Downloaded from http://sp.lyellcollection.

org/ at Schlumberger on November 29, 2013

Geological Society, London, Special Publications

Fault seal prediction: the gouge ratio method

B. Freeman, G. Yielding, D. T. Needham and M. E. Badley

Geological Society, London, Special Publications 1998, v.127;

p19-25.
doi: 10.1144/GSL.SP.1998.127.01.03

Email alerting click here to receive free e-mail alerts when

service new articles cite this article
Permission click here to seek permission to re-use all or
request part of this article

Subscribe click here to subscribe to Geological Society,

London, Special Publications or the Lyell
Collection

Notes

© The Geological Society of London 2013

 Downloaded from http://sp.lyellcollection.org/ at Schlumberger on November 29, 2013

Fault seal prediction: the gouge ratio method

B. F R E E M A N , G. Y I E L D I N G , D. T. N E E D H A M & M. E. B A D L E Y

Badley Earth Sciences Ltd, North Beck House, North Beck Lane, Hundleby, Spilsby,
Lincolnshire PE23 5NB, U K

Abstract: The reduction in pore throat size attributable to the enhanced clay content in the
fault rock between two juxtaposed reservoir bodies is able to produce an effective seal to
hydrocarbon migration. The gouge ratio, when applied to sand-shale sequences, indicates
the proportion of phyllosilicate material that is expected to be incorporated into the fault
rock, hence it provides a measure of the seal capacity. Gouge ratios for faulted reservoir
sequences are computed from data that are routinely available from fields at the appraisal
or production stage. We present a repeatable methodology that maximizes the use of
seismically based geometric data and well-based stratigraphic and compositional data. The
resulting gouge ratio calculations can be used comparatively at the scale of individual
faults, fields or basins. We examine an example fault from the Oseberg Syd Field that is
known to seal and we show that a threshold shale gouge ratio of 18% will support a
cross-fault pressure difference of c. 8 bar.

Structural geology has an important contribu-
tion to make to reservoir characterization, parti-
cularly the interpretation and analysis of faults.
Traditionally, oil industry interpretation has
focused on the delineation of horizon surfaces.
Faults are often either left uninterpreted or are
not interpreted in a manner which produces a
consistent three-dimensional model when com-
bined with horizon interpretation. The result of
such neglect is that fault planes are frequently
regarded as vertical for the purposes of reservoir
volume estimation and for hydrocarbon flow
simulation. Although faults have always been
recognized as key structures in defining the archi-
tecture of reservoirs, it is only recently that there
has been renewed interest in their behaviour as
lateral seals.
Fault seals fall into two broad categories: (1)
reservoir against non-reservoir in which case
the juxtaposition of permeable rocks against
non-permeable rocks provides a sealing mechan-
ism; and (2) reservoir against reservoir in which
case the fault itself must provide a barrier to
hydrocarbon migration. Knipe (1992) provides
a review of the mechanisms leading to fault
seal. This paper is concerned only with the
second type of seal - reservoir juxtaposed to
reservoir. In this case the estimation of the seal
potential of a fault has been approached from
at least three different angles. Work by Weber
et al. (1978) indicated that in some faulted
sequences, where reservoir rocks are interbedded
with shales, the shale units are drawn into the
fault zone. The shale layers showed normal-
drag geometries and it is implied that there
should be a barrier to hydrocarbon migration
where a reservoir unit is in contact with shale
that has been sheared into the fault plane.
Within the fault zone the shape of the shale
units tapers from the shale source bed towards
the centre of offset. The shape, hence thickness,
of the shale wedge can be computed as a function
of fault throw and shale bed thickness. In turn,
the fault zone shale thicknesses are used to
estimate seal potential. This type of method has
been used to explain the distribution of hydro-
carbon reservoirs in sand-shale sequences of
the Niger Delta (Bouvier et al. 1989; Jev et al.
1993).
A related method (Lindsay et al. 1993) predicts
the likelihood of shale source beds producing a
continuous smear of cIay material across a
reservoir-reservoir contact. This method too
has been applied successfully to stacked hydro-
carbon accumulations in the Columbus Basin
(Gibson 1994).
A second approach is to quantify selected
attributes on sets of faults which are known
either to seal or to leak and then to make predic-
tions about u n k n o w n behaviour by comparison.
This method has been outlined by Knott (1994).
Possible drawbacks to such a method are that an
extensive regional database is required and there
is some reliance on there being a relatively simple
and continuous stratigraphy. It also embodies a
notion that faults either seal or leak, whereas it
is more realistic to think of faults as having
spatially heterogeneous sealing characteristics.
Finally, a third approach is based on consid-
eration of the capillarity of the rocks which are
deformed in the fault zone (Knipe 1992). Watts
(1987) and Vavra et al. (1992) review the mechan-
isms by which capillarity is able to trap substan-
tial hydrocarbon accumulations. Watts (1987)
and Knipe (1992) argue that the equations of
capillarity used previously for vertical traps

