3d deghosting for full-azimuth and ultra-long offset ... · 3d deghosting for full-azimuth and...

3D deghosting for full-azimuth and ultra-long offset marine data Qiaofeng Wu*, Chang-Chun Lee, Wei Zhao, Ping Wang, Yunfeng Li, CGG

Summary

Removing ghost energy in marine streamer data is

important for both seismic processing and interpretation,

especially for 3D surface-related multiple elimination

(SRME) and improving bandwidth and resolution. Full

azimuth data have abundant azimuths, which require more

robust 3D deghosting techniques. The coarse and irregular

crossline sampling in full azimuth (FAZ) seismic data

creates challenges for 3D deghosting. A fully data-driven

3D deghosting technique using a progressive sparse Tau-P

inversion has proven to be able to overcome the sparse

sampling in the crossline direction and to be effective in

attenuating the receiver ghost in a 3D mode. We

demonstrate the benefits of 3D deghosting using a

staggered FAZ and ultra-long offset data set from Keathley

Canyon, Gulf of Mexico. Using the data-driven 3D

deghosting method, we observed less residual ghost energy

in shot gathers from a side-gun when compared with the 2D

pre-migration bootstrap deghosting method. The 3D

deghosting method subsequently improved the images of

steeply-dipping top of salt (TOS) and assisted with multiple

removal.

Introduction

Marine streamer data record both primary energy and ghost

energy. The ghost’s destructive interference with the

primary energy generates notches in the amplitude

spectrum and limits the usable frequency range. Removing

the ghost energy can fill the ghost notches and provide

broader spectrum bandwidth and an improved signal-to-

noise ratio (S/N), which are all beneficial for both seismic

imaging and interpretation. Several pre-migration (Wang

and Peng, 2012; Wang et al., 2013; Poole, 2013) and post-

migration (Soubaras, 2010) deghosting methods have been

proposed to remove the ghost energy and improve imaging

results.

To address the imaging challenges in the deepwater Gulf of

Mexico (GOM), Mandroux et al. (2013) developed a

staggered acquisition configuration with variable-depth

streamers. The configuration produces FAZ coverage up to

9 km and ultra-long offsets up to 18 km. The FAZ and

ultra-long offsets in this new acquisition design have shown

benefits in bandwidth extension, multiple suppression (Yu

et al., 2013), velocity model building (Mothi et al., 2013;

Wu et al., 2013), and improved subsalt illumination (Wu

and Li, 2014). At the same time, the staggered geometry

with FAZ coverage and ultra-long offsets gives large

variations in take-off angles, which creates challenges for

deghosting.

Wang et al. (2013) proposed a pre-migration deghosting

method using a bootstrap approach in the Tau-P domain, a

pseudo-3D method that determines slowness in the x-

direction through a 2D sparse Tau-P inversion and

determines slowness in the y-direction using a bootstrap

least squares inversion. This method is effective for most

2D and 3D data in NAZ or WAZ acquisition geometry.

However, in the staggered acquisition design, for the data

from the side-guns (i.e., large azimuth and take-off angle of

ghosts), the wavefield is strongly 3D and thus generates

challenges for deghosting using this bootstrap method.

Recently, Wang et al. (personal communication, 2014)

proposed a 3D deghosting method for pressure-only data.

This method is fully data-driven and uses a progressive

sparse Tau-P inversion to perform 3D joint deghosting and

crossline interpolation in one step. We applied this 3D

deghosting to FAZ data and observed the benefits of

removing the ghost where the bootstrap method suffers.

Better ghost removal subsequently improves seismic

images and provides better multiple suppression from 3D

SRME.

Study area

The study area is located in Keathley Canyon, in the central

GOM, which is to the interior of the Sigsbee Escarpment

and features complex salt structures. The data are acquired

using multiple vessels in a staggered configuration (Figure

1a) with variable-depth streamers towed from 10 m to 50

m. The rose diagram in Figure 1b shows that this

acquisition configuration provides full azimuthal coverage

up to 9 km and ultra-long offsets up to 18 km.

3D deghosting for side-gun data

Large azimuths and long offsets provide challenges for

deghosting. Figure 2a shows a shot gather example from a

1

3

2

5

4

~18000m

~6000m

Red circles = 10km, 18 kma b

Figure 1: (a) Staggered acquisition layout. (b) Rose diagram

considering reciprocity. The inner red circle is the 10-km

offset, and the outer red circle denotes the 18-km offset.

