Acoustic Impedance Inversion of the WASP data

Acoustic impedance inversion of vintage seismic data over a

proposed CO2 sequestration site in the Lake Wabamun Area,
Alberta
Abdullah Alshuhail, Don Lawton and Helen Isaac
ABSTRACT
Seismic inversion of the Nisku Formation in the Lake Wabamun area has revealed
contrasts in acoustic impedance caused by lateral changes in lithology and/or porosity of
the Nisku Formation. This interpretation is constrained by well control and supported by
seismic modelling, which suggests that changes in Nisku thickness over the range
encountered in the study area has only a small effect on acoustic impedance. Thus,
acoustic impedance mapping may provide one approach in pursuing favourable sites for
CO2 injection in addition to the conventional time structure, amplitude maps and other
seismic attributes. Our analysis revealed favourable low-impedance, high-porosity
locations that could be potential injection sites.
INTRODUCTION
The province of Alberta contributes more than 30% to the total Canadian CO2
emissions, according to recent statistics by Environment Canada (2007). In an effort to
offset the carbon emissions and meet up with potential increase in energy demands,
several CO2 sequestration projects were recently launched in the province. These include
the Alberta Saline Aquifer CO2 Project (ASAP), the Heartland Area CO2 Sequestration
Project (HARP), and the Wabamun Area CO2 Project (WASP).
In the Lake Wabamun Area, the Nisku Formation has been advocated as a sink for
large-scale CO2 sequestration project (Michael et al., 2008) with capacity in the order of a
megaton per year. Therefore, seismic characterization is being undertaken as part of
Phase I of WASP to: (1) generate detailed attribute maps of the Nisku Formation, (2)
provide indications of variations in its porosity and lithology, and (3) delineate any
potential geologic risks in the area that may compromise the integrity of the Nisku
Formation as a CO2 sink. This report outlines some of the results from the ongoing
seismic characterization pertaining to the acoustic impedance inversion.
STUDY AREA
The study area is located approximately 50 km southwest of Edmonton in the vicinity
of the Lake Wabamun Area in the central plains of Alberta, Canada (Figure 1). The
injection target is the Devonian brine-bearing Nisku dolostone formation. The formation
was deposited at the edge of carbonate shelf and its lithology becomes interbedded with
limestone and shale as we move to the northwest direction (Watts, 1987). The Nisku
Formation is confined between two aquitards (Figure 2) and its depth and thickness range
from 1800 - 2200 m and 40 - 80 m, respectively. Furthermore, the overlying Calmar shale
makes a good cap rock which prevents injected CO2 from seeping into the surface or into
shallow aquifers. In addition, the Nisku aquifer properties, such as temperature and
salinity, make it a favourable candidate for CO2 storage (Michael et al., 2008). The

CREWES Research Report Volume 21 (2009)

Alshuhail, Lawton, and Isaac

seismic characterization is constrained to a high-grade focus area, where favourable

conditions in terms of seismic coverage and other factors exist.

WASP Large-scale
Study Area
(LSSA)

(?!)
WASP High-grade
Focus Area
(HGFA)

Figure 1: Location map of Wabamun Lake study area (NRCan, 2009).

WASP SEISMIC PROGRAM

Seismic characterization of the Nisku Formation in the Wabamun Area CO2
Sequestration Project (WASP) high-grade focus area (HGFA) was based on analyzing
and interpreting poststack seismic datasets comprising more than two hundred 2D
seismic lines and seven 3D volumes. The study area is surrounded by two major
hydrocarbon resources in Alberta; the Leduc reef play (east) and the Moon Lake reef play
(northwest). Thus, part of the study area was mapped using vintage surface seismic data
that had been acquired as part of hydrocarbon exploration in the area. The 2D and 3D
seismic datasets have different acquisition and processing specifications and were
acquired over many years prior to this project being undertaken. Therefore, prior to
interpretation, inversion and attribute analysis, two primary steps were undertaken: data
calibration and amplitude normalization. These steps were necessary to account for the
vintage and datum differences within the data.