FREEMAN,B., YIELDING,G. NEEDHAM,D. T. & BADLEY,M. E. 1998. Fault seal prediction: the gouge ratio method.
In: COWARD,M. P., DALTABAN,T. S. • JOHNSON, H. (eds) Structural Geology in Reservoir Characterization.
Geological Society, London, Special Publications, 127, 19 25.
 Downloaded from http://sp.lyellcollection.org/ at Schlumberger on November 29, 2013
20 B. FREEMAN E T AL.
may also be applied to faults provided that
representative measurements can be obtained
from analysis of core. Ultimately this type of
analysis should be able to make 'standalone'
predictions of accumulation without recourse to
the 'calibration' required for the first two
methods. However, it must encounter scaling
problems. Since samples are analysed from
core, and reservoir-reservoir juxtaposition
areas can be several hundred thousand square
metres, there needs to be some means of up-
scaling, or averaging, of the capillarity informa-
tion. There is also a difficulty in obtaining
representative core. For a number of reasons,
oil companies avoid drilling through known
fault zones, for example to avoid early water
breakthrough. Therefore faults that are sampled
in core are at a scale that may be important in
controlling reservoir performance but are
unlikely to be responsible for the gross sealing
characteristics of the larger faults that trap
economic reserves.
Our approach is similar to that of Bouvier et al.
(1989) in that it considers the entire fault plane. It
also attempts to address some of the scaling
difficulties associated with the capillarity work.
We take the simple view that reservoir character-
ization is most useful if one is able to apply a
repeatable methodology to all faults in a pros-
pect, field, basin and between basins. This
means capitalizing on the primary data that are
routinely available, i.e. horizon and fault
interpretation for geometry, wireline logs for
reservoir zone thickness, content of clay or
other impermeable material, and, if available,
drilling and production information on reservoir
zone pressures. From the geometric data one can
derive a precise model of the fault's shape and
displacement distribution. Combining this with
stratigraphy and detailed lithology it is possible
to produce fault-wide maps predicting the distri-
bution of clay and other impermeable lithologies
(in the fault rock) as a proportion of the total
fault-rock volume (gouge ratio maps).
Unlike the work of Weber et al. (1978) and
Bouvier et al. (1989) that implies a single
mechanism for incorporating shale into a fault
zone, we simply infer that the composition of
the fault rock closely reflects that of the unde-
formed walls. Hence as the overall composition
of clay and other impermeable material increases
then we would expect the seal potential to
increase. For example, this means that faulted
sequences from dirty sandstones would have
similar seal potentials to sequences of inter-
bedded clean sands and shales, provided the
overall clay volume fractions are the same. In
the case where the impermeable material is
dominantly shale we refer to the calculation as
the shale gouge ratio. The rationale for this
stems from the observation that phyllosilicates
become concentrated along shear planes in the
fault zone and clog the larger pores. They are
then effective in reducing the maximum radius
of interconnected pores (e.g. Knipe 1992).
Gibson (1998) shows from a study of small-
displacement faults of varied mineralogy that
there is a clear inverse trend between clay content
and porosity. He also shows that of the struc-
tures observed in faults, clay gouge zones are
the most efficient at reducing pore throat size
and permeability.
The purpose of this note is to describe a
methodology that we have applied to the study
of lateral seals in many producing oil fields. We
do this with reference to a specific example
from the Oseberg Syd Field (Fig. 1). A full
description of the sealing characteristics of
faults in the Oseberg Syd Field is given in Fristad
et al. (1996).
Geological setting
The Oseberg Syd Field is in an area of Mesozoic
extension located within Block 30/9 on the
Norwegian Continental Shelf between the
Horda Platform and the Viking Graben. Most
faults within the Oseberg/Oseberg Syd region
strike N - S to N N W - S S E in an anastomosing
pattern sub-parallel with the Viking Graben.
The fault blocks range in size from less than
10 km 2 to 250 km 2. Almost all of the individual
