Seismic Illumination Analysis with One-Way Wave Propagators Coupled with Reflection/Transmission Coefficients in 3D Complex Media

Abstract

Seismic illumination analysis can help us to better understand how various seismic acquisition parameters and configuration affect seismic image qualities. It allows us to design effective seismic survey systems to provide reliable high-resolution seismic images for complex structures in both theoretical and exploration seismology. It is also a useful tool for estimation of the potential detecting power of a specific acquisition system for a given velocity structure of the medium. Most existing techniques for calculating illumination intensity distributions can be divided into two categories: one is based on ray tracing methods (Berkhout, 1997; Schneider, 1999; Bear et al., 2000; Muerdter et al., 2001a, 2001b, 20001c); and the other based on wave equation methods (Wu & Chen, 2002; Xie et al., 2006; Wu & Chen, 2006; Sun et al., 2007). The ray tracing is quite efficient and inexpensive. However, it bears large errors in complex areas because of the multi-path problem, the high frequency approximation and the singularity problem of the ray theory (Hoffmann, 2001). These factors limit its accuracy severely in complex areas. The wave-equation-based methods can provide reliable and accurate illumination intensity distributions in complex media. Full-wave finite-difference (FD) method is commonly employed for seismic forward modelling, i.e., wave propagation simulation for a known interval velocity model. But it is too expensive for illumination analysis in industrial application, especially for 3D case. Recently, One-way wave propagators have been widely used to seismic modelling and migration for its huge memory-saving and high computational efficiency. Localized illumination methods based on the Gabor-Daubechies frame decomposition or the local slant stack have been presented, which can give local information both in space and direction (e.g., Xie and Wu, 2002; Wu et al., 2000; Xie et al., 2003; Wu & Chen, 2006; Mao & Wu, 2007; Cao & Wu, 2008). However, its computational burden is still unacceptable due to the wavefield decomposition in every spatial point.