2.5D Time-Domain Seismic Wave Modelling in Viscoelastic Anisotropic Media

Abstract

Our study enhances seismic wave modeling through the development of 2.5-D timedomain viscoelastic anisotropic methods, tailored for efficiently and accurately simulating wave propagation in complex 2-D subsurface geological settings. These settings often include intricate features like free-surface topography, fluid-solid interfaces, and composite media, making our approach highly beneficial in cases where 3-D modeling is resource-intensive. Our methods, designed to handle the viscoelastic wave equations for heterogeneous tilted transversely isotropic media, incorporate newly-developed advanced techniques like the generalized recursive convolution and a real-domain fully parallelized computing strategy to effectively solve equations while keeping computational demands manageable. Experimental validation against complex 2-D models and application to benchmark geological models like BP2007 have proven our method's capability in accurately simulating 3-D wavefields in composite viscoelastic anisotropic media, aiding in seismic data interpretation and highresolution subsurface imaging. Additionally, we address the point-source to line-source wavefield transformation challenge for 2-D full waveform inversion, improving amplitude compensation in complex media through a simple filter and modified stretching factors, offering a more accurate transformation than traditional methods. This comprehensive approach significantly contributes to advancing seismic exploration, subsurface imaging, and geohazard detection by providing a robust and efficient tool for high-resolution imaging of the subsurface.

Date of Award	25 May 2024
Original language	American English
Supervisor	Zhou (Supervisor)