Azimuthal anisotropy? The time and depth imaging points of view: an imaging case history.

Transcription

1 Azimuthal anisotropy? The time and depth imaging points of view: an imaging case history. Serge Zimine 1, Gilles Lambaré* 1, Patrice Guillaume 1, Jean-Philippe Montel 1, Jean-Paul Touré 1, Nicolas Deladerrière 1, Xiaoming Zhang 1, Anthony Prescott 1, Didier Lecerf 1, Sylvain Navion 1, Jean-Luc Boelle 2, Ahmed Belmokhtar 3 & Abdeljebbar Ladmek 3 1 CGGVeritas Services SA; 2 Total; 3 Sonatrach Summary The superior imaging capabilities of wide azimuth (WAZ) are now well established. Processing such acquisitions requires adaptation of the processing workflows and tools. For example the recorded azimuthal information must be kept and the tools should be able to deal with the increased amount of data. Concerning the velocity model building, a question remains on the need of introducing azimuthal anisotropy. In this paper we address this point with a time and depth imaging case study for a high density wide azimuth (WAZ) land surface acquisition. We show on this dataset that azimuthally varying residual move-out observed on time migrated common image gathers (CIG) definitely disappears on depth migrated CIGs (in both cases no azimuthal anisotropy is introduced in the velocity model). This illustration highlights the limits of the assumptions of time imaging, thus promoting the use of depth imaging when processing high-density WAZ data, even in the context of mild geological complexity. Introduction The seismic industry is currently pushing towards the development of WAZ acquisitions. The processing of such datasets requires some adaptations of existing tools and workflows. For example the concept of offset vector tiles (allowing design of common offset vector (COV) cubes) has been introduced for processing WAZ acquisitions (Vermeer, 2002). Compared to narrow azimuth (NAZ) acquisitions the processing of WAZ data requires combination of information from the various azimuths associated to the various COV volumes. During this process the velocity model building step is critical for reconciling the kinematics observed on the various azimuths. In the frame of migration velocity analysis this firstly imposes to develop tools for picking azimuthal residual moveout (RMO) (Lecerf et al., 2009; Siliqi and Talaalout, 2009). The second need is to have adequate tools allowing tomographic update of the velocity suitable to handle WAZ acquisition (Montel et al., 2010; Zimine et al., 2010). Using such tools it becomes possible to investigate WAZ velocity model building addressing the problem of azimuthal anisotropy. In this paper we present a dense WAZ acquisition imaging case study. Our processing workflow involves non-linear slope tomography for both time and depth imaging (Guillaume et al., 2008; Lambaré et al., 2009). An azimuthal residual move-out picking is performed on initial Pres-Stack Time Migration (PreSTM) gathers. These gathers are characterized by a significant wobbling effect when sorted as snail gathers (Lecerf et al., 2009) (Figure 1). Should we interpret this wobbling effect as azimuthal anisotropy? Will it disappear after improving time velocity model or processing through depth imaging? These are the questions we address in this paper.

2 Figure 1 Gathering of COV PreSTM traces in a snail gather (Lecerf et al., 2009). Traces are ordered in offset range classes with increasing offset and within each offset range class according to increasing azimuth. On the snail gather on the right the RMO is characterized by a significant wobbling effect that can be interpreted as a parabolic elliptic effect. A high density WAZ land data For our investigations a subset was selected from a high density WAZ land orthogonal surface acquisition, over a faulty elongated anticline. All the figures presented in this paper are extracted from a central 7.5 km line, with a dip direction with respect to the main structure. This dataset was processed through offset vector binning to deliver 225 single-fold COVs with homogeneous offset and azimuth within each COV. These pseudo minimum datasets (Vermeer, 2002) represent cubes with minimum holes or over fold. The offset X and offset Y increment (400m) and range (-3000m to +3000m) used by the tiling process are directly related to the acquisition scheme (200m between shot lines and between receiver lines). Offset and azimuth information is preserved throughout the processing sequence. We start from an initial PreSTM done for the 225 COVs independently and ordered as snail gathers (Figure 1). This kind of data sorting is suitable for WAZ parabolic elliptical RMO picking. This approximation is reasonable as incidence angles are limited. Furthermore, Lecerf et al. (2009) have shown that this approximation is adequate to compensate for azimuthal velocity variations as the algorithm takes into account all azimuths simultaneously and is less sensitive to the noise than independent azimuthal sectoring analysis (Figure 2).

3 Figure 2: parabolic elliptical RMO corrections. Top left) wobbling effect observed on a snail PreSTM gathers. Bottom left) the same gather after parabolic elliptical RMO corrections demonstrating the relevance of the parabolic elliptical assumption. On the right the corresponding stacks where, as expected, the application of RMO corrections greatly improves focusing and clarify the fault delineation on the right flank of the anticline. The point is now to investigate if the azimuthally varying RMO can be interpreted in terms of azimuthally varying velocities. This can be addressed using a non-linear slope tomographic update of the velocity, with the advantage of using the same picking for time and depth velocity model building. Non-linear slope tomography for time and depth imaging Non-linear slope tomography (Guillaume et al., 2008) is an accurate tool for velocity model building that minimizes the local slope of migrated events in the common image point (CIP) gathers. Figure 3 Kinematic invariants and their relationships with depth and time tomography processes. The use of locally coherent events makes it particularly adapted to dense volumetric picking. Resolution of non-linear aspects provides the best possible update of the velocity model from a given picking. Here we use here the extension of the method to WAZ data (Montel et al., 2010) both for depth and time velocity update (Lambaré et al., 2009). It is coupled with a dense WAZ RMO picking (Lecerf et al., 2009). Note that picking of locally coherent events can be done indifferently in the depth or time-migrated domains (Lambaré et al., 2007) or even in the un-migrated domain (Figure 3). When picked in the migrated domain, a kinematic de-migration (Guillaume et al., 2001; Chauris et al., 2002) performs the conversion of the local event to the un-migrated domain, providing the so-called kinematic invariants. Application to the land WAZ dataset To update time and depth velocity models through slope tomography, we need to compute the kinematic invariants in the un-migrated domain via a de-migration process. Here we use the RMO

