Downloaded 04/24/15 to 190.216.251.44. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.
org/
3D VSP in Lake Maracaibo, Venezuela
Carlos Planchart1 *, Alexander Palacios1, Min Lou1, Xiaomin Zhao1, Javier Gonzlez2, Maria Gabriela
Montaez, 2 Rmulo Guedez 2
1
VSFusion, 2 PDVSA
Summary
Maracaibo Lake is one of the most significant
petroleum basins in Venezuela. Conventional seismic
data recorded in Lake Maracaibo generally shows a
significant resolution loss with penetration depth,
along with ringing and linear noise from guided
waves. A 3D Vertical Seismic Profile (VSP) was
acquired in the Lake to image an area where the
surface seismic data was of very poor quality. In
contrast to the surface data, the quality of the 3D VSP
data was excellent. A vector migration algorithm was
applied to image the 3 component VSP data. The
image result presented in this paper shows a classic
example where 3D VSP is able to identify detailed
stratigraphic and structural geologic features with
greater resolution than surface seismic data.
Introduction
Lake Maracaibo, located in the northwest of
Venezuela, is one of the most significant petroleum
basins in the country. Conventional seismic data
recorded in Lake Maracaibo generally shows a
significant resolution loss with penetration depth,
along with ringing and linear noise from guided
waves. In addition the seismic wave properties of
scattering and absorption are also thought to
contribute to the poor quality exhibited by the surface
seismic data Previous studies showed that the
scattering and high attenuation are likely caused by
wave propagation through a gas-charged mud layer
having very low-compressional and shear wave
velocities (Perez, 2003).
The study area is located in south central Maracaibo
Lake, in the eastern part of the Lama-Icotea fault. The
reservoirs are located in the Eocene sequence, and are
composed of C sands of the Misoa formation. The
depositional system is tide-dominated delta and
displays mostly distributary channels. Structures of
Eocene sequences are associated with longitudinal
faults trending NE, and many transversal faults
trending NW-SE. These faults form a dense grid
which causes segmentation of the faulted blocks. The
dominant frequency of the surface seismic is 17 Hz in
the study area.
SEG/San Antonio 2007 Annual Meeting
In geologic areas that exhibit complex surface and/or
subsurface geological structure, borehole seismic has
proven to be an important tool for acquiring high
resolution seismic data. With VSP geometries
(receiver placed in the well and seismic sources on the
surface), the seismic energy will pass through the
highly attenuating near-surface layer only once, as
opposed to conventional surface seismic that uses
surface receivers. As the seismic receiver is placed
closer to the geologic zone of interest, the VSP image
is generally of higher resolution than that recorded
using surface seismic techniques. Some economic
benefits derived from 3D VSP include risk reduction
in well placement, improvement in reserve calculation,
and better understanding of reservoir stratigraphy and
compartmentalization.
Since the first 3D VSP was acquired in 1986, many
acquisition and processing tools have been developed.
One example of a new data processing tool is the 3C3D migration (Wang, 2004) which can take advantage
of the full vector response of the three-component
VSP receivers recorded by borehole geophones.
In this paper we present an application of 3D VSP to
image a prospect area in Lake Maracaibo, Venezuela,
where surface seismic data failed to clearly resolve
both structural and stratigraphic features. We shall
discuss the VSP survey design, data acquisition, and
data processing of this 3D VSP project.
3D VSP Pre-survey Modeling
In order to meet the survey objectives within
budgetary constraints, pre-survey modeling was
performed to obtain the optimum positions of the
downhole VSP receivers and the surface source
locations. Ray trace modeling was used for pre-survey
design. For simplicity, the three major geologic
reflection zones of interest were modeled. Velocity
information was obtained from nearby well
information.
The main objective of the 3D VSP modeling was to
obtain a regular illumination zone over the area of
interest, located east of the well. The chosen source
geometry was a spiral shot pattern with the spiral
center located to the east of the well (Figures 1 and 2).
The illumination (or fold number) distribution is
shown in Figure 3.
3080
3D VSP in the Maracaibo Lake, Venezuela
Downloaded 04/24/15 to 190.216.251.44. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/
The shot spacing was 40 m and the distance between
spiral arms was approximately 120 m. A total of 3000
shots (Figure 2) were acquired. The total time for the
3D VSP acquisition was 17 hours.
Data pre-processing
Figure 1. 3D VSP acquisition geometry
Figure 2. Spiral shot pattern
Data acquisition
The acquisition was performed by Baker Atlas. The
source was a sleeve-gun array. A 3-component, 12
level Sercel SST-500 downhole array tool was used to
record the 3D VSP. The spacing between each tool in
the array was 10 meters.
The raw recorded VSP data had a very high signal to
noise ratio with signal frequencies in the 80 Hz range.
The initial processing steps consisted of digital start to
time break correction, QC and updating of survey
geometry information, and first arrival time picking of
the 3-component data. 3-component rotation was
performed to take advantage of the full vector
response of the receivers. We first estimated the
orientation of the horizontal axes at each receiver
position using hodogram analysis of the two
horizontal components. With the vertical component
as the axis of rotation, the horizontal components were
rotated so that one of the components was aligned in
the plane of the direct arrival energy. We refer to the
horizontal component pointing in the direction of the
direct arrival energy as the radial component. The
other horizontal component is referred to as the
transverse component.