Page 4238SEG Denver 2014 Annual MeetingDOI http://dx.doi.org/10.1190/segam2014-1297.1© 2014 SEG

Dow

nloa

ded

09/0

8/14

to 8

0.19

4.19

4.19

0. R

edis

trib

utio

n su

bjec

t to

SEG

lice

nse

or c

opyr

ight

; see

Ter

ms

of U

se a

t http

://lib

rary

.seg

.org

/

1

3

2

5

4

a

b

c

d

e

3s

6s

20 400Frequency (Hz)

DB

0-1

0

e

Figure 2: (a) Shot gather from red source and cable. Zoomed-in water bottom with (b) input, (c) bootstrap receiver deghosting, and (d) 3D

receiver deghosting. The same bootstrap source deghosting is applied on (c) and (d) before receiver deghosting. (e) Spectrum comparison with

input (red), bootstrap deghosting (green), and 3D deghosting (blue).

side-gun (source 3 and leading receiver marked in red)

before deghosting. This side-gun shot gather has a 3D

geometry and thus poses difficulties for the bootstrap

deghosting method that is based on a 2D Tau-P transform.

Figure 2b shows a zoomed-in section of the water bottom at

the apex of the shot gather from gun 3. The primary and

ghost energy are separated. Figure 2c shows the deghosting

result using the bootstrap method. Although this method

attenuated most of the ghost energy, obvious ghost

residuals remain. The 3D deghosting method, however, is

able to better attenuate the ghost energy and has no obvious

ghost residuals (Figure 2d).

Figure 2e shows the spectrum comparison for the shot

gathers before deghosting, after bootstrap deghosting, and

after 3D deghosting. The 3D deghosting method fills in the

ghost notches much better than the bootstrap deghosting

method.

Benefit of 3D deghosting for migrated images

To show the benefit of 3D deghosting, we ran Kirchhoff

sediment flood migrations using the side-gun data before

deghosting, after bootstrap deghosting, and after 3D

deghosting. Figures 3a and 3d show a depth slice and a

cross-section using the data before deghosting. The primary

and ghost energy both present in the migrated image. The

TOS is not clearly defined in the depth slice due to the

interference of primary and ghost energy. Figures 3b and 3e

show the same depth slice and cross-section after bootstrap

deghosting. In the cross-section, most of the ghost energy

at flat TOS areas is attenuated, and the depth slice shows a

better defined TOS event. However, obvious residual ghost

energy remains, especially at the dipping TOS region. The

migrated images with the data after 3D deghosting do not

have the residual ghost energy, and the TOS event is more

coherent in both the depth slice and cross-section (Figures

3c and 3f).


Dow

nloa

ded

09/0

8/14

to 8

0.19

4.19

4.19

0. R

edis

trib

utio

n su

bjec

t to

SEG

lice

nse

or c

opyr

ight

; see

Ter

ms

of U

se a

t http

://lib

rary

.seg

.org

/

3D deghosting for staggered ultra-long offsets and full azimuths

We also compared Kirchhoff gathers in the same locations

for input data before deghosting (Figure 3g), with bootstrap

deghosting, and with 3D deghosting. Although bootstrap

deghosting works well for near to middle offsets, there

were residual ghosts in far offsets (Figure 3h). The 3D

deghosting method removed the ghosts nicely from the far

offsets, and both the sediment layer and TOS events were

more coherent (Figure 3i).

Benefit of 3D deghost for multiple suppression

To demonstrate the deghosting effect on multiple

suppression, we used both bootstrap and 3D deghosted data

as input for 3D SRME. Ideally, SRME prediction requires

ghost-free data, and the 3D deghosting method provides

better deghosting in both the basin area and the dipping

TOS area (Figures 4a and 4c). Residual ghost energy from

the bootstrap method degraded the quality of the multiple

model prediction and thus resulted in residual multiples in

the SRME output (Figure 4b). The 3D deghosting method

resulted in better attenuation of multiples (Figure 4d). This

demonstrates that improved deghosting from the 3D

technique can more accurately predict multiple models and

subsequently produce better multiple suppression results.

Conclusions

Removing ghost energy provided a broader spectrum and

improved S/N in marine streamer data. Using FAZ and

ultra-long offset, variable-depth streamer data from the

GOM, we compared the deghosting results using bootstrap

a b c

d e f

1km

2 km

3 km

g h i

2 km

3 km

5km

Figure 3: Kirchhoff migration stack depth slice at 2200 m with input data (a) before deghosting, (b) after bootstrap deghosting, and (c) after 3D

deghosting. Kirchhoff migration stack cross-section with input data (d) before deghosting, (e) after bootstrap deghosting, and after 3D

deghosting(f); Kirchhoff migration gathers with input data (g) before deghosting, (h) after bootstrap deghosting, and (i) after 3D deghosting.


