Scalable seismic acquisition and algorithms for next-generation engineering geophysics

Abstract

Geoscientists and engineers are increasingly using denser arrays for continuous seismic monitoring, and often turning to ambient noise interferometry for low-cost near-surface imaging. While ambient noise interferometry greatly reduces acquisition costs, the computational cost of pair-wise cross-correlations can be prohibitively slow or expensive for a number of applications in engineering and environmental geophysics. Thus, we are motivated to avoid pair-wise cross-correlations when possible. By rearranging the operations involved in slant-stacks for dispersion image calculation, a quadratic operation becomes linear. However, these dispersion images tell us about surface-wave behavior, but the type of high-resolution near-surface imaging we aim to obtain from ultra-dense seismic arrays requires identification of body waves through double beamforming. This abstract extends this result to propose a new algorithm for double beamforming which is significantly more scalable, easily parallelizable, and does not require raw data to be exchanged between dense array patches.