See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/265289346

THE KUJUNG FORMATION IN KURNIA-1 : A VIABLE FRACTURED RESERVOIR

PLAY IN THE SOUTH MADURA BLOCK

Article

CITATIONS READS

0 300

3 authors, including:

Craig Buchan
Task Fronterra Geoscience
21 PUBLICATIONS   1,013 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Craig Buchan on 20 July 2015.

The user has requested enhancement of the downloaded file.

 IPA10-G-005

PROCEEDINGS, INDONESIAN PETROLEUM ASSOCIATION

Thirty-Fourth Annual Convention & Exhibition, May 2010

THE KUJUNG FORMATION IN KURNIA-1 : A VIABLE FRACTURED RESERVOIR PLAY IN

THE SOUTH MADURA BLOCK

Trevor Magee*
Craig Buchan**
Jeremy Prosser**

ABSTRACT Indonesia. Many of these formations are known to

host hydrocarbons, but the fine grained nature of
The Kurnia-1 well was drilled as the first test of the the carbonates results in generally poor reservoir
Kujung Formation in the South Madura exploration matrix properties (Wilson 2002; Park et al. 1995).
block on Madura Island in the East Java Basin. The However, as a result of the active tectonic history of
well was drilled to test a large anticlinal structure the region, many of these carbonates have
mapped on 2D seismic data. The well encountered experienced periods of faulting and folding,
very strong mudlog gas shows that persisted resulting in well developed fracture networks that
throughout the Kujung section. An open-hole enhance their permeability and increase their
production test flowed hydrocarbons to surface, but economic potential. Similar fractured carbonate
mechanical problems caused the test to be reservoirs are present in many large producing
abandoned without determining a flow rate. fields world-wide such as in Iran, Oman and
Evaluation of wireline and Logging While Drilling Canada (e.g. Stephenson et al. 2007; Wennberg et
(LWD) logs and sidewall cores from Kurnia-1 al. 2007; De Keijzer et al. 2007; Rawnsley et al.
indicated that the Kujung Formation carbonates are 2007). The Oligo-Miocene Kujung Formation of
characterised by low matrix porosity; however, the East Java Basin (Figure 1) is an Indonesian
borehole resistivity and azimuthal density images carbonate in which reservoir characteristics may be
reveal a high degree of fracturing, raising the enhanced by fracturing. This study focuses on data
possibility of a fractured reservoir play. The from a single well, Kurnia-1, located near the south
objective of this study was to evaluate the viability coast of Madura Island (Figure 1). The well was
of a fractured reservoir. Mudlog, LWD and wireline drilled through a Miocene clastic section to the
logging data and borehole image log analysis were primary objective Kujung Formation (Figure 1).
integrated in order to indentify and characterise Drilling proved difficult through the mid section of
fractures in the Kujung carbonates based on their the well and into the lower carbonate dominated
orientation and image characteristics. The study has section and associated testing difficulties precluded
confirmed strong fracture density in the Kujung a definitive production test, although hydrocarbons
Formation and that zones of greatest fracture were recovered to surface and several strong gas
intensity also had the strongest gas shows. Fractures shows were recorded while drilling.
have been assigned to three orientations and these
fracture sets have been integrated into a coherent In the absence of definitive production tests, a
structural framework. In the absence of production detailed analysis of open hole logs and geological
data, a qualitative assessment of permeability interpretation of Extended Range Micro Imager
potential has been made based on fracture (XRMI™ of Halliburton) and Azimuthal Density
characteristics and orientation relative to local and (AZD™ of Weatherford) borehole image data was
regional structures. The results of the study indicate carried out to assess the reservoir potential of the
the Kujung Formation in the South Madura block is Kujung Formation. The results reveal significant
a potentially viable fractured reservoir. fracture populations in the Kujung Formation that
coincide with the strongest gas shows. In addition,
INTRODUCTION borehole image derived dip data suggests that the
Deep water pelagic carbonates are common large scale structure is consistent with the current
throughout South East Asia and particularly in stress field and that fractures have orientations
conducive to flow within this model. Based on this
* Cooper Energy
** Task Geoscience evidence, it will be demonstrated that, despite the
apparent limited lithological potential of the Kujung through the Late Eocene to Early Oligocene, with
Formation, secondary porosity and permeability localised carbonate build up, referred to as the CD
provided by the fracture networks mean that the Carbonate. Renewed uplift ended this depositional
Kujung Formation may be a viable hydrocarbon phase and resulted in widespread erosion.
reservoir. Moreover, the study will show that
certain fracture orientations are more conducive to Basin sag, traditionally associated with a slowing of
flow, demonstrating the need to properly the northward migration of the Australian plate in
characterise the orientation and nature of fractures the early Oligocene (Hall, 2002) or due to newly
in order to fully exploit their potential. invoked northward subduction (Seubert and
Sulistianingsih, 2008), resulted in deposition of a
GEOLOGICAL SETTING thick Oligocene section. This sag phase was
followed by a tectonically quiescent period from
Tectonic history and stratigraphy Late Oligocene to Early Miocene times, during
which Kujung carbonate deposition was prevalent
The tectonic evolution of the East Java Basin has on stable land-attached platform areas and on fault-
been primarily controlled by the convergence of the controlled seafloor highs, while shales and marls
Indo-Australian and Eurasian plates. The basin is accumulated in basinal areas. In the Early Miocene
situated on the southern margin of the stable Sunda rapid carbonate deposition was restricted to seafloor
Craton to the north of a volcanic arc, running highs where reef growth could keep pace with the
through the centre of the island of Java. It rapidly rising sea level. In the South Madura area
experienced a complex tectonic history with initial seafloor highs, such as the CD Ridge, are orientated
extension followed by periods of differential parallel with the E-W structural grain. By mid
subsidence and later inversion (Hamilton, 1979). Miocene basinal shales and occasional thin
The structural grain of the basin was controlled by turbidites of the Tuban (or Lower OK) Formation
the fabric of the underlying basement. To the had replaced carbonate deposition.
northwest of Madura Island in the Muriah Trough,
Bawean Arch and Central Deep, this is A southward shift of the Australian plate boundary
predominantly NE-SW, whereas in the Kendeng in the Middle Miocene was followed by a phase of
Trough (onshore East Java) and the Madura Trough widespread magmatic activity across Java. The
the orientation is E-W (Figure 1; Sribudiyani et al., latest and most pronounced phase of compression
2005). began in the Mid Miocene: the main compressive
phase commenced in the Late Miocene and
The stratigraphy of the East Java Basin reflects a continued episodically through to the present day.
balance between carbonate and clastic deposition, Pre-existing graben faults were reactivated resulting
governed by the relative influences of tectonics, sea in large inversion structures. The North Madura
level and land-derived clastic input (Sharaf et al., Platform remained a relatively stable high, and was
2005). Many different stratigraphic schemes have bounded to the south by the inverted former basinal
been employed by different companies in the East region, now Madura Island. The main structural
Java basin. This paper has adopted the scheme grain over Madura Island at this time was East-
employed by Shell during their exploration of West as reflected by the Madura Uplift Zone. The
Madura Island between 1988 and 9992. This southern part of Madura Island to the south of the
stratigraphic scheme is shown in Figure 1; other Madura Uplift Zone remained basinal with
commonly used nomenclature may also be quoted deposition of the marine siltstones and claystones
