Auto-tracking: when it wins, when it fails

How auto-tracking actually works

Strip away the marketing and production auto-trackers do something very simple:

• Read the seed pick's amplitude sign, call it positive (peak) or negative (trough).
• Move to the next crossline.
• Within a small time window around the seed's time, search for the nearest local extremum of the SAME sign.
• If one is found AND the time jump from the seed is within a dip tolerance AND the amplitude is above a noise threshold, place a pick there.
• Use that new pick as the seed for the next crossline. Repeat.
• When any condition fails (no extremum, time jump too big, amplitude too small), stop.

That's the whole algorithm. No machine learning, no pattern recognition in the modern sense. Its power comes from the fact that real reflectors are laterally continuous in amplitude, polarity, and time, so a 1-D continuity follower tracks them well. Its limits come from the cases where that continuity breaks.

Where auto-tracking wins

• High signal-to-noise zones. When the target reflector is an order of magnitude brighter than the noise, the nearest-extremum search finds the right event every time.
• Continuous, gently-dipping reflectors. Normal sedimentary reflectors with no tectonic complications are the bread-and-butter case. Auto-track thousands of traces in seconds, then QC the map for anomalies.
• Dense, regular survey geometry. When consecutive traces are close in space (10 m to 25 m bin size) the reflector barely moves between traces and the tracker's small search window succeeds.
• Zero-phase data. The peak sits on the reflector, so snap-to-peak behaves. Phase-rotated or minimum-phase data confuses the tracker because "the peak" may not be on the reflector.

Where auto-tracking fails

• Faults. The reflector on the other side of the fault sits at a different time. The tracker, which only knows about local continuity, either drops the pick (if the time jump exceeds the dip tolerance) or walks onto a DIFFERENT reflector on the other side of the fault that happens to be near the tracker's last time estimate. The second case is especially dangerous, the tracker keeps going, confidently picking the wrong reflector across the entire downthrown block.
• Stratigraphic pinchouts. The reflector gets progressively dimmer as the bed thins below tuning. The tracker's amplitude threshold eventually rejects the pick and the horizon stops, often correctly, but sometimes in a way that hides a pinchout that is still geologically real.
• Low S/N or attenuation zones. In deep, attenuated parts of the volume, the "brightest extremum" may be noise rather than the reflector. The tracker picks noise.
• Overlapping reflectors of similar brightness. When two reflectors are near each other and comparably bright, the tracker flips between them, producing a "noisy" horizon that is really a hybrid of two different surfaces.
• Lateral facies change. A sand grading into a shale changes the reflector's sign or strength. The tracker stops (polarity flip) or walks onto a different event.

Exercise

• Seed. Click once on the brightest continuous reflector near the middle of the section, placing a single high-quality seed pick. Turn on Show trouble zones first if you want to see where the five failure modes live before you go looking for them.
• Auto-track the inline. Switch the mode to Auto-track and click on the reflector. Picks propagate outward in both directions from where you clicked, and the message underneath reports, separately for each direction, which of the conditions above stopped it and at which crossline. Read that line: it is the difference between watching a tracker stop and knowing why it stopped.
• Watch the confidence. Every auto-tracked pick carries its waveform correlation against your seed, and the horizon is drawn faded where that correlation falls. A faded stretch with healthy amplitude is the signature of a tracker that has walked onto a different event rather than lost the one you seeded.
• Propagate across inlines. Press Propagate to next inline repeatedly and watch how many picks survive each step.
• Grow it back. On any inline you reached by propagating, press Auto-track this inline again. It regrows every run that survived, outward, until each one meets a neighbour or stops for one of the five reasons above, and it never overwrites a pick you already have. Alternating propagate and regrow across forty-odd inlines takes the surface from about 700 picks to about 2,300, and coverage from 19 to 29 per cent. The holes it cannot close are the ones that are real: across the fault, and through the low signal-to-noise band. Pressing it a second time on the same inline does nothing, and says so, because regrowing from the frontier of the last regrowth would quietly reset the correlation reference and walk the horizon onto a different event, which is the failure this whole section is about.
• Observe what happens at the faults. There are many, and that is the point: two to ten of them cross any inline you display, with throws spanning more than an order of magnitude. Seed the tracker in different places and it stops for different reasons. Across ten seeds tried while building this section, the tracker ran between 33 and 93 traces and stopped on the dip tolerance seven times, on the amplitude gate eight times, and at the survey edge five times. You will not get the same answer twice, which is exactly the experience the taxonomy above is trying to give you a vocabulary for. Dropping the pick is the safer outcome. The dangerous one is when the tracker finds an event inside its tolerance on the far side of a small fault and carries on confidently down the wrong reflector.
• Grade. Press Grade. The RMS aggregates across every pick on the active horizon and is measured against the field's own truth function, the same one used to build the reflector, so the answer key cannot drift from the data. One sample is 4 ms.

The QC discipline

Auto-tracked picks are never final until they have been inspected. The time map is your primary QC tool. What to look for:

• Smooth colour gradients. Sedimentary reflectors produce maps that look like smooth topography. Sudden spots of very different colour in the middle of a smooth field are almost always pick errors.
• Sharp linear discontinuities. These may be faults (real) or tracker failures (not real). Cross-reference with your structural picks from Section 2.5 (when we build the fault picker) and with coherence attributes (Part 6). A lineament on the time map that has no corresponding fault on the structural interpretation is a bug, not a feature.
• Areas where picks drop out. The tracker gave up. Usually because the reflector became too dim or the dip too steep for the tolerance. You may need to re-seed in that area manually, or widen the tolerance and re-track.
• Near-well consistency. If you have wells, check that the auto-tracked horizon matches the well marker at each well within your expected depth uncertainty. Any systematic offset at wells means the tracker walked onto the wrong reflector somewhere between the seed and that well.

A working rhythm

The professional workflow combines manual and automatic:

• Seed manually at wells (highest-confidence anchor points).
• Seed manually at a few widely-spaced other inlines to form a skeleton.
• Auto-track from each seed, filling the volume.
• QC on the time map. Fix problem zones (delete bad picks, re-seed, re-track with adjusted tolerances).
• Cross-check at wells one more time.
• Ship the horizon to the downstream workflow (depth conversion, attribute extraction, volumetric calculations).

The step that junior interpreters skip is step 4, the QC. The interpreters whose maps get drilled through without nasty surprises are the ones who treat step 4 as the most important step, not the last one.