Delineation of thin bedded channel sand geometry by taking advantage of 90 degree phase conversion

and stratal slicing of 3D seismic Data: A case Study from Cambay Basin Area, India

Bidyut Mandal 1 and Rima Chatterjee 2

1
Gujarat State Petroleum Corporation Limited, Gandhinagar 382011, Gujarat, India
2
Dept. of Applied Geophysics, Indian School of Mines, Dhanbad-826004, India

E-mail: bidyutmandal1983@gmail.com and rima_c_99@yahoo.com

ABSTRACT

With the development of high quality 3D seismic, the exploration targets in Cambay Basin are being shifted
from the structural traps to stratigraphic traps. During recent years, some discoveries in thin bedded Middle
Eocene pay, indicates there is wide opportunity of further exploration if we can precisely locate thin bed
reservoir geometry which was over looked earlier in normal seismic amplitude data. In Seismic profiles channel
sands can be identified as “bright spot” with the low frequency, high amplitude and short reflection, however in
case of coal masking thin bedded reservoir it is very difficult to identify and some time misleading to interpreter
also. Unfortunately, although many reservoir-scale (well-to-well scale) features can be detected in vertical
seismic lines, few such features can be resolved and interpreted in the vertical perspective because of the data’s
limited bandwidth.

A case study from Cambay Basin Area provides an example that requires special attention for slicing, phase
character considerations for optimal depositional imaging. Being typical thin bedded reservoir type in nature,
thickness is below the resolution limit and hard to distinguish from the amplitude background, are difficult to
identify in vertical section of 3D Seismic. In inter-bedded sandstone, siltstone, shale and coal succession of
Lower Eocene Kalol Formation, a 90° phase shift of nearly zero-phase seismic data significantly improves litho-
logic and stratigraphic interpretation by putting the main lobe of the reflected wave in the centre of the thin
strata through rotating the seismic phase 90°. A match between seismic and synthetic profiles of phase shift
volume shows a better tie between seismic amplitude traces and lithology-indicative logs. Better geometric
imaging of amplitude units which was very difficult to identify in dual polarity seismic events. Combining the
stratal slice, cue from limited well data, we identified one thin bedded channel geometry with high amplitude
value. In contrast to time slices, which are created from the processed 3-D data volume at a constant time level,
stratal slices are created using the interpretation of key regional or local seismic reflectors. According to our
study, stratal slice mapping of phase shift data can delineate thin bedded channel sand geometry, thereby leading
to more optimal positioning of development and exploration well locations with minimized geological risk.

Keywords:
Thin bed, 90° Phase rotation, Stratal Slicing
1. INTRODUCTION

Cambay Basin, explored since 1950s, is one of the most prolific hydrocarbon basins situated in western part of
India. Exploration in Cambay Basin is fairly mature, most of the field-scale geologic and depositional prospects
based on structural component has already probed and substantial amount of hydrocarbon resources have
already been converted into initial in-place volume of hydrocarbons (S.Gupta 2006). With the development of
high quality 3D seismic, the exploration targets are being shifted from field scale to reservoir scale, industry is
turning its attention to reinterpretation of thin bedded stratigrahic features, which was over looked earlier.

3D Seismic Data (the two horizontal distance co-ordinates and the vertical co-ordinates that define the 3D
volume) can be sliced along three orthogonal planes. The vertical plane through a seismic line produces an
inline vertical section, the vertical line through a seismic cross line produces a cross line vertical section and the
horizontal plane for a given vertical coordinates produces a horizontal section, which is termed as time slice or a
depth slice based on vertical component parameter (Fig.1). A horizontal slice is made up of the 2D seismic data
lying on a flat level plane that cuts through a volume of 3D seismic data. If the vertical component is time, the
horizontal slice is time slice. In other words, a time slice is a display of the seismic amplitudes for a single time
arrival. Picking reflection on a time slice is equivalent to mapping a time contour, and successive contours can
be mapped directly by following the same event on successive time slices. This result is a structural contour
map. A horizon slice is a display from a 3D data set of the data elements that all lie on the same reflecting
horizon. Such a display brings out variations in amplitude (or some other attribute) on the horizon in question. A
horizon slice is different from a horizontal slice (or time slice). As Sheriff and Geldart (1995) point out, horizon
slice are useful in the study of stratigraphy and reservoir properties.

