Exploration Report

Operators’ choices for 4C seafloor acquisition are growing

Advances with autonomous nodes and OBC systems are adding to operator’s choices for acquiring multicomponent data.

Perry A. Fischer, Editor

Multicomponent seafloor data has been difficult and expensive to acquire and process. Over the last five years or so, there has been steady progress on both fronts. While the best processing schemes will continue to evolve, the acquisition logistics and costs are an equal, if not a greater, challenge, especially in deep water.

The two methods for acquiring 4C data are autonomous nodes and Ocean Bottom Cable systems. Typically, seafloor seismic in the North Sea has been limited to water depths of less than 400 m. Surveys elsewhere have usually been in water depths less than 150 m. Although there have been OBC systems with a 2,000-m depth rating on the market for years, the cost and logistics of actually acquiring multicomponent data at that depth appeared to be a barrier, which meant that the 2,000-m rating was never commercially tested – until now. Recently, a commercial record for ultra-deepwater OBC was set, tripling the previous commercial record.

RONCADOR RECORD BREAKER

The survey took place offshore Brazil in Campos basin. PGS was the contractor, working for Petrobras over the giant Roncador field. The company set a new industry record when the seafloor seismic crew completed a 4C/3D survey over a 45-sq-km area at water depths around 1,860 m (6,100 ft). The company used its FOURcE system, a 3-axis, gimbaled OBC system.

Topography was challenging, requiring the acquisition team to negotiate slopes and a ridge, Fig. 1. The data (sensor) density of this type of acquisition can be high, especially inline, but it’s not cheap. In shallow water, the positioning logistics and costs are much less. But this job required three vessels and relatively new positioning measures.

Fig. 1. Bathymetry for a large portion of the Roncador 4C/3D survey. The green shading indicates water depths deeper than 1,800 m.

This OBC acquisition comprised a source vessel – the Falcon Explorer, a recorder – the Ocean Explorer, and a cable-handling vessel – the Bergen Surveyor, which was furnished with 36 km of in-water equipment. Before beginning this project, the majority of the cables in inventory, many of them three years old, were replaced with new ones. Also, a specially designed deepwater cable-handling system was installed on the Bergen Surveyor.

The 6-km cables were deployed on the sea surface, while moving, from the stern of the Bergen Surveyor, made more difficult by strong and variable currents in the water column. Positioning during deployment was accomplished using a Kongsberg-Simrad HIPAP USBL system ranging into transponders located on the cables at every 12th ground station (300-m intervals). The transponders were attached to the cable with 1-m long tethers and housed in floats to ensure orientation to the surface. This system enabled cable placement to be consistently within 14 m of pre-plot location (less than 1% of water depth) without the added costs of ROVs. Final positions were generated for all ground stations using first-break pick times derived from production data. The ability to repeat receiver positions is an important consideration should Petrobras decide to use this survey as a baseline for future time-lapse seismic.

Avoidance of infrastructure can be one of the prime reasons for choosing the method of seafloor acquisition, but in the Petrobras shoot, it was more about the operator’s need for high density, multicomponent data.

Before this survey was undertaken, a 4C feasibility study was conducted for Roncador and Albacora fields, to determine whether 4C data would help in identifying hydrocarbon-bearing sands. According to Santi Randazzo, manager of multicomponent processing and interpretation for PGS, “Well log data with dipole sonics, together with P-wave streamer data, indicated that C-waves would provide a better tool than P-wave AVO analysis in identifying Class II sands, and in detecting false AVO anomalies.”

With the feasibility study indicating probable success, Petrobras decided to proceed with the 45-sq-km acquisition. Whether this will become a baseline for repeat surveys, that is, the world’s first deepwater OBC 4D, remains to be seen. Overall, this effort is part of a 4C Test Program under Petrobras Research Project PRAVAP 19 (Seismic Improved Oil Recovery Program). This program aims to acquire the more comprehensive 4C data in areas where conventional P-wave seismic are considered insufficient. Besides Roncador, 2D/4C surveys were also conducted in Violoa, Cachalote/ Jubarte, and Albacora. This was the first wide-scale application of multicomponent technology to offshore Brazilian oilfields.

