PSTM images after constant Q compensation of 175. GeoStreamer (right) shows much better resolution than the standard streamer (left).

The quest for the remaining oil and gas fields is getting harder. Reservoirs are deeper, covered with complex overburden and within increasingly complex stratigraphic traps. To meet this need, seismic exploration needs extended bandwidth: more low frequencies for better penetration (subsalt and sub-basalt) and more high frequencies for increased resolution.

Unfortunately, the bandwidth in marine seismic is constrained by ghosting (the result of interfering reflections from the water surface). Depending on the recording sensor depth, some frequencies are boosted while others are attenuated. A shallow streamer is favorable to high frequencies and detrimental to low frequencies. Conversely, a deeper streamer favors low frequencies at the expense of high frequencies. Marine seismic acquisition therefore involves a trade-off between high and low frequencies, which inherently limits data bandwidth.

GeoStreamer is a new streamer design proprietary to PGS that exactly removes the effects of ghosting. It uses two types of sensors, pressure and velocity, the combination of which allows the separation of the up-going and down-going waves.

Consequently, the full data bandwidth is restored, and there is no trade-off between high and low frequencies. This in turn benefits both penetration and resolution of seismic reflections to provide a clearer image of the subsurface.

Since its launch, the tool has been used in the North and Norwegian Seas, the North West Shelf of Australia, the Gulf of Mexico, and Trinidad and Tobago, as well as the Mediterranean (Cyprus and Egypt). Results from these surveys reinforce the anticipated benefits, namely better penetration, enhanced resolution, and improved multiple attenuation and imaging. Additionally, some exciting developments are occurring.

For the first time bandwidth is sufficient to determine the quality factor (Q) from surface seismic, the resulting images provide the required penetration for complex areas, and the resolution is ideal for reservoir characterization and 4-D.

Q estimation and compensation

Earth attenuation happens everywhere but is particularly noticeable in the North Sea. In a classic 1992 paper, Roy White demonstrated that Q cannot be determined from surface seismic data alone due to lack of bandwidth. The uncertainty is in fact inversely proportional to the cube of the bandwidth. The new streamer offers such a large bandwidth that Q estimation is now possible without the need for well or vertical seismic profile data.

An early test dataset acquired in the North Sea used two streamers in an over-under mode: a standard PGS solid streamer towed at 26 ft (8 m) and a GeoStreamer towed at 49 ft (15 m).

Both streamers recorded the same shots under very good weather conditions.

The spectral shapes are typical effects of earth attenuation Q. The broader bandwidth of the GeoStreamer and particularly the low frequencies allow Q to be measured for the first time directly from the surface seismic. In this case a Q value of 175 was extracted from the data, and the resulting spectra are correctly flattened after compensation for all depths. The final GeoStreamer image offers much better resolution.

Exploration objectives vs. dual-sensor streamer study results

A regional 2-D streamer survey acquired over the North West Shelf of Australia showed that the recorded pressure and particle velocity data from the GeoStreamer yield complementary information about the subsurface and together in processing result in significant data resolution, quality, and deep target imaging improvements.

Much of the North West Shelf is affected by relatively near-surface barriers to deeper target seismic imaging and resolution. Transpressional structures and complex swarm faulting affects the region and requires rich high frequency content for high-resolution seismic imaging. Analysis of the dual-sensor data shows that frequencies beyond the source ghost in the up-going pressure wavefield have much stronger amplitudes, up to at least 200 Hz. This particularly benefits thin-bed resolution and subtle stratigraphic interpretation.

The presence of carbonates produces complex seismic multiple noise events which create problems for velocity picking and seismic imaging. Production velocity picking was, however, significantly easier on up-going pressure data, in large part because of the characteristic low-frequency boost (see Figures 3 and 4). A clearer primary semblance trend is evident during analysis, and associated improvements in later stage velocity analysis subsequently derive from improved multiple removal.

Up-going pressure data have about 10 dB stronger amplitudes below 10 Hz through the target window in comparison to total pressure data from conventional streamer data.

Target Triassic pre-rift fault blocks have an improved frequency range of about 20 to 30 Hz using a 12 dB down criteria. Estimation of Q is more robust using up-going pressure data as a consequence. Post-rift fault reactivation is more clearly interpreted to affect internal reservoir and trap geometry. Finally, a thick pre-rift section contains dolerite intrusions, the mapping of which is key to constraining basin evolution. Figures 3 and 4 demonstrate the overwhelming improvements in deep target image clarity and character on up-going pressure data.

The increased GeoStreamer bandwidth is not only beneficial for final interpretation but also helps derive more accurate attributes. These attributes are used to optimize the seismic image but can also be used for interpretation. The high frequencies benefit the interpretation of subtle stratigraphic features and allow clearer imaging of features affected by transpressional fault reactivation. The typical low-frequency boost of at least 10 dB below 10 Hz has many benefits, including better velocity semblances (particularly in poor data areas affected by carbonates), more robust Q estimation, and significant signal penetration improvements for the seismic imaging of Triassic fault block plays and deeper dolerite intrusive features. Overall, the dual-sensor dataset has better resolution and is more interpretable at all depths.

On track and on target

Deployment of the GeoStreamer across the PGS fleet is on schedule, with Geo-Streamer currently operating on three PGS 2-D vessels and one 3-D vessel. The rest of the fleet will be gradually equipped with the technology in the next few years.

The tool has also proven to be quite rugged. Because it is towed deeper than a regular streamer, it is less sensitive to swell noise and can therefore operate when other vessels cannot.