SPECIAL TOPIC: MARINE SEISMIC

Source and streamer towing strategies for

improved efficiency, spatial sampling and near
offset coverage
Andrew Long1* presents a study of the fundamental relationships between cross-line source
configurations and towed streamer survey efficiency and spatial sampling using a set of
robust geometric relationships.

Introduction Source separation and CMP bin width

Several towing concepts have arisen in recent years that break Figure 1 shows a schematic source and streamer configuration
the convention of towing two source arrays between the inner- with two sources and ten streamers. The nomenclature used
most two streamers in a multi-streamer 3D configuration (‘dual- defines the basic geometry wherein the source separation is based
source shooting’). Three ‘arrays’ of one or more sub-arrays were on the nominal streamer separation, and the streamer separation is
towed more than 20 years ago to improve cross-line spatial divisible to a whole number by the subline separation.
sampling, but inline spatial sampling and fold were compro- For S sources (S = 2 or 3), a source separation that is an integer
mised by inefficient recycling times on air gun compressors and k multiple of the streamer separation L (i.e. kL) results in a nominal
limited recording lengths. Modern acquisition systems enable bin size equal to 1/S of the streamer separation, whereas towing
continuous recording, very short physical shot intervals, and up each source array with a separation of (k + 1/S)L results in a nomi-
to six source arrays being deployed between the innermost two nal bin width equal to 1/2S of the streamer separation. For example,
streamers; always with the ambition of improving cross-line the improved cross-line spatial sampling of triple-source shooting
spatial sampling. enables 12 streamers at 150 m separation to have equivalent
I discuss two newer variations to these scenarios using either cross-line spatial sampling and sail line efficiency to 18 streamers
dual-source or triple-source shooting, although the principles may at 100 m separation, but the inline shot spacing needs to be scaled
be expanded to more sources distributed in the cross-line direc- by a factor of 2/S to retain comparable CMP fold and inline spatial
tion: 1. Towing source arrays outside the innermost two streamers sampling in pre-stack gathers. If the sources for both dual-source
such that survey efficiency is improved courtesy of the nominal and triple-source shooting are between the innermost two streamers
sail-line separation being increased. It is demonstrated that (k=0) the only cross-line difference between dual-source and
cross-line fold becomes irregular as source separation increases triple-source shooting will be in terms of spatial sampling, and
in the cross-line direction, so complementary processing methods the nominal fold of each subline will be equal for all sublines in
may be required to reduce imaging artifacts. 2. Alternatively, if each sail line. When the sources for dual-source shooting or the
the sail-line separation is based on a conventional source towing outermost two sources for triple-source shooting are placed outside
scenario, some flexibility may be created in terms of near offset the innermost two streamers (k>0) the cross-line spatial sampling
sampling. is unchanged, but the nominal subline fold becomes irregular. I

Figure 1 Schematic example for dual-source

shooting. In the nomenclature used throughout this
paper the number of sources, S, is 2 in this example,
and the number of streamers, N, is 10. The near
offset is shown for trace 1 of subline 1. Geometric
relationships can be found between the subline
separation, the sail line separation, and S, N, and the
streamer separation, L.

1
PGS
*
Corresponding author, E-mail: [email protected]

FIRST BREAK I VOLUME 35 I NOVEMBER 2 017 71

 SPECIAL TOPIC: MARINE SEISMIC

Figure 2 (upper) Cross-line ray path schematic for

a conventional dual-source towing configuration
with 16 streamers and overhead perspective of the
sublines for two adjacent sail lines; (lower) Equivalent
plot for triple-source shooting and a 50% increase
in streamer separation. Sail line 1 in both scenarios
is represented by black sublines, and sail line 2 is
represented by blue sublines.

