Scattered-Wave Imaging Algorithms for Regional and Local Arrays

Abstract for proposal EAR0106836 (PH #36x)

Title: Scattered-wave imaging algorithms for regional and local arrays

PI's: J. Revenaugh, UC Santa Cruz

The scattering of elastic waves by heterogeneities in the Earth is an obvious and dominant source of waveform complexity at high frequencies, producing slowly decaying trains of energy that are visible in records for several minutes following primary (geometrical) arrivals. Studies attempting to characterize the nature of scattering traditionally have treated it either as a regionally stationary statistical process, or have restricted attention to the coherent portion of the scattered wavefield produced by crustal layering. Both approaches have yielded valuable results and will continue to do so, but each makes restrictive assumptions about the nature of scattering that ultimately limit their utility. We have developed an alternative approach, utilizing a variant of an algorithm developed by the oil industry to image scattering structures in the crust and upper mantle using regional array (aperture of order 100 km) recordings of teleseismic earthquakes. The method has been applied to data from the Southern California Seismic Network and the Hawaiian Volcano Observatory array. The results from both arrays reveal an intriguing and potentially very important geographic correlation of the strength of scattering with the density of background and aftershock earthquakes. More importantly, these results provide an all-important 'proof of concept'that it is possible to extract reliable structural information at scale-lengths less than average station separation.

In our proposal, we document a work plan to improve upon the imaging paradigm employed by our current method (referred to as Kirchhoff coda migration, or KCM) through better treatment of the physics of elastic scattering, more accurate wavefield summation, the incorporation of three-component recordings, and a regularized inversion. In so doing, we will continue to honor the usual exigencies of regional and local array recordings of earthquakes: irregular instrument deployment, data dropouts, extended source signatures (waveform complexities due to the earthquake rupture process and near-source structure), station-site effects and limited illumination back azimuths. Our goal is the production of 'canned' algorithms of broad-spread application to the imaging of crustal and uppermost mantle scattering structures by both the PASSCAL community and to users of permanent dense local and regional array data. We will apply the algorithms to data in hand from Southern California and the island of Hawaii.