Fig1: Data cube, time slice, inline section, cross line section (modified after Robinson and Treitel)

Traditionally, interpreters have analyzed vertical sections of 3-D seismic volumes line by line and found field-
scale geologic and depositional features. Sometimes, reservoir-scale features can be detected in these vertical
sections but many of these small-scale targets cannot be resolved and interpreted because of data bandwidth
limitations (Zeng, 2006). Generally with increased depth, seismic frequency decreases and seismic velocity
increases, leading to reduced resolution. Seismic signal-to-noise ratio typically deteriorates with depth as well,
and interpretation is further hampered at depth if the unit of interest is thin and below seismic resolution. In
addition, because of the reflection interferences and wavelet overlap, it is almost impossible to identify each thin
bed from the composite response of inter-bedded thin beds (Zeng, 2001). However, the seismic response of one
thin beds does not always interfere with other thin beds at all times. Auto tracing can also be a poor choice for
thinly inter-bedded sandstone/shale formations, which are less continuous and contain few strong seismic
anomalies. In these situations, horizontal slices are commonly applied to produce more desirable results. Time
slice and horizon slice are currently the most commonly used horizontal slice to extract stratigraphic
information. However, problems exist in the use of time or horizon slices for imaging of geologic time surfaces.
For depositional-facies analysis, both methods have limitations. Time slices can image a time surface only if
strata are flat-lying and sheet like. Horizon slices work well in a formation with structural dip, but the formation
still needs to be sheet like. Stratal slices (modified method of Horizontal slice) eliminate the conditions by
taking into account thickness gradient so that sediment wedges can be properly sliced (Zeng, 2001).

Standard seismic processing commonly produces zero-degree phase seismic data as the final products delivered
to interpreters. Interpretive advantages of zero-degree phase data include wavelet symmetry, centre lobe
(maximum amplitude) coinciding with reflection interface, and higher resolution. However, those advantages of
the zero-degree phase data are true only if the seismic reflection comes from a single reflection interface (e.g.,
the sea floor, a major unconformity, the top of a thick and blocky sandstone etc.). If the seismic reflection comes
from the thin beds, seismic event does not coincide with the sandstone bed directly. The aim of phase transform
is to provide seismic event with strata sense. The rotating degree depends on phase spectrum and phase of
interest strata in seismic data calibrated by well logging. When stratal thickness is close to half of wavelength,
zero phase data after processing just needs 90° rotation to ensure seismic section to correspond with lithology of
thin interbed strata. Generally, before interpretation, phase of seismic data is not definitive rather than just zero.
Therefore, it is required to analyze wavelet phase, and decide how much degree to rotate. For smaller thin
interbed such as a fourth order sequence interface, firstly it is needed to calibrate well logging and seismic, then
adjust phase of seismic data to ensure the major part of main waveform lobe to coincide with thin strata, so that
provides seismic event with geologic sense.

A case history from Lower Eocene Kalol Formation of Cambay Basin Area provides an example that requires
special attention for slicing, phase character considerations for optimal depositional imaging. Being typical thin
bedded reservoir type in nature, thickness is below the resolution limit and hard to distinguish from the
amplitude background, are difficult to identify in vertical section of 3D Seismic.In an inter-bedded sandstone,
siltstone, shale and coal succession in the Lower Eocene Formation of Cambay Basin area, a 90°phase shift of
nearly zero-phase seismic data significantly improves litho logic and stratigraphic interpretation by putting the
main lobe of the reflected wave in the centre of the thin strata through rotating the seismic phase 90°.A match
between seismic and synthetic profiles of phase shift volume shows a better tie between seismic amplitude
traces and lithology-indicative logs. Combining the stratal slice, cue from limited well data with the sedimentary
environment, one NNW-SSE channels with high amplitude value has been identified in our study area.Here we
shown stratal slice can delineate stratigraphic variations within a slab of the subsurface that is one-quarter
wavelength thick — the resolution of the seismic data.
2. GEOLOGY OF THE STUDY AREA

Cambay Basin has been characterized as a narrow elongated intra-cratonic rift-graben surrounded by Saurashtra
Uplift in the west, Aravalli-Delhi Fold Belt in north-east, and Deccan Craton in the south-east. The major
tectonic trend is roughly N-S, extending from Sanchor in the north to Gulf of Cambay in the south and further
opening up into Arabian Sea. The basin is about 425 km long and having width varying between 40 and 80 km.
About 5 to 7 km of sedimentary thickness is envisaged in the basin (Pandey et al., 1992). The major trend of the
lineament in North Cambay Basin is NNW-SSE and NE-SW. In the southern part, however, the main trend is
ENE-WSW (Chandra et al., 1969). The entire basin is divided into five tectonic blocks based on transverse fault
systems namely, Narmada–Tapti, Jambusar–Broach in south Cambay basin and Tarapur–Cambay, Ahmedabad–
Mehsana as well as Sanchor–Patan in north Cambay basin (Raju et al.,1971; Biswas, 1982, 1987).The study
area, very near to well known Sanand field, in Ahmedbad-Mehsana Tectonic Block(Fig.2)

