

124 24 00 E 124 26 00 E 124 28 00 E

100125

150

654000 656000 658000

8618000

8622000

8616000

8614000175

175200

150125

180

090

0030

060

120

150210

240

270

300

360

180

090

0030

060

120

150210

240

270

300

360

µ = 0.3Co = 2 MPa

µ = 0.5Co = 10 MPa

Lower-bound Cap Rock Strength Upper-bound Cap Rock Strength

∆P (MPa)0 10 15 25

Increasing risk of brittle failure

Figure 14. Polar diagrams showing the likelihood of fracturegeneration in the cap rock for lower and upper bound strengthscenarios. All points are poles to planes on a southern hemisphereprojection i.e. horizontal plane plots at centre of plot and verticalfracture striking north–south will plot on the boundary at either090° or 270°.

Cap Rock Geometry

Table 6. Parameters used to evaluate CapRock Geometry, the corresponding sealingprobability (j), the subjective description ofdata quality and quantity, and the modifiedsealing probability for trap volumes of100%, 10% and 1%, respectively.

Mildren et al. 287

Estimation of the fault seal capacity at Skua using SGRis problematic below the Callovian unconformity. Faultdisplacement is difficult to constrain and, furthermore,Jurassic (–Cretaceous) displacements were syn-sedimentary, which resulted in a great thickness of theshale-prone Lower Vulcan Formation being deposited inthe hanging-wall (Fig. 5). However, subsequent to, andalso possibly during this period, the foot-wall wasexposed and eroded, removing Jurassic–Cretaceousstrata, including the Vulcan Formation shale. Here, weconsider a notional fault that self-juxtaposes the Ploverreservoir and calculate SGR values using percent shaledata for the foot-wall (Fig. 16). Therefore, we are usingfault displacements that are far too low (100 m vs. km-scale) and the lowest possible percent shale. If, for this

and a ‘moderate’ uncertainty modifier is applied to all trapvolumes (Table 7).

Fault seal parameters

The Skua trap is bounded by a northeast-trending, northward-dipping fault along its northwestern margin. The fault wasinterpreted from the HDV11 seismic survey and althoughthe confidence of the fault plane interpretation is high, theresolution of the hanging-wall Jurassic sedimentscompromises the evaluation of fault throw. A detailed lookat the Skua Fault geometry indicates that the trap is onlybounded below the 2,230 m contour (Fig. 15). The upper 7%of the trap is bounded by the cap rock only and thereforeonly the cap rock seal parameters need be determined fortrap volumes between 0 and 7% (i.e. fault seal probabilitiesare set to 1 (Fig. 1).

Fault plane capacity is evaluated by assessing the sealingcapacity of the fault rock material. Skua-2 is the only wellin the area that intersects the Skua Fault. However, samplesof the fault rock material were of insufficient integrity toperform MICP analyses. In lieu of the MICP data,juxtaposition diagrams mapping SGR were used to predictthe sealing capacity of the fault, based on fault displacementmagnitude (Yielding et al. 1997).

Table 7. The corresponding sealing probability (k) for Cap RockIntegrity, the subjective description of data quality and quantity,and the modified sealing probability for trap volumes of 100%,10% and 1%, respectively.

Cap Rock Integrity

Considered Probability Quality & ModifiedVolume (%) (k) Quantity Probability (k)

100 0.325 Moderate 0.413

10 0.325 Moderate 0.413

1 0.325 Moderate 0.413

Limit of bounding Fault

Figure 15. Northeast-facing view of the Skua trap topography at the top of the Jurassic reservoir. Line of sight inset illustrates that trap isindependent of the rift fault (northwest margin of structure) above the 2,230 m depth contour.

2800

2750

2700

2650

2600

2550

2500

2450

2400

%Shale

0 100 0 50 100 150 200 250 300 350 400 450

Dep

th (m

)

Fault Throw (m)

0.0 to 0.10.1 to 0.20.2 to 0.30.3 to 0.4

0.9 to 1.0

0.4 to 0.50.5 to 0.60.6 to 0.70.7 to 0.80.8 to 0.9

100 m Fault Throw

Shale GougeRatio2363

Plover Formation Self Juxtaposition (Skua-6)

Figure 16. Juxtaposition triangle for self-juxtaposed Plover Formationreservoir at Skua-6. Note that above 100 m-throw SGRs > 0.20.Badleys, U.K. are thanked for provision of their Triangle software.


scenario, the results suggest that the fault may be sealing,we can expect that, given the likely juxtaposition seal anddisplacement of the higher shale content hanging-wall, thefault will in all probability be sealing to across fault flow.

