Fault Seal Analysis:

Constraining fault seal risk using seismic velocities

Tobias Colson

Supervisor: Prof. Myra Keep

Co-Supervisor: Adj. Prof. David Castillo

This thesis is submitted to fulfil the requirements for Master of Science (Geology) by way of
Thesis and Coursework
Faculty of Science
October 2014

i
ii
TABLE OF C ONTENTS

Acknowledgments ......................................................................................................................... v
Glossary ........................................................................................................................................ vii
Abstract.......................................................................................................................................... ix
1. Introduction ............................................................................................................................... 1
2. Background ................................................................................................................................ 2
2.1 The fault and capillary seal ............................................................................................... 2
2.2 The application of seismic data ......................................................................................... 7
2.3 Geological structure of the study area - Dampier Sub-basin and Rankin Trend,
North West Shelf ..................................................................................................................... 10
3. Aims (and objectives): ............................................................................................................ 16
4. Methodology: ........................................................................................................................... 17
5. Results ....................................................................................................................................... 24
5.1 Structural Characteristics of the Rankin Dockrell Keast Blocks - Rankin Trend . 24
5.2 Velocity Analysis ............................................................................................................... 44
6. Discussion/Analysis ................................................................................................................. 66
7. Conclusions .............................................................................................................................. 78
8. Further work ........................................................................................................................... 79
9. Appendix A -Tectonostratigraphic diagram ...................................................................... 80
10. Appendix B – Stratigraphic column for the West Dampier Sub-basin ........................ 81
11. Appendix C – Porosity and permeability in the NCB ..................................................... 82
12. Appendix D - Pore pressure ................................................................................................ 86
13. Appendix E - Empirical calculation of shear wave velocity ........................................... 89
14. Appendix F - Shear and normal stress .............................................................................. 91
15. Appendix G – Thermal gradient for NCB ........................................................................ 92
16. References .............................................................................................................................. 93

iii
iv
ACKNOWLEDGMENTS

This project was supported by BP Developments Australia Pty Ltd. Many thanks to Ben
Boterhoven, Vincent Dykman and Gerchard Pfau in the North West Shelf Joint Venture
Exploration team of BP for their assistance during this research.

Thanks to Prof. Myra Keep for supervising the project and co-supervisor Adj. Prof. David
Castillo for advice.

Finally to my MSc research peers in petroleum exploration geology, M.S. and C.vR.,

Prodeo Quo Victum !

T.Colson 2014

v
vi
GLOSSARY

Checkshot - Seismic Velocity Checkshot Survey - Used to calibrate time-depth curves by

conducting a series of sound shots with a receiver at various depths in the well.

CFF - Coulomb Failure Function.

FIT - Formation Integrity Test - used to test the strength of a formation by increasing the
bottom hole pressure in a well. Where LOT tests are conducted and 'Leak off' not achieved, a
FIT test has been conducted, often quoted as 'Max Pressure'.

LOT - Leak Off Test - A test to determine the strength or fracture pressure of the open
formation, usually conducted immediately after drilling below a new casing shoe. During the
test, the well is shut in and fluid is pumped into the wellbore to gradually increase the
pressure until the fluid enters the formation, or leaks off, either moving through permeable
paths in the rock or by creating a space by fracturing the rock. The results of the leak-off test
dictate the maximum pressure or mud weight that may be applied to the well during drilling
operations.

NCB - North Carnarvon Basin.

NWS - North West Shelf, Western Australia.

RFT/FIT - Repeat Formation Tester/Formation Interval Test - Tool/measurement used to

assess formation fluid pressure, density and other fluid properties (FIT is used in older Wells).

Sonic Log - Well Log used to measure interval transit time by sending out a high-frequency
sound signal and receiving a return signal at a receiver in the same tool. Resolution and
frequency are far higher than with a checkshot, so it has important scale dependent differences
compared with the checkshot.

Stacks - Stacking Velocities. Once seismic traces are collected according to a common mid-
point (CMP), their velocities are corrected for the normal hyperbolic moveout, and
combined/stacked to improve the signal-to-noise ratio. Stacks can be interval velocities, RMS
or average velocities.

Shmax - Maximum horizontal stress

Shmin - Minimum horizontal stress

vii
viii
ABSTRACT

Fault seal capacity is an important component in the conventional petroleum system.

Assessing the capacity for a fault to seal or leak can be difficult, particularly where well
constraint is lacking. In the frontier basin, in a marine setting, seismic velocities may be the
only data available. However, useful constraints on a faults sealing capacity can be extracted
from this data alone. This study investigates the robustness of a number of empirical relations
that can assist in extracting useful constraints from seismic velocities and amplitudes.
Reliable estimates on maximum and minimum stress tensors and pore pressures can be
calculated and combined with basic fault architecture analysis, to place practical constraints
on fault risk. In this study of an area on the Rankin Trend found good correlation between
well-based and seismic velocity-based pore pressures and stress magnitudes allowing a
coulomb failure function based only on stacking velocities to be calculated. Faults separating
the Rankin 1 well block from the Dockrell/Keast Field, were shown to be within a stable
stress regime. Calculated pore pressures match known RFT measurements and show that
overpressure can be assessed using basic time-migrated velocity stacks. Furthermore,
theoretical capillary pressure and hydrocarbon column calculations correlate with known
values and highlight the capillary seal potential of sand-sand juxtaposition seals which can
support 100 to 200 m column heights within the Triassic play. They suggest that many traps
on the Rankin Trend within the Triassic play are limited by a combination of pore pressure
and stress orientation. However, deeper intra-formational seals are likely to have increased
seal capacity by virtue of overburden stress and reduced porosity exceeding any capacity for
shear stress failure. Well data for this area confirms a Sh max orientation of approximately 110
+/- 10o and calculations show that faults striking within 20° of this direction may be at high
risk of failure within the neotectonic setting, where pore pressures and dip predicate fault slip.

ix
x
1. INTRODUCTION

The primary components of a petroleum system include source, seal, trap, reservoir and a
mechanism for migration of petroleum. These factors must all be assessed, not just for their
presence, but for the degree to which they are associated and the sequence in which they were
formed. Key to the effectiveness of a petroleum system is the extent to which secondary
migration and charge access can facilitate the linking of these primary components. An
element of assessing the risk factors associated with secondary migration and charge access is
the capacity for faults to inhibit or facilitate secondary migration, as well as their capacity to
contribute to trap formation.

Many petroleum exploration targets that are drilled on faulted traps and fail to return positive
results have a lack of petroleum accumulation attributed to charge access. Charge access
relates to the capacity for secondary migration to facilitate the 'filling' of a defined trap.
However, the contribution that fault seal has played in this is poorly understood.
Advancement in reservoir petroleum management, particularly with 4D time-dependent
seismic monitoring, has advanced our understanding of why faults fail when they do.
However, the ability to make predictions of fault seal capacity from the geometry and seismic
velocities alone allows more accurate risk factors to be applied in the 'green fields' stage of oil
and gas field exploration.

This study addresses some key constraints used in assessing whether a fault is likely to seal or
leak. Principal to these constraints is the typical lack of well control in frontier exploration. A
first-pass methodology based on seismic data alone enables better defined risking parameters
to be placed on the fault component of the petroleum system analysis where well control is
lacking.

1
2. B ACKGROUND

2.1 THE FAULT AND CAPILLARY SEAL

Secondary migration refers to the movement of petroleum from its source rock (in the
'kitchen') to a site where it accumulates (the trap). Secondary migration (movement through a
carrier bed) differs from primary migration (out of the source rock) via the porosity,
permeability and pore size distribution in which the petroleum migrates.

The process of secondary migration is essentially one of fluid movement. Density and
viscosity differences notwithstanding, the fundamental principles of fluid mechanics are
applied equally, whether the migration of gas, oil or water is being assessed. These principles
have being applied in studies on migration and seal capacity by Hubbert (1953), Smith
(1966), Berg (1975) and Schowalter (1979). Watts (1987) further developed cap (top seal) and
fault rock seal capacity based on capillary leakage in which cap rock seals are divided into
those that fail by capillary leakage and those that fail by hydraulic fracturing due to capillary
entry pressures being higher than the tensile strength of the sealing lithology.

Capillary leakage refers to interstitial pore pressure exceeding the capillary pressure of the
formation and subsequently leaking. Capillary pressure exists whenever two immiscible
phases are present in a fine bore tube or analogous void or pore space in a rock. Capillary
pressure is then defined as the pressure drop or differential across a curved liquid interface.

Watts (1987) shows how capillary pressure Pc in a pore space can be calculated if the fluid
interfacial tension γ, rock-fluid contact angle θ and the pore radius, r, are known;

𝟒𝜸 𝐜𝐨𝐬(𝜽)
𝑷𝑪 = ... Eqn. 1
𝒓

Alternatively, capillary pressure can be expressed as a hydrostatic head. It is equal to the

product of the height of the liquid (column height) the density difference between the two
liquids and the gravitational constant g ;

𝒉𝒄 = 𝑯𝒈(𝝆𝒘 − 𝝆𝒄)(Watts, 1987) ...Eqn. 2

For capillary seals, Watts (1987) noted that cap-rock thickness becomes unimportant once the
capillary slug pinch-off length is exceeded. The seal capacity of a hydraulic seal on the other

2
hand is highly dependent on thickness, together with the in-situ stress state of the sealing layer
and the degree of overpressure development.

The concept of cap rock failure by capillary leakage or hydraulic fracturing is directly
transferable to our understanding of whether a fault will seal and inhibit, or leak and/or
facilitate secondary migration. Sealing faults can thus be thought of as a capillary cap rock at
an angle.

Fault related seals may be further divided into those where the fault plane itself acts as a seal
and those where the sealing unit is laterally juxtaposed against reservoir sands (‘juxtaposition
seals’). Most if not all fault seals will preferentially leak via capillary leakage rather than
hydraulic fracture (not including leakage along the fault) (Aydin, 2000; Bailey et al., 2006;
Dodds et al., 2007; Shipton et al., 2005; Tueckmantel et al., 2010; Watts, 1987; Shepherd,
2009)

The concept of capillary pressure or membrane seal potential and associated pressure
differential is a direct reflection of the permeability of the fault in question. The permeability
is known to be highly dependent on the clay content of the fault (Aydin, 2000; Berg, 1975;
Bretan et al., 2003; Busch and Amann-Hildenbrand, 2013; Dehandschutter et al., 2005;
Dewhurst et al., 2002; Fisher et al., 2001; Fisher and Knipe, 2001; Kachi et al., 2005;
Ngwenya et al., 2000; Schowalter, 1979; Shipton et al., 2005; Watts, 1987) and clay content
in particular is a primary control on seal behaviour of faults in mixed clastic sequences
(Ngwenya et al., 2000; Torabi et al., 2013; Tueckmantel et al., 2010).

There are a number of empirical based algorithms that are widely used to attempt to predict
seal capacity based on the clay component of a fault. These include, the Shale Gouge Ratio
(SGR), the Smear Gouge Ratio (SMGR), the Shale Smear Factor (SSF) and the Clay Smear
Potential (CSP) (Allen and Allen, 2005; Childs et al., 2009; Kachi et al., 2005; Yielding et al.,
1997). The shale gouge ratio (SGR) is one of the more widely used algorithms and is based on
the assertion that the sealing capacity of a fault is directly related to the number of shale beds
and therefore amount of clay material within the slipped interval. If clay is incorporated in the
fault gouge, the clay will increase the membrane seal potential of the fault and increase the
chances of an effective seal.

3
The studies by Fisher and Knipe (2001) directly attributed increased clay percentage in the
fault gouge to decrease in permeability of the fault. However, the concept assumes a
homogeneous distribution of the percentage of clay within the fault, an assumption that
develops an inherent weakness in this approach for predicting seal potential. Studies have
shown consistently that there will generally be a high degree of variability in the amount of
clay within a fault (Bailey et al., 2006; Bretan et al., 2003; Fisher et al., 2001; Fisher and
Knipe, 2001; Torabi et al., 2013). Furthermore, this method inherently requires an
understanding of the amount of clay that has been incorporated within the fault gouge. Whilst
mature basins, where core analysis is available, benefit from a deterministic petro-physical
constraint placed on SGR estimates, new exploration fields generally lack this information.
Without robust clay component estimates, SGR parameters can become misleading. SGR
estimates suggest that faults with shale gouge of <15% clay component will tend to leak
(Yielding et al., 1997). However, quartz dissolution-related diagenetic processes can
contribute up to 2 orders of magnitude in permeability reduction in faults (Fisher and Knipe,
2001; Ngwenya et al., 2000; Tueckmantel et al., 2010). SGR estimates do not take this into
account, nor do they take into account increasing confining pressures with depth. The greater
a cap rock of any kind is buried, the greater the compaction and cementation contribution to
permeability reduction will be.

Empirically derived equations that attempt to incorporate factors such as cementation and
increased confining pressures are highly dependent on the data set from which they are based
and have limited applicability to different tectonic settings, fault types and lithologies.
Nonetheless the SGR and SMGR, where reliable estimates of clay content are available, can
provide a first pass assessment, and studies that show a reduced permeability in fault gouge
with increasing clay content suggest that higher clay content should indicate higher Pc. A
higher Pc alone, on the other hand, will not necessarily be indicative of a sealing fault,
because if the pressure differential, ΔP, across the fault exceeds Pc, the fault will still leak
(Fig. 1) (Fisher and Knipe 2001).

Thus, if the capacity for cross fault migration depends on the pressure differential, it then
depends on how well a trap on one side of the fault is charged as to whether a trap will
equilibrate on both sides of the fault (Fisher et al., 2001). Studies also show the sealing
capacity of a fault should be calculated from the difference in pressure between the

4
hydrocarbon and pore-water at the position along the fault where leakage is most likely to
occur (Fisher et al., 2001).

P2

Pc
P1 ΔP = P1-P2

Figure 1 Representation of a typical fault zone with fault gouge in the damage zone depicted
as varying from siliceous cataclastic gouge to shale smear gouge with a higher clay content
(from Shipton et al., 2005). Pc denotes the capillary pressure of the gouge, and p1 and p2
denote the host rock or reservoir pressure, where Δp is the difference in pressure from one
side of the fault to the other.

The pressure differential can therefore provide an indication as to whether there is fluid
communication on one side of the fault to the other. However, as (Fisher and Knipe, 2001)
and Fisher (2001) note, pressure differences and heights of hydrocarbon columns are often
more likely to be controlled by factors such as the capillary entry pressure differences of the
undeformed reservoir lithologies and the amount of hydrocarbons entering the reservoir than

5
by the actual capillary entry pressure of the fault rock itself. So whilst the pressure difference
will not necessarily reflect the capillary entry pressure of the fault gouge, it may provide an
indication of a potential effective fault seal. This can be complicated since a lack of pressure
difference does not necessarily mean that the fault is leaking. Without knowledge of absolute
hydrocarbon column height potential of the cap rock, a ΔP of zero may mean that that there
has been sufficient charge to fill the trap up to the maximum capillary pressure of the fault
gouge, and leakage has occurred whenever the pressure exceeds Pc max. This would present a
repetitive leak-seal sequence analogous to a ‘fill to spill’ scenario. Given such a scenario, an
assessment would ideally be made on pressure differences between hydrocarbon and pore
water pressures in areas with minimum shale/clay component as this would have minimal
capillary pressure potential and be where the fault was most likely to leak.

There are numerous studies (Aydin, 2000; Bense et al., 2013; Losh, 1998) that ascribe fault
gouge acting as a fluid conduit for hydrocarbon migration without providing significant direct
evidence. Rather, they conclude that fault formation and trap formation are related and
typically infer that fluid conduit migration ‘probably’ facilitated the migration pathway.
Studies do show that fault-related dilatant fractures can increase permeability (Aydin, 2000;
Dewhurst and Hennig, 2003; Dewhurst et al., 2002; Færseth et al., 2007; Ngwenya et al.,
2000; Shepherd, 2009), which, along with mode II failure, places a dependency of fluid
migration on the seismicity of the fault and a dependency on conditions necessary for
reactivation.

The magnitudes of the maximum principal stress and the minimum cohesive shear strength
necessary for reactivation sensitively depend on the orientation of the horizontal intermediate
principal stress. The reactivation becomes easiest when the intermediate principal stress axis
is rotated out of strike by 45°. Here it requires only one third of the minimum cohesion and
little more than half of the maximum compressive stress necessary for pure dip-slip reversal
(Mandl, 2000). This implies that the axis of maximum horizontal stress, σ H, during
reactivation is unlikely to coincide with that of the earlier axis of minimum horizontal stress,
σh. Hence, most inversion episodes for example, are non-coaxial, with oblique-slip or strike-
slip kinematics being the important modes of fault reactivation. In such a setting, shortening
strains will often be partitioned between faults displaying dip-slip, oblique-slip and strike-slip
kinematics (Turner and Williams, 2003).

6
The Coulomb Failure Function provides an assessment on the likelihood of mode I and II
failure where shear stress on a fault plane exceeds the effective normal stress. This function
takes into account the effect pore pressure has on the effective stresses on a fault plane. If
such a failure mode is prevalent, the in-situ stress field needs to be assessed and fault seal
reactivation risked (Castillo et al., 2000; Jones and Hillis, 2003; Reynolds et al., 2005;
Zoback, 2010; Zoback and Healy, 1984).

2.2 THE APPLICATION OF SEISMIC DATA

One of the principal indicators of fault seal potential is the pressure differential across the
fault. How then can we attempt to measure petro-physical properties such as pore fluid
pressure, rock strength and porosity? We know that when external forces are applied to a
body, balanced internal forces are set up within it. Stress is a measure of the intensity of these
balanced internal forces. The stress acting on an area of any surface within the body may be
resolved into a component of normal stress perpendicular to the surface, and a component of
shear stress in the plane of the surface. A fluid does not shear stress however, so no shear
stress is developed in a fluid. A body subjected to stress undergoes a change of shape and/or
size known as strain (Brown, 2004; Kearey et al., 2002; Yilmaz, 2008).

Up to a certain limiting value of stress, known as the yield strength of a material, the strain is
elastic and directly proportional to the applied stress. The linear relationship between stress
and strain in the elastic field is specified for any material by its various elastic moduli, each of
which expresses the ratio of a particular type of stress to the resultant strain.

It is the bulk elastic properties of a material that determine how much it will compress under
an external pressure. The ratio of the change in pressure to fractional volume compression is
referred to as the bulk elastic modulus (Bulk Modulus) and it is the bulk modulus that
influences the speed of sound such as seismic waves in a material, as well as being a factor in
the amount of elastic energy or stress that is built up in a solid. This bulk modulus can be
thought of as the state of incompressibility of a solid and it is this incompressibility that
allows the propagation of sound.