Fig.2: Location Map of Study Area

The lithostratigraphy of different blocks of the Cambay basin (Fig.3) has been described by various authors
(Bhandari and Chowdhary,1975; Biswas, 1982, 1987; Pandey et al., 1993;Kundu et al., 1997). In this
Ahmedabad–Mehsana tectonic block Deccan Trap forms the basement which is overlain unconformably by
Olpad Formation. It is further subdivided into two units designated as Older and Younger Cambay Shale
Formations. The Younger Cambay Shale has an unconformable relationship with the overlying Kalol
Formation. The Kadi Formation restricted to this tectonic block is the stratigraphic equivalent of the Younger
Cambay shale. The Kalol Formation has a conformable relationship with the overlain Tarapur Formation. The
Kalol Formation has been divided into four members. The upper three (Sertha,Kansari, and Wavel) were
deposited in an extensive area of this tectonic block and the lowermost Chhatral Member is restricted areally
(Bhandari and Chowdhary, 1975). The Olpad Formation was deposited over basaltic basement during the early
Paleocene under fluvial to shallow marine conditions. A major marine transgression from the southwest
deposited gray to black Cambay Shale during the late Paleocene. The sediments of Kalol Formation were laid
down in paralic environment under unstable shelf conditions with alternating transgression and regression
(Kumar et al., 2004). The Kalol Formation of the Middle Eocene age lies between Tarapur Shale and Younger
Cambay Shale. Thickness of Kalol Formation varies from 150 to 300 m. The major lithological units are
sandstone, shale, carbonaceous Shale and coal. The thickness of sandstone varies from 2 to 50 m (Das et al.,
2006).Kalol Formation consists of alternations of fine to medium grained sandstones, siltstones, silty shale and
coals. Facies map had been prepared by the previous author Sarraf et al. (2008) with three components of
lithofacies namely sand/silt, shale and coal. The entrapment style is strati-structural for Kalol pays. Tidal
channel, tidal flat and locally swampy environments were interpreted for the K-IV unit (Bhandari and
Chowdhary, 1975; Sarraf et al., 2008). The main reservoir rock is silt/sand of K-IV unit in the present study,
named as Eocene pay IV (EP-IV) sand of Kalol Formation.

Fig.3 Generalised Stratigraphy of Cambay Basin Area (modified after ONGC)

3. Methodology

3.1 Phase Rotation

Seismic data can be fragmented into components such as energy, frequency, amplitude and phase. It should be
remembered that a seismic trace is composed of only the real part of the complex domain. However, when we
speak of an attribute, we include the imaginary part as well. Tanner and Sheriff (1976) introduced the concept of
complex trace attributes for seismic signal analysis, and this bring us to an analytic signal which is
mathematically represented as

X (t)=f (t)+iH {f (t)}

f (t)is a real valued signal that is being transformed into complex analytical signal; and the operator H {f (t)}
represents the Hilbert transform given by:
 +∞
1 f ( s)
H {f (t)}=( ) P ∫ ❑ ds
π −∞ t−s

Where s is a variable of integration, P is Cauchy principal value and i is √ −¿ 1 ¿. For simplicity we will let:

g(t )=H {f (t)}

We can then redefine the analytic signal as

X (t)=f (t)+ig (t)

The Hilbert Transform filter, when applied to the complex analytic signal, introduces phase shift of
−π /2 at each positive frequency and + π /2 at each negative frequency, and this is in turn filters out all
negative frequencies.

In the frequency domain the seismic trace, (t) , can be expressed as:

X (ɷ)=R( ɷ)+iI (ɷ )

Where R(ɷ) is the real component and I (ɷ) is the imaginary component of the complex frequency signal.
Computing signals in Cartesian system is sometimes difficult and complex. To avoid the complication, Euler’s
method is commonly used and analytical seismic trace in frequency domain and can be expressed as.

t ɸ(ɷ)
X (ɷ)=¿ A ɷ∨e

The phase shift angle ɵ is applied to the entire trace using

t ɸ(ɷ+ ɵ)
X (ɷ)=¿ A ɷ∨e

An alternative and quite useful representation of the frequency signal is to express the coefficient as a magnitude
and phase angle (Fig.4).

Where the magnitude A(ɷ) is defined as:

A(ɷ)=√ R(ɷ )2+ I (ɷ)2

The Phase angle is computed as:

−1
ɸ (ɷ)=tan [ I (ɷ)/ R(ɷ )]
 Fig.4:Representation of frequency signal as a magnitude and phase angle

A phase rotation of 180° will produce a reverse polarity version of the input trace, while a + 90° or -90° rotation
will alter the seismic trace in such a way that peak or trough on the input trace will become a zero-crossing on
the output trace, and vice-versa (Fig, 5.2).

Standard seismic processing commonly produces zero-degree phase seismic data as the final products delivered
to interpreters. A zero-phase seismic trace is symmetrical to a single reflection surface, with the maximum
amplitude approximating impedance contrast. Interpretive advantages of zero-degree phase data include wavelet
symmetry, centre lobe (maximum amplitude) coinciding with reflection interface, and higher resolution.
However, those advantages of the zero-degree phase data are true only if the seismic reflection comes from a
single reflection interface (e.g., the sea floor, a major unconformity, the top of a thick and blocky sandstone
etc.). Zero phase data are ideal for lithofacies identification on an unconformity or in a thick bed. However, for
seismically thin depositional units, amplitude traces from zero-phase data become asymmetrical and are more
difficult to tie to lithology-indicative wireline logs. The simplest and most effective way to improve interpretive
value of amplitudes is to apply a 90° phase shift to the zero-phase data so that seismic traces are converted from
reflectivity series to relative impedance series (Sicking 1982; Zeng & Backus 2005a, b). In 90 °data maximum
amplitude correlates to the centre of a thin bed. A valuable by-product of this process is that stratigraphic
resolution of seismic data is improved.