Figure 16 shows the area of self-juxtaposition of thePlover reservoir at Skua-6. It is apparent that above relativelysmall throw values (ca. 100 m), compared to the expectedsize of the Skua Fault, SGR values are high and the faultrock material is expected to form a seal. Identical analysesfor other Skua wells yield similar results and we note that thegamma to % shale conversion used was conservative,compared to more ‘conventional’ conversion factors (e.g. forNorth Sea wells).

A conservative fault throw estimate of 100 m implies themajority of SGR values are between 0.4 and 1.0. SGR valuesof 0.2 or less are calculated at the base of the Plover Formationfor fault throws less than 120 m (Fig. 16). Cross-fault pressuredata has been used by Bretan et al. (2003) to calibrate SGRvalues to fault zone capillary entry pressures. For SGR valuesof 0.2, using Skua reservoir oil and gas densities, a maximumcolumn height of 47 m is predicted. This conservativeestimate is used as the maximum column height to estimateFault Plane Capacity for the range of trap volumes between7% and 100% (Table 8).

As a cautionary note, the fact that the fault has beenexposed and eroded during the Jurassic–Cretaceous may havean impact on seal integrity. Furthermore, the nature of thedetailed across-fault juxtapositions at the Callovianunconformity is unresolved, both stratigraphically andstructurally. This area of the fault surface constitutes a highdegree of uncertainty in terms of seal integrity, because post-unconformity fault reactivation (Cretaceous and Mio–Pliocene) will have resulted in juxtaposition of the Ploverreservoir against the unconformity itself and overlying units.Therefore, these probabilities are modified accordingly for a‘poor’ quality dataset (Table 1).

The juxtaposition fault seal analysis of the Skua Faultis hindered by two main factors. Firstly, accuratestratigraphic mapping in the hanging-wall is hindered bypoor seismic resolution and a lack of well control.Secondly, mapping fault displacements within the Jurassicsediments is difficult. The pre-Cretaceous displacementson the Skua Fault are large (km-scale) and much greaterthan the thickness of the reservoir (ca. 100 m-scale), whichresults in the Plover reservoir being juxtaposed againstdominantly shale-prone, non-reservoir Lower and UpperVulcan formations that should provide a juxtapositionseal (Fig. 16).

No samples of the Lower Vulcan Formation wereobtained from the Skua area wells for MICP analyses as apart of this study. However, a regional seal study by Kivioret al (2002) presented two maximum hydrocarbon columnheights, derived from Lower Vulcan Formation sealcapacities obtained from Montara-1 (13 m) and Oliver-1(563 m). The discrepancy between column heights atMontara and Oliver is attributed to the distribution of lateOxfordian to Kimmeridgian restricted marine claystonefacies, which are not present on the Ashmore Platform,Londonderry High and intra-graben highs within theVulcan Sub-basin (i.e. Montara and Skua structures (Kivioret al., 2002). A conservative maximum hydrocarbon columnheight estimate of 13 m is used to calculate JuxtapositionSeal Capacity and the corresponding sealing probabilities(Table 9; Equation 9). The large uncertainty and lack ofdata imply a ‘poor‘ dataset and the sealing probabilitieshave been modified accordingly (Table 1).

The likelihood of fault reactivation was previouslycalculated for the Skua Fault, using 2D structural data and ageneralised stress tensor (Mildren et al., 2002a). In order toevaluate post-charge reactivation for this study, the FASTmethodology has been applied to a 3D representation of theSkua Fault plane, with the use of a stress tensor determinedspecifically for the Skua Field. Without the advantage ofmechanical strength tests on Skua fault rock, a failureenvelope identical to the one used by Mildren et al. (2002a)has been implemented (cohesive strength = 5 MPa andcoefficient of friction = 0.7).

The FAST model shows a distinct difference in faultreactivation risk associated with the Mesozoic rift faults (Skua,Rowan, Spruce and Swift), relative to the post-rift faults. TheMesozoic rift faults are characterised by large

∆P values, predominantly greater than 20 MPa, but alsoas low as approximately 16 MPa (Figs 17 and 18). Thelowest ∆P values (higher reactivation risk) are associated withsteeper, shallower sections of the rift phase fault planes inthe vicinity of the cap rock base.

