The Salt Problem
Salt giveth and salt taketh away: it builds the traps that hold supergiant fields, then wraps them in a 4.5-kilometer-per-second refracting blanket that defeats ordinary seismic. This is the arms race in one path, the tectonics that shape the diapirs, the physics of why imaging fails beneath them, and the escalation that wins: depth migration, RTM, FWI, wide azimuths, and nodes on the seabed.
You can read diapirs, walls, canopies, and welds and say how each traps, explain exactly why the base of salt destroys a Kirchhoff image, choose RTM when the wavefield turns over, say what FWI adds to a salt model and where it stalls, justify wide-azimuth and ocean-bottom geometries by illumination rather than fashion, and read the Thunder Horse and pre-salt case files as the history of the problem being solved.
The rock that flows
Salt tectonics is one of the five great structural styles, and its habit of moving after deposition makes its traps, and its imaging problems, unlike anything brittle faulting builds.
The salt body's geometry is the trap map: flanks seal, welds leak, overhangs hide entire fields; reading the shapes is the interpretive skill the whole basin play rests on.
Why imaging breaks
Stacking assumes gentle structure and simple moveout; a salt flank violates both before breakfast. The failure of the simple machinery is the argument for everything that follows.
Depth migration through a velocity model is the first real answer to salt, and Kirchhoff's ray assumptions are both its speed and its undoing where the wavefield multipathes.
RTM runs the full two-way wave equation and does not care how many times the wavefield turns over; steep salt flanks and overhangs that Kirchhoff cannot dream of, RTM images as a matter of course.
The image is only as good as the salt geometry in the velocity model, and FWI, fed low frequencies and long offsets, sculpts that geometry from the data itself instead of an interpreter's guess.
SEAM and its siblings exist because of salt: community models with the answer known, where imaging algorithms are scored honestly before anyone bets a well on them.
Acquire around it
One boat dragging one azimuth cannot see under an overhang; WAZ, RAZ, and FAZ geometries exist because illumination under salt is an acquisition property before it is a processing one.
Nodes decouple receivers from the boat entirely: full azimuth, long offsets, and quiet data over the exact salt flank streamers cannot serve, at a price the prize must justify.
Migration aperture and illumination maps decide whether the target under the overhang receives energy at all; no algorithm images what the geometry never lit.
The Brazilian pre-salt is the full curriculum in one survey: nodes, full azimuth, FWI velocity, RTM imaging, a carbonate target under two kilometers of layered evaporites, acquisition design with everything at stake.
The trophies
The Gulf of Mexico giant that dragged subsalt imaging from Kirchhoff to FWI across two decades; every reprocessing vintage is a chapter in the arms race, read in one case file.
Tens of billions of barrels under a salt blanket that ordinary surveys could not pierce: the pre-salt is the existence proof that the whole escalation, nodes to FWI to RTM, pays for itself.