Document 11370296

advertisement
Velocity and Attenuation Analysis of
Gulf Coast Sediments using
Vertical Seismic Profiling
by
James Raymond Wingo
B.S., Massachusetts Institute of Technology
(1980)
SUBMITTED TO THE DEPARTMENT OF
EARTH AND PLANETARY SCIENCE
IN PARTIAL FULFILLMENT OF THE
REQUIREMENTS FOR THE
DEGREE OF
MASTER OF SCIENCE IN
EARTH AND PLANETARY SCIENCE
at the
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
JUNE,1981
©
James Raymond Wingo 1981
The author hereby grants to M.I.T. permission to reproduce and
to distribute copies of this thesis document in whole or in part.
A
'0
AI
!
Signature of Author
D
artment of
arth and Plan
y Science
"ay
Certified by
V W
T
Accepted by
22,1981
,Rfi M. Toksoz
sis Supervisor
W
Chairman, Departmental Graduate Committee
"WlI1SACHUSE
0 TTS NSTHUTE
PM
1
M
.If3W I1ES
Li5vgAB\s
Velocity and Attenuation Analysis
of Gulf Coast Sediments using
Vertical Seismic Profiling
by
James Raymond Wingo
Submitted to the Department of Earth and Planetary Science
on May 22, 1981 in partial fulfillment of the
requirements for the degree of Master of Science in
Geophysics
ABSTRACT
The elastic and anelastic properties of shallow sediments are becoming increasingly
important for detailed
understanding of earth properties.
In order to study the
velocity and attenuation properties of shallow sediments
near the Gulf Coast, a vertical seismic profile was completSignals from an impulsive
ed for a 1650 ft. deep gas well.
seismic source were received by a three-component geophone
clamped down the well. Arrivals were recorded in 10-ft.
intervals, from well base to the surface.
Attenuation analysis techniques included alignment and
then summation of traces from a series of depths to yield
average properties over comparison intervals.
Attenuation
computations were completed using a spectral ratio method.
For compressional waves, minimum interval size was a depth
range approximately equal to the wavelength of the dominant
frequency component of comparison waveforms.
P-wave attenuation increased markedly through the gas zone. The average
Qp dropped from about 35 to 5. The frequency content of the
source compressional waveform changed over time, so monitor
geophone calibrations were used.
Shear wave attenuation was relatively constant; average
Minimum s-wave attenuation computation
Os was about 20.
interval size was larger than that for p-waves, because of
source consistency problems.
Velocities were a function of depth more than rock type.
Average velocities for the three rock types encountered
Shear velocities increased more
ranged over only only 1%.
strongly with depth, as the ratio Vp/Vs decreased from about
6.8 to 3.6 from the surface to well base.
Thesis Supervisor:
Title:
Dr.
M. N. Toksoz
Professor
of Geophysics
Contents
Page
I
Title
1
II
Abstract
2
III
Contents
3
IV
Introduction
4
V
Geology
8
VI
Survey Geometry and Overview
9
VII
Field Procedure
10
VIII
Velocity Analysis
14
Processing
16
Determinations
18
Attenuation
22
Theory
22
Qp
25
Qs
31
X
Conclusions
35
XI
Recommendations
38
XII
Acknowledgements
39
XIII
References
40
XIV
List of Figures
44
XV
List of Tables
48
XVI
Appendix 1
49
XVI
Appendix 2
51
XVII
Figures
54
IX
XVIII Tables
84
Introduction
The seismic properties of shallow sediments are
for detailed
are increasingly important
yet
variable,
highly
complex and
examination of earth structure.
For example,
learn
local
about
more
and
velocities
in exploration
their
important to
seismology it is
in
variations
shallow
order to prepare common-depth-point stacks of s-wave data.
earthquake engineering,
damage depends
earthquake
of
the susceptibility
on the
structures
In
to
of near-surface
properties
In order to study shallow sediment properties,
sediments.
wave
velocities, in
p-wave
relationship to
shear
we
analyzed compressional and shear wave data acquired through use
of a vertical seismec profiling (VSP) technique.
of both p-and s-wave properties
Combined study
is superior
-to p-wave analysis alone for delineation of rock types and pore
fluid saturation
tics are
ties.
condition, because p-and
s-wave characterismedium proper-
differentially affected by changes in
Toksoz et al (1976) discussed the effects of changes in
rock type and pore fluid on rock velocities and
attenuation of
Tatham and
Stoffa (1976),
seismic waves
(Toksoz et al,1979).
and Gardner and Gregory (1974) discussed the value
both
p-and
s-wave
data
waves to
problems
in
of applying
exploration
seismology.
Previous work on shallow in-situ s-and
p-wave relationships
has been done by Tullos and Reid (1958), and Gardner and Harris
(1968), both for Gulf coast sediments.
Benzing (1976) present-
ed results of laboratory studies of p-and s-waves in carbonates
ratios for
and sandstones, finding higher velocities and Vp/Vs
sands than carbonates.
Considerable work has
in Japan
been done
engineering pur-
propagation properties (mainly for earthquake
Ohta et
poses).
used VSP in
al (1980)
p-and s-wave
on
near Tokyo,
a study
depth of
generating velocity structures for p-and s-waves to a
about 3
A review
km.
of research
on dynamic properties
of
sediments is given in a paper by Imai et al (1979).
In
p-and
near-surface
designed to determine
experiment
whether there is a detectible
This involved
a gas zone.
was
change in
as they
and velocity of seismic waves
the rate of attenuation
propagate through
this
velocities,
s-wave
information about
gain
intended to
to being
addition
the problem
of
finding resolution limits for the technique of vertical seismic
profiling under normal field conditions.
by
characterized
generated by
geophone
detection
a source
at the
surface.
the
below the surface.
source-generated
scattered
and
inhomogeneities.
seismic
converted
He found
phones well below the
Wuenschal (1974)
waves
noise
is
radiating
waveforms
Conventional surface
that signals
the
through
highly attenuation and relatively low-velocity
just
of
well
receivers
the
to
a
down
has the disadvantage
array positioning
travel
must
the
which is
acquisition
seismic data
technique for
VSP is a
weathered layer
noted that much of
caused
from
by
multiple
near-surface
that recording of signals by
surface can improve the signal
geo-
to noise
ratio markedly.
Because of the superior data resolution it provides, VSP has
been utilized
in analysis
frequencies.
Dix (1945)
tion techniques
properties
over seismic
discussed simple velocity determina-
for down-well tests.
completed a velocity
Coast well,
of medium
Tullos and
and attenuation study for a
obtaining values
of the attenuation
Reid (1958)
shallow Gulf
constant for
layered sediments despite significant reflection interference.
More
s-wave
recently,
VSP
Stewart et
velocity
Lash (1980)
study
focusing
al (1981)
for
on
presented results
converted
of
a p-and
wave
generation.
found increased attenuation
and reduced
before-and
after analysis
of
p-waves
passing
through a fracture zone.
A comprehensive
treatment of VSP, with discussion
of tech-
niques and results, is by Gal'perin (1974).
The potential
for VSP applications was further
explored by
Wyatt (1981), who utilized synthetic methods to generate
section which
displayed waveform
both time and depth.
transmission
a VSP
properties
in
Hardage (1981) compared synthetic seismo-
grams created from VSP data to that generalized from
data
and found that
tube
wave contamination, gave superior agreement
in one
instance VSP data,
sonic log
despite heavy
with surface
reflection data.
This
study yielded
structures
wave
for shallow
velocity
and
passing through a
p-and s-wave
velocity
Gulf Coast
sediments.
attenuation
gas zone,
were
notably
while shear wave
and attenuation
Compressional
affected
upon
properties were
less strongly altered.
For attenuation
analysis, traces from
consecutive depth points were aligned for the event of interest
and
then
were summed
Maximum resolution
to yield
for
wavelength for each
was at
average
interval properties.
least equal
to
wave type, due to effects
the
dominant
of interference
and source variation.
Velocity structures
dimensional
found
to
depended on
were generated
straight-raypath
correlate
well
wavelength and
through use
ray tracing
with
sonic-log
program,
data.
interval velocity, and
information was better constrained over depth.
of
and
a twowere
Resolution
shear wave
Geology
Sediments in the
survey took
region and
place were
marker layer
was
at depths
for
of Quaternary age.
at 2100
ft.,
below
the
which the
The shallowest
deepest
survey
depth.