Dow

nloa

ded

09/0

8/14

to 8

0.19

4.19

4.19

0. R

edis

trib

utio

n su

bjec

t to

SEG

lice

nse

or c

opyr

ight

; see

Ter

ms

of U

se a

t http

://lib

rary

.seg

.org

/

3D deghosting for staggered ultra-long offsets and full azimuths

and 3D deghosting methods for side-gun data. We showed

that the 3D deghosting method removed the ghost energy

more effectively than the bootstrap deghosting method. The

improved 3D deghosting result provided more accurate

TOS definition from the Kirchhoff sediment flood

migration and more coherent energy from mid to far offsets

in the migrated gathers. We also showed that with better

deghosting, SRME output has fewer residual multiples.

Acknowledgments

We thank Tony Huang for fruitful discussions and CGG for

permission to show this work. We also thank Xiao Huang

and John Aven for their work on the field data example.

a b

c d

3s

6s

Figure 4: Surface-related multiple elimination (SRME) (a) input and (b) output with bootstrap deghosted data. SRME (c) input and (d) output with 3D deghosted data.

241SEG Denver 2014 Annual MeetingDOI http://dx.doi.org/10.1190/segam2014-1297.1© 2014 SEG

Dow

nloa

ded

09/0

8/14

to 8

0.19

4.19

4.19

0. R

edis

trib

utio

n su

bjec

t to

SEG

lice

nse

or c

opyr

ight

; see

Ter

ms

of U

se a

t http

://lib

rary

.seg

.org

/

http://dx.doi.org/10.1190/segam2014-1297.1 EDITED REFERENCES Note: This reference list is a copy-edited version of the reference list submitted by the author. Reference lists for the 2014 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

Mandroux, F., B. Ong, C. Ting, S. Mothi, T. Huang, and Y. Li, 2013, Broadband, long-offset, full-azimuth, staggered marine acquisition in the Gulf of Mexico: First Break, 31, 125–132.

Mothi, S., K. Schwarz, and H. Zhu, 2013, Impact of full-azimuth and long-offset acquisition on full-waveform inversion in deep water Gulf of Mexico: Presented at the 83rd Annual International Meeting, SEG.

Poole, G., 2013, Premigration receiver deghosting and redatuming for variable depth streamer data: Presented at the 83rd Annual International Meeting, SEG.

Soubaras, R., 2010, Deghosting by joint deconvolution of a migration and a mirror migration: 80th Annual International Meeting, SEG, Expanded Abstracts, 3406–3410.

Wang, P., and C. Peng, 2012, Premigration deghosting for marine towed streamer data using a bootstrap approach: 82nd Annual International Meeting, SEG, Expanded Abstracts, doi: 10.1190/segam2012-1146.1.

Wang, P., S. Ray, C. Peng, Y. Li, and G. Poole, 2013, Premigration deghosting for marine streamer data using a bootstrap approach in tau-p domain: 83rd Annual International Meeting, SEG, Expanded Abstracts, doi: 10.1190/segam2013-0225.1.

Wu, Q., and Y. Li, 2014, Benefit of ultra-long offset data for subsalt imaging in deep water Gulf of Mexico: Presented at the 76th Annual International Conference and Exhibition, EAGE.

Wu, Q., Y. Li, Z. Li, and W. Han, 2013, Tilted orthorhombic imaging for full-azimuth towed streamer data in deep water Gulf of Mexico: 75th Annual International Conference and Exhibition, EAGE.

Yu, B., C. Ting, Y. Li, and Z. Gan, 2013, 3D SRME for marine broadband full-azimuth acquisition: 83rd Annual International Meeting, SEG, Expanded Abstracts, doi: 10.1190/segam2013-1043.1.


Dow

nloa

ded

09/0

8/14

to 8

0.19

4.19

4.19

0. R

edis

trib

utio

n su

bjec

t to

SEG

lice

nse

or c

opyr

ight

; see

Ter

ms

of U

se a

t http

://lib

rary

.seg

.org

/

3d deghosting for full-azimuth and ultra-long offset ... · 3d deghosting for full-azimuth and...

Documents