CREWES Research Report Volume 21 (2009)

Acoustic Impedance Inversion of the WASP data

Table 1 gives an outline of the volume and approximate areal coverage of the analyzed
seismic data as well as the number of wells with appropriate log curves and formation
tops that were available for integration into the seismic data analysis. No new seismic
data were acquired as part of Phase I of the project. The spatial distribution of the
available seismic data is illustrated in Figure 3, and shows that the data coverage is not
distributed uniformly. Thus characterization was constrained to those areas with good
coverage. The 2D seismic data were used primarily for identifying long-wavelength
structures, whereas the high-quality 3D data were used for detailed mapping, inversion
and generation of seismic attributes.
GR (API) DT (us/m)

Banff/Exshaw ~ 1505 m
Wabamun ~ 1555 m
Calmar ~ 1725 m
Nisku ~ 1729 m
Ireton ~ 1828 m

LEGEND
Sandstone
Carbonate
Shale
Evaporites
Crystalline
Aquifer
Aquitard
Aquitard/Aquiclude
(Bachu and Bennion, 2009)

Basal Sandstone ~ 2868 m

Figure 2: Stratigraphic model of Paleozoic strata in the study area using well 10-05-052-2W5
(Bachu and Bennion, 2009).

CREWES Research Report Volume 21 (2009)

Alshuhail, Lawton, and Isaac

Table 1: Summary of the available seismic and wireline data. HGFA refers to the WASP highgrade focus area. * The borehole data referred to here are those with both sonic log and Ireton
tops.

Seismic
data
Borehole
data*

2D

200 seismic lines; approximately 2432 km in length

3D

7 seismic volumes covering an area of 419 km2

Within HGFA

7 wells. Only 2 of which coincide with seismic coverage

Nearby HGFA

Over 15 wells. Only 6 coincide with seismic coverage

Moon Lake Reef

Play
Leduc Reef Play

Nisku Shelf

LEGENDS
WASP Largescale Study
Area

Reference 3D
volume

WASP Highgrade Focus

Area

3D Seismic
2D Seismic
Wells with Sonic
& Ireton Top
Water Source Well
Water Injection
Dry & Abandoned
Producing Gas
Suspended Gas
Water Source
Abandoned Oil

Pigeon Lake

0.0

20.0

40.0 km

Figure 3: Base map showing the distribution of the seismic and borehole data. Cyan shapes
indicate bodies of water.

CREWES Research Report Volume 21 (2009)

Acoustic Impedance Inversion of the WASP data

METHOD: ACOUSTIC IMPEDANCE INVERSION

Acoustic impedance is one of the most useful seismic attributes as it yields distribution
of pseudo-physical properties rather than a set of observations pertaining to the physical
properties distribution. However, acoustic impedance requires good estimation of
parameters as well as high quality seismic data and a good distribution of well control.
Furthermore, inversion suffers from non-uniqueness.
In this project, various post-stack acoustic impedance inversion techniques were tested
and only two were found to produce useful results: bandlimited and model-based
inversions. Bandlimited inversion, also known as recursive inversion, is constrained
mainly by the bandwidth of the seismic data, which typically falls between 7 and 60 Hz.
The method estimates the acoustic impedance recursively by first extracting an estimate
of the reflection coefficient from the seismic data and then re-arranging the normal
incidence reflection coefficient relation to solve for the acoustic impedance of the nth+1
layer (Zn+1) (Cooke and Schneider, 1983):
Zn 1 Zn solving for Zn 1
Zn 1 R n
Zn 1
1R n
Zn 1 Zn
Where Rn is the reflection coefficient of the nth layer, Zn is the acoustic impedance of the
nth layer (kg/m2.s), and n is a positive integer. Recall that the acoustic impedance of the
nth layer is its P-wave velocity (m/s) multiplied by its density (kg/m2), i.e. Zn = n n.
Rn

