HR: 16:45h
AN: S54B-04    [Abstracts]
TI: Hunting the Strike of a Dipping Interface With Receiver Function Moveout and Polarity Variations
AU: * Park, J
EM: jeffrey.park@yale.edu
AF: Yale University, Dept of Geology and Geophysics, New Haven, CT 06511, United States
AU: Katsuhiko, S
EM: katsuhiko.shiomi@yale.edu
AF: Yale University, Dept of Geology and Geophysics, New Haven, CT 06511, United States
AU: Vadim, L
EM: vlevin@rci.rutgers.edu
AF: Rutgers Univ., Dept of Geological Sciences, Piscataway, NJ 08854, United States
AB: P-to-S converted waves that originate at interfaces within Earth's crust and mantle (Ps phases) suffer directional variations in amplitude and polarity that depend on the dip of the interface and the presence of elastic anisotropy in the surrounding rock. The moveout of the Ps phase revealed by receiver-function analysis of data covering a range of backazimuths should constrain the dip and strike of a dipping interface, but Ps amplitude and polarity variations can cloud the determination. We report a simple technique for utilizing directional variation in Ps phase attributes that helps constrain its moveout, offering a way to distinguish a dipping-interface effect from an anisotropic one, e.g. in a subduction-zone environment. The radial and transverse components of Ps are both affected, in a similar way, by converting interface dip and anisotropic axis of symmetry tilt. In both cases a phase-shifted two-lobed back-azimuthal pattern of Ps amplitude on the radial and transverse RFs is expected. (Horizontal-axis anisotropy predicts a four-lobed pattern.) The relationship between radial and transverse RFs can be exploited in stacking to enhance signals, with one phase-relationship "correct" for anisotropy and interface dip, and the opposite phase-relationship "unmodelled." The low amplitude of the RF stack for the "unmodelled" phase is an indicator of the success in finding "correct" stacking rules. We employ a parameter search over interface strike and moveout values to find a minimum in the "unmodelled" RF stack. In cases tested, this minimum correlates well with the known strike and dip of a subduction-zone interface beneath a station of interest, even where the anisotropic symmetry axis is known to deviate from the slab dip e.g., station COR in Cascadia and station CUC in Calabria. Relating the optimal Ps moveout to slab dip depends on the assumed velocity model, though we observe that the optimal stacking moveout increases with slab dip, as expected.
DE: 3270 Time series analysis (1872, 4277, 4475)
DE: 7203 Body waves
DE: 7240 Subduction zones (1207, 1219, 1240)
SC: Seismology [S]
MN: 2007 Fall Meeting