4 information picked on PreSTM gathers and dips, and skeletons picked on a near-offset stack using an automatic structural mapping of coherent horizons (Siliqi et al., 2009). Effective parameters V eff and η eff will be updated in the time-migrated domain (Lambaré et al., 2009) whereas interval parameter V int will be updated in the depth-migrated domain from the same set of kinematic invariants (Figure 3). Figure 4: Time versus depth imaging after time and depth WAZ non-linear slope tomography. Top) The final PreSTM results where we observe on the left a remaining wobbling effect event if it has been significantly reduced compared to initial PreSTM (Figure 2). Bottom) the final PreSDM results where the wobbling effect has disappeared and where the faults are even sharper than on the PreSTM with RMO corrections from Figure 2. Figure 4 shows that the azimuthal kinematical variations (wobbling effects) observed on the initial PreSTM snail gathers is reduced after time tomography and is no longer visible after depth tomography, as depth imaging honours actual traveltimes. Depth velocity model building better solved azimuthal variations in the context of land WAZ acquisition, resulting in enhanced focusing as illustrated by flatter CIP gathers.another interesting observation is to consider the snail gather obtained with the initial PreSDM model (the one used as initial model for the non-linear slope tomography in depth). We observe on Figure 5 that the PreSDM snail gather do not show any wobbling effect even if significant RMO remains. The wobbling effect is obviously inherently associated to time imaging on the right flank of the anticline structure.conclusion We show with this application of WAZ non-linear slope tomography that interpretation of azimuthally varying RMO observed on PreSTM snail gathers can be tricky. It could be interpreted as effective azimuthal anisotropy as it remains after an accurate time tomographic update of the velocity. But since it disappeared as soon as we moved to depth imaging it has to be considered as a pure bias of time imaging observed on WAZ data. This bias seems to be essentially associated to structural dips rather than heterogeneity, layering, fractures, or intrinsic anisotropy sometimes advocated to explain effective azimuthal anisotropy (note that no interval anisotropy has been introduced for depth imaging). The resulting improvement in focusing is particularly helpful for fault delineation. As a conclusion we suggest that depth rather than time processing should be applied to high-density wideazimuth data, even in the context of mild geological complexity. Acknowledgements We are particularly grateful to Sonatrach, Cepsa, and Total for their permission to present these field data examples and publish results. We would also like to thank the CGGVeritas R&D Imaging team in Massy for their support and contribution to the WAZ tomography implementation.

5 Figure 5: Initial versus final PreSDM results after depth WAZ non-linear slope tomography. Top) The initial PreSDM results. Bottom) the final PreSDM results. We see that the tomographic update greatly enhance focusing and consequently delineation of the faults but also that for both initial and final PreSDM absolutely no significant wobbling effect can be observed. References Chauris, H., M. Noble, G. Lambaré, and P. Podvin, Migration velocity analysis from locally coherent events in 2-D laterally heterogeneous media, Part I : Theoretical aspects. Geophysics, 67(4): Guillaume, P., F. Audebert, P. Berthet,, B. David, A. Herrenschmidt, and X. Zhang, D Finite-offset tomographic inversion of CRP-scan data, with or without anisotropy, 71st annual SEG meeting, Expanded Abstracts, , INV2.2. Guillaume, P., G. Lambaré, O. Leblanc, P. Mitouard, J. Le Moigne, J.-P. Montel, T. Prescott, R. Siliqi, N. Vidal, X. Zhang and S. Zimine Kinematic invariants: an efficient and flexible approach for velocity model building, 78th annual SEG meeting, Las Vegas 2008, SEG workshop Advanced velocity model building techniques for depth imaging. Lambaré, G., P. Herrmann, P. Guillaume, S. Zimine, S. Wolfarth, O. Hermant, S. Butt, From time to depth imaging with Beyond Dix, First Break, 25 (9). Lambaré G., N. Deladerrière, Y. Traonmilin, J.P. Touré, J. Le Moigne, P. Herrmann, Non-linear Tomography for Time Imaging, Extended Abstracts, 71st EAGE Conference & Exhibition - Workshops and Fieldtrips. U029. Lecerf, D., S. Navion, J.L. Boelle, A. Belmokhtar, A. Ladmek, Azimuthal Residual Velocity Analysis in Offset Vector for WAZ Imaging, EAGE extended abstracts, V013. Montel J.P. P. Guillaume, G. Lambaré and O. Leblanc, Non-linear slope tomography: extension to MAZ and WAZ configurations, 72nd EAGE Barcelona. Siliqi R., Talaalout A., Structurally coherent Wide Azimuth Residual Move Out surfaces. 79 th Annual International Meeting, SEG, SEG, Expanded Abstracts, Vermeer, G., J., O., 2002, Processing with offset-vector-tile gathers: 63rd EAGE Amsterdam. Zimine, S., G. Lambaré, P. Guillaume, J.-P. Montel, J.-P. Touré, N. Deladerrière, X. Zhang, A. Prescott, D. Lecerf, S. Navion, J.-L. B., A. Belmokhtar and A. Ladmek, 2010.Can we correct for azimuthal variations of residual move-out in land WAZ context, using depth non-linear slope tomography? An imaging case history. 72nd EAGE Barcelona.