The Tertiary post-rift faults have ∆P values rangingbetween 0 and 25 MPa, which are distinctly higher risksthan the Mesozoic rift faults (Figs 17 and 18). ∆P valuesassociated with these faults are also depth dependent,ranging from 12–20 MPa at the base of the seal to 0 MPaabove the Johnson Formation. The Tertiary faults thatpropagate to the shallowest depths also correspond withthe lowest ∆P; i.e. the most recently activated faults areassociated with the highest risk. This predictioncorresponds with hydrocarbon shows observed withinTertiary reservoir targets, bounded by the post-rift faultsat Skua-2, -3 and -6.

Table 8. The corresponding sealing probability (a) for Fault PlaneCapacity, the subjective description of data quality and quantity,and the modified sealing probability for trap volumes of 100%,10% and 1%, respectively.

Fault Plane Capacity

Considered Probability Quality & ModifiedVolume (%) (a) Quantity Probability (a)

100 0.80 Poor 0.575

10 0.80 Poor 0.575

1 1.00 n/a 1.00

Table 9. The corresponding sealing probability (b) for JuxtapositionLithology Capacity, the subjective description of data quality andquantity, and the modified sealing probability for trap volumes of100%, 10% and 1%, respectively.

Juxtaposition Lithology Capacity

Considered Probability Quality & ModifiedVolume (%) (b) Quantity Probability (b)

100 0.10 Poor 0.40

10 0.19 Poor 0.42

1 1.00 n/a 1.00

Mildren et al. 289

Areas of low ∆P (12 MPa) are associated with the post-rift faults at the base of the seal Skua seal (Fig. 19). Thesezones are higher risk than the rift faults with which theyintersect and coalesce. Contemporary reactivation of theTertiary post-rift faults is therefore predicted to be a greaterrisk to trap integrity at Skua than the reactivation of theJurassic rift faults. Although the post-rift faults do not boundthe Skua structure, it is apparent that they cut through thesealing cap rock units (WGF package).

Brittle failure of the cap rock was considered, whenevaluating the Cap Rock Integrity parameter. However, thestrength of the cap rock is presumed to be weaker than the

fault rock. Therefore, new fractures are likely to be initiated inthe cap rock prior to fault reactivation. The probability thatthe post-rift faults compromise trap integrity is thereforesuperseded by cap rock fracturing and is automaticallyconsidered when evaluating the Cap Rock Integrity parameter.The Post-Charge Reactivation parameter is evaluated for thebounding Skua Fault and used to determine thecorresponding sealing probability (c).

The minimum ∆P of the Skua Fault (16 MPa) is convertedto a sealing probability using Equation 10 (Table 10). Thein situ stress tensor and structural geometry of the trap arewell defined. However, fault rock failure envelopes are basedon literature-derived strength estimates and, therefore, thesealing probability corresponding to Post-ChargeReactivation is modified, according to data quality andquantity described as ‘enough’ (Table 1).

Results

The volumetric distribution of probabilities for the six sealingparameters is presented in Figure 20 and summarised for trapvolumes of 1%, 10% and 100% in Table 11. Included are the

Figure 17. FAST 3D model of the Skua area,presented in section view, looking from west to east.See Figure 5 for reference. Fault reactivation ismapped using colour-graded ∆P values, such thathigh-risk areas (low ∆P) are in red and low risk areas(high ∆P) are in blue.

SF1 (4.3 km2)SF2 (2.3 km2)

Spruce Fault (73.6 km2)Skua Fault (68.5 km2)Swift Fault (87.6 km2)Rowan Fault (90.3 km2)RF6 (7.0 km2)RF5 (6.4 km2)RF4 (2.6 km2)RF3 (7.0 km2)RF2 (3.2 km2)RF1 (2.0 km2)SF3 (3.5 km2)

Total Fault Surface Area

0 10 20 30 40 50 60 70 80 90

0.5 km2

3.5 km2

3.0 km2

2.5 km2

2.0 km2

1.5 km2

1.0 km2

Faul

t Sur

face

Are

a (k

m2 )

∆P (MPa)

Figure 18. Surface area histogram for range of ∆P values calculatedfor all faults in the Skua area, illustrating lower ∆P values for post-rift faults (higher risk). Rift faults are purple and pink, post-riftfaults are orange, yellow, green and red, and correspond to faultsdisplayed in Figure 6. Legend indicates fault name and total surfacearea of individual fault.

Table 11. Combined sealing probabilities of the cap rock, the boundingfault and the Skua trap for 100%, 10% and 1% trap volumes.

Table 10. The corresponding sealing probability (c) for Post-ChargeReactivation, the subjective description of data quality and quantity,and the modified sealing probability for trap volumes of 100%,10% and 1%, respectively.

