Zero Crossing: Phase, Wavelet Polarity, and Seismic Horizon Picking
A zero crossing is the point on a seismic trace where the amplitude passes through zero, the null between a positive peak and a negative trough or vice versa, and it marks the instant where the phase of the underlying periodic signal is either zero or pi radians. Every reflection event recorded in a 3D seismic volume is a wavelet, not a simple spike, so a single subsurface interface produces a peak, a trough, and the zero crossings that bound them rather than one clean impulse at the reflector depth. Interpreters lean on zero crossings constantly because, for a zero-phase wavelet, the crossing itself, not the peak or trough, sits closest to the true position of an abrupt acoustic impedance boundary such as the top of a carbonate reef or the base of a channel sand. Autotracking software in interpretation platforms like Petrel or Kingdom snaps horizon picks to the nearest zero crossing of a chosen polarity, since crossings are less sensitive to amplitude tuning effects than peaks or troughs when bed thickness approaches or falls below a quarter wavelength. Polarity convention matters here: under SEG normal polarity, a downward-to-upward crossing (trough-to-peak) corresponds to a positive reflection coefficient, meaning acoustic impedance increases across the boundary, while a peak-to-trough crossing signals a decrease. In the WCSB, this distinction is routine practice when tying seismic to well control across the Montney, Duvernay, and Viking, where a hard carbonate or tight siltstone cap over a softer shale produces a reliably repeatable positive zero crossing that geophysicists use as a mapping horizon across an entire play fairway. Processing choices upstream of interpretation, including wavelet phase rotation, deconvolution, and migration, all shift where crossings land relative to true reflector depth, so a poorly phased zero-phase conversion can misposition a pick by several milliseconds, translating to real vertical error at typical WCSB seismic velocities of 3,000 to 5,000 m/s (roughly 9,800 to 16,400 ft/s). Because of that sensitivity, well-tie quality control against synthetic seismograms, generated from sonic and density logs under AER Directive 059 data requirements for well information, is standard practice before any zero crossing is trusted as a mapping surface for reserves booking or drilling location selection.
Key Takeaways
- Marks the phase-zero point: A zero crossing occurs where trace amplitude equals zero and signal phase equals zero or pi radians, sitting exactly between an adjacent peak and trough on the waveform.
- Best pick point for zero-phase data: For a properly zero-phased wavelet, the zero crossing aligns most closely with the true depth of a sharp acoustic impedance boundary, which is why horizon autotracking snaps to crossings rather than amplitude extremes.
- Polarity tells you impedance direction: A trough-to-peak crossing under SEG normal polarity indicates increasing acoustic impedance across the boundary, while a peak-to-trough crossing indicates decreasing impedance, a distinction routinely used to separate a carbonate cap from an underlying shale in WCSB mapping.
- Sensitive to processing artifacts: Deconvolution, migration, and phase rotation choices during processing can shift where a crossing lands relative to the true reflector depth by several milliseconds, which at typical WCSB interval velocities of 3,000 to 5,000 m/s translates into meaningful vertical mispositioning of a mapped horizon.
- Well ties validate the pick: Synthetic seismograms built from sonic and density logs, checked against AER Directive 059 well data, confirm which zero crossing on the seismic trace actually corresponds to the geological marker before it is used for structure mapping or a drilling location.
Zero Crossings vs. Peak and Trough Picking
Picking amplitude peaks or troughs instead of zero crossings is common where beds are thick relative to the seismic wavelet and tuning is not a concern, but as reservoir thickness approaches a quarter wavelength (commonly 15 to 25 m in WCSB Cretaceous clastics at typical frequencies), peak amplitude becomes distorted by constructive or destructive interference between the top and base reflections. Zero crossings degrade more gracefully under tuning because the null point is less affected by amplitude-domain interference than the peak itself, though timing accuracy still depends on wavelet symmetry.
Applying Zero Crossings to Montney and Duvernay Mapping
In Montney and Duvernay exploration, geophysicists commonly tie a strong, laterally continuous positive zero crossing to the top of a carbonate or dolomitized marker overlying the target shale, then use that crossing as the primary structural mapping horizon across a play fairway spanning hundreds of sections. Consistency of that crossing's character (amplitude, polarity, and two-way time) from well to well is itself a quality check on both data processing and structural continuity.
Fast Facts
The convention of picking zero crossings instead of peaks dates to the shift from analog paper seismic sections in the 1960s and 1970s, when interpreters physically traced wiggle traces with a pencil, to digital workstations in the 1980s that could compute the exact zero-amplitude sample and automatically propagate a pick across thousands of traces in seconds rather than hours of manual tracing.
Related Terms
Zero crossing only makes sense in the context of the seismic trace it belongs to, since the crossing is defined relative to that trace's amplitude and time axis. The shape of the crossing and how it relates to the true reflector depth is governed by the wavelet convolved with the earth's reflectivity, and interpreters specifically seek a zero-phase wavelet because it centers energy symmetrically around the crossing. Broader signal phase concepts and amplitude behavior at the crossing both feed directly into how confidently a geophysicist can pick and propagate a horizon.
Real-World WCSB Scenario: A Duvernay Well Tie Goes Wrong
A Duvernay operator in the Kaybob area mapped a positive zero crossing as the top of the Ireton shale marker across a 12-section 3D survey, then spaced four horizontal wells along the resulting structure map. The first two wells performed as predicted, but the third landed 8 m low of the intended target, and post-drill review found the original well tie had used a synthetic built from an older, poorly calibrated sonic log that mispositioned the pick by roughly 4 milliseconds.
Reprocessing the synthetic with a corrected sonic and rerunning the tie shifted the mapped horizon and flagged the structural low the third well had drilled into before the fourth location was finalized, avoiding an estimated CAD 6.5 million dry-hole risk on the remaining pad.