For simplification, Phase control of seismic waveform, corresponding to a geological object, as illustrated in
Fig.5.1 (modified after Zeng) described here. In case of Zero degree phase wavelet, seismic reflection is
symmetrical to a single surface (Fig.5.1.a), but is antisymmetrical to a seismically thin bed (Fig.5.1.c). In
contrast, a 90 ° phase wavelet results in a reflection that is antisymmetrical to a single surface (Fig.5.1.b), but
symmetrical to a thin bed (Fig.5.1.d).
 Fig.5.1.Phase control of seismic waveform, Fig:5.2 Original data (left) and Phase shift attribute (right)

corresponding to a geological object

The 90° transition of the phase position of the seismic data is equal to making simple integral of the seismic
data. Thus, the events of the seismic reflection correspond to the geological strata and the seismic amplitude has
some lithological stratal meaning. A rotation of seismic wavelet phase in thin-bedded section changes the micro
geometry of, and relationships among, seismic events, potentially leading to different interpretations of seismic
stratigraphy, lithofacies, and geomorphology. As a result, phase conditioning of seismic data can be critical for
reservoir-scale stratigraphic analysis and prediction.

3.2. Slicing Methodology

Unfortunately, although many reservoir-scale (well-to-well scale) features can be detected in vertical seismic
lines, however few such features can be resolved and interpreted in the vertical perspective, because of the
data’s limited bandwidth. Only in the horizontal perspective are such depositional features large enough to be
resolvable when displayed in map view on geologic time surfaces. Outcrop and subsurface studies show
depositional bodies have horizontal dimensions greater than their vertical dimensions. As a result, small
depositional bodies often can be resolved in plan view even if they can only be detected in vertical view (Zeng,
2006). There are many different ways to view a 3D seismic volume in a horizontal section. To implement
horizontal-view seismic interpretation, we must pick geologic-time surfaces (or stratal surfaces) from 3-D
seismic volumes so that seismic attribute maps across these fixed-geologic-time surfaces can be analyzed in
terms of depositional systems. Time slices (the earliest form of horizontal seismic display - dating from the
1970s), extracted from a data volume at a constant image-time coordinate, and is sufficient for sheet-like and
flat-lying formations. Horizon slicing (developed in the 1980s), Constructed by extracting a seismic attribute
parallel to a picked time-varying horizon, work well in a formation with structural dip , but the formation still
needs to be sheet like.
 Fig.6: Distinctions among time slices, horizon slices, and stratal slices (modified after Zeng)

Time slices (Fig.6.a) and horizon slices (Fig.6.b) are currently the most commonly used seismic-surface displays
to extract stratigraphic information. However, for depositional facies analysis (many depositional sequences are
characterized by thickness changes) both methods have limitations. Their strict application conditions have
prevented them from being used in complicated (e.g.,wedged) seismic sequences. Zeng (1994) and Zeng et al.
(1998 a,b) proposed the method of stratal slicing , where an extracted surface follows a fixed-geologic-time
surface and useful for high-resolution seismic imaging of thin depositional systems (Fig. 6.c). Posamentier et al.
(1996) offered a similar approach, known as proportional slicing. Stratal slicing improves seismic-surface
display mainly by making slices proportional between geologic time-equivalent seismic-reference events.
To implement horizontal-view seismic interpretation, we must pick geologic-time surfaces (or stratal surfaces)
from 3-D seismic volumes so that seismic attribute maps across these fixed-geologic-time surfaces can be
analyzed in terms of depositional systems. Sometimes, in case of strati-structural reservoir a horizon slice can
depict a part of thin bedded channel geometry in an area, where strata are almost sheet like and without faulting.
However, limitations occur due to thickness change across the area. Since Horizon tracking is time consuming;
only the key horizons are typically tracked .Therefore, only a small part of the available data is usually analyzed.
To analyze more data, horizons are often shifted, which is in most cases only valid in the vicinity of the original
horizon. A stratal slice, proportionally between two references horizon, enable a more accurate analysis and
useful for high-resolution seismic imaging of thin depositional systems (Fig.7). Selecting the right slicing
method for extracting geomorphic patterns properly from the 3D seismic data is therefore crucial.

