3-D Modellinqof a Reef-FaultBlock Structure
3-D modelling
of a reef-fault
block structure
Darran J. Edwards
ABSTRACT
The study presented here revolves around a 3-D zero-offset survey shot over a
reef-fault block model using the physical modelling facilities at The University of
Calgary. Dimensions and velocities for the model are taken from Keg River patch reefs
or bioherms flanking the eastern Peace River Arch, northwestern Alberta. These reefs
developed typically on or above prominent basement structures, which are thought to
be fault-controlled in this area. 2-D seismic data from the Panny Field are available for
interpretation, and a schematic 'reef-fault block (horst)' geological cross-section is
presented here. Direct identification of such faulting on 2-D data is shown to be
difficult. Erosion of such uplifted fault blocks produced basal clastic (Granite Wash)
fault-scarp deposits in adjacent margins. The main objective of the Panny model is to
image such a clastic layer beneath the reef. In real data, identification of any regular
trends of these deposits and correlation with overlying bioherms would substantiate any
theory on the tectonic control of such build-ups.
Such a scenario is used as the geological template for the physical model
construction. Physical modelling is described in terms of 3-D zero-offset acquisition,
processing and interpretation. A number of vertical and horizontal sections are
displayed. Timeslice investigation shows that the areal extent of the basal clastic fault
block is particularly well imaged, even beneath a large part of the reef. 3-I) numerical
modelling of a similar structure has also successfully delineated the regular margins of
this layer, thereby identifying fault positions.
Another aim of the model is to uncover interpretation pitfalls with 2-D data.
Through numerical modelling it can be shown that certain anomalies observed on
physical-model seismic sections are sideswipe reflections. Therefore, this study shows
that, in dealing with 2-D data reefs may be inferred along lines shot away from the
actual reef. This may frequently result in false imaging on 2-1) data.
INTRODUCTION
The Alberta Basin contains one of the best known Palaeozoic reef provinces in
the world, with the subsurface distribution of these carbonate formations mapped fairly
extensively. However, the mechanisms behind the initiation and development of such
reefs are still unclear. In other sedimentary basins, such as the Grand Banks,
Newfoundland (Tankard & Welsink, 1987), a close correlation has been observed
between basement structure and depositional patterns. A similar relationship between
carbonate build-up and basement elements in the Alberta Basin has been postulated by
numerous investigators (Sikabonyi & Rodgers, 1959; Mountjoy, 1980; and others).
Clearly, knowledge of the basement structure and identification of possible associated
faulting are major factors in determining whether any such relationship may be
postulated.
CREWESResearchReoort Volumo4 (1992J
4-1
Darran J. Edwards
_
5
I
NORTHWEST
TERRITORIES
/=la mv_ _M_t
_
I
:
/
!
!
\
/
\\
\
\
/
\\
/
\
115000 '
I
,
114o45 ,
_...!
•_'..
"
lid,,"
,
*
_'...L
;_
_,,_;,:.._;_,
tl,,
• * . _.-
L-"
° " '
" "
t..,"
1-3-_-6W$
+
i,'.
,_,,..
57°15'
_"
LinePa-1J,'.
R6W5M
T95
R5
I
O
.(p-
tch0t location
(ab_"idoned)
Well location
I
I
FIG. 1. Location of Panny (River) area wih 2-D seismic sections used in study.
4-2
CREWESResearchReoort Volume4 (1992)
3-D Modellinq of a Reef-Fault Block Structure
-115.40
-tl5,20
-115.00
I
q
57.70
I
I
-I
t4.50
-114.60
•
*
I
•
-114.40
I
i
-tt4.20
I
f
........
57.50
-1t4.00
•
I
_....
--T--"_:_---I
-113.80
57.70
=,\'_,_-_'-r"----'_._"<;-,-:--/,_C-"X_._
57.so
57,10
57.t0
5_,o
--115.40
/
I
/
f\'
--115.2Q
,,'-'.
I
,
--115.00
,
{:f...-T-,._L.___,.-"'_\_t_
• ', \_
-114.80
14.60
-114.40
--114,20
/
'_--115.8055_o
. well position
C.L20m
/
/
-114.95
57.40
P
_--114.00
-I 14.SrO
-I 14.85
-11'SO
FIG. 2. KB depth to Precambrian
(Latitude and longitude In degrees)
in the eastern Peace River Arch
(using available well data)
CREWES Research Reoort Volume 4 (19921
4-3
DarranJ. Edwards
:_oo
_tll
iJiii
i Lill
iii
lll,ii
iiii
Lil_ll
,
(after Cant, 1988);
note 90 degree 'block'
i
-I
I I I _=1"1 i [ i i i Ii i i i i [ i ; i i T [ ; i i i i r _]T_t
_0
6thMet,
20
tO
71
Tl_h _
Precambrian structure map
AG=axial graben
faulting (throws>4Om)
SP.¢¢LASTIC
_;_r;_
_.,,.,,-°"-
"_i_:,_,
"_'_-_--!i:::
.......
::::::::::"
:;il
Cross-section of the Keg River to Precambrian interval in the
Peace River Arch area as viewed from the northwest (Campbell., 1987).
Note location of Upper Keg River reefs over Precambrian horst.
FIG. 3. Geological setting for physical model: eastern Peace River Arch area
4-4
CREWES Research Report Volume 4 (1992)
3-D Modeltinqof a Reef-FaultBlockStructure
The study presented here revolves around a 3-D seismic model constructed
using the geology of the east Peace River Arch, northern Alberta as a template.
Velocities and depths from the Panny (River) Field were used from Anderson et al.