fault blocks that have been drilled contain oil
and gas. Furthermore, depths to the fluid con-
tacts differ between blocks, indicating that the
bounding faults are also sealing faults.
The fault discussed here separates two
hydrocarbon columns in Middle Jurassic Brent
reservoirs encountered in Wells 30/9-13s and
30/9-14 with a maximum cross-fault pressure
difference of 9.5bar (fault A in Fig. 1). The
throw maximum is close to the centre of the
fault, diminishing to zero at its southern end.
On the same fault there is considerable overlap
of Brent reservoirs, the maximum offset being
about half of the local Brent Group thickness.
The burial depth of the Brent Group at the
time of faulting is estimated to have been less
than 500 m.
Geometric information
It is important, before performing fault seal
analysis, to maximize the use of the basic
 Downloaded from http://sp.lyellcollection.org/ at Schlumberger on November 29, 2013
FAULT SEAL PREDICTION
/?
v '1 31/7
I
2kin41
Fig. 1. Location map of the Oseberg Syd Field off the western coast of Norway. Fault A, shown between Wells
30/9-13s and 30/9-14, is the fault referred to in this paper.
geological interpretation. This is usually in the
form of horizon surfaces from maps or cross-sec-
tions and from fault traces interpreted in cross-
section.
Bringing together these two sets of informa-
tion in a single, topologically consistent model
is non-trivial. First the fault surface has to be
modelled, in the algorithmic sense, from the
raw data, in this case the (x, y, z) values of the
traces picked on vertical sections. Ideally, the z-
values should be in depth (as in the example dis-
cussed here). We use a proprietary grid-based
method, referenced to a base plane, that esti-
mates the elevations at grid nodes from the pro-
jection of least-squares, best-fit planes through
the control points. Data points to be used in
both the plane fitting and in the final, weighted
estimation of values at the nodes are selected
by an octant search strategy. By modifying the
numbers of points required in each octant and
by changing the weights, we are able to exercise
a great deal of control over the resulting surface
model. This is important because it means that if
the starting data are good our fault model can be
forced to honour the control points very closely.
On the other hand, if data quality is poor, for
example when fault traces are mis-tied, then we
need to be able to smooth through the ambigu-
ities and picking errors. Typically, for good
quality data, we expect to see root-mean-square
(RMS) residuals between the modelled surface
and constitutive control points to be of the
order of a few metres or less. These errors are
smaller than the horizontal resolution of the
seismic data.
21
Despite the maturity of methods employed for
interpretation of horizon surfaces, these surfaces
are rarely in a form amenable to 'instant' analy-
sis. The main problem is usually an overestima-
tion of the heave which can lead to incorrect
throw computations and erroneous juxtaposi-
tions (see Needham et al. (1996) for a discussion).
During our modelling procedure we apply cor-
rections to horizon interpretation in the vicinity
of a fault which ensure a single, consistent
fault/horizon topology with no gaps or overlaps.
We refer to this model as the framework top-
ology (Fig. 2).
This framework topology forms the basis of an
analysis. From it we are able to derive quantities
such as throw, gouge ratio, thickness variations,
reservoir overlap areas, etc. It is used as a tem-
plate for refining stratigraphy from seismically
mappable scale to reservoir zone scale and for
computing pressure variations as projected
from well information.
It is possible to short-cut some of the steps in
the geometric analysis by producing strike pro-
jections of horizon terminations directly from
horizon maps (e.g. Allan 1989). However, there
are always vertical inaccuracies introduced by
using hand-drawn fault polygons, but more
importantly strike projections do not conserve
area. Hence reservoir-reservoir juxtapositions
that may be significant in a reservoir engineering
sense may be lost on a strike projection. In
our view it is better to maximize the geometric
data by producing a fully three-dimensional
topology rather than degrading it by using a
two-dimensional approach.
 Downloaded from http://sp.lyellcollection.org/ at Schlumberger on November 29, 2013