Fig:7. Schematic representation of Horizon slicing and Stratal slicing ( modified after Opendtect Software manual)
4. A Field-data example from Cambay Basin Area

Lower Eocene Kalol Formation consists of a combination of thinly inter-bedded sandstone, siltstone, shale and
coal. Reservoir intervals can be categorized regionally and locally into numerous corelatable units. Kalol pays
generally show strati-structural entrapment hence, the reservoir quality shows dependency on facies distribution.
A case history from Lower Eocene Kalol Formation of Cambay Basin Area provides an example that requires
special attention for slicing, phase character considerations for optimal depositional imaging. Being typical thin
bedded reservoir type in nature, thickness is below the resolution limit and hard to distinguish from the
amplitude background, are difficult to identify in vertical section of 3D Seismic.
Thin bedded reservoir zone of Lower Eocene Kalol Formation in our study area were tested as Hydrocarbon
bearing in Well A1, A2 and A3, however limited well control limits the heterogeneous reservoir geometry and
drilling risk of new location for further development of discovery area. To reach the project goal of finding
additional locations between and beyond wells, an advanced amplitude extraction method , based on stratal
slicing method was carried out on phase shift volume, so that thin bedded reservoir geometry can be
extrapolated to the entire area, throughout the section of interest. An important step is to optimize depositional-
facies mapping by automating depositional-surface picking in the 3-D volume and to link seismic-attribute
patterns directly to sedimentary rocks and depositional processes.

On well-logging, the thin bedded channel sand shows the characteristics of box curve and positive cycle with
bell-shaped. The channel sand body in the area is characterized by high porosity, low natural potential, low
natural gamma and low density (Fig.8). Wells in study area (Well A1, A2 and A3) penetrated fluvial sandstones
ranging from 3 to 5 m reservoir thickness. In the 29 Hz dominant frequency data, many of these sandstones are
below seismic resolution and most of them are below tuning. Most commonly they are only marginally
detectable in vertical section without resolving the tops and bases of the sand bodies, making a seismic facies
analysis challenging. Seismically, it is difficult to resolve the reservoir as they are beyond seismic resolution and
masked by underlying coal layers.

In an inter-bedded sandstone, siltstone, shale and coal succession in the Lower Eocene Formation of Cambay
Basin area, a 90°phase shift of nearly zero-phase seismic data significantly improves litho logic and
stratigraphic interpretation by putting the main lobe of the reflected wave in the centre of the thin strata through
rotating the seismic phase 90°. A match between seismic and synthetic profiles of phase shift volume shows a
better tie between seismic amplitude traces and lithology-indicative logs. Combining the stratal slice, cue from
limited well data with the sedimentary environment, a thin bedded channel geometry (which is very difficult to
identified in vertical section of 3D seismic data) has been delineated in study area. Here we shown stratal slice
can delineate stratigraphic variations within a slab of the subsurface that is one-quarter wavelength thick — the
resolution of the seismic data.

As for thin bedded reservoir, the identification and prediction of the heterogeneous channel-sand geometry is the
key to enhance drilling success rate and reduce the investment risk. According to our study, stratal slice
mapping of phase shift data can delineate thin bedded channel sand geometry, thereby leading to more optimal
positioning of development and exploration well locations with minimized geological risk.
 Fig:8. Correlation of Drilled Wells (A1,A2 and A3) in Study area

5. Results and Discussion

5.1. Estimating Vertical Resolution from Seismic

For practical purpose the tuning thickness can be considered as an indication of the vertical resolution.
Fundamentally, the tuning thickness is determined by the compressional velocity of the unit and the wavelength
of the seismic pulse (ʎ). A first order approximation towards making quick calculations for our study area:

ʎ
Tuning thickness =
(Widess, 1973) (1)
4
ʎ =V P (m/ s)/ F d ¿) (2)
F d=1 /T (3)
Where, ʎ =wavelength (m), F d = dominant frequency and T = ‘period’ (measured in seconds from trough to
trough or peak to peak). Seismic section of study area along with Amplitude Spectrum is illustrated in Fig. 9.
 Fig:9. Seismic Data Spectrum Analysis Diagram

A worked example based on our study area shows,

F d = 29 Hz
T=0.034 s
ʎ =80 m (if V P= 2320 m/s)
ʎ
Tuning thickness ( ) = 20 m
4

5.2 Well to Seismic Tie

Synthetic seismograms were generated for wells to link logs (in depth domain) to time domain seismic data and
to observe the seismic character of sands within the area. The synthetic seismograms were created by using the
extracted Wavelet at well locations. Well to seismic ties were performed by establishing correlation between the
seismic and synthetic seismograms by adjusting T-D functions (Fig.10). A match between seismic and synthetic
profiles of phase shift volume shows a better tie between seismic amplitude traces and lithology-indicative logs.
Better geometric imaging of amplitude units which was very difficult to identify in dual polarity seismic
events.The Gamma Ray (GR) curve is displayed for the reference of litho logical variations. Out of three wells
in study area only one well (Well A1) is vertical. Special concentration to tie the reservoir top between well and
seismic was attempted for A1 well .As per well log data, the thin bedded reservoir in tested wells (A1, A2 and
A3) is lying just above a thin coal marker, hence special concentration and Quality check has been carried out to
achieve a consistent tie in each coal layer between the wells and the seismic volume.