in brackets. of the Tawun Formation intercalated with
occasional mass flow sandstones of the Ngrayong
From the Early - Middle Eocene to Early Oligocene Member. The Tawun Formation is overlain by
a tensional regime, traditionally associated with sandstones and carbonates of the Bulu Formation of
back-arc spreading, resulted in development of a latest Middle Miocene age. The Bulu Formation is
series of generally NE-SW ridges and grabens overlain by the Late Miocene Pasean (or
along the southeastern margin of the Sunda shield. Wonocolo) Formation, which consists of a basal
In the eastern part of the basin the trend is E-W. shale unit overlain by sandstones and limestones.
These grabens filled with a thick sequence of pre- This unit outcrops or has been eroded over much of
Ngimbang and Ngimbang Formation clastics, the South Madura Block. The Pliocene Mundu and
including lacustrine shales that constitute the Paciran (or Kawengan) Formations are largely
principal source rock in the region. In the ensuing eroded over the South Madura Block as is the
sag phase, a fully marine shale sequence developed Pleistocene Lidah (or Pamekasan) Formation.
STUDY DATA were merged with the available wireline logs run to
produce a final suite of logs for interpretation of the
The data integrated during this study was wide entire Kujung section. A massive 250 metre thick
ranging and included a regional seismic grid and a carbonate, becoming shalier towards the base, was
broad suite of downhole reservoir evaluation tools encountered between 3064-3313m. Matrix porosity
run across the Kujung Formation in Kurnia-1. in this thick carbonate section is low, averaging 4-
These reservoir evaluation tools comprised: mudlog 6%.
data (lithology, gas, drilling parameters and
formation pressure); LWD (gamma ray, resistivity, Drill Stem Test of Fractured Intervals
neutron and density logs) and wireline logs (dual
laterolog, density, neutron, spectral gamma ray, full Despite the overall tight matrix conditions, strong
waveform sonic, formation pressure and dip image mudlog gas was encountered while drilling through
logs). Micropalaeontological analyses of cuttings the Kujung Formation and fluorescent shows were
and sidewall core samples provided age dating and also observed in these limestones in both cuttings
depositional setting, while thin sections from and sidewall cores. The strongest gas shows
sidewall cores provided information on rock fabric appeared to correspond to intervals of high fracture
and texture and qualitative assessment of porosity density as observed in the borehole image logs
and permeability. (XRMI) leading to a hypothesis that fractures were
the primary source of reservoir porosity in the tight
In addition to the more standard suite of wireline carbonates. A production test was run across the
tools, borehole image logs were collected, entire Kujung open-hole section to test hydrocarbon
providing a unique perspective on the geological deliverability. The test was conducted as an open-
evaluation of the Kujung Formation, because hole test to avoid potential damage to the fracture
structural, lithological and petrophysical network that may have resulted from cementing
information can be acquired at a much higher casing. Combustible hydrocarbons were detected at
spatial resolution than standard open hole logging surface, but mechanical problems caused the test to
tools. Adverse borehole conditions in some sections be abandoned without determining a flow rate. The
of Kurnia-1 meant that image logs were restricted test was aborted due to plugging of the test string
to two sections with a large (919 m) separation. The by intraformational shales.
upper interval was logged over 1237.5 to 2086 m
through the largely clastic Tawun and Tuban STRUCTURAL CHARACTER OF THE
formations. The lower image interval was logged KURNIA PROSPECT
over 3005 to 3350 m in the Kujung carbonates with
both AZD and XRMI logs available. Image logs in Seismic interpretation
both intervals were found to be of generally very
good quality with only minor stick and pull The Kurnia prospect was interpreted from seismic
artefacts and some reduced image quality in regions data and regional structural information as a large
where the borehole is enlarged. wrench-related anticline with two structural
culminations (Figure 3). The structure was
KUJUNG FORMATION MATRIX delineated on a 1.2 km x 1.8 km seismic grid
CHARACTERISATION comprising data acquired in 1983, 1984, 1986 and
1989. Although reprocessed using modern
The Kujung Formation comprises micro-crystalline technology and workflows, data quality remains
packstone to wackestone with very poor visible poor to fair. Figure 4 is a seismic line over the
porosity (Figure 2). In the upper part of the Kurnia structure.
formation (3004-3052 m) the carbonates are
intercalated with soft, silty claystones, with a highly The location of the structure close to the southern
interbedded nature as revealed by wireline logs. coast of Madura Island, restricted seismic coverage
Log evaluation shows the limestones to be over the crest of the Kurnia structure to a series of
generally tight with matrix porosities typically 3- line ends (Figure 3). Migration artefacts associated
6% with occasional beds with porosities of 8-10%. with these line-ends, have introduced some
uncertainty into the interpretation of the top Kujung
The lower interval from 3050 to 3350m was logged horizon. However, despite these issues, the Kurnia
using LWD gamma ray, resistivity and density- structure was mapped as a robust four-way dip
neutron. This section was also logged with cross closure, with several faults interpreted within the
dipole sonic and XRMI wireline logs. These logs structure, including a significant down-to-the-south
reverse fault immediately south of the Kurnia-1 direct detection in the XRMI image, or from
location (Figure 4). Interpretation of offshore lines interpretation of bedding trends. The majority of
confirmed closure to the south, resulting in an areal faults appear to have a WNW-ESE strike, parallel
closure of approximately 130 km2 with vertical to the regional fold axis, but a few were also
relief in the order of 400 ms TWT or 600 metres. interpreted to strike in a more E-W direction and
Kurnia-1 was located to test the higher and better may reflect basement trends (Sribudiyani et al.,
defined western culmination of the large-scale fold 2005).
(Figure 3).
The overall fold geometry described by bedding is
Kurnia-1 borehole image analysis of large-scale similar to that interpreted from seismic of an
structure antiform with a WNW-ESE axis, although the NNE
vergence of the fold based on well data is not
Although the main focus of this study is the Kujung apparent from seismic interpretation. A number of
Formation carbonates, bedding dip data was factors may contribute to this apparent discrepancy:
derived by manual picking of features in the XRMI the seismic data over the Kurnia structure consist of
images from both the Tawun/Tuban and Kujung a series of line ends, resulting in poor imaging and
image intervals in order to derive a large-scale migration; the structural dips measured in the well
structural model. The observed dip of bedding cannot be accurately imaged on seismic; and the
features in the upper and lower image intervals is steep NNE dips may be associated with drag
strikingly different. The clastics of the upper rotation close to the down-to-the-NNE reverse fault
interval have a variable dip orientation but are some 300 m to the NNE of Kurnia-1. It should also
generally shallow (<15°), dipping to the NW or be noted that there is about 100 metres uncertainty
WSW (Figure 5), whereas the limestone beds of the in the seismic line location. In addition the NNE
lower image interval dip steeply (>40°) to the NNE vergent closure evident from well data may be a
(Figure 5). The steep dips observed in the lower relatively minor reverse structure, resulting in an
section were unexpected based on seismic overall box-fold geometry, which is not evident at
interpretation, but are corroborated by AZD image seismic scale. The important factor is that both data
analysis over the same interval as the lower XRMI sets confirm the overall orientation of fold and fault
image. trends.

