layer stripping, Oil & Gas Glossary

Layer stripping is a seismic inversion strategy that solves for the subsurface one rock layer at a time, working from the shallowest interval downward, so that each layer is determined only after every layer above it has already been resolved and its effect removed from the data. The logic follows from the physics of reflection seismology: energy returning from a deep reflector has travelled twice through all the shallower rock, and its arrival time, amplitude, and waveform carry the imprint of that overburden. If the shallow section is solved first, its known velocity and structure can be stripped from the recorded wavefield, leaving residual data that respond mainly to the next layer down, which is then inverted in turn. The method appears in several guises. In refraction and first-arrival tomography, the top layer is determined from short-offset arrivals, then its delay is subtracted so the next layer can be estimated from longer offsets, building a near-surface velocity model step by step. In reflection-based velocity analysis and pre-stack depth migration, the macro-velocity model is updated top to bottom because deep reflections cannot be correctly positioned until the overburden velocity is right. In full-waveform inversion, layer stripping is imposed by progressively including deeper, later-arriving energy and by replacing the upper data with synthetics from the already-inverted shallow model, focusing each stage on the misfit from the deeper part. The appeal is stability: a global inversion of all layers at once is highly non-unique, because a velocity error in a shallow layer trades off against errors deeper down, and solving sequentially breaks that coupling. The weakness is error propagation, since any mistake in a shallow layer is carried into and amplified within every layer beneath it, so disciplined quality control at each step is essential. In the Western Canadian Sedimentary Basin, layer stripping underpins depth imaging of structurally complex Foothills thrust belts and of stacked plays such as the Montney, Duvernay, and Mannville, where a reliable near-surface and overburden model is the prerequisite for placing deep targets correctly. It connects directly to broader concepts of seismic inversion, velocity model building, and the migrated image that ultimately guides where a multimillion-dollar horizontal well is landed.

Key Takeaways

Top-down sequential solution: Layer stripping resolves the velocity and structure of each layer only after all shallower layers are known, then removes their effect from the data before tackling the next interval. This mirrors the physics that deep reflections carry the imprint of every overburden layer they pass through twice, so the overburden must be solved first.
Breaks inversion non-uniqueness: Inverting all layers simultaneously is unstable because a shallow velocity error trades off against deeper errors. Solving one layer at a time decouples these unknowns, giving a more stable and tractable problem. This is why top-to-bottom velocity updating is standard in pre-stack depth migration workflows for complex WCSB structures.
Appears in multiple methods: The same principle drives near-surface refraction and first-arrival tomography, reflection velocity analysis for depth migration, and layer-stripping full-waveform inversion that progressively adds deeper, later-arriving energy. In FWI, upper data are replaced with synthetics from the inverted shallow model so each stage minimizes only the deeper misfit.
Error propagation is the main risk: Because each layer builds on the one above, any error in a shallow interval is inherited and magnified in all deeper layers. Rigorous well ties, residual moveout checks, and image-gather flatness at every step are required to keep shallow mistakes from corrupting the deep target, where the economic payoff lives.
Critical for WCSB depth imaging: In Foothills thrust belts and stacked Montney-Duvernay-Mannville sections, a correct near-surface and overburden model is the prerequisite for positioning deep reflectors. Layer stripping builds that model in stable steps, and the resulting depth image directly controls horizontal well landing depth, where being off by even 10 m can miss the target bench.

Near-Surface Model Building from First Arrivals

The most common layer-stripping workflow in land seismic builds the near-surface velocity model from refracted first arrivals. Short-offset arrivals sample only the shallowest layer, so its velocity is solved first; that layer's traveltime delay is then subtracted from the data, and the residual first arrivals, now responding to the second layer, give its velocity. Repeating downward produces a layered near-surface model used for statics correction and as the starting velocity for migration. In the WCSB this matters because variable glacial till, muskeg, and weathering create large lateral velocity changes near surface that, if unresolved, smear deeper reflectors. A well-built layer-stripped near-surface model removes those distortions before the deeper Montney or Foothills targets are imaged.

Overburden Velocity and Deep Target Positioning

In pre-stack depth migration the macro-velocity model must be accurate from the surface down, because migration repositions each reflection using the cumulative velocity along its raypath. Layer stripping updates this model top to bottom: shallow velocities are refined until shallow image gathers are flat, that interval is fixed, and the next deeper interval is updated against its own residual moveout. Only when the full overburden is correct can a deep Duvernay or sub-thrust reflector be placed at true depth. Skipping a shallow correction and forcing a deep fix instead produces a velocity model that flattens gathers for the wrong reasons and mispositions the target, a costly error when a horizontal well is landed off the imaged depth.

Fast Facts

The layer-stripping idea is far older than modern depth imaging and traces to the foundations of seismology itself. The same top-down logic appears in the Herglotz-Wiechert inversion of the early 1900s, which recovered Earth's velocity-depth structure from earthquake traveltimes by progressively peeling away resolved shells. A century later the principle drives terabyte-scale full-waveform inversion on Montney and Foothills data, where supercomputers strip the overburden in stages, proof that a simple sequential strategy conceived for global seismology still anchors the most advanced exploration imaging.

Layer stripping is one approach within the broader field of seismic inversion, which converts recorded amplitudes into rock and velocity properties. It depends on building a reliable velocity model, the layered field of seismic speeds it constructs top to bottom, and that model feeds migration, the step that uses overburden velocity to place reflections at their true subsurface location. The finished product is a depth image whose accuracy at the target depends on every shallow layer having been stripped correctly first.

Real-World WCSB Scenario: Depth Imaging a Foothills Thrust Sheet

An operator targeting a sub-thrust gas play in the Alberta Foothills faced a classic imaging problem: steep dips, large near-surface velocity contrasts from outcropping carbonates, and stacked thrust sheets that scattered energy. An initial time-migrated volume mispositioned the target reflector and an exploration well drilled on it came in structurally low. The team reprocessed with layer-stripping pre-stack depth migration, solving the near-surface from first arrivals, then updating each thrust sheet's velocity in sequence, at a reprocessing cost near 1.2 million CAD.

The rebuilt depth image moved the target structure several hundred metres laterally and corrected its depth by tens of metres. The next well, located on the revised image, landed in the intended reservoir and tested commercial gas, turning a play that one mistie had nearly condemned into a producing pool and validating the sequential, top-down velocity discipline.