Typically sandstones have a low-AI and shale profile have high –AI. In an inter-bedded sandstone-shale-Coal
Eocene succession in the Cambay Basin area, well log data show coal layer has low AI as compare to sandstone
and shale , however sandstone have low acoustic impedance (AI) to shale or high porosity zone relative to low
porosity. Three key horizons H1, H2 and H3 namely Kalol Top, Coal Top and Coal Bottom were mapped
throughout the 3D post stack migrated seismic data volume and interpolated to obtain continuous horizon
surfaces for scanning of various seismic attributes. Well log analysis shows besides main coal layer (Coal Top
and Bottom) there is another prominent thin coal marker, near the reservoir zone of tested wells (A1, A2 and
A3).
 Fig:10. Well to Seismic Tie of A1 Well

Seismically, the reservoir zone is difficult to resolve as it masked by underlying coal layers and they are beyond
seismic resolution. When applying 90° phase rotation of this seismic data, then it can be found that the main
lobe of strong amplitude coincides with coal and sandstone detected by GR and Density log. Sandstone bodies
do not have simple relationship with seismic events. A match between seismic and synthetic profiles of phase
shift volume shows a better tie between seismic amplitude traces and litho logy-indicative logs. Based on well to
seismic relation, horizon H4 namely Coal marker was also attempt to mark, as a reference horizon of nearby
reservoir (Fig.10 and 11)

Fig:11. Correlation of Well tops and Interpreted Horizon

5.3. 90° phase shift of 3D Seismic Data

Phase transform is a technology that makes seismic waveform take a phase rotation to make sure reflection
event (wave peak or trough) corresponds with strata of thin interbed rather than with its top or bottom interface.
That endows seismic reflection event with lithostratigraphic sense. Although seismic data resolution is not
improved, the seismic profile is transformed into a superficially-geologic section similar with impedance one,
which makes strata interpretation more imaginable and more intuitive.

The aim of phase transform is to provide seismic event with strata sense. The rotating degree depends on phase
spectrum and phase of interest strata in seismic data calibrated by well logging. When stratal thickness is close
to half of wavelength, zero phase data after processing just needs 90°rotation to ensure seismic section to
correspond with lithology of thin interbed strata. Generally, before interpretation, phase of seismic data is not
definitive rather than just zero. Therefore, it is required to analyze wavelet phase, and decide how much degree
to rotate. For smaller thin interbed such as a fourth order sequence interface, firstly it is needed to calibrate well
logging and seismic, then adjust phase of seismic data to ensure the major part of main waveform lobe to
coincide with thin strata, so that endows seismic event with geologic sense (Mingchen,2009)

We examine 3D seismic data of Study area (Cambay Basin Area) to test the benefits of 90°-phase wavelets in
thin-bed interpretation. In an interbedded sandstone-shale-Coal Eocene succession in the Cambay Basin, a 90°
phase shift of nearly zero-phase seismic data significantly improves lithologic and stratigraphic interpretation. A
match between seismic and acoustic impedance (AI) profiles results in a better tie between seismic amplitude
traces and lithology-indicative logs.

After calibrating normal seismic data with well A1, it has been observed, oil-bearing sandstone centre is
consistent with the position of zero point, which means that the dislocation exists between the strong amplitude
detected in seismic profile and oil-bearing sandstone in study area. In order to get intuitive relationship between
sandstone and seismic reflection event, it is necessary to make a phase transformation of seismic data. After
application of 90°phase rotation, it can be found that the main lobe of strong amplitude trough perfectly
coincides with sandstone detected by GR (Fig.12). Therefore, geologic sense of strong amplitude is confirmed,
which is favour of sand-body tracking and describing. A match between seismic and synthetic profiles of phase
shift volume shows a better tie between seismic amplitude traces and lithology-indicative logs. Better geometric
imaging of amplitude units which was very difficult to identify in dual polarity seismic events.

Fig:12. Phase transformation of Seismic Data and its correlation between amplitude traces and log signature.

5.4. Application of Stratal Slicing Method

For thin bedded reservoir, where thickness is below the resolution limit and hard to distinguish from the
amplitude background, are difficult to identify in vertical section of 3D Seismic. One strategy to map
depositional systems with high resolution is to change the emphasis of seismic interpretation from vertical
sections to horizontal sections. In these situations, horizontal slices (time slices, horizon slices, stratal slices,
etc.) are commonly applied to produce more desirable results (Fig.14). Time slices and horizon slices are
currently the most commonly used seismic-surface displays to extract stratigraphic information. However for
depositional strata, where thickness changes occur, both methods show their limitations. A refinement of this
idea is to use “stratal slices” (Zeng et al., 2001).This kind of slice is obtained from a series of interfaces
proportionally interpolated between two interpreted isochronal interfaces.

In order to analyze thin bedded reservoir geometry in Lower Eocene Kalol Formation, conventional slicing
method was first applied to depict the channel geometry. As per our study we observe Stratal Slices produces
more desirable result, specially in case of thin bedded reservoir, where reservoir thickness is below the
resolution limit and are difficult to identify in vertical section of 3D Seismic. For depositional facies analysis,
Time slice and Horizon slice can be treated as special cases of stratal slicing when a flat lying or sheet lying
formation is encountered.

