Analysis of P-P and P-SV seismic data from Lousana, Alberta

Transcription

Analysis of P-P and P-SV seismic data from Lousana, Alberta
P-PandP-SVseismicdatafromLousana
Analysis of P-P and P-SV seismic data from Lousana,
Alberta
Susan L.M. Miller, Mark P. Harrison*, Don C. Lawton, Robert R.
Stewart, and Ken J. Szata**
ABSTRACT
Two orthogonal three-component seismic lines were shot by Unocal Canada
Ltd. in January, 1987, over the Nisku Lousana Field in central Alberta. The purpose of
the survey was to investigate a Nisku patch reef thought to be separated from the Nisku
shelf to the east by an anhydrite basin. These data have been reprocessed and are
currently being analyzed and used to develop methods for multicomponent seismic data
analysis. The data are of overall good quality and major events were confidently
correlated between the P-P and P-SV sections using offset synthetic seismograms. PSV offset synthetic models were also used to extract interval Vp/Vs values from the PSV data.
Forward log-based offset P-P and P-SV modelling shows character, isochron,
and Vp/Vs variations associated with lithology and porosity changes between reservoir
and non-reservoir rock in the Nisku. The modelling results indicate that
multicomponent seismic analysis is applicable in carbonate settings, even when the
reservoir intervals are relatively thin (23 m). To date, we have had difficulty applying
the modelling results to this data set because of multiple contamination in the zone of
interest and work is ongoing to resolve this issue. Full-waveform modelling is planned
to better understand the multiple activity on both components. However, there are
observable anomalies on the data which are currently under investigation.
Multicomponent
recording provides additional seismic measurements of the
subsurface to assist in developing an accurate geological model. Rock properties which
can be extracted from elastic-wave data, such as Vp/Vs, reduce the uncertainty in
predictions about mineralogy, porosity, and reservoir fluid type. Joint interpretation of
P-P and P-SV data was helpful in horizon picking and in providing feedback on the
validity of the interpretation. The interpretation techniques described in this paper can
be usefully applied to other areas, including other carbonate plays.
INTRODUCTION
In 1987, Unocal acquired two three-component lines across the Lousana Field.
Unocal donated these data to the CREWES Project, and in 1993 they were reprocessed
using IT&A's software and converted-wave code developed in CREWES. More
recently, the data were reprocessed again using the ProMAX processing package and
the same converted-wave algorithms. The reprocessed sections exhibit significantly
broader bandwidth on both the vertical and radial components. Modelling and
interpretive studies are ongoing, with the following objectives:
* SensorGeophysicalLtd.,Calgary,Alberta
**ChevronCanadaResources,Calgary,Alberta
CREWESResearchReport Volume6 (1994)
7-1
Miller,Harrison,Lawton,Stewartand Szata
• identify character, amplitude, or time interval variations which may be
associated with productive zones, and attempt to calibrate these variations to
changes in geology through forward modeling,
• use full-waveform modelling to study the signal interference caused by
interbed multi-converted multiples,
• develop and test techniques for analyzing multicomponent data,
• evaluate the benefit of multicomponent versus conventional recording.
Benefits
of multicomponent
recording
This discussion concentrates on the benefits of multicomponent recording as it
applies to data interpretation. Recording multicomponent data provides complementary
images of the subsurface for a relatively small additional cost, as conventional sources
and receiver geometries are employed. Due to the difference in travel path, wavelength,
and reflectivity, P-SV seismic sections may exhibit geologically significant changes in
amplitude or character which are not apparent on conventional P-wave data. Horizons
may be better imaged on one of the components because of different multiple paths and
wavelet interference effects.
This additional seismic measurement of subsurface rock properties can reduce
the uncertainty in predictions about mineralogy, porosity, and reservoir fluid type. The
P-SV seismic data can be used in conjunction with the P-P data to determine other rock
properties, such as Vp/Vs (or similarly, Poisson's ratio). Vp/Vs is sensitive to changes
in a number of rock characteristics, including lithology, porosity, pore shape, and pore
fluid (e.g. Pickett, 1963; Nations, 1974; Tatham, 1982; Eastwood and Castagna, 1983;
Miller and Stewart, 1990). A number of studies, some of which are described below,
have indicated that multicomponent seismic technology has practical applications in
carbonate settings (Rafavich et al., 1984; Wilkens et al., 1984; Goldberg and Gant,
1988; Georgi et al., 1989; Wang et al., 1991).
Robertson (1987) interpreted porosity from seismic data by correlating an
increase in porosity with an increase in Vp/Vs. He based his interpretation on the model
of Kuster and Toks6z (1974), which incorporates pore aspect ratio as a factor in
velocity response. According to this model, Vp/Vs will rise as porosity increases in a
brine-saturated carbonate with flat pores, but will decrease slightly if the pores tend to
have a higher aspect ratio (are rounder). If the carbonate is gas saturated, Vp/Vs will
drop sharply as porosity increases if pores are flat, and drop slightly if pores are
rounder.
Multicomponent data have been used successfully to differentiate tight limestone
from reservoir dolomite in the Scipio Trend in Michigan. Pardus and others (1991)
mapped the variation in the Vp/Vs ratio across the interval of interest on P- and S-wave
seismic data and related it to the ratio of limestone to dolomite. Limestone tends to a
Vp/Vs value of 1.9, dolomite of 1.8. Petrophysical studies suggest that anhydrite also
has a higher Vp/Vs ratio than dolomite and thus might be used in a similar manner to
track dolomite/anhydrite variations (Miller, 1992).
Finally, having two or more seismic measurements of the same geology gives
the interpreter additional hard data to incorporate into the geological model. There is
another data set to work with in areas where the data quality is poor or the interpretation
is unclear. The P-SV section can be used to guide and constrain conventional data
interpretation. Joint horizon interpretation offers the interpreter feedback in the form of
the resultant interval Vp/Vs values. Vp/Vs values which lie outside the range expected
7-2
CREWESResearchReport Volume6 (1994)
P-PandP-SVseismicdatafromLousana
due to geological variations indicate mispicks,
confidence in the interpretation.
Methods
for multicomponent
whereas reasonable
values increase
analysis
This paper discusses several techniques which are useful in the analysis of
multicomponent seismic data. These can be summarized as:
• correlation of vertical and radial components through P-P and P-SV logbased offset synthetic modelling,
• matching P-SV offset synthetic seismograms to P-SV stacked seismic data
and iteratively adjusting interval Vp/Vs values on the model to tie the data,
• workstation interpretation of horizons on P-P and P-SV components and
calculation of interval Vp/Vs values using P-P and P-SV isochrons,
• log-based forward modelling to determine amplitude, character, and traveltime response of each component to a variety of geological settings,
• use of this model to determine the expected variation in Vp/Vs across
different time intervals for a range of rock properties.