22 B. FREEMAN E T AL.

J a C,e*ooeous,]
Top LowerTarbertI

Bas___eBrent ]

Fig. 2. Three-dimensional model of fault A shaded in throw with illumination perpendicular to the page; vertical
exaggeration is a factor of three. Maximum throws (red) are approximately 250 in diminishing to zero (blue) where
the fault is mapped above the Base Cretaceous. Marked on the surface are hangingwall and footwall terminations
of the seismically mapped horizons: Base Brent, Top Lower Tarbert and Base Cretaceous.

Stratigraphic and compositional information m u c h degraded, e.g. in sub-salt plays, it is

In most cases it is not feasible to use the large-
scale stratigraphic model, defining the frame-
w o r k topology, directly to predict across-fault
reservoir characteristics. This is because a
m a p p a b l e seismic interval m a y contain m a n y
individual reservoir zones. M o r e o v e r , in geo-
logical settings where the seismic image is very
c o m m o n for there to be only one picked horizon,
say top reservoir. Hence, the f r a m e w o r k topol-
ogy needs to be a u g m e n t e d with stratigraphic
detail derived from alternative sources, namely
wireline logs. A l t h o u g h it is possible to divide
the log curves explicitly into reservoir and non-
reservoir units, this in itself can present a m a j o r
p r o b l e m of lateral correlation. Usually it is

Fig. 3. View, scale and illumination as for Fig. 2. (A) Modelled fault as a grid with the intersection of the
footwall reservoir zones (two shades of yellow). (B) Brent-Brent juxtaposition (red, blue and green). All the zones
within the Brent Group are potential reservoirs hence all the shaded region represents an area for potential leakage.
The overlap is shaded on a zone-by-zone basis to emphasize the complexity of juxtapositions that can arise from a
relatively simple, layer-cake, stratigraphic model.
 Downloaded from http://sp.lyellcollection.org/ at Schlumberger on November 29, 2013

FAULT SEAL PREDICTION 23

more practical to interpret the curves in terms of fault zone at that node. Practically, this means
reservoir zones, for which one can be confident quantifying the volume fraction of shale in the
that their lateral continuity exceeds the length window of rock that has slipped past the node,
of the fault in question. where the height of the window is equal to the
Having sub-divided seismic intervals into throw. In order to be able to deal with geological
zones and measured their relative (or absolute) situations that involve post-faulting erosion we
vertical thicknesses, we compute the locations are able to define the window to lie either entirely
of the zone intervals within the fi'amework topol- in one wall or to straddle the node with half in the
ogy. In the Oseberg Syd example the reservoirs of footwall and half in the hangingwall. Figure 4 is a
interest occupy the Brent Group, for which we schematic diagram describing the calculation at a
recognize eight zones of alternating good sands single node using a window of throw in the hang-
and poor sands. Zone compositions were derived ingwall. The gouge ratio method may be applied
from computer-processed petrophysical logs, to stratigraphies that isolate units explicitly as
specifically the shale volume fraction curve. All shale and reservoir units (Fig. 4a) or to strati-
zones contain some clay. The goods sands have
a clay content of c. 15% and the most shaly
sands have a clay content of c. 30%. Above the
a
Brent Group the average clay content is c. 50%,
beneath it, the clay content is c. 62%. Figure
3A shows the fault surface with intersecting
reservoir zones of the footwall. Note that the
upper footwall zones are eroded beneath the
Base Cretaceous unconformity. Combining the
footwall and hangingwall stratigraphies reveals
the extent of the B r e n t ' B r e n t juxtaposition
(Fig. 3B). Figure 3B also shows the complexity
of juxtaposition geometries that arise from rela-
thA~., r°w~[
tively simple stratigraphic relationships. There
is potential for cross-leak at all Brent-Brent ShaleGougeRatk
overlap, although it is the good sand-good
sand juxtapositions that represent the locations ~Nckness
throw xt00% ..
of greatest risk. The next stage in the process is
to quantify this risk by computing the shale
gouge ratio.

Gouge ratio computations

Gouge ratios can be computed for any imperme-
able material that is likely to be incorporated
into the fault zone. Most commonly the
impermeable material is clay, when we refer to
the parameter as the shale gouge ratio. The
gouge ratio calculation is performed, once
again, as a grid-based operation. The spacing
of the grid nodes is chosen such that the thinnest
stratigraphic units are adequately represented in
raster form. This grid is usually very much finer
Sh~
than that for the surface model. For conveni-
ence, the parameters of the gouge ratio grid
are always integer multiples of those from the
suface models. Hence it is straightforward to
interpolate structural information, i.e. throw
Fig. 4. (a) Computation of shale gouge ratio at a point
from the surface model onto the gouge ratio from a discrete sequence of sands (unlabelled) and
grid. shales (Z1 and Z2). (b) Computation of shale gouge
At every node on the gouge ratio grid we ratio at a point when the clay content is considered as
perform the same operation. We estimate the an average value over defined zones. Vcl indicates the
proportion of shale that is present within the volume fraction of clay in an interval of thickness Z.
 Downloaded from http://sp.lyellcollection.org/ at Schlumberger on November 29, 2013

