Annual Meeting - SCEC.org

Poster Abstracts | Group 2 – Seismology
Several new methods have been developed to retrieve local Green’s functions based on the crosscorrelation
of ambient noise (station-to-station) and conventional (source-to-station) inversions.
The latter methods provide the most broadband results but require accurate source parameters for
phase-delay recovery which depends on the starting model. We can avoid this trade-off by
applying the cut-and-paste technique which allows adjustable timing shifts to correct paths. We
use a tomographic model based on these delays from a large population of events to predict those
derived from the recent Chino Hills event and perform a test against direct inversion results.
Delays derived from ambient-noise relative to the nearest stations are generally compatible but
indicate multi-pathing issues along some paths. It appears that these two methods can be
combined to achieve more accurate results.
2-136
COMMUNITY VELOCITY MODEL FOR THE NEW MADRID REGION, CENTRAL
U.S. Boyd OS, Soble J, and Verbanaz R
In 1811-1812, a series of three major earthquakes struck the Central United States in the New
Madrid Seismic Zone. Having magnitudes near 7.5 and being located within the relatively stable
interior of the North American Continent, these events produced widespread strong shaking. If
these events were to occur today, there would be substantial devastation to people, buildings and
transportation and communication infrastructure. To better understand this threat and in
preparation for the upcoming bicentennial, the US Geological Survey is planning to produce and
support sophisticated numerical simulations of earthquake rupture and seismic wave propagation
due to a repeat of these events. To kick off this effort, the USGS began the development and
construction of a community seismic velocity model for use in these numerical simulations. We
have collected existing research regarding the p- and s-wave velocities, impedance contrasts and
densities of the lithosphere in the New Madrid region and synthesized these results into a single
model that can be used in earthquake simulations. We have identified areas of missing or
incomplete information for further study. The region covers an area of approximately 600,000
square km from Little Rock, Arkansas across to Nashville Tennessee, up to St Louis, Missouri. The
model has currently been gridded at 3 km lateral resolution and from 5-m resolution near the
surface to 10-km resolution at 100 km depth. Less but still substantial uncertainty exists for the
Mississippi Embayment where a majority of the research has been done. Newer regional models
such as those by van der Lee and others and Liang and Langston have improved regional
resolution beyond a 1-dimensional model, but for ground motion simulations, greater resolution
outside the Embayment is desired.
2-137
SURFACE WAVE CONSTRAINTS ON CRUSTAL S-WAVE VELOCITY STRUCTURE
Yano TE, Tanimoto T, and Alvizuri CR
S-wave velocity structure at shallow depths is crucial for accurate ground motion prediction and
seismic hazard. In order to improve crustal S-wave velocity structure, we have developed a new
procedure for joint inversion of Rayleigh wave short period data (about 0.1-0.4 Hz) using particle
motion (ZH ratio or equivalent to H/V) and phase velocity data.
Our particle motion data are the ratios between vertical and horizontal amplitudes of Rayleigh
waves (hereafter ZH ratios), measured from ambient noise. The most sensitive depth range is in the
upper 5 km, thus this data set will be crucial to constrain S-wave velocity in very shallow structure.
The most attractive aspect of this data set is that it is free from contamination from deep structure.
Basic procedures to obtain fundamental Rayleigh wave signals were described in Tanimoto and
Alvizuri (2006). Data were measured from continuous records typically for two years, for
212 | Southern California Earthquake Center