7
Seismic waves can then be thought of as parcels of elastic strain energy that propagate
outwards from the seismic source. Since rocks differ in their elastic moduli and densities, they
differ in their response to seismic waves measured through differing seismic velocities and
energy propagation. Information on the compressional P (Primary) and S (shear) wave
velocities, Vp and Vs , of rock layers measured through seismic wave response can therefore
provide an indication of the lithology of the rock through which they propagate and the nature
of the pore fluids contained within it. Because shear strain will not develop either in the pore
space or in the pore filling fluid, it is only the P-wave velocity that is influenced by pore
fluids. Thus, in principle, a variation in pore fluid type as well as pore fluid pressure should be
detectable through P-wave velocity variation (Brown, 2004; Kearey et al., 2002; Yilmaz,
2008).

Because the velocity of seismic waves is highly dependent on the density of the medium they
propagate through, subtle velocity changes can also give an indication of the densities and
pressure through which they propagate. This enables P wave velocity changes to give an
indication of not only the relative pressures and densities but also the bulk modulus and
subsequently the stress states of a rock (Hatchell and Bourne, 2005). The change in reservoir
pore-fluid properties due to pressure depletion such as gas coming out of solution can also
have a large impact on the seismic velocity, and studies show subtle variation in P-wave
velocities depending on whether the pore fluids are brine (water) or gas (Dvorkin, 2008;
Hatchell and Bourne, 2005)

The bulk modulus can be calculated from shear wave velocities based on empirically derived
relationships between Vp and Vs (Castagna, 1993). However, Vs can also be derived from
porosity velocity relationships (Castagna, 1993; Castagna et al., 1985; Dvorkin, 2008). Where
known porosity trends are available, this allows further deterministic assessment of the elastic
moduli response of a stress regime, as well providing comparison with bulk moduli
determined from the empirical relationship between Vs and Vp. Furthermore, we see that since
the pore fluid type should not affect the shear wave response, this would also allow a further
assessment on the impact on Vp by the pore fluid, independent of any anisotropic effect on the
Vs. The inter-relationship between Vp and capillary pressure becomes apparent when we
consider the empirical equation relation of porosity to pore throat radius (Pittman, 1992;
Torabi et al., 2013). Torabi et al. (2013) also show a porosity permeability power law
relationship that relates a lower porosity and permeability to fault core in relation to host rock.

8
However, apart from the exponent, the empirical relationship is the same, such that the
hydrocarbon column height can then be related to pore throat radius via;
𝟐𝜸
𝒉𝒄 = ... Eqn. 3
𝑹𝒈(𝝆𝒘 −𝝆𝒄 )

and where;

∆𝑷 𝟐𝜸
𝒉𝒄 = => ∆𝑷 = = 𝑷𝑪 ...Eqn. 4 (Hasegawa et al., 2005; Shipton et al.,
𝒈(𝝆𝒘 −𝝆𝒄 ) 𝑹

2005)

where Pc is the capillary pressure, ΔP the pore fluid pressure differential across the fault, g
gravitational constant, R the pore throat radius, γ fluid viscosity, ρ the density of w (water)
and c (hydrocarbon). Pore throat radius can be estimated from porosity through Pittman's
(1992) empirical relationship. Thus, through equations 3 and 4, an empirical determination of
capillary pressure for a given lithology can be assessed (Hasegawa et al., 2005; Pittman, 1992;
Shipton et al., 2005; Torabi et al., 2013; Watts, 1987).

Stacking velocities in a seismic survey is related to the normal-moveout (NMO) velocity

which in turn, is related to the root-mean-squared velocity from which the average and
interval velocities are derived. Interval velocity is the average velocity in an interval between
two reflectors. Pore shape, pore pressure, pore fluid saturation, confining pressure and
temperature all influence the interval velocity within rock unit (Yilmaz, 2008).

One of the most prominent factors influencing velocity in a rock of given lithology and
porosity is confining pressure. This type of pressure arises from the overburden and increases
with depth. It is generally true that velocity increases with depth, however, because of factors
such as pore pressure, there may be inversion in the velocity within a layer (Yilmaz, 2008).
There are a number of empirically derived equations relating the effect of overburden stress
and pore pressure to the Vp . Amongst the more robust relationships are the relationship of
pore pressure to vertical stress, first proposed by Terzaghi (1923) in 1923;

𝝈𝒗 = 𝝈𝒗 − 𝜶𝑷𝒑 ... Eqn. 5

9
and

𝒛
𝝈𝒗 = 𝝆𝒘 𝒈𝒛𝒘 + 𝒔𝒃
𝝆 𝒛 𝒈𝒅𝒛 ...Eqn. 6 (Gardner et al., 1974; King et al., 2010;
Tingay et al., 2003; Castagna et al., 1985)

Where σv is the vertical stress field, z is the depth or amount of overburden, g gravitational
constant and ρ the density of w; water, for the water column, or a density as a function of
depth below mudline. Density of the lithology is the dominant petrophysical property
affecting the velocity and Gardner et al's., (1974) equation 𝜌 = 1.741𝑉𝑝0.25 is typically used.
However, Castagna's (1993) separate empirical relations for sandstone, limestone and shale
(𝜌 = 𝑎𝑉𝑝2 + 𝑏𝑉𝑝 + 𝑐) show a closer relation to typical trends. The coefficients provided by
Castagna (1993) provide good results, notwithstanding the fact that if sonic log data is
available, these should always be corrected (Castagna et al., 1985; Gardner et al., 1974;
Tingay et al., 2003).

2.3 GEOLOGICAL STRUCTURE OF THE STUDY AREA - DAMPIER SUB-BASIN AND

R ANKIN TREND, NORTH WEST SHELF

The Dampier Sub-basin forms part of the North Carnarvon Basin (NCB) which itself
constitutes one of the four main basins of the North West Shelf (NWS) of Australia. The
NWS is a collection of marginal and offshore rift basins along the northern Australian margin
(Fig. 2). It is a marginal rift with pre-rift Permo-Triassic intracratonic sediments overlain by
Jurassic to Cenozoic syn- and post-rift successions deposited in response to rifting and
seafloor spreading throughout the Paleozoic and Mesozoic (Keep et al., 2000; Keep et al.,
2007; Longley et al., 2002). The NWS has pervasive NNE-SSW trending faults, some of
which show reactivation (Etheridge et al., 1991). The primary faults that influence the
structural trends are inherited from the final stages of rifting and breakup of Gondwana
throughout the Jurassic and early Cretaceous (Keep et al., 2007).

10
Figure 2. Regional geology of the North West Shelf showing the four main constituent basins
(Bonaparte, Browse, Canning and Carnarvon). Modified from (Keep and Moss, 2000).

Tectonic collision during the Neogene modified the pre-existing rift architecture, causing
localised fault reactivation (Longley et al., 2002). The change in tectonic regime from one of
overall extension to one of overall shortening, manifested as pulses of deformation during the
Miocene and Recent (Harrowfield et al., 2003). Bulk strain, estimated at less than 1%
shortening, was strongly partitioned on to a small number of favourably orientated structures
(Harrowfield et al., 2003; Keep et al., 2007). Areas of discrete reactivation/inversion occur in
some parts of the Bonaparte, Browse and Carnarvon basins (Keep et al., 2007). These plate
boundary forces continue to have an impact on the present day stress field and neotectonic
intraplate deformation of Australia (Fig. 3) (Hillis et al., 2008; Longley et al., 2002; Müller et
al., 2012).

11
Figure 3 Australian stress map showing the orientations (small black lines) of modelled
contemporary (6-0 Ma) maximum horizontal surface stress with small white lines showing
maximum horizontal stress measured from bore hole breakouts (colour bar is in MPa). In the
Carnarvon basin (Inset), the stress is shown to be in a NW-SE orientation (Müller et al.,
2012).

Differences in neotectonic activity along the NWS reflect rift-related compartmentalisation

which resulted in the division of the major basins. The compartmentalisation takes place
along NW-trending features that have been shown to be Proteozoic fracture systems. The
basement involvement in this compartmentalisation has resulted in the compartments
responding to regional stress fields in different ways (O'Brien et al., 1997). As a result, there
is a marked change in structural orientation on the NWS in the Carnarvon Basin from a
northeast orientation in the north (Dampier Sub-basin) to a north-northeast orientation in the
south (Exmouth Sub-Basin) (Baillie and Jacobson, 1995). Muller et al. (2012) modelled the
σH (horizontal stress) regime which changes considerably in the late Miocene with the onset
of collision at the Papua New Guinea margin. Late Miocene σ H orientations rotate in a

12
clockwise direction from east–west in the Carnarvon Basin to northeast in the Timor Sea,
correlating with structural observations.

The NCB which forms the westernmost basin of the NWS has undergone multiple
deformation episodes including Late Permian extension (Cathro and Karner, 2006), Late
Triassic inversion and further localised extension in the Early Jurassic (King et al., 2010) (see
Appendix A for tectonostratigraphic column of the NCB).

The two final phases of extension in the Mesozoic resulted in continental breakup and
formation of oceanic crust in the Argo (Callovian) and Gascoyne-Cuvier abyssal plains
(Cathro and Karner, 2006). Minor inversion at this time is interpreted to be the result of a
major plate re-organisation as India moved northwest and break up commenced between
Australia and Antarctica (Cathro and Karner, 2006; King et al., 2010).

Extensive faulting in the Jurassic produced the five main structural elements of the Dampier
Sub-basin, including the Rankin Platform, the Madeline-Dampier Trend, the Lewis Trough,
the Legendre-Rosemary Trend and the Enderby fault zone (Fig. 4). The Rankin Platform is a
relatively high structural platform of Triassic rocks between the Rankin Fault and the shelf
edge.

Since the Neogene, collision of the Australian Plate with the Java-Banda arc system to the
north has dominated tectonism along the NWS (Keep and Harrowfield, 2008; Keep et al.,
2007). As a result, present day stress fields (Fig. 3), determined from modelling, borehole
breakouts and seismicity, indicate that maximum horizontal stress trends NE-SW across the
Bonaparte, Browse and Canning basins, reflect the dominant control by plate collision in the
PNG region (Hillis and Reynolds, 2003; Hillis et al., 2008; Hillis and Williams, 1993a; Hillis
and Williams, 1993b; Müller et al., 2012). However, in the Carnarvon Basin the modelled
stress field changes orientation, with maximum horizontal stress oriented ESE, suggesting
dominant control by far field stress generated from the plate collision in the Himalayas (Hillis
et al., 2008; Revets et al., 2009).

13
 Madeleine trend

Rosemary trend

Figure 4 Dampier Sub-basin and Rankin Platform (modified from (Baillie and Jacobson,
1995)).

Cretaceous inversion concentrated along major boundary faults of the northeast–southwest

orientated Rankin, Madeleine, and Rosemary–Legendre trends, with the locus of Miocene
inversion located adjacent to the northwest limits of the Rosemary–Legendre Trend (Fig. 5).
Inversion has critically controlled the major structures for trapping hydrocarbons within the
Dampier Sub-basin (Cathro and Karner, 2006).

The Dampier Sub-basin can be divided broadly into two structures, flanking either side of the
main Oxfordian depocenter. Fields along the western flank include the large rift-related horst
block traps such as the Goodwyn, North Rankin, and Echo-Yodel wet gas fields, the single
unique Perseus lowside saddle trap and numerous other smaller horst and tilted fault block
traps such as the Searipple field beneath Perseus. Subcrop of different Early Jurassic and Late
Triassic sands within many traps at the level of the base regional seal produce complex fields
containing oil and dry gas and wet gas phase combinations (Longley et al., 2002).

14
The accumulation of significant volumes of hydrocarbons at multiple stratigraphic levels
along both the flanks and within the main basinal area demonstrates the laterally drained high-
impedance nature of the petroleum system within the central Dampier Sub-basin (Longley et
al., 2002). In the Dampier Sub-basin and Rankin platform in particular, the complex charge
history is reflected in at least 5 distinct fluid families, ranging from oil accumulations in the
Jurassic to dry gas accumulations suspected to have undergone a component of vertical
migration from the Triassic (Longley et al., 2002).

Figure 5 schematic cross-section of the West Dampier Sub-basin (Bennett, 2005) (see
Appendix B for typical stratigraphic column of the Dampier Sub-basin).

So we see the structural history of the Rankin platform, along with the numerous well-
defined, yet complex hydrocarbon fields, provide the ideal setting to better understand the
implications that the regional stress field and multiple episodes of fault reactivation have had
on sealing faults. The deep seated faults penetrating the Triassic are well suited as they have
typically been formed during the early stages of the Gondwana breakup and have undergone
multiple phases of reactivation. This is particularly relevant as industry now more actively
pursues the Triassic Play of the NWS.

15
3. AIMS (AND OBJECTIVES):

This study assesses parameters that can be used to define risk factors on fault seal when well
data is scarce. Fault seal risk analysis is not only important in the assessment of petroleum
reservoirs, but also in the assessment on the impact of any type of fluid flow within the
ground. The cross-discipline implications none more apparent than in the assessment of fluid
flow in ground water. Ground water management continues to of immense importance to the
broader economy and livelihood of those that depend on it. The developing science of CO 2
sequestration is yet another discipline where fault seal capacity can have significant
implications. The assessment of CO 2 reservoir storage potential, relies on the accurate
assessment of the impact that pre-existing faults and future potential seismicity may have on
its integrity.

The principle objectives focus on:

 Identifying the degree to which fault architecture and stress regime can contribute to
fault seal capacity assessment.

 Endeavouring to link fault seal capacity to formation pressure and seismic velocity.

 Evaluation of capacity to make passive evaluation (no wells) of fault seal risk factor.

Fundamental to these objectives is the assertion (hypothesis) that the risk factor on fault seal
can be better defined using seismic velocities alone, to a practical level, where well log data is
unavailable. This study aims to test this hypothesis.

16
4. METHODOLOGY :

The study involved the analysis of the Demeter 3D survey. Acquired in 2004 by Woodside
Pty Ltd on behalf of the North West Shelf Joint Venture, it is a SEG normal polarity pre-stack
time migrated volume. The survey covers approximately 3590 square km over the Echo-
Yodel, Rankin, Dockrell/Keast, Goodwyn and Perseus fields, to name a few, which sit on the
northern end of the Rankin Trend (Fig. 6). The survey has an inline bin spacing of 6.25m and
crossline bin spacing of 25m with a fold of coverage of 133. The survey was reportedly
optimised for maximum bandwidth to improve imaging of shallow targets (Bennett, 2005).

Figure 6 Map showing the blocks comprising the NWSJV and outline (red line) of the
Demeter 3D survey area, on the western edge of the Dampier Sub-basin, NCB (modified from
(Bennett, 2005).

A small subset of the area was selected, covering the Rankin and Dockrell/Keast fields and
associated fault blocks (Fig. 7). The subset area (summarised in Table 1) was selected as it
covers some well known fault blocks which could be used for assessment. Furthermore, the
associated stacking velocity file had to be small enough in size to be loaded on the software in
use, which has a 4GB restriction. Seismic analysis was conducted using the Landmark
software suite V5000.0.2 and final maps were generated using the Petrosys mapping software

17
suite. The stacking velocities were High Density Velocity Analysis (HDVA) ‘unsmoothed’
pre stack time migrated CMP stacks of a 12.5 m x 25 m grid. HDVA stacks were generated to
create a velocity field with higher spatial resolution for final NMO correction, rather than the
500m x 500m sparse array that may typically be generated. The stacking velocities are
sampled on every shotpoint at a spacing of 12.5 x 25 m with a vertical sampling of 32ms.

Table 1 Survey Subset details

Survey Inlines Crosslines Depth Range (ms) Bin Size Area (km2)
Demeter 3D -
1800 1400 0 - 6000 6.25 x 25 394
subset
32..sampling
HDVA Vels 900 1400 12.5 x 25 394
interval

Checkshot survey data was available for 32 wells in the Rankin Trend and Exmouth Plateau.
Exmouth Plateau wells were analysed for comparative purposes in a tectonic area with a
'similar' history. The checkshot data, sometimes referred to as the 'poor man’s’ sonic log,
provide accurate time depth conversion, as well as interval velocities that are based on Dix's
Equation (Dix, 1955). This allowed calculations to be compared to well-based data at
common depths. A quality check of the data was also performed, comparing HDVA,
checkshot and sonic velocities. Sonic logs have a vertical resolution greatly exceeding that of
the seismic velocities, thus, these were not used in calculation checks, rather the checkshots
represent a far more realistic approach to qualifying the validity of using the various empirical
relationships that may be employed on the stacking velocities.

18
 Kendrew
Trough
Rankin High

Rankin
Trend

Rankin
Escarpment

10 km

Figure 7 Base Oxfordian regional unconformity that intersects the Horst Graben structure of
the Rankin Trend. The escarpment fault which bounds the eastern edge of the Rankin Trend is
also indicated. The study area is bounded in the green box (modified from (Bennett, 2005)).

The empirical relationships used rely on established properties in standard rock physics
models. In basin analysis, the overburden stress Sv can be calculated as a function of thickness
Z and density ρ of the overlying rocks multiplied by gravity g.

𝑺𝑽 = 𝒁𝝆𝒈 ... Eqn. 7

The bulk density was calculated using the empirical relationship of (Castagna, 1993). This
polynomial fit allows density to be determined in the absence of sonic log, core sampling, or
locally calibrated trends:

𝝆 = 𝒂𝑽𝟐𝒑 + 𝒃𝑽 + 𝒄; ...Eqn. 8

Here, a = - 0.0261, b = 0.3730 and c = 1.4580 . These constants match the fit for shales and
feldspathic sandstones, and best approximate the marls, claystones and clay and shale inter-

19
bedded sandstones that dominate the lithology in the study area. The appropriateness of
Catagna's (1993) relationship was compared to (Gardner et al., 1974) relationship;

𝝆 = 𝟎. 𝟐𝟑 ∗ 𝟑. 𝟐𝟖𝟎𝟖𝑽𝟎.𝟐𝟓
𝒑 ...Eqn. 9

as well as a porosity based empirical relationship (Swarbrick and Osborne, 1998) relating
porosity to density;

𝝆 = 𝝆𝒎𝒂 𝟏 − ∅ + 𝝆𝒇𝒍 ∅ ...Eqn. 10

where ma and fl refer to the matrix and fluid density respectively. Equation 10, based on
(Wyllie et al., 1956) relationship, had the advantage of comparing density trends from two
different sources, thus comparing not just the closeness of 'fits'. Here, porosity is related to
depth via a porosity depth trend for the NCB (see Appendix C for details). The resulting
density allows Terzaghi's (Terzaghi, 1923) equation to calculate overburden stress;

𝝈 = 𝑺𝑽 − 𝑷𝒑 ; ...Eqn. 11 where Pp is the pore pressure.

