True-Amplitude Seismic Imaging Beneath Gas Clouds

More Info
expand_more

Abstract

A gas cloud is a region of gas accumulation in the subsurface, which can severely deteriorate the seismic data quality from deeper reflectors. Due to complex wave propagation through the anomaly and the resulting transmission imprint on the reflections from below this area, the image below the gas cloud is usually not properly recovered. The reflected events in this region appear with lower amplitude and lower frequency content, which is often refers to as ‘Q-attenuation’, ‘Q-absorption’, or anelastic or intrinsic loss. This thesis describes a new approach to imaging below such anomalies for seismic exploration purposes. The new full-waveform inversion-based approach can contribute to this largely unsolved problem. The approach differs from traditional solutions, which aim at effective amplitude and phase corrections of the seismic reflections (removing the ‘absorption’ effect), whereas the new developed approach treats the problem as a complex-scattering phenomenon. The basic principle of the proposed method is that the reflection response of the complex gas-cloud area, including its coda, carries detailed information on the gas-cloud properties, which can be translated into transmission-correction operators. We aim at constructing full-waveform transmission operators (including the codas) from the gas-cloud reflection response via an effective-medium representation, obtained through nonlinear full-waveform inversion. In our case we have used a Genetic Algorithm for this purpose. From the property model of the gas cloud, the transmission operators are determined by forward modeling. It has been shown that true-amplitude imaging of reflections below the gas cloud is achieved via multidimensional deconvolution of the seismic data for these full waveform transmission operators. For a 2D synthetic dataset, a suitable property model of the gas cloud region is obtained. Redatuming the seismic data using the full-waveform transmission operators from this effective medium model show that primary reflections underneath the gas cloud are very well recovered and the transmission codas are reduced. The method was also applied to a 2D real marine dataset, where we aim at inverting for the heterogeneous overburden above the gas-cloud region. The results after the true-amplitude redatuming show that the amplitudes of the target horizons are better recovered. These results are very promising, considering the fact that there are many effects that could limit the quality of the final result, such as the 3D geometry effects of the gas-cloud body itself, mode conversions that are neglected, limitation in the source estimation and restrictions in the acoustic forward-modeling algorithm. The stack after the full-waveform redatuming displays better and more continuous reflections. The results of this research has confirmed our belief that, what is often referred to as ‘absorption’, probably to a large extent is caused by complex wave-propagation effects. Although some energy really may disappear into heat (anelastic loss), we believe that a large part of the energy is scattered in elastic mechanisms (internal scattering, multipathing etc.) and, therefore, is recoverable by full-waveform deconvolution.