Fault Reactivation

Considered Probability Quality & ModifiedVolume (%) (c) Quantity Probability (c)

100 0.70 Enough 0.65

10 0.70 Enough 0.65

1 1.00 n/a 1.00

Closure Considered ColumnContour (m) Volume (%) Height

Pcap Pfault Ptrap

2300 100 129 0.00 0.10 0.00

2238 10 67 0.00 0.14 0.00

2200 1 29 0.00 0.00 0.00


combined sealing probabilities for the fault (Pfault), the caprock (Pcap) and the entire trap (Ptrap). Figure 20 is bestunderstood by referring back to Figure 2, which listssubjective descriptions of probability values with respectto the uncertainty of the result. The closer the probabilitiesare to 0.5 the more uncertain the result, i.e. possibly leaking,possibly sealing.

The distribution recognises that for all trap volumes, thesealing probability of the cap rock is lower than the sealingprobability of the Skua Fault. However, there is greateruncertainty associated with the fault-related parameters

compared to the cap rock parameters. There exists a dramaticdrop in the individual fault-related probabilities and Pfault atapproximately 7% volume, equivalent to the depth, belowwhich trap closure is dependent on the bounding Skua Fault.Volumes greater than 7% correspond with fault parameterprobabilities of less than 0.65, implying a high degree ofuncertainty. Pfault reflects this uncertainty and therefore theprobability of sealing is inconclusive (0.5).

The uncertainty associated with cap rock sealingprobabilities are more varied than those for faults (Fig. 20).Cap Rock Geometry maintains a high probability of sealing

Figure 19. Areas of low ∆P associated with post-rift faults at base of seal.

Figure 20. Distribution of parameter sealing probabilities, Pcap, Pfault and Ptrap or trap volumes considered at Skua.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0.00% 10.00% 20.00% 30.00% 40.00% 50.00% 60.00% 70.00% 80.00% 90.00% 100.00%

Trap Volume

Prob

abili

tyof

Seal

ing

Cap Rock Capacity (i) Fault Plane Capacity (a) PcapCaprock Geometry (j) Juxtaposition Seal Capacity (b) PfaultCaprock Integrity (k) Fault Reactivation (c) Ptrap

Mildren et al. 291

(‘highly likely to be sealing’) for all trap volumes, whileCap Rock Integrity is consistently uncertain. Theuncertainty associated with Cap Rock Capacity increases(sealing probability becomes closer to 0.5) for greaterhydrocarbon volumes. The distribution of sealingprobabilities associated with Pcap also varies withincreasing trap volume. At greater trap volumes, theuncertainty is reduced, probability is lowered and the caprock is considered to be leaking. The overall probabilitythat the trap is sealing, Ptrap, indicates that for hydrocarbonvolumes of less than 7%, the trap is ‘more likely to beleaking than not’, and ‘highly likely to be leaking’ whenconsidering greater volumes.

The probabilities can also be interpreted as chance ofgeological success implying that there is approximately a 28%chance that 7% of hydrocarbon can be retained and a 10%chance that the trap can fill to capacity.

Discussion

The probability distributions generated for the Skua Fieldimply that there is nearly a 90% chance the trap is leaking,i.e., ‘highly likely to be leaking’ (Figs 2 and 20). Thisprediction corresponds with empirical data that suggests thehydrocarbons are leaking from the structure. The result alsosuggests that leakage may be related to failure mechanismsassociated with the cap rock rather than the bounding fault,as previously suggested by many authors (O’Brien andWoods, 1995; O’Brien et al., 1996; Lisk et al., 1998). Theweak strength of the cap rock may permit brittle failure ofthe seal and subsequent leakage. If this mechanism has beenresponsible for leakage at Skua in the past, it could explainthe abundance of HRDZ on the foot-wall of the northwest-dipping Skua Fault and its absence from the hanging-wall.However, there remains a high degree of uncertaintyassociated with the Cap Rock Integrity parameter andadditional analyses are recommended to minimise theuncertainty.

A fault intersection at the northeastern end of the Skuatrap has also been suggested as a possible leakage mechanism(Gartrell et al., 2002). Gartrell et al. (2002) proposed thatthe intersection of the Skua Fault and a cross-trending pre-rift fault has been the principal control on hydrocarbonretention at Skua. Although not considered here, a static leakpoint such as the one proposed by Gartrell et al. (2002), couldalso be incorporated into the integrated work flow. Dynamic(reactivation) and static (fault intersection) leakage couldboth be evaluated for trap volumes where the intersectionlies within the trap closure. The lower of the twoprobabilities would correspond to the sealing probabilityfor Post-Charge Reactivation (c). The volumes where faultintersection influences trap integrity would correspond toa reduction in sealing potential.