Pore pressure can be measured in the well via a Repeat Formation Tester (RFT) tool and RFT
pressure data was collected from publically available Well Completion Reports (WCR) for 28
Wells along the Rankin Trend. This allows the effective stress to be calculated for depths
corresponding to the RFT measurement points. However, this also allows a QC check to be
performed on an empirical relation between seismic velocity and pore pressure. The common
pore pressure relationship given by Eaton (1969) relates stress measured to pore pressure from
the ratio of measured seismic velocities to the normal velocity trend for the basin. Although, it
is noted that a ‘normal’ trend can be difficult to confirm in many instances and possibly
erroneous where for example, as is the case for the Rankin trend, overpressure is so consistent
that the overpressure trend is 'normal'.

There are a variety of additional empirical relationships that reportedly provide good results
where good well calibration is available (Sayers, 2006; Soleymani and Riahi, 2012; Zhang,
2011). This study uses Bower's (1995) method, which required less calibration compared to
some others and was therefore considered more readily applied to the frontier basin. Bower's
(1995) modification relates velocity to pore pressure based on a virgin pressure curve (see
Appendix D) taken from a reference well, Rankin 1 in this case, and subtracts this from the
lithostatic pressure gradient;

20
 𝑽𝒊𝒏𝒕 −𝟏𝟐𝟓𝟎 𝟏/𝟎.𝟔
𝑷 = 𝟑. 𝟗𝟓𝟖 ∗ 𝒛 − ...Eqn. 12
𝟏𝟔.𝟓

where z is depth in meters, Vint the interval velocity in m/s.

As the stacking velocities only contained the P-wave velocities. Shear wave velocities are
calculated using the empirical relationship of Castagna (1993);

𝑽𝒔 = 𝒃𝑽 + 𝒄 ...Eqn. 13

where b = 0.77 and c = -0.867. Again this is for a lithology that best fits the dominant rock
type in the formations being studied (see Appendix E for the 'goodness of fit' for this
empirical relation). This allows Poisson's ratio to be calculated through the relationship;

𝑽𝟐𝒑 −𝟐𝑽𝟐𝒔
𝝂= ...Eqn. 14
𝟐(𝑽𝟐𝒑 −𝑽𝟐𝒔 )

And Poisson’s ratio can be used to equate the horizontal stress Shmin extent via;
𝑺𝒗
𝑺𝑯𝒎𝒊𝒏 = ; ...Eqn. 15
𝟏 𝝂 −𝟏

A range of different empirical relationships to calculate Shmin are available (Zoback, 2010).
Those employing an effective stress ratio consistent with Byerlee’s (Byerlee, 1978) range of
values for constant of internal friction are compared (Eaton, 1969; Hubbert and Willis, 1957;
Zoback and Healy, 1984). Figure 8 summarises the work flow utilising the relationships
described above.

This allows a range or properties to be assessed; importantly the Coulomb Failure Function
(CFF):

𝑪𝑭𝑭 = 𝝈𝒔 − 𝝁𝒔 𝝈𝒏 − 𝑷𝒑 ...Eqn. 16

where, σ n is the normal stress, σs the shear stress and Pp pore pressure. Appendix F describes
the three dimensional calculation of the normal and shear stress from the principal stress
vectors.

The CFF measures the difference between shear stress and effective normal stress. The effect
of increased pore pressure is to decrease the effect of overburden stress such that shear stress
is greater and Mode I and II failure is of higher risk.

21
If we know the normal and shear stress, we can calculate the angle of internal friction, or
angle at which the ratio of shear and normal stress puts the rock in a critical state of stress
through
𝝈𝟑
= 𝛍𝐜 = 𝐭𝐚𝐧 𝜽 ...Eqn. 17
𝝈𝟏

μs on the other hand refers to the coefficient of sliding friction, equal to;

𝑺𝑺
= 𝛍𝐬 ...Eqn. 18
𝑺𝒏

This is essentially Amontons’ law (Amontons, 1699), who had expanded on experiments on
friction that had first been conducted by Leonardo da Vinci (Gao et al, 2004). The coefficient
of sliding friction describes the ratio of shear to normal stress in a fault that will initiate slip or
movement. Byerlee (1978) found that for a range of rock types, the ratio of shear stress to
effective normal stress required to initiate slip was between 0.6 and 1, a phenomenon known
as Byerlee’s law. Further studies (Jaeger et al., 2007), also showed that the ratio, or μs was
remarkably consistent and approximate to 0.6. Some mudstones and clays have been shown to
have a μs of less than 0.6 and as low as 0.2 at low effective pressures (Lockner et al., 2011;
Moore and Lockner, 2004; Moore et al., 2009; Zoback, 2010). Low μs in the fault gouge of
some rock types is attributed to a high concentration of certain phyllosilicates, of which the
more common minerals have been shown to range between 0.2-0.5 (Moore and Lockner,
2004). However, aside from these minerals and the fault gouge that is comprised of them,
most rocks typically adhere to Byerlee's law thus, where μs is applied in this study, it is
assumed to be 0.6. This carries forward into the various empirical relations for Shmin
employed, where an Effective Stress Ratio (ESR) of 0.32 is equivalent to a value of μs of 0.6.

22
Figure 8 Workflow diagram summarising the process of assessing the validity of select
empirical relationships used to relate velocity to fault seal properties.

23
5. RESULTS

5.1 STRUCTURAL CHARACTERISTICS OF THE RANKIN DOCKRELL KEAST BLOCKS -

R ANKIN TREND

Spudded in July 1971, Rankin 1 (Fig. 9) tested a structural culmination on the Rankin Trend
structural high. The Rankin 1 well drilled through Pleistocene to Upper Triassic formations
with Cretaceous rocks lying unconformably on the Triassic with the entire Jurassic and much
of the Neocomian absent. The Triassic sandstones contained significant hydrocarbon
accumulations with 63m of gas/condensate pay underlain by a 3.4m oil column and an
oil/water contact at approximately 2955m, 274m higher than at North Rankin.

Io/Jansz Field

Rankin-Dockrell/Keast
Wells/Fields

Exmouth
Plateau

Gorgon Field

Barrow Island

Figure 9 The study area with main oil and gas fields and well locations. Inset: tectonic map
highlighting the Sub-basins and Rankin Trend in the NCB. The blue box over the inset relates
to the area covered by the larger picture (inset: (Geoscience_Australia, 2010)).

24
The Dockrell Field was discovered in 1973 with the Dockrell 1 well intersecting a stacked
reservoir of three hydrocarbon accumulations of gas and oil, with approximately 88m of oil
and gas in the top Brigadier reservoir, and 15m of gas/condensate in the lower two Mungaroo
Triassic intersections.

The Keast Field was discovered in 1996 and has two penetrations. Keast-1, drilled in 1996,
encountered a stacked reservoir with five hydrocarbon accumulations with the fluid
intersected being gas. Keast-2, drilled in 1997, was a dry hole. Whilst both Keast and
Dockrell penetrated the Upper Triassic, core data is limited to the Brigadier formation.

Together, the three wells, which are not under production, provide pressure information
across two faults separating the Rankin High, from the Keast Graben and Dockrell high (Fig.
10). The fault blocks of the Rankin High, Keast Graben and Dockrell High are comprised of
extensional graben, half graben blocks separated by moderately dipping (40-50°) planar
normal faults. These normal faults show differential offset in the strata indicative of
continuing reactivation throughout the Jurassic age (Figs. 10-12). The main southern
bounding Rankin Escarpment fault has been reactivated well into the middle Neogene with
hard and soft linkage between the Triassic and Cretaceous/Cenozoic overburden, alternating
along the NE-SW strike of the fault. The faults show significant throw in the Triassic strata
with the Rankin Keast fault and Keast Dockrell faults showing more than 300 ms TWT
(approximately more than 400 m). Pre-kinematic strata fill the Keast Graben within the
Triassic and along the Rankin Escarpment (Figs. 12-13).

Figures 10 and 11 both show two-way time slice through 2920 ms in the Triassic. Figure 11
displays amplitude variation across the time slice with red (negative) amplitudes principally
reflecting a shale-dominant lithology with a resulting high impedance contrast. The amplitude
time slice shows horizontal separation in the NW-SE faults, which, when considered with the
relative dip of reflectors suggests that a dextral strike-slip or oblique slip movement has
occurred.

Most of the Jurassic is missing with the reported Oxfordian sub-aerial exposure resulting in
the erosion of these sequences. The upper part of the Oxfordian is dominated by claystones
which form a seal over the underlying lowstand sands. This in turn is overlain by regional
transgressive and highstand marine claystones of the Early Cretaceous period.

25
 ms TWT

E-W faults within

the Rankin Block

Keast Graben

Dockrell Block
Rankin-Echo/Yodel Block

Keast/Dockrell N-S Fault

Rankin
Escarpment/Fault

Rankin/Keast N-S Fault

Figure 10 TWT structure map of the base Norian Late Triassic surface showing the main
structural elements in the study area. Triassic fault blocks have undergone extensive faulting
and reactivation with the southern bounding Rankin escarpment fault representing the
southern boundary of the Rankin Trend high. The Rankin, Keast and Dockrell wells have all
been drilled in fault bounded structural highs with recorded hydrocarbon accumulations.

These Oxfordian and Neocomian claystones dominate the seismic cross sections with high
impendence contrast. Only extensive and continual reactivation of the Rankin Escarpment and
Goodwyn fault block offset these regional claystone seals. However, whilst Neogene
reactivation is hardlinked to the escarpment fault, many Triassic fault blocks show some
minor evidence of reactivation, with soft-linking through the Jurassic and Cretaceous

26
claystones (Figs. 12-13). The claystone horizons themselves are dominated by classic
polygonal faulting that is typically associated with compaction and dewatering, but without
clear evidence of hardlinks through the more ductile shale dominant lithology (Fig. 13).

The semblance attribute time slice at 2910 ms TWT in figure 14 shows faulting within the
main fault blocks. The Keast Graben and Dockrell high have small faults cross-cutting the
blocks at an angle of approximately 60o to the bounding faults and with small vertical offsets,
typically less than 10 ms TWT. These cross-cutting faults terminate at the bounding faults.
The E-W faults within the Rankin/Echo-Yodel fault block are somewhat different, with very
little vertical offset and a E-W striking trend. The E-W faults also indicate horizontal
separation (Fig. 15). These E-W faults comprise a small horst and graben set cross-cutting the
Rankin/Echo-Yodel fault block and offsetting some N-S faults at an angle of approximately
60°.

Rankin/Keast fault
Keast/Dockrell fault

E-W faults within

Rankin Block

Rankin escarpment

Figure 11 2910 ms TWT amplitude time slice of the study area. The principal faults in the
study can be seen, along with horizontal separation in the east-west faults cross-cutting the
Rankin block.

27
SW NE
Reactivation of Triassic
Rankin fault into Neogene
1500 overburden

Fault bend fold 500

2000
m

Amplitude
2500
m
Rankin high

Keast Graben -500

Rankin escarpment
fault
400m

a)

SW

Eocene/Paleocene SB

KT boundary

Upper Norian

Middle Norian

100 msTWT
400m

b)
Figure 12 a) NE-SW cross section (inset shows location (inset: red lines show correct spatial
location of inline and crossline, but are not to scale in length) showing the Rankin High and
well bounded by the escarpment fault to the SW and Keast Graben to the NE. Both faults are
reactivated normal faults with fault bend folds in the upper Triassic and reactivation up into
the middle Neogene horizons b) showing principal stratigraphic boundaries (TWT, v/h x 2
and x 4 respectively).

28
 Eocene: Rankin Escarpment still present. Shore
line depositional features dominate with no
further evidence of reactivation of underlying
structure.

Shoreline channels in
siliciclastic-dominated
depositional sequence

Rankin Escarpment shows Neogene

reactivation

Rankin Escarpment fault

Goodwyn/Dockrell fault

Middle Cretaceous: Rankin Escarpment still

apparent with the geomorphology revealing little
about underlying structure in this calcilutite and
carbonate-dominated depositional sequence.

Base Cretaceous: Classic polygonal faulting in the

shale sequence with the Rankin escarpment
evident, but underlying architecture hidden under
the unconformity.

Figure 13 Semblance images show time

slices in the Eocene, Middle Cretaceous and
Base Cretaceous respectively. They highlight
the reactivation of the Rankin Escarpment
fault throughout, however, no reactivation of
the underlying horst-graben architecture is
evident within the geomorphology above the
regional unconformity which is controlled by
the depositional setting of the area.

29
 Keast Graben
3 km

E-W faults restricted

to the Triassic -
vertical offset < 5ms
TWT
Faults cross-cut the
graben and horst at
approximately 60o to
the bounding faults

Rhombohedral geometries
cross-cut the horst-graben
Complex faulting - architecture throughout
process zone the area.
crossover Rankin escarpment

Figure 14 Semblance attribute of the 2910 ms TWT slice (Late Triassic Carnian) through the study area highlighting the fault array. The main
Rankin Escarpment fault bounds the Rankin, Keast and Dockrell blocks to the south and the smaller East -West faults cross-cutting the Rankin
Block can be seen.

30
SW NE

2000

2250
m

2500
m
E-W faults within
the Rankin block
2 km

Figure 15 Cross section through E-W striking faults that manifest within the upper Triassic of
the Rankin block. These faults show small vertical offset (approx. 10 ms TWT) but 100s of
metres of horizontal separation (v/h x8).

The Keast/Graben and Rankin and Dockrell Highs are underlain by a large listric fault that
terminates at the escarpment and strikes along the uplifted basement structure that constitutes
the greater Rankin Trend (Fig. 16). Figure 16 shows that any tilting of the horst graben fault
blocks will not occur without influence from the underlying Rankin Trend structure.

Figure 17 also shows the Keast Graben displaying subtle strike slip features indicative of fault
reactivation with an oblique slip component, with strata peaking across offset (pseudo teepee).
Horizons with varying angle of dip and some reverse offset are also possible indications of
oblique slip. The regional base Cretaceous unconformity (Figs. 19 and 20) shows that whilst
the horst-graben bounding faults do not offset the shale, the geomorphology is still controlled
by the oblique slip reactivated horst-graben architecture. This architecture is also shown in
figure 18, which again exhibits oblique slip features on what would otherwise be considered a
classic extensional horst graben array.

31
 NW SE

Listric fault underlying horst graben architecture

Figure 16 Inline cross section of the Keast Graben highlighting listric faulting underlying the horst graben architecture (v/h x 2).

32
 Rankin Block Keast Graben Dockrell Block

SW NE

2000
500 500
Amplitude

Amplitude
2250

2500
-500

3 km

teepee transtensive / transpressive fault drag Rankin Escarpment

Figure 17 The horst graben architecture can be seen with transpressive/transtensive

reactivation generating teepee and inversion rollover along the main horst-graben bounding
faults. Varying angle of dip also indicate block rotation (oblique slip) rather than a simple
strike-slip sideways movement (v/h x8).

33
NW SE

2000

Fault drag across fault

with normal offset.

2500

Main escarpment fault maintains

net normal displacement - any
oblique slip reactivation is
transtensional
Varying dip in strata, plus reverse
3000 polarity in offset and teepee-like
conversion suggest inversion with an
oblique slip component has
reactivated the fault.

Figure 18 An inline cross section across the Keast Graben showing features indicative of
fault reactivation with an oblique slip component, with strata peaking across offset (pseudo
teepee). Horizons with varying angle of dip and some reverse offset suggest that horizontal
movement may be associated with the episodes of fault reactivation that have occurred (v/h
x8).

4 km

Figure 19 Map of the regional

Cretaceous unconformity: a
deep marine shale deposited on
top of the erosional Jurassic
unconformity. Whilst only the
main reactivated escarpment
faults offset the horizon, the
horizontal movement and
rotation of the Triassic fault
blocks are reflected in the
geomorphology of this surface.

34
 Dockrell 1 Dockrell 1
A A ´
B B´

Reactivation of Triassic
Rankin fault into Cretaceous
overburden Fault bend fold

2000 2000 500

2500 2500

Amplitude
Rankin escarpment
3000 fault 3000
Dockrell high
Dockrell high
A´
B
Keast Graben

500m 500m Rankin escarpment B´

fault

A
Figure 20 Dockrell a) shows the Dockrell horst fault block along strike NE-SW (inset: red lines show correct position of inline and crossline, not
to scale) which has NE dipping strata from the northerly tilt of the fault block. b) NW-SE cross-section on the edge of the escarpment again
shows the reactivation into the Neogene (v/hx2).

35
The implications of this fault reactivation on lithology and subsequently, pressure
communication across faults, may be further assessed from consideration of the velocity
profile across this fault array. Figures 21 to 25 show the stacking velocities as colour
overlaying the amplitude cross-sections for the fault blocks being studied. The cross sections
through the Rankin closure show significant velocity variation though the Triassic fault
blocks. In the Upper Triassic, velocity differential can be seen across the Rankin escarpment
fault, and interestingly within the Rankin High itself, where a small fault appears to be
delineating the southern and northern parts of the Rankin Closure (Fig. 21). Beyond 3000 -
3200 ms TWT, the velocity differential is too poor to make any significant observations
regarding relevant variation.

The Keast Graben and Dockrell High show distinct velocity shifts from the Cretaceous to
Triassic lithology (Fig. 22-23). Low velocities persist at greater depths around the bounding
faults. The velocity also shows a significant increase at the base of the Norian. In all the
velocity overlays, a low velocity zone can be seen dominating the Cretaceous formation. This
is a region consisting of a mixture of marls, claystones and calcilutite. The velocity increases
again over the claystone shale layers that make up the regional unconformity above the
Triassic. Lateral variation in velocity is minimal above the regional claystone seals and
relatively high below the seal within the Triassic fault blocks. Low velocities focus around
bounding faults and at the 'up dip' section of dipping strata (Figs. 22- 25). A distinct
differential in relative velocity across the faults is not evident. However, as figure 25
illuminates, the velocity gradient is highly sensitive to formation boundaries. The shift from
one major formation to another can almost be picked from the velocity gradient shift alone.

