EAGE/SEG Summer Research Workshop - Processing and imaging of seismic data

Cross-disciplinary communication between specialized E&P projects should be facilitated by a conceptual framework that can be used at different levels of abstraction. In this framework an unbiased strategy need be followed, meaning that parameterization of the subsurface model must be delayed as much as possible. One consequence of delayed parameterization is that velocity models should not drive the seismic processing algorithms.

The basic idea of this paper is to derive approximate equations for vertical slowness for P and SV waves and PP, SS and PS traveltimes in TIV media not using a weak-anisotropy approximation.

Starting with the concept of the surface-to-surface propagator matrix, we explain how the elements of the matrix are related to kinematic wavefield attributes.

The Common-Reflection-Surface (CRS) stacking technique provides a powerful time domain imaging especially in case of complex geology data with low signal-to-noise ratio. The macro-model approach does on principle save the effort to produce a macro-model. However, the method allows as well the input of a velocity model as a guide function, which is automatically improved. A coarse initial velocity model is helpful especially in case of strong multiple wavefields.

Wavefield redatuming can be very efficient to overcome the overburden imprint on the target area. Due to an increasing amount of 3D data it becomes more and more important to develop a feasible method for the redatuming of 3D prestack data. We propose an approach that can be used in general for wave-equation datuming and present a strategy to perform CFP-redatuming of sparse 3D data using this approach. Two tests applying the modified redatuming approach to synthetic 2D data show satisfying results implying, that this method, indeed, can be used for the redatuming of sparse 3D data.

In land environment, the proper definition of a complex near-surface is probably one of the most challenging and controversial issues. Attempts to attack this problem range from conventional statics, to the more sophisticated tomography and wave equation-based redatuming algorithms.

The Common-Reflection-Surface (CRS) Stack can be used to extract kinematic information for the construction of velocity models from seismic prestack data.

By means of the Common Focal Point (CFP) technology [1, 2], seismic two way traveltime data can be converted to one way focusing operators. These focusing operators are generally used for model independent imaging, so that the image will not be constrained by any (user-defined) model. Nevertheless, when a velocity model of the subsurface is requested, it can be obtained from these focusing operators by tomographic inversion. This abstract presents a method for the inversion of focusing operators using data-driven model parameterization. It is demonstrated that this inversion results in an accurate unconstrained estimation of both velocities and focal point locations.

Prestack migration velocity analysis in the vertical time domain reduces the velocity-depth ambiguity usually hampering the performance of prestack depth-migration velocity analysis.

Recent advances in imaging and computing technology have made prestack wave-equation depth migration feasible in a production environment. However, the state-of-the-art in velocity model building is still based on mainly Kirchhoff prestack depth migration and raybased tomographic inversion. This mismatch between migration algorithm and velocity model will ultimately limit the final image quality. This paper discusses some of the issues around the mismatch of velocity model building and migration tools and other issues that hamper the derivation of an optimal migration velocity model.