Recovering Nile Delta Messinian salt structure using Reconstructed Full Waveform Inversion

Juergen Fruehn, Stuart Greenwood; Ken McDermott, Sahil Mannick, Chao Wang, ION.
Downloaded 09/11/17 to 80.82.77.83. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

Summary increase lateral resolution in the imaging workflow (e.g.

The Messinian evaporitic sequence in the Mediterranean
Sea forms a complex screen which distorts and scatters
seismic energy travelling through it. Consequently,
attempts to build reliable velocity models and associated
depth images are greatly compromised. In this work, we
present part of a very large scale regional basement study,
using long-offset long 2D seismic lines, where we have
employed high resolution ray-tomography and
conventional full waveform inversion, in combination with
a new technique: reconstructed full waveform inversion
which better constrains the resultant models by using the
velocity model as an image constraint.
Introduction
As with much of the Mediterranean Sea, attempts to image
seismic data from the Nile Delta is beset with image
distortion due to the presence of an evaporate layer in the
shallow overburden (hereafter referred to generically as the
salt). The thickness of the Messinian evaporates varies
greatly, and in places is absent. Although model building is
often attempted by picking the top-salt features in great
detail, the complex inter-bedded structure and variable
thickness of the evaporate bodies, still often hampers both
velocity estimation and subsequent depth migration of pre-
salt structures (e.g. Jones & Davison, 2014).
Pratt et al., 1996; Plessix et al., 2010; Brittan et al., 2016).
However, due to the acoustic approximation usually made
with FWI, the data which can be reliably inverted against
are usually limited to the refracted (transmitted) wavefield,
as this is less sensitive to density change. Hence, classical
FWI only tends to give a meaningful velocity update in the
shallower parts of the section.
More recently, variants of FWI have been introduced that
use the image domain, or other criteria not simply restricted
to amplitude and phase behavior, as a constraint (e.g.
Warner and Gausch, 2016, Wang et al., 2016, Vigh et al.
2016). These techniques offer the possibility of achieving
deeper reliable model updates than conventional FWI.
Figure 1 shows a synthetic salt body example comparing
the results of FWI and RFWI, starting from a smoothed
version of the true model (Wang et al. 2017). As expected,
FWI is well able to update the shallow post salt section,
but fails to meaningfully delineate the salt body or the pre-
salt section. Conversely, RFWI successfully recovers the
salt geobody geometry and provides a reasonable update to
the sub-salt model.

The data used in this study are from a long offset deep
basin-study dataset of the type described by Horn (2015).
In this case the overall program includes several thousand
linear kilometers of 2D data, and the line shown here is
some 360km long. The maximum offset recorded in the
marine cable was 12km, with a record length of 18s.
Imaging was performed to a depth of 40km, so as to try to
capture Moho reflectors and to determine the boundary
between continental and oceanic crust.

Data processing included various forms of multiple

suppression as well as broadband deghosting, the latter
being of particular importance in recovering low-frequency
components of the data in advance of waveform inversion.

Velocity model building commenced with several iterations
of high resolution non-parametric inversion (Fruehn et al.,
2014), in conjunction with interpretation of the top and
base evaporite (salt) geobodies. It is well understood that
the resolution available with such ray-based methods is
limited to lateral scale lengths of several hundred metres
(e.g. Jones 2010, 2015) except for well-behaved clean long
offset data. Consequently, in recent years we have seen the
introduction of Full Waveform Inversion (FWI) in order to
Figure1. clockwise from top left: true model; smoothed
starting model; result of conventional FWI (the shallow
model is successfully recovered, but the salt body is not);
result of RFWI the shape of the salt body is reasonable
recovered, and the sub-salt model is beginning to emerge
(example courtesy of Chao Wang et al., 2017).

2017 SEG Page 1497

SEG International Exposition and 87th Annual Meeting
 Recovering Nile Delta Messinian salt structure using RFWI
Downloaded 09/11/17 to 80.82.77.83. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

Figure 2: image with model overlay after high resolution Figure 3: image with model overlay after tomo plus
ray tomography, and preliminary picking of the salt conventional FWI: the near surface and post-salt sequence
bodies, is nicely updated, but salt and below remain unchanged

2017 SEG Page 1498

SEG International Exposition and 87th Annual Meeting
 Recovering Nile Delta Messinian salt structure using RFWI
Downloaded 09/11/17 to 80.82.77.83. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

Figure 4: image with model overlay after tomo plus Figure 5: image with model overlay after further iterations
conventional FW, plus RFWI and residual tomo: salt and of RFWI and a final residual tomo: salt and below are now
below are now updated better defined

2017 SEG Page 1499

SEG International Exposition and 87th Annual Meeting
 Recovering Nile Delta Messinian salt structure using RFWI

Results on Nile Delta field data

Figure 2 shows a small segment of the 2D regional line,
Downloaded 09/11/17 to 80.82.77.83. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

with an interval velocity model overlay resulting from

several iterations of structurally constrained ray-based non-
parametric tomography, in conjunction with preliminary
picking of both the top and base salt. After fifty (internal)
iterations of FWI, we obtain the image in Figure3: here we
see as expected for FWI, that the near-seabed structure has
been updated, but that the salt and pre-salt section remains
essentially unchanged.

