Search

Parameterization lies at the center of anisotropic full-waveform inversion (FWI) with multiparameter updates. This is because FWI aims to update the long and short wavelengths of the perturbations. Recently, there has been an intensive effort to determine the optimal parameterization, centering the fundamental discussion mainly on the analysis of radiation patterns for each one of these parameterizations, and aiming to determine which is best suited for multiparameter inversion. We have developed a new parameterization in the scope of FWI, based on the concept of kinematically equivalent media, as originally proposed in other areas of seismic data analysis. The radiation pattern reveals that this parameterization combines some of the characteristics of parameterizations with one velocity and two Thomsen's parameters and parameterizations using two velocities and one Thomsen's parameter. The study of perturbation of traveltime with perturbation of model parameters shows that the new parameterization is less ambiguous when relating these quantities in comparison with other more commonly used parameterizations.

The aim of this study is to discuss the challenges for interbed multiple attenuation in the Asia-Pacific region. Many possible multiple generators such as seafloor, carbonate layer or volcanic flow are found in the data. Without any prior subsurface knowledge, the ISS method is used in our study to estimate all the interbed multiples . In order to obtain a high resolution multiple model, some data preparation can be required such as inverting near offset data for shallow water or broadband processing for better separation of thin layers. If the survey has strong crossline dip, 3D prediction is necessary to take in account the out-of-plane components. Through case studies from offshore Australia and offshore Vietnam, we demonstrate strategies for effective interbed multiple attenuation using the ISS method.

By adapting reverse time migration (RTM) and demigration as the migration and modeling operators to maximize the crosscorrelation between the simulated and the acquired seismic data, we introduced a new practical least-squares RTM (LSRTM) scheme and derived a steepest descent method in seeking the optimal image. Through synthetic and real data experiments, we determined that the proposed LSRTM provided high-quality images with balanced amplitudes, improved focusing, and enhanced resolution. The method was also capable of removing free surface ghosts caused by towed streamer acquisition, filling the structures and reducing crosstalk noise associated with simultaneous shooting.

Pre-stack elastic inversion requires a reliable wavelet model. This wavelet model must take into account the angle dependency of the wavelet, as well as the time dependency. This is even more important for broadband data, as the wavelets exhibit more variations, due to the fact that the maximum frequency is very high for small time and angles, but much smaller for large time and angle. We show how a continuously varying wavelet model can be estimated through a Bayesian inversion. This wavelet model can be used to pre-process the gathers and provide a zero-phase wavelet model for pre-stack elatic inversion. Synthetic and real data examples are shown to support this.

We investigated how current least-squares reverse time migration (LSRTM) methods perform on subsalt images. First, we compared the formulation of data-domain vs. image-domain least-squares migration (LSM), as well as methods using single-iteration approximation vs. iterative inversion. Next, we examined the resulting subsalt images of several LSRTM methods applied on both synthetic and field data. Among our tests, we found image-domain single-iteration LSRTM methods, including an extension from Guitton’s (2004) method in the curvelet domain, not only compensated for amplitude loss due to poor illumination caused by complex salt bodies, but also produced subsalt images with fewer migration artifacts (i.e., noise) in the field data. By contrast, an iterative inversion method showed its potential for broadening bandwidth in the subsalt, but was less effective in reducing noise. Based on our understanding, we will summarize the current state of LSRTM for subsalt imaging, especially between single-iteration and iterative LSRTM methods.

The importance of inverting jointly velocity and density parameters in full waveform inversion (FWI) is well established. In a former work we had proposed an innovative preserved amplitude FWI allowing improving the convergence rate of FWI. It was derived from preserved amplitude reverse time migration (RTM), involved a deconvolution imaging condition and was limited to the estimation of velocity perturbation from reflection data. We extend here the approach to a joint velocity and density preserved amplitude FWI. We present the theoretical derivation of the improved common shot FWI gradients and show how we can decouple the two parameters. We validate our approach on the Marmousi II synthetic model which shows that we can efficiently reconstruct the two parameters, and on a real data showing that we significantly reduce the data residual with the joint inversion.

Spectral broadening of migrated and stacked seismic images is a common method to enhance interpretability of reflection data. In this paper we propose a prestack AVA compliant spectral broadening approach based on non-stationary wavelet deconvolution. The algorithm employs AVA coupling in the prestack domain to shape the spectra of all traces in angle gathers simultaneously. Using synthetic and real data we show that the characteristics of all AVA classes are preserved and that the spectra of all angles are enhanced and better balanced.

Posing migration as an inverse or a least-squares problem can improve the quality of imaging. This class of techniques can resolve illumination issues and improve focusing. Standard iterative least-squares imaging can be expensive and results are often compromised. We present a procedure using matching filters operating in data-space rather than image space. Effective inversion results are demonstrated on synthetic and real data.

Marine seismic acquisitions record both primary and multiple wavefields. In a typical processing sequence, multiple energy is removed from the data before migration. However, there may be valuable information contained in the multiple wavefield. To discover this hidden information, reverse time migration of multiples (RTMM) was proposed. We evaluated the advantages of RTMM through three different real data processing projects and identified three key advantages. Additionally, we present a synthetic study of two types of crosstalk noise that hinder the full potential of RTMM as well as propose corresponding practical strategies to handle them.