GEOLOGY
The Lousana field is located west of the Fenn-Big Valley and Stettler Oil Fields
in Township 36, Range 21, West of the 4th Meridian, in central Alberta (Figure 1). The
target is a Nisku dolomite buildup which is separated from the Nisku carbonate shelf to
the east by an anhydrite basin, which forms the seal for the reef.
The Nisku is an Upper Devonian Formation in the Winterburn Group, which
conformably overlies the shales and carbonates of the Woodbend Group (Figure 2).
Immediately underlying the Nisku is the Ireton Formation, a calcareous shale. The
initial deposition of the Nisku occurred during a relative rise in sea level. The
antecedent topography of the Leduc reef complexes served as sites for deposition of the
Nisku carbonates, which was concentrated in shallow waters. Deposition was restricted
in deep-water areas to pinnacle reefs and small carbonate outliers (Stoakes, 1992).
During a later, regressive phase of deposition, sea water circulation was
restricted by the growth of the shelf margin in the northwest, the current location of the
West Pembina Field. Progressive evaporation resulted in the hypersaline deposits of an
upper evaporitic unit in the Nisku. Brining upward cycles of dolomudstones and
anhydrites infilled the lows between isolated reefs and platforms and capped the
succession. Basin infilling was completed by the deposition of the silts and shales of
the overlying Calmar Formation (Andrichuk and Wonfor, 1954).
The nearby Fenn West, Fenn-Big Valley, and Stettler Fields are structural traps
formed by the porous dolomitized Nisku shelf carbonates draping over Leduc reefs
(Rennie et al., 1989). There is no underlying Leduc at Lousana, which is west of the
Leduc edge, and the stratigraphic trap is a biohermal buildup of dolomite, perhaps one
of the carbonate outliers described by Stoakes (1992). The hypersaline deposits of the
upper phase of the Nisku filled the basin between the shelf and the buildup at Lousana,
and sealed the trap. The Nisku is 50 - 60 m thick in this area and the overlying Calmar
Formation is generally only about 3 m thick.
The Nisku at Lousana is a dolomite oil reservoir with up to about 25 m of
porosity in producing wells. The two oil wells in the field, 16-19-36-21W4 and 2-30-
CREWESResearchReport Volume6 (1994)
7-3
Miller, Harrison, Lawton, Stewart and Szata
/+
I
1
I
i
r
',..
°.
-o°I
Alberta
+i
. /,'_
Nisku _
\
__
i
+{
_,i
N
o
%_,°i
o"
• _
_._
/
1
_
.
_ . [+
,2-2o
+_ .
? /IBasin
'
,_Jl,ooe
?@'X: '°
+
u,,du
\
+......
R21W4
Carbonate edge
10 m porosity
Scale: 1:50,000
contour
FIG. 1. Shotpoint map of the Lousana survey showing seismic lines EKW-00I and
EKW-002, cross-section A-A', and well control. Carbonate edge and 10 m porosity
contours arc approximate only.
DEVONIAN STRATIGRAPHY
WABAMUN
GROUP
WINTERBURN
GROUP
WABAMUN
GRAMINIA2_
CALMAR
NISKU
Z
"'
a:
I,iJ
a.
a.
WOODBEND
GROUP
IRETON
°
LED
DUVERNAY
COOKING LAKE
BEAVERHILL
LAKE GROUP
-
WATERWAYS
SWANHILLS _
FIG.2. Stratigraphic
7-4
chart of the Upper Devonian
succession in central Alberta.
(after Stoakes, 1979)
CREWES Research Report Volume 6 (1994)
P-PandP-SVseismicdatafromLousana
36-21W4, have about 25 m of porosity and 10 m of pay. The 16-19 well went on
production in 1960 and has produced about 78,000 m3 of oil. The 2-30 well has
produced over 58,000 m3 of oil since it went on production in 1962. The porosity is
primarily vuggy, ranging to intergranular, and averages about 10% in the two
producing wells. Dry holes in the field are either tight or wet carbonate, or massive
anhydrite at the Nisku level. There is also gas production from the Viking and other
Lower Cretaceous formations in this area.
Geological cross-section A-A', shown in Figure 3, roughly parallels line EKW002, and is built from well logs from each of the shelf, basin, and reef environments.
DATA ACQUISITION
AND PROCESSING
Data acquisition and processing are described in greater detail by Miller et al.
(1993). The field acquisition parameters are summarized in Table 1. For both lines the
geophones were oriented at 45 degrees to the line direction, giving comparable amounts
of shear energy on H1 and H2. The H1 and H2 components were rotated into radial
and transverse directions. Vertical-, radial-, and transverse-component
gathers are
shown in Figures 4, 5, and 6 respectively for one sourcepoint from each of the two
lines. A time-variant gain function followed by individual trace-balance scaling has
been applied to these field records. The radial-component records are seen to have
good signal strength, with events that roughly correspond to those on the vertical
component. There is little signal evident on the transverse component.
Table 1. Field acquisition and recording parameters for the Lousana survey.
Energysource
Source pattern, EKW-001
Source pattern, EKW-002
Amplifiertype
Numberof channels
Samplerate
Recordingfilter
Notchfilter
Geophones per group
Type of geophones used
Number of groups recorded
Groupinterval
Normal source interval
NominalCMPfold
Spread
dynamite
single hole, 2 kg at 18 m
4 holes, 0.5 kg charges at 5 m
2 - SercelSN348
2 x 240
2 ms
Out-125Hz
Out
6 spread over 33 m
OYO 3-C, 10 Hz
110
33m
66 m
27
1799-16.5-SP-16.5-1799
m
The data were processed in 1993 on the IT&A system, as reported in Miller et
al. (1993). Subsequent to that report, tests for the removal of multiples were performed
using the parabolic radon transform method (Hampson, 1986) and predictive
deconvolution. The results for both methods were inconsistent, with some portions of
the data showing slight improvement while others were degraded.
The data have since been fully reprocessed on the ProMAX system using
similar flows to the IT&A processing for both the vertical and radial components.
Vertical and radial component flows are outlined in Figures 7 and 8 respectively. The
final migrated stacked sections are shown in Figures 9-12.
CREWESResearchReport Volume6 (1994)
7-5
Miller, Harrison, Lawton, Stewart and Szata
II
:::::::::
'_':::::::
'_'_: :::"_
-4>-
!
_
-...
:I
II" i
z
_<_
=,''_'=
,, ,,.o
!
!
If!
_ _°°_=_
i= ._
&
ii
I
'lliliilillilliiillilliililliil
IIII
'
IIIIIIIIl"_lllll
IIIl'_llllllll
,.I
,,,.I
i
IIII
IIIIIIIIlillll
li_iliiillliiiliiiiiiiiiiiiiii
i
!
i
I
IIII
_
_
I
iitllllllll_lllllllltl_lllllll!
I
O_i
I:_ 0
,-, _
_ ,_
'_
._
.
