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

net/publication/307880640

Low-frequency seismic analysis and direct hydrocarbon indicators

Conference Paper · September 2016

DOI: 10.1190/segam2016-13858356.1

CITATIONS READS
4 444

2 authors:

Bilal Ahmed Bhatti R. James Brown

University of Stavanger (UiS) University of Stavanger (UiS)
2 PUBLICATIONS   4 CITATIONS    105 PUBLICATIONS   984 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

The Low Frequency Project View project

Low frequency project View project

All content following this page was uploaded by Bilal Ahmed Bhatti on 15 October 2019.

The user has requested enhancement of the downloaded file.

Low-frequency seismic analysis and direct hydrocarbon indicators
Bilal Ahmed Bhatti* and R. James Brown, University of Stavanger, Norway

Summary Tai et al. (2009) consider the phenomenon of low-

frequency energy anomalies associated with reservoir zones
Using seismic and well data from the Johan Sverdrup field, (Figures 1 and 2). One sees in their analysis that the
Central North Sea and the Western Graben, Southern North ? connection between hydrocarbon occurrence and the
Sea, a low-frequency analysis has been carried out in the ? character of low-frequency anomalies is not always simple.
context of differences in lithological properties of
reservoirs. Spectral decomposition is applied to stacked
seismic traces to enable low-frequency analysis and
comparison of results from the two oil-bearing reservoirs.
To look for low-frequency anomalies, we have used
spectral sections of 5 Hz to 40 Hz. Additionally, in order to
analyze anomalous low-frequency responses, so-called
frequency gathers (amplitude vs time in various narrow
frequency bands) at a single seismic trace or well location
have been used. This study focuses on finding a connection
between the occurrence or non-occurrence of low-
frequency anomalies and the nature – heterogeneous or
homogeneous – of the reservoir. Here, Johan Sverdrup is Figure 1: Modified from Tai et al. (2009): (a) Low-
considered a relatively homogeneous reservoir while the frequency anomaly profile including two wells. CMPs and
2/7-31 discovery in Western Graben is considered a traveltime (ms) are plotted on the horizontal and vertical
relatively heterogeneous reservoir. axes, respectively. The red-orange colors represent low-
frequency anomalies in the section.
Introduction

Accumulating evidence suggests that hydrocarbons often

exhibit a low-frequency anomaly in seismic data.
Extraction of the frequency components of seismic waves
by spectral decomposition can be carried out by a number
of algorithms. Taner et al. (1979) observed anomalously
high-amplitude low-frequency evenets associated with a
gas reservoir. Since then, several studies, including:
laboratory studies (physical modeling), synthetic studies
(numerical modeling), and field examples have been
reported in the literature, in which low-frequency
anomalies indicate the presence of hydrocarbons. Physical
modeling studies have been carried out by e.g. Goloshubin Figure 2: Frequency-versus-time sections of traces;
et al. (2002), Batzle et al. (2006), Madonna et al. (2010) frequency (Hz) (horizontal axis) against time (vertical axes)
and Nakagawa et al. (2015). Numerical modeling studies in seconds. Large shifts in the peak frequencies (red-orange
have been published by e.g. Castagna et al. (2003), colors) are at about 11 Hz and 14 Hz. The left-hand
Carcione et al. (2010) Silin and and Goloshubin (2010), reservoir is commercial gas, the right-hand reservoir was
Chen et al. (2010), Lefeuve-Mesgouez et al. (2010) and dry (gas-fizz).
Huang et al. (2015). Several authors have observed low-
frequency anomalies associated with hydrocarbon Asymptotic numerical analysis of Biot’s theory of
reservoirs in reflection-seismic exploration data, e.g. poroelasticity by Goloshubin and Chabyshova (2012)
Goloshubin et al. (2002), Castagna et al. (2003), observed converted fast-slow-fast P-waves from a thinly
Goloshubin et al. (2006), Chen et al. (2012) and Zhong layered porous permeable fluid-saturated model of a
(2014). The mechanism responsible for these effects is still sandstone reservoir and the seismic frequency response.
not clearly understood. However, several investigators have According to their hypothesis, any low-frequency anomaly
attempted to find out the possible cause of these effects. is expected to be weakest in massive homogeneous thick
reservoirs and strongest for highly interbedded permeable
 Low-frequency seismic analysis and direct hydrocarbon indicators