Despite the large data gap between the upper and Drilling induced features and in-situ stress
lower image intervals (2086-3005 m), the
Drilling induced tensile fractures (DITF) were
relationships apparent when the data are viewed as
observed in both the upper and lower image
a whole appear to confirm the large-scale fold
intervals of Kurnia-1. Orientation of features in the
geometry suggested by the seismic analysis with
XRMI images showed that DITFs dominantly strike
some additional complexities (Figure 5). As
NNE-SSW (Figure 7). Breakouts were less
illustrated in the dip azimuth vector plot of Figure
common in both intervals and are generally low
5, there is a near continuous anti-clockwise azimuth
confidence picks as the observed hole enlargement
rotation from the NNE to WSW. This is
was often coincident with local concentrations of
accompanied by a gradual decrease in dip from
natural fractures, making it difficult to rule out hole
base to top of the well, albeit punctuated by some
enlargement by washout of fractured material. The
probable fault modifications between around 1740
interpreted breakouts, although of low confidence,
m and 1850 m in the upper image interval (Figure
do generally plot as expected at 90° to DITF strike
5). Taken together these trends are consistent with
(Figure 7). The orientation of the DITFs indicates
the strata from Kurnia-1 forming a large-scale NE
that SHmax is orientated NNE-SSW and Shmin
verging antiform (Figure 6). The orientation of this
WNW-ESE, consistent with the overall large-scale
fold was tested by plotting all data on a stereoplot
fold geometry and suggests that the stress regime
that shows that the bedding forms a linear girdle
observed today may be long-lived.
consistent with limbs of a fold with a calculated
axis plunging shallowly WNW (Figure 5). This The concentration of DITFs is greater in the Kujung
orientation was confirmed by plotting the data on Formation than in the upper clastic section. Whilst
an azimuth/depth histogram which shows a clear U- this may be partially controlled by lithological
shaped profile (Figure 5). properties, there is also a strong correlation with
drilling mud weight: the DITF frequency is highest
In addition to overall bedding trends, several faults where mud weight was highest, i.e. 16.5 ppg
were interpreted to intersect the wellbore either by between 3080-3114 m (Figure 10).
CHARACTERISTICS OF FRACTURES IN clustered, a conclusion that it is reflected in the
THE KUJUNG FORMATION fracture density distribution (Figure 10). To test
these conclusions, the orthogonal width of fracture
Fractures are common in the carbonates of the clusters was determined and the spacing of clusters
Kujung Formation and are easily recognised in the in each group calculated in order to fully
XRMI images collected in the lower interval of characterise fracture distribution (Table 1).
Kurnia-1. Fractures were classified primarily as
resistive or conductive relative to their host Figure 10 shows the approximate distribution of
lithology (e.g. Figure 8). In the case of conductive fracture clusters for each fracture group and Table 2
fractures, some show partial sinusoid traces on the shows statistics for the cluster density (number of
borehole image that are similar to DITFs (Figure 8). fractures within a cluster). These data show that the
The image trace tends to be strongest in the portion NNE-SSW striking fractures have the largest
of the image that corresponds to the tensile failure number of fractures per cluster (maximum 37 and a
direction (as determined from the occurrence of mean of around 4), although this is closely matched
DITFs) and diminishes away from this point by N-S striking fractures, which have a mean
(Figure 8). Previous studies have documented cluster density of around 3 fractures per cluster and
similar features elsewhere and suggest that the a larger total number of clusters at 77. The NW-SE
coincidence of these features with the DITF striking fractures have a mean of less than 2
orientation and their incomplete image trace are fractures per cluster and a median of 1, highlighting
caused by the fractures being drilling enhanced and the isolated nature of fractures in this orientation.
opened in response to borehole hoop stresses However, it should be noted that the mode for each
(Barton and Zoback, 2002; Barton et al., 2009). In group is one fracture indicating that isolated
view of this fact these fractures were classified fractures are common in each group (Table 2).
separately as drilling enhanced features.
Fracture aperture assessment
The imaged fractures are variable in orientation, but
can be divided into three groups based on strike The calculation of fracture aperture values from
orientation (Figure 9): group 1 strikes dominantly borehole resistivity images is complex. Several
NNE-SSW, group 2 strikes N-S and group 3 NW- methods have been suggested in different studies
SE. Conductive fractures are more common than based on measurement of excess conductance over
resistive (Figure 9), but this may be related to the the fracture interval. The most common algorithm
dominance of resistive calcareous lithologies in the used was devised by Luthi and Southaité (1990)
Kujung Formation making resistive fractures more and has many limitations that are discussed in detail
difficult to detect. No clear crosscutting relationship by the authors. The most important issue for the
was observed where one set of fractures is offset by current study is that the method is considered to be
another. Accordingly the temporal relationship of invalid for fractures with > 40° dip, which is the
the different strike groups cannot be determined. case for all fractures in the Kujung study interval.
To test this effect, the algorithm was used across
The fractures were further characterised by the lower image section of Kurnia-1, but was found
calculating spacing and density parameters to generate patently unrealistic apertures of >5 m.
specifically: orthogonal spacing of fractures in each Based on these results the method was considered
set; cluster thickness; un-fractured matrix thickness; to be inappropriate for the study well.
and along hole fracture and fracture cluster density.
Summary statistics for each of these parameters are As an alternative, a method was used in which the
provided in Tables 1 and 2 and discussed briefly interval of excess conductance associated with the