illustrate how this can be predicted below, comment on how wider consecutive shots to allow sufficient decay of shot energy
source towing can be used to improve sail line efficiency, and (Martin Landrø, personal communication).
comment on how wider source towing can be used to improve near The upper panel of Figure 3 shows the same 16 streamer
offset coverage for each subline. configuration where the source separation is now 1.5L, i.e.
k=1 and S=2 (dual-source shooting with the sources placed
Improved sail line efficiency outside the innermost two streamers), and the subline separation
The upper panel in Figure 2 shows a 16 streamer configuration remains L/2S, i.e. L/4. Note, however, that the nominal number
where the source separation is 0.5L, i.e. k=0 and S=2 (conven- of sublines for each sail line increases from SN (in this case
tional dual-source shooting), and the subline separation is L/2S, 32) to SN + Sk, i.e. 34, and the nominal sail line separation has
i.e. L/4. The black lines on the left represent the sublines for the correspondingly increased by 0.5kL. The number of nominal fold
first sail line, the blue lines on the right represent the sublines sublines remains unchanged, but Sk sublines have zero fold for
for the second sail line, and the nominal sail line separation is each sail line (the red lines), assuming no streamer fanning is
0.5(N+k)L, where N is the number of streamers. In the lower used and the shooting geometry is perfectly uniform. If this zero
panel of Figure 2 the number of sources has been increased by fold subline can be addressed by some form of regularization/
50% (S=3), the number of sublines has increased by 50%, the reconstruction in processing, the nominal sail line separation can
streamer separation has been increased by 50%, and as a con- be increased from 0.5NL to 0.5(N+1)L. In other words, for dual-
sequence the subline separation is unchanged but the nominal source shooting, the nominal sail line separation increases by half
sail line separation and therefore the ‘sail line efficiency’ has of the increase in the source separation if the zero fold sublines
been increased by 50%. If the vessel is capable of towing a 50% can be accommodated in processing.
wider streamer spread it follows that this sail line efficiency The lower panel of Figure 3 shows shows the same 16
can be realized, otherwise the number of streamers can be streamer configuration where the source separation is now 2.5L,
reduced to yield an achievable spread width that nevertheless i.e. k=2 and S=2 (dual-source shooting with the sources placed
has a roughly comparable sail line efficiency and lower streamer outside the innermost two streamers), and again the subline sep-
inventory usage. The penalty is that the shot interval must be aration remains L/2S, i.e. L/4. Following the principles observed
reduced by one-third to maintain equivalent CMP fold and in the upper panel of Figure 3, the nominal number of sublines
equivalent trace separation in the common receiver, common for each sail line increases to 36, the nominal sail line separation
offset, and common midpoint domains. Shorter shot intervals has increased by another 0.5L, and there are 2k (i.e. 4) zero fold
necessitate recording overlap, particularly when long records sublines. This pattern will continue until the sources are placed
are desired. Separation of overlapping/blended shot gathers in outside the outermost two streamers in the spread such that k=N,
signal processing inevitably have some cost to data fidelity, the source separation is (N + 0.5)L, there are 4N sublines, half
and residual shot energy from preceding shots can also create of which are zero fold (so the effective subline separation has
unwanted noise, particularly when the shot interval is short. doubled), and the nominal sail line separation is NL, i.e. sail line
Anecdotal estimates state that 18 seconds is required between efficiency is twice that for conventional dual-source shooting

72 FIRST BREAK I VOLUME 35 I NOVEMBER 2 017

 SPECIAL TOPIC: MARINE SEISMIC

with the sources towed between the innermost two streamers. Source separation
These same principles also apply to triple-source shooting
(S=3) where the centre source remains midway between the
innermost two streamers but the outer two sources are increased Subline separation (bin width)
in separation (e.g. Figure 4). In such scenarios there will be
various pairs of adjacent zero fold sublines incurred as the outer Sail line separation
source separation increases, and the increase in nominal sail line
separation is half the increase in outer source separation. Table 1 Total number of sublines per sail line
summarizes the geometric relationships discussed here, assuming
in all scenarios that the acquisition geometry is uniform and Number of zero fold sublines per sail line
streamers are parallel with no feathering. Table 1 Relationships between geometric parameters for towed streamer
Overall, we see that sail line efficiency for dual-source acquisition with two or more sources. L = streamer separation, N = number of
and triple-source shooting varies between ‘1’ (conventional streamers, S = number of sources, k is an integer.
configuration) and ‘2’ (outer sources towed outside the outermost
streamers); in other words, sail line efficiency can be doubled if streamer spread with L/2 separation in the case of S=2 (four
the (outer) sources are moved outside the streamer spread. This sources in total), or triple-source arrays outside the streamer
has logistical challenges discussed later, and effectively doubles spread with L/3 separation in the case of S=3 (nine sources in
the subline separation. A solution to the compromised cross-line total). Such considerations would inevitably involve significant
spatial sampling is to tow either dual-source arrays outside the shot blending due to the necessity for short shot intervals.

