Slant Stack: The Tau-P Transform, Plane-Wave Decomposition, and Multiple Suppression in Seismic Processing
A slant stack is a seismic-processing operation that sums, or stacks, seismic traces after shifting each one in time by an amount proportional to its source-to-receiver offset. The result transforms data from the conventional offset-versus-time domain into the tau-p domain, where p is the ray parameter or horizontal slowness, the inverse of horizontal phase velocity, and tau is the intercept time at zero offset. Because the operation maps a family of events with a common slope onto a single point, the slant stack is mathematically the linear Radon transform, and it performs a plane-wave decomposition: it breaks a wavefield made of spherically spreading energy into the set of plane waves that compose it, each plane wave indexed by its slowness p. The technique is valuable precisely in settings where reflectors are not flat. By applying a linear time shift and summing along sloping trajectories, a slant stack concentrates dipping, linear seismic events into compact, well-separated features that are easy to isolate, mute, or model. This makes it a workhorse for several processing tasks. In multiple suppression, primaries and multiples that overlap in the offset-time gather separate cleanly in tau-p because they have different moveout, so a high-resolution Radon demultiple can attenuate water-bottom and intrabed multiples that ordinary stacking cannot remove. The transform also underlies tau-p deconvolution, dip filtering, trace interpolation to fill acquisition gaps, refraction analysis using the linear moveout of head waves, and velocity analysis. In the Western Canadian Sedimentary Basin the slant stack earns its keep in the structurally complex Alberta Foothills, where Mesozoic and Paleozoic strata are folded and thrust into steeply dipping panels above fault-propagation folds; the strong dips and crossing events in foothills data are exactly the conditions for which tau-p processing was designed. It is equally useful in the resource plays of the basin, where suppressing multiples and interpolating sparse 3D geometries improves the imaging of subtle Montney and Duvernay targets, and where clean plane-wave gathers feed amplitude-variation-with-offset work and prestack migration.
Key Takeaways
- It is the linear Radon transform: A slant stack applies a time shift proportional to offset and sums along that slope, which is the definition of the linear Radon, or tau-p, transform. The offset axis becomes the ray-parameter (slowness) axis p, and the time axis becomes the zero-offset intercept time tau, decomposing a spherical wavefield into its component plane waves.
- Built for dipping reflectors: Because the operation stacks energy along linear trajectories of varying slope, dipping and linear events that smear across a conventional gather collapse to compact points in tau-p. This is why the technique excels in structurally deformed settings such as the Alberta Foothills, where steep dips defeat methods that assume flat layering.
- Premier multiple-suppression tool: Primaries and multiples with different moveout separate in the tau-p domain even when they overlap in offset-time. High-resolution Radon demultiple exploits this to attenuate water-bottom and short-period multiples, and tau-p deconvolution removes reverberations, producing a cleaner stack and more reliable amplitudes for interpretation.
- Enables interpolation and AVO: Plane-wave gathers from the slant stack support trace interpolation to regularize sparse or aliased 3D acquisition and feed amplitude-variation-with-offset analysis and prestack migration. Clean, regularized gathers are essential for the subtle amplitude work used to derisk Montney and Duvernay drilling targets.
- Reversible with care: The forward slant stack and its inverse let processors move data into tau-p, apply filtering or muting, and transform back to offset-time. High-resolution and sparse-inversion variants reduce the smearing and aliasing artifacts of the basic transform, which is critical when the goal is to preserve primary amplitudes while removing only the unwanted multiple energy.
Slant Stack in Alberta Foothills Imaging
The Alberta Foothills present some of the hardest imaging in the WCSB: imbricate thrust sheets, overturned limbs, and reflectors dipping 30 degrees or more put crossing linear events in every gather. A conventional normal-moveout-and-stack flow blurs these, but a slant stack separates the steeply dipping thrust-sheet reflections from gentler events by their slowness, letting processors filter coherent linear noise such as ground roll and refracted head waves while preserving the structural signal. The cleaned plane-wave gathers then feed depth migration that can position a Mississippian or Devonian target beneath a thrust with enough confidence to spot a well, where a mispositioned image could place a costly deep well off the structure entirely.
Demultiple and Interpolation in Resource Plays
In the basin proper, the slant stack is most often deployed for demultiple and trace regularization. Modern WCSB 3D surveys can have irregular fold and spatial aliasing from access and budget constraints, and tau-p interpolation fills those gaps so prestack migration does not generate operator artifacts. High-resolution Radon demultiple then strips intrabed multiples that would otherwise mimic or mask thin Montney and Duvernay reflectors. The payoff is measured in interpretation confidence: cleaner amplitudes mean AVO and inversion products more faithfully track porosity and fluid, reducing the chance of drilling a seismic artifact rather than a reservoir.
Fast Facts
The transform underlying the slant stack is named for Johann Radon, the Austrian mathematician who described it in 1917, decades before any seismic recording existed and in a purely abstract paper on integral geometry. The very same mathematics that lets a slant stack pull a dipping reflector out of foothills noise is what reconstructs a medical CT image from X-ray projections. Geophysics and radiology independently rediscovered Radon's 1917 result and built two entirely different multibillion-dollar industries on the same equation.
Related Terms
The slant stack is a core operation within seismic processing, the workflow that turns field recordings into interpretable images. Its main payoff is multiple attenuation, removing the reverberations that contaminate the primary reflections geophysicists actually want. It operates on common depth point and shot gathers organized by offset, and its strength in steeply dipping settings makes it indispensable to imaging structurally complex foothills plays.
Real-World WCSB Scenario
A processing contractor reworking a legacy 3D survey over a Foothills gas prospect southwest of Calgary applies a tau-p demultiple and dip-filter flow to suppress strong intrabed multiples and refracted noise that had obscured a Mississippian target beneath a thrust. The reprocessing package, including slant-stack demultiple, interpolation, and prestack depth migration, runs about CAD 90,000 for the survey, a fraction of the cost of a single deep Foothills well.
The cleaned image repositions the target culmination roughly 250 m laterally from the legacy interpretation, and the operator moves the proposed surface location accordingly. The well finds pay on structure, an outcome the original multiple-contaminated volume would have missed, validating the modest reprocessing spend many times over.