(1988). The Keg River Formation, where structurally draped or closed across
underlying Precambrian highs, is the principal reservoir facies in this area. Keg River
Formation patch reefs or bioherms developed typically on or above prominent basement
structures (Campbell, 1987). Basal clastic or Granite Wash deposits are frequently
associated with these Precambrian highs. Erosion of Peace River Arch uplifted faulted
blocks produced the coarse-grained clastic deposits in adjacent fault-bounded margins.
The primary purpose of the Panny River Model was to analyze the seismic
response of lower-velocity basal clastic deposits underlying Keg River patch reefs in an
attempt to delineate any basement faulting. Identification of regular trends or
orientations through detailed mapping of these deposits by timeslice investigation, and
correlation with overlying carbonate bioherms would substantiate the theory behind
tectonic development of such build-ups. Direct detection of such faulting on 2-D
seismic data is normally difficult because of the limited offset and vertical or subvertical nature of the basement displacement. Furthermore, this faulting is also thought
to have a strong horizontal, strike-slip component to it. Mitchell (1987) concludes
(from an oral presentation) that "3-D seismic has been an effective tool in locating
Precambrian highs and their trends (in this area), even when their seismic expression is
subtle". The model's secondary purpose was to uncover interpretation pitfalls in 2-D
data (particularly those associated with sideswipe), and to illustrate the advantages of
3-D seismic data.
The first section of this paper describes the field acquisition, processing and
interpretation of two 2-D seismic sections shot over part of the Panny Field. The two
sections intersect at a well drilled into the flank of a patch reef draped over one such
Precambrian high. The second section outlines the 3-D physical seismic modelling of a
'reef-fault block' at The University of Calgary in terms of model construction, 3-D data
acquisition, processing and interpretation. Finally, 3-D numerical modelling of a
corresponding structure is discussed, with particular attention given to the effects of
sideswipe.
GEOLOGICAL
FIELD SETTING
Since the discovery of oil at Amoco's 2/3-11-96-6W5
test well at Panny in
1983, the eastern Peace River Arch in northern Alberta (Figure 1) has been an active
exploration area. An estimated 55 million bbl (8.7 x 106 m3) of recoverable oil has
already been discovered. The Keg River Formation is the principal reservoir facies and
is typically productive where structurally closed across Precambrian highs. It is
commonly considered to be a structural trap as these sediments are productive only
where draped across underlying basement structures and the Upper Keg River Reef
Member, respectively. Although porosity and permeability within these latter two units
are stratigraphically controlled, the off-structure facies generally being tighter, these
effects are considered as secondary to the drape resulting from differential compaction.
The reservoir facies is capped by the basal anhydrite unit of the Muskeg Formation.
The Precambrian surface consists of anomalous structural highs prevalent
throughout the area, with relief typically ranging from 10 to 100m. Generally, such
CREWESResearchRenort Volume4 (19921
4-5
DarranJ. Edwards
structures have been mapped from the available seismic and well data as being areally
closed, a pattern consistent with the idea of a surface fractured by conjugate pairsets of
faults oriented in NW-SE and SW-NE directions. There is evidence, both geological
and geophysical (Anderson et al., 1988; Cant, 1988), of renewed faulting since Keg
River time, either on new surfaces or as reactivations of former faults. The first
documented episode of reactivated normal faulting affecting the Arch occurred during
deposition of the Elk Point Group in the mid-Devonian. Conglomeratic Granite Wash
detritus or basal elastics (and related elastics such as the Gilwood Formation) are
eroded from upfanlted blocks and interbedded with Elk Point sediments indicating
episodes of tectonism. It is thought that basement relief influenced the locations of
bioherms in onlapping carbonate units (Cant, 1988). The Slave Point and Keg River
reefs occur on upfaulted blocks near the eastern edge of the Arch, for example at Slave
Field (Dunham et al., 1983). A schematic cross-section of the east Peace River Arch
area is displayed in figure 3.
Panny River example:
2-D seismic
sections
The Keg River Formation is the principal reservoir facies of the Panny Field,
the basal Palaeozoic clastics being a secondary target. Figure 1 shows part of the Panny
Field with the approximate orientation of the two 2-D seismic lines, PA-1 and PA-6,
and wells used in this study. Wells 1-3-96-6W5 and 3-11-96-6W5
both produce
from the Keg River Formation.
The seismic data were acquired by Enertec Geophysical Services Ltd. in March
1984 using a 1000m - 60m * 60m - 1000m, 96-trace, spread, with source and
receiver intervals of 40m and 20m respectively. A hydrapulse source with 18 pops per
source point was used. Data acquired using hydrapulse sources typically have poor
penetration below 800-1000ms with an abundancy of lower frequencies. This would
seem to present a problem since it is the aim of this paper to image basement faulting,
and its subsequent effects.
The data were processed up to stack by Geo-X
following processing parameters:
Systems Ltd. using the
Demultiplex
Amplitude
Recovery
Phase
Compensation
Spiking
Deconvolution
Elevation
Initial
(80 ms operator, 1% prewhitening)
and Refraction
Velocity
Statics
Analysis
Statics (Surface Stack Residual)
Velocity
Normal
4-6
Analysis
Moveout
Mean Scaling
Mute
(400-1200ms)
Trace Gather
(24 fold)
CREWESResearchReport Volume4 (1992)
3-D Modellinq of a Reef-Fault Block Structure
0
4000m
I
!
S.P. 2801
3001
3201
0.5
-.I
PA-1
m
1.2
{projectedIocatlccl)
1-3-96-6W5
S,P. 3121
+
3-11.96-6W5
÷
2921 UnePA-1
2721
0.5
-t
m_
PA-6
1.2
FIG. 4.