The well
penetrated two
shallow gas zones
which were
surrounded by layers of poorly consolidated sediments.
For
velocity analysis four classifications of sediments, determined
from
facies.
well -log
analysis, were
They were:
to
near-surface sediments,
stone, and clayey shales.
roughly 65% shales,
used
In percentages,
20% sands,
distinguish
sands, silt-
the strata were
and 15% silts.
Sands were
dominant near the surface, while shale was most
common near
the well bottom.
The
strata
were very
unconsolidated.
Core
samples
taken below 1,000 feet added little
lithologic information.
Washout
compromising
was
a
recurrent problem,
information and causing vibration of the
well
log
horizontal compon-
ents of the downhole geophone at several locations.
velocity
analysis,
thirty-one layers with an average thickness of 50
feet were
The
strata
were
thin.
In
the
Boundaries were often defined across gradual changes
used.
in sediment
causing
a
content.
"bright
There was
spot"
lithology is best described
at
only one
1,320
as a
feet.
good reflector,
Overall,
series of thin
the
beds with
continuously varying proportions of sand, mud, and silt.
Figures 3-4 are plots of strata types down the well.
Experimental Overview and Geometry
Testing was carried
out for
positions, with
a geophone
consistent
for
each
procedure,
describes
three sources fixed
moving up
source.
the
The
through
next
in surface
depth intervals
section,
experimental
on
process
survey
in
greater
detail.
The
survey
machine
utilized three
called the
vibroseis truck.
thumper,
sources, a
a vacuum
sliding
gun, and
weight drop
a
shear wave
The thumper was the only source used
for the
velocity and attenuation analysis in this paper, because
of the
superior spatial resolution it provided. (Appendix 1 describes t
the source and the recording apparatus.)
The three
sources generated
the well, at an average
2).
The vacuum
fifty feet,
azimuth
distance of about 250 feet
thumper
270 feet
and the
57"W,
(see figure
from the
vacuum gun
well
northwest
thumper, S360 12"W relative to and 260 feet away from
The
vibrator was
244 feet from
south of
a separation of about
gun and thumper were at
with the
of S250
pulses from positions
the well
at
of
an
the
the well.
at an angle
of S400
was
halfway
57"E, and 283 feet northwest of the thumper.
A three-component
between
the thumper
between
thumper
and
and
monitor geophone
well, about
well.
It
was
buried
ten feet
at
a
off
depth
the
of
line
about
twenty-five feet.
The well itself was cased and cemented , and extending to
a depth of 1650 feet.
Source geometry was partly dictated by
fiel
conditions.
Obstacles
included
the
presence
well,
woods,
and
thumper and
sources
were
of a
large
uniformly muddy
throughout the
kept stationary
mudpit
between
terrain.
survey,
The
and dry
cement was used to preserve and improve the coupling between the
thumper and ground.
10
Procedure
Field
10-ft.
procedure
downhole
was
to complete
geophone
spacings,
thumper
and
shots
for
gun
and
vacuum
vibroseis measurements every 100 feet.
For each depth, the thumper generated four
for each of
line from
pulses, two
two weight ramp orientations symmetric
the thumper
to the well.
ramp was positioned at
an angle
In each
of 450 to
instance the
the horizontal.
The two positionings will here be referred to as
"west," according
to the
about a
weight ramp position
"east" and
relative to
the line to the well.
The thumper
generate
orientations were
both compressional
so chosen
order to
waves;
to minimize
generation of SV-waves, and to produce SH first
arrivals of
opposite polarities
calculations).
and shear
in
(for travel-time and
Figure
5
is an
shear attenuation
example of
the
amplitudes of shear and compressional waves produced
relative
by the
thumper as recorded by the downhole geophone at 1,500 feet.
After
the thumper had completed pulse
for the first
clamping of
the downhole
generations for
geophone
feet, the downhole phone was moved up to 1,640 feet,
on
until it reached 1,330 feet.
The thumper
at 1,650
and so
retained the
same polarity for the first pulse at a new geophone depth as
for the last
every time the
pulse at
the preceding
weight chassis
the operator fired a
depth.
moved to a
In addition,
new orientation,
few test shots until he
was satisfied
that
the
couple
After the
alternation
between
source and ground was adequate.
thumper reached
with the
movements
of
surface.
During this
100
1,330 feet, it
vacuum gun
feet
until
for
the
successive
geophone
sequence, one
all of its rounds, followed
operated in
by the
geophone
reached
the
source would complete
other, and at
the next
depth the order of shooting would reverse.
The geophone was lowered again to 1,320 feet
final vacuum gun run at
the
experiment
Ten-foot
depths
the
30 feet,
thumper
geophone movements
where
gravity-weight
and for the
the
was
after the
remainder of
only
source
were standard except
arrivals
had
used.
to avoid
already
been
recorded during the vacuum gun sequence.
Recording
instrument gains
were changed several
for
thedownhole
geophone
times near the surface, but
only once
for shots below 600 feet (the shallow cutoff for attenuation
analysis
in this study):
they were
reduced 12 db
for all
three channels when the downhole geophone reached 620 feet.
Monitor
geophone gains
were changed
once
during the
survey, just after all vacuum gun tests were completed.
The
(excluding
combined
effects
tests) and
of 330
shots
muddy field conditions
for steps to improve the source-ground
Dry cement
pool when
thumper operator
After each
orientation
created need
coupling frequently.
was therefore placed beneath the
necessary.
per
thumper ground
addition of
cement, the
initiated several test shots to prepare the
couple.
12
The survey lasted
time
was
spent
three days,
resolving
but about half
logistical
of that
difficulties;
shooting itself ran continuously for about 36 hours.
the
Velocity Analysis
Compressional Waves Travel Times
Arrival time
picks from
the thumper source
the inputs for compressional and shear wave
sis.
Plots with
a time scale of 50
initial p-wave picks,
for
checking.
data were
velocity analy-
ms/inch were used for
and 20 ms/inch plots were
This approach
gave arrival
later used
times
within a
range of about + 1.5 ms.
However,
the usual
measurement was
in first-break
human error involved
accompanied by uncertainties
recording instrument
zero time
engendered by
variations.
The recording
instruments often cut in before or after the source began to
generate its signal;
of fifteen
time difference
for
different
the worst case was an
ms. between two
thumps traveling
Usually the range of
apparent travel
to the
first arrivals
depth point.
same
variation in travel times to
the same
point was +2 ms.
This problem was resolved to within human picking error
by examination of
the monitor
geophone arrival times.
average arrival time of 23 ms. from source to
was used to
monitor phone
determine zero time differences, and
computed varied
in a manner consistent with
in travel time to the downhole geophone.
An
lags thus
the variations
The close downhole
geophone spacings and the small and monotonic p-wave moveout
of
2 or 3
ms.
per
spacing (except for
those near-surface
readings where head waves arrived first) helped
to minimize
error.
still
The net
time picks
result is confidence in travel
to within about 2 ms.
Figures 6-9 show the pattern of
compressional waveform
arrivals down the well.
Shear Waves Travel Times
Shear wave arrival
times were less well fixed.
There
and zero
were the usual difficulties, including human error
time, as for p-waves.
But since accurate shear wave picks require overlays of
rotated
arrivals of
figure 10),
opposite polarities, (see
the zero-time uncertainties were doubled.
There were
because of
difficulties in identification of shear arrivals
low level noise due to tube wave and
sional wave interference.
not generate
late-arriving compres-
Finally, the
but reversed
identical
also
thumper source did
waveforms
upon
re-
orientation; this problem increased with time and differentthe downhole
ial compaction, and therefore was largest when
geophone was shallowest.
p-wave moni-
The zero-time problem was corrected using
tor
correction parameters, and arrivals were
overlay
plots scaled
to 40 ms/inch.
weak near the surface,
arrivals and
because of
measured from
Reversals were
perhaps due to effects of
surface wave masking.
tube wave
There were
also roughly 20 ringy, clipped, or dead depth points
horizontal
components of the downhole geophone.
very
on the
(There was
only one bad depth
cementing
phone
point for the vertical component.
allowed much
oscillation.)
more horizontal than
The
shear
interpolated values in some cases.
wave
vertical geo-
arrival
patterns recorded over small spacings
greatly
to increase
the surface.
again helped
time
arrival
measurement
near
But spatially the shear wave error ranges are
than for
compressional waves;
the
average
ratio is 4, while the errors range over roughly a
two.
are
Error ranges are +3 ms. at depth and +4 ms.
accuracy.
smaller
times
Nevertheless, consistent
moveout
shear wave
Poor
Vp/Vs
factor of
(This means that the shear wave velocity structure was
better
constrained
in terms
of layer
depths
and
thick-
nesses.