Fig:13. Representation of Stratal slice on Seismic Section (with reference to Coal Top and to Kalol top )

A time slice, sampling the thickest part of the reservoir sand at TWT 1152 ms, unable to depict the reservoir
geometry in most of the study area because of significant structural dip and faulting. A horizon slice, made 24
ms above reference Coal Top, depicted part of channel geometry in an area where strata are almost sheet like
and without faulting, limitations occur due to thickness change across the area. As the thickness changes occur
between two isochronal surfaces, a diachronous phenomenon is observed (Fig.13). So, the slice is unable to
reflect the planar variation of sand-body. However, stratal slice is proportionally interpolated between two
interfaces according to thickness ,it considers sedimentary space variation with position and decreases
error ,therefore, it is more reasonable than time slice and horizon slice. Stratal slicing improves seismic-surface
display mainly by making slices linearly between geologic time-equivalent seismic-reference events. A stratal
slice picked proportionally between two references horizon, i.e. Kalol Top and Coal Top (as well log analysis
indicates the reservoir zone is in between kalol and coal Top) depicted clear channel geometry throughout the
study area (Fig.10, 13, 15.1 and 15.2).
 Fig:14. Comparison of various slicing methods (Time Slice, Horizon Slice and Stratal Slices) for study area

Based on well to seismic correlation 19 stratal slice has been generated (roughly 3ms interval each) in between
Kalol Top and Coal Top. The slices are highlighted on the well to seismic section (Fig.13) to show position of
representative stratal slices in two-way time and their relationship with seismic events, and well log
characteristics. The thin bedded channel sand body in the area is characterized by low natural potential, low
natural gamma and low density. Sliced based amplitude anomaly (RMS Amplitude and Extracted Value
Amplitude), predict a better channel geometry. The highlighted slices demonstrate, out of 30 slice, only 2 slices
(Slice no7 and 8 ) are coincide with reservoir trough as identified from well to seismic relation, depicted a
better view of channel geometry (Fig.13, 15.1 and 15.2).

Fig: 15.1. Stratal Sliced based Extracted value Amplitude (with reference to Coal Top and Kalol top )
 Fig:15.2. Stratal Sliced based RMS Amplitude (with reference to Coal Top and Kalol top )

The selection of the time window is also a key factor of extracting seismic attributes. If the selected time
window is too small, the slice which is used to extract seismic attributes will lose useful geological information.
Conversely, if the time window is too large, the slice will include too much unnecessary geological information.
As the reservoir is very thin (3m-5 m) and underlain by a thin coal layer, seismically it is very difficult to
resolve, however based on well to seismic tie, underlain thin coal layer can be predicted and marked in 90
degree phase shift seismic data. Therefore, it is more reasonable to choose Coal Marker and Kalol top as the
upper and lower limits of the time window in order to extract seismic attributes. For illustration, Fig.16 is the
slice selecting the Kalol Top and Coal Marker as top and bottom boundary to extract the slice based amplitude
attributes. Out of 13 slice, slice no 2 (6 ms above coal marker), coincide with reservoir trough as predicted from
well to seismic relation, depicted a better view of channel geometry in our study area.

Fig16. Representation of Stratal slice on Seismic Section ( with reference to Coal Marker and Kalol Top)
Fig:17.1. Stratal Sliced based Extracted value Amplitude (with reference to Coal Marker and Kalol top )

Fig:17.2. Stratal Sliced based RMS Amplitude (with reference to Coal Marker and Kalol top )

A set of amplitude stratal slices in the formation (Fig. 15.1, 15.2, 17.1 and 17.2) reveals the presence of a
meandering fluvial system. Several tests were done to ensure that they are favourable seismic images for
delineate thin bedded reservoir geometry in study area.

Conclusion
Standard zero-phase seismic data are less desirable for litho logic interpretation of thin bed. A rotation of
seismic wavelet phase in thin-bedded section changes the micro geometry and relationships among seismic
events. As a result, phase conditioning of seismic data can be critical for thin bedded reservoir-scale analysis
and prediction.

Stratal Slices (a series of interfaces proportionally interpolated between two interpreted isochronal interfaces)
produces more desirable result, where reservoir thickness is below the resolution limit and are difficult to
identify in vertical section of 3D Seismic. For depositional facies analysis, Time slice and Horizon slice can be
treated as special cases of stratal slicing when a flat lying or sheet lying formation is encountered.

As for thin bedded reservoir, the identification and prediction of the heterogeneous channel-sand geometry is the
key to enhance drilling success rate and reduce the investment risk. According to our study, stratal slice
mapping of phase shift data can delineate thin bedded channel sand geometry, thereby leading to more optimal
positioning of development and exploration well locations with minimized geological risk.

Acknowledgments
We gratefully acknowledge the Gujarat State Petroleum Corporation Limited, Gandhinagar, Gujarat, regarding
the various data support and analysis. We also acknowledge M/s Schlumberger for using Petrel Seismic to
Simulation Software for stratal slicing attribute and HLS Asia Ltd for using DSD application towards well to
seismic tie.