reservoirs. Therefore, gas shadows can potentially indicate Well Analysis

not only hydrocarbons but reservoir quality as well.
The objective of this study is to test their hypothesis by
comparing the low-frequency response of the relatively
homogeneous massive sandstone reservoir of the Draupne
Fm (Johan Sverdrup field, Central North Sea) with the
relatively heterogeneous permeable reservoir containing
sand lenses of the Ula Fm (Western Graben, Southern
North Sea).
Data and methodology
The database for this study comprises well data from two
wells and two post-stack full-offset migrated 3D broad
band seismic data, CGG-R14 located in Johan Sverdrup
field, Central North Sea and MC3D-CGR2012 located in
the Western Graben, Southern North Sea. Well logs
provided for this study include gamma-ray, caliper, neutron
porosity, density, resistivity and sonic logs, etc. (Figure 3).
Analysis of the two wells has been carried out in this study:
The 16/3-4 well is situated in Johan Sverdrup field, Central
North Sea. It was drilled to prove the extension of the
Avaldsnes Discovery to the north-east of the structural crest
of the Avaldsnes structure. Figure 4a is the log response of
well 16/3-4. The gamma-ray log shows a smooth trend in
reservoir having massive homogeneous sandstone having
high porosity values and exceptionally high permeability.
Resistivity values abruptly increase where the reservoir is
hydrocarbon saturated. The 2/7-31 well was drilled in
Western Graben to test the hydrocarbon potential of Upper
Jurassic fluvial sand lenses of Ula formation. From the
gamma-ray log shown in Figure 4b, the reservoir consists
of mostly sandstone but interbedded with thin layers of
shale. Moreover, the reservoir is of higher impedance than
the overlying shales. The resistivity values changes from 35
to 50 ohm-m where the reservoir is hydrocarbon saturated.

After loading the well and seismic data into the

interpretation software, well-log analysis was carried out to
differentiate between the lithological properties of the two
reservoirs. To tie the well to the seismic data, synthetic
seismograms were generated, followed by seismic
interpretation and time-structural mapping. Finally, spectral
decomposition was performed to analyze the frequency
responses of two different reservoirs.

Geological setting

Sørlie et al. (2012) describe the Johan Sverdrup discovery,

made in August 2010 in well 16/2-6, PL 501. It is defined
as a fault-bounded trap against the Southern Utsira
Basement High and overlying Jurassic shales and/or
Cretaceous Åsgard marls. The main reservoirs consist of a
sequence of Middle and Upper Jurassic sandstones of
excellent quality. The Volgian age sandstones of the
Draupne Fm consist of unconsolidated, massive
homogeneous sands and are thought to represent an
amalgamated marine, high-density gravity flow or bedload
deposits.

MC3D-CGR2012 is located in the Western Graben of the

Southern North Sea, approximately 260 km SW of Farsund
on the south-west coast of Norway, and it is adjacent to the
United Kingdom and Denmark sectors. Geologically, this
area is located between the Lindesnes Ridge to the north-
east and Grensen Nose to the south-west. This area has
proven hydrocarbon potential and several oil and gas fields
have been discovered on Lindesnes Ridge, e.g. Eldfisk,
Embla and Vallhall, among other, from Jurassic and
Cretaceous reservoirs (Gennaro et al., 2013).
Figure 3: Location map of study area
 Low-frequency seismic analysis and direct hydrocarbon indicators

anomalies are observed at and below the reservoir on 10 Hz

spectral section, whereas, the anomaly is weaker at the
reservoir top and absent below the reservoir.

Figure 4: Well-log responses of wells: (a) 16/3-4 and (b)

2/7-31. The logs include: gamma-ray, neutron porosity and
density (cross-plot) and resistivity log. Seismic data for
16/3-4 is taken from seismic cube CGG-R14 and 2/7-31 is
taken from seismic cube MC3D-CGR2012.

