UKCS Cornerstone: a variable-depth streamer acquisition case study

Transcription

1 first break volume 30, November 2012 special topic UKCS Cornerstone: a variable-depth streamer acquisition case study George Moise, 1 Geoff Body, 1 Vincent Durussel, 1 Fabrice Mandroux1 and Jo Firth 1* fill in the details of how a broadband marine acquisition survey deploying variable-depth streamers was undertaken in the North Sea. B roadband variable-depth streamer acquisition is now a mature technology, with over 21,000 km of 2D, 52,000 km 2 of 3D and 8000km 2 of wide-azimuth data having been acquired, both multi-client and proprietary. In this article, we describe a case study of the acquisition of a shallow-water, variable-depth streamer, multi-client 3D survey in Quad 21 of the UK North Sea. The variable-depth streamer broadband acquisition solution uses differences in the receiver depths to produce receiver ghost notch diversity, allowing the streamer to be towed deeper to improve the low-frequency signal-to-noise ratio without compromising the high frequencies. This ghost notch diversity is exploited by proprietary deghosting and imaging techniques (Soubaras, 2010), to produce a wavelet with both high signal-to-noise ratio and maximum bandwidth. This solution capitalizes on the extremely low-noise characteristics and precise low-frequency response of solid streamers (Dowle, 2006). Solid streamers are quieter and can be towed deeper (up to 50 m) than other streamers, so that better low-frequency signal-to-noise ratios are achieved. The new generation of solid streamers have lower instrument low-cut filters, which allow recording down to 2 Hz, providing a full extra octave of data at low frequencies over some other systems. Conventional wisdom was that seismic sources did not produce sufficient low frequency to warrant recording below 5 Hz, but even using a conventional source we routinely observe coherent signal as low as 2.5 Hz. The six octaves of bandwidth provided by this broadband solution produce sharp wavelets with minimal sidelobes, enhancing the fine stratigraphic detail and revealing the genuine seismic reflection response of geologic formation boundaries. Low frequencies have been found to play a major role in suppressing wavelet sidelobes, as well as their obvious application for improving penetration or the imaging of deeper targets. In a mature basin such as the North Sea broadband seismic provides significant enhancements for exploration of stratigraphic plays and subtle structural closures, as well as further enhancing hydrocarbon recovery by providing more detailed information about local facies variation and reservoir compartmentalization. Location The multi-client survey was acquired by CGGVeritas in UK waters covering several blocks in Quad 21 of the Central North Sea. The 2118 km 2 programme (full fold) is situated in water depths varying between m and covers four fields with various obstructions including platforms, FPSOs, and subsea installations (Figure 1). The survey was acquired in two phases which have since been merged to form one BroadSeis dataset. The survey overlaps the existing conventional multi-client Cornerstone survey, which will be merged with the Quad 21 data to create a contiguous dataset. Figure 1 Survey location map. The survey outline is shown in red, with existing conventional Cornerstone multi-client library data shown in pink. 1 CGGVeritas. * Corresponding author, jo.firth@cggveritas.com 2012 EAGE 91

2 special topic first break volume 30, November 2012 Unconformity reflector can be a challenge, especially around salt diapirs. The Gannet B field is associated with both a gas cloud and a salt-induced structure. Figure 2 The CGGVeritas Oceanic Challenger, towing ten variable-depthprofile streamers. Geological background and challenges The CGGVeritas Quad 21 survey covers a large part of the Tay fan sandstones, a reservoir sand for many fields and discoveries in the area, including the Guillemot and Gannet field complexes. Some of the key challenges in this area include seismic imaging through local gas clouds and around salt diapirs. One of the main objectives of this survey is to resolve thin beds and facies variations within the Tay turbidite complex. The deep Jurassic is also of interest and the Quad 21 survey covers the Teal oil field complex producing from the Fulmar sandstone. At depths of about three seconds Two-Way Time (TWT), imaging the Base Cretaceous Acquisition parameters The first part of the survey was acquired by the CGGVeritas Oceanic Challenger during March and April For various operational reasons, including poor weather during mobilization and excessive seismic interference from nearby ocean bottom cable and node surveys, only the western, priority part of the survey was fully covered in the first stage. The remainder was acquired by the Oceanic Endeavour in September and October The survey was acquired using 10 streamers (6 km), separated by 75 m, towed in a variable-depth configuration with receiver depths varying from 6 m to 50 m. Two 4410 cu in, 2000 psi gun arrays were used with a separation of 37.5 m and a depth of 6.0 m, fired at m shot point intervals (37.5 m per array), to provide nominal 80 fold data in 6.25 x m bins. The nominal inline near offset was 115 m (Figure 3). No difficulties were encountered in controlling variabledepth profile streamers during line-turns on this or any similarly acquired survey. The streamers are flattened at an intermediate depth before the turn to ease operations, and returned to the correct profile on the run-in. It takes approximately five minutes to make these depth adjustments on either side of the turn. Line turns and streamer maintenance are managed in exactly the same way as for conventional acquisition. Currents and obstructions The lines were recorded in a North-South direction (1.5 o North), roughly parallel to the residual current in the central North Sea (associated with the North Sea circulation pattern) which is typically 0.2 m/s towards the south. The tidal current flow in this area is generally southeast and Figure 3 Streamer configuration (not to scale), shown without streamer fanning EAGE