Model-based inversion uses a different approach. First, the well control and the
seismic data (horizons) are used to build an initial low-frequency estimated model of the
acoustic impedance distribution. Using an estimate of the source wavelet, the model is
then perturbed and the model response, in the form of synthetic seismogram, is measured.
The model responses are then compared to the actual seismic traces, usually by means of
crosscorrelation. The process is iterated until the model converges, i.e. the model
response becomes within a predefined acceptable range from the actual observation. The
misfit error between the two is quantified through various means; one of the most
commonly used measures is the sum of the squared differences (Lines and Treitel, 1984).
There are many elements that could degrade the reliability of the inversion results,
some of which could not be controlled, such as noise, whereas others could not be
precisely calculated, such as the source wavelet. However, each method has its own
advantages and disadvantages in regard to those limitations and it is suggested that by
using both methods some of the ambiguities associated with the inversion results could
be minimized. More information about those and other inversion methods can be found in
Waters (1978), Aki and Richards (1980), Lines and Treitel (1984), Russell (1988), and
the STRATA software theory manual. Figure 4 depicts some of the major steps adopted
in the acoustic impedance inversion framework.

CREWES Research Report Volume 21 (2009)

Alshuhail, Lawton, and Isaac

Seismic data (from Seisware)

calibrated + horizons picked
Well logs (from database)
Ireton top + sonic

Wavelet extraction I
statistical
Edit logs
median filteretc

Inversion analysis:
inversion method(s) + parameters optimization

Execute the inversion

1
Seismic-to-well tie 4
appropriate CDPs
2
Wavelet extraction II
using well

Build the acoustic impedance

model using appropriate wells

Extract the acoustic impedance

map of the Nisku

Porosity/Lithology
indications

Figure 4: Flowchart outlining the major steps followed in the seismic inversion to extract the
acoustic impedance map of the Nisku event.

RESULTS
Field Data
A critical factor to achieving good inversion results is the seismic-to-well tie. The
correlation coefficients associated with the wells used in the inversion are shown in
Figure 5. Figure 6 shows example of the seismic-to-well tie near the water source well
(1F1-11-29-45-2W5), at which a very good correlation is obtained (0.92). Prior to
showing the Nisku acoustic impedance map, two examples were selected to illustrate the
performance of each of the inversion methods. The first example is from the 2D seismic
line near the water source well (Figure 7) while the other is from an inline extracted from
the reference 3D volume (Figure 10). Figure 8 and Figure 11 show the initial guess
model for the 2D and the extracted inline seismic sections while the inversion results
using bandlimited and model-based inversion for those two sections are illustrated in
Figure 9 and Figure 13, respectively.
The bandlimited inversion appears to produce a more detailed acoustic impedance
model than the model-based inversion. For instance, the Wabamun and the Nisku
formations are clearly separated by low impedance in the bandlimited inversion whereas
they are hardly separated in the model-based inversion results (Figure 9 and Figure 13).
However, for the acoustic impedance of the Nisku Formation, both methods yield similar
results (Figure 13 and Figure 14), the only apparent difference being in the magnitude of
the impedance. This is probably due to scaling differences. Furthermore, because there is
a lack of well control, it is crucial that the impedance maps are interpreted only in terms
of relative rather than absolute changes in acoustic impedance.
Several interesting low impedance zones are highlighted in Figure 13 and Figure 14.
The impedance determination from the 2D seismic line near the water source well is also
shown for comparison. By examining those maps, there seems to be two categories of
low impedance: one that is associated with lithological/porosity changes in the Nisku
Formation and another which is associated with discontinuities in the overlying

CREWES Research Report Volume 21 (2009)

Acoustic Impedance Inversion of the WASP data

Wabamun event. Even though the results are not shown here, discriminating against the
two groups of anomalies was achieved using coherency-sensitive methods. However,
differentiation between acoustic impedance changes caused by enhanced porosity and
those associated with a possible increase in shale content remains tenuous. With respect
to the lithological/porosity changes, low acoustic impedance (Figure 13 and Figure 14)
appears to normally correspond to low NRMS (Figure 15) while the time structure of the
Nisku (Figure 16) is rather smooth and does not exhibit any significant variations within
the WASP focus area expect for following the regional dip in the northeast-southwest
direction.