Condensed
scale displays of lines PA-I
and PA-6
CREWES Research Reoort Volume 4 (1992)
4-7
Darran J. Edwards
S.P. 2921
2871
2821
sec
Beaverhill
Slave Point
m
Muskeg Fm
FIG. 5. WELL 03-11-096-06-W5
Synthetic Seismogram tie with line Pa-6
4-8
CREWES Research Report
Volume 4 (1992)
3-D Modellinqof a Reef-FaultBlock Structure
CDP
Cross-correlation
Stack
Post-stack processing was undertaken on ITA's INSIGHT software using the
following modules:
Karhunen-Loeve
FK Filter
Stack
(SNR enhancement)
(20-80 Hz)
GappedPredictive
Deconvolution
(operator lOOms, gap 50ms)
2-D interpretation
Figure 4 consists of condensed scale seismic displays of lines PA-1 and PA-6
in order to illustrate the subtle undulatory nature of the Precambrian basement on such
sections, together with the draping of overlying formations over topographic highs.
Well 3-11-96-5W5
is found at the point of intersection of these two lines and is used
to tie well depths to the seismic data. The synthetic and field data directly correlate at the
Precambrian, Keg River, Muskeg, Slave Point, and Beaverhill Lake events (Figure 5).
Throughout this paper, such expressions as 'the Keg River reflection event' refer to the
top of the named unit.
Well 1-3-96--6W5 is found approximately 0.5 km to the south of line PA--6 but
well depths may be projected to intersect it. The synthetic seismogram produced from
this well corresponds well with the events interpreted from the previous well. Campbell
(1987) describes the core from well 1-3-96_W5
at depths of 1224m to 1260m. The
Keg River in this cored sequence consists of a high-energy bioherm rooted directly on
top of Precambrian granite gneiss basement. Overlying the basement rock are a series
of siliciclastic sands, conglomerates and carbonate interbeds. Bioclastic gravelstromatoporoid sequences appear to form the bulk of the deposits associated with the
bioherm facies. Figures 6 and 7 show the seismic and schematic geological crosssections surrounding the well. The bioherm or patch reef is interpreted as overlying a
Precambrian basement high and adjacent basal clastic deposits. The Keg River
Formation is draped across these structural highs, as a result of differential compaction
of lower Palaeozoic sediments, producing a reservoir where structurally closed against
the flank of the underlying high. Vertical or near-vertical faulting has been interpreted in
the Precambrian hut with limited confidence. It would appear these die out below 1 sec
(possibly due to the poor resolution of the data). Nevertheless, faults have been
identified where the basal clastic unit, thought to be a fault-scarp deposit, is known to
exist through well data.
Such a 'reef-fauh block (horst)' scenario is used as the geological template for
the physical model construction. However, this situation is simplified in the model by
the fact that the Keg River interval is only represented by an elongate patch reef with
sloping flanks; whereas in reality the reefal facies and adjacent off-reef sediment are
normally of similar velocity. Indeed, sometimes the reefal facies is of slightly lower
velocity. Hence, the patch reef top and flanks are frequently difficult to image on
seismic data shot this area.
CREWESResearchReoort Volume4 (1992)
4-9
1-3-96-6W5
(projected
0
I
800 rn
_L
1p-
m
S.P. 3051
location)
3001
2961
0.75
Muskeg Fm
_¢
o_
KegRIL,
UrF_
Keg
Precambrian
1.00
m
-%
1.25
FIG. 6. Enlarged seismic section of line PA_
around well 1-3-96_W5
c_
3-D Modellinq of a Reef-Fault Block Structure
_.LS.P. 601
1-3-96..6W5 "_-
501
0.85
Top
Keg
River
Fm
_
Top Precambrian
1.05
0
I
800m
I
1-3-96-6W5
I
1261 m
FIG. 7. Geological
interpretation
of patch reef around well 1-3-96-6W5
CREWES Research Reoort Volume 4 (1992)
4-11
DarranJ. Edwards
PHYSICAL
MODELLING
Physical seismic modelling involves the generation of seismic responses over
scaled geological models in the laboratory. The success of physical modelling is
directly dependent on the choice of modelling materials and scaling factors used for the
simulation. The materials used must have acoustic velocities appropriate for scaling and
must be readily formable into complex shapes. Scale factors are chosen for the model
experiments such that the value of a model parameter multiplied by the appropriate scale
factor results in the dimensions of the field prototype. In this study, the physical
seismic modelling method was employed to study the seismic response of an reef
partially overlying a set of vertical faults.
Reef-fault
block
(horst)
physical
model
The geometry of the model is shown in plan view, cross-section, and
perspective 3-D outline in figures 8 and 9 respectively. Scaling factors for velocity of
1:2, for depth and distance of 1:10,000, and for time of 1:5,000 were all used such that
the model parameters multiplied by the scale factor resulted in the dimensions of the
field prototype. The dimensions for the Keg River reef were taken from Anderson et al.
(1988) and are typically elongate, around 1000 m by 500 m and 50 m thick (10 cm x 5
cm x 0.5 cm scaled). Such reef are found at a depth of 1200m (12cm scaled) within the
Panny and Trout Fields associated with the eastern Peace River Arch. A characteristic
maximum throw for the observed normal faulting can be estimated from the contoured
Precambrian surface (Figure 2) at approximately 100 m (lcm scaled).
The materials used for model construction were those that had velocities
appropriate for scaling with respect to the field velocities observed in the Peace River
Arch area. The stratigraphy and velocity structure is contained in Table 1. Using a
velocity scaling factor of 1:2, a suitable model substance for the reefs was found to be
Canadian Tire Plastic resin with 50 micron diameter glass beads (at a volume of 1:1)
with a model P-wave velocity of 2840 m/s, giving an unscaled velocity of 5680 m/s
(compared to 5600 m/s field value). The basal clastic or Granite Wash unit was
constructed from only the Canadian Tire plastic resin, having a P-wave velocity of
2452 m/s scaled (4904 m/s unscaled, compared to the field prototype of 5000 m/s).
Finally, the base (Precambrian) was cut out of a Plexiglas sheet with an unscaled model
value of 2750 m/s (both unscaled and field P-wave velocities are 5000m/s).
