Tail Mute: Ground Roll Suppression, Offset-Time Cutoffs, and WCSB Foothills Seismic Processing

A tail mute is a processing operation applied to seismic reflection data that zeroes out, or "mutes," a region of each shot or common-midpoint gather defined by a cutoff in record time, source-to-receiver offset, or both, with the specific goal of removing slow, coherent surface-wave energy before the data are stacked. The most common target of a tail mute is ground roll, a low-frequency, low-velocity, high-amplitude Rayleigh wave that travels along the near surface and arrives progressively later on the far-offset traces of a gather. Because ground roll plots as a steeply dipping linear fan in the time-offset domain, a tail mute drawn as a sloping line that follows the leading edge of that fan can excise the noise while preserving the underlying reflection signal. The operation gets its name from the part of the trace it acts on: where a front-end or top mute removes the early "head" of the record (first breaks and the refraction cone), the tail mute removes the late "tail" on far offsets where ground roll, air blast, and other slow surface modes dominate. Tail muting is one tool in a broader noise-attenuation workflow that also includes frequency filtering, f-k or tau-p velocity filtering, and surface-wave-adaptive subtraction, and it is usually applied conservatively so that genuine far-offset reflection energy, which carries the amplitude-versus-offset information used in fluid and lithology prediction, is not thrown away with the noise. In Western Canadian Sedimentary Basin work the tail mute is routine on 2D and 3D land surveys over the Montney, Duvernay, and Cardium, where strong ground roll from the unconsolidated glacial till and muskeg of the Alberta and northeast British Columbia near surface can otherwise swamp the deeper reflections geophysicists need to map drilling targets. Processors define the mute as a set of time-offset pairs picked interactively on representative gathers, then interpolate that geometry across the survey, often tapering the mute edge over a window of 50 to 150 milliseconds rather than applying a hard zero so that the abrupt amplitude step does not itself create artifacts in the stacked section. The cutoff values are velocity-aware: because ground-roll apparent velocity in the WCSB near surface is commonly 300 to 900 metres per second (roughly 1,000 to 3,000 feet per second), the mute line is steeper than the moveout of a deep reflection traveling at several thousand metres per second, which is exactly what lets the two be separated.

Key Takeaways

  • Targets slow surface waves: A tail mute removes coherent low-velocity noise, principally ground roll (a Rayleigh wave at roughly 300 to 900 m/s, or 1,000 to 3,000 ft/s) plus air blast and guided waves, from the far-offset, late-time corner of a gather where that energy crosses and masks the deeper primary reflections used to map drilling targets.
  • Defined by offset, time, or both: The mute is specified as a series of time-offset coordinate pairs picked on representative gathers, then interpolated across the survey. The geometry is chosen so the mute line is steeper than reflection moveout, exploiting the velocity contrast between slow surface waves and fast deep reflections to separate them.
  • Applied before stack: Tail muting happens during prestack processing, ahead of normal-moveout correction and the summation that forms the final stack. Muting noise on individual gathers prevents that energy from being smeared into the stacked trace, where it would be far harder to remove after summation has averaged it into every sample.
  • Tapered edges prevent artifacts: Rather than a hard zero, the mute boundary is usually ramped over a 50 to 150 millisecond window. A sharp amplitude step would introduce its own high-frequency ringing and edge artifacts into the section, so a smooth cosine or linear taper at the mute edge keeps the processed data clean.
  • Conservative to protect AVO: Because far-offset traces carry the amplitude-versus-offset signal used in fluid and lithology prediction, processors mute conservatively. Over-aggressive tail muting discards legitimate far-offset reflection energy and degrades AVO analysis on Montney and Duvernay programs where that signal drives interpretation.

Picking the Mute on Time-Offset Gathers

A processor picks a tail mute by displaying a representative common-shot or common-midpoint gather and identifying the boundary between the fast hyperbolic reflections and the slow linear ground-roll fan. Ground roll appears as a wedge of high-amplitude, low-frequency energy whose leading edge dips down toward later times on far offsets, following an apparent velocity of a few hundred metres per second. The analyst clicks a series of time-offset points along that leading edge, and the software interpolates a smooth mute function between them. On a WCSB 3D survey with thousands of gathers, the picks are made on a sparse grid of control gathers, perhaps every five hundredth gather, then spatially interpolated, with quality-control passes confirming the mute neither cuts into shallow reflections nor leaves residual ground roll on the longest offsets.

Tail Mute Versus Frequency and F-K Filtering

The tail mute is a geometric, surgical tool: it removes everything inside a chosen time-offset polygon regardless of frequency content, which is both its strength and its limit. Where ground roll and signal overlap in the same time-offset region, a hard mute would also remove the reflection, so processors pair the mute with frequency-selective methods. Low-cut filtering removes the sub-15-Hz band where ground roll is strongest, while f-k (frequency-wavenumber) and tau-p filters separate events by dip and apparent velocity, suppressing the linear ground-roll moveout while passing the flatter reflection moveout. A typical WCSB land-processing sequence applies a conservative tail mute first to remove the clearly isolated far-offset noise corner, then uses f-k or adaptive subtraction to attack the residual ground roll that still overlaps the reflection cone at intermediate offsets.

Fast Facts

Ground roll can reach amplitudes 100 to 1,000 times larger than the deep reflections a survey is trying to record, which is why it must be addressed early. The Rayleigh wave that produces it was described mathematically by Lord Rayleigh in 1885, decades before reflection seismology existed. In the muskeg and glacial-till near surface of northeast British Columbia's Montney fairway, ground-roll velocities can drop below 400 m/s (1,300 ft/s), producing some of the steepest, most aggressive mute geometries seen in North American land processing.

A tail mute is most directly tied to ground roll, the surface-wave noise it is designed to remove. It is one variety of mute, the general operation of zeroing portions of a seismic trace, and it works hand in hand with the stack that follows, since muting noise before summation keeps the stacked trace clean. The end product it protects is the interpretable depth section, where any ground roll left unmuted would appear as low-frequency smearing that obscures the reflections geoscientists use to plan wells.

Real-World WCSB Scenario: Cleaning a Montney 3D Near Dawson Creek

An operator shooting a 120 square kilometre 3D survey over a Montney development block near Dawson Creek, British Columbia, finds that raw shot gathers are dominated by ground roll traveling at roughly 380 m/s (1,250 ft/s) through the muskeg, with amplitudes large enough to clip the deeper Montney and Doig reflections near 1,800 to 2,400 metres depth. The processing contractor, working a budget of roughly CAD 18,000 to 25,000 per square kilometre for acquisition plus a separate processing fee, picks a tapered tail mute on control gathers every 500 CMPs, ramped over 80 milliseconds, then follows with a 6 Hz low-cut filter and f-k dip filtering on the residual.

The combined sequence lifts the signal-to-noise ratio on the far-offset traces enough to recover the amplitude-versus-offset response across the Montney, letting the geophysics team map a brittle, gas-charged interval and high-grade three horizontal well pads. Without the conservative tail mute preserving far-offset signal, the AVO attribute that flagged the best rock would have been buried in surface-wave noise.