Figures 11-16 are plots of shear wave arrival waveforms
for
the
indicated geophone
locations down
figure 17 shows incidence times for shear
the
well, and
and compressional
waves as a function of depth.
Ray Tracing
After
arrival time
picking was
completed,
travel times were input to a flat-layer ray
the
tracing program
as standards for comparison with computed travel times.
velocity structures were calculated by
one
lithology,
and
one
which was
Two
ray-tracing methods:
whose layer boundaries correlated with
boundaries,
real
generated
real lithologic
independent
of
with interfaces marked every 50 ft.
The software propagated
rays to
the well for
a range
and
density
of initial
by the
specified
incident angles
It then interpolated to yield travel times at 10-ft.
user.
intervals (for which real travel times were also available).
Another
times
travel
computed
compared
program
with
real travel times, and it generated files with residuals for
The investigator used trial and
10-and 100-foot intervals.
Appendix
error methods to minimize travel time differences.
2 contains a listing of the ray-tracing program.
structure, layer
For the lithologically-based velocity
boundary
analysis of
the
modeling program
based
were
A thirty-one layer model, with four general
classes (including the near-surface region
class) was
on
sonic and
well log resistivity, SP, induction,
density charts.
sediment
to
inputs
then used
However,
analysis.
for ray-tracing
as one
although the average layer thickness is 50 feet, only eleven
and thirteen
of the layers had thicknesses of over 90 feet,
the
latter
Velocities for
of ten to twenty feet.
were of thicknesses
were determined
layers
beyond
time picking
of resolution for the travel
program's limits
modelling
the
constraints, and encountered uniqueness problems resolved by
examination
of
information
initial velocities
from
were based
the
on sonic
.18 is a
plot of
Their
analysis, and
log
were varied in a consistent manner for depth
Figure
logs.
well
and lithology.
final computed velocities
for the
litholic model.
Comparison of
shows
that
the
the lithologic
two
are very
17
and
similar,
fi-'d-block
Lth
models
differences
primarily due to smoothing where lithologic layers are thin.
For instance, in the 1300-1400 ft. region, where the average
layer thickness for
ft.,
the blocked
model, like
lower velocity for
yields
a
the lithologic
model was less
than 20
the lithologic model
shows a
the block including the gas
smoothed average
velocity for
region, but
the
deeper zone,
contrasting with
the lithologic model's ( and
sonic log's)
greater acoustic
differences.
the lithologic
resolution of
straints in
The
model
in
the
ray-tracing
The higher
resolution from
second zone
modelling
is
for
the
beyond
error
the
con-
this study.
proceeding discussion
summarizes
velocity trends
with rock type based on lithologic model analysis.
Average
velocities were based on layer velocities weighted according
to layer thickness,
is
there
greatest
so that the thickest layers,
certainty
in
interval
for which
velocity,
are
weighted most heavily.
Compressional Wave Velocities
Table
1
shows
average velocities
various sediment types.
the three
classes are
computed
almost equal,
slightly lower,
spread in velocities was only 1%.
with
sand velocities
at 6050 ft/s;
well,
where
all
and shale
the overall
However, over 80% of the
sands and almost all of the silts are in the
of the
the
The table shows that velocities for
averaging 6090 ft/s, silt speeds about 6075 ft/s,
velocities only
for
velocities are
shallower half
lower.
For
the
the well,
rock type
partly isolated from
sand and
can be
averages of
6090 and
compaction effects,
were equal to
silt velocities
their whole-survey
while shale
6075 ft/s, respectively,
For all
velocities were significantly lower, at 6010 ft/s.
classes, the
ft/s, the
velocity was 5930
average top-half
bottom-half average was
due to
where velocity changes
shallower half of
6210 ft/s,
Vp was
and the overall
6070 ft/s.
Shear Wave Velocities
Variation between shear wave velocities was
of depth more than of strata type.
a function
For shales, the overall
Vs was 1710 ft/s, with a top half average of 1600 ft/sec and
a lower-half mean of 1770 ft/s.
ft/s through
averaged 1450
Silts
the
survey depths,
while sands were near 1300 ft/s.
Shear wave
The top-half
average was
velocities increased
average speed
1750
ft/s, and
strongly
the bottom-half
was 1320 ft/s,
the overall
with depth.
average
was 1540
ft/s.
Vp/Vs Relationships
values
Vp/Vs
were
most strongly
indicative
of
the
relatively rapid shear wave velocity increase with depth and
compaction.
(See figure 19.)
Shales, most
of 3.6.
prevalent at depth, had an
overall Vp/Vs
Sands, the dominant shallow sediment, had a
19
Vp/Vs
of 4.6, while for silts the value was 4.2.
For the top half of the well, the average Vp/Vs was 4.5,
and for the bottom portion it was 4.0; from top to
base the
mean was 4.2.
Sonic Log Comparisons
The
velocity
waves through
velocity
structure
ray-tracing methods
structure derived
Correspondence was
were generally
generated
close.
for
was later compared
from sonic
log
travel
Velocities from
slightly lower
compressional
than for
to a
times.
the sonic
the
log
layered model
(see figure 18), particularly near the surface.
As a further
test, the
averaged over ten-foot
165-layer
case to
sonic log velocity
intervals and
the ray-tracing
picks were
were then input
program.
as a
Travel times
computed using sonic log velocities were consistently greater
than
real
travel
times,
differences at shallow depths.
with
the
greatest
Below 1,000 feet the travel
time differences were within picking error.
porosity were
sources
abnormally large
indicate,
surprising.
then
Figure 20
this
transit
If washout and
near the surface,
trend in
shows residuals
as other
residuals
as a
is
function
not
of
depth for the sonic model.
Even assuming that in situ velocities are well-known, a
straight-ray
program
like that
used for
this
study will
compute from the true velocities smaller transit
times than
would arise in
real field tests, because real
20
raypaths are
somewhat
curved. Velocities discussed here for
the layered
model in p were chosen so that computed times would generate
slightly negative travel time residuals within the
picking
error.
The positive
residuals for
range of
the sonic log
model, which were consistently greater than picking error at
shallow
depths, indicate
that sonic
log
velocities
were
probably significantly lower than real velocities in shallow
regions.
Attenuation
is
Attenuation
results from
interactions including reflection
spreading, scattering,
tion, geometric
In addition,
of
wave propagates.
constructive and destructive
interference can
in measured
apparent variations
cause
It
and refrac-
absorption
and
the
through which the
the material
energy by
describing
it travels through a medium.
loss of a wave as
energy
term
a comprehensive
attenuation
over a
range of frequencies.
of these attenuation agents except
All
elastic
properties, whereas
losses,
of which
absorption
absorption are
includes anelastic
frictional interactions are
probably the
dominant component (Johnston et al,1979).
Early field
by
work on seismic wave attenuation
that the
(1945),who found
Born
frequency
arrivals through shallow earth decreases
time.
was done
content
of
exponentially with
This led Born to assume propagation of plane seismic
waves with amplitudes of the form:
A(f)=G(f,z)(e-azt)ei(2nft-kz)
2
where G depends on geometry,including reflections and (1/z )
geometric
spreading;
k=27r/X=2rf/V; and V
The
a
is
is phase
attenuation coefficient
with frequency
the
attenuation
velocity.
usually
over a wide range of
22
coefficient;
increases
linearly
frequencies, including
the seismic range:
a= vf
and v is a constant which is characteristic of a
type,
and
which
varies
with
given rock
saturation
and
pressure (Toksoz et al, 1979).
The
value of v
for a
particular rock
sample
can be
determined by comparing the frequency spectrum of the sample
with
that of a known reference.
The ratio
of the natural
logarithms of the respective amplitudes is:
ln(Ai/A 2 )=( v2 -v 1) zf+ln(Gi/G 2 )
Assuming that Gi/G
does not depend on frequency,
the slope
of of the spectral ratio plot should be constant.
If v1 is
2
known or is very small, then the attenuation constant of the
sample can be found directly (Toksoz et al,1979).
