The Gold Standard of Archaeology | H/V Directionality, Wavefield Composition and Two-Dimensional Structure

1. Keep the directional quantity explicit

Back to contents

State whether the analysis rotates the horizontal time series before spectra, projects spectral amplitudes by azimuth, uses radial and transverse components or compares component-specific H/V curves. These operations are not interchangeable. Report the azimuth convention, step size, smoothing and horizontal denominator or numerator rule. A coloured polar plot is not self-explanatory unless the calculated quantity is defined.

2. Orient and level the instrument before interpreting azimuth

Back to contents

A directionality study depends on trustworthy component orientation. Record true north, magnetic declination, installation marks, levelling, coupling and any later correction. A swapped or reversed component can rotate the apparent maximum. Check sensor response and timing across channels. Reoccupy at least some stations with a known rotation when the directional pattern will carry archaeological weight.

3. Distinguish structure from wavefield composition

Back to contents

The ambient field can contain Rayleigh waves, Love waves and body waves arriving from unequal directions. Rayleigh motion contributes vertical and radial energy; Love motion is transverse and has no vertical component. A dominant road, coastline or industrial source can therefore create a directional H/V maximum without a directional subsurface resonance. Arrays, beamforming, source logs and time-window comparison help identify the wavefield.

4. Look for repeatability across time, station and frequency

Back to contents

A structural effect should usually show spatial and frequency behaviour that makes geological sense. Compare day and night, weekdays and quiet periods, and independent windows. Map peak frequency, amplitude and preferred azimuth across neighbouring stations. A compass direction that follows a moving machine schedule is weak structural evidence; a pattern that rotates with a mapped basin edge across stations is more persuasive.

5. Recognise signatures of two-dimensional structure

Back to contents

A dipping interface, valley edge or narrow sediment body can alter resonance frequency, particle motion and amplification with azimuth. Two-dimensional resonance may align with or across the long axis depending on geometry and mode. Peak splitting, azimuth-dependent frequency and systematic spatial rotation deserve modelling beyond a one-dimensional column. Directionality alone does not identify which 2-D geometry produced it.

6. Worked example: a polar maximum beside a buried ditch

Back to contents

A grid survey shows the largest H/V amplitude east–west at 5 hertz, apparently parallel to a suspected ditch. Night recordings retain the frequency but lose most amplitude directionality, while an array indicates daytime Love-wave energy from a road west of the site. Stations crossing the ditch still show a small, coherent frequency shift. The team attributes the strong polar amplitude to the source field and the spatial frequency change to shallow structure, then tests that interface with electrical resistivity.

7. Use arrays and active checks when one station is ambiguous

Back to contents

A small array can estimate arrival direction, phase velocity and wave type over useful bands. Active surface-wave data, boreholes, refraction, electrical methods or repeat sensors can constrain depth and geometry. Choose the check that separates the competing explanations. More H/V stations improve spatial description but do not by themselves resolve whether the energy is Love, Rayleigh or body wave.

8. Changed-context transfer check: a deep sedimentary valley

Back to contents

Move from a shallow archaeological site to a kilometre-scale valley. The transferable principle is to compare directional observations with source and structural models. The changed scale brings longer wavelengths, regional sources, basin-edge waves and stronger three-dimensional effects. A method calibrated for a ten-metre target cannot simply be stretched to a valley, and valley directionality does not prove a buried wall at a local station.

9. Model alternatives rather than one preferred azimuth

Back to contents

Forward-model plausible one-dimensional, dipping-interface and two-dimensional structures, and compare them with directional frequency and amplitude patterns. Vary source azimuth and Love–Rayleigh mixture where possible. Treat amplitude fits cautiously because damping and wavefield assumptions matter. The useful result may be a set of geometries that survive tests, not one sharp cross-section.

10. Know the limits and common misconceptions

Back to contents

A stable preferred direction is not automatically anisotropy. It may reflect geometry or sources. The maximum H/V azimuth is not necessarily the strike of an interface, and a polar amplitude lobe does not give depth. Averaging horizontals can conceal information but may also be more robust for a basic site frequency. Directional analysis adds diagnostic dimensions; it does not remove non-uniqueness.

11. Audit acquisition, rotation and interpretation together

Back to contents

Archive raw three-component records, orientation notes, rejected windows, rotated spectra, polar plots, source observations, station maps, array products and alternative models. Publish the frequency bands in which directionality is stable and the bands rejected as source-dominated. Show how the archaeological conclusion changes when azimuth step, smoothing, window selection and component combination change.

12. The gold standard uses directionality to test the story

Back to contents

Directional H/V is strongest as a challenge to simple interpretation. The gold standard earns the compass pattern through verified orientation, repeatable windows, spatial coherence and wavefield evidence. It compares one-dimensional and two-dimensional explanations, uses independent checks and keeps source effects visible. Only then can a directional feature contribute to a defensible archaeological model.