The velocity distribution map, for the base Norian surface (Fig. 26), shows how the velocity is
principally dependent on density and therefore depth, with the deeper parts of the surface
correlating with higher velocities. The surface is within the Mungaroo formation, and shows
lateral variation in lithology that is not seen in the TWT surface.

36
 SW Rankin 1 NE

Little velocity
variation across the
Rankin/Keast fault
5000
Velocity variation noted
across minor fault transecting
the Rankin high in which the
Rankin 1 Well has been
drilled.
2000

Vint m/s
2500

Keast Graben
Rankin escarpment
3000 fault

3500

Figure 21 Cross section (location inset) showing velocity variation across the Triassic fault blocks of the study area. A velocity differential is
observed between the minor fault transecting the Rankin structural high. This variation may suggest a pressure differential associated with fault
seal. This delineates the Rankin accumulation from the south-bounding escarpment, however, little differential is evident between the Rankin
block and Keast Graben (v/h x2).

37
NW SE
Dockrell 1

5000

Relatively slower
velocities appear
around the faults

2000

2500

Vint m/s
3000

3500

Keast Graben
Dockrell high
Rankin block

Figure 22 Crossline through the Dockrell and Keast Graben. The velocity profile shows relatively low velocities around the main horst-graben
bounding faults. The velocities also appear to delineate the main formations with high consistency (v/h x1.5).

38
 Dockrell 1 Dockrell 1
A A ´
B B´

Rankin escarpment
fault 5000

2000 2000

2500 2500

Vint m/s
3000 3000 Dockrell high
Dockrell high

Rankin escarpment A´
B 3500
fault

B´
500m 500m

Figure 23 Inline and crossline across the Dockrell accumulation showing significant velocity variation across the Triassic fault blocks of the
study area. The Rankin escarpment fault displays a velocity differential across it, however, again there is little variation between the Keast
Graben and Dockrell high. Both the inline, B, across the Dockrell Horst and the crossline A, aligned with the strike of the horst show evidence of
slow velocity patches within the Carnian which may suggest fluids or gas within the Middle Triassic (v/h x4).

39
 SW NE

5000

Vint m/s
Structural culmination over otherwise
extensional architecture is highly suggestive of
inversion
Rankin escarpment Rankin/Keast fault
fault

3500

Figure 24 This crossline across the Rankin, Keast and Dockrell fault blocks shows evidence of inversion, principally through the structural
culmination within the greater Keast, Dockrell Graben where sediment has been deposited prior to extension, extensionally listrically faulted and
inverted. The varying dip within each fault block is itself a possible result of fault block rotation associated with oblique slip. Unlike the inlines,
this cross line at a smaller scale does show some velocity variation across the Rankin/Kea st fault, however in this instance, it is difficult to
separate this from fault shadow effectsf (v/h x4).

40
 Interval Velocity (m/s)
0 1000 2000 3000 4000 5000 6000 7000 8000
SW NE

claystone
calcareous
marl and
5000

and marl
claystone
Calcilutite
Rankin 1 HDVA SPT Stack
claystone
sandstone
2000 &
claystone

Vint m/s
claystone
Coarse Sst with interbedded
2500

3500
3000

600m

Figure 25 Whilst velocity differential across faults is ambiguous, the velocity profile shows remarkable consistency with the formation
boundaries. ‘Formation’, as being distinct from ‘lithology’ (such as shale or sandstone), means the overall gross lithological properties that define
a formation such as the Haycock Marl or Mungaroo Formation (v/hx2).

41
TWT ms

Velocity m/s

2 km

Figure 26 a) Base Norian Horizon in TWT showing the principal fault blocks of the study, including Goodwyn 6 and the eastern edge of the
Goodwyn fault block. b) showing the P-wave velocity profile of the Base Norian horizon. The velocity profile is a good indicator of lithology
and density, and therefore the fault block architecture, however, lateral resolution is too poor to infer any pore pressure differential.

42
Figure 27 shows the Depth Structure map of the Base Norian Horizon. It shows that the
general structural geomorphology is consistent with the TWT maps with no culminations or
depressions removed during depth conversion. The depth map is used in the pore pressure
calculation for the base Norian horizon.

Depth mTVDSS

5 km

Figure 27 Depth Structure map of the base Norian (Triassic) horizon. The structure remains
consistent with the time structure maps.

43
5.2 VELOCITY ANALYSIS

The velocity profile of the Rankin Trend is broadly consistent along its entirety (Fig. 28).
However, when compared to a similar region within the same basin; the Exmouth Plateau,
each displays a different rate of increase with depth. The exception is Gorgon 1, which is
known to sit over an anomalous 'hot spot' within the basin (see Appendix G). In addition, the
interval velocities (Fig. 29) as opposed to the instantaneous velocity, show a velocity reversal
at approximately 2000 mBML which is also different when compared to the Exmouth
Plateau, which has no reversal. The velocity reversal, present over the entire Rankin Trend, is
clearly shown in figure 30.

Velocity Profile of the Rankin Trend and Exmouth Plateau Jansz 3

Nickols Bay
OWT (ms) Jupiter
0.00 500.00 1000.00 1500.00 2000.00 2500.00 Vinck 1
Eendracht 1
0.00 Goodwyn 6
Goodwyn 9
Goodwyn 1
Zeewulf 1
Zagreus 1
1000.00 Athena 1
Dockrell 1
Exmouth Plateau Wells Dockrell 2
Rankin 1
2000.00 North Rankin 1
North Rankin 2
Lady Nora 1
Depth (m)

Dixon 1
Dixon 2
3000.00 West Dixon 1
Rankin Trend Wells Sculptor 1
Miller 1
Haycock 1
4000.00 Yodel 1
Echo 1
Keast 1
Perseus 1
Dampier 1
5000.00
Malmsey 1
Eastbrook 1
Gorgon 1
Mecury 1
6000.00 Saturn 1

Figure 28 The velocity profile of 33 wells in the NCB comparing the Rankin Trend to the
Exmouth Plateau, two sub-regions of the Basin that are relatively uplifted. Each has a distinct
velocity trend of its own.

44
 Rankin Trend Wells

Exmouth Plateau Wells

Figure 29 Profile of the interval velocity in the Rankin trend and Exmouth Plateau for
comparison. Unlike the Exmouth Plateau, the Rankin trend has a distinct change in slope at
approximately 1200m and 2500m TVDSS.

Velocity Profile of the Rankin Trend

Interval Velocity (m/s)
Dockrell 1
0 1000 2000 3000 4000 5000 6000 7000 8000
0.00
Dockrell 2
500.00

1000.00 Rankin 1

1500.00
North Rankin 1

2000.00
Depth (m)

North Rankin 2
2500.00

3000.00
West Dixon 1

3500.00
Sculptor 1
4000.00

4500.00 Keast 1

5000.00

Figure 30 The interval velocity profile of the wells within and adjacent to the blocks of the
study area on the Rankin trend. A decrease in interval velocity with increasing depth is seen
between approximately 2200m and 2700m.

45
The Vp/Vs ratio is approximately 2, albeit with a slow decrease (increasing Vs) with depth.
The first 1000m below mudline has a slow decrease in the ratio in both the Rankin trend and
Exmouth plateau. The ratio remains remarkably constant beyond 1000m BML, with the
exception of significant lithology variation or overpressure/fluid contact zones.

Vp/Vs Trends for the Rankin Trend and Exmouth Plateau Jansz 3
Nickols Bay
Vp/Vs Jupiter
1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 Vinck 1
0.00 Goodwyn 6
Goodwyn 9
Goodwyn 1
500.00
Rankin Trend Wells Zeewulf 1
Rankin Trend water depth approx. Zagreus 1
1000.00
<200m Athena 1
Dockrell 1
Dockrell 2
1500.00 Rankin 1
North Rankin 1
North Rankin 2
2000.00 Lady Nora 1
Depth (mTVDSS)

Dixon 1
Exmouth Plateau Wells Dixon 2
2500.00 Exmouth water depth approx 1400m West Dixon 1
Sculptor 1
3000.00 Miller 1
Haycock 1
Yodel 1
3500.00 Echo 1
Keast 1
Perseus 1
4000.00 Dampier 1
Malmsey 1
Eastbrook 1
4500.00
Gorgon 1
Mercury 1
5000.00 Saturn 1

Figure 31 The Vp/Vs ratio with depth is rather consistent, at approximately 2, and 2.3 in
depths correlating with overpressure.

When comparing the stress and pore pressure differential across a fault block it is important to
ensure that the stacking velocities have been 'picked' to sufficient accuracy. Figure 32 shows a
comparison between the sonic log and checkshot velocities from four wells within the study
area. These well velocities are compared to the stacking velocity over the seismic shotpoint
that is nearest to the respective wells (within 75m of the well by horizontal distance). The
velocities show a good correlation: however, the sonic logs display vertical resolution of less
than 1 m whereas the vertical resolution of the checkshot and stacking velocities are on the
order of tens of metres. This is important to consider within the context of differentiating

46
individual channel sands or intra-formational seals that may themselves be less than 1m thick.
However, within the context of determining gross formation stress and pore pressure
variation, the stack and checkshot interval velocities provide a workable resolution.

Hence, we contend that the stacking and checkshot interval velocities are valid data sources
for estimating empirically related properties such as pore pressure and stress in a geological
formation. Figure 33 shows the relationship between velocity and density, as described in
Section 4: (Methodology). The densities were derived from the velocity profiles over the
Rankin Trend and Exmouth Plateau. These velocity-based densities broadly match the density
trend derived from the NCB porosity trend.

RFT/FIT pressure tests (Fig. 34-35) in the Rankin and Keast/Dockrell field suggests that these
accumulations have a pressure differential of up to 750 psi. The Rankin 1 well RFT data is old
(1973) and perhaps not as precise as newer tests, however, it suggests that there is capillary
seal of at least 750 psi across the Rankin/Keast fault. The Keast 1 and Dockrell 1 Wells
indicate a pressure differential of 280 psi, suggesting that the N-S fault separating Dockrell
and Keast has a capillary potential of at least 280psi. Most wells follow the hydrostatic
gradient with notable exceptions including Zagreus, Orthrus and Gorgon (Fig. 34). With
respect to the Rankin Trend wells, Lady Nora 1 and Rankin 1 stand out, both overpressured
relative to the hydrostatic gradient at shallower depths. Most wells show a deviation from the
hydrostatic gradient around 3500 mTVDSS, with subtle variations, for example, Goodwyn 6
is slightly deeper, around 3600 m, whereas Rankin 1, located structurally on one of the
shallowest locations within the Triassic, has elevated pore pressure from 2900m.

47
 Comparison of velocity data for Comparision of velocity data for
Dockrell 1 Rankin 1
Interval Velocity (m/s) Velocity (m/s)
0 2000 4000 6000 8000
0 2000 4000 6000 8000
0 0
0 0

500 Sonic 500

500 Sonic Log
Checkshot 500
1000 1000 Checkshot Vint
HDVA SPT Stack
HDVA SPT Stack
1500 1000 1500 1000
Depth (mTVDSS)

Depth (mTVDSS)
TWT (ms)

TWT (ms)
2000 2000
1500 1500
2500 2500

3000 3000
2000 2000

3500 3500

2500 2500
4000 4000

4500
4500

Comparison of velocity data for Comparison of velocity data for

Keast 1 Goodwyn 6
Interval Velocity (m/s) Interval velocity (m/s)
0 2000 4000 6000 8000 0 2000 4000 6000 8000
0 0 0 0

500 Sonic
500
Sonic Checkshot 500
500 1000
Checkshot
1000 HDVA SPT Stack
HDVA SPT Stack 1500
1000
Depth (mTVDSS)

1500 1000
Depth (mTVDSS)

2000

TWT (ms)
TWT (ms)

2000 2500 1500

1500
2500
3000
2000
3500
3000 2000
4000 2500
3500
4500
2500
4000
3000
5000

Figure 32 Wells from the main fault blocks being studied show a comparison between the
interval velocities extracted from the HDVA stacks being examined and the checkshot and
sonic log velocities from each respective well. Good correlation supports the validity of using
either velocity source to derive stress and pore pressure relationships.

48
Pore fluid retention leading to overpressure is restricted to the Triassic, below the regional
claystone seals. The over pressure can be seen (Figs. 34-36) with the pressure deviation from
the hydrostatic trend occurring at approximately 3600 mTVDSS, which is the typical depth of
the regional claystone seal within the study area.

The effect of this overpressure can be seen in figure 36 where the effective pressure is not a
simple linear trend with depth. Rather, the effective pressure trend is concave downwards
with further reduction in effective pressure around 3500 mTVDSS. Further evidence for
effective stress reduction being due to overpressure is assessed through the LOT data from the
NCB. Figure 37 shows this data including FIT tests that recorded maximum pressures. A
linear trend throughout the NCB is evident and the maximum pressure data points scatter with
the normal distribution around the actual LOT data. Figure 37b shows just the LOT data and
the consistency of the trend. Figure 38 shows the methods based on an empirical Poisson’s
ratio (Eqn. 15) and porosity trend underestimate the Shmin at depth, when compared to values
derived from LOT data. The methods of (Hubbert and Willis, 1957; Zoback and Healy, 1984)
and (Eaton, 1969) provide a closer fit, with the Zoback and Healy (1984) and Hubbert and
Willis (1957) relations underestimating Shmin at depth and Eaton overestimating the stress. In
these relations, Eaton’s (1969) and Zoback and Healy’s (1984) relations use an effective
stress ratio of 0.32 (coefficient of internal friction (CIF) of 0.6), whilst that of Hubbert and
Willis’ is 0.38 (CIF 0.5). Eaton’s (1969) relationship provides a poorer fit for the Rankin
Trend alone (Fig. 39), although may fit better at greater depths.

The difference in Sh min derived from the velocity and from LOT is plotted in Figure 40. The
power law relationship resulting from the underestimation of Sh min from the velocity allows a
'correction factor' to be calculated. This correction factor assumes a calibration well is present,
nonetheless, a clear trend is discernable.

Figure 43a-d show the coulomb failure function described for the principal fault blocks in the
study. Described in Section 4, these functions are graphed for a fixed orientation
representative of the main graben bounding NE-SW faults that separated Rankin from Keast
and Dockrell. A constant coefficient of sliding friction of 0.6 is assumed. The CFF suggests,
that under the current stress regime, these fault blocks should not consist of a stress regime
that would lead to fault slip for a plane representing the Rankin/Keast/Dockrell N-S faults
(Fig. 10). However, when the Shmin derived from the P-wave interval velocity is employed, an

49
erroneous data set is found. When the velocity-based CFF with a correction factor applied
shows a broad consistency with the LOT-based CFF.

Pore pressures are calculated from Bowers’ (1995) relation with a Biot factor of 0.8 (see
Appendix D for details) and pore throat radius calculated from (Pittman, 1992) relation. With
empirically derived pore pressure, Eaton’s (1969) or Zoback and Healy’s (1984) equation can
be used to calculate the CFF in place of a ‘corrected’, well-dependent CFF. The calculated
pore pressure correlate well with the RFT measured pore pressures (Figs. 44-45). The pore
pressure variation across the area is broadly consistent, with some scatter coming from the use
of checkshot velocities, particularly in the case of Keast (Fig. 46). However, the calculated
pore pressures from stacks, mapped on to the base Norian surface in the case of figure 47,
suggest there is a lateral variability.

The degree of this lateral variability further is assessed by looking at pore pressure
distribution for a constant depth(time) over a time slice of 2300 ms TWT. Eliminating the
depth-dependent pressure effects that occur with a depth-varying surface, it shows that a
lateral variability in pore pressure remains. However, anomalous velocities (and hence
pressures) are observed.

The CFF was subsequently calculated using Bowers (1995) pore pressure and LOT based
Shmin and Zoback and Healy (1984) based Shmin (Fig. 49). Both methods suggest that even
within the overpressure window there is little risk of reactivation on the modelled N-S fault.
Final CFF’s used stacking velocities rather than checkshot velocities (Figs. 50-52). The three
primary fault planes representing the Rankin escarpment, Rankin/Keast fault and small East-
West Rankin Block faults are modelled. The effects of pore pressure in the overpressure
window can be seen. Whereas the Rankin/Keast fault is stable, the escarpment fault has a
positive CFF at depths equivalent to the top of the Triassic. Within the current stress regime,
the escarpment fault may still be subject to failure in places.

50
 Density Relationships
Density (g/mm3)
1.5 1.7 1.9 2.1 2.3 2.5 2.7 2.9
0.00

500.00

1000.00

1500.00
Depth (mBML)

2000.00

2500.00

3000.00

3500.00

4000.00 Gardner 1974

Castagna 1993
4500.00
Swarbick and Osborne 1998
5000.00

Figure 33 The density relationship with depth for the Rankin Trend, NCB. Scatter in the
density derived from interval velocities reflects differing lithologies and pore pressures,
whereas Equation 10 (Swarbrick and Osborne, 1998) relates an empirically derived porosity
with depth trend to density, hence, the smooth exponential type curve.

Figure 34 The formation pressures of selected wells from the Rankin trend and Exmouth
plateau. The lithostatic gradient is based on core measured Rankin trend porosities resulting in
the Exmouth plateau well, Jansz 3 exceeding it.

51
 Rankin Trend Formation Pressures
Pressure (PSI)
2000.00 3000.00 4000.00 5000.00 6000.00 7000.00 8000.00 9000.00 10000.00
1000.0
Dixon 2

1500.0 Dixon 1

Exmouth Plateau wells Rankin 1

2000.0 Dockrell 1
Zagreus 1*
2500.0
Depth (mBML)

Orthrus 1*

3000.0 Maenad 1*
Goodwyn 1
3500.0 Goodwyn 6
Keast 1
4000.0 example of pressure gradient Sculptor 1
variation due to overpressure.
4500.0 Keast 2
Peublo 1
5000.0

Figure 35 With the exception of the *wells from the Exmouth Plateau, the Rankin Trend
wells all show changes in formation pressure gradient associated with the overpressure from
disequilibrium compaction. Goodwyn 6 is pointed to as its numerous data points make the
gradient change easier to see.

Rankin Trend Effective Pressures

Pressure (PSI)

2000 3000 4000 5000 6000 7000 8000 9000 10000

2000

2500

3000
Dixon 2
Depth (mBML)

Dixon 1

3500 Rankin 1
Dockrell 1
Goodwyn 6
4000
Keast 1
Keast 2
Lithostatic Pressure Gradient
4500
Hyrdrostatic Pressure Gradient
Linear (Lithostatic Pressure Gradient)
5000

Figure 36 Effective pressures from wells in the study area. They show a concave downwards
trend as overpressure at depth of 3400m and below reduces the effective stress and associated
stress to depth pressure gradient.