Physical
modelling
system
The physical modelling system at the University of Calgary was developed by
Cheadle et al. (1985). The major components of the system are: a water-filled tank (3 m
wide, 4 m long, and 2 m high); two perpendicular beams containing motorized
carriages; two spherical ITC-1089C ultrasonic piezoelectric transducers; a pre-amplifier;
a pulse generator; an IBM-XT PC; and a digital storage oscilloscope.
During operation, the reef-fault block model is submerged in the tank on a
levelled platform for the experiment. The two transducers act as source and receiver and
are moved across the model on the motorized carriages. The acquisition geometry of the
survey is programmed using a PB 386/25 which controls the positions of the
transducers. A zero-phase signal is obtained by the summation of three wave trains
generated from the pulse generator. The received signal is digitized by a high speed
storage oscilloscope. A direct link between the oscilloscope and a Perkin-Elmer allows
the transfer of the seismic trace, containing a maximum of 4096 samples plus the trace
4-12
CREWESResearchReport Volume4 (1992)
3-D Modelfin,qof a Reef-Fault Block Structure
FIG. 9.
SEISMIC
(PEACE
Formation
RIVER
ARCH
desired
field velocity
m/s)
' cover'
basal cl_tics
Preeambrian
view of physical model
DATA
IP-wove;
reef
3-D perspective
PHYSICAL
AREA)
SCALED
thlcknes,,
(m)
vel(m/s)
' water '
5600
50
3anadian Tire plastic
resin with 0.13cm
glass beads; vol 1:1
50(X)
100
2anadian Th'c plastic
resin
model
S-wave
model
Vp/Vs
vel(m/s)
-2840
(x2=5680)
Scalino parameters : time 1:5000 ; distancedepth
-2452
x2=4904)
thlcknas_
(¢m)
1480
plexi-glass
Table 1. Modelling
MODELLING
PARAMETERS
model
P-wove
moterlol
1200
5500
3-D
12
-1480
1.92
0.5
,,1182
2.07
1,0
1375
1.19
2,0
2750
(x2=5500
1:10000 ; velocity 1:2
parameters
CREWES Research ReDort Volume 4 (19921
4-13
3-D Modellinqof a Reef-FaultBlock Structure
header, onto magnetic tape in SEG-Y format. At this stage, the raw data can be
processed using ITA's Insight 3-D software.
Acquisition
parameters
A 3-D zero-offset survey was carried out over a scaled area of 25 cm by 20 cm
(2500 m x 2000 m unscaled). 3-D zero-offset surveys have a 'stacked' trace at the
centre of every bin, and for this survey 5 stacks per trace were used. The survey
consisted of 100 lines and 80 shots per line (100 inlines, 80 crosslines). The scaled bin
spacing used was 0.25 cm or 25 m unscaled (line, shot, and receiver spacing
consistent). The inlines were shot parallel to the short side of the reef (shooting
direction). The nearest offset possible was 100 m because of the spatial nature of the
source and receiver transducers. A sample rate of lms was used over a record length of
2 seconds.
3-D
processing
Following acquisition, the raw 3-D zero-offset data is loaded onto ITA's
INSIGHT system for 3-D processing. The 3-D zero-offset processing flow for the
reef-fauh block model is summarized in figure i0. 3-D geometry in the form of
observer and survey note files are then generated and the dataset is updated. The data
were muted from 0 to 700 ms in order to remove the direct arrivals which are of a very
high amplitude nature. This is directly related to the small source-receiver offset of 100
m. Effects due to water reverberations are also removed.
The 3-D data volume is sorted on the basis of full 3-D composite CDP number
(combined line and station number). This sort map is used to pad the 3-D dataset to a
full regular 3-D cube. Data records corresponding to either inline or crossline (station)
data segments can then be easily extracted from the data cube.
Inline 50 (reef and fault block) was extracted from the padded 3-D dataset for
2-D migration testing. A frequency band range of 15 to 100 Hz was obtained for input
into the migration from analysis of the Fourier spectra (Figure 11). RMS velocities
needed for the migration were computed from the known interval velocities of the
model using the following equation:
v.,,=
,.,
61
where Vi and ti are the known velocity and travel time through the ith layer.
Confirmation of these velocities was provided through migration, with the
interval velocities recalculated from the RMS velocities using the Dix equatio n. Figures
11 and 12 illustrate the unmigrated and 2-D migrated inline 50 respectively.
A one-pass 3-D phase-shift depth migration was applied to the entire 3-D
dataset, using downward continuation with the two-way wave equation. Time-RMS
velocity pairs input into the migration correspond to known interval velocities over the
central part of the reef model, since the main aim of the migration is to collapse the
CREWESResearchReoort Volume4 (1992)
4-15
Darran J. Edwards
(physical
modelling tank) DATA1
RAW
3-D ZERO-OFFSET
(3-D GEOMETRY_
,_==-_,,_,,
ITA SEG-Y format data _verticnl/timaslice
J
_-setup end analysie J
3--0
axtraction)
_uslng Landmark '3-.0 Plus'.)
)
MUTE
data muted 0 to 700ms |
J
I
_OSORTMAP
)
data sorted on the basie of full 3-D
composite CDP number
|
|
(combined line and station number))
(,nline50extracted)_.
(
3-D DATA CUBE
_ j
Output
detaset padded to full 3-D cube
_for 2-D migration tasting)
'
J-_selected
(Analysis of Fourier J
Spectra
L(for max I min frequencies)
1
(from
known interval
velocities)
1
one-pass
3-Dmigration
phase-shift
depth
t
Online5o)
2-D MIGRATION I
using Landmark '3-D Plue'
_
Lsalectadinlinee/croesline_
_(
I
out.ut
FIG. I0.
