dip moveout, Oil & Gas Glossary

Dip moveout, abbreviated DMO, is the seismic processing step that corrects for the reflection-point smear introduced by dipping reflectors so that traces in a common-midpoint gather genuinely image the same subsurface point and can be stacked coherently regardless of dip. The problem it solves arises from a subtle geometric fact: for a flat reflector, every source-receiver pair sharing a common midpoint reflects off the same subsurface point, so a normal-moveout correction followed by stacking works perfectly. For a dipping reflector, however, the reflection point shifts updip as offset increases, so a common-midpoint gather over a dipping bed contains energy from a smear of different subsurface points rather than one. Plain normal-moveout and stack therefore blur dipping events and, worse, mis-stack crossing reflections of different dips that share a midpoint, because each dip requires a different stacking velocity. DMO repositions the prestack energy so that each output trace corresponds to the zero-offset, normal-incidence reflection point, decoupling stacking velocity from dip and letting events of all dips stack together at a single dip-independent velocity. Conceptually DMO is a partial prestack migration: it performs the offset-dependent part of migration before stack, leaving the remaining zero-offset migration to be done conventionally after stack, which yields a large data-compression benefit because post-stack migration is far cheaper than full prestack migration. The technique was developed through the early and mid 1980s. Yilmaz and Claerbout introduced prestack partial migration in 1980, Deregowski and Rocca worked the constant-offset and layered-media formulation in 1981, and Dave Hale's 1983 Stanford doctoral thesis, Dip Moveout by Fourier Transform, gave the elegant frequency-domain operator that made DMO practical and widely adopted on production seismic. DMO can be implemented several ways: as prestack partial migration, in the time domain by Kirchhoff integral summation along DMO ellipses, as a finite-difference offset-continuation, or in the Fourier domain by Hale's method. In the Western Canadian Sedimentary Basin, DMO mattered most for imaging steeply dipping and structurally complex plays: the thrust-faulted Alberta Foothills, dipping Cardium and Viking clinoforms, reef flanks on Leduc and Nisku carbonate buildups, and fault-bounded compartments where conflicting dips overlap on a single gather. By the late 1990s and 2000s, full prestack time and depth migration grew cheap enough on modern compute to subsume the DMO step in many workflows, since prestack migration handles dip and reflection-point smear directly. DMO nonetheless remains an important concept and a still-used tool for legacy reprocessing, for fast turnaround, and for understanding why dip and stacking velocity are coupled, and a great deal of the WCSB's archived 1980s and 1990s seismic that operators still mine for infill targets was processed with DMO at its heart.

Key Takeaways

Reflection-Point Smear Is The Problem: Over a dipping reflector, the reflection point migrates updip as offset grows, so a common-midpoint gather samples a smear of points, not one. Plain normal-moveout and stack blur these events. DMO repositions the energy to the zero-offset reflection point so a clean, dip-correct stack is possible.
DMO Decouples Dip From Velocity: Without DMO, each reflector dip needs its own stacking velocity, so crossing events of different dips cannot stack at one velocity. DMO removes the dip dependence, letting all dips stack coherently with a single dip-independent velocity, which is essential where steep Foothills thrusts cross gentler bedding.
It Is A Partial Prestack Migration: DMO performs the offset-dependent part of migration before stack, leaving cheap zero-offset migration for after stack. This data compression made dip-correct imaging affordable on 1980s and 1990s compute, long before full prestack migration was economically routine for large WCSB surveys.
Hale's Fourier Method Made It Practical: Dave Hale's 1983 Stanford thesis gave a clean frequency-domain DMO operator that became the industry standard. Earlier work by Yilmaz, Claerbout, Deregowski, and Rocca framed the partial-migration concept; Hale's transform made it fast enough for production processing.
Largely Superseded But Still Relevant: Modern prestack time and depth migration handle dip directly and have absorbed the DMO step in many flows. DMO still matters for legacy reprocessing, quick-turnaround projects, and teaching the dip-velocity coupling, and underpins the archived WCSB seismic operators still use to find infill and step-out targets.

Why Conflicting Dips Defeat A Plain Stack

In the Alberta Foothills, a steeply dipping thrust sheet often overlies gently dipping footwall bedding, so a single common-midpoint gather contains two events with very different dips. A normal-moveout velocity tuned to flatten the gentle event over-corrects the steep one, and vice versa, so a plain stack smears or kills one reflector. DMO resolves this by moving each event's prestack energy to its true zero-offset reflection point first, after which both dips stack coherently at one velocity. The result is a far cleaner image of the imbricate thrust structures that host Foothills gas, where conventional processing would leave the steepest, often most prospective, fault blocks poorly imaged.

DMO In Legacy WCSB Reprocessing

Much of the basin's 2D and early 3D seismic from the 1980s and 1990s was processed with DMO before full prestack migration was affordable. Operators chasing infill and step-out targets in mature Cardium, Viking, and carbonate-reef plays frequently reprocess this vintage data, and understanding its DMO history matters because a DMO stack already carries dip corrections that must not be double-applied during modern migration. A reprocessing project might cost CAD 80,000 to CAD 250,000 for a township-scale survey, cheap insurance before committing a CAD 6 to 12 million horizontal well to a reinterpreted structure.

Fast Facts

The geometric heart of Kirchhoff DMO is the DMO ellipse: a single input sample is smeared into an elliptical impulse response whose shape depends only on offset and time, not on velocity, which is the surprising property that lets DMO decouple dip from velocity in the first place. Dave Hale was a graduate student at Stanford's Exploration Project when his 1983 thesis turned this messy operator into a clean Fourier transform, and the method spread so fast through the industry that within a few years DMO had become a near-default step in marine and land processing worldwide.

Dip moveout is the dip-correcting companion to Normal Moveout, which flattens reflection hyperbolas for flat beds but fails on conflicting dips without DMO. It is a partial form of Seismic Migration, the broader process that moves dipping events to true position, and both depend on an accurate Velocity Layering model to position energy correctly. The end product feeds the Common Midpoint stack, whose coherence DMO exists to protect over dipping reflectors.

Real-World WCSB Scenario: Imaging A Foothills Thrust Play

A gas explorer reprocesses a 1990s 2D line across an Alberta Foothills thrust belt where conventional stacks left the steep hanging-wall reflectors smeared. Applying DMO before stack and then post-stack migration sharpens an imbricate thrust block previously invisible, revealing a fault-bounded culmination at about 3,100 m (10,170 ft). The reprocessing and reinterpretation cost roughly CAD 140,000, a small fraction of the well it informs.

The clarified structure supported drilling a CAD 11 million sour-gas test that found the predicted closure. Without DMO, the conflicting dips of the thrust sheet and footwall would have stacked incoherently, leaving the prospect unmappable and the target undefined.