Although it is interesting to speculate on the failuremechanisms responsible for leakage at Skua, it is not arecommended practice using the integrated work flowalone. The work flow is presented as a methodology forevaluating the geologic success of the trap within the playconcept described by Otis and Schneidermann (1997). Theremaining three elements of the play, source, reservoir anddynamics (timing/migration), must be evaluated, beforeleakage can be quantified and failure mechanisms

identified. For example, although Skua is predicted to beleaking, if the rate of charge is significantly greater thanthe rate of leakage, then hydrocarbons will still be retained.This directly impacts on recoverable hydrocarbon volumeestimates and highlight the need to incorporate all elementsof the play concept.

Summary

This paper presents an integrated work flow for evaluatingtrap integrity and demonstrates the approach by applying itto the Skua Field in the Timor Sea. Although the results permitspeculation about the sealing/leaking mechanisms at play,the core of this paper demonstrates the usefulness andflexibility of the methodology.

The integrated work flow is a tool that ensures all sealfailure mechanisms are considered, when predicting sealintegrity for the purpose of estimating hydrocarbon reserves.It aids explorers to identify the uncertainties associated witheach sealing mechanism and therefore focus their attentionwhere it is most required. The flexibility of the work flowpermits explorers to use their preferred analysismethodologies when evaluating each seal parameter and, whenconsistently applied across perspective structures within abasin, it can be used as part of a portfolio ranking scheme.

Acknowledgments

The authors would like to thank the contribution of APCRCSeals Program consortium members Anadarko, BHP Billiton,Chevron, ExxonMobil, Globex, OMV, Origin, Santos andWoodside, and for their permission to publish these results.Thanks also to Pete van Ruth, Jerry Meyer, Ric Daniel, GeoffO’Brien, Wayne Bailey and Anthony Gartrell for their inputand to Schlumberger and Badleys for use of their software.Grant Ellis, David Dewhurst, and Hege Nordgård Bolås arealso acknowledged for their detailed comments on the originalversion of this manuscript.

References

ALLAN, U.S., 1989, Model for hydrocarbon migration andentrapment within faulted structures. AAPG Bulletin, 73,803–811.

ANTONELLINI, M. AND AYDIN, A.1994, Effect of faultingon fluid flow in porous sandstones: petrophysicalproperties. AAPG Bulletin, 78, 355–377.

BALCH, A.H. AND LEE, M.W., 1984, Vertical seismicprofiling: techniques, applications and case histories,Boston, International Human Resource DevelopmentCorporation, p. 488.

BHP PETROLEUM, 1992, Skua-8 well completion report,Basic Data Volume, unpublished.

BOUVIER, J.D., SIJPESTEIJN, K., KLEUSNER, D.F.,ONYEJEKWE, C.C. AND VAN DER PAL, R.C., 1989,Three dimensional seismic interpretation and fault sealinginvestigations, Nun River, Nigeria. AAPG Bulletin, 73,1,397–1,414.

BRECKELS, I.M. AND VAN EEKLEN, H.A.M., 1982,Relationship between horizontal stress and depth insedimentary basins. Journal of Petroleum Technology,34, 2,191–2,198.


BRETAN, P., YIELDING, G. AND JONES, H., 2003, Usingcalibrated shale gouge ratio to estimate hydrocarboncolumn heights. AAPG Bulletin, 87, 397–413.

CASTILLO, D.A., HILLIS, R.R., ASQUITH, K. ANDFISCHER, M., 1998, State of stress in the Timor Sea area,based on deep wellbore observations and frictional failurecriteria: application to fault-trap integrity. In: Purcell, P.G.and Purcell, R.R. (Eds), 1998, The Sedimentary Basinsof Western Australia 2: Proceedings of the PetroleumExploration Society of Australia Symposium, Perth, WA,1998, 325–341.

COWLEY, R. AND O’BRIEN, G.W., 2000, Identificationand interpretation of leaking hydrocarbons using seismicdata: A comparative montage of examples from the majorfields in Australia’s North West Shelf and GippslandBasin, The APPEA Journal, 40, 121–150.

CUNDY, H. AND ROLLETT, A., 1989, MathematicalModels, 3rd Ed., Tarquin Publications, Stradbroke,England, 255–256.

DOWNEY, M.W., 1984, Evaluating seals for hydrocarbonaccumulation. AAPG Bulletin, 68, 1,752–1,763.

FITTALL, A.M. AND COWLEY, R.G., 1992, The HV11 3-Dseismic survey: Skua-Swift area geology revealed? TheAPEA Journal, 32(1), 159–170.

