Marine exploration in overdrive

In 1999 I had the pleasure of organizing a joint SEG/SPE conference in Kananaskis with Dan Ebrom of BP on time-lapse measurements for reservoir monitoring. To provide context, the oil price was USUS$18/bbl, modestly up on US$12/bbl from the previous year, largely due a faster-than-trend rise of 1.6 percent in global consumption coupled with a 5.4 percent fall in OPEC production to 29.3 MMbbl/day. Mexico and Norway both announced cuts in output to reinforce the OPEC action and, in the US, output also fell by 3.8 percent. Global oil production was around 72 MMbbl/day and the reserves/replacement ratio was 41 years. BP had launched the round of mega-mergers with Amoco and subsequently appeared to be going through the alphabet of competitors, merging with Arco, Aral, Burmah, and Castrol. With the consolidation of exploration portfolios, BP had confidently announced that it saw no need to explore for a decade (possibly unintentionally prophetic) and that the industry emphasis was on using seismic methods to squeeze more oil out of existing reserves with reservoir-focused technologies such as 4D seismic surveys. Hence, the Kananaskis meeting in which, during a discussion on the feasibility of carrying out 4D surveys with legacy exploration seismic data as a baseline, one contributor remarked that legacy seismic data could be defined as those data for which the navigation tapes had been lost.

Today, the oil price is hovering around US$135/bbl, global production is around 86 MMbbl/day, the reserves/replacement ratio has fallen marginally to 40.5 years, the Chinese and Indian economies accelerate, and analysts debate whether Hubbert’s peak is yesterday, today, or tomorrow. Climate change, carbon sequestration, and the switch to renewables dominate the agenda, but the lead-time for non-hydrocarbon energy sources is such that oil will continue to dominate the energy mix for quite a while. So, what of seismic exploration in the present context? Should I still be looking for those lost navigation tapes? The economic crystal ball is so clouded by the sub-prime uncertainty that prediction is foolhardy – we might expect that an ongoing firm oil price will continue to drive strongly the quest for fresh reserves from the few remaining unexplored basins as in the Arctic, from tough targets such as sub-salt and sub-basalt, from unconventional reserves as in heavy oil, tar sands, oil shale, and hydrates, and from deep gas reservoirs. But, the need to maximize recovery factor of existing reserves remains and will return to importance, so not only should we be looking for lost navigation tapes, but we might also consider future-proofing today’s exploration seismic data for tomorrow’s 4D baseline, whether for production enhancement or for carbon storage monitoring.

Access to acreage is a key issue for the international industry, pushing companies into inhospitable Arctic waters and deepwater sub-salt/sub-basalt plays in the Gulf of Mexico, West Africa, and offshore South America. With exploration wells in the Gulf of Mexico costing up to US$100m and a deepwater exploration success rate of 11 percent in 2006, minimization of exploration risk is critical. The Gulf of Mexico has led the introduction of multi-azimuth, wide-azimuth, and rich-azimuth surveys in which illumination from as many angles as possible is being used to throw (seismic) light into the dark corners below high-velocity-contrast, irregular salt bodies. Careful design ensures that azimuth/offset distributions are as uniform as possible to minimize acquisition footprint, and the resulting surveys, which are acquired with two or more vessels, show significant benefit in sub-salt imaging compared to narrow-azimuth, single-vessel acquisition. Not only is the primary, sub-salt reflectivity imaged better and more consistently from the higher fold and improved illumination, but the multiple problem is also noticeably less severe in wide-azimuth surveys. While multiple source vessels shooting into a rectangular streamer array will permit a rich azimuthal raypath distribution, an intriguing alternative is to acquire data on wide turns with long streamers, which can also improve azimuthal coverage with better efficiency.

A development that has brought benefit to sub-basalt imaging on the Atlantic margin, where scattering loss from rough, high-impedance-contrast basalt flows, associated multiples, mode conversions and strong geometrical spreading makes imaging of deeper targets problematic, is the use of low-frequency seismic acquisition. Deep-towed streamers, long-offsets, and large sources optimized for low frequencies can generate and record data to penetrate and improve imaging below top basalt; however, a single, deep-towed streamer limits access to high-frequency bandwidth for the post-basalt sequence because of the low ghost notch frequency. Recently, dynamically-positioned streamers have enabled the use of vertically separated twin-streamer combinations, whereby each streamer’s ghost notch is filled by data from the other streamer. The resulting deghosted data have a broader bandwidth at both high and low frequencies compared to a single streamer and this over/under technique, complemented with vertically displaced gun arrays to deal with the source ghost, has been employed with good results in both the basalt-covered Atlantic margin and the sub-salt targets of the Gulf of Mexico and the Barents Sea.

While some 3D over/under projects have been completed, acquisition efficiency is reduced by the twin streamer combination, motivating a search for a single-streamer approach. Initial results from 2D trials of a twin-sensor streamer, similar to a towed ocean-bottom cable (OBC) and comprising hydrophone-geophone pairs, were presented at SEG 2007. Using data processing analogous to hydrophone-geophone summation in OBC data, the twin-sensor streamer data were deghosted from about 20 Hz upwards, with good uplift for the high-frequency bandwidth data. In contrast to twin-streamer data, frequencies below 20 Hz relied on the deep-towed hydrophone sensors alone to optimize the signal-to-noise ratio because cable-borne noise affected the geophone signals increasingly at lower frequencies.

