RAI Attribute
RAI Attribute
Department of Geosciences,
Universiti Teknologi Petronas
Email : fathiyah.jamaludin@utp.edu.my
ABSTRACT
This study is focusing on the characterisation of Pleistocene to present day fluvial system found in Sepat Field in the
offshore of Terengganu. It aims to characterise the fluvial system in Sepat Field and correlate the fluvial systems found
in Sepat Field with the larger scale fluvial system of the Malay Basin. The methodological approach used for this
subsurface study includes geological interpretation of 3-dimensional seismic data and wireline logs interpretation.
Frequency spectral decomposition combined with variance and relative acoustic impedance attributes had provided
a precise prediction on the dimensions, flow orientation and type of the fluvial systems in this field. This study shows
that the fluvial system in this field is dominated by meandering channels that flow from northwest to southeast.
Points bars, straight channels, channels boundaries and the evolution of the channels can be imaged and interpreted
in this study through frequency spectral decomposition technique. A “hook” shape point bar stands out clearly in the
data with depth up to 1.5km and is interpreted to be related to the Malay basin fluvial system. The fluvial system in
Sepat Field is interpreted to deposit in the midstream river system and highly associated with the Chao Phraya-Johore
River, with sediment sources come through this drainage basin and distributed to the tributaries fluvial channels
including those in Sepat Field.
Keywords: seismic attributes, fluvial characterisation, Malay Basin, spectral decomposition, meandering river
Sepat Field is made even difficult by the existence Malay Basin has gone through three major tectonic
of gas clouds that degraded the seismic data quality events that form the observed structure in the basin
[5]. With that, this study was conducted to utilise the today (Figure 2). The first tectonic event was extension
seismic attributes in characterising the fluvial system phase started during Late Cretaceous to Early Miocene
of Sepat Field and correlate it the larger scale fluvial followed by the second event, the thermal subsidence
system of the Malay Basin. which took place from Early Miocene to Middle Miocene
Figure 1 The location map is showing Sepat Field within the Malay Basin, in offshore of Terengganu
GEOLOGICAL SETTING OF THE STUDY AREA and the last tectonic event was compression stage which
happened from Late Miocene to Pliocene [9]. The Malay
The Malay Basin is situated at the centre of Sundaland, basin experienced inversion during the Middle to Late
the cratonic core of Southeast Asia. It elongates Miocene, which took place during the 3rd tectonic event,
from NW – SE and spreads a zone of around 500km the compression stage [8],[6],[9].
long and 250km wide [6]. It is found seaward in the
South China Sea, east of Peninsular Malaysia with the The Malay Basin strata are subdivided informally into
latitude of 4 – 8 N and longitude of 102 – 106 E [7]. It is seismo-stratigraphic units [10] refer as “Group”, derived
one of the deepest continental extensional basins in exclusively from the oil companies working in this region.
the region and the biggest among the hydrocarbon- The “Group” is named according to alphabetical orders,
bearing Tertiary basin of the Sunda Shelf. The Tenggol starting from Group A as the youngest (Pliocene-recent)
Arch isolates the Malay Basin from Penyu Basin [9], and Group M as the oldest (Lower Oligocene). Figure 2
while the Narathiwat High separates the Malay Basin summarises the stratigraphic units used in the Malay
from Thailand’s Pattani Basin [8], [3]. Basin, in relation to the structural history of this basin.
Figure 2 Stratigraphy of Malay Basin from Group A-M showing the dominant lithofacies in fluvial,
estuary sand and marine with delta front sands. The right-most column is displaying the major
tectonic events that had affected this basin, modified after [3]
A 3D seismic volume and three sets of wireline logs The first step taken in the interpretation of the data
from Sepat Field in offshore Malaysia were used is establishing the seismic to tie well. Data acquired
throughout this study. The data was provided by in the geophysics exploration is in the two-way-
Malaysia Petroleum Management (MPM), PETRONAS time domain while the data from the good logs is
for this study. The data was acquired in the year an in-depth domain. Seismic to well tie enables the
2002 with a seismic grid of 12.5m inline spacing and comparison of the data quality in depth and time
25.0m crossline spacing. The seismic data quality domain. Sonic and density logs from Sepat Deep-1 well
ranges from poor to good, with the poor quality is the seismic to well tie was performed with a dominant
mainly caused by the shallow gas clouds effect. The frequency of 30Hz to 40Hz. A moderate to good well
minimum bed resolution is estimated at 17m with to seismic tie with minor bulk shift was generated in
a dominant frequency of 43Hz. Since this study is this study.
focusing on the younger stratigraphic units of Group
A and Group B, the seismic data was cropped only to Seismic Attributes Interpretation
the interest section. Well, data from Sepat-3, Sepat-8,
Sepat Barat-1 and Sepat Deep-1 were used to tie the Horizons are picked within Group A and Group B
geological interpretation across the Sepat Field area. sections from -100ms to -1250ms in a two-way time
domain. The time slices at a time -100ms to -500ms Relative Acoustic Impedance Attribute
clearly show the traces of rivers and channels.
