Using Ambient Seismic Noise to Determine Short-Period Phase Velocities and Shallow Shear Velocities in Young Oceanic Lithosphere

Abstract

Using 10 broadband ocean bottom seismometers from the 11-month-long deployment of the Gravity Lineations Intraplate Melting Petrologic and Seismologic Expedition (GLIMPSE) passive seismic experiment located in the south central Pacific, we have estimated the seismic impulse responses from ambient seismic noise for 45 station-to-station paths. The raw impulse responses show moveout with station-to-station distance, and there is a trend of decreasing signal-to-noise ratios as the station-to-station distance increases. The decrease in signal-to-noise ratio is expected as a smaller range of azimuths of propagating surface waves will contribute constructively to the cross-correlated signal with increasing distance, although scattering may also play a role in the coherence of seismic noise at periods less than 16 sec. From these station-to-station paths, we determined group velocities for the fundamental mode Rayleigh waves of a 2–16-sec period and the second mode Rayleigh wave of a 3.5–7-sec period. We calculate phase velocities for the fundamental mode and second mode Rayleigh wave over the same period range as the group velocities by applying a time variable filter to the noise correlation function and carefully unwrapping the phase spectrum of the resulting filtered impulse responses. Within this period range, there is a transition from waves at short periods, whose energy is mostly in the water column, to waves sensitive to crustal and upper mantle structure. The phase velocities for the second mode, which have peak sensitivity in the lower crust and shallow mantle, show evidence for azimuthal anisotropy. The average phase velocities of the station-to-station paths in the east–west direction are 2% faster than the north–south paths at the 4–7-sec period, consistent with the fast directions determined from SKS wave splitting measurements of N100°E. By incorporating the short-period fundamental and higher mode phase velocities from ambient seismic noise with longer period (16–100 sec) fundamental mode Rayleigh-wave phase velocities determined from teleseismic events, we inverted for the average crustal and upper mantle shear velocity structure and water column depth and velocity. The predicted phase velocities are extremely sensitive to the water column compressional velocity. We determined the average water column velocity to be 1466±3 m/sec, in contrast to the average of 1500 m/sec estimated from shipboard measurements weighted according to the Rayleigh-wave sensitivity kernel. The difference may be due to the dispersive effects of scattering by bathymetry or by the thin variable thickness sediment layer. The inversion also produces a Vp/Vs ratio of 1.88 for the crust. This method can provide useful information about the shallow seismic structure of the oceanic crust and uppermost mantle and is an important complement to longer period studies.