•
::::::::::::::::::::::::_
___
li
_i
_
_r_
_ '_ _ I:_
........
i4O.i
._
_'
_
........
_l
_
I
I_tlllllllll_lllll
0
_
"i_':-_-!
r_
i_:
_
o
_i
_
_
,, i,l,i
7-6
" '=
.......' .............
i
i
CREWES Research Report Volume 6 (1994)
_ _
_ ._
P-P and P-SV seismic data from Lousana
I
21
41
GI
81
I01
121
141
181
181
201
0,00
0.215
0.50
0.75
LO0
1.25
N
1.75
2.76
3,00
FIG. 4. Vertical-component
records from line EKW-00I,
line EKW-002, shotpoint 185.5 (right).
1
il
41
61
81
101
_
I_.1
shotpoint
Im
181
181.5 (left) and
_1
0.01
o.'1!
o_1
0.71
I.OO
1.21
1,7!
UI
2_,
2.21
2,71
3.01
3,2,!
3.11
!11
FIG. 5. Radial-component
records from line EKW-001,
EKW-002, shotpoint 185.5 (right).
shotpoint
CREWES Research Report Volume6 (1994)
181.5 (left) and line
7-7
Miller, Harrison, Lawton, Stewart and Szata
I
2[
41
61
_
1Ol
I_
141
161
NI
201
0.I0
0J5
0A0
0.15
1,00
1.2S
1.80
A
UJ 2.,o
m
_"
2.15
2.10
2.15
3.10
8.10
4,10
FIG. 6. Transverse-component
records from line EKW-001,
line EKW-002, shotpoint 185.5 (right).
7-8
shotpoint
CREWES Research Report Volume6 (1994)
181.5 (left) and
P-P and P-SV seismic data from Lousana
t2 spreading
gain
Demultiplex
I
Surface-consistent
deconvolution
Source, receiver & CDP components
80 ms operator, 0.1% prewhitening
Elevation & refraction statics
Initial velocity analysis
I
I
Automatic surface-consistent
statics
I
700-1800 ms window, maximum 20 ms shift
F
Normal moveout removal
Velocity analysis
I
Zero-phase spectral whitening
4/8 - 100/110 Hz
500 ms AGC
by CMP,
offsets
0-2640 m
I Stack Apply
mute
function
I
Zero-phase spectral whitening
4/8 - 100/110 Hz
500 ms AGC
I
Phase-shift
f-x predictionmigration
filter
Trace equilization
10/14-70/80 Hz bandpass filter
FIG. 7. Processing
flow for vertical-component
data.
CREWES Research Report Volume 6 (1994)
7-9
Miller, Harrison, Lawton, Stewart and Szata
I 45 degDemultiplex
geophone rotation ]
I
I
Velocity analysis
Normal moveout removal
I
I
Reversepolarity of leadingspread
Zero-phasespectralwhitening
I
Apply final P-P static solution
700 ms AGC
Applyreceiver-stackstatics
6/10-45/55 Hz
[
Initial velocity analysis
I
I Compute ACP trim statics
equilization
t Trace
z.5spreading
gain ]
Offsets33-2640
Depth-variant
CCPmstack [
Normal moveout removal
Shot-mode f-k filter
Restore normal moveout
Surface-consistent
Apply mute function
Zero-phase spectral whitening
2/6-50/60 Hz
700 ms AGC
deconvolution
] f-x prediction
filter
]
Source,receiver& CDPcomponents
120 ms operator, 0.1%prewhitening
Phase-shift migration
Modified velocity function
Asymptotic-conversion-point binning ]
Vp/Vs value of 2.29
I
1100-2300 ms window, maximum 20 ms shift
I Automatic surface-consistent
statics
I
FIG. 8. Processing
7-10
flow for radial-component
] 4/8-40/50
I
Hz bandpass filter
Traceequilization
I
data.
CREWES Research Report Volume 6 (1994)
]
I
P-PandP-SVseismicdatafromLousana
sw
NE
l--
FIG. 9. Migrated P-P stacked section for line EKW-001 after reprocessing.
NW
16"19
12"20
se
ID|
E
I=
FIG. 10. Migrated P-P stacked section for line EKW-002 after reprocessing. The wells
shown are used in correlation and modelling.
CREWESResearchReport Volume6 (1994)
7-11
Miller, Harrison, Lawton, Stewart, and Szata
$W
N:E
i=
FIG. l 1. Migrated
,M_ _
P-SV stacked section for line EKW-O01 after reprocessing.
16-19
12-20
se
IQZ
20a
xal
I--
a,**l
_na
2_m
2z_u
2,L01
FIG. 12. Migrated P-SV stacked section for line EKW-002
wells shown are used in correlation and modelling.
7-12
after reprocessing.
CREWES Research Rel_ort Volume 6 (1994)
The
P-PandP-SVseismicdatafromLousana
For both components, the most recently processed data has significantly
broader bandwidth. There is also improved imaging of some of the reflectors, most
notably the Wabamun at about 1775 ms. Figure 13 shows a comparison of 1993 and
1994 processing on the radial component of line EKW-002.
19B
EKW-_2 Radial18B3
i_
_BG
i|_
LT_
2 ZTSZ_ ZTZZ?O r _4 Z_tZSSZSSZ]
1_
Z9
_P
IgS
CP Z_rZ_ Zqtzqe
LgO
EKW-_P_2
Ri_zl I_B4
iJ5
iBO
Z_ zn 3_
iTS
L_
FIG. 13. Comparison of the radial component stacked section from line EKW-002
processed in 1993 (left) to the same data reprocessed in 1994 (right). In addition to the
broader bandwidth, note the improved imaging of the Wabamun horizon, the peak at
about 1775 ms. The peak at 1900 ms is the Nisku event.
MODELLING
Correlation
of P-P and P-SV
AND INTERPRETATION
sections
Correlation between P-P and P-SV events was initially done on the IT&Aprocessed data, and did not need to be repeated for the data reprocessed on ProMAX.
Thus, the figures in this section show data from the 1993 processing, but the technique
applies to any data set.
Horizons were interpreted on the vertical (P-P) component in the conventional
manner of matching a synthetic seismogram created from a sonic log to the seismic
data. The data shown use the polarity convention that a peak represents a positive
increase in acoustic impedance. The P-wave offset synthetic seismogram, from 16-1936-21W4, matches the seismic very well through most of the section (Figure 14).
However, the Wabamun, a high amplitude peak on the synthetic seismogram, is poorly
imaged on the actual data. This is mostly likely due to interference from short path
interbed multiples, probably originating in the coal beds of the Mannville Formation.
This multiple energy also appears to interfere with the primary reflections from deeper
in the section. A mistie occurs at the Nisku, causing uncertainty in the seismic pick for
this horizon. Problems with the data quality below the Mississippian.motivated the
reprocessing of the data. The Wabamun event was improved on both components,
particularly the P-SV data, but the Nisku mistie is still unresolved.