Nonetheless, the rivers and channels become more Relative Acoustic Impedance works based on the
difficult to trace after -500ms due to chaotic and physical properties of the rocks, computed by taking
disruption on the reflections of the seismic data due to the density and interval velocity of the strata. It is
gas seepages. Application of True Amplitude, Relative usually used in the inversion to get the correct value
Acoustic Impedance and Coherence-Variance seismic to acoustic impedance [11]. The application of this
attributes on the seismic data that is affected by attribute usually beneficial for thin beds, but in this
shallow gas seepages have been proven to improve the study, it is applied to show the better boundary of the
quality of seismic images as presented on the paper by channels. Relative acoustic impedance measures the
[5]. In this study, we had emphasis the application of running sum of the trace to which a low-cut filter is
Frequency Spectral Decomposition and Red-Green- used where is it applied to remove the shift which is
Blue (RGB) and Cyan-Magenta-Yellow (CMY) colours typical in impedance data [11]. The calculated trace is
blending, along with the application of the Variance composed of the simple integration of the complex
and Relative Acoustic Impedance attributes. trace and represents the approximation of the
high-frequency component of the relative acoustic
Variance Attribute impedance. Therefore, relative acoustic impedance
helped in the detection of thin channels beds at a
Variance is the measurement of differences between minimum area.
waveforms and traces. It operates on a spatial window
based on the adjacent traces [11]. Variance detects Frequency Spectral Decomposition
discontinuities that produce trace- to- trace variability & RGB Blending
over a vertical analysis window of a certain number
of waveforms samples which was later normalised by Frequency decomposition replaces the single input
the average energy traces obtain from the seismic [12], trace with a gather of traces corresponding to the
[11]. A high value of variance indicates heterogeneity spectral decomposition of the input attribute. The raw
which is usually caused by faults and rivers. Variance is seismic volume is used as an input, and the output is
a suitable seismic attribute to be used in this study. a few volumes with different frequency band as the
Figure 3 Bandwidth of 10Hz to 80Hz set for frequency spectral decomposition attribute in this study
user defined. Spectral decomposition allows discrete volumes. The decomposition is achieved through the
frequency responses to be isolated and combined application of a set of bandpass filters to the seismic
with red, green and blue (RGB) blending that traces. Figure 3 shows the frequency coverage from
enables detection of structures which have different 10Hz to 80Hz to ensure the seismic bandwidth of
frequency bands as resolution thickness depends both low and high-frequency amplitude. Figure 4
on the wavelength. Lower frequency has a longer shows the workflow for spectral decomposition using
wavelength that produces thicker tuning thickness, GeoTeric ® 2017 software that was applied in this
and higher frequency has a shorter wavelength that study. This technique used was uniform constant Q
produces thinner tuning thickness. Therefore, it is which is analogous to wavelet transform. Using this
necessary for the presence of both low and high method, the frequency modulation is kept constant,
frequency to ensure thick beds and thin beds are and the bandwidth scale is varied.
detected respectively.
tops from Sepat-3 Well was used as a reference to pick The frequency and amplitude are low with chaotic
the horizons. The focus group of this study starts from reflectors in the middle section due to gas seepage
100ms to 1250ms consisting of Group A (Pleistocene effect. Normal faults are commonly observed in this
to Recent) and Group B (Upper Miocene). However, section. The reflectors within Horizon B have sub-
due to poor seismic data at the upper section, which parallel to chaotic reflections, and the presence of
is caused by the gas seepage, only four horizons were Direct Hydrocarbon Indicator (DHI) is frequent in this
picked that falls in Group A and B. Group A is divided group due to gas seepage. Anticlinal domal structure
into A1 and A2 since it is a thick bed. Another four is also observed in this group.
seismic horizons were selected below time 1250s to
represent the structure of the Sepat field.
Figure 5 Map view of the arbitrary line A-A’ showing all the stratigraphic horizons picked in this study.
The dashed lines are representing the fluvial traces in Group A and Group B
Figure 5 is showing the interpreted seismic cross Fluvial Characterization through Relative
section A-A’ that crossed all the four wells available Acoustic Impedance & Variance Attributes
for this study. Horizons A1, A2 and B belongs to Group
A and B while Horizon D and E belong to Group Detection of fluvial features including river and
D and E. The seismic characteristics for reflectors channels is compared using Relative Acoustic
within A1 and A2 show discontinuous bed at the top Impedance and Variance attributes. As the depth
followed by a more continuous, parallel reflectors. of the data goes more in-depth, the usage of both
Figure 6 Time slice at -225ms using (a) Relative Acoustic Impedance attribute and (b) Variance attribute.
