HR: 0830h
AN: S11E-0344 [PDF]
TI: Coherence-Weighted Wavepath Migration of Teleseismic Data
AU: * Sheng, J
EM: sjm@geophys.utah.edu
AF: University of Utah, Dept of GG
WBB RM# 719, Salt Lake City, UT 84112 United States
AU: Schuster, G T
EM: schuster@mines.utah.edu
AF: University of Utah, Dept of GG
WBB RM# 719, Salt Lake City, UT 84112 United States
AU: Pankow, K L
EM: pankow@seis.utah.edu
AF: University of Utah, Dept of GG
WBB RM# 719, Salt Lake City, UT 84112 United States
AU: Pechmann, J C
EM: pechmann@seis.utah.edu
AF: University of Utah, Dept of GG
WBB RM# 719, Salt Lake City, UT 84112 United States
AU: Nowack, R L
EM: nowack@purdue.edu
AF: Purdue University, Dept. of Earth and Atmospheric Sciences, West Lafayette, IN 47907 United States
AB:
Migration of teleseismic records from earthquakes
can image a selected set of reflections or converted phases
to their correct locations of origin, but will also image
other phases to incorrect locations. For example, a
migration algorithm tuned to ps conversions can correctly
image the Moho but may also focus ghost reflections from
the moho such as pPs or pSs to incorrect reflector locations. (Here,
lower and upper case letters indicate upgoing and downgoing waves,
respectively.) This coherent "noise" from other phases can be
suppressed by stacking images migrated from many teleseisms
but, in practice, the timely
acquisition of well-recorded events can be impractical.
To partly remedy this problem, we propose coherence-weighted
wavepath migration of teleseismic data. The idea is to first apply
wavepath migration to, say, ps, pPs, and pSs arrivals in receiver
functions and pPp arrivals in autocorrelograms.
Then, a local semblance analysis is applied to the resulting
ps, pPs, pSs and pPp images to calculate the coherence among
these images and a weighting factor proportional to it.
Correctly migrated events are coincident at the same locations
resulting in high coherence weights at these points.
Smaller weights are computed elsewhere. The coherence
weights are applied to the ps image to give the final image.
In addition, selective migration of the ps arrivals
is partly achieved by wavepath migration of receiver functions
using special imaging conditions that account for the expected
particle motion of ps conversions and their apparent moveout.
To demonstrate the viability of this procedure, we computed
elastic synthetic seismograms for a 4-layer crustal Utah model
using a plane P wave source incident at a 40 degree angle.
Receiver functions were computed and ps conversions were
migrated to give the reflectivity distribution $r({\bf x} )^{ps}$.
The pPp events were migrated from
autocorrelograms to give $r({\bf x} )^{pPp}$.
Semblance weights were computed for the $r({\bf x} )^{pPp}$ and
$r({\bf x} )^{ps}$ images, and a final image was obtained by
applying the semblance weights to
the $r({\bf x} )^{ps}$ image.
Comparing this image to the unweighted ps image shows
the efficacy of this approach.
Our poster will also show results from migrating
teleseismic data recorded by the Utah Regional Seismic Network.
DE: 7203 Body wave propagation
DE: 7205 Continental crust (1242)
DE: 7260 Theory and modeling
SC: Seismology [S]
MN: 2003 Fall Meeting