Imaging and Inversion of Reflected Surface Waves
Hyslop, Craig W. 1978-
MetadataShow full item record
Surface waves that reflect, or backscatter, from lateral changes in the near surface provide information about the location and depth of anomalies as well as their properties such as reflectivity and velocity. We develop 2D and 3D algorithms that forward model surface waves and image surface-wave reflectors. Our forward model is based on surface-wave propagation in a vertically layered model. We extend solutions to account for lateral heterogeneity in 2D and 3D. By using high-resolution dispersion-curve estimation and array-windowing schemes, we improve upon existing methods of determining shear-wave velocity from direct surface waves. Surface-wave reflection imaging along the survey line provides a 2D estimate of lateral surface-wave reflectivity as a function of spatial location and frequency (or depth). The 2D reflectivity image has sharper lateral resolution than shear-wave velocity models estimated from direct surface-waves. We verify our method of modeling and extracting reflectivity using numerical and physical modeling. The normalized root-mean-square deviation between our forward modeling and extracted reflectivity from data generated by SPECFEM2D is 14%. Extracted reflectivity from a physical model with a vertical fault also matches reflectivity predictions. Reflectivity maps from Hockley Fault near Houston, Texas are consistent with evidence for faults in traditional body-wave reflection images. By modifying the dispersion relations for wavefield extrapolation methods, we model and migrate surface-waves across the 2D free surface to create a 3D reflectivity image. Given a single survey line we show that the wavefield within the direct surface-wave cone, often identified as noise, contains valuable reflections that should be used for imaging. Migrated surface-wave images from the Bradford 3D seismic survey in Pennsylvania and seismic data from the Arctic Slope of Alaska correspond to topographical features in the area. Using synthetic models, we show that having sources and receivers within the image greatly improves migration results. We use reflectivity to update surface-wave phase-velocities (determined by picking dispersion curves) across the survey line. By inverting the updated phase-velocities, we estimate a shear-wave velocity model that is sharply defined in the lateral direction. The updated velocity model for the Hockley Fault system near Houston, Texas helps identify normal faulting and a small graben feature near the main fault. Imaging and inversion of reflected surface waves promises to deliver an augmented picture of the near surface.