The Gold Standard of Archaeology | Seismic Attribute Analysis, Instantaneous Phase and False Edges

1. Begin with the archaeological boundary you expect

Back to contents

State whether the target is a wall face, trench margin, floor, cavity roof or sediment contact, and predict its orientation, depth and seismic contrast. Attribute analysis should test a physical expectation, not search an unlimited colour palette for attractive shapes. Record what the raw traces ought to contain if the feature is real and which alternative structures could create the same response.

2. Build the analytic trace from a documented signal

Back to contents

The recorded trace is treated as the real component of a complex signal; a quadrature component is commonly produced with a Hilbert transform. Instantaneous amplitude is the envelope, while instantaneous phase is the angle between the components. These quantities inherit the data’s sampling, bandwidth, deconvolution and phase character. They are derived views, not new independent measurements.

3. Understand what phase removes—and what it does not

Back to contents

Instantaneous phase deemphasises absolute amplitude, so a continuous event can remain visible where coupling or reflection strength changes. Yet phase becomes unstable when the envelope approaches zero, and every sample still depends on the filtered waveform. A phase image cannot be called amplitude-free in the broader evidential sense: low signal-to-noise zones, wavelet changes and interference continue to shape it.

4. Read edges as signal transitions, not object outlines

Back to contents

A constant-phase band may follow a reflector, while a sharp phase change may mark interference, termination or a different waveform. The coloured edge is not automatically the edge of a wall. Compare the attribute with wiggle traces, envelope, frequency content and expected travel-time geometry. Archaeological identification requires continuity across lines and support from acquisition geometry or another method.

5. Identify common sources of false edges

Back to contents

Band-pass choices can move zero crossings; automatic gain can promote weak noise; migration can focus energy; muting can introduce abrupt boundaries; phase wrapping creates colour jumps at the chosen display interval. Surface waves, multiples and out-of-plane arrivals can also remain coherent. Recalculate attributes across reasonable preprocessing choices and mark any boundary that moves, appears only after one filter or follows acquisition footprints.

6. Worked example: a suspected ditch beside a stone wall

Back to contents

A profile shows a faint break beside a strong shallow reflector. Instantaneous phase produces a crisp vertical colour boundary that looks like a ditch edge. On raw gathers, the boundary begins where a ground-roll mute changes slope. A gentler mute and an alternative band-pass shift the phase edge by two metres, while the stone-wall reflection remains stable. A resistivity line supports the wall but not the ditch. The defensible map retains the wall and labels the second boundary as a processing-sensitive anomaly.

7. Use a small attribute panel with independent checks

Back to contents

Choose attributes that answer distinct questions: envelope for local energy, phase for continuity, frequency for waveform change and coherence for lateral similarity. Highly correlated displays are not independent confirmation. Validate a candidate on adjacent lines, forward models, known test features or withheld profiles. Prefer the simplest attribute set that distinguishes the archaeological hypothesis from named alternatives.

8. Changed-context transfer check: ground-penetrating radar

Back to contents

Move from shallow seismic reflection to ground-penetrating radar. Analytic-signal attributes can still show envelope, phase and frequency, and false edges still arise near low amplitude or processing boundaries. The changed physics adds antenna ringing, electromagnetic velocity, attenuation and polarity conventions. A useful transfer preserves the audit logic while recalibrating what phase continuity means for the instrument and material contrast.

9. Know the limits and common misconceptions

Back to contents

Instantaneous phase does not reveal material type, and a sharp phase boundary does not guarantee a sharp geological boundary. Attribute displays can improve interpretation without increasing true resolution. Colour scales influence perception, while a two-dimensional line can misplace three-dimensional energy. A machine classifier trained on attributes inherits every preprocessing choice. Excavation or independent geophysics remains necessary for identity.

10. Audit the route from waveform to mapped edge

Back to contents

Archive raw gathers, geometry, instrument response, coupling notes, preprocessing order, filter corners, wavelet and polarity conventions, Hilbert implementation, attribute parameters and colour scales. Export the same section with several defensible choices and overlay candidate edges on raw data. Record which features persist across lines and methods. The audit should let another analyst reproduce both the boundary and its uncertainty.

11. Frequently asked questions

Back to contents

Why use instantaneous phase? It can make event continuity easier to follow when amplitude varies. Is it independent of amplitude? The phase formula is normalised, but low envelope and processing still matter. Does a colour jump equal a wall edge? No. Should many attributes be stacked for confidence? Only when they contribute genuinely different evidence. Can attributes repair poor acquisition? No; they reorganise the signal that was recorded.

12. The gold standard keeps the waveform beneath the colour

Back to contents

A credible attribute interpretation begins with a physical target and remains traceable to raw seismic evidence. It shows why a derived edge is stable across reasonable filters, lines and independent checks, and it labels low-energy or processing-sensitive zones honestly. The gold standard uses phase to guide attention, not to turn display contrast into archaeological certainty.