First-arrival traveltime tomography at Long Beach California using ambient seismic noise and the adjoint-state method

Jorge A. Castillo Castellanos, & Robert W. Clayton

Submitted July 25, 2018, SCEC Contribution #8196, 2018 SCEC Annual Meeting Poster #095

Knowledge of the fine-scale velocity structure of the subsurface is crucial for predicting ground motion and, thus, hazard assessment. In a recent study, Lin et al. (2013) showed that ambient noise surface wave correlations from a high-density oil-company survey could be used to produce a detailed 3D shear-wave velocity model for the top 800 m of Long Beach, CA. In this study, we extend their analysis and use the P-wave portion of the Green’s function to generate multiple refraction profiles within Long Beach and determine the velocity structure to a depth of about 3 km. To overcome the inherent challenge of extracting body waves from ambient noise, we divide the Long Beach network into small-aperture subarrays and apply a double beamforming technique between all group of stations to isolate P waves from surface waves. The resulting beamed traces allow us to identify prominent direct and weak scattered arrivals traveling between all subarray pairs. To produce the final velocity profiles, we use a stacked seismic section to derive an initial 1D model and the adjoint-state method to construct the gradients of the misfit functional without the necessity of calculating and saving sensitivities (Fréchet derivatives). The inverted velocity profiles show velocity variations of the order of 0.3 km/s and strong lateral discontinuities caused by the presence of sharp structures such as the Newport Inglewood Fault. These profiles also provide us with a set of static corrections that will be applied to a follow on reflection analysis, which will in turn allow us to generate an even more detailed image of the subsurface.

Key Words
Ambient Noise, Long Beach, Tomography

Citation
Castillo Castellanos, J. A., & Clayton, R. W. (2018, 07). First-arrival traveltime tomography at Long Beach California using ambient seismic noise and the adjoint-state method. Poster Presentation at 2018 SCEC Annual Meeting.


Related Projects & Working Groups
Seismology