CREWESResearchReport Volume6 (1994)
7-13
Miller,Harrison,Lawton,StewartandSzata
There are no full-waveform sonic logs or 3-C VSPs available from this area to
serve as control for the P-SV, or radial component, interpretation. Therefore, offset PP and P-SV synthetic seismograms were created from the 16-19 well, using the offset
synthetic modelling program described by Lawton et al. (1992). Ricker wavelets were
used, with a central frequency of 30 Hz for the P-P seismogram and 15 Hz for the PSV seismogram. The offsets are from 0 to 1650 m, with a group interval of 66 m.
Normal moveout corrections and mutes were applied prior to stacking the offset traces.
For the initial correlation, the P-SV seismogram was calculated using the 16-19 sonic
curve for the P-wave velocities and assuming a constant Vp/Vs of 2.00.
Our polarity convention is that a peak on the P-SV section represents an
increase in acoustic impedance, as do peaks on the P-P data. Because both synthetic
seismograms were created from the same depth model, the correlation between the
offset synthetic stacks was straightforward. The P-SV synthetic seismogram was then
used to identify the events on the P-SV seismic section (Figure 15).
Vp/Vs
through
offset synthetic
modelling
Figure 15 indicates a time-variant shift between the events on the P-SV
synthetic stack and the P-SV seismic data. This occurs because the synthetic model was
created using a constant Vp/Vs value, whereas Vp/Vs varies with depth in the earth.
The interval Vp/Vs can be adjusted to stretch or squeeze the synthetic stack in a timevariant manner such that the events line up with those on the data. To determine the
correct interval Vp/Vs, a suite of P-SV offset synthetic stacks was generated using
constant Vp/Vs values ranging from 1.60 to 2.20 in steps of 0.05 (Figure 16). Vp/Vs
was taken from the synthetic stack which best fit the stacked section across a given
interval. From this analysis, an S-wave transit time curve was generated from the Pwave sonic curve and the depth-variant Vp/Vs values. The final P-SV offset synthetic
seismogram was generated using both P- and S-wave sonic curves and properly ties the
P-SV data (Figure 17). The resultant interval Vp/Vs is interesting in that it is quite low
in the Nisku to Cooking Lake interval, although the mistie at the Nisku horizon calls
into question the validity of this result. Future plans include repeating this analysis on
the newly reprocessed data.
Forward
P-P and P-SV
modelling
Well log curves were modified to simulate a variety of geologic conditions in
the Nisku formation. The base P-wave sonic curve was from the 12-20-36-24W4 well
which is on line EKW-002 (S.P. 175) and matched the P-wave seismic data reasonably
well. The 12-20 well is in the basin with anhydrite at the Nisku level. This well only
penetrated to the Ireton Fm, and so a composite curve was made using the 16-19-3624W5 sonic log from the Ireton to the base at the Beaverhill Lake Fm. Since there was
no density log available, the density log from 8-20-36-24W5 was modified to match the
12-20 depths. The values in Table 2 were used in the Nisku Formation to simulate
anhydrite, tight limestone, tight dolomite, and dolomite with 10% porosity, both oilfilled and gas-filled. The values were obtained from well logs from the Lousana field,
the Mobil Davey well at 3-13-34-29W4 (see Miller, 1992), Schlumberger (1989), and
petrophysical relationships such as the time average equation (Wyllie et al., 1956). The
Nisku porosity is primarily vuggy, so Roberston's Kuster-Toks6z models (1987) for
rounded pores in dolomite were used to estimate the variation in Vp/Vs with porosity
and pore fluid.
7-14
CREWESResearchReport Volume6 (1994)
P-P and P-SV seismic data from Lousana
P.P OffSet
Olfllel(m)
0
seismogram
528
1858
Shot point s
L91 tgB 186
...............
1584
..........
.
......
308 _,;;
_.,,..
....
.......................
ggg
, 1111
................
1
I
.....
_
. ...
m_
•
i
)_
e_
.....................
• •
._..m._.t_.
: ;; ;..:;;......
"_.lllll;;;;;;;::lT;
....
)..
......
....
"..
_ ..._ ............
148 146 143 141"
...........
;" 280
......
)1..l**....iph.i
_1_'_" ....
- _ _8o
.L.'.................
::_
.......
_ .'_.,
',: :_"_.',, ,_
.....
l,pm...I._b
;51
_ .......
Db b.
._.,._.,.._.;
_ ..........
...
..,_.....
o. ,._ _.
.I _..._
_,._._;_;,, ,,:.... ,...
,.:,
,l_;;_-t)l)i"'di;ilp._l_L_i;_i;i,_'li_l..*,,,,,t_-Illli-,,,,,,,",;,
NIII_I"III_I)MIil
, "_
I
l
O_
I
'
r
- :
ii;;;;i;;il)i:::_":::::
:_'-__ L_;t|il_i_l
_
[ : :: :: ::::_:
::: ::: : :::::::'_':'_'.':
:::
:::::2
-
-
"
,
.;
.......
1
1;llI¼11;lliiilil li,,lililiPillPill_illlll|l|li_il_i_illllhll,_tlH|llllli,l|tll,h_fff_
t;_o
'"'
tl,,lllllllllli,,_llllli_lilltllllu)litmmmt_ttll,llill_lm_m
--
N,s_u :=::=::::"'""'"'''
"
.....................
1380
I
;81 15B 156 153
, ".*_'_.0bo,*'lb,,_,'"."
.........................
lOgO-
Wa0_r1"l%_1
•
. ....... ,..._, ,,.,_..... ...... •....
*,:_llllIIlllllllllllllh
:
L/! 168 186 863
• .,_,_.,.
.
)).).
.._ ...... .....:
.......
2rid White S_ck_
173
i._...._..m
.N_.Jl41-_.
soo
808
)k...
tO1 175 |78
..,..
--_...
;;;. -_
.......
_._F,_r_oa;;
......... _"""
-..
.; ! _ ......_ ..............................
;.: _
_.;__..',....
.,.
..............
.I,_,_-_-
........
;
m_--
183
I
• _1_ .............
;i_'_ 't:t
................
:: ;'_';::;
''lll;_l
' .........
rrr _ _- _ r .......
"l
'"lil"_"lililllmltllllll''_'tiltiilllltllliil''''';i_'illlllililllliiil
"
1380
' ",
il
11gO
EKW-002
1260
,,llii:i*lliliilIMllli'l'l_,,
Vertical
FIG. 14. Correlation of P-P offset synthetic seismogram and synthetic stack with P-P
seismic data. The synthetic is generated from the 16-19 well, and spliced into line
EKW-002 at S.P. 191. The tie is good except for the Wabamun and the mistie at the
Nisku. The seismic data were processed using IT&A software.
Shotpoint_
Constant Vii/Vii2.0
,,,l,,,_
......
,,i,,
.....
191 lOS
,llI_l,
Hr,
.................