Survey objectives varied depending upon the particular area. In Roncador, the survey targeted relatively deep sands (3,000 – 3,700 m, ~11,000+ ft). Some of these reservoirs are transparent to conventional (stacked) P-wave data, making them difficult to map, use AVO techniques on or perform elastic inversion. Petrobras also wanted to validate several oil-water contacts.

This OBC ultra-deepwater survey could change the acquisition landscape for companies wrestling with the difficult decision of whether and how to acquire multicomponent data. That is, how best to determine feasibility, including the value, for multicomponent data acquisition, which immediately leads to whether to use nodes or cables in deep and ultradeep water. Data quality will be a huge determining factor, and one that will be closely watched as the industry moves toward common solutions. A report on the Roncador shoot with more details will be presented at the upcoming SEG conference.

NODE ACQUISITION

Autonomous nodes are independent 4C seismometers that are deployed on the seafloor for a matter of days to months. They comprise a 3-axis geophone, a hydrophone, tilt meter, and are battery operated with precision clocks that mark the time, so events from each node can be tied together. Data is stored on hard drives or solid-state memory. Nodes may be – and usually are – placed much farther apart than sensors on OBC systems. The much sparser node density allows for improved economics and wide-azimuth acquisition, while the sparse data density can be enhanced, to a large extent, by a denser shooting pattern. In ultradeep water, certainly below 2,000 m, it may be the only viable method for getting multicomponent data.

The previous world record for node acquisition, in terms of area, was the work done by Seabed for Pemex. That survey planted 1,600 receivers, deployed in seven, 232-node swaths in a 400 x 400-m grid, for a total survey covering 220 sq km. To achieve wide azimuths, 520 sq km of sea surface was shot. This allowed offsets up to 8,000 m with regular fold coverage of over 100. The shooting area was nearly 160 sq km for each swath. Sail line interval was 150 m, with dual source shooting flip-flop every 50 m.²

The drivers in the Pemex case were the need for multicomponent data combined with the avoidance of dense seafloor infrastructure. Thus, nodes were employed even though it was a shallow-water job, and each one was positioned by an ROV with crane support.

BP has been active with some small, deepwater-acquisition experiments using nodes. The most recent was a six-node test over the Atlantis field in the Gulf of Mexico, to examine data quality and feasibility issues. The trial used an ROV to plant the six nodes along a 300-m (990-ft) straight line. Spacing varied, with two nodes placed just 3-m apart.² The shooting pattern and a sample of the data are shown in Fig. 2.

Fig. 2. Wagon wheel shooting pattern (top) used over Atlantis trial. Common receiver gather (bottom) for each component of 4C data for a single shooting line.1

One of the nodes was lifted and replanted, with one of the sail lines being shot three times, to investigate any effects any such movement/ replanting might have on repeatability. The effect of having the ROV push the node deeper into the seafloor was also investigated, to see whether seafloor coupling was improved, which was indeed the case. Other experiments looked into the effect of planting nodes without regard to orientation and inclination.

There was no variation in unit orientation vs. azimuth or, at least, it was small enough that vector fidelity, while always a concern, was not an issue. In all respects, including orientation, fidelity, repeatability and coupling, the trial succeeded, sufficient for BP to order 900 of Fairfield’s Z-3000 nodes (3,000-m rated), which are being custom-built for the job. These nodes are designed for continuous 4C sampling at 2 ms intervals over a 28-day period. Batteries contribute significantly to its 200-lb weight.

OTHER DEVELOPMENTS

This experiment has led to what will be the largest commercial, deepwater multicomponent acquisition: a 150-sq-km node survey by Fairfield over BP’s Atlantis field, at water depths over 1,250 m (4,100 ft). The job should take about four months, and is slated to begin next month.

This article is limited to retrievable systems and, thus, will not include discussion of the permanent 4C cable installations at BP’s Valhall field in the North Sea and the pilot at Shell’s Mars platform in the Gulf of Mexico. However, Multiwave Geophysical will conduct the world’s first, quasi-permanent 4C seabed acquisition.