52
 Depth TVDSS
Depth TVDSS

a)

Depth TVDSS
Depth TVDSS

b)

Figure 37 XLOT and ISIP pressures from 47 wells from the NCB show a linear trend
between depth and inferred Sh min at 0 to 4 km depth below sea floor. It suggests a broad
consistency in regional Sh max throughout the NCB and well as a consistent vertical to
horizontal stress ratio (data from the CSIRO pressure plot TM database).

53
 Comparison of Shmin empirical methods to Shmin from LOT
Shmin (psi)
0 2000 4000 6000 8000 10000 12000
0.00
Shmin from empirical
500.00 Poisson
Shmin from Porosity
1000.00 relation
Shmin from Hubbert &
Willis (1957)
1500.00 Shmin from Zoback & Healy
(1984)
Depth mTVDSS

2000.00 Shmin from LOT trend

2500.00 Shmin Eaton (1969)

Linear (Shmin from LOT

3000.00
trend)
3500.00

4000.00
Increase in uncertainty due to increase in
4500.00
lithological heterogeneity (see Appendix 11) Underestimation in Shmin by
5000.00 bilateral constrain based relation

Figure 38 Comparison of the main empirical relationships used to determine Shmin from
interval velocities. Bilateral constraint based relations such empirical Poisson and Porosity
based trends underestimate the Sh min at depth.

Comparison of Shmin empirical methods accounting for bilaterial

constraint on the Rankin trend
Shmin (psi)
0 2000 4000 6000 8000 10000 12000 14000
0.00
Eaton (1969)
500.00
Shmin from Hubbert & Willis
1000.00 (1957)
Shmin from Zoback & Healy
(1984)
1500.00 Shmin from LOT
Depth mTVDSS

2000.00 Overburden Stress psi

2500.00

3000.00

3500.00

4000.00

4500.00

5000.00

Figure 39 Based on the assumed effective stress ratio of 0.32, Zoback and Healy's and
Hubbert and Willis' relations provides a good fit to LOT trends, Eaton's tends to overestimate
at depth.

54
Figure 40 The ratio of Sv and Shmin is plotted for the Rankin trend wells in the study area and
compares Shmin estimated from LOT data to Shmin calculated from an empirical relation. The
green calculated values have an average value of 0.62, however, tend to underestimate the
ratio compared to the LOT based data which has an average of 0.64. Apart from a deviation
around 3000m, the ratio has a small decrease with depth due to the gradual increase in
overburden stress.
Difference plot between velocity-based and LOT-based
Shmin
6000.00

5000.00
difference Shmin (psi)

4000.00

3000.00

2000.00
𝑦 = 0.0002𝑥 2 + 0.1355𝑥
1000.00
R2=0.78

0.00
0.00 500.00 1000.00 1500.00 2000.00 2500.00 3000.00 3500.00 4000.00 4500.00 5000.00

Depth mTVDSS

Figure 41 plots the difference in Shmin where σ3 is derived from shear wave velocity using
Castagna's (1993) relation and σ3 is inferred from the LOT measurements. This presents a
'correction factor' that can account for the error in the assumption of bilateral constraint. The
validity of using shear wave based Sh min may be appropriate where this difference is known,
e.g. from a reference well (see Appendix E for validity check).

55
 Comparison between Pore Pressure, σv and Shmax for selected Wells on the Rankin Dixon 1
Trend Pressure (PSI) Rankin 1
1000.00 6000.00 11000.00 16000.00 21000.00 Dockrell 1
2000.0 Goodwyn 6
Keast 1
Keast 2
2500.0 Lithostatic Pressure Gradient
Hydrostatic Pressure Gradient
sHmin
3000.0
Dixon 1 sHmax
Depth (mBML)

Dockrell 1 sHmax

3500.0 Rankin 1 sHmax

Goodwyn 6 sHmax
Echo 1 sHmax
4000.0 Keast 1 sHmax
Dixon 1 σv
Rankin 1 σv
4500.0 Dockrell 1 σv
Goodwyn 6 σv
Echo 1 σv
5000.0
Keast 1 σv

Figure 42 Comparison between pore pressure and vertical and minimum horizontal stress
components. The graph shows effective stress reduction where pore pressure increases.
Furthermore, σ1 remains in the vertical direction, suggesting a normal/strike slip tectonic
regime.

56
 a Coulomb Failure Function using LOT stress b Coulomb Failure Function using Vs from Castagna's (1985) relation
CFF (psi)
``
CFF (psi)
-1400.00 -1200.00 -1000.00 -800.00 -600.00 -400.00 -200.00 0.00 -600.00 -400.00 -200.00 0.00 200.00 400.00 600.00 800.00
0.0 0.0
Rankin 1 Rankin 1
500.0 500.0
Dockrell 1 Dockrell 1
1000.0 1000.0
Keast 1 Keast 1

Depth (mTVDSS)
Depth (mTVDSS)
1500.0 1500.0

2000.0 2000.0

2500.0 2500.0

3000.0 3000.0

3500.0 3500.0

4000.0 4000.0

Coulomb Failure Function using Vs with empirical correction factor Coulomb Failure Function using Shmin derived from Zoback and Healy's
c applied
d (1985) relation
CFF (psi) CFF (psi)
-1600.00 -1400.00 -1200.00 -1000.00 -800.00 -600.00 -400.00 -200.00 0.00 -1400.00 -1200.00 -1000.00 -800.00 -600.00 -400.00 -200.00 0.00
0.0 0.0
Rankin 1 Rankin 1
500.0 500.0
Dockrell 1 Dockrell 1
1000.0 1000.0
Keast 1 Keast 1

Depth (mTVDSS)
Depth (mTVDSS)
1500.0 1500.0

2000.0 2000.0

2500.0 2500.0

3000.0 3000.0

3500.0 3500.0

4000.0 4000.0

Figure 43 Coulomb Failure Function for a fault plane that is representative of the N-S steeply dipping faults that separate the Rankin/Keast and
Dockrell blocks (Schematic of the theoretical plane used can be seen in figure 56). The CFF is calculated using Pp and LOT measurements at a
depth which has Norian (Triassic) Mungaroo formation geology in the Rankin Trend. a) The shear stress is measured from LOT data, are all
below zero suggesting no prevalence for failure. b) The shear stresses are derived from the seismic velocities using Vp to Vs conversion for an
estimation of Poisons ration. C) a correction factor is applied accounting differences between LOT-based shear stress and Vp based shear stress
d) shear stress calculated assuming a fixed principal stress ratio i.e. fixed ESR of 0.32, as per Zoback and Healy (1984) or Eaton's (1969)
equations. The figures show the degree to which the 'corrected' c, or calculated d, CFF broadly correlate with a, the LOT based CFF.

57
 Pore pressure comparison between RFT and velocity
measurements on Keast 1
Pore Pressure (psi)
0.00 1000.00 2000.00 3000.00 4000.00 5000.00 6000.00 7000.00 8000.00
0
Velocity based pore pressure
500
RFT
1000
TVDSS(mTVDSS)

HHDVA stack SPT

1500
(ms)

2000
Depth

2500

3000

3500

4000

Figure 44 Comparison of pore pressure from checkshot interval velocity, RFT and HDVA
velocity stack. The velocity picks from the stack closely match the least squares fit of the
checkshot.

Pore pressure comparison between RFT and velocity

measurements on Rankin 1
Pore Pressure (psi)
0.00 1000.00 2000.00 3000.00 4000.00 5000.00 6000.00 7000.00 8000.00
0.00
Velocity based
500.00
pore pressure
1000.00 RFT measured
pore pressure
1500.00
(mTVDSS)
(ms)

2000.00
TVDSS

2500.00
Depth

3000.00

3500.00

4000.00

4500.00

Figure 45 Comparison between Rankin 1 RFT measured pore pressure and calculated pore
pressure using the Bowers’ (1995) relationship with velocity. RFT measurements are biased
towards the target reservoir (only measured at the depth corresponding to reservoir), but
show a reasonable correlation.

58
 Pore Pressure from Velocity on the Rankin Trend
Pore Pressure (psi)
0.00 2000.00 4000.00 6000.00 8000.00 10000.00
0.00 Goodwyn 6

Rankin 1
500.00
North
Rankin 1
1000.00 Yodel 1

Keast 1
1500.00
Depth mTVDSS

Sv
2000.00

2500.00

3000.00

3500.00

4000.00

Figure 46 The estimated pore pressure trend based on checkshot velocities from the
respective wells in the list is broadly consistent across the different fault blocks within which
the wells are located.

59
Pore pressure (psi)
Figure 47 Pore pressure variation across the
base Norian horizon. Higher pore pressures are
focussed around fault boundaries, as well as
broadly within the Rankin High and Dockrell
Highs and near the Goodwyn block. Pore
pressures on the scale bar are in psi. This map
does not indicate pressure differential across
faults as the horst and graben are at different
depths.

Higher pore pressure in the shallower section of the base

Norian horizon- localised around the Rankin, Dockrell and
Goodwyn 6 blocks

60
 Poor velocity picks in the lower
(red) corner of the study area
or everywhere south of the
Velocities anomalously high in this area
Rankin escarpment.
produce anomalously low pore pressures
Velocity m/s

Pore Pressure (psi)

Comparison of velocity data for

Dockrell 1
Interval Velocity (m/s)
0 2000 4000 6000 8000
0 0

500 Sonic
500
Checkshot
1000
HDVA SPT Stack
1500 1000
Depth (mTVDSS)

TWT (ms)

2000
1500
2500

3000
2000
As Dockrell 1 velocity comparison shows, 5 km 5 km
3500

4000
2500
the stacking velocities in the SE corner of
4500
the study area poorly match calibration
below 3100 m(2000 ms TWT).

Figure 48 Observation of lateral variation in a) velocity and b) pore pressure from a time slice at 2300 ms TWT shows good consistency with
values for most of the study area. However, anomalous results are evident in the lower corner (see schematic top left corner) of the study area
that is south of the Rankin escarpment. This is shown to be due to poorly picked stacking velocities in this part of the study area (inset). Whilst
relative variation is useful, absolute values are only as good as the velocity picks.

61
 Coulomb Failure Function using Shmin from LOT and Pore Pressure from
Bowers relation
CFF (psi)
-3500.00 -3000.00 -2500.00 -2000.00 -1500.00 -1000.00 -500.00 0.00
0.00
Rankin 1
500.00

1000.00

1500.00 Dockrell 1
Depth (mTVDSS)

2000.00

2500.00
Yodel 1
3000.00

3500.00

4000.00 Keast 1

4500.00

a)
Coulomb Failure Function using Shmin from Zoback & Healy (1984) and Pore
Pressure from Bowers relation
CFF (psi)
-3000.00 -2500.00 -2000.00 -1500.00 -1000.00 -500.00 0.00
0.00 Rankin 1

500.00

1000.00
Dockrell 1
1500.00
Depth (mTVDSS)

2000.00

2500.00
Yodel 1
3000.00

3500.00

4000.00 Keast 1

4500.00

b)
Figure 49 a) Coulomb Failure Function using LOT-derived Shmin shows the fault plane
representing the Rankin/Keast fault should be stable, albeit with relatively reduced stability
between 2000 and 3000m BML due to increased pore pressure. The Zoback and Healy (1984)
relation is similar, but with reduced variation due to pore pressure.

62
 CFF for fault plane representing the Rankin/Keast Fault with Rankin Block
Pore Pressures
CFF (psi)
-3000.00 -2500.00 -2000.00 -1500.00 -1000.00 -500.00 0.00
0

1000

2000
Depth (mTVDSS)

3000

4000

5000

6000

Figure 50 Coulomb failure function profile using stacking velocities and the empirical
relations from Zoback and Healy (1985) and Bowers (1983).The stacks have reduced
variability, but are broadly consistent with the checkshot CFF.

CFF comparison for three fault planes representing the main fault within
the Rankin Fault Block with Rankin Block Pore Pressures
CFF (psi)
-3000.00 -2500.00 -2000.00 -1500.00 -1000.00 -500.00 0.00 500.00 1000.00 1500.00 2000.00
0

1000
Rankin BlocK: Fault plane 010/50/E
Rankin Block: Fault Plane 070/40/E
Rankin Bloock: Fault Plane 080/60/W
2000 Depth (mTVDSS)

3000

4000

5000

6000

Figure 51 The CFF for three fault planes in the study area. The escarpment fault in the only
one within the current stress regime that high risk of mode I or II failure. In this calculation,
the intermediate stress is approximated as 0.9σ1.

63
Further analysis of the possible range of hydrocarbon column heights was made. Based on
mercury injection permeability and porosity tests on core sample at STP conditions.
Combined with NCB porosity trends for the Norian Mungaroo (Appendix C), it presents a
range of possible capillary pressures and HC heights for shale and sandstone. At depths
equivalent to typical oil/gas-water contacts, sandstone capillary pressures measure
approximately 500 psi, with HC heights up to 126m. Shale on the other hand results in
calculated capillary pressure of up to 3700 psi and HC heights of over 1000m. Mixed
lithology trend lay somewhere in between and are representative of low Vshale sands due to
the statistical bias in the porosity trends (see Appendix C).

Capillary Pressure from Mungaroo Porosity Trends on the NCB

Capillary Pressure (psi)
0 500 1000 1500 2000 2500 3000 3500 4000
0.00
Sandstone
500.00

1000.00 Shale

1500.00
Depth (TVDSS)

Sst
2000.00 dominant
mixed lith
2500.00

3000.00

3500.00

4000.00

4500.00

Figure 52 Theoretical capillary pressures for sandstone, shale and sandstone dominant mixed
lithologies calculated using mercury injection STP trends and Mungaroo core based porosity
trends. Maximum capillary pressures increase with depth and are significantly higher for
shale vs. sandstone.

64
 HC column heights from NCB porosity trends in the Norian Mungaroo
HC column height (m)
0 200 400 600 800 1000 1200 1400
0.00 Sst

500.00
Shale

1000.00
Sst
Depth mTVDSS

1500.00 dominant
mixed lith
2000.00

2500.00

3000.00

3500.00

4000.00

Figure 53 Resulting permissible HC column heights based on theoretical capillary pressures

calculated from Mungaroo porosity trends. This shows that even for moderately porous
sandstone, significant HC heights can be sustained at depth.

65
6. DISCUSSION/ANALYSIS

Structurally, the Rankin and Dockrell/Keast fault blocks show evidence for fault reactivation
through to the Neogene (Fig. 12-18). However, whilst multiple episodes of reactivation are
evident, this reactivation is highly partitioned, with main horst-graben bounding faults
showing differential offset within the Triassic, up to the Jurassic/Cretaceous unconformity.
Within the Cretaceous and Cenozoic, reactivation is focused on NE-SW bounding faults of
the Rankin escarpment and Goodwyn/Dockrell blocks (Fig. 12- 18).

The faults show characteristic features associated with oblique slip which, along with varying
dip of strata indicates that oblique slip is associated with fault block rotation during extension.
The orientation of fault arrays and classic rhombohedral geometries within the fault blocks
also suggests an element of oblique slip associated with seismicity during the formation of the
horst-graben architecture (Fig. 14). A rose diagram of the fault orientation also has an array
consistent with oblique extension (Fig. 54).

Horst Graben orientation

Escarpment fault orientation

Figure 54 The fault orientation within the Upper Triassic (Rose diagram) is consistent with a
typical array that results from oblique extension. As the range of angles of obliquity shown
modelled by (Withjack and Jamison, 1985). The angle of maximum extension shows
inversion oblique to extension. The pink line is in the direction of maximum extension, red
arrows indicate Sh max orientation: approx. 110 o.

66
Whilst fault offset maintains net normal displacement, subtle inversion features are evident
around the Rankin escarpment which shows evidence of Neogene reactivation. Fault
reactivation along the horst-graben bounding faults that separate the fault blocks show signs
of late reactivation to be predominately via oblique slip. Stereonets in figure 57 using
Angelier's (Angelier, 1979) method, show the likely direction of slip for these reactivated
faults is consistent with the right lateral horizontal movement that is evident (Fig. 14, 17, 19,
20). However, this remains comparatively minor and the faults have broadly accommodated
very little strain throughout the Cenozoic. This evidence suggests little seismicity along these
faults, which are likely to be relatively stable in the present day. Some indication of the
validity of this assertion comes from passive seismic data recorded on the NWS. Figure 55
shows all earthquakes recorded at a depth of less than 10km since 1954 (when recording
began). The only earthquake measured near the study area was in proximity and likely
associated with reactivation of the Rankin escarpment fault.

Rankin escarpment bounds the

southern edge of NWSJV fields
3.7
1991 3.0
1990

1976 2.7
1997 3.1 3.2
5.6
1986 1997 1991
3.7 3.2
1986 4.0
1987
3.3 19791984
2002 3.4 4.8 3.2
1979

Figure 55 Yellow circles are earthquakes recorded since 1954 which were located at depths
of 10km or less. Magnitudes 2 or greater (non labelled <2) (data from (Geoscience_Australia,
2014)).
Local borehole breakout data from around the study area show present day Shmax orientations
to be approximately NW-SE. This is relatively consistent with the far field Shmax for the NCB,
reported to be approximately 110o (Hillis et al., 2008; Hillis and Williams, 1993b). The fault
array shows a geometry that is suggested to be consistent with far field stress dominance,
which is typical for the NCB (Dyksterhuis and Muller, 2008; Hillis et al., 2008; Keep et al.,
2002).

67
Whilst there is lateral variation in velocity within the Triassic, the variation across the faults
themselves is subtle, if not typically indistinguishable. A notable exception is within the
Rankin Block near the Rankin 1 well, across a small synthetic fault (Fig. 21). Whilst there is
little differential across the faults, this is a relative velocity reduction around the fault damage
zone. This suggests that there is some change in the petrophysical properties in the lithology
immediately adjacent to the fault surface. As that change is likely to be a combination of
density and elastic response, it may be indicative of the greater damage zone envelope of the
fault. The fact that there is little velocity variation across the faults separating the Rankin
High and the Keast Graben is interesting in so far as the Rankin/Keast fault juxtaposes fluvial
dominated Mungaroo units with dominantly deltaic Brigadier lithology. This suggests that the
two distinct formations must have similar gross petrophysical properties as well as reducing
the prospect of the velocities being able to show subtle pressure differences, which the RFT
data (Fig. 35) indicates should exist.