For VSP analysis,
one can
measure v by
assuming that
waveforms propagate over the same, nearly vertical, raypaths
until
reaching the neighborhood of the
shallower receiver.
A comparison of spectra of the deeper and
ers should
be indicative of the attenuation' through strata
between
the
similar
raypaths
two
measuring
points.
improves with
geometry such as for this study.
p-wave
shallower receiv-
analysis
starting with
depth for
assumption
VSP
of
surveys of
One reason for completing
depths below
600
(Other reasons
because
of raypath differences.
quality
deterioration and increasingly
incident angles.)
This
ft.
was
were data
non-vertical p-wave
One can define a quantity Q, called "quality", which is
absence of
is independent of frequency in
related to v and
dispersion:
Q=w/vV
wavelet broaden-
Physically, Q is inversely proportional to
loss
strain
the
to
and
ing,
per
energy
in
strain
cycle:
Q=AW/21TW
W is
and AW is
work completed,
energy lost.
Q values
for
various rock types are given in Table 2.
Fluid
decreases
saturation
Qs,
and
both Qp
motion along
because fluids facilitate relative rock matrix
cracks, and also
absorption
of and
because motion
partly
fluid itself causes some energy loss (Johnson et
by the
al, 1979).
In a relative sense, Qs is more affected by fluid saturation
Gas saturation
than Qp.
quality para-
also decreases both
meters.
Taken
together,
Qs
and
can
Qp
indicate
help
saturation characteristics of a reservoir.
the
The question is
whether in situ studies can provide resolution sufficient to
isolate anelastic interval attenuation effects.
Processing Procedure
The
spectra
Fortran software
from
an
used for
interactively
24
this
specified
study
window
generated
of
the
comparison waveforms.
It calculated
the logarithmic ratio
of the spectra of the waveforms, and then yielded a
these spectral ratios.
which
to take a
figures 24-27).
plot of
The user then picked the range over
least-squares fit
to determine
v.
(See
This software was written by Marc Willis of
M.I.T.
P-Wave Attenuation Analysis
The initial approach taken for processing
the compres-
sional wave data was to align the waveforms from
(accuracy
of each
depth
breaking
waveform), and
interval, normalizing
number
for
was +1.5 ms, computed from
stacking:
to correct
ary to compensate
sum them
over
according
variability and
for error introduced by these
over
the
to
source and
to be about 150 ft., which is
close to the dominant compressional wavelength of
Attenuation
to
The minimum stack size necess-
reflection problems was found
ft.
a desired
There were two motivations
for source
cancel reflection arrivals.
high-resolution plots
the output
of traces in the stack.
all depths
smaller
intervals
was
about 190
less
than
experimental error.
Stacking
did not
reflection arrivals.
the
completely destroy
Ganley and
Reflections, like
elastic
attenuation, so
effects
of
Kanasewich (1979) discuss
effect of reflection arrivals on
ions.
the
attenuation computat-
primary arrivals,
experience an-
that complete cancellation
due to
destructive interference incurred by st-cking is not possible.
Nevertheless, summation does great y reduce the effects
of reflections.
Major elements
Source consistency was a major concern.
in
thumper ramp
acclivity, and coupling problems introduced by
each thumper
included variation
source problem
of the
each addition
and
reorientation
of dry
Extreme
cement.
shot-to-shot source variations were identifiable by means of
inspection of the trends in monitor geophone versus downhole
strengths.
geophone arrival
The
data
was run
the energy
Fortran program which flagged those depths where
was markedly
compressional wave window
through a
arriving
through a
different from that coming in at nearby depths and from that
at the
arriving
depth from
same
set
the thumper
at the
Where amplitudes were anomalous on both
opposite polarity.
downhole and monitor geophones, that depth point was removed
from the stack.
In this way, five depth points were removed
the section for each thumper polarity.
from
more difficult problem was variation
A much
strength due to cumulative effects.
did
stacking
not
cause any
It was
cancellations.
frequency content of the source waveform
ently over
uphole.
ground
time, and
equivalently, as
in source
not random, so
Rather,
the
increased consistthe
geophone moved
This was probably due to general compaction of the
near the source.
Figure 21
shows monitor geophone
vertical component spectra which correspond to the shots for
the indicated stacks of eastward downhole traces.
in frequency content
was greatest
The shift
at the beginning
of the
survey, and by the time the geophone had moved up to a depth
of 1,100 feet the frequency content of the waveform reaching
the monitor geophone had almost stabilized.
The frequency shift was probably caused more by altera-
task
This made
the source itself.
character of
change in
the
because the
phone calibrations difficult,
of monitor
by a
and ground than
between source
tions in the coupling
monitor and downhole geophones sensed different parts of the
probably varied spatially
the
less,
possible
only
the quality of
the couple
as well as over time.
Neverthe-
pattern, and
source's radiation
from the
that
geophone with
the monitor
information from
incorporate
to
was
correction
downhole arrivals.
The correction technique used assumed that the spectral
amplitudes of the vertical component of the monitor geophone
corresponded
directly to
waveforms,
Als and
spreading,
the
A2 s.
After
amplitudes
the
thos-e of
at
source
comparison
for spherical
correcting
a certain
and
depth
time
can be written as:
A(zi)=Ais(e-az y
A(z 2 )=A2 s(e-az 2 )
Then
A(z2)=(A2s)(e-a(z2-zl)
A(zl
Aid
Removal
of
variation
source
cancelling (A2 s/Als):
downhole spectrum
effects
the algorithm
by the
was
then
done
by
multiplied the deeper
monitor spectrum for
the shallow
stack, and the deep monitor spectrum by the shallow downhole
spectrum.
This correction was completed for those (deeper)
stacks for which frequency shift was significant.
The spectral ratio method used had significant limitatof
Inaccuracy
ions.
Q
the
as
measurements increases
Q
logarithm
of the
spectral ratios of the traces of interest decreases.
Noise
slope of
increases, and the
the natural
Also, in such instances,
begins to dominate in these cases.
the Q value computed is significantly affected by the choice
Windows
of the spectral window for attenuation computation.
chosen for the listed values were picked for spectral ranges
where compressional wave energy was strongest, and covered a
frequency band centered near 30 Hz, with a width of about 16
Hz.
example
An
of the
effects of
window
spectral
for 150-foot
variation is the monitor-corrected comparison,
feet, for
stacks, of the arrivals centered at 975 and 1,125
which the Q goes from -28
(the value of 80 was
22-39 hertz,
while
size
to 700 to 80 as the window widens.
width of
picked with a typical window
the first
two values
were
for narrow
strong ampli-
windows
which excluded some frequencies with
tudes).
In cases such as this, attenuation is probably low,
within
the range of experimental error.
this case that monitor geophone
It is possible in
corrections overcompensated
for downhole source radiation patterns, or that constructive
interference strengthened the deeper waveform.
Time window length of the comparison waveforms can also
influence
calculations.
Tests
showed
that
too
window yields a spectrum dominated by the properties
short a
of the
sine taper used to prepare the trace for
while too long a window includes more
Time
windows
used
for
the
spectral analysis,
interference effects.
compressional
wave
analysis
discussed here were all four cycles long.
Results for Qg
Tables 3-5
show values
200-foot stacks
geophone.
of
Qp
computed
from 150-and
of the vertical component of
the downhole
Figures 22 and 23 are plots of
compressional
the spectra for
waveforms which were sums of
fifteen traces
whose depths were centered on the indicated depths.
One
result immediately
attenuation
the
is markedly
main gas sand.
variability,
as
stands out,
higher across that
Elsewhere, values
one
that
value
is
being that
zone including
for Qp show notable
negative,
while
comparison of Q values computed
shows a
other values are relatively high.
Nevertheless,
rough correspondence between the values for stacks
thumper
with eastward versus westward orientation.
from the
As dis-
cussed in the geologic section, lithology from the depths of
600 to about 1,320 feet consists mainly of shaly strata with
small
sandy and silty layers interspersed.
moderate through the shallowest shale layers,
with depth and compaction.
and decreases
Q values vary from 12 to 85, but
center around 35, with four of the eight data
10 of that value.