References:

Banerjee, A., Jha, M., Mittal, A.K., Thomas, N.J., Misra, K.N., 2000. The effective source rocks in the north Cambay basin.
Mar. Pet. Geol. 17, 1111–1129.
Bhandari, L.L., Chowdhary, L.R., 1975. Stratigraphic analysis of Kalol and Kadi formations. Am. Assoc. Pet. Geol. Bull. 59,
856–871.
Biswas, S.K., 1982. Rift basins in the western margin of India and their hydrocarbon prospects. Am. Assoc. Pet. Geol. Bull.
66 (10), 1497–1513.
Biswas, S.K., 1987. Regional tectonic framework, structure, evolution of the western marginal basins of India.
Tectonophysics 135, 307–327.
Chopra S, Marfurt K J. Seismic attributes — a historical perspective [J]. SEG 75th Anniversary,Geophysics, 2005, 70: 350-
2850.
Das, S., Singh, H., Tiwari, D., Parulkar, S.N., 2006. Reservoir classification and geological remodeling of Kalol Sands of
Sobhasan Complex, North Cambay Basin, India. 6 th International Conference and Exposition on Petroleum Geophysics.
Kolkata, India, pp. 1062–1070.
Davies, R.J., (ed.), 2007, Seismic geomorphology: applications to hydrocarbon exploration and production: Geological
Society of London, 274 p.
Gupta, S., Chakraborty, Kanungo, U.S., Bharsakale, A., Dattaray, K., 2006. Reservoir delineation of Lower Kalol pays and
field growth in Wadu and Paliyad areas, Cambay basin. 6th International Conference and Exposition on Petroleum
Geophysics. Kolkata, India, pp. 1358–1365.
Jia Mingchen, Chen Yongjun and Guo Xuguang. Application of Seismic Sedimentology Method to Lithologic Trap
Identification of Shinan Area in Junggar Basin.
Juye Shi, Tailiang Fan, Hongyu Wang, Dongfang Yu, Xiao Ai. Application of Seismic Zedimentology in the Channel
Interpretation: A Case Study from the Block Doshan in South Turgay Basin, Kazakstan
Kundu, J., Prakash, A., Prasad, B., 1997. Structural framework and distribution of synrift depocentres in Cambay rift basin,
western India. Am. Assoc. Pet. Geol.Bull. 81, 1392–1393

Liu Bao-guo, LIU Li-hui. Application of applied seismic sedimentology in sedimentary facies analysis [J]. Geophysical
Prospecting for Petroleum, 2008, 47(3): 266−271.ol. 22 [2015], Bund. 22 12351
Mao zhiguo, Fan Tailiang, Ling Zongfa: Application of seismic sedimentology high-resolution sequence stratigraphic
frameworks: Taking block A of South Turgay Basin, Kazakstan as an example.

Sarraf S.C.*, A. Chatterjee, S.M. Qureshi, P. Chaturvedi, P.K.Gupta,Roopchand, D.P. Sahasrabudhe. Geological Modeling
and Facies Analysis in Sanand Field,Cambay Basin, India, 7th International Conference and Exposition on Petroleum
Geophysics. Hyderabad, India, p-249.

Rob Simm and Mike Bacon, 2014, Seismic Amplitude: An Interpreter’s Handbook: Cambridge University Press, 271 p.

Schlumberger : An interpreters Guide to seismic attributes , Petrel 2009, For Western Geco internal use only , 205 p.

Xu, L., 2006, Using Spectrum Decomposition Technique for Prediction of Thin Reservoir: Oil Geophysical Prospecting, v.
1/3, p. 299-302.
Yueying Hong, Hongmei Wang, Wenyuan Tian, and Aling Mo. Identification and Prediction of the High Heterogeneous
Channel Sand in Southern Turgay Basin*
Zeng, H. 2010. Stratal Slicing: benefits and challenges. The Leading Edge, 29, 1040–1047.
Zeng, H., Henry, S. C., and Riola, J. P. 1998b.Stratal Slicing, Part II: Real Seismic Data.Geophysics, 63, 514–522.
Zeng, H., and T.F. Hentz, 2004, High-frequency sequence stratigraphy from seismic sedimentology: Applied to Miocene,
Vermilion Block50, Tiger Shoal area, offshore Louisiana: AAPG Bulletin, v. 88/2, p. 153-174.
Zeng, H., M.M. Backus, K.T. Barrow, and N. Tyler, 1998, Stratal slicing, Part 1: Realistic 3D seismic model: Geophysics, v.
63/2, p. 502-513.
Zeng Hong-liu, Backus M M. Interpretive advantages of 90-phase wavelets (Part 1): Modeling [J]. Geophysics, 2005, 70(3):
7-15.
Zeng, H., and T.F. Hentz, 2004, High-frequency sequence stratigraphy from seismic sedimentology: Applied to Miocene,
Vermilion Block50, Tiger Shoal area, offshore Louisiana: AAPG Bulletin, v. 88/2, p. 153-174.

Zeng, H., M.M. Backus, K.T. Barrow, and N. Tyler, 1998, Stratal slicing, Part 1: Realistic 3D seismic model: Geophysics, v.
63/2, p. 502-513.