After horizontal rotation, wavefield separation was
applied using a 3-component parametric wavefield
separation algorithm. Parametric separation is
conceptually a beam-steering operation, where the
geophone array is directionally tuned to receive the
various P and S wave modes of interest. Using a
velocity model, incident angles are computed at each
receiver location for the upcoming and downgoing P
and S wave modes. Mode converted downgoing S
wave angles are computed from the model interface
immediately above the receiver. The incident angles of
upcoming reflected P waves and upcoming
(converted) S waves are computed for those waves
reflected from the model interface immediately below
the receiver. This velocity and angle information was
used to create a parametric model used to construct the
P and S wavefields in the frequency domain. The
decomposed P and S wave data were output in vector
format; each decomposed wavefield is output in the
vertical and radial directions.
Predictive deconvolution was applied to the separated
reflected P wavefield to compress the wavelet and
attenuate multiple arrivals.
Tomographic Velocity Model
Figure 3. Fold distribution of 3D VSP survey
A non-linear travel time tomography algorithm was
employed to determine the subsurface velocity field.
Processing steps include first arrival time picking,
generating an initial velocity model, and tomographic
inversion.
Figure 4 displays the velocity model after tomographic
inversion. The inversion converged quickly after 10
iterations (within 5 ms RMS tolerance error). The
SEG/San Antonio 2007 Annual Meeting
3081
Downloaded 04/24/15 to 190.216.251.44. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/
3D VSP in the Maracaibo Lake, Venezuela
RMS travel time error after the tomographic inversion
is presented in Figure 5. This velocity model was used
in the depth migration of the processed 3D VSP data.
Results
Figure 4. Velocity model after tomographic inversion
3D VSP can also help in the interpretation of the
sedimentation environment. For example, Figure 9
shows a sandstone, with shaly interbeds between
12068 and 12133 ft. The blocky response of the
gamma ray log indicates a probable distributary
channel. Figure 10 shows depth slices of trace
amplitude for the same interval every 20 feet between
12063 and 12163 ft. (red: positive amplitude, blue:
negative amplitude). In the depth slices we can see the
lateral variation of amplitude, delineating the
distributary channel.
The area in the blue circle in Figures 7 and 8 is a
comparison between the surface seismic and the 3D
VSP image results. . The vertical resolution of surface
seismic data was estimated to be 195 ft and the
resolution of the VSP data to be 80 ft. Estimates are
based upon the velocity of the Misoa sands (13300
ft/sec) and the central frequency of 3D VSP data being
40 Hz. The 3D VSP image clearly exhibits
significantly higher resolution than the surface seismic
section in both the vertical and horizontal directions.
The detailed thin stratigraphic units and small faults
are easily identified in the 3D VSP image result,
greatly helping the interpreters and geologists map the
reservoir and to identify new well locations for
drilling.
Conclusions
Figure 5: Tomographic inversion RMS errors vs.
number of iterations.
3C-3D Migration
The processed 3D VSP data was migrated using a 3component 3D vector pre-stack Kirchhoff depth
migration (3C-3D). The 3C-3D can image the full 3component reflected wave field allowing the
reconstruction of true reflection amplitudes and at the
same time reducing unwanted migration smearing and
other false mirror image reflections inherent to single
component Kirchhoff migration. The program can
handle both 3D isotropic and 3D tilted transverse
isotropic (TTI) velocity model by using an efficient
and accurate fast-marching method to calculate travel
times (Lou, 2006). The 3C-3D can image any multicomponent (1C, 2C, or 3C) VSP data with various
wave-modes (P-P, P-S, or S-S). In this project,
vertical and horizontal radial component vectormigration was applied to the separated up-going Pwave reflection wavefield. Figure 6 shows several
vertical and horizontal slices in depth from the
migrated 3C-3D image cube.
SEG/San Antonio 2007 Annual Meeting
This is the first 3D VSP that has been acquired in
Venezuela. A 3D VSP seismic cube of approximately
2.3 Km2 was obtained in an area where previously
acquired surface seismic data had poor image quality.
Our 3C-3D migration provided excellent image results
for the 3D VSP data, which allows identification of
detailed structural and stratigraphic structures with a
significantly greater resolution than surface seismic
data. This 3D VSP image has helped to greatly
improve the geological interpretation and to reduce the
risks related to placement of new wells in the area.
Acknowledgements
We thank PDVSA for allowing us to publish this
work. We also thank Simon Shaw, leader of the Baker
Atlas seismic team, for his teams excellent work
during the data acquisition.
3082
Downloaded 04/24/15 to 190.216.251.44. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/
3D VSP in the Maracaibo Lake, Venezuela
Figure 6. - Vertical and horizontal slices
Figure 7: A comparison of 3D VSP (right panel) and
surface seismic data (left panel)
Figure 8. Comparison between surface seismic data (left) and 3D VSP image (right).
Figure 10. Depth slices of amplitude trace between
12063 and 12163 feet, every 20 ft.
Figure 9. Well logs with top and base of sandstone
related to distributary channel
SEG/San Antonio 2007 Annual Meeting
3083
Downloaded 04/24/15 to 190.216.251.44. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/
EDITED REFERENCES
Note: This reference list is a copy-edited version of the reference list submitted by the author. Reference lists for the 2007
SEG Technical Program Expanded Abstracts have been copy edited so that references provided with the online metadata for
each paper will achieve a high degree of linking to cited sources that appear on the Web.
REFERENCES
Lou, M., 2006, Efficient and accurate traveltime computation in 3D TTI media by the fast marching method: 76th Annual
International Meeting, SEG, Expanded Abstracts, 184188.
Perez, J. A., 2003, Study of the seismic attenuation generated by the mud layer in Lake Maracaibo, Venezuela: M. S. thesis,
Texas A&M University, http://handle.tamu.edu/1969.1/414, accessed March 15, 2007.
Wang, D., 2004, Vector 3C3D VSP Kirchhoff migration: 74th Annual International Meeting, SEG, Expanded Abstracts, 2458
2461.
SEG/San Antonio 2007 Annual Meeting
3084
