The Gold Standard of Archaeology | Shear-Wave Splitting, Polarisation Rotation and Fracture Orientation

1. Define the splitting measurement before naming fractures

Back to contents

Shear-wave splitting is observed as two polarised shear components with a time delay. The fast direction describes the measured wavefield under a model; it is not yet a mapped fracture strike. Record the phase, frequency band, source–receiver path, coordinate system and correction method. A directional interpretation begins after the measurement survives waveform and geometry checks.

2. Understand what fast direction and delay time contain

Back to contents

Fast polarisation can align with a symmetry direction created by fractures or material fabric, while delay time accumulates along the anisotropic portion of the path. Both are path-integrated. A large delay may reflect stronger anisotropy, a longer anisotropic path or several layers. One receiver cannot usually place the responsible layer at a unique depth.

3. Acquire enough components and azimuths

Back to contents

Reliable rotation requires correctly oriented multicomponent sensors and controlled source polarisations or suitable natural arrivals. Repeat shots, reciprocal geometry and several azimuths help separate stable subsurface patterns from source radiation and coupling. Orientation errors can rotate the estimated fast direction; poor horizontal coupling can create false component differences. Preserve raw coordinates before any rotation.

4. Use polarisation rotation as a test, not a reveal

Back to contents

Methods such as Alford-style rotation seek axes that minimise cross-energy or separate fast and slow modes. The result depends on window, frequency, noise, offset and layering. Scan plausible angles and delays, inspect objective surfaces and retain competing minima. A single visually clean rotated pair can be manufactured by flexible choices; stability across shots and windows is stronger evidence.

5. Distinguish one anisotropic layer from several

Back to contents

If successive layers have different symmetry directions, the first split wave can be split again. The combined waveform may be frequency- and backazimuth-dependent and cannot be represented faithfully by one fast direction and one delay. Compare single- and multilayer models, use varying paths where possible and ask whether near-surface anisotropy overwrites deeper information.

6. Worked example: a fast axis beside a buried wall

Back to contents

A multicomponent line beside a suspected wall produces a fast direction of 035° and a six-millisecond delay. The wall on the plan trends 030°, so the first interpretation calls the result masonry fractures. Repeated shots from the opposite end keep the delay but rotate the optimum toward 055° at higher frequency. A second line shows the same 055° direction away from the wall. The evidence supports shallow sediment fabric plus a local perturbation, not a uniquely fractured wall. Excavation is used to test the anomaly rather than confirm a label already assigned.

7. Relate splitting to fracture orientation cautiously

Back to contents

In a simple medium with one aligned, subvertical fracture set, fast polarisation may parallel fracture strike. Multiple non-orthogonal sets, bedding, stress-aligned microcracks, intrinsic mineral anisotropy and oblique incidence can change that relation. State the assumed symmetry and propagation angle. Where the claim matters, compare exposed structure, cores, excavation or independent geophysics rather than translating fast direction directly into a compass bearing for fractures.

8. Changed-context transfer check: a crustal earthquake record

Back to contents

Move from a shallow controlled-source profile to local-earthquake or teleseismic splitting. The transferable workflow still tests component orientation, windows, fast direction, delay and path integration. The changed scale introduces unknown source radiation, deeper paths and much poorer archaeological relevance. A mantle splitting tutorial cannot donate its depth or fracture interpretation to a twenty-metre site without rebuilding the acquisition model.

9. Know the limits and common misconceptions

Back to contents

Splitting demonstrates anisotropy somewhere along the sampled path under adequate data, but it does not identify cause or depth uniquely. Fast direction need not equal fracture strike in complex media. Delay time is not a direct fracture-density meter without path and rock-physics assumptions. Null measurements may arise from noise, unfavourable initial polarisation, symmetry or cancelling layers; they do not prove isotropy.

10. Audit waveforms, rotations and structural alternatives

Back to contents

Retain raw three-component data, sensor orientations, source geometry, coupling notes, filters, time windows, polarisation plots, angle–delay surfaces, quality grades and rejected records. Show results across frequencies and paths. Compare one-layer, multilayer and non-fracture explanations, and document independent site evidence. The final map should carry uncertainty in direction and spatial support, not only an arrow.

11. Frequently asked questions

Back to contents

What splits the wave? Elastic anisotropy along its path. Does the fast direction always follow fractures? No; that requires a specific structural model. What does delay time measure? Accumulated differential travel time. Can one shot locate the anisotropic layer? Usually not. Why do windows matter? Mixed phases and noise change the optimum rotation. What makes a fracture interpretation strong? Stable splitting plus compatible geometry and independent structural evidence.

12. The gold standard earns a fracture direction by exclusion

Back to contents

A credible splitting study starts with raw components, verifies orientation and tests whether the fast direction is stable across windows, frequencies and paths. It treats rotation as a model-dependent diagnostic, keeps depth non-uniqueness explicit and distinguishes anisotropy from its possible causes. The gold standard calls a fracture orientation only after simpler acquisition, layering and fabric explanations have been challenged.