Inline

LEGENDS

100-10-21-50-2W5

WASP Largescale Study

Area

100-05-20-50-2W5

100-16-16-50-2W5

Reference 3D
volume

WASP Highgrade Focus

Area

100-08-20-48-27W5

3D Seismic
2D Seismic
Wells with Sonic
& Ireton Top

100-02-21-48-1W5

Water Source Well

Water Injection
Dry & Abandoned
Producing Gas

2D seismic

Suspended Gas
Water Source
Abandoned Oil

1F1-11-29-45-2W5

Well
1F1-11-29-45-2W5
100-10-21-50-2W5
100-16-16-50-2W5
100-05-20-50-2W5
100-02-21-48-1W5
100-08-20-48-27W4

0.0

Seismic Reference
2D
3D
3D
3D
3D
3D

20.0

40.0 km

Correlation Coefficient
0.92
0.80
0.83
0.72
0.63
0.80

Figure 5: Correlation coefficient, over designed correlation window, between seismic data and
synthetic seismogram from selected wells within the study area. The wells location are shown in
the base map. The blue lines show the location of the 2D seismic (Figure 7) and the inline (Figure
10) invoked in the inversion.

CREWES Research Report Volume 21 (2009)

Alshuhail, Lawton, and Isaac

LEGENDS
WASP Largescale Study
Area

WASP Highgrade Focus

Area

Reference 3D
volume

3D Seismic
2D Seismic
Wells with Sonic
& Ireton Top
Water Source Well
Water Injection
Dry & Abandoned
Producing Gas
Suspended Gas
Water Source

Water source well

1F1-11-29-045-2W5

Abandoned Oil

2D seismic line
0.0

20.0

40.0 km

P-wave
Deep
Gamma
Caliper
Synthetic
Density
SP
Velocity
Induction
Ray
(cm)
Seismogram
(kg/m3)
(mV)
(m/s)
(ohm-m)
(API)
0
40 -400
239
50 1500
3000 1000 6000
0 0.1
1x104 0

Seismic
(CDP)
247

255

300

295

400
500

700
800

1215

Depth (m)

Two-way Time (ms)

600

900
1000
Viking

1634

1100
1200
1300

Nisku

2250

1400

Figure 6: Seismic-to-well tie at the water source well (1F1-11-29-45-2W5). The location of the
well and the seismic (blue) line are shown in the top image. The correlation coefficient is 0.92
over the outlined zone (dashed rectangle) in the bottom image.

CREWES Research Report Volume 21 (2009)

Acoustic Impedance Inversion of the WASP data

SE

110

150

200

1100

Trace
1F1-11-29-045-2W5

300

350

NW

Viking

Two-way Time (ms)

1200

1300

Wabamun

Nisku
1400
NRMS Amplitude
-1.0

0.0

1.0

1500
BH Lake

Figure 7: Seismic section near the water source well (Figure 5). The green curve at the well
location is the correlated synthetic seismic trace. The correlation coefficient is 0.92.
SE

110

150

200

1100

Trace
1F1-11-29-045-2W5

300

350

NW

Viking

Two-way Time (ms)

1200

1300
Wabamun

Nisku
1400
Acoustic Impedance (kg/m2.s)
6.0 x 106

1.25 x 107

2.0 x 107

Figure 8: The initial acoustic impedance model of the seismic section in Figure 7. The blue curve
at the well location is the correlated synthetic seismic trace while the black curves are the actual
seismic traces.

CREWES Research Report Volume 21 (2009)

Alshuhail, Lawton, and Isaac

SE

110

150

200

1100

Trace
1F1-11-29-045-2W5

300

350

NW

Viking

Two-way Time (ms)

1200

1300
Wabamun

Nisku
1400
Acoustic Impedance (kg/m2.s)
6.0 x 106

SE

110

150

1.25 x 107

200

1100

Trace
1F1-11-29-045-2W5

2.0 x 107

300

350

NW

Viking

Two-way Time (ms)

1200

1300
Wabamun

Nisku
1400
Acoustic Impedance (kg/m2.s)
6.0 x

106

1.25 x 107

2.0 x 107

Figure 9: Estimated acoustic impedance along the 2D seismic line near the water source well
(Figure 5) using bandlimited (top) and model-based (bottom) inversion methods. The inserted
blue curve at the well location represents the computed acoustic impedance from the sonic and
density logs. The black curves represent the acoustic impedance from the bandlimited inversion
whereas in the model-based inversion they represent the misfit error.