Attenuation is
points within
In the region around the larger gas zone,
Q values for both sources drop dramatically to about
2, and
values computed after monitor phone
corrections show
skew, but qualitatively
indicate, again,
greater east-west
in the
intermediate attenuation
shales, reduced
shallower
attenuation in
through the deeper shales, high
attenuation
the
sources,
west
and
for the next interval.
respectively,
Q
for east
5
and
-15
diverge to
then
at 1,320
gas zone
the
region of
and moderate
feet,
attenuation in the lowest zone which includes the second gas
The most anomalous Q values
zone and a high-velocity shale.
versus those
centered on 975
stacked arrivals
are for the
Perhaps
centered on 1,125, where both Q's are close to -30.
structive
interference affected
versus
825
differences in a
975
the deeper waveform.
ft.
for
perhaps con-
positive Q values for east
high negative and high
about
compensated
these regions, or
shifts in
source frequency
stacks
incorrectly
geophone corrections
monitor
and west
only
represent
very flat slope centered about
The
small
zero; both
indicate very small attenuation, in the range of experimental error.
for 200-foot
More measurements,
region just above the
below
the
gas, and
gas sand
they also
stacks,
to that zone
indicated
through the region including the gas.
and associated
including and
high attenuation
Q values were about 2
before source variation corrections, while
ues centered on 5.
compared the
compensated val-
Figures 24-27 show the spectral windows
log ratio
plots used for
ions.
30
these determinat-
Shear Wave Attenuation
Processing
mainly
the two
on
geophone,
the downhole
of
horizontal components
and their preferred horizontal polarity
pendicular to the line from
since the
were received
(primarily SH) arrivals
shear wave
The
to the well.
the thumper
the horizontal
as its cable twisted during each relocation,
the geophone
But
re-oriented itself
downhole geophone continually
components of
was per-
needed to be
artificially
re-
oriented in order to isolate the SH-arrivals consistently on
one channel.
the assumption
specified
The rotation
that the
energy arriving
a
within
shear.
was primarily
time window
iterated through 1800
algorithm operated according to
in 10 steps, and chose
The
userprogram
that rotation
angle which maximized shear energy arriving at the designatThis channel, the SH-maximized record, will be
ed channel.
referred to as the transverse trace.
Figure 5
shows the
three components
geophone at a depth of 1500 ft.
a greater
amount of
of
the downhole
The transverse channel has
shear energy,
and
less compressional
wave energy, than the other rotated horizontal channel.
A problem with the rotation algorithm was that, depending on
real downhole geophone orientation, it
chose either
shear arrival or
its reverse.
the correct polarity
The
best
of the
in-stream correction
is to
compare
the rotated
a
with
waveform
previously
having
arrival
rotated
the
correct polarity.
comparison
A
of east
wave
shear
with west
rotation
angles for the same depth point (fixed geophone) showed that
the
angles usually
rotation
computed
in part
Variation was due
over
ranged
of
to limitations
+150.
the rotation
algorithm, but real differences in source realignments could
also have contributed to variations.
The
was
algorithm
rotation
effective
when
incident
energy was indeed primarily shear, and compressional reflection
and
tube
wave
minimal.
interference was
Careful,
trace-by-trace, rotation window selection largely alleviated
the
interference problem,
pleted in
order to
but further processing
correct for small
was com-
rotation differences
caused by contamination.
The first step in the rotation correction procedure was
to align all of the horizontal traces according to the shear
Alignments were next
wave arrival times (picked to +3 ms).
fine-tuned
ence began.
Starting with the deepest traces (least likely
to be contaminated by
depth point
calibration sequ-
by computer, and the rotation
was
tube waves),
compared to
the SH arrival
a summed
composite
at each
of three
previously re-rotated SH arrivals from points directly below
it, and was then re-rotated through +15 degrees in
maximize the semblance between it and the
order to
reference stacked
trace.
Once rotation, alignment and rotation
calibration were
over 400-foot
completed, the transverse, traces were summed
A stack was again necessary in order to account
intervals.
for possible shear wave reflections, other wave-type contamination,
and source
variations.
Attenuation measurements
were tried for a
variety of
window lengths,
two-cycle window
was chosen as a compromise
finally a
and
between desire
and need
shear arrival information as possible
for as much
Figure 29
to minimize depth-dependent interference effects.
shows the impact of
a larger
window size of
seven cycles,
versus a more typical length of four periods.
Nevertheless,
though
simple
interference was
analog
was
two-cycle windows
showed a
Hz.
Attenuation
methods:
indicated
trace analysis
primary frequency
with humps around ten
still a problem.
about 20
and
Hz, the
calculations
that
the
from the
spectra
spectral
distribution
hertz, and
a node at 20
bimodal
thirty
Al-
were
then
with a short spectral window
made
about
using
the
two
10-Hz
hump, and with a much longer window including both humps.
Qs Results
Table 6
shows the
measurements calculated
results of
shear
wave attenuation
over a series of intervals.
and west-thumper orientation SH-sections were
arately,
as for
rough correlation
compressional waves,
and
between the Q's found for
East-
analyzed sepresults
showed
each polarity.
Attenuation of the shear waves was consistently greater than
for compressional waves,
as Q's values ranged from a minimum
of 10 to a maximum of 50.
Six of the ten Q values tabulated
and the high Q's are
were closely spaced about 11,
all for
narrow spectral windows (centered about 10 hertz).
Source waveform variability, geophone ring,
ference combined
over
depth, so a
centered
roughly
the shear
check was
the best data
computed
from
comparison of
15,
For a
and
(5-30 hz), Qs was about 25.
and log ratio plots for
smaller stacks
the west section.
those
Q
stacks
agreed
short spectral
window
for a long
spectral window
(See figure 30 for the spectral
these stacks.
34
two
spectra variable
including approximately
points from
with previous runs.
(5-15 hz), Qs was near
wave
run using
about 810 and 1,200 feet,
twenty of
values
to make
and inter-
Conclusions
order to
A vertical seismic profile was completed in
investigate
the seismic
The
sediments.
properties of
experimental
both
generated
equipment
velocity and
shear wave data, from which
compressional and
Gulf Coast
shallow
attenuation determinations were made.
10
of
Use
constrain
to
sufficient
measurements
ft.
to
of
about
a
was 6070 ft/s.
and
wavelength,
The
velocity calculations to zones of one-half wavelength.
dominant p-wave wavelength
was
attenuation
wave
compressional
intervals
downhole
offsets
clamping
*was 190 ft., and the
average Vp
with a dominant wavelength of
Shear waves,
about 90 ft. and velocities averaging 1540
ft/s, one-fourth
of the p-wave velocities, had potential for
greater spatial
resolution, but
problems
source consistency
increased the
minimum interval size for attenuation calculations.
Analysis of well-log data enabled the investigator to
divide
strata
straight-raypath,
into
four
two-dimensional
sediment
A
classes.
raytracing
program
was
then used to generate a velocity structure which yielded the
following sediment velocity relationships:
for the near-surface, Vp was 4860 ft/s, while
Vs was
700 ft/s;
for sands, Vp was 5960 ft/s, and Vs was 1300 ft/s;
for silts, Vp was 6075 ft/s, while Vs was
and for shaley clays,
Vp was
1430 ft/s;
6050 ft/s, and
Vs was
1710 ft/s.
However, the dominant factor affecting velocities was
compaction.
Both p-and s-velocities increased consistently
with
though
depth,
shear velocities
were
more
strongly
altered with burial depth.
Through
the
significantly,
much.
This
gas
zone,
p-wave
velocities
while s-wave velocities were not
result
was
apparent
from
dropped
changed as
both
ray-tracing
modelling and sonic log measurements.
Attenuation
processing
points
sequence which
then
and
interest.
steps
measurements
summed
Shear
of
wave
completed
aligned traces
them
ov'er
geophone
after
from
all
a
depth
of
intervals
depth
processing included
artificial
energy-maximization
were
the additional
rotation
an
using
technique for a shear wave
window, and
re-rotation using semblance methods.
s-waves traces were checked by
Both p-and
which
flagged shots for which arrivals
a program
exhibited anomalous
arrival strength.
kind
Another
continual
change
increase
in
the
monitor geophone
for data
to
of
source
in signal
frequency content
source-ground couple
calibrations.
which this
involving
variation,
over
time,
The following
series processing
due
a
to a
required
results are
steps
had been
applied.
Q p values centered about 35, but
a maximum of 90
to a minimum of 5.
ranged from
As the geophone passed
through the
major gas
sand,
Qp
dropped
significantly,
to
about 5, even after monitor geophone corrections were added.