here. Calculated spacing statistics for individual fracture is measured and a pixel distance derived
fractures indicate that fractures in all three from the image that is equivalent to the fracture
orientation groups are on average closely spaced height calculated by Luthi and Southaité (1990).
(Table 1). Minimum spacing is around 1 cm and the The method is cruder than that of Luthi and
majority of fractures have spacing of less than 15 Southaité (1990), but is effective in giving a
cm. The full range of spacing values is variable up qualitative estimate of average fracture aperture. A
to a maximum of around 50 m for NW-SE striking summary of the derived apertures is provided in
fractures. Spacing statistics for all three groups Table 3.
have a lognormal distribution with spacing values
heavily skewed towards small spacing in each Based on visual examination of the XRMI images,
group, indicating that the fractures are strongly it is unlikely that any of the fractures have apertures
greater than 1 cm, and the apertures determined Relationship of fractures to large-scale structure
using the excess conductance technique appear to
provide reasonable results for the majority of the As outlined, there is strong evidence from bedding
fractures measured with a median calculated dip trends from Kurnia-1 to support the case that
aperture of 1 cm and mode and mean of 3 mm. If the strata form a large-scale anticlinal structure
the effects of apparent versus true thickness (due to (Figures 5 and 6). Asymmetric folds of this type are
angle of intersection of the fracture with the common in the over-riding plate in fold and thrust
wellbore) are taken into consideration, then it is belt settings at subduction boundaries as is the case
more likely that the true fracture aperture is on the for Madura Island. Seismic interpretation of the
millimetre rather than centimetre scale. The very Kurnia structure suggests that folding is fault
large calculated apertures of as much as 51.5 m are controlled in the form of a forced fold (Cosgrove
generated where the algorithm detects more and Ameen, 2005) with folding induced in the
conductive responses from adjacent carbonate overlying strata as the fault progressively breaks
mudstone beds and interprets these as a fracture through from below.
response. These excessively large apertures are
clearly erroneous and should be ignored. The mechanical nature of such folds has been
studied in great detail and the fracture patterns
expected in these features can be readily predicted
VIABILITY OF THE KUJUNG AS A based on the variance of outer arc extension and
FRACTURED RESERVOIR inner arc compression around the fold hinge (e.g.
Price 1966; Stearns 1978; Cosgrove and Ameen
The viability of a fractured reservoir depends on the 2005). Based on the method of Price (1966) a
fractures being open and connected. In Kurnia-1 the theoretical fracture orientation set was generated to
first qualitative indication of an open fracture compare to the fractures observed in Kurnia-1 in
network was obtained when the well repeatedly order to assess whether the fracture orientation sets
―kicked‖ when drilling through the upper Kujung could be generated purely from folding (Figure 11).
section, and the associated gas had to be circulated In this theoretical scenario the upper clastic section
out through the choke manifold while the would sit on the gently SW dipping limb and the
mudweight was increased. A production test was lower image interval, Kujung Formation, would fall
performed in an attempt to quantify the production on the steeply dipping NE limb (Figure 11). It can
potential from the fractured Kujung reservoir: this then be demonstrated that the theoretical fracture
flowed combustible hydrocarbons to surface, but patterns depicted in Figure 11 closely match all of
testing difficulties prevented the flow test from the defined sets observed in Kurnia-1. In addition,
being completed conclusively. the model predicts that the tensional (T) fractures,
which form parallel and perpendicular to the fold
In the absence of conclusive test results, qualitative hinge, will be orientated normal to bedding.
predictions of flow potential have been made based Therefore, the T1 and T3 fractures, formed parallel
on the image log characteristics of the fractures and to the hinge, should dip in the opposite direction to
their orientation relative to present day stress and/or bedding in each of the fold limbs and show reverse
known larger scale structures. In the following dip on either side of the fold limb (Figure 11). This
sections the character and orientations of the is indeed the case for Kurnia-1: in the upper image
fractures in the Kujung Formation are considered in interval (on the SW dipping limb) bedding dips
an attempt to qualitatively rank fracture orientations broadly to the W or SW whilst the hinge-parallel
with respect to their potential for fluid flow. fractures dip to the NE; and in the lower image
Kujung interval beds dip to the NNE whilst hinge-
It is very difficult to assess the openness and parallel fractures dip to the SW (Figure 11).
connectivity of a fracture network based purely on
wireline logs and borehole image logs. It is often An additional feature of forced folds observed in
assumed, when drilling with conductive water type examples such as the Zagros Mountains of Iran
based muds, that conductive fractures are open (e.g. Satterzadeh et al. 2000; Stephenson et al.
because they give a mud response. However, a 2007), is that they tend to form periclinal or doubly
conductive response would also be obtained from plunging non-cylindrical folds, which can cause
closed fractures filled with conductive enhancement of the fractures that strike
mineralisation such as pyrite or shale gouge. perpendicular to the main fold hinge (T2 and T4 in
Conversely, resistive fractures suggest cementation Figure 11). This could explain the preponderance of
and yet may transmit fluid. NNE striking fractures observed in both of the
image intervals of Kurnia-1 (Figures 10 and 12). often referred to as being critically stressed
Several lines of evidence support a periclinal fold (Hickman et al 1997; Finkbeiner et al. 1997; Barton
geometry for the Kurnia structure, but the most et al. 1998; Trice 1999; Rogers 2003). Figure 12
compelling is Kujung structure map that clearly shows that the majority of conductive fractures fall
shows a doubly plunging structure (Figure 3). within the 30° preferential window and that those