However, as figure 25 illuminates, the velocity gradient is highly sensitive to formation

boundaries. The shift from one major formation to another can be picked from the velocity
gradient shift alone. Particularly prominent is the velocity reversal on the Rankin Trend due to
the 'siliciclastic' formation in the Eocene (see Fig 5), which is thin to absent in the offshore
Exmouth Plateau, hence no velocity reversal manifests over it. This suggests that whilst the
particular stacking velocities used in this study may not have sufficient resolution to detect
minor fluid pressure changes, the velocities are sensitive to the contrast between different
bulk petrophysical properties of different formations. The Vp/Vs ratio (Fig. 31) is a typical
indication of lithology, particularly when correlated with acoustic impedance. Whilst the
Vp/Vs ratio approximates 2, it also shows subtle variation, particularly at depths that
correspond to overpressure zones and significant lithology shifts such as we see in a
formation change.

The 'siliciclastic' formation referred to is part of the Winning Group where claystones at the
base of the Wilcox formation grade into limestones of the Walcott formation with a mix of
claystones and minor interbedded standstones (Apthorpe, 1988; Hocking et al., 1987). Whilst
this lithology is not likely to be particularly slow, it is in relation to the underlying and
overlying lithology, hence the velocity reversal that is evident. This also highlights caution in
making simplistic interpretations of lithology based on velocity, because any velocity reversal
or increase is going to be relative to the overlying/underlying geology.

68
Rather than absolute velocities, an understanding of the Poisson ratio extracted from these
velocities provides a more robust insight into the type of lithology, be it carbonate, shaly or
clean sands. Furthermore, studies have reliably correlated Vp/Vs vs. Acoustic Impedance (AI)
correlations to infer general lithology type (Pendrel et al.; Saussus and Jarvis; Vernik et al.).
This is also the premise to studies that use impedance inversion to either deterministically
match elastic properties of lithology to seismic data, or model seismic data based on lithology
type with a stochastic type feedback to acquire a match to a set of calibration wells (Pendrel et
al.; Rowbotham et al., 2003; Spikes and Dvorkin, 2003). Lithology determination from a
stochastic model-based method is independent of velocity error and far more accurate,
although, requires a calibration well in the first place and is computationally demanding. So a
simple Vp /Vs vs. AI relation can still be more useful in the frontier basin.

Figure 26 shows the base Norian TWT structure map and associated velocity map of the
varying interval velocity for a constant lithology. It shows that within a single formation,
there remains significant velocity variation which is highly dependent on depth. The deeper
the horizon, such as in the Keast Graben, the faster the velocity. The variation is great enough
to exceed any possible distinct observation of a pore pressure differential across a fault, yet
again does provide a distinct variation with depth and pressure.

The density can be seen in figure 33 where the variation in lithology creates a significant
variation in bulk rock density through the formations. This density is empirically derived from
the velocity and broadly correlates with the porosity-based trend of equation 10. The porosity-
based trend is derived from core log porosities over the entire NCB. The velocity based
densities from the Rankin Trend and Exmouth Plateau reflect the general NCB trend, and
confirm a high degree of validity for the density relationships investigated.

It is through these velocity-derived densities that we are able to calculate Sv and Shmin to place
some preliminary geomechanical constraints on the lithology. This is compared to the LOT
measured Shmin which, as figure 37 shows, is not only consistent across the entire NCB, but
increases linearly. The implications of the linear relation between stress and strain, i.e. the
elastic moduli of the material, can be seen in the linear increase in Sh min with depth
(overburden stress).

Nonetheless, as figure 38 shows, calculating Shmin from a porosity-based or velocity-

determined Poisson ratio, underestimates the Shmin at depth. This is because they rely on
bilateral constraint to determine the horizontal stress. Bilateral constraint refers to the

69
expansion of a material and hence increase in horizontal stress, when it is squeezed. This is
the property of elasticity that the Poisson ratio measures, hence the reliance on the empirically
determined Poisson ratio to assess the Shmin. Calculating the stress strain ratio from bilateral
constraint relies on a linear relationship, however, as Appendix C highlights, porosity, and
therefore density and the associated bulk modulus of a material will depend on a negative
exponential relation with increasing pressure (depth). Thus the stress strain ratio appears non-
linear where it is determined from P-wave velocities. Zoback and Healy’s (1984) relation,
Hubbert and Willis’s original (1957) relation and Eaton’s (Eaton, 1969) relation result in
more accurate values for horizontal stress and generally confirm the coefficient of friction’s
nominal value of 0.6, for a vast array of circumstances, hence its validity in use for a frontier
region for example. This is because these empirical relations fix the effective stress ratio. This
fixed ESR probably begins to break down at depth as the crust transitions from the brittle to
plastic elastic moduli regime or where nonlinear viscoelastic solids, such as certain clays and
salt, dominate the lithology. However, in sedimentary basin lithology, a constant ESR is
shown to remain particularly valid. This is confirmed in figure 40, where the ratio of Shmin to
Sv, is a little over 0.6. The ratio of vertical to horizontal stress is assessed as approximately
0.62 for the Rankin, Dockrell and Keast blocks.

The determination of Shmin by bilateral constraint underestimates true values and there is an
increased uncertainty in the bilateral-based empirical relation where there is a significant
heterogeneity in the lithology or pore fluid (see Appendix E). However, the relation is
remarkably consistent. The difference between the LOT-based and bilateral constraint
velocity-based Shmin shows a power-law relation with depth (Fig. 41). This consistency
facilitates the determination of a 'correction' factor to be applied to the velocity- based ratio.
Shmin can then be calculated using this correction factor, as an alternative method to the
‘assumed’ effective stress ratio of 0.32. However, as figure 30 shows, a distinct difference in
the trend between Rankin and the Exmouth Plateau is evident, which suggests that the
underestimation will vary according to broad differences in the petrophysical properties area
to area.

Furthermore, variation from area to area is likely to be enhanced by the fact that Castagna's
(1993) empirical relation used to determine Vs has not had its constants calibrated for the
formations being studied. Whilst this correction factor can account for underestimation in
certain circumstances, figure 30 shows that care must be taken not to assume a trend over an
entire basin; the sub-basins are likely to vary in their shear wave velocity response. If a well

70
calibration cannot be carried out for a study area, such a method should be avoided, and one
of the empirical relations using a fixed effective stress ratio used instead.

The distinct effect of overpressure on the effective stress (Fig. 42) manifests as a sharp
effective stress reduction on the Rankin trend at approximately 3000m and 3500 mTVDSS.
Whilst the CFF for the Rankin/Keast fault does not suggest any fault instability, it shows a
distinct deviation from the trend with depth where overpressures are reported. This is an
indication of the effect that overpressure and pore fluid retention can have on the stability of a
fault. The capacity to assess this for a range of fault blocks using seismic velocities alone, can
provide a useful initial constraint on fault risk and reservoir integrity. As figure 43 shows, for
a number of fault blocks, a similar pattern is observed, all being in the stable regime, but with
less stability around the 3000m mark where overpressure is reported.

Figure 43b shows the effect of using the incorrect Shmin from equation 15. The under
estimation of Shmin result in an incorrect shear stress and an unrealistic conclusion of fault
instability. We can apply a correction factor and in so doing acquire a reasonable
approximation on the CFF from velocities alone (Fig. 43c). The ‘corrected’ CFF shows
similar relative pressures and trends with a deviation around 3000m due to overpressure. As
previously noted, however, the problem with the corrected CFF is that the correction factor is
highly variable from area to area and cannot be applied on a general basis. Hence, where well
calibration for the study area is not available, the constant ESR methods should be applied.

Thus a confidence can be established in the calculated CFF’s used in velocity stacks. Figure
51 shows this for the theoretical planes which are illustrated in the schematic of figure 56 . It
is important to remember that the CFF is dependent on the relative strike and dip of a fault
with respect to the principle stress orientations. Faults will vary in dip and strike and a
thorough CFF analysis needs to ‘segment’ a fault to account for this variation. This study only
employed a single plane for each fault to establish the validity of the stack-based calculations.

71
 Max Max
Rankin/Keast fault ΔP 750psi ΔP 280psi
Rankin Keast Dockrell

East-West faults

water contact
relatively higher

Rankin Escarpment

a) b)

Figure 56 Schematic showing the theoretical planes representing the Rankin Escarpment,
Rankin/Keast/Dockrell faults and E-W faults modelled in the CFF plot using the stacking
velocities. The CFF plot suggests that the green planes are stable and orange planes have a
high mode I/II failure risk in the current Sh max regime. The escarpment fault, represented by a
orange plane, is the only fault to show evidence of reactivation in the Neogene. Maximum
pressure differentials are both in the Triassic. b) schematic showing base depths of Rankin,
Keast and Dockrell gas columns. Oil-water and gas-water contacts are in blue, depths are
mTVDSS. Heights of columns are relative only and not actual heights. Adjacent fault block
water contacts at different heights also suggest a degree of sealing capacity to the faults.

Nonetheless, the implications of the stack-based CFF’s are twofold: the capacity to acquire a
CFF assessment from seismic velocities alone, and the necessity to have either a correction
factor based on a reference well or the assumption of a fixed ESR. It is rare to not have at
least one exploration well in a basin, and whilst the comparison between the Rankin Trend
and Exmouth Plateau shows that petrophysical properties and seismic anisotropy should not
be over-extrapolated, a correction factor from a similar region could be adequate to provide an
initial approximate CFF. It is noted that a CFF based on LOT and sonic log data will always
provide a superior assessment of the CFF and stress conditions in general, and should not be
superseded by an approximation from seismic velocities. However, the contention remains
that where well based data is lacking, this framework provides far greater constraints on stress
and fault failure assessment than would otherwise be available.

72
 Goodwyn/Dockrell fault

Between 2500 and 4500 m

Base regional Cretaceous Uc TVDSS

Brigadier Fm

Mungaroo Fm

4 km

Above 2500m and Below

4500m TVDSS Whilst pore pressure reduces stability
on the Rankin/Keast fault, it does not
touch the failure line (for CIF =0.6)

Figure 57 Cross-section across the study area and TWT map of

the regional Cretaceous unconformity. The figure shows that with faults orientated NNW-SSE and ENE-WSW at depths between 2500 and 4000
mTVDSS (approximately 1600-3000 ms TWT) there is a higher risk of fault reactivation. Pore pressure plays a large part in making this range
unstable. The only faults that show evidence of reactivation in the Cenozoic are the Rankin escarpment and Goodwyn/Dockrell fault: both are in
higher risk orientations. Most fault-seal field limitations therefore appear to be due to capillary seal potential which ranges between 500 and 3500
psi depending on the clay content of the fault gouge.

73
The availability of pore pressure trends is a further constraint that can be made using the
seismic velocities. Bowers (Bowers, 1995) formula shows overlap between the RFT pressures
and velocity-based pore pressure. This indicates that the use of velocities to infer a pore
pressure range or shift may be useful where further constraints are absent. There remains a
considerable amount of uncertainty due to the scatter in the velocity data in some locations, be
it well location or shotpoint from the stack. This scatter is a product of the lithology varying
the velocity, rather than pore pressure, and leads to a standard deviation of up to 1800 psi in
the case of Keast 1 (Fig. 44). However, this deviation is lost in the stacking velocities, which
are representative of the least square mean value of the interval velocity at any given depth.
The checkshot velocities show the standard deviation to vary with depth, a reflection of the
level of heterogeneity in different lithological formations.

A pore pressure deviation from the normal trend could be due to overpressure or fluid
accumulation or may be a reflection of significant heterogeneity in the lithology. Furthermore,
whilst lateral variability is significant, figure 48 shows that poor velocity picks have a
significant effect on calculated pore pressure, giving rise to anomalous results. Hence caution
is required before taking velocity-based pore pressures as an absolute direct indicator without
well constraint. It is noted that good correlation between velocity-based pore pressure and
RFT pore pressure was only achieved where the Biot factor was taken into account (Appendix
D). This suggests that always assuming the Biot factor to be 1 may be erroneous in many
circumstances.

The degree of correlation between RFT and velocity-based pore pressures, LOT-based and
velocity-based CFF's, and log-based and velocity-based formation heterogeneity and density
variation is high. Using stacking velocities to calculate CCF and pore pressure presents an
inherent uncertainty that comes with the method by which they are picked. However, error
remains from the empirical relationship utilised to extract density and shear wave velocity.
This error could be reduced by calibrating the empirical relationships although this is
somewhat difficult in frontier basins. Error could also be reduced by using measured Vs rather
than using a calculated Vp:Vs relation, although this is often not available.

These same sources of error place uncertainty in pore pressure prediction, however, resolution
is also a key issue. Where vertical sampling of the stacking velocities is, in this case 32ms,
this can amount to more than 60m, combined with the half wavelength of the seismic wave;
on the order of 30m, single pay pressure differences can be quickly lost. Absolute pore

74
pressure determination remains important for effective stress calculations and providing
useful information for drilling in frontier regions where mud weights and pressure kick
tolerance may be poorly constrained. However, capillary seal potential remains dependent on
the pressure differential across the seal i.e. relative pressure, rather than absolute pressure.
Thus, in the instance that a calibration well was available, theoretical capillary pressures and
hydrocarbon column heads can be constrained. Pittman's (Pittman, 1992) equation was used
as it enables a constraint to be placed on a possible range in capillary pressures, where fluid
type and clay content in fault gouge is unknown. Sand based column heights give an
indication of possible sand-sand juxtaposition, arguably the weakest link in the fault seal
capacity. Sand capillary pressures based column heights closely match measured values (Fig.
58).

Comparison of recorded HC column heights to Hg capillary based

theoretical min and max heights (based on Mungaroo porosity
trends in the NCB)

10000
HC column height (log m)

1000

100

10

1
Rankin 1 Dockrell 1 Keast 1 Goodwyn 6

Sst/Sst HC Column potential Shale/Shale HC Column potential Recorded HC column

Figure 58A comparison of theoretical HC column heights calculated from mercury-based

capillary pressures and NCB Mungaroo (Norian) porosity trends. It suggests that where sand-
sand juxtaposition is taken as the ‘weakest’ link in fault seal, significant column heights can
still be achieved, assuming geomechanically stable lithology.

75
One must expect that if the fault is stable, a potential HC column height will sit somewhere
between the sand and shale values, depending on the clay content of the fault gouge. Where
column heights are less than the sand value, if the fault is stable, charge access may well be an
issue. These values are based on mercury capillary threshold pressures, thus whilst they can
provide a useful range, once the source rock was defined and fluid type assessed, they could
be constrained using relationships that take this into account (Berg, 1975; Busch and Amann-
Hildenbrand, 2013; Watts, 1987) Nonetheless they proved a useful first pass assessment,
particularly when considered in conjunction with the stress regime and CFF of the fault.

Thus we can see that seismic velocities and amplitudes provide useful constraints on fault seal
assessment where good well constraint is lacking. A correlation matrix (Fig. 59) summarises
that match between measured values and seismic velocity-based empirically calculated values
in this case study. A summary of an effective workflow is presented in figure 60. This shows
that depending on the well constraint available, constants can be refined over time to make
better predictions of the fault seal capacity.

Correlation Matrix
Correlation Key

Good correlation between Well -

based and Velocity-based risk
assessment

Figure 59 Correlation matrix highlights the good correlation between directly measured
values and seismic-based empirically calculated values. It supports the applicability of
calculating these values where well constraint is lacking.

76
Figure 60 Recommended work flow for using stacking velocities to calculate CFF, pore
pressure and capillary pressure for fault seal constraint.

77
7. CONCLUSIONS

This study has shown that a number of empirical relationships that are available enable a
robust approximation of pore pressures and stress magnitudes to be calculated from seismic
velocities. When combined with a basic fault architecture assessment (dip, strike, sense of
offset, and evidence of reactivation), an assessment of fault seismicity can be made. This
study successfully calculated a coulomb failure function using P-wave seismic velocities as
the only source of input. A subsequent porosity evaluation also enabled a theoretical
calculation to be made on the capillary pressures which further constrain fault seal risk. A
case study of the Rankin and Dockrell/Keast fault blocks shows that there is robust correlation
between measured data and empirically derived values. They show that the main faults
separating the Rankin and Dockrell and Keast fault blocks are in a stable stress regime and
not subject to further reactivation. They have an increased risk of failure around the base of
the top-sealing regional Jurassic and Cretaceous unconformity, where overpressure is
prevalent and principal stress ratios are large. This highlights the importance of taking all of
the following factors into consideration in fault seal risk: pore pressure, stress ratio, and dip
and strike of fault. All of these will affect whether a fault is high or low risk for either slip or
capillary-based leakage.

Rankin and Dockrell/Keast faults orientated approximately NNE-SSW in strike, with dips of
approximately 50o show no risk of slip, and have a minimum hydrocarbon column head
sealing potential of 120m .

The final uncertainty remains with the clay content of the fault gouge and fluid type, both of
which affect capillary seal potential. Whilst these two factors cannot be assessed from seismic
data alone, theoretical boundaries can be accommodated through assessment of potential
sand-sand and shale-shale potential. In this case, sand-sand capillary pressures were found to
be approximately 500 psi in the Late Triassic, corresponding to some capillary pressures
measured in the study area. Where source rock presence, maturity and type were all within
ideal windows, if the hydrocarbon column head was less than this, charge access should be
considered the primary accumulation risk.

78
8. FURTHER WORK

In using the velocity stacks for CFF and pore pressure range determination, a systematic risk
profile could be implemented for an entire region. This is in the spirit of studies such as those
conducted by (Castillo et al., 2000; Reynolds et al., 2005). Whilst general strikes and dips
may put groups of faults in a high risk or low risk category, this process would need to be
refined for prospect-level evaluation whereby trap-relevant faults would need to be
segmented. This is because a fault will typically vary in dip and strike and this variation
places each respective dip and strike segment under a slightly different stress. In addition pore
pressure plays an important role in creating fault instability and, as we have seen, pore
pressure also varies both laterally and vertically. Hence, the level of resolution considered in a
fault study should be selected based on whether a risk profile is needed for a regional or
prospect level.

This should be compared to a mature basin province for a statistical Vs empirical faulted trap
comparison. This would allow a statistically robust deterministic Vs empirical comparison to
be conducted between faults that have been recorded to have sealed with their risk profile, as
calculated from the seismic velocity empirical relations.

Vp/Vs ratio's also lend use for the purpose of lithology determination. Regional velocity cubes
can be inverted and Vp/Vs vs. acoustic impedance plots compared with velocity-based
Poisson ratio's to allow a relatively robust constraint on lithology in the frontier basin.

Finally, this study presents an example of the potential application of model-based full
waveform inversion (FWI) velocity data. The accuracy of FWI velocities can reduce the
uncertainty associated with standard velocity picks and present a broader scope for the
utilisation of the FWI process in future surveys.