FREEMAN, B., YIELDING, G. AND BADLEY, M., 1990,Fault correlation during seismic interpretation. FirstBreak, 8(3), 87–95.

GARTRELL, A., LISK, M. AND UNDERSCHULTZ, J.,2002, Controls on the trap integrity of the Skua Oil Field,Timor Sea. In: Keep, M. and Moss, S. (Eds), 2002, TheSedimentary Basins of Western Australia 3: Proceedingsof the Petroleum Exploration Society of AustraliaSymposium, Perth, WA, 2002, 389–407.

GIBSON, R.G., 1994, Fault zone seals in siliciclastic strataof the Columbus Basin, offshore Trinidad. AAPG Bulletin,78, 1,372–1,385.

GOLDSTEIN, B.A., 1994, Explicating a gut feel:benchmarking the chance for exploration success. TheAPEA Journal, 34(1), 378–417.

HILLIS, R.R., MILDREN, S.D., PIGRAM, C.J. ANDWILLOUGHBY, D.R., 1997, Rotation of horizontalstresses in the Australian north west continental shelfdue to the collision of the Indo-Australian and Eurasianplates. Tectonics, 16, 323–335.

JONES, R.M., BOULT, P., HILLIS, R.R., MILDREN, S.D.AND KALDI, J., 2000, Integrated hydrocarbon sealevaluation in the Penola Trough, Otway Basin. TheAPPEA Journal, 40(1), 194–211.

JONES, R.M., DEWHURST, D.N., HILLIS, R.R. ANDMILDREN, S.D., 2002, Geomechanical faultcharacterisation: impact on quantitative fault sealrisking. SPE 78213 Presented at SPE/ISRM RockMechanics Conference, Oilrock 2002, Dallas, Texas,October 20–23, 2002.

JONES, R.M. AND HILLIS, R.R., 2003, An integrated,quantitative approach to assessing fault seal risk. AAPGBulletin, 87(3), 507–524.

KALDI, J.G. AND ATKINSON, C.D., 1997, Evaluating sealpotential: example for the Talang Akar Formation,offshore northwest Java, Indonesia. In Surdam, R.C. (Ed.),Seals, Traps and the Petroleum System. AAPG Memoir67, 85–101.

KIVIOR, T, KALDI, J.G. AND LANG S.C., 2002, Cretaceousseal potential and sequence architecture in the VulcanSub-basin, North West Shelf Australia. The APPEAJournal, 42(1), 203–224.

KNIPE, R.J., 1992, Faulting process and fault seal. In:Larsen, R.M., Breke, H., Larsen, B.T. and Talleraas, E.(Eds) Structural and tectonic modelling and itsapplication to petroleum geology. NPF SpecialPublication 1, Stavanger, 325–342.

KNIPE, R.J., 1997, Juxtaposition and seal diagrams to helpanalyse fault seals in hydrocarbon reservoirs. AAPGBulletin, 81, 187–195.

LISK, M. AND EADINGTON, P., 1994, Oil migration inthe Cartier Trough, Vulcan Sub-basin. In: Purcell, P.G.and Purcell, R.R. (Eds), 1994, The Sedimentary Basinsof Western Australia: Proceedings of the PetroleumExploration Society of Australia Symposium, Perth, WA,1994, 301–312.

LISK, M., BRINCAT, M.P., EADINGTON, P.J. ANDO’BRIEN, G.W., 1998, Hydrocarbon charge in the VulcanSub-basin. In: Purcell, P.G. and Purcell, R.R. (Eds), 1998,The Sedimentary Basins of Western Australia 2:Proceedings of the Petroleum Exploration Society ofAustralia Symposium, Perth, WA, 1998, 288–302.

McGARR, A. AND GAY, N.C., 1978, State of stress in theEarth’s crust, Annual Review of Earth and PlanetarySciences, 6, 405–436.

MacDANIEL, R.P., 1988, The geological evolution andhydrocarbon potential of the western Timor Sea region,The First Century. APEA, Sydney, 270–278.

MacKAY, J.A., 1996, Risk management in internationalpetroleum ventures: ideas from a Hedburg Conference.AAPG Bulletin, 80(12), 1,845–1,849.

MARTIN, B.A. AND CAWLEY, S.J, 1991, Onshore andoffshore petroleum seepage: contrasting a conventionalstudy in Papua New Guinea and airborne laserfluorosensing over the Arafura Sea. The APPEA Journal,31(1), 333–353.

