Signal: Seismic Reflection Energy, Signal-to-Noise Ratio, and Bandwidth Preservation in WCSB Imaging
Signal, in seismic and geophysical work, is the portion of the recorded wavefield that carries desirable information about the subsurface, the reflection energy that tells an interpreter where geologic interfaces lie, how deep they are, and what the rocks across them are like. Every seismic trace is a sum of two things: signal, the coherent reflections and refractions that obey the physics of wave propagation through layered earth, and noise, the unwanted energy, random or coherent, that accompanies the signal and partly masks it. The job of acquisition design and processing is to maximize the signal relative to the noise, a relationship captured by the signal-to-noise ratio (S/N), the single most important measure of seismic data quality. A high S/N record shows continuous, interpretable reflections; a low S/N record buries those reflections under ground roll, wind, swell, powerline hum, multiples, or random background, and no amount of clever interpretation can recover structure that the noise has erased. Signal in the reflection method originates when a controlled source, a vibroseis sweep or a dynamite charge on land, an air-gun array offshore, sends elastic energy into the ground that partly reflects at each contrast in acoustic impedance, the product of rock density and velocity. The reflected energy returns to geophones or hydrophones as the signal, and its amplitude, frequency content, and arrival time encode the depth and properties of the reflector. Because the earth absorbs high frequencies faster than low ones, signal bandwidth narrows with depth, so a shallow Mannville reflection may carry usable energy to 120 Hz while a deep Montney or Duvernay event is limited to perhaps 50 Hz, and preserving every hertz of that diminishing band is a central goal of amplitude-friendly processing. The contrast between signal and noise drives nearly every processing decision. Stacking, the summation of many traces that sampled the same subsurface point at different offsets, is the workhorse for improving S/N, since coherent signal adds constructively while random noise partly cancels, and the improvement scales roughly with the square root of the number of traces, or fold. Filtering, deconvolution, and migration all aim to enhance signal or reposition it correctly without destroying the amplitude and frequency information it carries. The danger in every step is that overly aggressive processing meant to suppress noise also attenuates real signal, particularly the low-frequency and far-offset energy that quantitative methods depend on. In the Western Canadian Sedimentary Basin, where interpreters chase thin Cardium and Viking sands, stratigraphic traps in the Mannville, and resource shales in the Montney and Duvernay, the usable signal bandwidth often decides whether a thin bed can be resolved or an amplitude anomaly trusted, so survey parameters such as fold, offset distribution, and source effort are chosen specifically to deliver the strongest, broadest signal the target depth allows.
Key Takeaways
- Signal Versus Noise: Signal is the coherent reflection energy carrying subsurface information; noise is everything else recorded alongside it. Their ratio, the signal-to-noise ratio, is the master measure of data quality, because reflections buried under ground roll, multiples, or random background cannot be interpreted no matter how skilled the interpreter or how advanced the workstation.
- Origin In Impedance Contrasts: Reflection signal is generated where seismic energy meets a change in acoustic impedance, the product of velocity and density. The strength of the returning signal scales with the size of that contrast, and its amplitude, frequency, and travel time encode the depth and rock properties of the reflector being imaged.
- Bandwidth Narrows With Depth: Earth absorption strips high frequencies faster than low ones, so signal bandwidth shrinks downward. A shallow Mannville event may carry energy to 120 Hz while a deep Duvernay reflection is limited near 50 Hz. Preserving the diminishing high-frequency signal is what lets interpreters resolve thin WCSB beds.
- Stacking Builds Signal-To-Noise: Summing many offset traces that sampled the same subsurface point, the stack, adds coherent signal constructively while random noise partly cancels, improving S/N roughly as the square root of fold. This is why high-fold surveys are specified over low S/N targets and why fold is a primary acquisition-design lever.
- Processing Must Protect Signal: Filtering, deconvolution, and migration aim to enhance or correctly reposition signal, but each can attenuate real reflection energy if applied too hard. Amplitude-preserving workflows guard the low-frequency and far-offset signal that AVO and seismic inversion require, treating signal fidelity as more important than a cosmetically clean section.
Signal-to-Noise Ratio and Fold
The practical handle on signal quality is the signal-to-noise ratio, and the most reliable way to raise it during processing is stacking. When many traces that illuminated the same common midpoint are corrected for moveout and summed, the coherent signal reinforces while incoherent noise tends to cancel, and the S/N improvement scales approximately with the square root of the fold, the number of traces in the stack. Doubling fold from 60 to 120 yields about a 41 percent gain in S/N, which is why surveys over deep, low-amplitude WCSB shale targets are designed for high nominal fold and broad offset distribution. Fold cannot fix a survey that never recorded the signal, but it is the cheapest insurance against random noise.
Protecting Far-Offset and Low-Frequency Signal
Quantitative interpretation depends on parts of the signal that are easy to lose. Amplitude-versus-offset analysis reads the change in reflection amplitude from near to far offset, so the far-offset signal must survive processing intact, not be muted or over-filtered. Seismic inversion needs the low-frequency end of the signal band to constrain absolute impedance, energy that a careless low-cut filter can strip. Processors therefore use gentle, well-documented filters and amplitude-preserving migration when the goal is rock-property work, accepting a slightly noisier-looking section in exchange for trustworthy amplitudes. Guarding these vulnerable bands of the signal is what separates a structural product from one fit for quantitative WCSB reservoir characterization.
Fast Facts
The square-root rule that governs how stacking strengthens signal is the same statistical law that lets astronomers pull a faint star out of a noisy image by combining many exposures, and seismologists exploit it relentlessly: a modern WCSB 3D survey may stack well over a hundred traces per bin, turning records where the raw reflection is invisible to the eye into a clean image. The deepest Duvernay and Montney targets are routinely interpreted from signal that, on any single field trace, is completely lost beneath the noise and only emerges once the fold is summed.
Related Terms
Signal is meaningless without its counterpart noise, since data quality is defined by the ratio between them, and a filter is the tool used to suppress noise so the signal stands clear. Signal is built up during seismic acquisition and strengthened by stacking many traces, and its preservation matters most for amplitude versus offset work, where the far-offset and low-frequency signal carries the rock-property information that turns a seismic image into a reservoir prediction.
Real-World WCSB Scenario
An operator planning a Montney development near Wapiti, Alberta, must image a target near 3,000 m where the reflection signal is weak and the overburden generates strong multiples. The geophysics team specifies a high-effort 3D survey, nominal fold above 120, long offsets to capture the far-offset signal needed for a later AVO screen, and a vibroseis sweep tuned to push energy into the high frequencies the deep target still supports, on an acquisition budget near CAD 2.1 million.
In processing, the contractor protects the signal with surface-consistent deconvolution and gentle, documented filtering, then stacks the high fold to lift the buried Montney reflection above the multiple noise. The resulting volume carries usable signal bandwidth and trustworthy far-offset amplitudes, and the AVO anomaly that survives supports a pad of horizontal wells that come in on prognosis, validating the upfront spend on signal quality.