10

CREWES Research Report Volume 21 (2009)

Acoustic Impedance Inversion of the WASP data

N

1
1000

100-10-21-050-2W5

200

Crossline
300

400

NRMS Amplitude
0.0

-1.0

1.0

Viking

1100

Two-way Time (ms)

500

1200

Wabamun
1300
Nisku

1400

BH Lake

1500

Figure 10: Inline extracted from the reference 3D seismic volume (Figure 5). The green curve at
the well location is the correlated synthetic seismic trace. The correlation coefficient is 0.80.
N

1
1000

100-10-21-050-2W5

200

Crossline
300

400

Acoustic Impedance (kg/m2.s)

6.0 x 106

1.25 x 107

2.0 x 107

Viking

1100

Two-way Time (ms)

500

1200

Wabamun
1300
Nisku

1400

BH Lake

1500

Figure 11: The initial acoustic impedance model corresponding to the inline in Figure 10. The blue
curve at the well location is the correlated synthetic seismic trace while the black curves are the
actual seismic traces.

CREWES Research Report Volume 21 (2009)

11

Alshuhail, Lawton, and Isaac

N

1
1000

100-10-21-050-2W5

200

Crossline
300

400

Acoustic Impedance (kg/m2.s)

6.0 x

106

1.25 x 107

2.0 x 107

Viking

1100

Two-way Time (ms)

500

1200

Wabamun
1300
Nisku

1400

BH Lake

1500
N

1
1000

100-10-21-050-2W5

200

Crossline
300

400

Acoustic Impedance (kg/m2.s)

6.0 x

106

1.25 x 107

2.0 x 107

Viking

1100

Two-way Time (ms)

500

1200

Wabamun
1300
Nisku

1400

BH Lake

1500

Figure 12: Estimated acoustic impedance of the inline in Figure 10 and Figure 11 using
bandlimited (top) and model-based (bottom) inversion methods. The inserted blue curve at the
well location represents the computed acoustic impedance from the sonic and density logs. The
black curves refer to the acoustic impedance from the bandlimited inversion whereas in the
model-based inversion they represent the misfit error.

12

CREWES Research Report Volume 21 (2009)

Acoustic Impedance Inversion of the WASP data

Low acoustic impedance associated

with discontinuities footprint
Low acoustic impedance
Ip (kg/m2.s)
1.30x107

1.65x107

2.0x107

Low impedance near the

water source well

0.0

20.0 km

Figure 13: Estimated Acoustic impedance (Ip) map of the Nisku Formation using bandlimited
inversion. Due to lack of well control, the inversion was not performed on the entire dataset.

CREWES Research Report Volume 21 (2009)

13

Alshuhail, Lawton, and Isaac

Low acoustic impedance associated

with discontinuities footprint
Low acoustic impedance
Ip (kg/m2.s)
1.30x107

1.65x107

2.0x107

Low impedance near the

water source well

0.0

20.0 km

Figure 14: Estimated Acoustic impedance (Ip) map of the Nisku Formation using model-based
inversion. Due to lack of well control, the inversion was not performed on the entire dataset.

14

CREWES Research Report Volume 21 (2009)

Acoustic Impedance Inversion of the WASP data

Discontinuity
footprint

Karsting
footprint

Low amplitude

NRMS
Amplitude
-3

0.0

20.0 km

Figure 15: Nisku NRMS amplitude map. Some of the major amplitude anomalies are specified:
Wabamun karsting footprint, Wabamun discontinuities footprints, and the Nisku amplitude low. It
should be emphasized that the karsting and discontinuities shown here are just due to the
footprint of those anomalies and do not indicate that the Nisku has been physically affected.

CREWES Research Report Volume 21 (2009)

15

Alshuhail, Lawton, and Isaac

Discontinuity
footprint

Karsting
footprint

Low time

Time (ms)
1200

1300

1400

1500

0.0

20.0 km

Figure 16: The Nisku time structure map. The locations of some of the major anomalies are
specified: Wabamun karsting footprint, Wabamun discontinuities footprints, and the Nisku local
time low. It should be emphasized that the karsting and discontinuities shown here are due to the
footprint of those anomalies and do not indicate that the Nisku has been physically affected.