Although this summary deals primarily with marine exploration, it should not go unnoted that land seismic technology is also addressing similar issues of broader bandwidth, higher fidelity, and acquisition efficiency for cost-effective, deeper penetration and subsurface characterization. Vibrators with greater output, improved linearity, and lower frequencies are being deployed to address deep clastic oil and gas reservoirs in the Middle East that lie beneath thick, high-velocity, multiple-generating carbonate sequences. Imaginative approaches are being deployed for flexible, adaptive array forming to attenuate ground roll, and ways to reduce the acquisition cycle time for improved productivity are always needed.

Continuing the theme of de-risking targets in areas where the seismic imaging is challenged, several companies have experimented extensively with other remote-sensing techniques including four-component ocean-bottom seismometers for ultra-long offsets, better velocity resolution, shear-wave analysis and base-basalt estimation, and gravity or gradio-gravimetry for better velocity modeling and location of salt flanks and base salt. However, the current remote-sensing favorite is undoubtedly electromagnetic (EM) methods of one form or another. Marine magneto-tellurics (MMT) uses the natural fluctuations in the Earth’s EM field, caused by the solar wind and thunderstorms, to estimate the frequency-dependent effective impedance of the Earth over wide bandwidths of EM data recorded by autonomous sensors on the seabed. MMT can detect conductive sediments below resistive basalts or salt bodies and help to determine the depths of their bases. Used alone, the resolution is usually insufficient for reservoir-scale studies, but in combination with seismic data, it can both improve velocity model building for prestack depth migration and even help to high-grade drilling opportunities through combined inversion.

MMT has a natural dead band, which happens to be around 0.1-10 Hz, that would be needed for reservoir studies, and so controlled-source source electromagnetic (CSEM) techniques have been developed that use receivers similar to those in MMT surveys to record EM fields from a deep-towed dipole current source transmitting in this dead band. In several published case studies, amplitude analysis of the received signals has succeeded in identifying thin, resistive bodies (hydrocarbon reservoirs) embedded in conductive bodies (shales), thereby offering an alternative to seismic amplitude variation with offset as a way to de-risk target structures.

CSEM surveys transmit a continuous source signal and analysis is carried out in the frequency domain but, as their name implies, transient electromagnetic (TEM) surveys use a transient, stepped current to record and subsequently analyze EM data in the time domain. The method can be used both on land and offshore and has the benefit that, by appropriate selection of the acquisition parameters, marine surveys can be carried out in shallow water depths by subtracting an estimate of the air wave, which directly couples source and receivers, but has no subsurface information. The airwave amplitude increases with decreasing water depth and decays more slowly with source/receiver offset than the ground-coupled field component. At long offset, the TEM signal is dominated by the airwave, allowing it to be estimated and subtracted from signals at shorter, reservoir-targeted offsets. The CSEM method works best in deep water, which strongly attenuates the airwave, but more sophisticated inversion techniques are being developed to invert the complete CSEM signal, including the airwave, in shallower water depths as well.

To deal with tough targets, a more integrated approach is being taken to allow simultaneous joint inversion of short- and long-offset seismic data, wide-azimuths, and combinations of gravity and EM data to improve definition of the base and edges of salt (and basalt) bodies and to image better the sediments below. Multi-data synthetic models have been developed for testing inversion algorithms and real data case studies are now being presented. No doubt EM data, integrated with seismic inversions for improved structural imaging and subsurface characterization, will continue to be featured strongly in future EAGE and SEG meetings.

To characterize the acquisition parameters, modern exploration seismic methods acquire a volume of auxiliary data that is as large as the seismic data volume itself. These GPS time-stamped auxiliary data include the 3D position of each sensing element, each source element and their depths of immersion; the near-field signature from each gun element and the real-time computation of the far-field source signature; water depths at the source point, tidal information, current information, streamer feathering and tension, water velocity, and many more. In effect, the seismic data are qualified and supported by an amazing variety of other pieces of information, all of which contribute to our knowledge of the survey and all of which are useful to establish a reference for future 4D surveys for reservoir monitoring. So, do we need to archive the navigation data of our exploration surveys? The answer is an emphatic “yes”; not only the navigation data, but also the rest of the data that go with the acquisition, as there will surely come a time when we will need to reconstruct each and every change that has occurred in the subsurface to guide our understanding of reservoir performance, either for hydrocarbon production or for carbon storage. Exciting times are with us and, as an industry, we must respond to the triple challenges of finding oil, maximizing its recovery, and minimizing the climate impact of its combustion products; otherwise, future generations will

Philip Christie
An Oxford physics graduate, Christie joined Schlumberger in 1972 as a wireline engineer in West Africa. After three years in operations, he undertook a PhD at Cambridge University. Following a post-doc in high-resolution surface and in-seam seismics for coal mapping, he returned to Schlumberger in 1981 as the borehole geophysicist for their European Business Unit. From 1985 to 1996, he launched a borehole seismic department in Schlumberger’s technology centre in France; directed the Geoacoustics department at Schlumberger-Doll Research in Connecticut, and established the Seismic department at Schlumberger Cambridge Research. From 1996 to 1997, Christie worked in BP’s Atlantic Margin group, jointly co-ordinating the Foinaven reservoir monitoring experiment with BP, Shell and WesternGeco.

After leading WesternGeco’s Reservoir Geophysics group in Gatwick, supporting multi-component and time-lapse seismics, Phil returned to Schlumberger Cambridge in October 2000 as a scientific advisor and a leader of Schlumberger’s Geophysics Technical Community. Christie is currently VP of the EAGE, a Co-Editor of Petroleum Geoscience, sits on NERC and UK University committees and is an active member of Society of Exploration Engineers (SEG).