282
`
`Marine II
`
`time
`
`ISECONDS)
`
`02
`I
`
`03
`!
`
`04
`I
`
`05
`/
`
`06
`1
`
`07
`\
`
`08
`,
`
`09
`I,
`
`10
`
`11
`I,
`
`12
`
`FIG. 4. Signature 33 degrees behind array. Scale factor =
`0.495.
`
`(Figure 2). The scale factors shown in the figures give the
`relative amplitude scaling between these three figures.
`
`Modeled radiation patterns
`There are many parameters that may have an effect on the
`observed radiation patterns. Some of these are the 2-D array
`geometry, air gun distribution within the array, variations in
`the firing times of the air guns, and variations in the depths of
`individual air guns. The radiation pattern due to the array
`geometry can be modeled using equally weighted ideal
`dipole sources (point source plus free-surface ghost) at each
`air gun position within the array, and summing these at
`desired observation points. Such modeling predicts an iso-
`tropic radiation pattern for the square array, symmetrical
`directionality with more energy in the in-line vertical plane
`for the wide array, and symmetrical directionality with more
`energy in the crossline vertical plane for the long array.
`Other observed radiation pattern properties for these three
`arrays must therefore be due to other factors.
`Modeling of the far-field radiation patterns using measured
`near-field air gun signatures allows the effect of the gun
`distribution within the arrays to be observed. Because the
`subarrays are designed with the largest guns in the front and
`the smallest guns in the rear. some forward directionality
`results from gun distribution. This elect does not completely
`account for the strong forward directionality observed for
`the wide array.
`The firing times of each individual air gun and its depth
`can also be incorporated into the model. Because the firing
`times were closely controlled and monitored during data
`collection, they will not have a significant effect on the
`radiation patterns. The depths of the air guns were also
`monitored during data collection and displayed a rather wide
`variation within the arrays. Variations of more than 2 m
`above or below the nominal air gun depth of 10 m were
`observed. even within a single subarray. These gun depth
`distributions were also observed to be slowly varying from
`shot to shot. Including these observations in the radiation
`
`pattern modeling scheme provides yet a better prediction for
`the observed radiation patterns.
`In conclusion, the observed data confirm our ability to
`make field measurements of 3-D radiation patterns, allowing
`us to observe the effectiveness of various array designs. For
`instance, wide arrays are designed to minimize out-of-line
`scattered energy and maximize in-line energy. The observed
`data show that the range of take-off angles for which the
`energy is less than 6 dB down from the peak energy is nearly
`twice as wide in the in-line plane for the wide array as it is for
`the square array. Similarly, long arrays are designed to
`minimize in-line multiple energy. The observed data show
`that the radiation pattern, as measured above, is almost one-
`fourth as wide in the in-line plane and nearly twice as wide in
`the crossline plane for the long array as it is for the square
`array. These data can also be used to demonstrate our ability
`to model source array radiation patterns. Such modeling,
`incorporating monitored air gun depths, allows us to observe
`the importance of various parameters, particularly air gun
`depth control. in tailoring the source array signature.
`
`MAR2.8
`Three-Dimensional Air Gun Arrays
`G. C. Smith, Southern Oil Exploration Corp., South Africa
`In a marine air gun array composed of subarrays, the
`depth of each subarray can be different. To ensure a vertical-
`ly downgoing wave field, the firing of each subarray is
`delayed by a time which depends on the depth. In this way
`the ghost reflection can be suppressed and the peak-to-
`bubble ratio can be improved. Care must be taken in the
`arrangement of such subarrays to ensure acceptable energy
`emission characteristics in directions away from the vertical.
`The experiment described in this paper can be extended to
`the design of very broad-band high resolution sources.
`
`Introduction
`The historical development of air gun arrays followed a
`number of stages. The oscillatory signature of a single air
`gun was largely overcome by the use of a number of guns of
`different sizes, to a greater or lesser extent interacting, to
`produce a signature with a large primary-to-bubble ratio
`(Giles and Johnstone. 1973; Nooteboom, 1978; Brandsaeter
`et al.. 1979).
`The arrangement of the air guns (or other sources) into
`spatial arrays became a subject of much interest. Arrays in
`yhich sources have been arranged in the in-line direction to
`act as spatial filters were described by Newman et al. (1977),
`Lofthouse and Bennett (1978). and Ursin (1978. 1983).