3 first break volume 30, November 2012 special topic northwest, so currents tended to be approximately inline. There were several obstacles in the survey area, requiring the vessel to perform many deviations and close passes, and although there was some reversal of current when shooting close to the Gannet platforms, the feathering was kept under control by the streamer steering systems. No undershooting was performed, resulting in holes in the final coverage in the vicinity of the obstructions. In one area of the survey, near the Anasuria FPSO, echo soundings at -39 m and -42 m were recorded, indicating a possible subsea obstruction. In the absence of any information to confirm or refute these readings, the maximum depth of the streamer was adjusted to 15 m (whilst retaining the variable-depth shape with near offsets at 6 m) for three sail lines in this vicinity. Variable-depth streamer acquisition is not restricted to deep waters as it is possible to vary the streamer shape over a survey to accommodate shallow-water areas, even to the extent that a flat streamer profile may be adopted in certain areas. Processing is seamless for the data recorded at different depths, although in areas where the streamer depth variation is very limited, some of the broadband data starts to resemble deghosted conventional data. Prioritizing data quality In order to increase both survey efficiency and improve the fold of coverage (particularly at longer offsets) and therefore the data quality, a strategy of streamer fanning was employed, where the tails of the 6000 m streamers were separated by m rather than the 75 m at the heads. This technique combats poor far-offset coverage due to variations in feathering, by combining a 25% increase in the spread of the streamer tails with advanced 5D Fourier interpolation during processing. Since higher frequencies are attenuated at longer offsets and greater depths, larger interpolation distances to infill missing traces are generally allowed for longer offsets in processing. Fanning the streamers applies this principle to the acquisition, by increasing the streamer separation towards their tail ends. This leads to better sampling of the mid-to-far offsets and smaller holes in the CMP coverage, allowing standard offset binfold specifications to be met with a drastic reduction in infill, which in turn leads to reduced survey time and HSE exposure, with no degradation of data quality. In tests in the extreme currents of the Gulf of Mexico, it was not possible to see the change between adjacent swaths after processing, where one was acquired using conventional acquisition with 42.8% infill and the other was acquired using streamer fanning and 7.4% infill. Streamer fanning is fully compatible with variable-depth streamer acquisition, as only hydrophones are used, so there is no increase in susceptibility to noise from streamer steering. Infill for the first part of the Quad 21 survey was only 9%, which includes deviation around a weather buoy east of Anasuria and returning to infill with short lines after its removal. This compares with infill of 38% on the nearby Quad 44 survey recorded under similar current conditions, but without fanning or streamer steering. The second part of the survey required even less infill. Passive acoustic monitoring Following changes to environmental protection legislation on 1 October 2011, it became permissible to start seismic lines at night provided passive acoustic monitoring (PAM) devices are deployed and no marine mammals are detected in the area. Previously lines could only be started during daylight hours after trained marine mammal observers (MMOs) had declared the area clear by visual monitoring. The Oceanic Endeavour is equipped with advanced PAM systems and was able to take advantage of this change in legislation to increase productivity while ensuring high environmental standards. Equipping vessels with the latest environmental protection technology not only helps to protect endangered species, but also increases acquisition efficiency and therefore reduces the duration of the survey. Weather-related noise Due to the deep tow of the majority of the receivers with the variable-depth profile, only the nearest offsets are significantly affected by weather-related noise. In practice, such noise is confined to the very near offsets and for this survey, due to the noise-resistant characteristics of the solid streamers, was always recorded at either modest or negligible amplitude. The noise appeared fairly contiguous in the shot domain, but in the receiver domain it was more incoherent and therefore easier to attenuate. Impulsive noise attenuation, applied in the receiver domain, followed by a gentle FK mute was found to successfully attenuate the front-end noise with no impact on the unaffected lines (Figure 4) and with no degradation of the low-frequency signal. Approximately 9% of the sail-lines were identified as having a higher level of swell-related noise, whilst the remaining lines had ranges of a few noisy shots or were unaffected. The mid and far offsets were not affected by weather-related noise on any line. Seismic interference This survey was recorded in an area of intense seismic activity, with seismic interference being recorded from an ocean bottom cable survey in the east and a node survey to the southeast. No time-sharing could be agreed, since the node survey was being recorded using nodes that have a short battery life (15 days), resulting in all three vessels shooting simultaneously. Seismic interference is therefore present on almost all sequences for the first part of the acquisition. An attempt was made to clean the data for QC purposes with moderate success during onboard processing and no line was rejected due to high levels of seismic interference EAGE 93