4-16
inlines / croaslinas )
3-D zero-offset
processing
flow for reef-fault
CREWES Research Reoort Volume 4 (1992)
block model
•
1.00
_E
__
o.50
x_
"=8" o.2s
0.11
25
r_
Trace
10
20
30
40
50
50
75
,
Frequency (Hz)
60
100
70
125
80
1.0
i
1.1
FIG. 1l. ]nlinc50 with frequencyrangefor 2-D migrationtesting
CDP
T[ME
10
V-RMS
CDP
V-|NT
TIME
32
V-RMS
V-INT
CDP
COP
|480
IO
1480
1480
840
910
1480
1988
1480
5000
800
840
14110
1892
1480
5600
I0
_
1480
1480
1480
1400
10
1480
LOlO
2559
5508
810
2286
50QQ
_0
1892
5600
840
1480
1480
|OtO
2776
5SQO
2041
5500
lOtO
_28
55_
Trace 10
20
30
1010
40
V-flMS
50
Y-INT
TIME
70
1460
I
TIME
50
|O
60
V-R"S
70
V-|NT
1480
80
!
0.7
i
-I
3
0.8
,_,_
-'--
•
rt--
0.9
,_-r..4--.
10
.i
:=
i
1.1
"=
FIG. 12. Test 2-D migration of inline 50 (with RMS velocities used)
3-D Modellinqof a Reef-FaultBlock Structure
diffractions associated with the reef flanks and vertical faults. This is because only one
set of velocities can be input into this 'fast' depth-migration process. The timefrequency band definition used for the 2-D migration is also included in this 3-D
version.
3-D INTERPRETATION
Both the raw and migrated 3-D zero-offset datasets were interpreted using
output from ITA's Insight system and '3-D Plus" on the Landmark workstation. Two
types of seismic display were produced on both systems: vertical sections; and
horizontal or timeslice sections.
Vertical
seismic
sections
Figures 13 and 14 illustrate representative inline sections from the unmigrated and
migrated datasets respectively. They represent situations where :
- only the faulted platform is present
- both the reef and faulted platform are evident
- the reef overlying a simple platform
- a simple unfaulted platform.
A number of events can be picked on the recorded data in a quantitative manner
because we know the model geometry and velocity. Top of the reef, platform and basal
clastic unit reflectors can all be readily identified due to their high amplitude, continuous
nature. The base of the model is also evident, just below 1000 ms. The highest
amplitude visible appears to correspond with the top reef event at 810 ms, which is to
be expected since the velocity and density contrast between the two media is greatest
(water 1480 rn/s and reef 5680 m/s - unscaled velocities). On sections such as inline
70, the platform is imaged beneath the reef. Slight pull-up can be identified due to the
high- velocity nature of the Canadian Tire plastic resin/glass bead material (reef) with
respect to the surrounding water. The reflections from the platform under the reef body
arrive earlier than those to the side. A basal clastic event is distinguishable at around
900ms from u'aces 1 to 40 on lines where the fault block is present (inlines 10 and 40).
At first glance it may be possible to correlate this with an adjacent lower-amplitude
event at the same time. This is particularly evident on inline 10 where a continuous
event could be interpreted. This lower-amplitude arrival is thought to be the result of a
bond in the Plexiglas (platform) layer. The reflector that represents the model base
appears to vary in amplitude and arrival time. On lines where the reef is present (inlines
40, 70), it is possible that pull-up can be identified for this arrival. Where the fault
block is present (inlines 10, 40), a push- or pull-down effect can be recognised beneath
the lower-velocity basal clastic unit with respect to the platform. The reflections from
this part of the model base therefore arrive later than those adjacent.
On the unmigrated sections, a number of diffractions may be inferred: those
associated with the edge of the reef (with a variation in diffraction curvature); and less
prominent curves delineating the vertical fault. The later would appear to have phase
reversals from branch to branch of the curve. However, no phase reversal is apparent
CREWESResearchReoort Volume4 (1992_
4-19
Darran J. Edwards
Trace
._)-t_t
0,8
10
20
30
40
.t.)-:_-t_-t-_-H
50
-_.t-H-t
_::_,"_"__
_
0,9 ...........
60
.t._*.½.],-}-_-t.t .i .), +H.
70
80
t .i.,I-t-_*._4-_-_4 _.V)_._
-*.
___:;_:_::::-_-,':-,-_
'_'--_:-'-'-_-:-"-_-_.
_:.:-_-'--'.e__
,= 1.o
1.1
.......
Inline
__._/Pf.._-r_¢
................
._
0.7
I.-
..... .............
_,- -H-_ ............_-_:11_:I_.::;;;::;:;;:;;+_H-,+,+_+_=====================
:::::::::::::::
::::::::::::::::::::::::::::::::::::::::::::::::::::::
1.1
......
10 (fault-block)
.._,_-_)_.'p_n_,-,--._.'_-;
;.._._._.-.-._-_
_._
-.-.__-:.-:__,,,,,,,_-6&,.
.... _.-.,.;_:.-;
....................
._-.n,_._
.............................
,.-,_
_*,.__:'_'-._.
--.._---_`_`_`_`_'`_`_`__;;r`:_:_`_-_`_
...... ,,,,, ...............
Inline 40 (reef and fault-block)
===================================
...... .:,_.......
::::::::::::::::::::::::::
....
1 0 ___;.._.;._..._...s_._
11
J..........
: :_'_)_'_'_:_3:t_:_3_-H-_-_-_
=:,.'.
:,,....................................
,_
_-_._-_-_-_:t:_d#l:tct=
-__._-._._._._.__
_,_,_,,_.__,._d,,'_t,F_;
.......................
Inline 70 (reef only)
0.7
• ,_;_._...,,,.,,._,,,-_,,_b_,,,_,,,,_"
-| ......... I ......... I ......... t ......... I ......... !
