Rayleigh-Wave Analysis and Removal Using a Novel Weighted Median Filter

David G. Schieck

In this study a non-linear, hybrid f-k filter is used to remove aliased dipping events and random noise glitches from a seismic shot gather. This weighted median filter uses time-domain coefficients, from the inverse Fourier transform of a dip or velocity filter developed the frequency domain, as the filter weights. This results in a filter which has median characteristics but also rejects dips analogous to an f-k velocity filter. A fast 2-D median filter algorithm is developed and applied to synthetic and real P-SV data from the Springbank area of Alberta.

Rayleigh waves travelling in the near-surface are often dispersive (phase velocity dependent on frequency). Three methods of estimating dispersion parameters are evaluated in this study: namely, narrowband filter analysis, shear-wave refraction, and the -p transform. A linear approximation to the dispersion curve is used to compress dispersive noise for synthetic data, Springbank data, and a multicomponent data set from Wyoming.