4 special topic first break volume 30, November 2012 During the processing of the survey, further testing of seismic interference attenuation was performed and an improved result was obtained (see Figure 5). This was achieved using a proprietary algorithm applied in the tau-p domain. This technique detects and subtracts seismic interference noise from shot gathers belonging to a single subsurface line of a 3D marine recording. The method assumes that, as different acquisition crews are not synchronized in terms of shooting interval, either noise arrival times or noise frequency content, or both, are not predictable from one shotpoint to another. This is particularly the case when a small time-window is considered, because noise may only hit specific shot records. This assumption is generally valid unless the interfering noise source continuously radiates. Each subsurface line was transformed into the tau-p domain in the shot direction to create tau-shotpoint-ray parameter cubes. These cubes were then processed in small overlapping time-space windows so that the interference was seen as coherent in the shot direction but random in the ray parameter direction. Using this property, prediction filters were used to detect and remove the noise, after Fourier transformation to the spatial fxy (frequency-shotpoint-ray parameter) domain. Using this method it was possible to attenuate almost all the interference without causing any attenuation of the primary signal. Residual seismic interference remaining after this process was attenuated by conventional prediction filtering in the fxy (frequency-shot-receiver) domain. This technique proved to be extremely successful in attenuating the seismic interference and preserving the full bandwidth of the data. Processing sequence After initial processing of various fast-track datasets, the full survey was processed through the sequence shown in Table 1. This is a standard sequence for the Central North Sea. Only the designature and proprietary deghosting are specific to variable-depth streamer acquisition, although some modifications have been made to other modules such as 3D SRME (Sablon, 2011; Lin, 2011) in order to properly handle the receiver ghost data and make full use of the extended bandwidth. Data examples The deghosting is performed in the image domain and is fully 3D, so we achieve high-resolution images with outstanding detail. The data showed considerable improvement Figure 4 Shots before and after weather-related noise attenuation. The noise is limited to the very near offsets and is attenuated without damaging underlying primaries. Figure 5 Shots before and after seismic interference noise attenuation EAGE

5 first break volume 30, November 2012 special topic Figure 6 Timeslices showing the detail of the polygonal faults and the continuity of deeper horizons around a salt dome in Quad 21. over previous data in the area, especially better continuity of the seismic reflectors through the gas cloud above the salt diapirs and exceptional clarity of the radial faulting (Figure 6). The frequency spectrum is broader everywhere, so that details are better imaged and the continuity around the salt dome is greatly improved. The dramatic reduction of sidelobe interference provided by the broadband wavelet gives much higher resolution, especially of the Paleocene channels, and clearer fault definition, especially in the shallower section where contourites and polygonal faulting are observed (Figures 7 and 8). The low frequencies provide the characteristic broadband appearance to the data which allows sedimentary packages to be clearly identified, as well as providing good penetration below the salt. Thanks to the optimized curved streamer profile, which gets deeper quickly, there are broad bandwidths even in the shallow data in the mute zone, where only the near part of the cable is included in the stack. This allows incredible resolution, for example, the Pliocene iceberg scours which are visible in the data around 500 ms (Figure 9). It could be thought that the ultra-low frequencies obtained by variable-depth streamer acquisition are not real, but some migration artefact or NMO stretch. However analyses of the raw shot records for the Quad 21 survey (with only impulsive noise attenuation applied) clearly show signal being recorded below 3 Hz (Figure 10), demonstrating that we are both emitting and recording these frequencies. The broad bandwidths provide exceptional prestack time migration images for this survey and prestack depth migration is expected to provide even more spectacular results. Table 1 Processing Sequence EAGE 95