Red boxes are highlighting the point bars and older channel cross-cutting the younger channel. The channels’
boundaries are precisely detected using Variance attribute while the sand body distribution within the point
bar is better determined using Relative Acoustic Impedance attribute
attributes is more difficult, because it also delineates distribution of sand bars within the point bars since
some noise from the gas seepage. In general, both it works relatively with the amplitude values (Figure
seismic attributes can show the body of channels, 6a). At the same time, Variance attribute provides
point bars and cross-cutting tributary channels. a reliable set of data for the interpretation of the
Relative Acoustic Impedance attribute acts as a more boundaries between the rivers and channels (Figure
powerful tool to give a better interpretation of the 6b).
Figure 7 Spectral decomposition analysis on time 225ms. (A), (B) and (C) shows frequency at 30Hz, 40Hz and
50Hz respectively that was produced as the output from spectral decomposition. (D) shows red – green –
blue (RGB) blend volume that is a combination of (A), (B) and (C). RGB channels of different frequency in one
volume. (E) shows Cyan Magenta Yellow (CMY) blending that is more frequently used for structural analysis.
The yellow, red and black box represents clear, medium and smeared image respectively
Fluvial Characterization through Frequency a meandering river eroded due to moving water in
Spectral Decomposition & Colors Blending the stream and created outer banks and widens the
valley [13]. The inner section of the point bar contains
Frequency bands of 30Hz, 40Hz and 50Hz were used less energy where silt will deposit and the exterior
respectively to separate the seismic characters within part exhibit higher energy where sand is stored. The
these bands. Figure 7a-7c represent the differences analysis from Sepat-3, Sepat-8 and Sepat Deep-1
in channels enhancement at time slice -225ms with wells shows that the fluvial system in Sepat Field is
a yellow, red and black box representing the good, dominated by silty sand and sandy lithologies.
medium and smeared image quality respectively.
Figure 7a, representing 30Hz bandwidth shows wider
channel (central section of the image) is at a good, DISCUSSION
satisfactory quality, while the smaller meandering
channels (east side) produce poor, smeared quality Dimension of the fluvial in Sepat Field
image at this frequency band. Figure 7b, representing
the 40Hz bandwidth shows a medium quality Most of the channels detected on the shallow surface
image of the time slices on all channels’ detection. of Group A and Group B are elongated, a long river
At bandwidth of 50Hz, the imaging of smaller with shallow depth.
channels (east side) is improved compared to within
bandwidth of 30Hz and 40Hz. The frequency spectral At the time of -225ms, the observation from RGB
decomposition technique shows that lower frequency colours blending had imaged five river systems
value gives a better result for wider channel while (annotated as no 1-5 in Figure 8a) flowing in several
higher frequency value is more suitable to delineate directions, channel cross-cutting each other and a
the smaller channels. “hook” channel at the edge of the map (annotated
as no.1 in Figure 8). As we go deeper to time -402ms
Figure 7d and 7e show two different colours blending, (Figure 8b), the map view of the river system in Sepat
which are Red-Green-Blue (RGB) and Cyan-Magenta- Field shows rivers labelled as 3, 4 and 5 were not
Yellow (CMY), respectively. Colours blending reveal deposited yet at this time. One wide river labelled
the geological features in seismic data more clearly as no 6 is observed at this time -402ms and it is
compared to single attribute viewing. RGB colours related A wider river appears at this time range, and
blend highlights the variations, heterogeneity and seismic images become harder to interpret due to gas
differences in the seismic data. It allows detection seepages, but the “hook” shaped meandering river
with similar amplitude characteristic, while CMY (labelled as 1) is still clearly visible. It is interpreted as
colours blend focuses more on the structural features an incised valley was tracked at the northeast section
such as faulting. In this study, RGB colours blending of Sepat Field (Figure 8a-8d). This “hook” shaped incise
produced a better time slice image for rivers and valley is detected from time 90ms to 1400ms, which is
channels interpretation compared to CMY colours approximately 1.5km deep and 0.65km width (Figure
blend. 9). At the time -606ms, the only river labelled as no 1,
and no 2 appears, which means these two rivers were
Type and lithology of the fluvial in Sepat Field' deposited much earlier compared to the other rivers
(no 3, 4, 5 and 6). At this time, we can also observe a
Channels that are observed from Sepat Field are fault that cuts through river 2 (Figure 8c). As we go
mainly meandering river with different sinuosity level. deeper (time -729ms (Figure 8d)) only “hook” shape
Few straight channels are also observed. Meandering incise valley can be seen, proving that it is a deep
channels are mostly found in lower coastal plain valley.
