Directly Imaging Steep Fault Zones Using Multicomponent Seismic Data

Abstract

Elastic reverse-time migration (ERTM) is the most powerful tool for imaging complex subsurface structures. However, conventional imaging conditions used in ERTM often produce significant artifacts caused by the cross correlation of full wavefields and the polarization of shear waves (S-waves). We use a novel 2D ERTM imaging technique to reduce these artifacts and directly image fault zones. We employ a wavefield separation scheme to obtain downgoing and upgoing waves, and leftgoing and rightgoing waves for both compressional (P) and shear waves within an imaging region. Imaging with these wavefields eliminates the low-wavenumber image noise caused by cross-correlation of full wavefields. We then adopt an imaging condition in the angle domain to properly account for the S-wave polarization, and obtain compressional-to-compressional (PP), compressional-to-shear (PS), shear-to-compressional (SP), and shear-to-shear (SS) images. Our efforts are primarily focused on imaging steeply-dipping fault zones. We validate our ERTM imaging technique using synthetic elastic-wave reflection data for a geophysical model built using geologic features found at the Soda Lake geothermal field, and obtain significantly improved images of the fault zones compared to those obtained using conventional ERTM.