79
9. APPENDIX A -TECTONOSTRATIGRAPHIC DIAGRAM
Overview shows the general stress regime and the main tectonic event association in the
NWS(Baillie and Jacobson, 1995; Cathro and Karner, 2006; Keep et al., 2007; Longley et al.,
2002).
Stress Regime Epoch (Age) Tectonic Event Deposition Event in Barrow Sub-basin
the Carnarvon Basin
Compressional Pliocene Generation of Timor Island
and Trough
7 Ma. Miocene Inversion
(Messinian)
10 Ma. Miocene Inversion Delambre Fm
(Tortonian)
13 Ma. Miocene Inversion Yardie Gp
(Serravallian)
18 Ma. Miocene Inversion Cape Range Gp
(Burdigalian)
Miocene Minor siliciclastic
deposition
Oligocene Australian and Eurasian Carbonate
(Banda Arc) Plates Collide sedimentation
Extensional Late Paleogene Final Separation of
(Oligocene) Australia and Antarctica
Early Paleogene Onset of Coral Sea Rifting
(Paleocene)
Late Cretaceous Minor Inversion due to Toolonga Calcilutite
intraplate deformation
resulting from AA breakup
and India migration.
Creation of Barrow
Anticline.
Early Cretaceous Australia/Antarctica Regional Transgression
Breakup deposits marine
(Muderong) Shale.
Australia/India Rift Crust formation of
Separation Argo & Gascoyne-
Cuvier Abyssal plains
Australia/India Rift Onset Transgressive
highstand deltaic
complexes
Late Jurassic West Burma Block III East and West
Separation Gondwana Break up,
West Burma Block III Rift submarine fans
Onset accumulate large
West Burma Block II volumes of sediment in
Separation syn-rift and eustatic sea
level rise and fall.
Middle Jurassic West Burma Block II Rift
Onset
Early Jurassic West Burma Block I Localised extension
Separation initiates NCB
Late Triassic West Burma Block I Rift Interbedded sand and Inversion
Onset muds deposited in a
fluvial deltaic coastal
(Mungaroo) complex
Lhasa Block Separation Pangea Breakup -
Locker Shale deposited
in marine settings
Permo-Triassic Regional Transpression Permian sediments
Boundary Tectonism comprise marine and
marginal marine
successions
Compressional Mid-Carboniferous Gondwana and Laurussia Late Carboniferous
collide to form Pangea glacial conditions
Devonian Uplift and warping, marine 1 Km of Devonian
conditions. Sediments with
Frasnian reef complex.
Ordovician-Silurian Carnarvon intracratonic
basin formation in
Gondwana

80
10. APPENDIX B – S TRATIGRAPHIC COLUMN FOR THE WEST DAMPIER
S UB-BASIN
(Bennett, 2005)

81
11. APPENDIX C – POROSITY AND PERMEABILITY IN THE NCB

The porosity and permeability trends for the NCB were acquired from 30 wells within the
NCB, biased towards wells along or near the Rankin Trend (Fig. 61). All porosity and
permeability data are from core analysis on samples from the Mungaroo formation of the
Triassic Period, NCB, NWS. This selection method was chosen so as not to 'skew' trends by
mixing formations. The Triassic was most pertinent to this study because fault seal risk
uncertainty is targeted in the Triassic. In addition, both mixed and single lithology trends have
been acquired, however, due to the core sampling bias towards the 'reservoir', sandstone has
the most reliable statistics.

Averaged Porosity vs Average Depth for the Triassic (Norian) Mungaroo Formation in the North Carnarvon Basin
Porosity %
5.00 10.00 15.00 20.00 25.00
2000.00
Sirius 1 Briseis 1
Maenad 1

2500.00
Vinck 1 North Rankin 1
Gorgon 1 Glencoe 1
Noblige 1 Yodel 1 Keast 1
Gorgon 3 Echo 1
Griffin 1 Lady Nora 2
3000.00
Jupiter 1 North Rankin 2 Orthrus 1 Dockrell 1 Athena 1
Rankin 1
Depth (m)

Observation 1 Nimrod 1ST3 North Gorgon 4

Hilada 1A
Long Island 1
3500.00 Dailey 1 North Gorgon 3 West Tryal Rocks 1
Goodwyn 6
Malmsey 1

Bowers 1 Griffin 1
4000.00

Bluebell 1

4500.00

Figure 61 The arithmetic average for 32 Wells in the NCB. 4 Wells were excluded as their
inboard and shallow setting mark a statistically significant separation from the rest.

Broad trends in porosity with depth relations are very similar, irrespective of the lithology
(Fig. 62). The trends confirm to the typical (Mondol et al., 2007; Weaver, 1989; Wood, 2010)
exponential-type relationship between porosity and depth. Shales will tend to have a much
faster reduction in porosity with depth as the clay mineralogy dewaters and compacts. This

82
then slows where further porosity reduction is due to diagenetic reactions. Sandstone
compacts much less in the first instance, then slows where further porosity reduction is due to
diagenesis and cementation. The Mungaroo Formation is an intra-formational claystone
sandstone formation and it is the depositional environment that is by far the greatest
determining factor of the ultimate porosity trend and porosity permeability relationship in a
formation. Sorting and maturity, along with clay content, have a large impact on the ultimate
porosity permeability trend. We see this in figures 63 and 64 for example where the 'slope' of
the porosity permeability plot is different depending on well location and its depositional
setting, with wells in steep slope setting having coarser and typically cleaner sandstones.
Results from this study suggest that good correlation and control should allow depositional
setting to be inferred from the porosity permeability trend alone. However, in this instance,
the porosity permeability trend used in Pittman’s (Pittman, 1992) relationship was taken from
the Rankin Trend only wells, for sandstone lithology, as this lithology type is biased in the
RFT data and hence allows for more robust comparison. Capillary pressure determination and
subsequent hydrocarbon column height calculation were assessed using sandstone, shale and
mixed lithologies for comparison.

Porosity Permeability
100 m average Trends fordifferent
porosity for sandstone in the
lith types Triassic
in the Triassic (Norian) Mungaroo
(Norian) Mungaroo Fm Fm. -
North Carnarvon
Porosity % Basin
1.00 6.00 11.00 16.00 21.00 26.00 31.00
0
y = 3313.7e-0.047x
R² = 0.4176
500
y = 3262.8e-0.084x
R² = 0.9229
1000

1500
mBML

2000
y = 8389.8e-0.078x
R² = 0.5405
Depth (m)

2500 Sst
Depth

Sh
Slst
3000 Clst

3500

4000

4500

5000

Figure 62 Porosity with depth trends show typical exponential curves. Shale porosity
reduction occurs quicker than with sandstone, however, broad trends are similar. These
relationships are used for the Shmin and capillary pressure relations in the study.

83
 Poro Perm Trends for sandstone in the Triassic (Norian) Mungaroo Fm
- North Carnarvon Basin
100000

10000

1000
Permeability (mD)

100

10

Sst Sandstone
1 Calc Sltst Calcic Siltstone
Sst & Sh Sandy Siltstone
Ls Limestone
0.1 Sltst Siltstone
Clst Claystone
Sh Shale
0.01 Slty Sh Slty Clst
Slty Sst Silty Sandstone
Marl limy mudstone
0.001
0 5 10 15 20 25 30 35 40
Porosity %

Figure 63 Porosity vs. Permeability plot showing that lithology is not a primary determining
factor in the resulting porosity-permeability relationship.

84
 Nimblefoot 1
Clio 1

Bluebell 1
Poro Perm Trends for sandstone in the Triassic (Norian) Mungaroo Fm - NWS Bowers 1
100000.00 Briseis 1
Clio 2
Glencoe 1
10000.00 Gorgon 3
Gorgon 1
Coarse and/or Clean
Long Island 1
Nimblefoot 1 Glencoe 1
1000.00
North Gorgon 3
North Gorgon 4
North Gorgon 1
100.00
Observation 1
Permeability (mD)

Onslow 1
Sirius 1
10.00
West Tryal Rocks 1
Zeepaard 1
Gorgon 3 Goodwyn 6
1.00
Malmsey 1
Nickols Bay 1
Noblige 1
0.10 Fine and/or Shaly Vinck 1
Athena 1
Dockrell 1
0.01 Echo 1
Gaea 1
Keast 1
0.00 North Rankin 1
0.00 5.00 10.00 15.00 20.00 25.00 30.00 35.00 40.00 North Rankin 2
Porosity % Yodel 1

Figure 64 Porosity/log permeability plot highlighting how well location and associated
depositional environment area greater determining factor in the porosity-permeability
trend. Inset: Gamma ray and resistivity logs indicating (shale content) and GR trend
matching the depositional environment. Briseis 1

85
12. APPENDIX D - PORE PRESSURE

This study uses Bowers (1995) relation for pore pressure calculations using seismic velocities.
Bowers (1995) relation employs constants derived from a so-called ‘virgin’ curve. This
describes the increase in velocity effective stress relation with depth where fluids do not affect
the trend by reducing the effective stress. The virgin curve constants in this were derived from
a corrected sonic log from the Rankin 1 Well (Fig. 65). This sonic log curve was ‘corrected’
to account for hydrocarbon accumulations, hence giving the background or ‘virgin’ relation.

Virgin Pore Pressure trend based on a hydrocarbon corrected Vp

Sonic Log from Rankin 1
6000
y = 16.5x0.6
5000 R² = 0.6
Pore pressure (psi)

4000

3000

2000

1000

0
2000 4000 6000 8000 10000 12000 14000 16000 18000
Stress (psi)

Figure 65 Pore pressure Vs, effective stress from sonic log values, corrected for effects of
background fluid pressure. The power regression fit provides constants applied to Bowers
pore pressure empirical relationship.

In Bowers’ (1995) relation the mudline velocity is subtracted from the interval velocity and is
typically approximated as 1500 m/s; seawater velocity. In this study, the value of 1250 m/s
was used, as this relates to the velocity extrapolated to zero pore pressure, because the sea
column at the mudline exerts 100's of psi in pressure, depending on depth. The typical water
depth in this study was 120 m, relating to 190 psi, so such an extrapolation is considered to
better account for the effect of seawater column height on hydrostatic pressure.

86
Poroelasticity refers to the time-dependant deformation of a porous elastic solid.
Poroelasticity results in pore pressure within a porous solid responding to an external pressure
applied and affecting the response to that external pressure. Because of this response, the pore
pressure can affect the measured bulk modulus of the material. The Biot coefficient attempts
to account to the poroelastic effect on the bulk modulus (seismic response) of a material. This
study infers this poroelastic response effect on the seismic-based pore pressure calculations by
multiplying the seismic-based pore pressure by a poroelastic (Biot) correction factor. This
correction factor is estimated by taking the ratio of RFT pore pressure to seismic-based pore
pressure. The average correction factor is approximately 0.8. Whilst there is up to a 50%
variation in this factor, the results of applying 0.8 across all values provide a very good
correlation for wells and fault blocks, whereas assuming a correction factor of 1 leads to
erroneous results.

Poroelasticity correction factor

0.40 0.50 0.60 0.70 0.80 0.90 1.00 1.10 1.20
0.0
Rankin Trend and Exmouth Plateau Wells
500.0
Rankin
1000.0 Dockrell
Keast
1500.0
Depth mTVDSS

2000.0

2500.0

3000.0

3500.0

4000.0

4500.0

5000.0

Figure 66 A value for poroelasticity is taken from the ratio of empirically based pore pressure
and RFT based pore pressure. An average is taken as 0.8, seen to be the approximate median
in this scatter plot and which results in consistently good agreement between calculated and
measured pore pressure.

87
 Pore pressure comparison between RFT and velocity
measurements on Keast 1
Pore Pressure (psi)
0.00 1000.00 2000.00 3000.00 4000.00 5000.00 6000.00 7000.00 8000.00 9000.00
0

500 Velocity based pore

pressure
1000 RFT
TVDSS (ms)

1500 HDVA stack SPT

2000 Power (Velocity based

pore pressure)
2500
y = 96.684x0.3921
3000 R² = 0.7658

3500

4000

Figure 67 This pore pressure plot using a Biot factor of 1 does not fit the RFT data, rather it
is anomalously high. This is consistent with all the blocks, and the effect of applying the
correction factor of 0.8 can be seen in the results (section 5.2).

88
13. APPENDIX E - EMPIRICAL CALCULATION OF SHEAR WAVE VELOCITY

Figure 68 compares Vs measured from the sonic log of the Rankin 1 well to Vs calculated
using Castagna's (1993) Vp relation. The relation is based on an empirical 'fit' from laboratory
tests on feldspathic sandstone. The comparison plot shows good correlation with the linear
regression fit (pseudo mudrock line) shows a Vp of approximately 1500 m/s where Vs is zero.
The correlation plot (Fig. 69) is another qualifier of the appropriateness of using Castagna's
(1993) relation. Some scatter persists in the correlation plot, as is expected with heterogeneity
in the lithology. More scatter (a poorer fit in the empirical relation) occurs at depths where the
Mungaroo formation is penetrated, as well as at depths with significant gas flow is recorded.
Considering the lack of calibration of this empirical relation, the correlation was remarkably
good, considering the context of having a relation that could be applied in a frontier basin
where no calibration well is available.

Vp Vs plot comparing measured and calculated Vs

6000

5000

4000
Vp m/s

3000
Sonic log measurement

Vs calculated from Castagna's

2000 (1993) Vp relation
Linear (Sonic log
measurement)
1000

0
0 500 1000 1500 2000 2500 3000 3500
Vs m/s
Figure 68 Comparison of measured to empirically calculated S-wave velocity plotted vs. the
P-wave velocity. The plot show good correlation with the linear regression fit representing a
pseudo mudrock line. The mudrock line confirms a P-wave velocity of approximately 1500
m/s where shear wave velocity is zero.

89
 Correlation plot of measured vs. calculated Vs
4000
Vs (calculated from Castagna (1993) Vp relation)

3500

3000

2500

2000

1500 Depth range mTVDSS

● 2500-2800
1000
● 2800-3100
● 3100-3400
500 ● 3400-3700
● 3700-4000
0
0 500 1000 1500 2000 2500 3000 3500 4000
Vs (measured from sonic log)

Figure 69 The correlation plot of calculated vs. measured shear wave velocity. Increased
scatter indicates poorer correlation and therefore poorer validity of the empirical trend used.
There is a poorer fit at depths correlating with greater lithological heterogeneity and
overpressured zones.

90
14. APPENDIX F - SHEAR AND NORMAL STRESS

σn and σs are the normal and shear stress components of the principal stress tensors σ1 σ 2 and
σ3 . As described in the method, the principal stresses can be determined from a combination
of overburden stress, LOT measurements and either mud weight recordings, where well data
is available, or constraining the intermediate stress to one that fits the Andersonian tectonic
regime, as defined by evidence in the fault architecture. The three dimensional vector
components of these principal stress vectors acting on a fault plane, i.e. the normal and shear
stress, is calculated via a third order tensor, basis vector transformation, whereby the vectors
are rotated to the coordinate system of the fault plane from that of the orthogonal Cartesian
system of the principal stress vectors. The equations used in this study for the calculation of
σn and σs, as described by Zoback (2010) and Jaeger et al. (2007) are as follows:

𝜎𝑆 = 𝑎11 𝑎12 𝜎1 + 𝑎12 𝑎22 𝜎2 + 𝑎13 𝑎23 𝜎3 ...Eqn. 17

2
𝜎𝑛 = 𝑎11 2
𝜎1 + 𝑎12 2
𝜎2 + 𝑎13 𝜎3 ...Eqn. 18

where aij are the matrix components as follows,

𝑐𝑜𝑠𝛾𝑐𝑜𝑠𝜆 𝑐𝑜𝑠𝛾𝑠𝑖𝑛𝜆 −𝑠𝑖𝑛𝛾

𝐴= −𝑠𝑖𝑛𝜆 𝑐𝑜𝑠𝜆 0 ...Eqn, 19
𝑠𝑖𝑛𝛾𝑐𝑜𝑠𝜆 𝑠𝑖𝑛𝛾𝑠𝑖𝑛𝜆 𝑐𝑜𝑠𝛾

where γ is the angle between the fault normal and minimum principal stress vector, and λ is
the angle between the strike of the fault and the plane perpendicular to the minimum
principal stress vector.

91
15. APPENDIX G – THERMAL GRADIENT FOR NCB

Temperature Gradient - North Carnarvon Basin

Temperature ( C)
0.00 20.00 40.00 60.00 80.00 100.00 120.00 140.00 160.00 180.00
0.00
Gorgon and
Larsen Deep 1
1000.00

2000.00
Depth (mBML)

3000.00

4000.00

5000.00

6000.00

Figure 70 A thermal gradient of approximately 29.3 + 3.0 oC/km is measured from 28 wells
in the NCB. This thermal gradient is measured from a least squares linear regression fit
excluding values measured within the Gorgon field as it is consistently assessed as being
anomalously ‘hot’ compared with the rest of the basin. Larsen Deep 1 off the Exmouth
Plateau is another that appears relatively hot in comparison.