24 B. FREEMAN E T AL.

GOC .........
. . . °' ,. ,.-~ ~ ~"~v"'":~"~"
U p t h r o w n reservoir z o n e s
OWC;~ ~,,~-,~,~........... ~-~- ~m~.-....................-, - "...... . outlined in blue

GOC ~ " ' "'~ . . . . . ~"'~ .......... '

". . . . . . . . . . . . . . . ~ Heather
Sst
Z

Lwr

/ , ".~----%~%,j.

~' / [" "~ '.., ,.-~ Shale Gouge Ratio (%)

100m .... '.."'a ! °
I m
500m
D o w n t h r o w n reservoir
z o n e s outlined in b l a c k "~---, 20[i nC£s2~c~i~g
~jj Base
Brent

Jo
IBNIII I I I

Fig. 5. Isometric view of fault A seen from the west. Vertical exaggeration is a factor of five. Footwall and
hangingwall interval boundaries are shown as solid lines and the Brent-Brent overlap is shaded in shale gouge
ratio. Fluid contacts (GOC, gas-oil contact; OWC, oil-water contact) are projected from Well 30/9-13s in the
footwall and from Well 30/9-14 in the hangingwall.

graphies that are defined more generally as value, 15%, lies in the saddle towards the south-
shaly members and reservoir members (Fig. ern end of the structure. This point is deeper than
4b). In the latter case the shale gouge ratio is the structural spill point, hence it is of little inter-
computed from the sum of the volume fractions est. The critical region, with a shale gouge ratio
of clay in both reservoir and non-reservoir of 18%, lies towards the crest of the structure
units. in the oil leg of the footwall and in the gas leg
This flexibility in the interpretation of the of the hangingwall. Repeat formation tester
gouge ratio model means that it is suitable to data from wells on either side of the fault are
apply to a broad range of scales. One may also summarized in Fig. 6. This clearly demonstrates
represent the probability of encountering shale that fault A seals and that the shale gouge ratio
beds of given thickness as a volume fraction of of 18% is a threshold value able to support a
clay in a broader zone. Although we do not minimum pressure difference of c. 8 bar.
discuss this aspect any further, it opens the In order to make maximum use of shale gouge
possibility of using gouge ratio calculations in ratio calculations it is important to review the
stochastic models. threshold values on all of the faults in the field
Figure 4 shows that thegouge ratio computa- that are known to seal. This provides a calibra-
tions involve throw and vertical thicknesses. tion. It is then possible to calculate the shale
Geometrically this is equivalent to dip separation gouge ratio for faults that have not been drilled
and bed thickness measured in the direction of in one or both walls and to predict whether or
fault dip. Furthermore, providing there are no not the fault is likely to be able to support a pres-
significant lateral thickness changes on the scale sure difference, hence trap hydrocarbons. Fristad
of the displacement, the throw and vertical thick- et al. (1996) provide a complete description of all
ness model can be applied to faults of dip-slip, the sealing faults in the Oseberg Syd Field. They
oblique-slip or strike-slip origin. find consistently that a shale gouge ratio less than
For our Oseberg Syd example, Fig. 5 shows 15% will not seal, between 15% and 18% will
shale gouge ratio computed over all of the support small pressure differences (c. 0.5bar)
Brent-Brent juxtaposition. Shale gouge ratio and that above 18% will support pressure differ-
varies between 15% and 30%. The lowest ences in excess of 8 bar.
 Downloaded from http://sp.lyellcollection.org/ at Schlumberger on November 29, 2013
FAULT SEAL PREDICTION
Pore pressure (bars)
305 310 315 320 325 330
FW 9.5 bars HW
. . . . . . . . N,~_ Ha_ngingwall_GOC
10!m ........... ~ i
l Petroleum Society (eds) Hydrocarbon Seals -
Fig. 6. Summary of the repeat formation tester data
showing the maximum pressure difference at the top of
the structure between the footwall and hangingwall gas
legs.
Remarks
(1) G o u g e ratio calculations show that the