____--r:::::
":: ": : :: "_¥_¥_¥_";";'_'_Y'¥(('Y_'_:_'_Y_Y_'_'_:e_'_:_¥_'_'_'_5_
0.8
_i22__i_'_"_'_'L:'"_'_'_.':._._.._-_tt_-_tt-t_t_t-t__
1.1
-_,-_-,-_:-_-:.:._.;.;4_.:.:-:.:.:-:_._
_:.:.:-:.___
Inline 90 (no reef nor fault block)
FIG. 13.
4-20
Representative
unmigrated
CREWES Research ReDort Volume 4 (1992)
sections
_ ___ _ __ -
3-D Modelfinq of a Reef-Fault Block Structure
Trace
10
.........
_.........
:::::::::::::::::::::::::::::::::::::::::::::
0.7
20
40
t.:. ...... i .....
50
60
70
;;;!
....... ::!:::::::::!
..................................
80
.........
I
_.-..__},:_--_-----_:-_:-----.:=--L._._._
::::::::::::::::::::::::::::::::::
............
: = : : : : : :.- _ £_ _: : : : : : :..-- (£,_ / ([ f i i i¢_E
::.-::::::7::::7::
................
..............................
_.E'_j._.gE._
.......
(_'
E 1.0
30
i.........
. . , . =_=_:>__3r---:_v_.
.__._._._ H '+,, --kkl._kkkkd_k.kkLkk
p_)_._:_:;
.... _,_>_;;;._
:::-.-._-_.;_-_'_';.':_=_-
___
.........
kk_k.k:,.-,,,,,,
"..................
Inline 10 (fault-block)
0.7
0.8
_l-_J_-'-"
;_"_"'" "_"___.:_:,_:,'}_HH __
_
._.._._._.
...;:,;;;::;:----:-
+ o.8---_-
""
.-%_.<-_ ....................................
...........
,'+,'++"-+÷_,_,"_
"
-_
-
-.-,.....
If
..............
ff=lff:t_
_,m__+_
......... _._ .
.... _._.d.?,._.,..,+_,._.__
_:_,__.-::
1.1
}i_ i H _iH i __ __
-
- ---_- -- :_,k_i_.._._-_-_
.-_
"
...............................................................
Inline 40 (reef and fault-block)
.........
.......
I .........
_-_'_
I .........
H _Hiiiiit,.'iii
I .........
H _
I .........
I .........
_ _'_ii!!iii_'_
I .......
..I
.........
I
_
::::_:_:_::_/:::/_:._/._/:_::::::::::::_._::_:`:._:_.!_2:_!_!!_::_/_..:_.i._:_t.._..
--_t-<-_-._.p_.
..........
. .. ._._c-.:#£_-.
. . ._x:_T.-.#£F£..£.-,.--f_ f
,_<_.
. ........
_._-_._-._.e_-_-._.-<
. ... . . . .. . . . . _.__
.. ._
. ._
Inline 70 (reef only)
.........
I .........
0.7 _-.h'._?;
0.8
I .........
I .........
I .....
"r.,.I
.........
I .........
I ..........
_??_?_my..?
Y:Y_iY+__.-?.--i-.---.:
____ ii _i_ _i_ _ _i_????i_
i i J ; ; ; _;_U_U`;_;_;_:_:3_`_:'`L:_:_L_'``_:,`_.L:_':_:_:_L_':,`:_:_```V:_:_:_:_:_:_:_:_;_;_;_:_:_:_:_:_:_:_:_:_:_:_:_;_:``"_
Inline 90 (no reef nor fault-block)
FIG. 14. Representative
migrated sections
CREWES Research Reoort Volume 4 (1992)
4-21
Darran J. Edwards
on the reef edge diffraction with the steeper, narrower curvature and shorter 'tail'. Such
diffraction curves should be removed by migration. The migrated sections (Figure 14)
illustrate that those associated with the fault and the diffraction with the wider curvature
both disappear; but the so-called 'diffractor' with the shorter tail and no evidence of
phase reversal still remains. This feature is not a diffraction but is a reflection from the
flank of the reef, probably because of the zero-offset nature of the survey. The
reflection has a steeper dip than the diffraction. This can be further illustrated by
numerical modelling of such a reef, using the Sierra modelling system, which does not
compute diffractions. Yet, after Sierramodelling, this feature is still present - indicating
its origin as a reflection. Unmigrated _nd migrated timeslices further illustrate the areal
extent of these distinct diffractions and reflections
An apparent 'diffraction' may also be inferred on unmigrated inline 90 (Figure
13) with the curve apex at, or just below, the platform event. Here, no irregularity nor
diffraction source is present. This feature can be interpreted to be a sideswipe reflection
from the flank of the reef. A sideswipe reflection is one whose reflection points lie
outside the plane of the seismic section. It may be shown through numerical modelling
that this is an out-of-plane reflection (Figure 21). If a reflector has a component of dip
(reef flank) in the direction normal to the plane of the seismic section, the dip will give
rise to sideswipe. Without this understanding, it would not be possible to properly
correlate a reflection through a 2-13 grid and produce a correct interpretation.
Timeslice
displays
One aim of this paper has been to test the usefulness of timeslice sections in
interpreting the reef-fault block model. Four characteristic pairs of unmigrated and
migrated timeslices are presented in this paper (Figures 15 to 19), illustrating
particularly well the nature of the fault-block arrival amplitudes and propagation of
diffraction curves and reef-flank reflections with time and distance.
Above the platform, the extent of the reef outline is well imaged on the timeslice
for time 820 ms (Figure 15), with surrounding water having a zero amplitude. The
focussing effects seen on the reef flank comers of the unmigrated data are not present
when migrated.