Following this, the RFWI was run for 5 iterations, which

modified the salt boundary and velocity within the salt
geobody. Subsequent tomography for the deep section
refined the pre-salt model (Figure 4). Finally, a further 18
iterations of RFWI improved the salt geobody and the
intra-salt velocity distribution, and more ray-tomography
successfully updated the very deep section (Figure 5).

Figure 6 shows a well log superimposed on the interval

velocity models for the starting model (which was the final
product of the previous 2016 regional isotropic model
building) and then on the final anisotropic new model. In
the new results, the fine-detail of the salt geobodies is
evident, as is the near-seabed detail.

Conclusions
Although FWI has offered great promise as a mean of
updating near-surface structure, pushing past the limits of
ray-based tomography, the acoustic approximation, when
applied in the data-domain is still essentially limited to the
transmitted wavefield. Extending FWI to include image-
domain constraints can greatly expand the range of
usefulness of the wave-based inversion techniques, also
helping to mitigate the effects of cycle skipping on Figure 6: Top interval velocity from legacy 2016
convergence. conventional isotropic tomographic model update.
Bottom new anisotropic tomography + FWI + RFWI
model. Both shown with the associated well log.
Acknowledgements
We thank BP staff for their helpful insights into the
expected geological structure of the region, and to ION for
permission to present this work. We also thank Ian Jones
for his help in preparing the manuscript, and guidance in
the model building.

2017 SEG Page 1500

SEG International Exposition and 87th Annual Meeting
 EDITED REFERENCES
Note: This reference list is a copyedited version of the reference list submitted by the author. Reference lists for the 2017
SEG Technical Program Expanded Abstracts have been copyedited 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
Downloaded 09/11/17 to 80.82.77.83. Redistribution subject to SEG license or copyright; see Terms of Use at http://library.seg.org/

Brittan, J., J. Bai, H. Delome, C. Wang, and D. Yingst, 2013, Full waveform inversionThe state of the
art: First Break, 31, 7581.
Fruehn, J., S. Greenwood, V. Valler, and D. Sekulic, 2014, Resolving small-scale near-seabed velocity
anomalies using non-parametric autopicking and hybrid tomography: CSEG Recorder, 39.
Jones, I. F., 2010, An introduction to velocity model building: EAGE.
Jones, I. F., 2015, Estimating subsurface parameter fields for seismic migration: Velocity model building:
Encyclopedia of Exploration Geophysics, U1-1U1-24.
Jones, I. F., and I. Davison, 2014, Seismic imaging in and around salt bodies: Interpretation, 2, SL1
SL20, http://doi.org/10.1190/INT-2014-0033.1.
Plessix, R.-E., G. Baeten, J. W. de Maag, M. Klaassen, R. Zhang, and Z. Tao, 2010, Application of
acoustic full waveform inversion to a low-frequency large-offset land data set: 80th Annual
International Meeting, SEG, Expanded Abstracts, 930934, http://doi.org/10.1190/1.3513930.
Pratt, R. G., Z.-M. Song, P. Williamson, and M. Warner, 1996, Two-dimensional velocity models from
wide-angle seismic data by wavefield inversion: Geophysical Journal International, 124, 323
340, http://doi.org/10.1111/j.1365-246X.1996.tb07023.x.
Vigh, D., K. Jiao, X. Cheng, D. Sun, and W. Lewis, 2016, Earth-model building from shallow to deep
with full-waveform inversion: The Leading Edge, 35, 10251030.
Wang, C., D. Yingst, P. Farmer, and J. Leveille, 2016, Full-waveform inversion with the reconstructed
wavefield method: 86th Annual International Meeting, SEG, Expanded Abstracts, 12371241,
https://doi.org/10.1190/segam2016-13870317.1.
Wang, C., D. Yingst, P. Farmer, G. Martin, and J. Leveille, 2017, Reconstructed full waveform inversion
with the extended source: 87th Annual International Meeting, SEG, Expanded Abstracts.
Warner, M., and L. Guasch, 2016, Adaptive waveform inversion: Theory: Geophysics, 81, no. 6, R429
R445, https://doi.org/10.1190/geo2015-0387.1.

2017 SEG Page 1501

SEG International Exposition and 87th Annual Meeting