Search

Multiple contamination in seismic data from the south of the Sultanate of Oman is challenging due to the combined presence of short-period surface and internal multiples generated in the upper section. We present an innovative workflow to attenuate multiples in the pre-migration pre-stack domain based on a dataset from the south of the Sultanate of Oman. The workflow involved least-squares multiple imaging to obtain a reliable reflectivity of multiple generators in the near surface.

Multi-dimensional deconvolution (MDD) is an attractive method to remove free-surface multiples for OBN datasets. Multi-dimensional implementations are able to produce accurate multiple models in areas of structurally complex reflectivity. For most OBN acquisitions, the receiver density is not sufficient for receiver-side MDD. Consequently, source-side methods have been introduced to take advantage of the often denser source sampling.

Acoustic full-waveform inversion (FWI) is widely employed to produce high-resolution subsurface models, especially for the P-wave velocity. However, more complex geological challenges, such as shallow chalk and salt, generate strong impedance contrasts that exhibit elastic effects, causing acoustic approximations to struggle. In parts of the Southern North Sea, fast shallow chalk outcrops at the water bottom and Zechstein salt geometries sit above target intervals, both severely challenging velocity model building and imaging. This shallow-water case study uses short-offset, dual-azimuth, towed-streamer data where the near-surface geological environment and maximum acquired offset limit the diving-wave penetration.

This paper is an attempt to fill the technology gap existing between pure P- and PS-wave imaging. Full
wavefield extrapolation techniques are well developed for P-wave and RTM has been now available for
almost a decade. Conversely, ray based migration algorithms are still the workhorse for converted-wave
(PS-wave) depth imaging. Here, we introduce a new converted-wave anisotropic RTM, using a low-rank
decomposition of mixed-domain space-wavenumber propagators for quasi-P and quasi-S waves. These
operators are formal integral solutions of the pure-mode wave equations which guarantee stable and
dispersion-free time extrapolation for coarse time steps in anisotropic, heterogeneous media. The puremode extrapolators are attractive for both PS-wave structural imaging and velocity analysis. An ocean
bottom cable synthetic example illustrates the effectiveness of low-rank PS-wave RTM when compared
against state-of-the-art Gaussian beam and finite difference RTM algorithms.

A good estimation of velocity anisotropy is important in a pre-stack depth migration project in order to obtain proper focusing and accurate depthing of the seismic image. Considering that the velocity along the bedding is expected to be faster than across the bedding, the anisotropy parameter epsilon ( Thomsen, 1986 ) is typically positive. We present an example of shallow water imaging from offshore Suriname, in which using negative epsilon was found necessary in order to achieve both seismic well tie and flat pre-stack depth migration image gathers. From well information, the negative anisotropy correlates well with the mostly unconsolidated sand layers.

We demonstrate how the use of multiples in imaging may provide improved shallow illumination and potentially reduce the requirement for extensive site-survey acquisition in some areas.

Recent discoveries in Namibia’s Orange Basin have raised industry expectations that a similar level of success could be replicated in its conjugate margin, Uruguay. Still considered as a frontier basin, the Uruguay deepwater has similar geological elements at basin scale and unique aspects at prospect scale that could help to achieve exploration success at Cretaceous stratigraphic levels.

Least squares migration (LSM), like interpolation, has the potential to address sampling issues and generate images with better amplitudes than migration. Although both techniques share the same goal and often the same formulation, they differ on the nature of the model that is used to predict the data. For LSM the cost of the migration/modelling operator bring limitations. In this abstract I discuss some of these limitations, possible ways to overcome them, and analyse some similarities and differences with respect to interpolation. Also I present results using a LSM implementation for Kirchhoff Depth Migration for both PP and PS data.

With a new name and its guiding vision to “see things differently”, Viridien (formerly CGG) focuses on human ingenuity to usher in a bright future across multiple sectors.