Figure 3 (upper) Cross-line ray path schematic

for a wide source towing with two sources (source
separation = 1.5 x streamer separation) configuration
with 16 streamers and overhead perspective of the
sublines for two adjacent sail lines; (lower) Equivalent
plot for source separation = 2.5 x streamer separation.
Sail line 1 in both scenarios is represented by black
sublines, sail line 2 is represented by blue sublines,
and red represents zero fold sublines.

Figure 4 Cross-line ray path schematic for a wide source towing with three sources (source separation = 1.5 x streamer separation) configuration with 16 streamers. Note the
larger fold gap at outer sublines by comparison to the upper panel of Figure 3.

FIRST BREAK I VOLUME 35 I NOVEMBER 2 017 73

 SPECIAL TOPIC: MARINE SEISMIC

Figure 5 (upper) Cross-line ray path schematic

for a wide source towing with two sources (source
separation = 1.5 x streamer separation) configuration
with 16 streamers and overhead perspective of
the sublines for two adjacent sail lines. Sail line
separation is the same as for conventional dual-
source shooting; (lower) Equivalent plot for source
separation = 2.5 x streamer separation. Again, sail
line separation is the same as for conventional
dual-source shooting. Sail line 1 in both scenarios
is represented by black sublines, sail line 2 is
represented by blue sublines, and red represents zero
fold sublines. Note in both scenarios that there are no
longer zero fold sublines at each sail line boundary.
Compare to the upper panel of Figure 2.

Mitigation of zero fold CMP sublines fold in each subline shown as having ‘zero fold’ in the schematic
Continuing the simplistic assumptions of no streamer fanning illustrations of Figures 3 to 5, and interpolation/reconstruction
being used and that the shooting geometry is perfectly uniform, in processing may yield uniform CMP fold for all offsets and all
it is shown that retaining the nominal sail line separation of sublines.
k=0 for each source separation scenario will yield nominally
uniform CMP fold everywhere, but the sublines around each Conclusions
sail line boundary will alternately correspond to each sail line I have illustrated the fundamental relationships between source
as illustrated in the upper and lower panels of Figure 5. In other configurations and towed streamer survey efficiency and spatial
words, the sublines overlap in a manner at each sail line boundary sampling for the ‘conventional’ scenario where all source arrays are
that ‘cancels’ the zero fold sublines. The upper panel of Figure 5 towed between the innermost two streamers, and then for the sce-
is the configuration in the upper panel of Figure 3 with nominal narios of increasingly large source separations and sources being
sail line separation reduced back to 0.5NL, and the lower panel of towed outside the innermost two streamers. ‘Sail line efficiency’
Figure 5 is the configuration in the lower panel of Figure 3 with increases with increasing source separation if a predictable pattern
nominal sail line separation reduced back to 0.5NL. Note that of zero fold sublines centered around the sail line boundaries can
the source-receiver azimuth will vary in an alternating manner be accommodated in signal processing and imaging. The number
in this ‘overlap’ region as adjacent sublines correspond to source of zero fold sublines increases with increasing source separation.
locations from different sail lines, and irregular streamer and sail Alternatively, if the sail line separation is not adjusted, being based
line geometry will affect CMP fold uniformity too. upon the nominal sail line separation for ‘conventional’ source
So why ‘undo the efficiency gain’ by reducing sail line sep- towing, the zero fold sublines are mitigated by the finite fold contri-
aration? The answer is that near offset coverage for each subline butions from the sublines of the adjacent sail lines in an interleaved
can be improved by comparison to the upper panel of Figure 2 manner. Sail line efficiency is therefore not changed, but the near
while maintaining (relatively) uniform CMP fold on all sublines. offset distribution will be changed for each subline.
This issue will be addressed within a future paper.
In practice the use of streamer fanning and natural variations Acknowledgements
in streamer and sail line geometry will probably result in finite I would like to thank PGS for permission to publish this material.

74 FIRST BREAK I VOLUME 35 I NOVEMBER 2 017