`The deployment of extended source arrays in the crossline
`direction was also developed (Parkes et al., 1981; Tree et al.,
`1982). Such arrays act as spatial filters in a direction across
`the survey line, and serve principally to suppress noise
`scattered from near-surface anomalies (Lamer et al.. 1983).
`This paper extends the concept of spatial arrangements of
`sources into the third, \ :rtical dimension.
`
`Vertical distribution of seismic sources
`In the case of marine recording with an array of sources.
`individual elements making up the array can be towed at
`different depths, and fired at different times. The shallowest
`
`Downloaded 11/19/15 to 69.140.6.94. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/
`
`PGS Exhibit 1006 - PGS v. WG
`
`

`
`Marine II
`
`283
`
`element is fired first, the next shallowest second, with a time
`delay equal to the depth difference divided by the velocity of
`sound through water, and so on. The result is that vertically
`beneath the array the primary energy sums in phase while
`the ghost is spread out in time
`The idea of distributing the elements of a seismic source in
`the vertical direction, with time delays, is not a new one. It
`was the subject of a patent (Prescott, 1935) and variations on
`the same theme have been described for seismic surveys on
`land many times. the result being achieved by the progres-
`sive detonation of a long charge or multiple charges, or by
`summing separate recordings made with shots at different
`depths in the same shot-hole (Shock, 1950; Van Melle and
`Weatherburn, 1953; Musgrave et al., 1958; Seabrooke, 1961;
`Hammond, 1962; Sengbush, 1962; Martner and Silverman,
`1962; Fail and Layotte, 1970).
`
`Field parameters
`The energy source parameters used in the experiment
`described here were as follows. The air gun array consisted
`of four identical subarrays, each consisting of 7 different
`sized guns so chosen and arranged to provide a signature
`with a large peak-to-bubble ratio at a range of depths. The
`total capacity of the array was 5 560 inch’.
`Each subarray was 19 m long, with large guns at the front
`and small guns at the back. This geometry was considered to
`provide a good in-line spatial antialias filter for 25 m channel
`
`.w
`
`,
`
`-2e.Y
`
`DB
`
`I
`
`-u’-
`
`I/
`
`FIG. 2. (a) Amplitude spectrum of Figure la. (b) Amplitude
`spectrum of Figure 1 b.
`
`separation in the cable, and therefore the subarrays were not
`spread out in the in-line direction.
`Across the direction of the line, the subarrays were
`positioned 15 and 37.5 m to port and starboard of the center
`line, making a four element array 75 m wide. Newman (1983)
`pointed out that the CDP stack suppresses noise scattered
`from directions directly to the side of the line, and that the
`“dangerous” direction is closer to the in-line direction than
`to the crossline direction. However, we felt that the noise
`should be suppressed early in the acquisition and processing
`sequence in order to provide maximum signal-to-noise ratio
`for prestack processing. Thus a wide array was chosen with
`dimensions appropriate to the suppression of side-scattered
`noise in the main seismic frequency band.
`The vertical distribution was obtained by towing the four
`subarrays at depth of 5.4, 7.2, 9.1, and I I.0 m with delays of
`0, 1.25,2.5 and 3.75 ms, respectively. These delays are small
`enough not to affect significantly the spatial filtering action
`of the wide array.
`
`Far-field signatures
`Far-field signatures, together with their amplitude and
`phase spectra, were computed for the vertically distributed
`array and for an array towed at a uniform depth of 9.1 m. The
`far-field signatures were calculated from near-field signa-
`tures by the method described by Ziolkowski et al. (1982).
`The signatures are shown in Figure 1. The field filters used
`were 5.3 Hz, I8 dB/octave lowcut and 64 Hz, 18 dB/octave
`high-cut. The amplitude and phase spectra are illustrated in
`Figures 2 and 3, respectively. The uniform depth array
`shows a notch in the amplitude spectrum with an associated
`
`FIG. 1. (a) Far-field signature of vertically distributed array.
`(b) Far-field signature of uniform depth array. Both recorded
`with field filters 5.3 Hz, 18 dB/oct; 84 Hz, 18 dB/oct.
