HR: 12:00h
AN: S32A-07    [Abstracts]
TI: Model-Based, Near-Surface Scattering Corrections for Short-Period Teleseismic Data
AU: * Campman, x
EM: xander@erl.mit.edu
AF: Earth Resources Laboratory, MIT, 42 Carleton St., Cambridge, MA 02142 United States
AU: rondenay, s
EM: rondenay@mit.edu
AF: Dept. of Earth, Atmosph. and Planet. Sc., MIT, 77 Massachusetts Avenue, Cambridge, MA 02139 United States
AB: Scattered surface waves can cause artefacts in images produced with short-period teleseismic data and may limit the maximum obtainable resolution of these images. We describe initial results from the application of a method for the suppression of scattered waves caused by surface topography in teleseismic data. Contrary to wavenumber-frequency domain filtering, tau-p transforms and polarization filters, which are based on certain characteristics of the noise (in this case the surface waves), our method is a model-based approach that takes some of the complexity of wave propagation into account. The basic idea of this approach is that, if we know the distribution of scatterers near the surface, we can predict and subsequently subtract the scattered waves from the data. To estimate of the distribution of scatterers from the data, we first select a window around a strong arrival that contains the incident P-wave and scattered coda phases. Next, we separate the scattered coda phases from the P-wave using multi-channel filtering. The resulting scattered wavefield is the input for an inversion algorithm that maps the distribution of surface scatterers by fitting the data to synthetics produced by an efficient model for body-to-surface wave scattering. The map is then used to predict and remove scattered surface-waves from the entire record. This algorithm has already been successfully applied to densely sampled, single-component controlled-source seismic data. The advantage of this approach is that it only removes that part of the data which fits the body-to-surface wave scattering model. In contrast, wavenumber-frequency domain filters, for example, remove any wavefield component within a certain range of apparent velocities, including phases that can be used for imaging. Data from the Los Angeles Region Seismic Experiment (LARSE93) are used to validate the present algorithm. In particular, we analyze these data to understand the scattering of waves incident upon the San Gabriel Mountains. We describe pre-processing of the data, then present images obtained with the inversion algorithm, and discuss issues related to the assumptions made in the formulation of our scattered noise model.
DE: 0900 EXPLORATION GEOPHYSICS
DE: 0902 Computational methods: seismic
DE: 0935 Seismic methods (3025, 7294)
DE: 7200 SEISMOLOGY
DE: 7255 Surface waves and free oscillations
SC: Seismology [S]
MN: Fall Meeting 2005