Qp
values
(40),
were highest
and were
less large
that layer including
in the
intermediate
through the
both the
shallow
deeper gas zone
depth shales
shales and
and deepest
shales (25).
Average Qs was about 25. Qs showed
smaller variation
with depth and source orientation than did Qp.
Interference
effects, probably due to near-surface multiples, caused some
frequency-dependant variations in attenuation.
Recommendations
More
patterns
study
of the
is
needed to
source. This
determine
will
constrain
measurement
error with respect to time
and
(spacial
depth
relationship between
the
variations).
attenuation
(source variations)
In
monitor geophone
radiation
and
particular,
the
downhole signal
strengths needs to be resolved.
Use of a source with stationary baseplates might help
to improve source consistency. Deeper burial of
the monitor
geophone
in
will
help
geophone corrections.
to
increase
confidence
monitor
Acknowledgements
I would like to thank Dr. Bob Lucas at Amoco Production
Research
Company for
this study.
Also, my special thanks to Lana Nelson at Amoco
kindness in
Orleans for
in New
Company
Production
support for
providing assistance and
her
patience
his first
helping this fledgling programmer in
Bob Melton
that awful monster called CMS.
encounters with
was a great help
and
in interpretation of the geology
from the
Dr. Jeff Johnston, my advisor
at Amoco,
well log data, and
was a valuable source of information and advice.
Dr.
At M.I.T.,
in
invaluable
encouragement.
was
which
used
Roger Turpening and Robert
generating
Marc Willis
for
ideas,
provided the
attenuation
Stewart were
approaches,
and
Fortran software
analysis.
Pam
Stanley
Paul
Cunningham
Especially, I would like to thank Professor
Toksoz for
assisted
with
velocity
analysis,
and
didn't.
his recommendations, support and understanding.
I was very
lucky to meet him through field camp in 1979; I believe that
he likes eggnog almost as well as lard bread.
Finally, my thanks to my bro's, who embody
essence of ROGS.
the eternal
References
Aki, K. (1976) Signal to noise ratio of seismic
measurements, Volcanoes and tectonosphere, edited by
Hitoshi Aoki and Susumu Iizuku, Tokai University Press.
Anderson, D. L.,
et al., Attenuation of seismic energy in
the upper mantle: J. Geophys. Res., 70,1441-1448,
1965.
Auberger, M. and J. S. Rinehart, Method for measuring attenuation of ultrasonic longitudinal waves in plastics
and rocks: J. Acoust. Soc. A.,
Birch,
F.
and D. Bancroft,
32, 1698-1699, 1960.
Elasticity and internal friction
in a long column of granite: Bull. Seis. Soc. Amer.,
28,
243-255, 1938.
Born, W. T., The attenuation constants of earth materials:
Geophysics, 6, 132-138, 1941.
Collins, F. and C. C. Lee, Seismic wave attenuation characteristics from pulse experiments:
Geophysics,
21,
16-
40, 1956.
Dix, C. H., The Interpretation of well-shot data, Geophysics
10, 2,
160-170,
1945.
Gal'perin, E. I., Vertical Seismic PLofiling, Society of
40
Exploration Geophysicists Special Publication No.
12, Tulsa, Okla., 1974.
GanleyD.C., and KanasewichE.R., Measurement of absorption
and
dispersion from check-shot surveys: Institute of
Earth
and
Planetary Physics, University of Alberta,
Edmonton, Alberta, Canada,1977.
Gardner,G.H.F., Wyllie,M.R.J., and DroschakD.M., Effects
of pressure and fluid saturation on
of elastic waves in
the
attenuation
sands: J. Petroleum
Technology,
p.189-198,1974.
Gordon, D. H.
,
Velocity and attenuation in imperfectly
elastic rock: J. Geophys. Res., 73, 3917-3395, 1968.
Horton, C. W., A loss mechanism for Pierre shale: Geophysics
24, 667-680, 1959.
Johnston, D. H., The attenuation of seismic waves in dry and
saturated rocks: Ph.D. Thesis, MIT, Dept. of Earth,
Space ond-Planetary Sciences,1978.
and Timur,A.,
Johnston, D. H., ToksozM.N.,
seismic waves in dry
Toeoretical models
and
and
Attenuation of
saturated
rocks, Part II,
mechanisms,
Geophysics, 44,
p.691-711,1979.
KennetP., IresonR.L., and Conn,P.J., Vertical sesimic
profiles and their application in exploration
&WAM"Mm"
,
geophysics: Geophisical Prospecting, p. 676-699,1980.
Knopoff, L.,
Q:Rev. Geophys., 2, 625-600, 1964.
Kuster, G. T.,
Toksoz, M.N. (1974) Velocity and attenuation
of seimic waves in two-phase media:
Part I and Part
II, Geophysics, 39, 587-618.
Lash, C. C., Shear waves, multiple reflections, and
converted waves formed by a deep vertical wave test
(vertical seimic profiling), Geophysics, 45, 9, 13731411,
1980.
McDonal, F. J. et al., Attenuation of shear and compressional waves in Pierre shale, Geophysics, 23, 3, 421-439,
1958.
O'Brien, P. N. S.
,
A discussion on the nature and magnitude
of elastc absorption in seismic prospecting: Geophys.
Prosp., 9, 261-275, 1961.
Pelesnick, L. and I. Zietz, Internal friction of fine
grained limestones at ultrasonic frequencies: Geophysics, 24, 285-296, 1959.
Ricker, N., The form and laws of propagation of seismic
wavelets: Geophysics, 18, 10-40, 1953.
SchoenbergerM., and LevinF.K., Apparent Attenuation due to
intrabed multiples,II: Geophysics,43,730-737,1978.
42
Strick, E., The determination of Q, dynamic viscosity and
transient creep from wave propagation measurements:
Geophys. J. Roy. Astro. Soc., 13,197-218, 1965.
Stewart, R. R., R. M. Turpening, and M. N.Toksoz, Study of a
subsurface fracture zone by vertical seismic profiling,
submitted for publication in Geophysical Research
Letters,
1981.
Toksoz, M. N., R. M. Turpening, and R. R. Stewart, Assessment of the Antrim Oil Shale fracture zone by seismic
profiling,
U. S.
Dept.
of Energy Report No.
FE-2346-
91, September, 1980.
Toksoz, M. N., Johnston, D.H., and Timur, A., Attenuation of
seismic waves in dry and saturated rocks, Part I, Lab
measurements, Geophysics, 44, p.681-690, 1979.
Tullos,F.H., and Reid,A.C., Seismic attenuation of gulf
coast sediments: Geophysics,34,516-528,1969.
Walsh, J. B., Seismic wave attenuation in rock due to
friction: J. Geophys. Res. 71, 2591-2599, 1966.
White, J. E.,
"Seismic Waves: Radiation, Transmission, and
Attenuation", McGraw-Hill, NY, 1965.
Wuenschel, P. C., Dispersive body waves-an experimental
study: Geophysics, 30, 539-551, 1965.
Figures
1.
Survey geometry.
2.
The thumper source with weight ramp in its three primary
configurations.
The bottom two positions were used in this
survey (from Toksoz et al,
3-4.
5.
1981).
Geologic section.
Waveform
arrivals recorded
the downhole geophone for
traces
a depth
were rotated
shown
isolate shear
on the three components of
of 1500 ft.
Horizontal
during processing
wave energy on the trace
in
order to
displayed uppermost
here.
6-9.
Arrivals to
the vertical
component of the
downhole
geophone for the depths indicated.
10.
Overlay plot of the transverse, or shear wave-maximized
trace
for
the
configurations,
Compressional
shear waves
thumper
from the
source
east
and
downhole geophone at
arrivals break
come in
in
with the
with reversed first
same
west
1,500 feet.
polarity, but
particle motions,
due to source reorientation.
11-16.
by
Shear
subtracting
thumper source
different
wave traces.
These waveforms were computed
corresponding transverse
in east
and west
polarities signals
records
orientations.
retain the same
for
the
Since for
polarity for
compressional waves, but reverse polarities for shear waves,
this subtraction
technique enhances
shear
wave
The traces were calibrated for variations in
strength
Scaling
with
source
reorientation
was done over
a p-wave
arrivals.
p-wave arrival
before
subtracting.
window around the
time of
first-break p-arrivals.