" ""
t1tl
183
181
.._.._.?l:l.rlii-,..
_
_
iillillllilli
5oo -,.._H.;m--..,,-
186
llli)illl>ill)llll
';.:.]:;]';;:7]:::
_;i;;;;_;71;;/;_3-_
.....
_,,i_,,.,,,,_,,+,,,,,,
,,,,i
*i_,
,,.
Hllllllll_:llllltllllll,-..llll.
8o8
2r!<lW_lsSpeckli:
IIII
"
:2 ::"?:::"
"l)l))lt))l,)lltlltl)lll
VlMng gO0
_i-i
I))'
I I
tit
'1)11
:_/_
i i i i i i ) ) ill)
'
.....
III
....
IIII
lt)l*))lt))ll
:
'
"
:
+"
,_i
lilll
:.....................................
:
Cookinl
...............
;:+
i
ili_
I ....
[
,_",,,,
''
."
16_ tO1
,..,_*,.i.._.-;;,;',:"
.--.
,,._
i._,_,r
:_l'_,'t.
"
"
i_
708
IW
IH_,,
t
Jill41111|il
I fill
.....
t
1100
._J.IIP 1200
,ll
'
_l_i||
I_*+
._,,_lill
,,;,*,
III
tO08
" "'"'
""'
7;_ %: :_. _.;i
I*.,,,,'
';, ....
lliil
"_
---_.
'
•
"
._*_" ,w, ....
,...tll.
'>
"
_,
"-lll_
,,.
"
_'_'"
llll|
l I, ........
,.
"wr_
|1))
._,,,,_m.,,.m,,,,,'
Im' ............................
wai_i_un
tll'''l
_|iPlmlmlllmlllilllll_--_/;,
"
-,
_,.____._.
108
'
,,'%
lllllll,
iii+
.,llllllllillll|lll|l)h
'"""_:::::::::L::L_
." _"
. al_i,
II i
,,,l,,,*tl,l_Ul
.
;68
"l'_'_rllll'-i"l_ll[i_.>"_l_"_-;''_C"r_<"_Ll'll;;_lli_
i'll--"_lil'l_'lt_
780
171
_I,,_,,_,_,,*,_
:_"LI I'i-_'ti*ltllllli_ltl'::::_''")"
lt_l, ti_, .l t I _ _, I ._ll_, i llillllll,
....
173
...... ,
.
_"._t,_...'-.,._--:
,lt_,,'l,,lll,_.
680
178 176
'
.....
"
v
--
r
"
--
"lllmi1480
; I":
"7.':
'
i .......
I
.....
,*lltli__i4_.'flt)l_lllilip _lltPIt'ti' t,,I;;i IJ_!It-i;lth i_; i88,
+IMP
ltli_llt
I
......
ttlll;
.....
lll,lllil_'lt,,llllllllllllllli
,,,,_-llmllm_tlil_liulliiiliiiliiftilil_
I
I
IIi
i
,_88
.......
,III
;
"
IJIke
130R
..............
"
_i
'_'
...................
_t_ ii,
. II
ll'll'tlr
. "
I,.,,,,.)**,,I,l_l_t)_)l)l
...........
EKW-002
ii ;ttP,I, wt
i
;;_,_'lllilliltlllLli_tI IIit 2_08
Radial
FIG. 15. The /'-/7 seismogram is first tied to the P-SV seismogram, which was
generated using the P-wave sonic log from 16-19 and a constant Vp/Vs of 2.0. It is
plotted at 2/3 the scale of the P-P seismogram. The P-SV seismogram is then tied to the
P-SV data at shotpoint 191. The correlation is straight forward, but there are time
misties due to the assumption of constant Vp/Vs. Processing was on IT&A's software.
CREWES Research Re_ort Volume 6 (1994)
7-15
Miller, Harrison, Lawton, Stewart and Szata
vp/vg
Offset (m)
0
m
1056
528
1584
')DPPJp_+_'_pl'
_ill
8_e )l/lltll'lll'l'lll'l'llll
,0
)+i!
'HDt
ItilL
'1'1'11
1.6
1.7
1.8
MIsslsslpptan
2.1
2,2
tttt
: liB,
,- ;;_5ill+l+
i ,
I
',k,.. ii P
----"-'--"
....
"'PPPPPPPPPPPPPHq' H'H';'_, PI P
'"
2.0
_
--- _'"""""'"''""'
i'i'
+...,.,.',.,,,i,,,,,
i,',i '.,,nnv,,,. _..... '--'"'/L..........
1.9
_
liD'
p
Inn,
_p
--
+_."
r__
)))l
++
.......
,+++.
coo+,+
fl????:
li/'
Wabamun
l_OgOI |,
I'_;tt'(t';l_ll;l('_
I,I,_
!
I
I_-
)
)
I
:,
r
P-P offset seismogram
I'1//
P-SV offset seismograms
FIG. 16. The P-P offset synthetic seismogram and stack is compared to a series of PSV offset synthetic stacks. P-SV stacks were generated using the P-wave sonic log
from 16-19 and constant Vp/Vs values ranging from 1.60 to 2.20 in steps of 0.05;
every other one is shown here. The seismogram stretches in time as Vp/Vs is increased.
P-S V offBe! lellmogrllm
191 _..=_J
1_6 t_3 l_l
v.,t._, v_v.
!.,li Jill
II
•
""
_l.l_q__Ik
"'._'"-.....
ill)
.'.-,,,,_ ,._;..+.,_'
..
, IlK _--q-I_._"::
,L .,
-_ .....
, d',
•.....
V_ldh(I
Co_nng _ko
FIG. 17.
determine
generate
resultant
the Nisku
7-16
- '"'
IIII
...............
"'"'"
+I III llllllllll)lI)lll
ili)[_"
...........
............
,_it, t,ll [.:_L ............
'
I
+_" ""kHk_--'"'+'';
_,_p)+ffmv , m,l_+|_,.
,.,
"
"_ i+_Hlllll)llN_l_l_i;)ll
'-- " ;! t,mt"_-._
_+ _2J'n'_E_l_el_-
7_
_,,...._,,._,.,._'._::__..,;kn_i_'_._._q_,,_-=ee
, ".....
r_r,"7 ....
.... ,._,'11,'ff.__
......
, " ,..+,," .,t;_,_,,
'-' -"k
.,._"+_."_
4,1-, I,_
))L_t*l)) k+It'.'rl_li.111,111111 + i_l__
•
343 141
,,:,-'._...,..._,..,.+-,::,
.;-_+...;::_,_:-:'-+_",,:;_ :',' _.;. ': :,, "",;+:'_'_.
•__-.-.... ;_:_:.._,.,,.,___.._..,
.., ,' ,'d..'_;.,.._._._...+__-.:,_-,u-._= so.
.'"_)_'
;E
ShotpoIntl
178 _76 173 171 168 166 t63 t6t 158 $5e t53 LSL :4_ l_
"_+_.' _lt'" "
+iL_m,...,
tE,,+,;iilt+.-..
