Ground roll: Rejection using adaptive phase‐matched filters

Abstract

The technique of phase‐matched filtering dispersive surface waves is extended to permit an adaptive, iterative process by which the signal itself in a seismic trace designs a filter to remove the surface wave. The technique is robust and well‐behaved and requires the specification of only simple parameters for its operation. The technique is applied to data sets from three regions, representing a wide range in the ratio of surface‐wave noise to exploration signal. The technique works very well with poor data sets and also improves good data sets. Since the technique is applied to individual traces, it works in situations for which f‐k filtering might not be feasible due to poor spatial sampling. The technique is computationally more intensive than recursive digital band‐pass filtering of individual traces, but is less intensive than filtering in the f‐k domain.