that were classified as drilling enhanced tend to fall
The fact that the Kurnia structure bears in the critically stressed orientation and would
resemblance to examples such as the those from the therefore be expected to more easily dilate during
Zagros Mountains is non-trivial because the drilling. Of course, this assessment is purely
lithologies that make up the Asmari Formation in qualitative because modelling of the full stress
the Iranian example are also dominated by pelagic tensor and knowledge of the mechanical nature of
carbonates with little matrix porosity and the rocks is required to accurately predict the
permeability, but which have high hydrocarbon critical stress orientation, but it serves as a
yields through secondary fracture porosity (e.g. reasonable means of ranking fracture orientations
Sattarzadeh et al. 2000, Stephenson et al. 2007 and that would be favourable for hydrocarbon
Wennberg et al. 2007). Further examples of similar extraction.
forced fold carbonate reservoirs can be found in
Oman (De Keijzer et al. 2007). Although the Fracture porosity and permeability
Kurnia-1 interpretation is based on only one well,
the comparison with the examples in Zagros and Using the calculated statistics for fracture
Oman is potentially important as both host fields distribution it is possible to make a semi-
with large hydrocarbon reserves. quantitative estimate of the secondary porosity that
may be derived from fractures and also a qualitative
Fractures and in-situ stress estimate of permeability potential. The estimates
are calculated in two ways both using the mean
As well as the comparison to proven fractured matrix block size for each fracture group but using
reservoir analogues, some other features of the different values for the assumed aperture: method 1
Kurnia fracture orientation make them attractive for uses the mean cluster width and assumes fractures
hydrocarbon production. As outlined, the fracture are open for the full width; and method 2 uses the
orientations compare well with the overall large- mean aperture combined with the mean number of
scale structure and are therefore consistent with fractures per cluster to derive a flow aperture
having formed under the same stress field that (Figure 13). The results of these calculations are
resulted in folding and faulting. However, a perhaps given in Table 4.
more important feature is the fact that the fracture
orientations are consistent with the orientations of The statistics show that the porosity values
present day in-situ stress based on breakout and calculated purely on cluster width (method 1) are
DITF orientations. Figure 12 shows a very very high in the order of 20% for NNE-SSW and
simplified orthogonal plain strain ellipse for N-S striking fractures. This estimate is considered
fracture mode based on a NNE-SSW orientated excessive as the cluster width calculation is based
maximum horizontal stress. This allows a basic only on the distance between the first and last
orientation ranking to be made whereby fractures fracture in each cluster and does not take into
that are NW-SE trending i.e. with strike parallel to account the matrix rock that lies between each
the minimum horizontal stress (Shmin) orientation fracture. The porosity estimate based on method 2
are the least likely to be open as they are held shut is considered to be more realistic and gives mean
by orthogonal compression (Figure 12). By secondary porosity estimates of around 0.5%, in
comparison, fractures striking parallel to SHmax line with the worldwide average for fractured
(NE-SW) are considered more likely to allow fluid reservoirs (e.g. Nelson 2001). In the most intensely
flow as they can orthogonally dilate in the direction fractured zones, however, secondary porosity
of least compressional stress (Figure 12). estimates, calculated using the maximum number of
Additionally, previous studies have shown that fractures per cluster, can approach 5% in the case of
whilst fractures parallel to SHmax may be open, they NNE-SSW striking fractures. The N-S and NW-SE
are not necessarily the most permeable (e.g. Barton striking fractures also have localised secondary
et al 1998; Rogers 2003). Field testing and porosity estimates approaching 3% (Table 4). If
modelling suggests that in many cases the most more than one set of fractures is open and
permeable fractures are those that lie oblique to connected then the combined porosity would be
SHmax and are experiencing shear failure and are high.
To convert these estimates into permeability during drilling and the distribution of fractures in
potential it is inherent that assumptions are made the well, and an incomplete drill stem test showed
that all of the fractures are open to fluids (i.e. are that it was possible to produce combustible gas to
not cemented or filled with gouge) and are surface. All of these features combine to suggest
connected away from the wellbore and have infinite that the Kujung Formation is indeed a viable
extension along their length. If this were the case fractured hydrocarbon reservoir.
then NNE-SSW striking fractures alone would
provide reasonable permeability potential and if Information gathered on the nature of the fracture
connected to either of the other groups then network will be useful in well design for future
potential for flow is good. appraisal and exploitation of the fractured Kujung
reservoir. The structural model can be used to
Fracture occurrence compared to mudlog gas predict zones of most intense fracturing. Production
shows wells should be drilled as horizontal or highly
deviated wells in a northwesterly orientation in
All of the discussions presented suggest that there is zones of greatest predicted fracture intensity to
indeed very good potential for the fracture sets in maximise the intersection of open fractures. Further
the Kujung Formation to form a viable hydrocarbon geomechanical studies should also be performed to
reservoir. However, all of these methods are in one optimise mudweight to control formation pressure
way or other indirect tests of flow potential. Despite while avoiding invasion and damage of fracture
the problems encountered during production network.
testing, it is very encouraging that combustible
hydrocarbons flowed to surface, although it is not ACKNOWLEDGEMENTS
clear that these were sourced from the fractured