The survey is for BP Indonesia and will be carried out by Elnusa Geosains, which is Multiwave’s Indonesian partner company. The unique shallow-water 4C survey will bury the OBC cables in trenches to reduce the effect of current-induced noise on the seafloor sensors. The seafloor in the survey area is relatively soft clay and sand, and should allow burial of 40 km of OBC, followed by retrieval, relocating, and re-burial. The survey was scheduled to begin as this is was written. The company recently acquired a shallow-water 4C/3D survey equal in area to the Pemex record discussed above. The OBC survey took place offshore Qatar in ISND and ISSD fields. Results have not been released.

PROCESSING

Processing OBC data to optimize its value to operators is far from standardized, and is evolving quickly enough that conclusions reached a few years ago may not be relevant today. Further, the desire to process PS converted waves presents an added level of complexity that must be understood and addressed in processing. OBS acquisition geometry is more flexible than streamer acquisition, because the source can move independent of the seafloor receivers. However, one consequence of this and other differences is that they are not well suited to many conventional towed-streamer processing methodologies.³

An OBC survey was shot offshore Trinidad by WesternGeco in 2000. It was processed in 2000 – 2001, and then reprocessed in 2004, in both cases by WesternGeco. The interpreters’ view of the results in 2001 was that the PS data was no better than the existing towed streamer data and did not succeed in imaging through the gas-obstructed zone. Equally disappointing was the fact that the PZ data did not look as good as the towed-streamer data outside of the gas-affected zone. The question was asked, “Could the processing scheme be improved?”

BP’s experience in 4C seafloor data elsewhere led it, some three years later, to reprocess the Trinidad data. The results showed some improvement, mostly through better pre-processing. In addition, by then, pre-stack time migration had become available for both PZ and PS data. Reprocessing paid particular attention to signal retention in bandwidth, necessarily richer in the lower frequencies due to the ocean bottom setting.³

There was clear improvement via the more sophisticated P-wave reprocessing. However, the PS-wave reprocessing did not bring the same level of improvement. There was improvement in the 2004 PZ OBC data showing that signal was getting through the gas obstruction, Fig. 3. The results of the PS reprocessing were less impressive. The gas affected zone caused the same amount of disruption in both 2004 and 2001 results, suggesting a very shallow cause that is not being addressed by statics corrections.³ These observations suggest that OBS processing algorithms are not yet as well understood as those used on conventional seismic data, but are a key element that is steadily improving.

Fig. 3. (a) PZ in-line from 2000 OBC processing; (b) 2004 reprocessed PZ in-line. (c) PZ x-line (at red arrow) from 2000 processing; (d) 2004 reprocessed PZ x-line (at red arrow).3

CONCLUSION

So, is the industry about to rapidly acquire multicomponent seafloor data in large quantities? Not quite. It is still an emerging value proposition. But in certain deepwater reservoirs, operators might have little choice. Considering the level of expense involved, together with the variability of reservoir, seafloor topography, equipment, and deployment and retrieval schemes, industry is doing a good job in taking up the R&D banner in seafloor acquisition. But it’s still likely to be viewed as too slow from the contractors’ viewpoint. Ultimately, the stakeholders on both sides – those who acquire data and those who use it – will decide which technologies move forward.

In general, the only way to answer the difficult questions of “when, where and how” to acquire, process and use this data is for operators such as BP, Pemex, Shell and others to share their experiences with the public, especially with data over non-competitive fields. Paraphrasing the Chinese adage, “We live in interesting seismic times.” 

 LITERATURE CITED

 ¹  Garcia, M. V., et al., “Imaging Sihil with full azimuth ocean bottom 4C node data,” EAGE 67th Conference & Exhibition, Madrid, Spain, June 13 – 16, 2005.

 2  Docherty, P, et al., “Multicomponent ocean bottom seismic data acquired with an autonomous node system,” EAGE 67th Conference & Exhibition, Madrid, Spain, June 13 – 16, 2005.

 3  Johnston, R., et al., “The maturity of ocean-bottom seismic: An example from Trinidad, ”EAGE 67th Conference & Exhibition, Madrid, Spain, June 13 – 16, 2005.