spatial sealing characteristics of a single
fault are likely to be heterogeneous, i.e. it
may seal over some of the s a n d - s a n d con-
tacts but leak over others.
(2) In the example from the Oseberg Syd Field
we show that a shale gouge ratio in excess of
18% will support a pressure difference of
c. 8 bar. In our experience from the Oseberg
Syd Field and several other fields we find, as
a rule of thumb, that shale gouge ratios less
than 15% are likely to leak. As the shale
gouge ratio increases above 20% the likeli-
h o o d that the fault will seal also increases.
(3) A n important step in the context of apprai-
sal/production is to constrain the threshold
value as closely as possible by considering
as m a n y sealing faults as possible from the
same field. However, using the rule of
t h u m b (above), shale gouge ratio may also
be used in exploration provided rudimen-
tary information on zone thicknesses and
composition is known. Various possibili-
ties, using different well prognoses, could
be used in an exploration setting to assist
risk calculation for a prospect.
(4) Nearly all fields in appraisal or production
have, as standard, the relevant information
required to perform a shale gouge ratio
analysis, i.e. seismic interpretation (in time
or depth), wireline logs, drilling and pro-
duction pressure data.
25
(5) The gouge ratio m e t h o d embodies a notion
335 of upscaling, is straightforward to apply
and produces usable results very rapidly.
References
ALLAN, U. S. 1989. Model for hydrocarbon migration
and entrapment within faulted structures. AAPG
Bulletin, 73, 803-811.
BOUVIER,J. D., KAARS-SIJPESTEIJN,C. H., KLUESNER,
D. F., ONYEJEKWE,C. C. & VAN DER PAL.,R. C.
1989. Three-dimensional seismic interpretation
and fault sealing investigations, Nun River Field,
Nigeria. AAPG Bulletin, 73, 1397-1414.
FRISTAD, T., GROTH, A., YIELDING,G. & FREEMAN,B.
1996. Quantitative fault seal prediction a case
study from the Oseberg Syd area. In: Norwegian
Importance for Exploration and Production (Con-
ference Abstracts). Norwegian Petroleum Society,
Oslo.
G~BSON, R. G. 1994. Fault-zone seals in siliciclastic
strata of the Columbus Basin, offshore Trinidad.
AAPG Bulletin, 78, 1372-1385.
- - 1 9 9 8 . Physical character and fluid-flow properties
of sandstone-derived fault zones. This' volume.
JEV,B. I., KAARS-SIJPESTEIJN,C. H., PETERS,M. P. A. M.,
WATTS, N. L. & WILKIE, J. T. 1993. Akaso Field,
Nigeria: use of integrated 3D seismic, fault-slicing,
clay smearing and RFT pressure data on fault
trapping and dynamic leakage. AAPG Bulletin,
77, 1389-1404.
KNIPE, R. J. 1992. Faulting processes and fault seal.
117: LARSEN,R. M., BREKKE,H., LARSEN,B. T. &
TALLERAS,E. (eds) Structural and Tectonic Model-
ling and its Application to Petroleum Geology.
Elsevier, Amsterdam, 325-342.
KNOTT, S. D. 1994. Fault seal analysis in the North Sea.
AAPG Bulletin, 77, 778 792.
LINDSAY, N. G., MURPHY, F. C., WALSH, J. J. &
WATTERSO~, J. 1993. Outcrop Studies of Shale
Smear on Fault Surfaces. International Associa-
tion of Sedimentologists Special Publication, 15,
113-123.
NEEDHAM,D. T., YIELDING,G. & FREEMAN,B. 1996.
Analysis of fault geometry and displacement
patterns. In: BUCHANAN, P. G. & NIEUWLAND,
D. A. (eds) Modern Developments in Structural
Interpretation, Validation and Modelling. Geological
Society, London, Special Publication, 99, 189-199.
VAVRA, C. L., KALDI, G. G. & SNEIDEe,, R. M. 1992.
Geological applications of capillary pressure: a
review. AAPG Bulletin, 76, 840-850.
WATTS, N. 1987. Theoretical aspects of cap-rock and
fault seals for single- and two-phase hydrocarbon
columns. Marine and Petroleum Geology, 4, 274-
307.
WEBER, K. J., MANDL,G., PILAAR,W. F., LEHNER,F. &
PRECIOUS, R. G. 1978. The role of faults in hydro-
carbon migration and trapping in Nigerian growth
fault structures. Offshore Technology' Conference
No. 10. Paper OTC 3356, 2643 2653.