`
`Downloaded 11/19/15 to 69.140.6.94. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/
`
`PGS Exhibit 1006 - PGS v. WG
`
`

`
`Marine II
`
`1
`
`FIG. 3. Phase spectrum of Figure la. (b) Phase spectrum of
`Figure 1 b.
`
`in the case of the
`is not present
`phase ambiguity which
`vertically distributed array. A further advantage shown by
`the vertically distributed array
`is the bubble suppression.
`This is because the depth of a gun influences the bubble
`period. so that the variation
`in gun depths gives further
`variation
`in bubble periods beyond
`that obtained by using
`guns with dif‘erent capacities.
`The depths chosen for thi, experiment. 5.3. 7.1. 9.1, and
`I1 .O m were chosen to fit certain constraints.
`If subarrays
`are too shallow. too much energy
`is lost. and if they are too
`deep. peak-to-huhhlc
`ratio3 become poor. Also the streamer
`depth wah constrained by difficult sea conditions, and thi\
`limited the frequencies which it was useful to introduce
`into
`the ground. However,
`combinations of depths can be de-
`signed to make high resolution sources, especially with
`wx~~xxs where bubbles do not need to be taken into account.
`For injtancc. sources at depth\ of 3.75. 7.5. 11.75. and 1 m
`with delays of0. 5. IO and 15 mc. reqxxztively. give rise to a
`ghost operator whose ilmplitude spectrum is less than 3.5 dB
`IX7 Hz. The ghost operator t’or a
`down between
`13 and
`uniform depth of3.75 m has an amplitude spectrum less than
`3.5 JB down only between 37 and I53 HL
`The danger in this approach
`IIt‘5 in the possibility of setting
`up undc\irahle energy cmisGon characteristics
`in directions
`from the vertical.
`
`way
`
`Horizontal arrangenwnt of subarrays
`
`in which it is possible
`the 12 ditl2rcnt wily\
`Figure 4 \how\
`to arrange the I'CILII. \trb;uxlyj
`at the li)ur JiiTerent depths.
`
`FIG. 4. The 12 different ways of arranging four subarrays in
`a wide array with depths of 5.4, 7.2, 9.1, and 11 .O m. Vertical
`exaggeration is 2 : 1.
`
`,*a*aaaa..b8blia;b;
`”
`I""
`
`AzI""I* 11111111,
`”
`”
`
`”
`
`"'1
`
`(b)
`
`1
`FIG. 5. The relative times of primary and ghost arrivals
`according to the angle from the vertical in the plane perpen-
`dicular to the seismic line. Primary arrivals are represented
`by the solid lines, ghosts by the dashed lines: (a) corre-
`sponds to arrangement 1 in Figure 4; and (b) corresponds to
`arrangement 10 in Figure 4.
`
`Downloaded 11/19/15 to 69.140.6.94. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/
`
`PGS Exhibit 1006 - PGS v. WG
`
`

`
`Marine II
`
`Conclusion
`
`285
`
`Suppression of ghost reflections from air gun arrays can be
`obtained by the vertical distribution of subarrays. The same
`will apply to other sources, also, and if such sources do not
`generate bubbles then individual sources can be considered
`rather than subarrays.
`In the case of air guns the vertical
`distribution of subarrays also improves
`the peak-to-bubble
`ratio.
`
`Acknowledgments
`I would like to thank Gregg Parkes of Merlin Profilers Ltd
`for providing several of the diagrams in this paper.
`
`References
`Brandsaeter, H., Farestveit, A., and Ursin. B.. 1979. A new high-
`resolution or deep penetration air gun array: Geophysics, 44.865-
`879.
`Fail. J. P., and Layotte, P. C., 1970. MCthode de filtrage du
`fant6me: application g des cas reels: Geophys. Prosp., 18, 434-
`464.
`Giles. B. F.. and Johnston, R. C.. 1973, System approach to air gun
`array design: Geophys. Prosp.. 21, 77-101.
`Hammond, J. W., 1962, Ghost elimination from reflectionrecords:
`Geophysics, 27,48-w.
`Lamer, K., Chambers, R., Yang, M., Lynn, W., and Wai, W., 1983,
`Coherent noise in marine seismic data: Geophysics, 48, 854-886.
