[First Hit]

AAPG/SEG International Conference & Exhibition

Datapages, Inc.Print this page

Recent Advances and Experiences in Subsalt Seismic Imaging



Over the last decade we have seen enormous advances in subsalt imaging thanks to simultaneous advances in seismic acquisition, imaging algorithms, and high-performance computing: wide-azimuth, long-offset streamer configurations and node acquisitions improved illumination of targets, and long-published wave-equation-based methods, notably reverse-time Previous HitmigrationNext Hit (RTM) and Full Waveform Inversion (FWI) finally became practicable. Early experiences with such methods confirmed their value for subsalt imaging but also revealed limitations some of which have been resolved by advances in the algorithms. For example, we are now able to calculate RTM gathers in offset, similar to those from Previous HitKirchhoffNext Hit Previous HitmigrationNext Hit, allowing similar post-processing and RMO analysis, in a cost-effective way. This improves the output image and offers enhanced input for subsalt tomographic updates. We are now studying combining this method with a true-amplitude imaging condition for subsalt AVO analysis. This imaging condition amounts to applying an approximate inverse of the Hessian operator for an underlying least-squares data misfit function, to the standard Previous HitmigrationNext Hit. We can go further and explicitly aim to solve the inverse problem iteratively, that is, do a least squares Previous HitmigrationNext Hit. This further improves amplitudes and resolution, but each iteration costs about as much as a Previous HitmigrationNext Hit so it is extremely expensive; preconditioning by true-amplitude Previous HitmigrationNext Hit may significantly reduce the number of iterations, and so the cost. Our gather calculation further optimizes the benefits of least-squares Previous HitmigrationTop.

FWI is good for high-resolution velocity estimation, but its application to improving models with salt bodies is thought to be problematic, requiring complex workflows. However our recent experience suggests that in some cases a simple FWI methodology may benefit subsalt imaging by improving the suprasalt velocities. Direct subsalt velocity updating by FWI seems seductive but probably requires a combination of increased acquisition effort and better algorithms to achieve the required penetration while avoiding cycle-skipping. On the other hand recent progress on image-domain tools such as wave-equation tomography based on RTM gathers may have positioned them to succeed classical tomography for subsalt contexts.

In this paper we shall present a number of these advances that are contributing to the ongoing improvement in our subsalt imaging, with illustrative examples.