VSP: AN IN-DEPTH SEISMIC UNDERSTANDING
Robert R. Stewart
Department of Geology & Geophysics, University of Calgary, Calgary, Canada
SUMMARY Wireline and Survey Well
Vertical seismic profiling has been a useful measurement to recording truck Source 1 Source 2
obtain rock properties (velocity, impedance, attenuation,
anisotropy) in depth as well as to provide a seismic image of the
subsurface. The VSP can also give insight into seismic wave propa-
gation and provide processing and interpretive assistance in the
analysis of surface seismic data. New multi-level receivers and
hydrophone strings have improved the acquisition efficiency of the
survey. Detailed interpretation, phase-matching work, AVO efforts,
and elastic-wave analysis can all benefit from VSP information. 3-D
images from an area of sources recorded in a VSP show consider- Multi-level
able promise. Similarly, the use of borehole seismic measurements tool
to monitor hydraulic fracturing and perform repeated surveys is
Drilling a well indicates a significant interest in the subsurface of Upgoing Multiple
an area. Hopefully, the drilling results support the original enthusi-
asm. The well might have been a stratigraphic test, a step-out devel-
opment well, an expensive offshore well, or a land exploration Figure 1. Schematic diagram of a VSP survey indicating a survey well,
borehole. It will likely have been logged with various tools. seismic source, receiver, wireline and recording trucks (from DiSiena et
Whatever the well’s type, its location will often have been selected al., 1984).
on the basis of seismic images. In most cases (even the embarrass-
ing ones), we will want to derive as much information from the
borehole as possible - either to find more (or less) of the same. The
vertical seismic profile is both a well log and a seismic imaging tool.
As such, it can help in the geophysical appraisal of the region
around a well. The VSP has four important roles to play in assess-
ing the rock and fluids close to the borehole: 1) to provide in situ
rock properties in depth, particularly seismic velocity, impedance,
anisotropy, and attenuation, 2) to assist in understanding seismic
wave propagation (e.g., source signatures, multiples, and conver-
sions), 3) to make well understood reflectivity images in depth, and
4) to use all of the above in further surface seismic data processing
The basic components of a VSP survey are a seismic source,
wireline and downhole receiver array, and a recording/wireline Figure 2. Conducting a VSP survey in the Pikes Peak heavy oilfield,
truck (shown schematically in Figure 1). Saskatchewan. A vibrator source is in the foreground with
wireline/recording truck and support vehicles in the background. A
A photograph of a VSP survey in progress, at the Pikes Peak crane is used to deploy the wireline and tool in the well.
heavy oilfield in Saskatchewan, is shown in Figure 2. If a single
source position is used within several tens of metres of the bore-
hole, then the survey is called a, “zero-offset” VSP. The typical A full areal set of shots on the surface would constitute a 3-D
objectives of the zero-offset survey are to provide a seismic time-to- VSP. The 3-D VSP can be accomplished very economically if shot in
depth relationship, interval velocities in depth, and a normal-inci- conjunction with a 3-D surface seismic survey. The goal of both the
dence reflectivity trace. If there are a number of regularly offset VSP and surface seismic surveys then would be to make a full vol-
sources from the well-head, then the survey is often called a “walk- umetric picture of the subsurface. A schematic diagram of 2-D and
away” VSP. The walkaway VSPs are usually conducted to deter- 3-D configurations is shown in Figure 3.
mine AVO behavior or to create a 2-D reflectivity image away from
the borehole. Several sources deployed at various offsets might be ACQUISITION
just called a multi-offset VSP or if at different directions from the Downhole seismic data are typically acquired using tools con-
well, a multi-azimuth survey. These surveys would give 2-D sec- taining three-component geophones clamped to the borehole
tions away from the well.
Continued on Page 80
September, 2001 CSEG Recorder 79
VSP: AN IN-DEPTH SEISMIC UNDERSTANDING
Continued from Page 79
Note the relatively straight line of the first arrivals with depth.
Moving to farther offsets the shallow first arrivals are actually
upcoming refractions. We interpret the curved arrivals later in
the data to be source-generated shear waves - which are fairly
common even on the vertical geophones!