6 special topic first break volume 30, November 2012 Figure 7 The broadband texture of the image highlights stratigraphy and provides insight into subtle lithology variations. In this example the Paleocene channels targets are clearly defined. Figure 8 The sharp broadband wavelet with minimal sidelobes provides clear, high-resolution imaging of details such as the polygonal faulting and the contourites. Conclusion and further developments Despite being the first large-scale 3D survey to be acquired using variable-depth streamer acquisition, the programme was successfully completed without any operational difficulties, demonstrating the robustness of the technique. Furthermore, no issues were encountered with controlling the streamers in fan mode with the variable-depth profile in an area with many obstacles, even when they had to be raised to avoid the unidentified obstruction near the Anasuria FPSO. The lines that were recorded with a shallower profile were seamlessly integrated during the proprietary deghosting without affecting image quality, demonstrating the flexibility of the acquisition technique in accommodating variations in water depth. Fanning the streamers allowed the survey to be performed in an efficient manner, with minimal infill being required to achieve the fold specifications at all offsets. The proprietary deghosting algorithm can accommodate not only the variations in the streamer profile during acquisition but also other acquisition geometries altogether. Therefore the holes in the coverage near the platforms could be infilled at a later date, if required, by the use of long-endurance broad-bandwidth nodes to provide data of similar character to be merged with the towed streamer data. The data examples clearly show the benefits that may be obtained by broadband acquisition, in terms of high-resolution definition of fine structures and deep penetration for imaging below complex overburdens. The deghosting is trueamplitude, so the data provide good coverage of the AVO anomalies observed in the area. The improved low-frequency content has been shown (Soubaras and Lafet, 2011; Michel et al., 2012) to provide greater stability and more quantitative results for both post- and prestack seismic inversion. In conventional seismic data, the lack of low frequencies means EAGE

7 first break volume 30, November 2012 special topic Figure 9 Timeslice showing Pliocene iceberg scours at about 500 ms. The optimized curved streamer profile means that the streamer gets deep quickly, providing notch diversity and broad bandwidths even in the mute zone where only near offsets are used for imaging. Figure 10 Shots before and after low pass filtering, with their respective FK spectra, demonstrating that signal with frequencies below 3Hz is being recorded. that an a-priori low-frequency model must be incorporated in the inversion process. With broad-bandwidth, variabledepth streamer data, high-resolution seismic velocities can be used to define the low frequency model in the 0-5 Hz range, whilst the reflectivity provides information from 2.5 Hz. Prestack elastic inversion has already been demonstrated on North Sea BroadSeis data. Whilst variable-depth streamer data can largely be processed using conventional algorithms, new techniques, such as joint noise attenuation, are being developed that capitalize on and make use of the recorded ghost information, rather than simply attenuating it. It is expected that these new algorithms will provide an advantage at key stages in the processing sequence and further improve seismic processing. In terms of tackling the source ghost, a broadband source and source-deghosting algorithm which complements variable-depth streamer acquisition has been successfully tested. This enables us to remove the source 2012 EAGE 97

8 special topic ghost notch as well as the receiver ghost notch, thus eliminating constraints on the highest frequencies that we can record. As we are already recording data down to 2.5 Hz, the removal of the source ghost is much more effective at the high-frequency end of the spectrum than at the low frequencies. As the deghosting is fully 3D in the image domain, the variable-depth streamer broadband solution is implicitly compatible with wide- and full-azimuth acquisition. We have already acquired a proprietary high-density wideazimuth survey in the Gulf of Mexico and are currently acquiring an ultra-long offset, full-azimuth, multi-client programme in the Keathley Canyon using this technique combined with a proprietary staggered vessel configuration. This first commercial 3D variable-depth streamer survey in the North Sea brought a range of acquisition and processing challenges which were successfully addressed. It demonstrated the technique as a robust, efficient, and very effective solution for broadband marine streamer seismic. It also provided the first real insight into the potential of the technique to provide high-resolution images full of geological detail and better deep imaging. The images benefit from a rich broadband texture and a sharp, clean wavelet with minimal sidelobes which enhances the interpretation of potential drilling hazards, faulting, stratigraphy and subtle lithology and fluid effects throughout the section. FairfieldNodal ZLAND Sercel 428 L A N D / T Z / M A R I N E Acknowledgements The authors would like to thank the CGGVeritas Multi- Client & New Ventures division for permission to show this data, Gregor Duval for providing geological input, and crews of the Oceanic Challenger and Oceanic Endeavour, and the CGGVeritas processing crews in Massy and Crawley for their help in the preparation of this article. Source Vessel Aram Aries Cases References Dowle, R. [2006] Solid streamer noise reduction principles. 76 th SEG Annual International Meeting, Expanded Abstracts, 25, Lin, D. et al. [2011] Optimising the processing flow for variable depth streamer data. First Break, 29(9), Michel, L. et al. [2012] Variable Depth Streamer Benefits for Rock Property Inversion. 74th EAGE Conference & Exhibition, Extended Abstracts. Sablon, R. et al. [2011], Multiple attenuation for variable-depth streamer data: from deep to shallow water. 81 st SEG Annual International Meeting, Expanded Extracts. Soubaras, R. [2010] Deghosting by joint deconvolution of a migration and a mirror migration: 80 th SEG Annual International Meeting, Expanded Abstracts, Soubaras, R. and Lafet, Y. [2011] Variable-depth Streamer Acquisition Broadband Data for Imaging and Inversion. 81 st SEG Annual International Meeting, Expanded Extracts. Visit us at SEG - Las Vegas, Nevada 4-9 November Booth # EAGE