',...-",
....
++'+" '
i*,
I
..k_'-."_- _
.,.L;Ptli+m_ip_l.lk._T;T,,._],vr_p_iTl
_ t
•
.,
;_ " , .-:_._II_
h,_,,,,
' ....
'
li_l_ll_+iit+l+li)lql_N_)bil''*b'l+'l'''i_tpllll_ll_IPlli''+.+_
+
,,,_r i.)Ih(:
"'."_..+L+-::3-'C_.
.......
,rFIT.., ""
_ ' _ _'_"_ 21_e
mllp---.--ww,m-_l, peel,
l,.,,,+L ..... :- +. ........
_l_'_P lk:''; -+._;,,,.,,'h,,,l,_h,,,,,l''_
_m ...... ,' .......
*_i,l@L"
-'-_)l
+_.ol+.l,
......
""";::_l_41111_$_Ll_rlPI)l
tYp_lq
Ilulllls_
I Llllllllllll_
:?_.a
EKW-Q02 Radllll
P-SV synthetic stacks from Figure 16 were compared to the P-SV data to
optimum intervalVp/Vs
values. The interval Vp/Vs values were used to
a new S-wave sonic curve for input to the P-SV synthetic seismogram. The
P-SV seismogram shown here ties the data correctly. In this case, the mistie at
(Figure 14) still causes some uncertainty in the results.
CREWES Research Report Volume 6 (19941
P-PandP-SVseismicdatafromLousana
Table 2. Rock property values used in P-P and P-SV offset synthetic models.
Anhydrite
Tightlimestone
Tightdolomite
Dolomite:10%
porosity - oil filled
Dolomite:10%
porosity - gas filled
A tp (_ts/ft)
[Its/m]
50 [164]
47 [154]
43 [141]
57 [187]
65 [213]
A ts (_ts/ft)
[_s/m]
1001328]
90 [295]
77 [253]
1001328]
Vp/Vs
2.00
1.90
1.80
1.75
density (g/cm 3)
[kg/m3]
2.96 [2960]
2.70 [2700]
2.85 [2850]
2.65 [2650]
100 [328]
1.55
2.56 [2560]
The Nisku interval is 162 ft (50 m) thick, from 5828 to 5990 ft (1776-1826 m).
Two porosity thicknesses were simulated: 75 ft (23 m) from 5900 to 5975 ft (17981821 m) as is found in the 16-19 well, and 125 ft (38 m) from 5850-5975 ft (17831821 m) where depths are relative to kelly bushing. The porosity was overlain by
anhydrite and underlain by tight dolomite within the Nisku. All the logs are identical
outside of the Nisku formation.
The P-P model consisted of 25 receivers with 66 m spacing, with a far offset of
1650 m. Because of phase changes at far offsets, the spread was reduced for the P-SV
model to 22 receivers at 66 m spacing for a far offset of 1452 m. A 40 Hz zero-phase
Ricker wavelet was used for the P-P model and a 25 Hz zero-phase Ricker for the PSV model to best match the reprocessed data.
The offset synthetic stacks are shown for the P-P case in Figure 18 and the PSV case in Figure 19. Both components exhibit changes at the Nisku level with
changes in lithology, porosity, and pore fluid. When porosity is introduced, a trough
followed by a peak develops below the peak at the top of the Nisku, indicating porosity
and the base of porosity respectively. The amplitude of this lower peak brightens either
by thickening the porosity from 23 m to 38 m or by the replacement of oil with gas.
The modelled P-P response is similar for anhydrite and tight limestone: a tight doublet.
A slight difference in the character of the doublet is evident on the P-SV model. In tight
dolomite a single broad peak/trough pair is evident on both models.
Horizons were picked on both models and P-P and P-SV isochrons used to
calculate Vp/Vs values across several time intervals for each stack. For converted-wave
data, Vp/Vs is equivalent to (2Is/Ip) -1, where Is and Ip are the P-SV and P-P interval
transit times respectively (Garotta, 1987). Horizons were chosen which bracketed the
Nisku and can be readily identified on the real data. Figures 20 and 21 show Vp/Vs
variations across six intervals with the following thicknesses:
Mississippian- CookingLake:
1906ft
581 m
Wabamun- CookingLake:
1484ft
452 m
Wabamun salt marker - Cooking Lake:
1076 ft
328 m
Mississippian- Ireton trough:
1200ft
366 m
Wabamun- Iretontrough:
778ft
237m
Wabamun salt marker - Ireton trough:
370 ft
113 m
There are several observations to note from these plots. Vp/Vs varies with
lithology, porosity, and pore fluid, with the degree of variation diminishing as the
thickness of the interval increases. The general trend is for Vp/Vs to decrease to the
right as the model changes from nonreservoir rocks to reservoir rocks. Thickening the
porous zone from 23 m to 38 m decreases Vp/Vs more than replacing the oil with gas.
CREWESResearchReport Volume6 (1994)
7-17
Miller, Harrison, Lawton, Stewart and Szata
(h)
Ck¢
FIG. 18. P-P model results showing (a) offset synthetic seismogram for anhydrite, and
offset synthetic stacks for (b) anhydrite; (c) tight limestone; (d) tight dolomite; (e) 23 m
oil-filled porous dolomite; (f) 38 m oil-filled porous dolomite; (g) 23 m gas-filled
porous dolomite; (h) 38 m gas-filled porous dolomite.
Miss'
Wab,
Salt.
Nisku
Ireton
FIG. 19. P-SV model results for the same geological
7-18
settings as in Figure 18.
CREWES Research Report Volume 6 (1994)
P-P and P-SV seismic data from Lousana
1.84
1.82
1.76.
1.74,
-&
o
I
(b)
.
Miss - Ckg Lk m,
Wab - Ckg Lk
Salt - Ckg Lk
I
• I
•
(c)
(d)
(e)
(f)
(g)
(h)
FIG. 20. Plot of Vp/Vs measured across intervals on the model data in Figures 18 &
19. The lower horizon is the Cooking Lake (Ckg Lk) and the upper horizons are the
Mississippian (Miss), Wabamun (Wab), and Wabamun salt marker (Salt). The Nisku
simulates (b) anhydrite; (c) tight limestone; (d) tight dolomite; (e) 23 m oil-filled porous
dolomite; (f) 38 m oil-filled porous dolomite; (g) 23 m gas-filled porous dolomite; (h)
38 m gas-filled porous dolomite.
,
2.00.
1.95.
-,t
Miss
- Ireton
Wab - Ireton
o
Salt- Ireton
[
.
1.90..
\
_
1.70'
1.65'
1.60'
(b)
(c)
(d)
(e)
(f)
(g)
(h)
FIG. 21. Plot of Vp/Vs measured across three more intervals on the model data with
the Ireton as the lower horizon. The upper horizons and the geological settings are the
same as in Figure 20.
