Capstone: OBN deep-water imaging in salt

Ocean-bottom nodes (OBN) are the premium acquisition choice for sub-salt exploration in the Gulf of Mexico, offshore Brazil and similar settings. Nodes sit on the sea floor, record the full four-component wavefield and deliver clean low frequencies, down to about 1.3 Hz in the figure, that streamer surveys cannot match. That low-frequency content is what lets FWI start without cycle skipping.

Project setup

A Gulf of Mexico deep-water Miocene target under a salt canopy 2 to 3 km thick. 2 km water depth; nodes at 400 m spacing under a dense 50 m by 50 m shot carpet, so each node records on the order of $10^5$ shots out to about 10 km offset in every azimuth. Target depth: 6 to 8 km below sea level.

The pipeline

On the input data the target, logged at 7000 m in the well, images 780 m too shallow in (b). The flow builds the velocity model the way a salt project does: multiples out, a tomographic start, FWI from the lowest usable frequency, a salt flood, a base-salt pick and an anisotropic update. Started at 3.0 Hz, FWI fails the half-period test: the predicted far-offset error of 235 ms is more than $T/2$ = 167 ms. Started at 2.0 Hz it passes; the salt picked in the isotropic model still sits too deep, so the TTI update re-picks top and base salt in the anisotropic model, and only then does the target tie the well.

Why this project needs the full physics-based toolkit

The salt canopy has $V\_{\\mathrm{P}}\\approx 4500$ m/s against sedimentary $V\_{\\mathrm{P}}\\approx 2500$ m/s, a contrast of about 1.8 that bends rays severely at the salt boundary; beyond the critical angle, \\theta\_c=\\arcsin(2500/4500)\\approx 34^\\circ, no P wave is transmitted into the salt at all. Multiple arrivals, turning waves and prism reflections at the salt flanks all contribute. Kirchhoff PSDM with a single-arrival traveltime table cannot represent this multi-pathing. One-way wave-equation migration cannot propagate turning waves or dips steeper than about 70°. Two-way RTM handles all of them when the salt boundary is in the velocity model.

The sediments around and above the salt, especially the shales, are anisotropic, with Thomsen parameters $\\epsilon\\approx 0.05$ to 0.15 and $\\delta\\approx 0.02$ to 0.08 in typical Gulf sections; the salt itself is nearly isotropic. Migrating with an isotropic model mislocates sub-salt reflectors by hundreds of metres, so anisotropic (TTI) RTM is standard, with TTI FWI updating velocity and anisotropy together. That the salt is close to isotropic is one reason velocity-model building focuses on the surrounding sediments.

Low-frequency FWI is the control variable

FWI at 1.5 to 3 Hz is what makes this project feasible. Streamer data usable from about 4.4 Hz, as in the figure, are fine for imaging away from salt but cycle-skip through it. The extra 1.5 to 2 octaves of usable low frequency that OBN delivers (down to about 1.3 Hz in the figure instead of 4.4 Hz, $\\log\_2(4.4/1.3)\\approx 1.8$) is what lets FWI converge on the sub-salt model: the starting model's traveltime error only has to stay below half a period, $1/(2f\_{\\min})$, which is 333 ms at 1.5 Hz against 113 ms at 4.4 Hz. The exercise Record on a streamer shows the difference in the figure.

Cost profile

• Acquisition: US$50 to 150 million per 4000 km² (node deployment cost dominates).
• FWI processing: 3 to 6 months on a 200-GPU cluster.
• RTM: 2 to 4 months on the same cluster, run in the common-receiver domain by reciprocity (each node gather migrated as one shot).
• Human interpretation and QI: 6 to 12 months after imaging.

Total project cycle 18 to 30 months. The reward is a new play in a mature basin: a single successful sub-salt discovery can justify a decade of sub-salt exploration investment.

Where this goes next

Section 10.3 covers a different deep-water scenario: a wide-azimuth marine streamer survey whose primary deliverable is a high-resolution velocity model for exploration mapping, not a specific sub-salt target.

References

• Yilmaz, Ö. (2001). Seismic Data Analysis (2 vols.). SEG.
• Etgen, J., Gray, S. H., Zhang, Y. (2009). An overview of depth imaging in exploration geophysics. Geophysics, 74, WCA5.
• Baysal, E., Kosloff, D. D., Sherwood, J. W. C. (1983). Reverse time migration. Geophysics, 48, 1514.
• Virieux, J., Operto, S. (2009). An overview of full-waveform inversion in exploration geophysics. Geophysics, 74, WCC1.