17. First arrival times for compressional and shear waves as
a function of depth.
18.
Plot of compressional, shear, and sonic log velocities
with - depth.
The
two
former
velocity
structures
were
computed using a flat-layer ray tracing program.
19.
The
ratio
velocities
between
compressional
as a function of depth.
and
shear
Shear wave
wave
velocities
increase relatively quickly with depth.
20.
Travel time residuals for the sonic log velocity model.
Inputs were
averaged sonic velocities.
the difference
range for
model
between computed and real times.
p-wave travel time
travel
The residuals
times
picks is
are significantly
The error
about 2 ms.
greater
are
than
Sonic
real
travel times for shallow depths.
21.
Amplitude spectra for sums of arrivals at the vertical
components
of the
wave time window.
monitor geophone during
a compressional
The stacks are for source shots for which
the downhole geophone was positioned as indicated.
22.
Amplitude
spectra
for
first
compressional
wave
arrivals received on the vertical component of
geophone
for
values.
All arrivals
clamping
depths centered
on
the downhole
the
indicated
were from the thumper source
in its
eastward orientation.
23.
Downhole
22,
but
geophone compressional
received
from
the
spectra as in figure
in
thumper
its
westward
orientation.
21-24.
Plots
of
spectra, and
amplitude
the
the
compressional
ratio of
wave
the natural
coefficients for
first
arrival
logarithms
each frequency,
for
vertical downhole geophone traces centered on 1260
of the
stacks of
and 1410
feet (including clamping points from 1,200 to 1310, and 1320
to 1510 feet,respectively).
to 1,340
feet.
The main gas zone was at 1,320
Chosen spectral
window limits
are marked
with arrows.
Figures 24 and 25
are for arrivals which had
not been
corrected for source variations.
Figures
Attenuation
26
and
decreased
27
source-corrected
show
with
the
corrections,
spectra.
but
was
nevertheless hogher than through any other section of strata
near the well.
28.
P-wave Q-values.
These calculations
east and west orientations.
are averages for
Monitor-phone calibrated values
were used for the deepest intervals.
29. A comparison
of shear
wave spectra for
different time
windows, of seven and four cycles, starting with first shear
wave arrivals.
30.
Interference is greater for longer windows.
A sample spectral comparison for two shear wave stacks
centered
on 810 and 1200 feet.
for both short (usually
wave
5-14 hz)
windows. Interference
signal near 20 hz.
Attenuation was calculated
and long (5-30
destroyed much of
hz) shear
the arriving
Tables
1.
Velocity relationships
waves
computed for
between compressional and shear
three sediment classes
determined from
well log analysis.
2.
Q values for various rock types.
3-5.
Computed Qp
values for
From Johnston (1978).
15-member sums of
waveforms
arriving to the vertical component of the
downhole geophone
for clamping levels distributed about the
indicated average
depths.
Figure
3
shows
computations
found
without
source
corrections.
Figure
4 shows calculations with source
corrections, as
determined from the monitor geophone arrivals.
Figure 5 gives Qp values found for the indicated
sums of
traces just above and just below the primary gas zone.
6. Computed Qs values for the indicated intervals.
Appendix 1
Equipment Description
The seismic source
was
a
used for the analysis of
gravity weight-drop
Figure 1 shows
device known
as
this paper
a
"thumper".
the thumper in various configurations.
the tests run on
the Gulf Coast, the thumper
For
operated with
the weight ramp in the two non-vertical orientations.
The
weight used was
eight foot
long
horizontal.
ramp
1500 pounds,
which
inclined
450
maximum extension
The weight's
from its base,
was
and it slid
protruding one
foot above
the
the
ramp's end.
square metal
Whenever the thumper changed
the operator
also moved the baseplate, and
orientations it
from
was nine feet
The force of impact was distributed by a 2 ft.
base plate at the ramp's base.
down an
tried to minimize the resultant coupling changes
by running
a few test shots.
monitor and downhole geophones had
Both
Signals arriving at the geophones
perpendicular components.
were recorded
12
gains
db
milliseconds,
three mutually
banks with
by Minie-Sosie instruments on two
apart.
and
the
rate
sampling
The
total
listening
used
period
was
was
2
two
seconds.
A
vacuum-driven weight
vibroseis truck
only
for
operated in
100-foot
drop machine
and
a shear-wave
addition to
the
thumper, but
downhole
geophone
spacings.
In addition,
record the
a
surface geophone
arrivals from
array was
the vibroseis
also
source.
used
The data
from these sources was not used for the discussion in this
paper.
50
to
C
c
vat is the arrm of velocity vat**&
c
thick is
9
C
c
C
4a or
with iatorfac. depths.
It bad all othr
alr
ieratioas.
iaforatioa is measured from the v'!rtLicsl.
Thtat is the last seagle for rayVtraciag Iteratioss.
TThiAcr Is the iscremeat betwomm aciest aingles for
re
x
successive Jtoetioas.
ep'n
C
Walkis the source-to-well offset.
C Inz Isthe svrvey bottom depth
clasplug depth&
t
C
9
level
This is
I~~
Lii
HC
is sansrraw storing depth. et which ras& or* lxtideatp
C
c
C~h
roIa
G
c
9
steH
$: travel time vaiufo.
ritfte.
The program traces rets from a succession of incident
anles cointroled byV he user. It records
the
deptt
Which the revs arrive at the mall as well astahleir travel
ie,
n Interpolates to find travel times at depths
for which there ae rest trayel tie measurements*
£
G
C
aa
9rey
3
pro mrs.ra
stores these travel times, and the veet too
is
a into a residual Cosn *rise* progrem
of his liking. latertpolatioa Cteau& inaccurate whorl
ffseta ra 'rV aarg a. The proroa recor-ds the
arrival informatioa for the user to laspect to decide
it he needs a &salter sngular increment.
Imsz-i8sa.
watk*261*
character respoass
write(6,i )
i foraat(' *Ater I If ou sant structure parseeters read from fie')
resd(S.2) Armed
Oformst(v)
I? (*oot(' eater &umber of lawers')
iir adl~i go to 3
call read Ithick,ve,almag)
9
£
£
Subroutine read reads In velkocity structure values
from storage files*
writes,
)
6 format(, eater layer thiohsesso')
rasd(S~a) thick(um)
writa(6 ' )
19 fermat(O eoter
layer votocity')
rosd(5.l) valtse)
4 wrlio(B.111
It formet(' enter thotae1)
read(5.2.) thtel
fae.01743SH9
thaeis focothetaeI
13)
brttolt.
13 forest(' *sar thetef
read(S.2) theit
writetS. 15)
1S
enter *alle
formeiC
lxcrsueat)
read (5.2) thincr
thiacr-I&CMthlCr
tketas .tetal
ltot*S
Lii
utote.
thet
tasw
tJ
&/t
dx-%kick( I )tas(
tkaw)
xiktoxtotfdx
it~t~t.a~vtk)go to 31
dlsthtck( 11 )ioa(1l1ue
go to 35
I1I(doet.gI.smax9)
Stoaew'eltota
diextkack thaew)
how
desltedia/velIM)
%ttie(j1;trthee~j)4delt
34to30929
11.ait.
3theaswo thutsac
3Scoatinmue
lsvel(
34
lies.of
gk eti
0
are$
ii-6
dolie461.11
wrteC?, 115) tevita),6rtlaata),sagleta)
116 foreatt' sa'.16,' trtiae-',f12.5,' S acident 6agle**,012.6)
1tsteve(a)-level(a*11
go to 112
ttillt.4axI
tt
area.
contaiue
anxe
112
*.xeaage(mr+1),aagleter)
urlte(6,113)
between sagte',112.s,
113 forest(' mex-',15,
Ihow for
9
C
638 do 7
ad,12.s)
laterpeotatoa
k=1.165
k1eaktiacr
do 80 ame1.ti
lfikIS~mq.Ievetta)) go to IS
1(kil.g.teveota)) go6o 79
7
re teveitt)-Ieve(a+1
go to 69
79 mes-I
1t(IteveIta).eg.0).and.(ki0.g6.1I))
i
go to
kI6
It write(6,830)
130 fereat' no raga arrived below',16,' feet')
71
to
to
15 travik)*trtieu(t)
£9 coati&%e
70 continue
71 coatine
do 26S i-Ik
116*161
write(4,261) 110,6rayv()
261 foraet(17,(12.5)
250 coatiAuI
stop
C
e
subroatlas read(6kick~veltrs)
dieseslon jdeep(165),v*1(I6S),tkick165)
deep(1),vei)
raed(3,9)
Is()-$
WhI1)&
Ifta.q.1
go
to 11
do
I 1..
read(3,9) jdeep(),vetI()
9 Ioreet(1?,f13.5
-
tik(Ihtledsep Ail-jdesp(i-1)
10 cotIh~AVO
II return
and
11
Well
Vibrator
Vacuum Gun
Thump
routc
-,8
.