Timeslice 860 ms (Figure 16) provides a good example of the areal extent of the
basal clastic fault block. The regularity if the fault block margin is particularly well
imaged in the migrated timeslice. The Canadian Tire plastic resin (basal clastics)Plexiglas (platform) interface shows up as a marked positive amplitude anomaly. The
amplitude decreases significantly toward the centre of the reef where it becomes masked
by transmission losses due to the high amplitude top reef event. Also on this timeslice,
a small amount of pull-up may be inferred by the presence of anomalous amplitude
values around the projected position of the reef, whilst elsewhere values are essentially
negligible.
In the 900 ms and 1000 ms timeslices (Figures 17 and 18 respectively), both
pull-up and push-down are encountered. As mentioned earlier, pull-up is located
beneath the reef and push-down effects are associated with the lower velocity faulted
layer. The effect of the push-down is again to successfully delineate the faulted
margins.
On timeslices beneath the top-platform event, the diffraction wavefront from the
reef flank can be observed to propagate laterally away with time. These diffraction
curves are removed when migrated, leaving a (zero-offset) flank reflection which has
4-22
CREWESResearchReport Volume4 (1992)
Cross line
20
100
100
40
40
20
20
Unmigrated
40
Migrated
FIG. 15. Unmigrated and migrated timeslices for time 820ms, imaging the reef outline
60
80
100
100
20
20
Unmigrated
Migrated
3
FIG. 16. Unmigrated and migrated 860ms timeslices
(note imaging of reef-fault block)
Crossline
Cross line
20
CO
40
60
20
80
40
100
100
80
o
D:
60
Q
40
40
20
20
Unmigrated
FIG. 17. Unmigrated and migrated 9Mms timeslices
Migrated
60
80
Crossline
20
40
60
100
100
20
20
Unmigrated
PlG. 18. Unmigrated and migrated 1000ms timeslices
Migrated
to
4
3-D Modellin.qof a Reef-FaultBlockStructure
similar concentric
diffraction).
pattern but is of less width (reflection
NUMERICAL
has steeper dip than
MODELLING
Sierra's 3-D ray-tracing software was used to shoot a similar 3-D zero-offset
survey over a reef-fault block. The only difference in model design is that vertical faults
are not allowed in the Sierra software. However, the spread length is sufficiently small
for any effect to be negligible. An example of an inline and perspective 3-D view of the
reef-fault block model is shown in figure 19. By using 3-D ray tracing seismic
modelling, it is possible to confn-m theories established during physical modelling.
Diffraction computation is not possible using this software, thereby confirming the
zero-offset reflection origin of the reef flank anomalies imaged beneath the platform in
selected inlines (Figure 20). Timeslices (Figure 21) were produced through numerical
modelling showing the areal extent of these reef-edge reflections. Timeslices below the
platform event again successfully delineate the regular margins of the basal clastic layer,
thus identifying the fault.
By examination of such timeslices, it is evident that a line located away from the
reef is going to contain these reef edge reflections - constituting sideswipe or out-ofplane reflections. Figure 22 illustrates this with the raypaths for crossline 15
superimposed on a contour map and 3-D perspective plot. Due to sideswipe, an
anomaly is present on the unmigrated seismic section . Therefore, 3-D ray-trace
modelling also shows that reefs are detected by seismic sections shot off the buildup. In
such cases, 2-D migration may then result in the false imaging of a buildup along a line
where none occurs.
CONCLUSIONS
In this study, a 3-D zero-offset survey shot over a reef-fault block model was
carried out using the physical modelling system at The University of Calgary.
Dimensions and velocities for the model were taken from Keg River patch reefs or
bioherms flanking the eastern Peace River Arch, northwest Alberta. These reefs
developed typically on or above prominent basement structures (Campbell, 1987).
Such Precambrian relief is thought to be controlled by normal block faulting. However,
the vertical and limited offset associated with such faulting makes their direct
identification on 2-D seismic sections somewhat difficult. The situation may also be
complicated by pull-up or push-down effects associated with the reefal facies and offreef sediment. Erosion of uplifted, faulted blocks produced basal clastic (Granite Wash)
or fault-scarp deposits in adjacent margins. The base of such a lower-velocity basal
clasfic layer can be interpreted to a certain degree on the inline sections of the 3-D zerooffset physical model survey. Nevertheless, identifying the layer directly beneath the
reef still remains a problem.
Investigation of timeslices through the 3-D physical model dataset show that the
areal extent of the basal clastic fault block is particularly well imaged, even beneath
most of the overlying higher velocity reef. 3-D numerical modelling of a similar reeffault block also successfully delineated the regular margins of the basal clastic layer,
CREWESResearchReoort Volume4 (1992)
4-27
Darran J. Edwards
DISTANCE: METERS
400
BOO
1200
1600
VELOCITY M/SEC
1475. Ill
5000.
5500.
5600.
FIG. 19. 3-d perspective view and cross-section of input to numerical modelling
4-28
CREWES Research Report Volume 4 (1992)
3-D Modellinq of a Reef-Fault Block Structure
._l.J-lJJ-J.Jl_J_J.J.l-J.JJJ_-_J_J._JJJ
Illlllllllllillll
-II/11/ll/1111I.ll}/}lll///IlllI/l/I]IIllIl/IlllI/l//lllllI/IIlIl
_-II__II_I _.I I I I I I_I.I I I I I I_I _I I I I_I I_I I I I I I I I I I_I_I _I I __
Inline 10 (fault-block)
i-lillll]llllilllITlllrll'_"iil_l_rlii_l'H
iirl/TrillliTllllillllTllll
__/llllllllllllll]lllliJ_lllllllllllllllll/ll
=--1_111111t__1///111_
-i-II/l/l /////lil/l l l l l l l l l l l l l l l l l l l l /l/l l/l l
__________________________________________________________________________________
Inline 30 (reef and fault-block)
:-_l_il'___..III'T.