Figure 3. Ray-tracing models of 2-D and 3-D VSP surveys showing
the acquisition geometries and reflection coverage.
wall. A typical geophone tool will have 5 levels, although there
are tools with up to 80 levels of 3-C geophones that can clamp to
the borehole wall. VSP surveys can be conducted in open as well
as cased holes, but cased holes are often preferred due to mag-
netic clamping tools and avoidance of borehole stability prob-
lems. Schlumberger Canada’s 5-level 3-C tool is shown in Figure
4. In practice, the tool is usually lowered to the bottom of the well
and records a source shot or shake. The tool will be moved its
length up the hole and the source is reactivated. This continues
up the hole until time or budgets expire. Recording over the
whole vertical range of the well is advantageous to provide the
most complete depth and offset coverage. The cost of the VSP is
assessed according to factors that include: the number of depth
levels recorded, total vertical aperture of the operation, number
and type of source offsets, time on site, tool rental, and of course
mob/demobilization costs. We note that hydrophone receiver
strings can also be effectively used in VSP surveys. This form of
acquisition has the advantage that many receivers can be
deployed with minimal effort. Borehole waves are a major source
of noise with hydrophone cable surveys, but much of this noise
can be removed with various filtering operations.
Figure 5. Vertical channel VSP data from six offsets of a survey in the
Pikes Peak heavy oilfield, Saskatchewan.
Once we have acquired our borehole seismic data, we need to
extract the properties of interest from it as well as make some kind
of image. The most basic information that we will want from the
VSP is a seismic time-to-depth relationship. This information is
Figure 4. Photograph of Schlumberger Canada’s 5-level three-compo-
acquired by just picking the first-breaking energy of the zero-offset
nent geophone tool. This tool magnetically clamps to a cased borehole.
survey. The time-depth curve may be used in sonic log calibration
or stretching surface seismic data into depth. From the time-to-
depth picks, we can also extract an actual seismic interval velocity.
The kind of data that are recorded are shown in Figure 5. This velocity might be used in building a rock physics data base,
These are the vertical channel recordings for a six-source VSP surface seismic imaging, or as a constraint for seismic inversion. In
survey in the Pikes Peak heavy oilfield, Saskatchewan. The addition, by finding an angular dependence of velocity (or split
source offsets were 23m, 90m, 180m, 270m, 360m, and 450m. shear waves), we may be able to estimate anisotropy in the rock.
Continued on Page 81
80 CSEG Recorder September, 2001
VSP: AN IN-DEPTH SEISMIC UNDERSTANDING
Continued from Page 80
This anisotropy might be attached to the state of stress in the for-
mations or fracture intensity and direction. We can also use the fre-
quency loss, in the first few cycles of the downgoing energy, over
several levels to calculate an attenuation. This absorption or Q fac-
tor can be used as a rock property or in a Q-compensation tech-
nique with surface seismic data (Figure 6).
Figure 7. Reflectivity section generated from a vertical hydrophone
cable and surface dynamite shots (from Gulati, 1998).
Figure 8. We note that the gamma ray and sonic logs are in depth
(600m-1800m) as is the horizontal scale of the VSP. The VSP how-
ever, is also in two-way time on the vertical axis. This allows a direct
correlation of surface seismic data as well as synthetic data to the
well logs in depth. Because this particular VSP used an offset
source, we have created an offset image - both for P waves and con-
verted (PS) waves. These VSP images can also be correlated and
integrated into the interpretation.
Figure 6. The logartithmic spectral ratio plot to determine a Q value
from a 90 m offset VSP survey in the Pikes Peak heavy oilfield,
Saskatchewan (from Xu et al., 2001).
Another example of a composite plot, this time highlighting
The next level of sophistication would have us process a zero- Figure 8. A composite plot including well logs, VSP (depth and time),
offset survey for its reflectivity. In this case, the VSP is really pro- synthetic seismogram, VSP extracted trace (VET-a stack of the depth
viding a one-dimensional image. However, we do know quite a bit and time plot), a section of surface seismic data, and offset VSP section
about this single trace. It can be made largely multiple free and (in offset distance and time), a stack of converted-wave VSP data, and a
zero-phase. We’re confident with the zero-phase estimate because PS VSP section. The data are from Rolling Hills, Alberta.
we have measured the downgoing wave and can thus accomplish
a deterministic (signature-type) deconvolution. In addition, via the
composite or L plot, the VSP provides a unique mapping between converted waves, from the Cold Lake area, Alberta is shown in
seismic reflectivity in time and rock properties in depth. Figure 9.