I
Grouc 2
Group 3
200
0
SCALE
Group 4
Group 5
Group 6
Figure 1
Group 7
-54-
200
IN
FEET
400
,-%
-ir
* ~
- .;
*1
r..
-.
-
7
srr44
~
.1d
*
-
fl'
~
i4E-tC.
Figure 2
-55-
I
0
460
60
520
70 ctC
560
160
180
c
710
270
740O
290
840r7
3
440
460
Figure 3
-56-
860
13140
1360
1380
'o.
1390
1410
370
C
I 000
020
I
1030
1500
1510
1550
1550
1600
1650
1190
1200
-
c//////
30
Figure 4
340 1a
-57-
shales
sands
siits
radial
rtical
p
2d0
aido
edo
z=500
sWo
lio
time
b
ins)
d~o
aob
o
sco
MILLS
Sao
Sac
1200
1500
Sao
GO
610
620
6qO
6410
ES50
660
670
680
690
700
710
720
730
V
7-!0
750
760
770
780
790
800
Figure 6
-59-
1900
2100
MILLS
=ao
9CI
soc
1200
15cc
810
I
!a
820
830
850
8SD
870
-
880
Soo
-
910
20
IG
850
968
970
980
990
tocooJ
1010
Figure 7-
-60-
2100
I ell0
~
I
---
MrLLS
Ma
600
9CD
12 a C
151-140
1230
Ito
Figure 8
-61-
C
M
a
L
doo
eca
r
I 5~2O
I ULM&
Figure9
-62-
2:30
transverse component
z=1500 ft
amplitude
1000
500
time
(ms)
1500
M LLS
saa
Soo
so
-1200'
I80
1500o
-
S40
710
720
730
'4-A
740
Figure 11
-64-
MILLS
SCQ
~oO
12Cr
9Wo
I Sco
770
7130
7SO
~Aivv~
820
A
4ft
A
680
~J
870
V
910
Figure 12
-65-
2:cM
MILLS
So
I
Co4
I
S200
2500o
I
950
97C
ISM
1000
1010
1020
1030
I C14 0
1070
1 080
1100
Figure
-66-
l3
Leco
MtLLS
Goo
sOO
1200
ieco
2500
1110
1120
1160
1170
1200
1210
1220A
1230
A
12F0
Figure 14
-67-
21t
MILLS
0
oo
2co
1200
loco
2500
I
t
I
1~8
1350
o
w-I"%
I
1410
Naps
1420
1390
^
"A3 0a
S
1430
1420,
Figure 15-68-
2ItO
w
MILLS
S son
1450
00
603
I0 a
12CO
v, f
-
1560
1580
VV
150
1510
1520
150
l
-
v
15=0
1610
-69-
Figure 16
LSO
210
time (m s)
900
300 600
50
depth
(ft)
1000-
1500
Figure 17
-70-
velocity (ft/sec)
0
4000
2000
6000
40 a
500
Vp
Vs
1000
1500-
Owm-@
-
Figure 18
vsp
--- sonic log
-71-
Vp/Vs
1.
3.
5.
500
1000 -
1500 Figure 19
-72-
7-
time (m s)
3.0
.5
4.5
5007
depth
(ft)
1000-
1500-
travel time residuals
for sonic log model
-'i<rare
-73-
20
600-750
750-900
900-1050
10)0-1200
120 0-13 50
1350-1500
-
1500-1650
--
U)
Crl
-
2
4
frequency (hz)
750-900
~ 600-750
900-1050
C>
1050-1200
-- 1200-1320
4cl
1320-1500
~~
(D
-
1.500-1650
ci
cl
20n
40
frequency (hz)
0
Li
£3
r'i
750-900
w
C)
--
--
3
1050-1200
1200-1320
n)
320-1500
1500-1650
11
-
4/,
0
frequency (hz)
1200 vs 1400 west uncalibrated
-2
Ii
f-I
S1O~0*
0
303
FREQUENCY,
-I0
kiz
60
1200 vs 1400 east uncalibrated
L
N
R
AT
-1-
'100 0-
010
b30
20
FREQUENCY,
Hz
0
1200 vs 1400 west calibrated
X10
60
30
1
0
10
-
- -
-'10
FREQI EICY,
H2
50
GO0
1.200 vs 1400 east calibrated
X~ 10
Q~'
I
G.a
--
I
0 --
-r-
r
-
i0
20
10
FREGI ENCY,
HZ
C50
GO
P-wave Q
40
20
60
1
600
1000
depth
(ft)
I
I
1500"
Figure 28
-81-
-
SOOO
__
_
_
PI
S 250-0 -
A
)
O
-r-Tr-
-rj
g
10000-
t'J
--
-
10
r'
VT~
20
r
TT
3040
50
60
50
60
0-
0)5
500-
1
40
0
V
FREQUENCY,
Hiz
WdE ST
L
MIEDIUM
8 1 0 JS 1200
SHIIE AR
JNDd
0 .3
H
P
.- ,
10
20
C)
30
FREQUENACY,
HZ
Table 1
Velocity Relationships
Sediment Type
Vp
V-q
VP/Vs
Sands
6090
1300
4.7
Silts
6075
1430
4.3
Shales
6050
1700
3.5
Top half
6010
1600
3.8
Lower half
6090
1820
3.4
Top half
5930
1320
4.5
Bottom
6210
1750
3.6
Overall
6070
1540
4. 0
All Types
84
Mea3ured Body Wave
Rock
Quincy Granite
S&lenlhofen Limestone
1-1 Limestone
Q
125
166
112
188
165
liz
Frequency,
Q For Several Rock Types
Reference
Method
long. resonance
Birch and Bancroft
(1938)
Peselnick and Zietz
(1959)
Peselnick and Zietz
(1959)
Born (1941)
long. resonance
Born
P and S wave pulses
Toks6z et al.
120
flexural vibrations
450
P waves in situ
Bruckshaw and Mahanta
(1954)
McDonel et al. (1958)
x 10 3
(.14-4 .5)
(3-15)
x 10 6
(5-10)
x 10 6
lluntLon Limestone
65
(2.8-10.6)
Amherst Sandstone
52
(.930-12.8)
Berca Sandstone
(brine Laturated)
10
(.2-.8)
Navajo Sandstone
21
50 -
Pierre Shale
32
10
50 -
x 10 3
x 103
x 106
long resonance
tors. resonance
P wave pulses
S wave pulses
P wave pulses
S wave
in
(1941)
situ
Table 2
From Johnson
(1978)
(1978)
Table 3
Depth
675
825
vs
Depth
825
975
Polarity
West
1125
Uncorrected
East
it
Wes t
if
East
975
Comments
90
it
Wes t
of
East
It
West
it
",
1275
",
East
1275
1425
West
if
East
1425
1575
West
East
-15
Table 4
Depth
675
825
975
1125
1275
1425
vs
Depth
825
975
1125
1275
1425
1575
Polar ity
Comments
West
10
Corrected
East
10
"I
West
87
"'
Eas t
-76
",
West
-32
"I
Eas t
-25
"
We s t
35
",
Eas t
83
"I
We s t
5
",
East
5
",
West
15
"i
Eas t
32
",
Table 5
Depth
1220-1310
vs
Depth
Polarity
1320-1510
West
Cormments
Uncorrected
East
West
Corrected
East
i
Table 6
Shear Wave Q Values
Depthl
600-990
810-1,200
970-1,320
vs
Depth2
Q
Polarity
Spectral Window
13
East
5-13 hz
25
East
5-39 hz
12
West
5-37 hz
1,200-1,600 41
East
5-14 hz
10
East
5-33 hz
50
West
5-15 hz
1,320-1,650 50
East
5-14 hz
10
East
5-33 hz
10
West
5-14 hz
970-1,320
89
Download