MILDREN, S.D., HILLIS, R.R., FETT, T. ANDROBINSON, P.H., 1994, Contemporary stresses in theTimor Sea: implications for fault-trap integrity. In: Purcell,P.G. and Purcell, R.R. (Eds), 1994, The SedimentaryBasins of Western Australia: Proceedings of the PetroleumExploration Society of Australia Symposium, Perth, WA,1994, 301–312.

MILDREN, S.D. AND HILLIS, R.R., 2000, In situ stressesin the southern Bonaparte Basin, Australia: Implicationsfor first- and second-order controls on stress orientation.Geophysical Research Letters, 27(20), 3,413–3,416.

MILDREN, S.D., HILLIS, R.R. AND KALDI, J., 2002a,Calibrating predictions of fault seal reactivation in theTimor Sea. The APPEA Journal, 42, 187–202.

MILDREN, S.D., BOULT, P. AND CAMAC, B.A., 2002b,Risking brittle failure of cap rocks in the Otway BasinAustralia. In: Evaluating the hydrocarbon sealingpotential of faults and cap rocks. AAPG HedbergConference, December 1–5, Barossa Valley S.A.

MURRIS, R.J., 1980, Middle East: stratigraphic evolutionand oil habitat. AAPG Bulletin, 64, 597–618.

NAKANISHI, T. AND LANG, S.C., 2001, Visualisation offluvial stratigraphic trap opportunities in the Pondrinie3D Survey, Cooper-Eromanga Basin. In: Hill, K.C. and

Mildren et al. 293

Bernecker, T. (Eds), 2001, The Eastern Australasian BasinsSymposium – A refocused energy perspective for thefuture, Petroleum Exploration Society of Australia SpecialPublication, Melbourne, 301–312.

O’BRIEN, G.W., BICKFORD, G., BISHOP, J., HEGGIE,D.T. AND MARSHALL, J.F., 1992, Light HydrocarbonGeochemistry of the Vulcan Sub-basin, Timor Sea: RigSeismic Survey 97. BMR Record 1992/62.

O’BRIEN, G.W., ETHERIDGE, M.A., WILLCOX, J.B.,MORSE, M., SYMONDS, P., NORMAN, C. ANDNEEDHAM, D.J., 1993, The structural architecture ofthe Timor Sea, northwestern Australia: implications forbasin development and hydrocarbon exploration. TheAPEA Journal, 33(1), 258–279.

O’BRIEN, G.W. AND WOODS, E.P., 1995, Hydrocarbon-related diagenetic zones (HRDZs) in the Vulcan Sub-basin, Timor Sea: recognition and explorationimplications. The APPEA Journal, 35, 220–252.

O’BRIEN, G.W., QUAIFE, P., COWLEY, R., MORSE, M.,WILSON, D., FELLOWS, M. AND LISK, M., 1998,Evaluating trap integrity in the Vulcan Sub-basin, TimorSea, Australia, using integrated remote-sensinggeochemical technologies. In: Purcell, P.G. and Purcell,R.R. (Eds), 1998, The Sedimentary Basins of WesternAustralia 2: Proceedings of the Petroleum ExplorationSociety of Australia Symposium, Perth, WA, 1998,237–254.

O’BRIEN, G.W., LAWRENCE, G.M., WILLIAMS, A.K.,WEBSTER, M., LEE, J., COWLEY, R. AND BURNS, S., 2001,Hydrocarbon migration and seepage in the Timor Sea andnorthern Browse Basin, North West Shelf, Australia: Anintegrated SAR, geological and geochemical study. AGSORecord 2001/11 and GIS.

OSBORNE, M., 1990, The exploration and appraisal historyof the Skua Field, AC/P2 – Timor Sea. The APEA Journal,30(1), 27–50.

OTIS, R.M. AND SCHNEIDERMANN, N., 1997, A processfor evaluating exploration prospects. AAPG Bulletin, 81,1,087–1,109.

PATILLO, J. AND NICHOLLS, P.J., 1990, A tectono-stratigraphic framework for the Vulcan Graben, TimorSea region, The APEA Journal, 30, 27–51.

ROSE, P.R., 2001 Risk analysis and management ofpetroleum exploration ventures. AAPG Methods inExploration 12, 164 pages.

SCHOWALTER, T. T., 1979, Mechanics of secondaryhydrocarbon migration and entrapment. AAPG Bulletin,63, 723–760.

SERRA, O., 1984, Fundamentals of well-log interpretation.Translated from French by P. Westaway and H. Abbott,Development in Petroleum Science, Elsevier, Amsterdam.