intervals. In this regard the record of mudlog gas The authors wish to thank the management and
proves extremely valuable in demonstrating that the geotechnical staff of Cooper Energy and our joint
extracted gas is most likely sourced from the venture partners in the South Madura PSC, Nations
fractured intervals. Figure 10 shows fracture Petroleum and PT Eksindo for their support and
density and cluster occurrence for each fracture set encouragement to conduct this study and to draft
plotted alongside mudlog gas. This plot highlights this paper and to MIGAS for their permission to
that there is a direct correlation between fracture publish this paper.
occurrence and high mudlog gas levels. In addition
the mudlog gas is highest over intervals of NNE- REFERENCES
SSW and N-S striking fractures, confirming
predictions that fractures within 30° of SHmax are Barton, C.A., D. A. Castillo, D. Moos, P. Peska,
more likely to be open to flow. However, gas shows M.D. Zoback, 1998, Characterizing the full stress
do not seem to correlate with the occurrence of tensor based on observations of drilling-induced
DITFs: the interval of greatest DITFs (3080 to 3115 wellbore failures in vertical and inclined boreholes
m) has very low mudlog gas (Figure 10). This is leading to improved wellbore stability and
encouraging as it suggests that whilst some natural permeability prediction, APPEA Journal 1998, 29-
fractures may be enhanced during drilling, they are 53.
open anyway and not required to be held open
under mud pressure to produce gas. Cosgrove, J. W. and Ameen, M. S., 2005. A
comparison of the geometry, spatial organisation
SUMMARY and fracture patterns associated with forced folds
and buckle folds. In: COSGROVE, J. W. &
Based on the data outlined in this study, it has been AMEEN, M. S. (eds.), Forced Folds and Fractures.
shown that, despite poor primary matrix porosity Special Publication of the Geological Society,
and permeability in the carbonates of the Kujung London, 169, 7-21.
Formation, fractures within the formation appear to
have the necessary porosity and permeability to De Keijzer, M., Hillgartner, H., Al Dhahab, S. and
host and produce hydrocarbons. The fractures Rawnsley, K., 2007. A surface-subsurface study of
appear to have formed subject to a consistent long- reservoir-scale fracture heterogeneities in
lived stress field and are optimally orientated with Cretaceous carbonates, North Oman. In: Lonergan,
respect to present day in-situ stress to be open to L., Jolly, R. J. H., Rawnsley, K. and Sanderson, D.
fluid flow. There is extremely good correlation J. (eds.), Fractured reservoirs. Special Publication
between the occurrence of observed gas shows of the Geological Society, London, 270, 227-244.
Finkbeiner , T., Barton, C., Zoback, M., 1997. Seubert, B.W. and Sulistianingsih, F., 2008. A
Relationships among in situ stress, fractures and proposed new model for tectonic evolution of South
faults, and fluid flow: Monterey Formation, Santa Java, Indonesia, Proc. Indonesian Petroleum
Maria Basin, California. AAPG Bulletin 81, 1975- Association 32nd Annual Convention and
1999. Exhibition, digital paper IPA08-G-034.
Hall, R., 2002. Cenozoic geological and plate Sharaf, E.F., Simo, J. A., Carroll, A.R. and Shields,
tectonic evolution of SE Asia and the SW Pacific: M., 2005, Stratigraphic evolution of Oligocene–
computer based reconstructions, model and Miocene carbonates and siliciclastics, East Java
animations, Journal of Asian Earth Sciences 20, Basin, Indonesia, AAPG Bull., V. 89, No. 6, 799-
353-431. 819.
Hamilton, W.B., 1979. Tectonics of the Indonesian Sribudiyani, Muchsin, N., Ryacudu, R., Kunto, T.,
region: USGS Prof. Paper 1078, 345 p. Astono, P., Prasetya, I., Sapiie, B., Asikin, S.,
Harsolumakso, A.H., and Yulianto, I., 2003. The
Hickman, S.H., Barton, C.A., Zoback, M.D., collision of the East Java microplate and its
Morin, R., Sass, J. and Benoit, R. 1997. In situ implications for hydrocarbon occurrences in the
stress and fracture permeability along the Stillwater East Java Basin, Indonesian Petroleum Association,
fault zone, Dixie Valley, Nevada. International Proceedings 29th Annual Convention, 335-346.
Journal of Rock Mechanics and Mining Science 34,
414. Stearns, D. W., 1978. Faulting and forced folding in
the Rocky Mountains foreland. Geological Society
Johansen, K.B., Depositional geometries and of America Memoir, 151, 1-38.
hydrocarbon potential within Kujung carbonates
along the North Madura Platform, as revealed by Stephenson, B. J., Koopman, A., Hillgartner, H.,
3D and 2D seismic data. Proc. Indonesian McQuillan, H., Bourne, S., Noad, J. J. and
Petroleum Association, 29th Annual Convention, Rawnsley, K. 2007. Structural and stratigraphic
digital paper IPA03-G-174. controls on fold-related fracturing in the Zagros
Mountains, Iran: implications for reservoir
Luthi, S. M. and Southaité, P., 1990. Fracture development. In: Lonergan, L., Jolly, R. J. H.,
apertures from electrical borehole scans. Rawnsley, K. and Sanderson, D. J. (eds.), Fractured
Geophysics, 55, 821-833. reservoirs. Special Publication of the Geological
Society, London, 270, 1-21.
Nelson, R.A., 2001. Geological analysis of
naturally fractured reservoirs. Butterworth- Trice, R.A. 1999. Application of borehole image
Heinemann, Woburn. 93pp. logs in constructing 3D static models of productive
fracture networks in the Apulian Platform, Southern
Park, R.K., Matter, A., Tonkin, P.C., 1995. Porosity
Appennines. In: Lovell, M. A. et al. (eds.) Borehole
evolution in the Batu Raja Carbonates of the Sunda
imaging: applications and case histories. Geological
Basin—Windows of opportunity. Indonesian
Society, London. Special Publication 159, 77-90.
Petroleum Association, Proceedings 24th Annual
Convention, 63–184. Wennberg, O. P., Azizzadeh, M., Aqrawi, A. A. M.,
Blanc, E., Brockbank, P., Lyslo, K. B., Pickard, N.,
Price. N. J., 1966. Fault and joint development in
Salem, L. D. and Svana, T. 2007. The Khaviz
brittle and semi-brittle rock, Pergamon, Oxford.
Anticline: an outcrop analogue to giant fractures
Rogers, S.F., 2003. Critical stress-related Asmari Formation reservoirs in SW Iran. In:
permeability in fractured rocks. In: Ameen M. (ed.) Lonergan, L., Jolly, R. J. H., Rawnsley, K. and
Fractures and in-situ stress characterizations of Sanderson, D. J. (eds.), Fractured reservoirs.
hydrocarbon reservoirs. Geological Society, Special Publication of the Geological Society,
London. Special publication 209, 7-16. London, 270, 23-42.