and have fining sequence upwards. Point bars are
tracked in the seismic attributes and were form when
Figure 8 Different depth of time slices showing the orientation of the rivers found at the depths. The
shallowest section in (A) at -225ms shows five different rivers flow, but as the time goes deeper to (B) – 402ms,
the rivers labelled as no 3, 4 and 5 cannot be seen on the time slices. This means these rivers form only after
the deposition of sediments at -402ms. Fault dividing the river can be seen at the time (C) -606ms and just
point bar (meandering river) no 1 can be observed in the deeper section at (D) -729ms
Tributaries
Channel
Indication of deep
incised valley
5 km
Figure 9 Time slice at -108ms twt shows a highly sinuous channel, which is flowing from north to south. It
shows the characteristic of a deeply incised valley and can be tracked up till time slice -1400 twt. Point bars are
well developed that gives bright amplitudes
Fluvial System of Sepat Field in relation to Malay With the access to the map of river system of Malay
Basin Fluvial System Basin produced from the study of [7] and [6], the
position of Sepat Field is located next to the Noring
The main orientation of the channels observed in Oil Field (Figure 10). Since the major trunk river flows
Sepat Field in from northwest to southeast and from the northwest to southeast, the “hook” features
west to east. This direction is relatively similar to found in the meandering river of Sepat Field is has
the flow direction of the Chao Phraya-Johore River high possibility to be part of the developed point
that flows from high ground in Thailand, which then bars along the major river system of the South China
passes through the axial of the Malay Basin [7]. Our Sea (Figure 10). The fluvial system in Sepat Field is
interpretation relates this major “hook” (Figure 8a-8c) noted to be part of the larger fluvial system of Malay
to be a part of the significant trunk channel of the Basin and falls in the midstream. It is located within
Late Pleistocene rivers [7]. However, since it is located the transitional zone between the high grounds
at the edge of the seismic cube, the continuation of (upstream) in the north (Thailand) and the marine
the valley could not be tracked further. environment in the downstream (South China Sea) as
presented in Figure 11.
Figure 10 (A) shows the merged 3D dataset of the Malay Basin with the location of Sepat Field from the
study of [7]. The two blue lines in (A) represent the valley margins that bound the large meandering channel
observed in research conducted by [7]. The red box in (A) is later interpreted as shown in (B) that focuses the
connectivity of the trunk “hook” channel from Sepat field to the rest of the data, and it shows good continuity
of the Chao Phraya – Johore River
Figure 11 Position of Sepat Field within the Malay Basin fluvial system, relative to the interpretation by [7]
CONCLUSION the time slices. Sepat Field and the whole Malay Basin
fluvial system developed as midstream rivers in the
The characterisation of the fluvial channel in context the upper stream is in the northern side and
Sepat Field was achieved by applying geophysical the downstream is the marine system of the South
technique mainly focusing on attributes application. China Sea.
The application of seismic attributes in characterising
stratigraphic features has shown its effectivity
following the presence of the gas cloud. If seismic ACKNOWLEDGEMENT
attributes were not applied, the characterisation of
channels would have been tough and incomplete. We want to thank PETRONAS’s Malaysian Petroleum
The interpretation that was done using variance Management (MPM) for the permission to use
attributes was compared with frequency spectral the subsurface data for this study, Schlumberger
decomposition that enhanced the interpretation. for providing the academic license for Petrel, an
The utilisation of colours blending also improved the interpretation software and Foster Findlay Associate
interpretation in time-slices. (FFA) for providing the GeoTeric® license.
[3] M. Madon, P. Abolins, M. J. Hoesni, and M. B. Ahmad, [12] D. Subrahmanyam, “Seismic Attribute - A Review,”
“Malay basin,” U. S. Geological Survey, Wyoming, WY, presented at the 7th International Conference &
Open-File Report 99-50T, 2002. Exposition on Petroleum Geophysics, Hyderabad,
India, 2008.
[4] S. Schumm, “River variability and complexity,”
Cambridge University Press, Cambridge, 2005. [13] D.U. Kr, D.C. Pan, G. Parida, R. Rakesh, B. Ashutosh,
“Application of Multi – Attributes and Spectral
[5] A.H. Abdul Latiff, D.P. Ghosh, and S.N.F. Jamaludin, Decomposition with RGB Blending for understanding
“Structural Enhancement in Shallow Gas Cloud the strati – structural features: A Case Study,”
Region” IOP Conference Series: Earth Environmental presented at 10th Biennial Internation Conference &
Science, vol. 23, 012005, Oct. 2014. Exposition,Kerela, India, 2013.