CREWES Research Report Volume 6 (1994)
7-19
Miller, Harrison, Lawton, Stewart and Szata
If one of the bounding horizons is too close to the zone of interest, wavelet
interference from that zone may affect the picks and cause unexpected results. This is
probably the case for the sharp decrease in Vp/Vs which occurs for tight dolomite when
the bottom horizon is the Ireton, which underlies the Nisku. This may also be the cause
of the increase in Vp/Vs across the Wabamun salt to Cooking Lake interval when gas
replaces oil with 23 m of porosity. Using this salt as the top marker and the Ireton as
the base marker produces the thinnest interval and thus the largest variations in Vp/Vs,
however the values may be in error due to wavelet interference.
Our experience is that changes of 0.05 in Vp/Vs are observable on seismic
sections. The intervals which use the Ireton as the lower bounding horizon show at
least this much change between the anhydrite and 23 m of porous dolomite (as occurs
in the oil wells). This suggests that the difference may be detectable on seismic data,
although possible wavelet interference from the Nisku should be considered. The
greatest contrast for all intervals occurs between tight anhydrite or limestone and the 38
m thick gas-filled porous dolomite. This effect might be observable even in oil plays if
the gas to oil ratio is high enough. Gas saturation as low as 5-10% will result in a sharp
decrease in Vp/Vs (Gregory, 1977). The modelling results indicate that Vp/Vs analysis
is applicable in carbonate settings, even when the reservoir intervals are relatively thin
(23 m) and comprise only 15-20% of the total measured interval.
Workstation
interpretation
The reprocessed migrated sections of both components were loaded onto the
Landmark workstation and the correlated events interpreted on both lines. Interpretation
of the new data is in progress, but P-P and P-SV interpretations to date for line EKW002 are shown in Figure 22. The P-SV data are plotted at 2/3 the scale of the P-P data,
and the excellent tie between the two components is evident.
Isochrons and Vp/Vs ratios were calculated for several intervals between the
interpreted horizons. If the event correlations between the components are accurate, the
dimensionless ratio will be free of effects due to depth or thickness variations, as these
will affect both components equally. Variations may be caused by factors such as
changes in lithology, porosity, pore fluid, and other formation characteristics (Tatham
and McCormack, 1991 ).
Figure 23 shows Vp/Vs along line EKW-002 for two intervals. The top curve is
the interval Vp/Vs for the portion of the Cretaceous section from the Second White
Specks to the Mannville. The high Vp/Vs values from about 2.2 to 2.4 are consistent
with a shaley section. By contrast, the Vp/Vs values from an interval within the
Paleozoic, the Mississippian to the Nisku, average about 1.8, a reasonable value for
carbonate rocks. In the Paleozoic interval, Vp/Vs decreases to about 1.6 between
shotpoints 182 and 202. Examination of the isochrons shows a thickening of the
isochron on the P-P data and a thinning on the P-SV data, both contributing to the
overall drop in Vp/Vs. The 16-19 well at shotpoint 191 is located on a structural high,
suggesting that the P-SV time structure is consistent with the geological structure.
Pushdown of the Nisku horizon on the P-P section may be a velocity effect caused by
features in the shallower part of the section. This Vp/Vs anomaly occurs in the section
immediately above, but not including, the Nisku reservoir. With further analysis we
hope to determine if it is related to the dolomitic buildup in the underlying section.
To date, we have been unable to measure Vp/Vs across an interval which
includes the Nisku reservoir. Several of the horizons which bound the Nisku have been
damaged by multiple interference and cannot be picked with the level of certainty
required for meaningful results. We are still working with the data
7-20
CREWESResearchReport Volume6 (1994)
P-P and P-SVseismic data from Lousana
FIG. 22. Interpretation of the P-P data (left) and P-SV data (right) is shown for part of
line EKW-002. To facilitate interpretation, the sections are displayed on the workstation
simultaneously and the P-SV data plotted at 2/3 the scale of the P-P data.
3.0
2.8-
Cretaceous
2.62.42.2(O
i
>
2.01.81.61.4-
Paleozoic
1.2-
1.0O J O J < N C \ J O J C M C \ J C \ J C S J C \ J C S J C \ J ( N C J O J
Shotpoint
FIG. 23. Vp/Vs values on line EKW-002 for two intervals: Second White Specks to
Mannville in the Cretaceous section and Mississippian to Nisku in the Paleozoic. This
Cretaceous interval is primarily shales, whereas the Paleozoic consists mainly of
carbonate rocks. The difference in lithology is reflected in Vp/Vs.
CREWES Research Report Volume 6 (1994)
7-21
Miller, Harrison, Lawton, Stewart and Szata
in an attempt to use horizons which bound the Nisku and can be picked with a
reasonable level of confidence.
The advantages of having two seismic data sets to work with became evident
during the course of this study. It was useful to display both components on the screen
simultaneously while picking horizons, especially in zones where the horizons were
difficult to track. In the case of the Wabamun and Nisku horizons, the pick was clearer
on the P-SV data than on the P-P data. Also, the calculated Vp/Vs values were used as
a quality control check on the horizon interpretations. Vp/Vs values which were outside
the range expected due to geological variations indicated mispicks and the need to
revisit the data. For example, on both components there were two possible picks for the
Wabamun event. Choosing the wrong cycle on either section resulted in unreasonable
Vp/Vs values for the intervals above and below the pick. Vp/Vs provided a useful
constraint on the interpretation of this and other horizons. By providing the interpreter
with feedback on the validity of horizon picks, Vp/Vs can increase confidence in the
final interpretation.
SUMMARY
Two orthogonal three-component seismic lines were shot by Unocal in January,
1987, over the Nisku Lousana Field in central Alberta. The purpose of the survey was
to investigate a Nisku patch reef thought to be separated from the Nisku shelf to the east
by an anhydrite basin. These data have been reprocessed and are currently being
analyzed and used to develop methods of multicomponent seismic data analysis.
Several techniques which are useful in the analysis of multicomponent data are
described and applied in this study. The data are of overall good quality and major
events were confidently correlated between the P-P and P-SV sections using offset
synthetic seismograms. P-SV offset synthetic modelling was also used to extract
interval Vp/Vs values from the P-SV data.
Forward offset synthetic modelling showed character, isochron, and Vp/Vs
variations associated with changes in lithology, porosity, porosity thickness, and pore
fluid. The modelling results indicate that multicomponent seismic analysis is applicable
in carbonate settings, even when the reservoir intervals are relatively thin (23 m) and
comprise only 15-20% of the total measured interval. To date, we have had difficulty
applying these results to this data set because of multiple contamination in the zone of
interest, however work is ongoing to resolve this issue.