16

CREWES Research Report Volume 21 (2009)

Acoustic Impedance Inversion of the WASP data

Synthetic Data
There are several factors that could affect the acoustic impedance of the Nisku event
as observed in the field data. Therefore, some sensitivity analyses were undertaken and it
was found that thickness and average P-wave velocity are the primary parameters
affecting the Nisku event impedance. Subsequently, zero-offset synthetic seismograms
were generated using a 25 Hz Ricker wavelet and the convolutional model to further
understand the effect of these two parameters on the Nisku event acoustic impedance.
Figure 17 is a side-by-side panel display illustrating the thickness and average P-wave
velocity effect on the Nisku event amplitude whereas Figure 18 shows a display of the
Nisku event amplitude as a function of those two parameters. Changing the average Pwave velocity (vertical axis) will cause over a 60% change in the amplitude whereas
changing the thickness (horizontal axis) will only cause about an 8% variation in
amplitude (Figure 15).
h

10

20

30

40

50

60

70

80

90

100

avg. 5500 6500 5500 6500 5500 6500 5500 6500 5500 6500 5500 6500 5500 6500 5500 6500 5500 6500 5500 6500 5500 6500

Two-way Time (ms)

1000

1000

Wabamun

1100

Calmar

Nisku

Ireton

1200

1200

Blue annotation (h) is thickness in m.

Red annotation (avg.) is average P-wave velocity in m/s.

Figure 17: Sequential display of normal incidence synthetic seismograms in which the Nisku
event amplitude is modelled as function of thickness and average P-wave velocity. The top of the
Nisku event is identified as the peak at approximately 1130 ms (the green bracket along the time
axis approximates the frequency cycle through the formation). In each blue bracket (thickness
effect), there are 11 traces, each representing the seismic amplitude associated with that
thickness and an average Nisku P-wave velocity increasing from 5500 m/s to 6500 ms/ at an
increment of 100 m/s. The modelling was undertaken using well 100-10-05-052-02W5. The actual
Nisku thickness and average P-wave velocity at this well are 100 m and 6100 m/s, respectively.

CREWES Research Report Volume 21 (2009)

17

Alshuhail, Lawton, and Isaac

To relate variations in the acoustic impedance to the two primary physical parameters
of interest, i.e. thickness and average P-wave velocity, the acoustic impedance of the
synthetic seismogram in Figure 17 was reconstructed using bandlimited and model-based
inversion methods (Figure 19). The modelling results suggest that the acoustic impedance
variations in the study area, excluding those associated with karsting and discontinuities
footprints, are most likely due to variability in the average P-wave velocity (i.e. lithology
and/or porosity) rather than thickness.
6500
6400

Amplitude
0.06

Average P-wave Velocity (m/s)

6300
6200
6100

0.1

6000
5900
5800
5700

0.12

5600
5500
0

10

20

30

40

50

60

70

80

90

100

Thickness (m)

Figure 18: Nisku event amplitude as a function of thickness and average P-wave velocity
synthetic modelling. The actual Nisku thickness and average P-wave velocity are 100 m and
6100 m/s, respectively. The maps represent the Nisku amplitude map that resulted from picking
the peak amplitude corresponding to the Nisku event in Figure 17. The black dashed rectangle
outlines the most likely Nisku thickness and average velocity values within the study area based
on well control. The results suggest that the Nisku average P-wave velocity (or impedance) effect
is more significant than the thickness effect. Thus, the highest sensitivity is along the vertical axis.

