Mute: NMO Stretch Removal, Air Wave and Ground Roll Suppression, and WCSB Seismic Stacking
In seismic data processing, to mute is to set the amplitude of selected portions of seismic traces to zero so that they contribute nothing to the final stacked image. The purpose is to remove energy that would degrade the stack rather than improve it: air waves travelling directly through the atmosphere at roughly 330 m/s, ground roll (Rayleigh-wave surface energy) that dominates the low-frequency near-offset part of a shot record, refracted first arrivals, and the badly distorted far-offset data created by normal moveout (NMO) stretch. Muting is one of the most consequential editing decisions in a processing flow because a stack is simply the summation of many NMO-corrected traces at a common midpoint, and any noise left in those traces is summed along with the signal. Three families of mute are applied in a typical Western Canadian Sedimentary Basin processing sequence. The front-end or top mute zeroes everything above a picked time-offset curve to eliminate first breaks, the direct wave, and refracted arrivals that would otherwise smear across the shallow section. The surgical or inside mute removes a windowed patch of coherent noise, most often ground roll and air-blast energy that forms a triangular fan of low-frequency, low-velocity arrivals on land records shot with dynamite or vibroseis sources. The stretch mute, applied after NMO correction, removes the far-offset, shallow-time part of each gather where the moveout correction has stretched the wavelet so severely that its frequency content has dropped and it no longer matches the zero-offset wavelet. NMO stretch is unavoidable because the correction is larger at long offsets and short times, and stacking a stretched wavelet against an unstretched one destroys high-frequency resolution. Processors define a stretch-mute percentage, commonly 25 to 35 percent, that automatically zeros samples once the fractional frequency change exceeds that threshold, then refine it with hand-picked mute functions. The art of muting lies in the trade-off it forces: an aggressive mute removes more noise but discards fold, lowering the signal-to-noise ratio of the stack, while a permissive mute keeps fold high but lets distorted or noisy energy bleed into the image. On WCSB targets such as the Montney, Duvernay, and Cardium, where the exploration goal is to map thin, high-frequency reflectors and subtle stratigraphic edges, mute design directly controls how much usable bandwidth survives into the final volume that an interpreter loads in Calgary.
Key Takeaways
- Zeroing, not filtering: A mute hard-sets trace amplitudes to zero over a defined time-offset region, removing that energy from the stack entirely. Unlike a frequency filter, it discards the data rather than reshaping it, which is why mute picking is treated as an irreversible editing step in the processing flow.
- Front-end mute kills first arrivals: The top mute follows the picked first-break and refraction curve, zeroing the direct wave and head waves that arrive before the reflections of interest. Without it, high-amplitude refracted energy at far offsets would dominate the shallow stack and obscure genuine reflectors.
- Stretch mute protects bandwidth: NMO correction stretches the wavelet most at long offsets and shallow times, lowering its frequency. A stretch-mute percentage of typically 25 to 35 percent removes samples once distortion exceeds that limit, preserving the high-frequency content critical for resolving thin WCSB reservoirs.
- Surgical mute attacks ground roll: An inside or surgical mute excises the triangular fan of low-velocity, low-frequency ground roll and air-blast energy on land shot records. This is especially important for vibroseis and dynamite surveys over the foothills and plains of Alberta and northeast British Columbia.
- Mute trades fold for cleanliness: Every muted sample reduces the fold (number of traces summed) at that time. Aggressive muting raises noise rejection but lowers signal-to-noise ratio and weakens AVO analysis, so processors balance mute severity against the offset range needed for amplitude studies.
Stretch Mute and Far-Offset AVO
The stretch mute is in direct tension with amplitude-versus-offset (AVO) analysis, a workhorse technique for predicting gas and fluid effects in WCSB sands and shales. AVO needs the longest possible offset range to measure how reflection amplitude changes with angle, but those long offsets are exactly where NMO stretch is worst. Processors working a Montney or Falher gas play often relax the stretch-mute percentage to 30 to 40 percent on AVO-targeted volumes, accepting some wavelet distortion to retain the far-offset amplitudes, then apply a separate tighter mute for the structural stack. Choosing the wrong stretch limit can either kill the AVO signal or contaminate it with stretched noise.
Mute Picking in a Land Processing Flow
On a 3D land survey over the Alberta plains, mute functions are picked interactively on representative supergathers rather than trace by trace. The processor displays NMO-corrected gathers, identifies where the wavelet visibly broadens, and digitizes a smooth time-offset curve that the software interpolates across the survey. Ground-roll surgical mutes are picked on raw shot records before deconvolution, while the stretch mute is applied after velocity analysis. Quality control involves stacking with and without each mute and comparing the brightness and continuity of target reflectors, since an over-aggressive mute can erase a genuine dimming or stratigraphic pinch-out.
Fast Facts
NMO stretch is governed by the relation that the fractional stretch equals the change in moveout divided by the event time, so a reflector at 0.3 seconds recorded at a 3,000 metre offset can be stretched by more than 50 percent, while the same reflector at 2.0 seconds and the same offset stretches by only a few percent. This is why stretch mutes always carve a curved, time-dependent notch out of the far-offset corner of every gather, and why deep targets keep usable offsets that shallow targets lose entirely.
Related Terms
Muting is inseparable from several adjacent processing concepts. It is applied around normal moveout correction, since the stretch mute exists specifically to clean up NMO distortion before summation. The end product it protects is the seismic stack, where muted traces simply contribute zero fold. Mute design also interacts with ground roll suppression, because the surgical mute is one of the primary tools, alongside frequency and velocity filtering, for removing that coherent surface-wave noise from land records.
Real-World WCSB Scenario: Mute Design on a Duvernay 3D
A processing house in Calgary reprocessed a legacy 3D survey over a Duvernay shale play near Fox Creek, Alberta, where the operator needed crisp imaging of a 30 metre target at roughly 3,300 metres depth. The original 1990s mute had used a conservative 20 percent stretch limit that discarded far offsets and flattened the AVO response. Reprocessing with a 35 percent stretch mute, plus a carefully repicked surgical mute to strip residual ground roll, recovered roughly 8 Hz of high-frequency bandwidth at the target and restored the offset range needed for fluid prediction.
The improved volume let the interpreter distinguish a porous, overpressured fairway from a tight calcareous interval, guiding the placement of a CAD 9 million horizontal well into the better-quality rock. The reprocessing cost, near CAD 180,000, was trivial against the value of landing the lateral in the productive window rather than a marginal one.