Online Journal for E&P Geoscientists

Caselitz, Bertrand ^*1; Monir, Emad ¹; Halim, Abdallah ¹
(1) PGS, Nasr City, Cairo, Egypt.

Noise generated in the water column, such as seismic interference (SI), rig noise and ship noise presents a significant challenge to processing conventional streamer data (Brittan et al., 2008). The interference noise is characterized as linear events of variable frequency bandwidth with an arbitrary apparent velocity on the seismic record. It is these characteristics which make the attenuation of such noise a processing challenge.

Interference noise attenuation processing techniques can be divided into two groups: the first group of methods model or predict the undesired noise and then subtract it from the input data; the second group of methods apply noise and signal separation methods in some domain where the noise or the signal is less coherent. Both groups of techniques work well when the interference noise has significantly larger amplitude and a different apparent velocity than the source signal.

In 2009, an interference noise technique using a dual sensor streamer was proposed (Cambois et al., 2009). Seismic interference noise generally travels through the water column only and is thus most likely to be incident upon the streamer with an angle that is closer to the horizontal plane than the vertical plane. As streamer geophones are sensitive to wave direction, the interference noise will be preferentially recorded by the omni-directional streamer hydrophones. The technique of Cambois et al. (2009) is based on cross-ghosting each coincident pair of sensors, in order to derive the events that are not common to both phones and produce a noise model. In this paper we both illustrate the use of this approach and propose a second methodology to tackle the interference noise based on a dual sensor streamer acquisition. The second technique utilizes the up-going and down-going pressure field that is obtained by summation and subtraction of a hydrophone and geophone. The down-going field is re-datumed to produce a ‘pseudo’ up-going pressure field. The difference between the up-going and the re-datumed down-going pressure fields produces an interference noise model which is then adaptively subtracted from the up-going pressure field. Both the Cambois et al. (2009) method and the method described in this paper have the merit to depend only on the noise difference between hydrophone and geophone data. This allows a successful attenuation of the seismic interference noise even when the noise has weak amplitude and/or no dip discrimination.