Sattarzadeh, Y., Cosgrove, J. W. and Vita-Finzi, C., Wilson, M., 2002. Cenozoic carbonates in
2007. The interplay of faulting and folding during Southeast Asia: implications for equatorial
the evolution of the Zagros deformation belt. In: carbonate development. Sedimentary Geology, 147,
Cosgrove, J. W. & Ameen, M. S. (eds.), Forced 295-428.
Folds and Fractures. Special Publication of the
Geological Society, London, 169, 187-196.
 TABLE 1 - SUMMARY STATISTICS OF FRACTURE SPACING

Individual fracture spacings Cluster width Matrix block width

Group 60°/118 56°/270 45°/211 60°/118 56°/270 45°/211 60°/118 56°/270 45°/211
orientation ° ° ° ° ° ° ° ° °
Count 345 314 86 50 77 42 50 77 42
Min
Spacing
(m) 0.013 0.008 0.013 0.072 0.090 0.070 0.558 0.149 0.163
Max
Spacing
(m) 35.433 36.548 53.126 3.086 2.030 1.013 35.350 36.420 30.830
Geometric
Mean (m) 0.164 0.270 0.329 0.485 0.493 0.161 2.294 1.559 1.671
Median (m) 0.115 0.197 0.231 0.565 0.507 0.116 2.231 1.507 1.269
Mode (m) 0.060 0.088 0.047 0.087 0.135 0.082 0.565 0.807 N/A
St Dev 2.497 2.439 6.588 0.848 0.561 0.231 5.559 4.444 6.820
Error Wε 0.012 0.011 0.020 0.012 0.011 0.020 0.012 0.011 0.020
Acute angle 34.790 32.179 48.795 34.790 32.179 48.795 34.790 32.179 48.795
Kurtosis 115.834 149.200 43.798 0.455 -0.294 2.540 21.405 44.957 6.807
Skewness 9.453 10.793 6.121 1.187 0.852 1.756 4.180 6.130 2.627

TABLE 2 - SUMMARY STATISTICS OF CLUSTER DENSITY FOR EACH ORIENTATION GROUP

Group
60°/118° 56°/270° 45°/211°
orientation
Number of
clusters 50 77 42
Min number of
fractures 1.000 1.000 1.000
Max number of
fractures 37.000 15.000 15.000
Mean 3.860 2.997 1.604
Median 4.000 3.000 1.000
Mode 1.000 1.000 1.000
St Dev 7.973 3.223 2.352
Kurtosis 4.30 2.033 21.81
Skewness 2.08 1.526 4.23
 TABLE 3 - SUMMARY OF CALCULATED FRACTURE APERTURES

statistics - all fractures

number of
fractures 633.000
Min aperture (m) 0.003
Max aperture (m) 51.506
Mean aperture
(m) 0.003
Median aperture
(m) 0.010
Mode 0.003
St Dev 2.685
Kurtosis 220.717
Skewness 13.168

