Nonlinear surface wave prediction and separation

Removal of surface waves is an integral step in seismic processing. There are many standard techniques for removal of this type of coherent noise, such as f-k filtering, but these methods are not always effective. One of the common problems with removal of surface waves is that they tend to be aliased in the frequency domain. This can make removal difficult and affect the frequency content of the reflector signals, as this signals will not be completely separated. As seen in (Hennenfent, G. and F. Herrmann, 2006, Application of stable signal recovery to seismic interpolation) interpolation can be used effectively to resample the seismic record thus dealiasing the surface waves. This separates the signals in the frequency domain allowing for a more precise and complete removal. The use of this technique with curvelet based surface wave predictions and an iterative L1 separation scheme can be used to remove surface waves from shot records more completely that with standard techniques.