Multicomponent recording provides additional seismic measurements of the
subsurface to assist in developing an accurate geological model. Rock properties which
can be extracted from elastic-wave data, such as Vp/Vs, reduce the uncertainty in
predictions about mineralogy, porosity, and reservoir fluid type. Joint interpretation of
P-P and P-SV data was helpful in horizon picking and in providing feedback on the
validity of the interpretation. The interpretation techniques described in this paper can
be usefully applied to other areas, including other carbonate plays.
FUTURE
WORK
The results of forward modelling encourage us to continue working with this
data set. We hope to obtain further results from interpretation methods described in this
paper, such as interval Vp/Vs analysis. Multiple contamination has complicated the
application of these methods, so we plan to pursue full-waveform modelling to try to
7-22
CREWESResearchReport Volume6 (1994)
P-P and P-SV seismic data from Lousana
understand
the effects of interbed multiple energy on both components.
Continued
AVO modelling
and AVO inversion may help us further understand
the effects of
mineralogy,
porosity,
and pore fluid in carbonate
settings.
In keeping
with our
objective to develop multicomponent
interpretation
techniques,
we plan to enhance the
offset synthetic
modelling program as a tool for extracting interval Vp/Vs values.
ACKNOWLEDGMENTS
We would like to acknowledge
Unocal Canada Ltd. for donating
the seismic
data to the CREWES
Project. We are grateful to Chevron
Canada Resources
and
Sensor Geophysical
Ltd. for participating
in this study, particularly
for the time and
effort spent by Mark Harrison on reprocessing
the data at the University
of Calgary.
Our thanks to the companies
which donated software used in this work: Landmark,
GMA, and Hampson-Russell.
Thank you to Andrea Bell of Norcen Energy Resources
Ltd. for useful discussions
on the geology of the Lousana Field. We thank the
CREWES
Sponsors for their financial support.
REFERENCES
Andrichuk, J.M. and Wonfor, J.S., 1954, Late Devonian geologic history in Stettler area, Alberta,
Canada: AAPG Bull., 38, 2500-2536.
Domenico, S.N., 1984, Rock lithology and porosity determination from shear and compressional wave
velocity: Geophysics, 49, 1188-1195.
Eastwood, R.L. and Castagna, J.P., 1983, Basis for interpretation of Vp/Vs ratios in complex
lithologies: Soc. Prof. Well Log Analysts 24th Annual Logging Symp.
Garotta, R., 1987, Two-component
acquisition as a routine procedure, in Danbom, S.H., and
Domenico, S.N., Eds., Shear-wave exploration:
Soc. Expl. Geophys., Geophysical
development series 1,122-136.
Georgi, D.T., Heavysege, R.G., Chen, S.T., and Eriksen, E.A., 1989, Application of shear and
compressional transit-time data to cased hole carbonate reservoir evaluation: Can. Well
Logging Soc. 12th Formation Evaluation Symp.
Goldberg, D. and Gant, W.T., 1988, Shear-wave processing of sonic log waveforms in a limestone
reservoir: Geophysics, 53,668-676.
Gregory, A.R., 1977, Aspects of rock physics from laboratory and log data that are important to
seismic interpretation, in Seismic Stratigraphy - Applications to Hydrocarbon Exploration:
AAPG Memoir 26, 15-45.
Hampson, D., 1986, Inverse velocity stacking for multiple elimination: Can. J. Expl. Geophys., 22,
44-45.
Kuster, G.T. and Toks6z, M.N., 1974, Velocity and attenuation of seismic waves in two-phase media:
Part 1. Theoretical formulations: Geophysics, 39, 587-606.
Lawton, D.C. and Howell, C.E., 1992, P-SV and P-P synthetic stacks: Presented at the 62th Annual
SEG Meeting.
Miller, S.L.M. and Stewart, R.R., 1990, Effects of lithology, porosity and shaliness on P- and S-wave
velocities from sonic logs: Can. J. Expl. Geophys., 26, 94-103.
Miller, S.L.M., 1992, Well log analysis of Vp and Vs in carbonates: CREWES Research Report 4.
Miller, S.L.M., Harrison, M.P., Szata, K.J., and Stewart, R.R., 1993, Processing and preliminary
interpretation of multicomponent
seismic data from Lousana, Alberta: CREWES Research
Report 5.
Nations, J.F., 1974, Lithology and porosity from acoustic shear and compressional wave transit time
relationships: Soc. Prof. Well Log Analysts 15th Annual Symp.
Pardus, Y.C., Conner, J., Schuler, N.R., and Tatham, R.H., 1990, Vp/Vs and lithology in carbonate
rocks: A case study in the Scipio trend in southern Michigan: Presented at the 60th Annual
SEG Meeting.
Pickett, G.R., 1963, Acoustic character logs and their applications in formation evaluation: J.
Petr.Tech., June, 659-667.
CREWES Research Report
Volume 6 (1994)
7-23
Miller, Harrison, Lawton, Stewart and Szata
Rafavich,
F., Kendall, C.H.St.C., and Todd, T.P., 1984, The relationship between acoustic properties
and the petrographic character of carbonate rocks: Geophysics, 49, 1622-1636.
Rennie, W., Leyland, W., and Skuce, A., 1989, Winterburn (Nisku) Reservoirs in Anderson, N.L.,
Hills, L.V., and Cederwall, D.A., Eds. The CSEG/CSPG Geophysical Atlas of Western
Canadian Hydrocarbon Pools, 133-154.
Robertson, J.D., 1987, Carbonate porosity from SIP traveltime ratios: Geophysics, 52, 1346-1354.
Schlumberger, 1989, Log Interpretation Principles/Applications: Schlumberger Educational Services.
Stoakes, F.A., 1979, Sea level control of carbonate-shale deposition during progradational basinfilling: the Upper Devonian Duvernay and Ireton Formations of Alberta, Canada: Ph.D.
dissertation, Univ. of Calgary.
Stoakes, F.A., 1992, Winterburn megasequence in Wendte, J., Stoakes, F.A., Campbell, C.V.,
Devonian-Early Mississippian carbonates of the Western Canada sedimentary basin; a sequence
stratigraphic framework: SEPM short course No. 28,207-224.
Tatham, R.H., 1982, Vp/Vs and lithology: Geophysics, 47, 336-344.
Tatham,R.H. and McCormack, M.D., 1991, Multicomponent
Seismology in Petroleum Exploration:
Soc. Expl. Geophys., Investigations in geophysics no. 6.
Wang, Z., Hirsche, W.K., and Sedgwick, G., 1991, Seismic velocities in carbonate rocks: J. Can.
Petr. Tech., 30, 112-122.
Wilkens, R., Simmons, G., and Caruso, L., 1984, The ratio Vp/Vs as a discriminant of composition
for siliceous limestones: Geophysics, 49, 1850-1860.
Wyllie, M.R.J., Gregory, A.R., and Gardner, L.W., 1956, Elastic wave velocities in heterogeneous and
porous media: Geophysics, 21, 41-70.
7-24
CREWES Research Report
Volume 6 (1994)