92
16. REFERENCES

Allen, P.A. and Allen, J.R., 2005. Basin Analysis: Principles and Applications. Blackwell Publishing,
Malden, MA.
Angelier, J., 1979. Determination of the mean principal directions of stresses for a given fault
population. Tectonophysics, 56(3–4): T17-T26.
Apthorpe, M., 1988. Cainozoic Depositional History of the North West Shelf. In: P.G. Purcell and R.R.
Purcell (Editors), Petroleum Exploration Society of Australia Symposium. PESA, Perth, pp. 55-
84.
Aydin, A., 2000. Fractures, faults, and hydrocarbon entrapment, migration and flow. Marine and
Petroleum Geology, 17(7): 797-814.
Bailey, W.R., Underschultz, J., Dewhurst, D.N., Kovack, G., Mildren, S. and Raven, M., 2006. Multi-
disciplinary approach to fault and top seal appraisal; Pyrenees–Macedon oil and gas fields,
Exmouth Sub-basin, Australian Northwest Shelf. Marine and Petroleum Geology, 23(2): 241-
259.
Baillie, P.W. and Jacobson, E., 1995. Structural Evolution of the Carnarvon Terrace, Western Australia.
Australian Petroleum and Exploration Association Journal, 35: 321-332.
Bennett, K., 2005. Demeter 3D Seismic Interpretation Report, Northwest Shelf Venture Area,
Carnarvon Basin, Australia., Woodside Petroleum Ltd.
Bense, V.F., Gleeson, T., Loveless, S.E., Bour, O. and Scibek, J., 2013. Fault zone hydrogeology. Earth-
Science Reviews, 127(0): 171-192.
Berg, R.R., 1975. Capillary pressure in stratigraphic traps. AAPG Bulletin, 59: 939-956.
Bowers, G.L., 1995. Pore Pressure Estimation From Velocity Data: Accounting for Overpressure
Mechanism Besides Undercompaction. SPE Drilling and Completion, 27488: 89-95.
Bretan, P., Yielding, G. and Jones, H., 2003. Using calibrated shale gouge ratio to estimate
hydrocarbon column heights. AAPG Bulletin, 87(3): 397-413.
Brown, A.R., 2004. Interpretation of Three-Dimensional Seismic Data. AAPG Memoir 42

SEG Investigations in Geophysics, No. 9. AAPG and the Society of Exploration Geophysicists, Tulsa,
Oklahoma.
Busch, A. and Amann-Hildenbrand, A., 2013. Predicting capillarity of mudrocks. Marine and
Petroleum Geology, 45(0): 208-223.
Byerlee, J., 1978. Friction of Rocks. Pure and applied geophysics, 116(4-5): 615-626.
Castagna, J.P., 1993. Petrophysical imaging using AVO, The Leading Edge. Society of Exploration
Geophysics, Tulsa, Oklahoma, pp. 172-178.
Castagna, J.P., Batzle, M.L. and Eastwood, R.L., 1985. Relationships between compressional-wave and
shear-wave velocities in clastic silicate rocks. Geophysics, 50(4): 571-581.
Castillo, D.A., Bishop, D.J., Donaldson, I., Kuek, D., deRuig, M., Trupp, M. and Shuster, M.W., 2000.
Trap Integrity in the Laminaria High-Nancar Trough Region, Timor Sea: Prediction of Fault
Seal Failure Using Well-Constrained Stress Tensors and Fault Surfaces Interpreted from 3D
Seismic, APPEA. APPEA, Brisbane, pp. 151-173.
Cathro, D.L. and Karner, G.D., 2006. Cretaceous–Tertiary inversion history of the Dampier Sub-basin,
northwest Australia: Insights from quantitative basin modelling. Marine and Petroleum
Geology, 23(4): 503-526.
Childs, C., Sylta, Ø., Moriya, S., Morewood, N., Manzocchi, T., Walsh, J.J. and Hermanssen, D., 2009.
Calibrating fault seal using a hydrocarbon migration model of the Oseberg Syd area, Viking
Graben. Marine and Petroleum Geology, 26(6): 764-774.
Dehandschutter, B., Gaviglio, P., Sizun, J.P., Sintubin, M., Vandycke, S., Vandenberghe, N. and
Wouters, L., 2005. Volumetric matrix strain related to intraformational faulting in
argillaceous sediments. Journal of the Geological Society, 162(5): 801-813.

93
Dewhurst, D.N. and Hennig, A.L., 2003. Geomechanical properties related to top seal leakage in the
Carnarvon Basin, Northwest Shelf, Australia. Petroleum Geoscience, 9(3): 255-263.
Dewhurst, D.N., Jones, R.M. and Raven, M.D., 2002. Microstructural and petrophysical
characterization of Muderong Shale: application to top seal risking. Petroleum Geoscience,
8(4): 371-383.
Dix, C.H., 1955. Seismic Velocities from Surface Measurements. Geophysics, 20(1): 68.
Dodds, K.J., Dewhurst, D.N., Siggins, A.F., Ciz, R., Urosevic, M., Gurevich, B. and Sherlock, D.H., 2007.
Experimental and theoretical rock physics research with application to reservoirs, seals and
fluid processes. Journal of Petroleum Science and Engineering, 57(1–2): 16-36.
Dvorkin, J.P., 2008. Yet another Vs equation. Geophysics, 73(2): 35-39.
Dyksterhuis, S. and Muller, R.D., 2008. Cause and evolution of intraplate orogeny in Australia.
Geology, 36: 495-498.
Eaton, B.A., 1969. Fracture gradient prediction and its application in oilfield operations. Journal of
Petroleum Technology, 246: 1353-1360.
Etheridge, M., McQueen, H. and Lambeck, K., 1991. The Role of Intraplate Stress in Tertiary (and
Mesozoic) Deformation of the Australian Continent and Its Margins: A Key Factor in
Petroleum Trap Formation. Exploration Geophysics, 22: 123-128.
Færseth, R.B., Johnsen, E. and Sperrevik, S., 2007. Methodology for risking fault seal capacity:
Implications of fault zone architecture. AAPG Bulletin, 91(9): 1231-1246.
Fisher, Q.J., Harris, S.D., McAllister, E., Knipe, R.J. and Bolton, A.J., 2001. Hydrocarbon flow across
faults by capillary leakage revisited. Marine and Petroleum Geology, 18(2): 251-257.
Fisher, Q.J. and Knipe, R.J., 2001. The permeability of faults within siliciclastic petroleum reservoirs of
the North Sea and Norwegian Continental Shelf. Marine and Petroleum Geology, 18(10):
1063-1081.
Gardner, G., Gardner, L. and Gregory, A., 1974. Formation velocity and density - The diagnostics
basics for stratigraphic traps. Geophysics, 39(6): 770-780.
Geoscience_Australia, 2010. Barrow Sub-basin, Northern carnarvon basin release areas w10-15, w10-
16, w10-17, w10-18 and w10-19. In: Department, of, Resources, Energy and Tourism
(Editors). Australian Government, pp. 1-42.
Geoscience_Australia, 2014. Earthquakes@GA, Australia.
Harrowfield, M., Cunneen, J., Keep, M. and Crowe, W., 2003. Early-stage orogenesis in the Timor Sea
region, NWAustralia. Geological Society of London, 160: 991-1001.
Hasegawa, S., Sorkhabi, R., Iwanaga, S. and Sakuyama, N., 2005. Fault-seal Analysis in the Temana
Field, Offshore Sarawak, Malaysia. In: R. Sorhabi and Y. Tsuji (Editors), Faults, Fluid Flow and
Petroleum Traps. AAPG Memoir 85, pp. 43-58.
Hatchell, P. and Bourne, S., 2005. Rocks under strain: Strain-induced time-lapse time shifts are
oserved for depleting reservoirs, The Leading Edge. Society for Exploration Geophysics.
Hillis, R.R. and Reynolds, S.D., 2003. In situ stress field of Australia. In: R.R. Hillis and M. R.D. (Editors).
Geological Society of Australia, pp. 49-58.
Hillis, R.R., Sandiford, M., Reynolds, S.D. and Quigley, M.C., 2008. Present-day stresses, seismicity and
Neogene-to-Recent tectonics of Australia’s ‘passive’ margins: intraplate deformation
controlled by plate boundary forces. In: H. JOHNSON, A.G. DORE´ , R.W. GATLIFF, R.
HOLDSWORTH, E.R. LUNDIN and J.D. RITCHIE (Editors), Geological Society London, Special
Publications. The Geological Society of London, pp. 71-90.
Hillis, R.R. and Williams, A.F., 1993a. The contemporary stress field of the Barrow-Dampier Sub-Basin
and its implications for horizontal drilling. Exploration Geophysics, 24(4): 567-576.
Hillis, R.R. and Williams, A.F., 1993b. The Stress Field of the North West Shelf and Wellbore Stability.
Australian Petroleum and Exploration Association Journal, 33: 373-385.
Hocking, R.M., Moors, H.T. and Van de Graaff, W.J.E., 1987. Geology of the Carnarvon Basin Western
Australia, Geological Survey of Western Australia
Hubbert, M.K., 1953. Entrapment of Petroleum under hydrodynamic conditions. AAPG Bulletin, 37:
1954-2026.

94
Hubbert, M.K. and Willis, D.G., 1957. Mechanics of Hydraulic fracturing. Petroleum Transactions of
the American Institue of Mechanical Engineers, 210: 153-169.
Jaeger, J.C., Cook, N.G.W. and Zimmerman, R.W., 2007. Fundamentals of Rock Mechanics. Blackwell
Publishing, USA.
Jones, R.M. and Hillis, R.R., 2003. An integrated, quantitative approach to assessing fault-seal risk.
AAPG Bulletin, 87(3): 507-524.
Kachi, T., Yamada, H., Yasuhara, K., Fujimoto, M., Hasegawa, S., Iwanaga, S. and Sorhabi, R., 2005.
Fault Seal Analysis Applied to the Erawan Gas-condensate Field in the Gulf of Thailand. In: R.
Sorhabi and Y. Tsuji (Editors), Faults, Fluid flow and petroleum traps. AAPG Memoir 85, pp.
59-78.
Kearey, P., Brooks, M. and Hill, I., 2002. An Introduction to Geophysical Exploration. Blackwell
Publishing, London.
Keep, M., Bishop, A. and Longley, I., 2000. Neogene wrench reactivation of the Barcoo Sub-basin,
northwest Australia: implications for Neogene tectonics of the northern Australian margin.
Petroleum Geoscience, 6: 211-220.
Keep, M. and Harrowfield, M., 2008. Elastic flexure and distributed deformation along Australia’s
North West Shelf: Neogene tectonics of the Bonaparte and Browse basins. In: H. JOHNSON,
A.G. DORE ´ , R.W. GATLIFF, R. HOLDSWORTH, E. LUNDIN and J.D. RITCHIE (Editors), The
Nature and Origins of Compressive Margins. Geological Society, London, pp. 185-200.
Keep, M., Harrowfield, M. and Crowe, W., 2007. The Neogene tectonic history of the North West
Shelf, Australia. Exploration Geophysics, 38: 151-174.
Keep, M., Longley, I. and Jones, R., 2002. Sumba and its effect on Australia's north-western margin.
Geological Society of Australia Special Publication, 22: 303-312.
Keep, M. and Moss, S.J., 2000. Basement reactivation and control of Neogene structures in the Outer
Browse Basin, North West Shelf. Exploration Geophysics, 31: 424-432.
King, R.C., Neubauer, M., Hillis, R.R. and Reynolds, S.D., 2010. Variation of vertical stress in the
Carnarvon Basin, NW Shelf, Australia. Tectonophysics, 482(1–4): 73-81.
Lockner, D.A., Morrow, C., Moore, D. and Hickman, S., 2011. Low strength of deep San Andreas fault
gouge from SAFOD core. Nature, 472(7341): 82-85.
Longley, I., Buessenschuett, C., Clydsdale, L., Cubitt, C.L., Davis, R.C. and 2002. The North West Shelf
of Australia - a Woodside perspective. In: M. Keep and S. Moss (Editors), Western Australian
Basins Symposium. Petroleum Exploration Society of Australia, Perth, pp. 27-86.
Losh, S., 1998. Oil Migration in a Major Growth Fault: Structural analysis of the Pathfinder core,
South Eugene Island Block 330, Ofshore Louisiana. AAPG Bulletin, 82(9): 1694-1710.
Mandl, G., 2000. Faulting in Brittle Rocks. Springer-Verlag, Germany.
Mondol, N.H., Biolykke, K., Jahren, J. and Hoeg, K., 2007. Experimental mechanical compaction of clay
mineral aggregates—Changes in physical properties of mudstones during burial. Marine and
Petroleum Geology, 24: 289-311.
Moore, D.E. and Lockner, D.A., 2004. Crystallographic controls on the frictional behavior of dry and
water-saturated sheet structure minerals. Journal of Geophysical Research: Solid Earth,
109(B3): B03401.
Moore, D.E., Lockner, D.A., Ito, H., Ikeda, R., Tanaka, H. and Omura, K., 2009. Geometry of the Nojima
Fault at Nojima-Hirabayashi, Japan - II. Microstructures and their Implications for
Permeability and Strength. Pure & Applied Geophysics, 166(10-11): 1669-1691.
Müller, R.D., Dyksterhuis, S. and Rey, P., 2012. Australian paleo-stress fields and tectonic reactivation
over the past 100 Ma. Australian Journal of Earth Sciences, 59(1): 13-28.
Ngwenya, B.T., Elphick, S.C., Main, I.G. and Shimmield, G.B., 2000. Experimental constraints on the
diagenetic self-sealing capacity of faults in high porosity rocks. Earth and Planetary Science
Letters, 183(1–2): 187-199.
O'Brien, G.W., Morse, M., Wilson, D., Quaife, P., Colwell, J., Higgins, R. and Foster, C.B., 1997. Margin-
Scale, Basin-Involved Compartmentalisation of Australia's North West Shelf: A Primary

95
 Control on Basin Scale Rift, Depositional and Reactivation Histories. Australian Petroleum
Production and Exploration Association Journal, 39: 40-61.
Pendrel, J., Debeye, H., Pedersen‐Tatalovic, R., Goodway, B., Dufour, J., Bogaards, M. and Stewart,
R.R., Estimation and interpretation of P and S impedance volumes from simultaneous
inversion of P‐wave offset seismic data, SEG Technical Program Expanded Abstracts 2000, pp.
146-149.
Pittman, E.D., 1992. Relationship of Porosity and Permeability to Various Parameters Derived from
Mercury Injection-Capillary Pressure Curves for Sandstone. AAPG Bulletin, 76(2): 191-198.
Revets, S.A., Keep, M. and Kennett, B.L.N., 2009. NW Australian intraplate seismicity and stress
regime. Journal of Geophysical Research: Solid Earth, 114(B10): B10305.
Reynolds, S.D., Paraschivoiu, E., Hillis, R.R. and Obrien, G.W., 2005. A Regional Analysis of Fault
Reactivation and Seal Integrity Based on Geomechanical Modeling: An Example from the
Bight Basin, Australia. In: P. Boult and J. Kaldi (Editors), AAPG Hedberg Series. Evaluating fault
and cap rock seals, pp. 57-71.
Rowbotham, P.S., Marion, D., Lamy, P., Insalaco, E., Swaby, P.A. and Boisseau, Y., 2003.
Multidisciplinary stochastic impedance inversion: integrating geological understanding and
capturing reservoir uncertainty. Petroleum Geoscience, 9(4): 287-294.
Saussus, D. and Jarvis, K., Extracting detailed lithology from seismic data, ASEG Extended Abstracts
2009: 20th Geophysical Conference, pp. 1-5.
Sayers, C.M., 2006. An introduction to velocity-based pore-pressure estimation. The Leading Edge,
25(12): 1496-1500.
Schowalter, T.T., 1979. Mechanisms of secondary hydrocarbon migration and entrapment. AAPG
Bulletin, 63: 723-760.
Shepherd, M., 2009. Structural Geology: Faults. In: M. Shepherd (Editor), Oil field production geology.
AAPG Memoir 91, pp. 107-122.
Shipton, Z.K., Evans, J.P. and Thompson, L.B., 2005. The Geometry and Thickness of Deformation-
band Fault Core and its Influence on Sealing Characteristics of Deformation Fault Zones. In: R.
Sorhabi and Y. Tsuji (Editors), Faults, fluid flow and petroleum traps. AAPG Memoir 85, pp.
181-195.
Smith, D.A., 1966. Theoretical considerations of sealing and non-sealing faults. AAPG Bulletin, 50:
363-374.
Soleymani, H. and Riahi, M.-A., 2012. Velocity based pore pressure prediction—A case study at one
of the Iranian southwest oil fields. Journal of Petroleum Science and Engineering, 94–95(0):
40-46.
Spikes, K.T. and Dvorkin, J.P., 2003. Model-based prediction of porosity and reservoir quality from P-
and S-wave data. Geophysical Research Letters, 30(20): 2029.
Swarbrick, R.E. and Osborne, M.J., 1998. Mechanisms that Generate Abnormal Pressures: An
Overview. In: B.E. Law, G.F. Ulmishek and V.I. Slavin (Editors), AAPG Memoir, pp. 13-34.
Terzaghi, K., 1923. Die Berechnung der Durchlassigkeitsziffer des Tones aws dem Verlanf der
Hydrodynamischen Spannungserscheinungen. Sb. Akad. Wiss, Wien: 132-135.
Tingay, M.R.P., Hillis, R.R., Morley, C.K., Swarbrick, R.E. and Okpere, E.C., 2003. Variation in vertical
stress in the Baram Basin, Brunei: tectonic and geomechanical implications. Marine and
Petroleum Geology, 20(10): 1201-1212.
Torabi, A., Fossen, H. and Braathen, A., 2013. Insight into petrophysical properties of deformed
sandstone reservoirs. AAPG Bulletin, 97(4): 619-637.
Tueckmantel, C., Fisher, Q.J., Knipe, R.J., Lickorish, H. and Khalil, S.M., 2010. Fault seal prediction of
seismic-scale normal faults in porous sandstone: A case study from the eastern Gulf of Suez
rift, Egypt. Marine and Petroleum Geology, 27(2): 334-350.
Turner, J.P. and Williams, G.A., 2003. Sedimentary Basin Inversion and Intra-Plate Shortening. Earth
Science Reviews, 65: 277-304.
Vernik, L., Fisher, D. and Bahret, S., Estimation of net‐to‐gross from P and S impedance: Part II — 3D
seismic inversion, SEG Technical Program Expanded Abstracts 2001, pp. 211-214.

96
Watts, N.L., 1987. Theoretical aspects of cap-rock and fault seals for single- and two-phase
hydrocarbon columns. Marine and Petroleum Geology, 4(4): 274-307.
Weaver, C.E., 1989. Clays, Muds and Shales. Elsevier, Amsterdam.
Withjack, M.O. and Jamison, W.R., 1985. Deformation produced by oblique rifting. Tectonophysics,
126: 99-124.
Wood, L.J., 2010. Shale Tectonics. AAPG Memoir, 93. AAPG, Tulsa, OK.
Wyllie, M., Gregory, A. and Gardner, L., 1956. ELASTIC WAVE VELOCITIES IN HETEROGENEOUS AND
POROUS MEDIA. GEOPHYSICS, 21(1): 41-70.
Yielding, G., Freeman, B. and Needham, D.T., 1997. Quantitative Fault Seal Prediction. AAPG Bulletin,
81: 897-917.
Yilmaz, O., 2008. Seismic Data Analysis - Processing, Inversion and Interpretation of Seismic Data, 1.
Society of Exploration Geophysicists, USA.
Zhang, J., 2011. Pore pressure prediction from well logs: Methods, modifications, and new
approaches. Earth-Science Reviews, 108(1–2): 50-63.
Zoback, M.D., 2010. Reservoir Geomechanics. Cambridge University Press, UK.
Zoback, M.D. and Healy, J.H., 1984. Friction, Faulting and in-situ stress. Annales Geophysicae, 2: 689-
698.

97