18

CREWES Research Report Volume 21 (2009)

Acoustic Impedance Inversion of the WASP data

6500
6400

Ip (kg/m2.s)
1.5x 107

Average P-wave Velocity (m/s)

6300
6200
6100
6000

1.65 x 107

5900
5800
5700
1.8 x 107

5600
5500
0

10

20

30

40

50

60

70

80

90

100

Thickness (m)
6500
6400

Ip (kg/m2.s)
1.5x 107

Average P-wave Velocity (m/s)

6300
6200
6100
6000

1.65 x 107

5900
5800
5700
1.8 x 107

5600
5500
0

10

20

30

40

50

60

70

80

90

100

Thickness (m)

Figure 19: Acoustic impedance (Ip) of the synthetic seismogram in Figure 17 using bandlimited
(top) and model-based (bottom) inversion methods. Similar to the seismic amplitude response,
the acoustic impedance shows high sensitivity toward variations in the Nisku Formation average
P-wave velocity rather than thickness. The maps illustrate the direct proportionality between
average P-wave velocity and acoustic impedance. Note the similarity in the results except for the
small difference in the magnitude of the acoustic impedance due to scaling issues.

CREWES Research Report Volume 21 (2009)

19

Alshuhail, Lawton, and Isaac

CONCLUSIONS
The acoustic impedance map of the Nisku event shows strong variations across the
WASP study area. Our interpretation indicates that there are two categories of low
impedance: one that is associated with lithological/porosity changes in the Nisku
Formation and another which is associated with discontinuities in the overlying
Wabamun event. In addition to the time structure, NRMS amplitude and coherencysensitive attributes, forward and inverse seismic modelling were undertaken in
delineating anomalies caused by lithological variations in contrast to those induced by
geological discontinuities. The modelling results suggest that variation in the average Pwave velocity is the primary physical parameters steering the acoustic impedance of the
Nisku event. Consequently, several favourable zones of low acoustic impedance are
identified in the study area. However, differentiation between acoustic impedance
changes caused by enhanced porosity and those associated with a possible increase in
shale content remains tenuous.
AKNOWLEDGEMENT
We thank WASP sponsors and the Natural Sciences and Engineering Research
Council of Canada (NSERC) for funding the project and the WASP technical team for
valuable discussions, ENCANA kindly provided access to the seismic data, and Jimmy
Haszard (ENCANA), Jay LeBlanc (ENCANA), and Ian Reglar (formerly ENCANA) provided
valuable technical support. Hampson-Russell, Seismic-Micro Technology and Zokero are
thanked for providing seismic interpretation software. We also acknowledge the
continuing support of CREWES sponsors.
REFERENCES
Aki, K., Richards, P.G., 1980. Quantitative Seismology, Freeman and Co., New York.
Bachu, Stefan and Bennion, Brant, 2009, Effects of In-situ Conditions on Relative Permeability
Characteristics
of
CO2-Brine
System:
<http://www.gwpc.org/elibrary/documents/co2/LBNL%20CO2%20Effects%20of%20Insitu%20Conditions%20on%20Per
meability.pdf>.
Cooke, D. A., and Schneider, W. A., 1983, Generalized linear inversion of reflection seismic data:
Geophysics, 48, 665-676.
Lines, L. R., and Treitel, S., 1984, A review of least-squares inversion and its application to geophysical
problems: Geophysical Prospecting, 32, 159-186.
Russell, B.H., 1988, Introduction to seismic inversion methods: Society of Exploration Geophysicists,
Course Notes Series, No. 2, S. N. Domenico, Series Editor.
Sherriff, R. E., 1991. Encyclopedic Dictionary of Exploration Geophysics: Society of Exploration
Geophysicists, 4th edition, Tulsa, Oklahoma.
Michael, K., S. Bachu1, B. Buschkuehle1, K. Haug, and S. Talman, 2008, Comprehensive Characterization
of a Potential Site for CO2 Geological Storage in Central Alberta, Canada: AAPG Manuscript, 40
p.
Natural
Resources
of
Canada,
2009,
<
http://atlas.nrcan.gc.ca/site/english/maps/reference/provincesterritories/alberta/referencemap_imag
e_view>.
Waters, K. H., 1978, Reflection seismology: John Wiley and Sons.
Watts, N., 1987, Carbonate sedimentology and depositional history of the Nisku Formation in south-central
Alberta, in F.F. Krause and O. G. Burrowes, eds., Devonian lithofacies and reservoir styles in
Alberta: 2nd International Symposium of the Devonian System, 13th CSPG Core Conference, 87152.

20

CREWES Research Report Volume 21 (2009)