Ground roll rejection via f-v filtering

Yi-Yang Song, Robert R. Stewart

Ground roll is often apparent on seismic records and as it can obscure events of greater interest, it generally needs to be suppressed. This paper discusses the use of f-v falters operators in the 1-D Fourier transform of the (,p) domain to reject ground roll. Several implementations of -v filters are used: a surgical filter to suppress low velocity events; a phase difference filter (supposing the ground roll has 90 degree phase difference in both vertical and radial component data); and a cross-correlation filter ( assuming ground roll has a similar amplitude signature on both vertical and radial channels). These filters are developed and applied to synthetic data and field data from the Crystal area of central Alberta Canada. We find that f-v filters outperform f-k filters on the cases discussed here.