SLUIJK, D. AND PARKER, J.R., 1986, Comparisons ofpredrilling predictions with postdrilling outcomes,using Shell’s Prospect Appraisal System. In: Rice, D.D.and James, W.R. (Eds), Oil and gas assessment:methods and applications. AAPG Studies in Geology,21, 55–58.

VAVRA, C. L., KALDI, J.G. AND SNEIDER, R.M., 1992,Geological applications of capillary pressure: A Review.AAPG Bulletin, 76, 840–850.

WATSON, P., 1998, A process for estimating geological riskof petroleum exploration prospects. The APPEA Journal,38, 577–583.

WATTS, N., 1987, Theoretical aspects of cap-rock and faultseals for single- and two-phase hydrocarbon columns.Marine and Petroleum Geology, 4, 274–307.

WOODS, E.P., 1994, A salt-related detachment model forthe development of the Vulcan Sub-basin. In: Purcell, P.G.and Purcell, R.R. (Eds), 1994, The Sedimentary Basinsof Western Australia: Proceedings of the PetroleumExploration Society of Australia Symposium, Perth, WA,1994, 259–274.

YIELDING, G., FREEMAN, B. AND NEEDHAM, T., 1997,Quantitative fault seal prediction. AAPG Bulletin, 81,897–917.

ZOBACK, M.L., 1992, First and second-order patterns ofstress in the lithosphere: the world stress map project,Journal of Geophysical Research, 97(B8), 11,703–11,728.


Scott Mildren completed his PhDat the University of Adelaide in1997 and took up a position asstructural geologist with Z&S(Asia) Ltd. (now Baker AtlasGeoscience). At Z&S, heinterpreted resistivity and acousticimage data with application toin situ stresses and fracturecharacterisation. From 1999 until2004, he was among the academic

staff at the National Centre for Petroleum Geology andGeophysics and the Australian School of Petroleum,researching various geomechanical related issues, such asfault seal integrity, fractured reservoirs and wellbore stability.He is currently a managing director of JRS PetroleumResearch Pty Ltd, providing geomechanical and image logservices to the petroleum industry. Scott is a member ofAAPG, ASEG, PESA and SPE.

Tom Kivior graduated with a BSc(Honours) from the NationalCentre for Petroleum Geologyand Geophysics (Adelaide, 1997),and is completing a PhD from theAustralian School of Petroleum(University of Adelaide). Tomcurrently works for Schlumbergerand is a Director of ArchimedesConsulting Pty Ltd. He is a

member of AAPG, ASEG and PESA.

Richard Hillis is the State ofSouth Australia Professor ofPetroleum Geology and theMawson Professor of Geologyand Geophysics at the Universityof Adelaide. He graduated with aBSc (Honours) from ImperialCollege (London, 1985), and aPhD from the University ofEdinburgh (1989). His main

research interests are in petroleum geomechanics andsedimentary basin tectonics. He leads a group of nine, whoare working on these topics at the Australian School ofPetroleum (University of Adelaide), has published ~70papers, and has consulted to many Australian and internationaloil companies. Richard is a Director of JRS PetroleumResearch Pty Ltd and of Petratherm Ltd. He is a member ofAAPG, AGU, ASEG, EAGE, GSA, GSL, PESA, SEG & SPE.

John Kaldi is head of theAustralian School of Petroleum inAdelaide. He studied forBachelors and Masters Degrees inGeology at Queens College, CityUniversity of New York, andreceived a PhD from CambridgeUniversity, England. From 1980to 1982, he worked as a ResearchGeologist for the Saskatchewan

Geological Survey in Regina, Saskatchewan, Canada,specializing in Mississippian Carbonate reservoirs. He wasthen with Shell Canada as a Senior Research Geologist,focusing on production geology in carbonates and wassubsequently with ARCO in Plano, Texas, as a SeniorReservoir Geologist. Between 1991 and 1997, he wasGeological Specialist/Chief Development Geologist withARCO, Indonesia, in Jakarta, working mainly on theevaluation of reservoirs, seals and pay. In 1997, he joinedVICO Indonesia as Chief Geologist. John Kaldi’s specialitiesinclude carbonate sedimentology and diagenesis, sealevaluation, reservoir geology and multi-disciplinary studies,and he has more than 40 publications on these topics. He is afrequent contributor to APPEA proceedings and is activelyinvolved in several professional societies including AAPG,IPA, SPE, SEPM, IAS and PESA.

Integrated seal assessment and geologic risk with …...Integrated seal assessment and geologic risk with application to the Skua Field, Timor Sea, Australia Scott D. Mildren 1,2,3,

Documents