Spectral analysis

Spectral decomposition is used to convert seismic data into

its frequency components. A number of spectral
decomposition techniques have been proposed in the
literature, for example, the short-time Fourier transform
(STFT), the continuous wavelet transform (CWT), the S-
transform, matching-pursuit decomposition and
instantaneous spectral analysis (ISA) (Mallat and Zhang,
1993; Qian and Chen, 1994; Castagna et al., 2003; Lu and
Li, 2012). In this analysis, a sub-type of CWT, the Gabor-
Morlet transform is used because of its high resolution
(S.S. Ahmad, personal communication).
Figure 5 is an example of Inline 2167 from CGG-R14
(Johan Sverdrup field, Central North Sea), which is
spectrally decomposed into narrow bands around 10 Hz
(Figure 5a) and 35 Hz (Figure 5b). The top reservoir is
marked at 1850 ms with an arrow on both the low- and
high-frequency spectral sections. However, no significant
low-frequency anomaly response is observed associated
with the massive homogeneous hydrocarbon-bearing
sandstone of the Draupne Fm.
Figure 6 is an example of the spectral analysis that was
carried out on Inline 2617 from MC3D-CGR2012 (Western
Graben, Southern North Sea), which is spectrally
decomposed into narrow band spectral sections of 10 Hz
(Figure 6a) and 30 Hz (Figure 6b). The top reservoir is
marked with an arrow at 3680 ms in both low- and high-
frequency spectral sections. Significant low-frequency
Figure 5: Spectrally decomposed, low-frequency anomaly
profile (Inline 2167) including the well 16/3-4. This section
is taken from the seismic cube, CGG-R14. CMPs and
traveltime (ms) are plotted on horizontal and vertical axes,
respectively. The colors represent amplitude in the section
(blue = high) at two frequencies: (a) 10 Hz and (b) 35 Hz.
Red arrow indicating reservoir top.
Discussion and conclusion
In this study, we have analyzed the low-frequency response
from two different reservoirs having distinct lithological
properties. Based on our analysis, the low-frequency
anomalous response is found to be strong at and below the
interbedded heterogeneous Ula reservoir, but the response
was more-or-less absent at and below the massive
homogeneous Draupne reservoir at Johan Sverdrup. These
results support the hypothesis proposed by Goloshubin and
Chabyshova (2012), according to which, low-frequency
anomalies associated with hydrocarbons are expected to be
weakest in massive homogeneous thick reservoirs and
strongest for highly interbedded pearmeable reservoirs.
 Low-frequency seismic analysis and direct hydrocarbon indicators

This suggests that the relatively heterogeneous nature of the Acknowledgements

Ula reservoir and its relatively low permeability (though
quite high, but not nearly as permeable as the Draupne
reservoir) leads to greater attenuation of the high-frequency
content than at Johan Sverdrup. However, such a
mechanism needs further investigation and other factors,
like e.g. resonance effects due to conversions at reservoir
interfaces involving fast-slow-fast P-waves could also be a
contributing element.
This research represents part of a thesis project for the
Master of Science degree in Petroleum Geosciences
Engineering at the University of Stavanger, Norway.
The authors would like to acknowledge Lundin AS,
Norway and PGS AS, Norway for providing the data sets.
The authors would like to thank Børge Rosland at
Skagen44 AS for his abundant assistance throughout the
study as an external supervisor for this thesis project.
Finally, the author would like to thank Professor Alejandro
Escalona and PhD candidate Sayyid Suhail Ahmad for
guidance throughout the study.

Figure 6: Spectral sections at 10 Hz (a) and (30 Hz) (b) are

taken from the seismic cube MC3D-CGR2012 are low-
frequency anomaly profiles at Inline 2617 including well
(2/7-31). This section is taken from the seismic cube, CGG-
R14. CMPs and traveltime (ms) are plotted on horizontal
and vertical axes, respectively. The colors represent energy
at two frequencies. Red arrows indicate reservoir top.

View publication stats