If we have a source or sources offset from the well-head then Various other geometries can be employed to create specific pic-
we can make a lateral reflectivity section from the data. One such tures. For example, having receivers in a deviated or horizontal
image is shown below (Figure 7), where we used a 48-level verti- well can provide a high-resolution image below the borehole.
cal hydrophone cable and regularly spaced dynamite shots, to
offsets of 1500 m, to create an image a Glauconitic reservoir at a 3-D VSP
depth of 1500m. By using an areal distribution of shots (perhaps from a simul-
taneously conducted 3-D surface seismic survey), we can create a
By processing the horizontal channels, we can also make an 3-D image. Such a 3C-3D VSP was acquired over the Blackfoot oil
image of the converted waves. As mentioned previously, we can field in Alberta, Canada. This 3-D VSP was recorded simultane-
assemble well logs, synthetic seismograms, VSP data, and surface ously with a surface 3C-3D seismic program. Dynamite shots (4
seismic data into a compelling interpretive presentation. One such
composite plot from the Rolling Hills, Alberta area is shown in Continued on Page 82
September, 2001 CSEG Recorder 81
VSP: AN IN-DEPTH SEISMIC UNDERSTANDING
Continued from Page 81
Figure 9. Composite plot from measurements in the Cold Lake area.
Note the very high frequency content of the VSP as compared to the slice
of surface seismic section (from Sun, 1999).
kg at 18 m) from the surface 3C-3D seismic survey, that fell with-
in 2200m offset from the recording well were used in the VSP
analysis (Figure 10). As these 431 shots for the surface 3-D seismic
were being taken, a 5-level borehole tool moved seven times (75m
each) recording over a receiver depth range from 400m to 910m.
The 3-D VSP data were processed using basic VSP processing
techniques that included hodogram analysis, wavefield separation
using median filters, and VSP deconvolution. The final P-P and P-S Figure 10. Map of the Blackfoot surveys showing shot points for the sur-
image volumes were obtained by VSPCDP stacking the upgoing face 3C-3D and the 3C-3D VSP. A previous broad-band and a recent high-
resolution 3-C line are also shown in the figure (modified from Margrave
wavefields in 3-D cells followed by f-xy deconvolution (Figure 11).
et al., 1998).
The P-P and P-S sections from the 3-D VSP correlate well with those
from the surface 3C-3D seismic survey.
The discovery of PanCanadian Petroleum’s Blackfoot oilfield
was assisted by anomalies on P-wave 3-D seismic slices (from a
1993 survey). These anomalies may be related to gas in the upper
reservoir. However, many non-reservoir anomalies exist, too. A 3C-
3D seismic survey was conducted in 1995 to attempt to differentiate
reservoir from non-reservoir rock using converted waves. Figure 12
shows time slices at the reservoir (channel level) from the resultant
P and converted-wave volumes of the surface seismic survey. Both
show a north-south trend of the interpreted Blackfoot Glauconitic
channel body. The box on the time slices in Figure 12 indicates the
area of coverage of the 3-D VSP. The time slices in Figure 13 of the
3-D VSP indicate anomalies (reds and oranges are interpreted as
sand indicators) and possible additional targets for drilling.
There have been some remarkable images developed from bore-
hole sensors that monitor hydraulic fractures. Earthquake hypocen-
tral location techniques are used to plot the fractures as they occur
in time. This can give an excellent indication of the extent of a
hydraulic fracture. Furthermore, the is great promise in perma- Figure 11. P-wave section (north-south line) after VSPCDP stack, trace
nently emplacing motion sensors in a well to repeatedly monitor equalization, time-variant spectral whitening, and f-xy deconvolution
fracturing or image the area of interest for fluid and pressure (from Gulati, 1998).
