Summary

Slanted cable marine data acquisition has gained popularity in recent years as an aid to receiver deghosting and up-down wavefield separation. The rationale for using slanted streamers as opposed to a conventional flat configuration is related to the well-known ghost model that predicts that no signal will be recorded at certain frequencies, known as notch frequencies. Various authors report that, by varying the cable depth, the associated notch frequencies vary, which aids in producing a final deghosted image. In this paper, we study the need for slanted cable acquisition in the context of wave-equation deghosting, which has shown that the notch frequencies, while attenuated, are not zero. We find that, in the case of noise-free data, slant cables offer no advantage; however, with the introduction of noise, we find the slant configuration shows improved deghosting and is less sensitive to the noise.

Introduction

The concept of using a slanted marine cable as a mitigation strategy for the receiver ghost problem has been known for some time, but has only recently seen commercial application (Ray, 1982; Dragoset, 1991; Soubaras, 2010). The rationale for slanting the cables in depth is based on the traditional ghost model that indicates that the ghost notch is a function of receiver depth. Changing receiver depth along the cable by means of slanting the cable, thus, changes the receiver ghost-notch frequency. As a result, when traces with different source-receiver offsets are combined, as they are in processes such as stacking and migration, frequencies attenuated at one offset are complemented by other traces with different offsets where these frequencies are not so attenuated. Moreover, prestack deghosting algorithms can possibly take advantage of this so-called notch diversity by means of interpolation or other types of estimation of the missing frequencies based on nearby traces.

Causal deghosting approaches, in theory, can deghost prestack data for even the notch frequencies without resorting to interpolation or estimation procedures, which has been demonstrated for both flat and irregularly shaped receiver arrangements (Beasley et al., 2013a; Beasley et al., 2013b; Ferber and Beasley, 2014; Robertsson et al., 2014). However, in the real world of seismic data acquisition, we observe significant noise both in the seismic data itself as well as in the estimates of physical parameters needed for deghosting, e.g., errors in estimating water velocity and receiver depths. Because these techniques are based on a wave-equation extrapolation procedure, the process uses nearby receivers to perform the extrapolation. In a slanted cable configuration, the receivers exhibit different ghost characteristics – in particular, different signal-to-noise (S/N) ratio characteristics near the notch frequencies – so it is natural to investigate if slanted cables would also benefit prestack wave-equation deghosting. In this paper, we model the problem using synthetically generated seismic data to which we add realistic seismic acquisition noise. We find that, in this particular case, a slanted cable arrangement offers a slight advantage in prestack deghosting over a flat cable acquisition geometry.

This content is only available via PDF.
You can access this article if you purchase or spend a download.