l]._l_lirlllITlllrilll
rrlllll
_ ll__Jllllllllllllllllll
i-l_lllll{lll/tt{/////////1_l
-IllIIllIIlIIlIl/IlIl l l l l l l l l l tlil l l l l l l l l l l l l l l lIl l
_-____I____I___i_Ii____iI___________I__I__I_I___I______________II_I_____i_____iI_I__
Inline 70 (reef only)
IllllllI- II-lllll-lllIllllll
,,,,,,,,,,
__
/llllllI/I'/lI/1]]l
Inline 85 (simple platform)
- note sideswipe effects
FIG. 20. Selected inlines from the numerical
modelling
of reef-fault
CREWES Research Reoort Volume 4 (1992_
block
4-29
Da/ran J. Edwards
10
20
3O
CMP NUMBER
40
50
60
7O
8O
7O
80
90
100
Timeslice at 1650ms (unsealed)
• note imaging of basal clastic layer
CMP NUMBER
10
BO
20
30
40
50
60
9O
1OO
7060
I
I
50 +
4O
i
30-20-
Timeslice at 1580ms (unsealed)
- note imaging of reef outline
FIG. 21.
Selected timeslices from the numerical modelling
of reef-fault block model
4-30
CREWES Research Report Volume 4 (1992)
1200
3-D perspective of ray-tracing for line
.PROFILE ALOMO
ICOO
Contour map of ray-tracing for line
LINE NO,
1
Seismic section of crossline 15
illustrating sideswipe
I
IX
1>
*O
X*
U
33
»•
-U)
4-4
FIG. 22. Numerical modelling of crossline 15 illustrating sideswipe effects
Darran J. Edwards
thus identifying the approximate fault positions. It is therefore possible to infer the
presence of a fault bounded layer (basal clastics) beneath a patch reef. This has
significant implications for real data around the Peace River Arch area in terms of the
possible tectonic development of Keg River patch reefs.
The model's secondary purpose was to uncover possible interpretation
pitfalls
in 2-D data and to illustrate the advantages of 3-D seismic data. 3-D ray-trace
modelling confirmed that certain anomalies identified on the physical-model
vertical
sections and timeslices constituted sideswipe or out-of-plane
reflections. It became
apparent that lines located off-reef contained reef-edge or flank reflections. Therefore,
numerical modelling shows that in dealing with 2-D data such reefs may be inferred
along lines shot away from the actual reel This may frequently result in the false
imaging of a reef on 2-D seismic sections.
ACKNOWLEDGMENTS
I am grateful for the valuable suggestions of Dr's R. James Brown and Don
Lawton, particularly in the early stages of this study. The generous donation of the 2-D
Panny River data by Mr. Barry Korchinski of Sigma Explorations
is sincerely
acknowledged.
Mr. Eric Gallant is thanked for his technical assistance in the physical
model building. Thanks also to Kelly Hrabi for his help in 3-D processing. Finally, I
would like to thank the sponsors of the CREWES project for their continued support.
FUTURE
WORK
In continuing the study of determing whether tectonic control on carbonate
morphology is viable, a next step would be to look at both regional and local structural
elements in reefal facies areas. Lithoprobe has recently acquired deep seismic profiles
(down to 22 seconds) across the Rimbey-Leduc trend of central Alberta, as part of the
Alberta Basement Transect program. It is anticipated that this data, together with data to
be shot in the Peace River Arch area will be reprocessed and interpreted with this
basement control objective in mind. Deep seismic data across the nearby Swan Hills
area are also available for this study..
REFERENCES
Anderson, N.L., Brown, R.J., and Hinds, R.C., 1988, A seismic perspective on the Panny and Trout
Fields of North-Central Alberta: Can. J. Expl. Geo.,24,54-165
Campbell, C.V., 1987, Su'atigraphy and facies of the Upper Elk Point Subgroup, northern Alberta: in
Krause, F.F., and Burrows, O.G. (Eds.), Devonian Lithofacies and reservoir styles in Alberta,
Second Intemat. Symp. Devon. Syst., Calgary, 243-286.
Cant, DJ., 1988, Regional Structure and Development of the Peace River Arch, Alberta: A Palaeozoic
failed-rift system? Bull. Can. Petr. Geol., 36, 284-295.
Cheadle, S.P., Bertram, M.B., and Lawton, D.C., 1985, Development of a physical seismic modelling
system, University of Calgary, In: Current Research, Part A, Geol. Surv. Can., Pt 85-1A,
499-504.
Dunham, J.B., Crawford, G.A., Panasiuk, W., 1983, Sedimentology of the Slave Point Formation
(Devonian) at Slave Field, Lubicon Lake, Alberta, In: Harris, P.M. (Ed.), Carbonate
Buildups-A Core Workshop, No. 4, 73-111.
4-32
CREWES Research Report Volume 4 (1992)
3-D Modellinq of a Reef-Fault Block Structure
Mitchell, K.I., 1987, A random walk through a Keg River carbonate shoal using 3-D seismic to
examine a typical anomaly in the Kidney/Panny area (Abstract): Can. J. Expl. Geo., 24, 89.
Mountjoy, E.W., 1980, Some questions about the devlopment of Upper Devonian carbonate buildups
(reefs), Western Canada Basin: Bull. Can. Pet. Geol., 315-344.
Sikabonyi, L.A., and Rodgers, J., 1959, Palaeozoic Tectonics and Sedimentation in the northern half
of the West Canadian Basin: Jnl. Alba. Soc. Pet. Geol., 193-216.
Tankard, A.J., and Welsink, H.J., 1987, extensional tectonics and stratigraphy of Hibernia oilfield,
Grand Banks, Newfoundland: A.A.P.G. Bull, 71, 1210-1232.
CREWES Research Reoort
Volume 4 (1992_
4-33