Continued on Page 83
82 CSEG Recorder September, 2001
VSP: AN IN-DEPTH SEISMIC UNDERSTANDING
Continued from Page 82
the CREWES Project for their continued technical and financial
DiSiena, J.P., Byun, B.S., Fix, J.E., and Gaiser, J.E., 1984, F-K analysis and tube
wave filtering, in Eds. Toksoz, N.M. and Stewart, R.R., Vertical Seismic
Profiling, Part B-Advanced Concepts, Geophys. Press.
Gulati, J.S., 1998, Borehole seismic surveying: 3C-3D VSP and land vertical
cable analysis: M.Sc. Thesis, University of Calgary.
Margrave, G.F., Lawton, D.C., and Stewart, R.R., 1998, Interpreting channel
sands with 3C-3D seismic data: The Leading Edge, 4, 509-513.
Sun, Z., 1999, Seismic methods for heavy oil reservoir monitoring and characteri-
zation: Ph.D. Thesis, University of Calgary.
Xu, C., Stewart, R.R., and Osborne, C.A., 2001, Walkaway VSP processing and
Q estimation: Pikes Peak, Sask.: Presented at the CSEG 2001 Ann. Nat.
Convention, Calgary, Alberta. R
Figure 12. P-wave time slice at the channel level from the 3-D surface seis-
mic (modified from Margrave et al., 1998). The black rectangle in the
Figure shows the outline of the region mapped by the 3-D VSP survey.
ROBERT R. STEWART
Rob graduated from the University of
Toronto with a B.Sc. in physics and mathe-
matics and completed a Ph.D. in geo-
physics from the Massachusetts Institute of
Technology. He has been employed with
the Chevron Oil Field Research Company
in La Habra, California; Arco Exploration
and Production Research Centre in Dallas,
Texas; Veritas Software Ltd., Calgary, and
since 1987 has run his own geophysical
(a) (b) consulting company, GENNIX Technology
Figure 13. Time slices from the 3-D VSP, (a) P-P VSP time slice and (b) Corp.
P-S VSP time slice at the channel level. We interpret the red and orange Rob is a professor of geophysics at the University of Calgary
colours (amplitude values) as sand indicators and blues as indicative of and held the Chair in Exploration Geophysics from 1987-1997. He
shales. is the director of the CREWES Project, an industry-university con-
sortium studying advanced seismic methods in exploration that
was honoured with APEGGA’s Achievement Award in 1993 and
Borehole seismic measurements can be used for a number of
NSERC’s University-Industry Synergy Award in 1999.
purposes including the estimation of rock property values, seismic
propagation understanding, interpretive assistance, and stand- Rob is a past editor of the Canadian Journal of Exploration
alone imaging. The use of hydrophone receivers promises faster Geophysics, past associate editor for GEOPHYSICS, and a lecturer
and cheaper acquisition as do many-level geophone tools. The for the SEG Continuing Education Program. Rob and colleagues
have received Best Presentation, Paper, and Poster Awards from
VSP is continuing to develop as a sophisticated seismic log (for in
the CSEG and SEG. He was the CSEG Convention Technical
situ values including anisotropy). Converted-wave analysis is Chairman in 1992 and subsequently awarded the CSEG Medal.
considerably helped by 3-C VSP measurements. 3-D VSP imag- Rob was President of the CSEG in 1997-98 and completed the
ing is providing some very useful images. Permanently emplac- SEG’s inaugural Distinguished Educator Program: a 6-month
ing downhole sensors may provide a whole new realm of world lecture tour of 12 countries in 1999.
He is a member of the Canadian Space Agency’s Space
Exploration Advisory Committee, the SEG’s Continuing
ACKNOWLEDGEMENTS Education Committee, APEGGA’s Council, and belongs to a num-
I would like to thank Husky Energy Inc., Calgary, especially ber of organizations including Sigma Xi, AGU, and the EAGE.
Larry Mewhort, for expert counsel and support of the Pikes Peak Rob’s current professional interests include: motion sensor
effort as well as AOSTRA (Project #1296) for its assistance. The design, exploration seismology, borehole geophysical methods,
Blackfoot Seismic Project and PanCanadian Petroleum Ltd. made geostatistics, and natural hazards.
the 3C-3D seismic surveys possible. Finally, I thank sponsors of
September, 2001 CSEG Recorder 83