`Lofthouse. J. H.. and Bennett. G. T.. 1978. Extended arravs for
`’
`marine seismic acquisition: Geophysics, 4j. 3-22.
`Martner, S. T., and Silverman, D., 1962. Broomstick distributed
`charge: Geophysics. 27, 1007-1015.
`Musgrave, A. W.. Ehlert. G. W., and Nash. D. M.. Jr.. 1958.
`Directivity effect of elongated charges: Geophysics, 23, 81-96.
`Newman, P., Small, J. O., and Waites, J. D., 1977, Theory and
`application of water gun arrays in marine seismic exploration:
`Presented at the 47th Annual SEG Meeting, Calgary.
`Newman, P.. 1983, Seismic response to sea floor diffractors: Pre-
`sented at the 53rd Annual SEG Meeting, Las Vegas.
`Nooteboom, J. J.. 1978. Signature and amplitude of linear air gun
`arrays: Geophys. Prosp., 26. 194-201.
`Parkes, G. E., Hatton. L.. and Haulzland. T.. 1981. Marine source
`array directivity - A new wide air&n array system: Presented at
`the 5lst Annual SEG meeting Los Angeles.
`Prescott, H. R.. 1935, Method”bf making geological explorations:
`U.S. Patent No. 1,998,412: filed March 29, 1934.
`Seabrook, D. S., 1961, Anomalous events on the reflection seismo-
`gram: Geophysics. 26, 85-99.
`Sengbush, R. L.. 1962. Stratigraphic trap study in Cottonwood
`Creek field, Big Horn basin, Wyoming: Geophysics, 27.427-444.
`Shock, L., 1950, The progressive detonation of multiple charges in a
`single seismic shot: geophysics 15. 208-218.
`Tree;E. L., Lugg. R. D.. &I Brummitt, J. G. IYX?. The attenua-
`tion of source generated noise in marine seism’ic using areal arrays
`of water guns: Presented at the 52nd Annual SEG Meeting.
`Dallas.
`Ursin, B.. 1978, Attenuation of coherent noise in marine seismic
`exploration using very long arrays: Geophys. Prosp., 26, 7X-749.
`Ursin, B., 1983, Spatial filtering of marine seismic data: Geophysics.
`48, 161 I-1630.
`Van Melle. F. A.. and Weatherburn. K. R.. 1953. Ghost reflections
`caused bv energy initiallv reflected above the’level of the shot:
`Geouhvs&s. IX. 7Y%-X04:
`Ziolkowski, A.. Parkes, G., Hatton. L.. and Haugland. T.. 1982.
`The signature of an air gun array: Computation from near-field
`measurements including interactions: Geophysics, 47. 1413-1421.
`
`FIG. 6. Directivity plots for the array: (a) corresponds to
`arrangement 1 in Figure 4; (b) corresponds
`to arrangement
`10 in Figure 4.
`
`times of primary and ghost
`Figure 5 shows the relative
`arrivals in the far-field as a function of angle in the plane
`to the seismic line. for two of the possible
`perpendicular
`arrangements.
`In Figure 5a the subarrays are arranged
`in
`order of increasing depth from left to right (5.4; 7.2; 9. I; I I .O
`m), and it can be seen that a wave
`is set up at about 84
`degrees to the vertical, caused by the constructive
`interfer-
`ence of ghost reflections.
`In Figure 5b the depths of the
`(9. I; 5.4; I I .O; 7.2 m), and the
`subarrays are “randomized”
`ghost reflections are well spread out in time at all angles.
`Directivity plots for these two arrangements are shown in
`Figure 6; they are noticeably different. There is more energy
`emission close to the vertical
`in 6a (corresponding
`to the
`sequential depth arrangment)
`than in 6b (corresponding
`to
`the random depth arrangement). Of course, the total energy
`emitted by the two configurations is the same, however,
`the
`distribution with angle differs. Consequently,
`the decreased
`in Figure 6b i5 balanced by increased
`mainlobe emission
`sidelobe emission. However, because of its narrower main-
`lobe emission, particularly at low frequencies,
`the random-
`ized arrangement of \ubarray depths was chosen for the
`experiment.
`
`Downloaded 11/19/15 to 69.140.6.94. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/
`
`PGS Exhibit 1006 - PGS v. WG