TABLE 4 - SUMMARY OF FRACTURE POROSITY ESTIMATES

Group 1 - Group 2 - Group 3 -

Units 60°/118° 56°/270° 45°/211°
Mean cluster separation m 2.294 1.559 1.671
Mean cluster width m 0.485 0.493 0.161
Mean number of fractures 3.9 3.0 1.6
Max number of fractures 37 15 15
Mean fracture aperture m 0.003 0.003 0.003
Effective aperture m 0.012 0.009 0.005
Effective Max aperture m 0.111 0.045 0.045
Porosity estimate from cluster
% 17.4 24.0 8.8
width
Porosity estimate from
fracture aperture and mean % 0.5 0.6 0.3
number of fractures
Maximum Porosity from
aperture and max number of % 4.6 2.8 2.6
fractures in cluster
Figure 1 - (a) Map of the East Java Basin showing dominant tectonic trends reflecting the dominant
influence of subduction along the Java Trench. Location of the study well Kurnia-1 is also
shown on the south coast of Madura Island. (b) General stratigraphic column for the East Java
Basin, note that stratigraphic unit names in this area are not standardised and the nomenclature
of Shell is adopted here for convenience.

Figure 2 - Photomicrographs of thin sections taken from sidewall cores in the Kujung Formation, showing
typical matrix characteristics of the carbonates.
Figure 3 - Seismic depth map of top Kujung horizon. The Kurnia structure is characterised as a doubly
plunging anticline with a dominant NW-SE axis. The positions of two reverse faults controlling
formation of the structure are shown in relation to the collar position of Kurnia-1.
Figure 4 - SW-NE Seismic line showing location of Kurnia-1 relative to reverse faults.
Figure 5 - Combined vector azimuth walkout plot, stereonet and dip azimuth/depth plot of all bedding in
the Kurnia-1 image intervals. Note that there is a data gap from 2086-3005 m between the
intervals. The vector walkout plot indicates a near continuous anticlockwise azimuth rotation
from NNE dip to SW dip from base to top of section. This relationship is further confirmed by
the stereoplot which shows a well defined elongate girdle indicative of a cylindrical fold feature
and the azimuth/depth plot which shows a clear U-shaped distribution. The calculated fold axis
is 8°/286° from the stereoplot and 10°/285 from dip/azimuth. Vector walkout plot created using
Task Geoscience Attitude™ software, dip/azimuth plot created using Quickdip™ software.
Figure 6 - Comparison of bedding fabric orientations in XRMI and ALD imagery. Both image types
confirm steep NNE dipping bedding in the Kujung Carbonate section. Figure 6 - Possible fold
geometry for strata in Kurnia-1 incorporating imaged NW-SE trending faults from seismic
interpretation. In this scenario it is interpreted that steep NNE dipping beds in the Kujung
Formation are formed by frictional drag along the red reverse fault. The SW dips that are
dominant in the seismic interpretation on the south limb of the structure are formed by
movement across the blue fault. Well dip projection to section line created using Task
Geoscience Attitude™ software.
Figure 7 - Examples of DITFs from lower image interval in Kujung Carbonate section. Left image shows
typical DITF forming two conductive stripes 180° apart in XRMI imagery. Note that AZD
response appears largely unaffected. DITFs appear to be offset along natural fracture traces and
natural fractures may be enhanced at points of intersection. In right image DITFs can be seen to
branch into so called petals. Note that DITFs appears bed bound in this case. Rose diagrams
show DITF and breakout orientations for features picked in the upper and lower image intervals
of Kurnia-1: these dominantly indicate a NE-SW orientation for SHmax and NW-SE orientation
for Shmin.
Figure 8 - Examples of image characteristics of fractures in the Kujung Formation. Note scale varies
slightly in each image.
Figure 9 - Summary of fracture orientations observed in the lower image interval of well Kurnia-1.
Figure 10 - Log plot showing a comparison of fracture density distribution, as raw density (black) and
corrected for borehole sampling bias (red); fracture cluster occurrence per group; DITF
density and mudlog gas. The comparison indicates that the gas shows all occurred at fractured
intervals and most commonly where fractures strike N-S or NNE-SSW. There appears to be
an almost negative correlation with DITF occurrence: where DITFs are most numerous
mudlog gas is low.
Figure 11 - (a) Predicted orientation of fractures within the limbs and hinge of an asymmetric forced fold
(after Price 1966; Cosgrove and Ameen 2005). Fractures labelled T are tensional while R
fractures are shear fractures. (b) Example of predicted fracture orientations adapted for a
doubly plunging periclinal fold geometry (after Stearns 1978). Note that the R or shear
fractures (red and green) change orientation around the outer flanks of the dome (right of
image) reflecting the change in orientation caused by the periclinal folding and tending to
parallelism with the T fractures (magenta) that form along the main hinge line. (c) Correlation
of predicted fracture patterns with those observed in Kurnia-1. The correlation fit is very good
with all of the predicted fracture orientations observed in the well data.
 Figure 12 - Simple plane strain ellipse for the Kurnia structure illustrating expected shear displacements
under NE-SW orientated maximum compressive stress. Rose plot shows a qualitative
prediction of fractures that may be under critical stress and therefore more likely to be open to
fluid flow.

Figure 13 - Summary of methods used to calculate fracture porosity.

View publication stats