Broadband Processing of Conventional Marine Streamer Seismic Data

Abstract

We have developed WiBand, a de-ghosting and broadband processing method that can be applied to conventional streamer data acquired with single component streamers. We recorded, over one sail line, multi-streamer data with the streamers towed at different depths, obtaining data over multiple subsurface lines (tracks) parallel to one another. We apply WiBand to each track independently to remove the effects of the source and receiver ghosts. We compare the WiBand results from a deep tow track to conventionally processed result from a shallow tow track to validate the phase fidelity of the de-ghosting process. The results lead us to conclude that WiBand correctly removes the ghosts and recovers broadband data from deep tow streamer data. Introduction The vast majority of marine seismic data are acquired with conventional towed streamer cables which are equipped only with hydrophones and are all towed essentially at the same constant depth for any single survey. The receiver ghost zeros out the spectral response of the recorded data at notch frequencies equal to any integer divided by the ghost delay time, limiting both the top end of the usable spectrum as well as attenuating the low frequency response. The source ghost has a similar effect. The combined effect of the ghosts results in greatly reduced resolution in the subsurface image. A conventional "solution" for this problem is to tow the sources and streamers shallow (at less than 7m of depth below the surface) to obtain higher frequencies, at the expense of attenuating low frequencies and exposing the sensors to a noisier environment near the sea surface. There are currently several acquisition methods that can mitigate or remove the effects of the ghosts and obtain broadband seismic images (Carlson, 2007, Posthumus, 1993, Soubaras, 2010). These methods require nonconventional data acquisition, either with streamers equipped with multi-component sensors, or with streamers configured in unconventional geometries such as over-and-under or slanted configurations. Such methods are not applicable to conventional streamer data. We have developed an effective broadband processing method based on a new de-ghosting technique. It can remove most of the ghost effects from conventional streamer data. In this paper we will refer to this method as WiBand. This method is designed to address both the amplitude attenuation and the phase distortion introduced by the ghosts to obtain nearly flat spectral response in the typical range of 4Hz to 150Hz, as well as a compact, well focused seismic wavelet. In order to validate our method, we carried out an experiment where multiple streamers were towed at different depths, and evaluated the phase reconstruction fidelity of the algorithm by comparing the WiBand result from the deep tow data with the standard processing result from the shallow tow data.