research / time integrity

Can a shared geophysical field provide an independent timing observable?

A preregistered six-hour O3 experiment using public LIGO Hanford and Livingston vault magnetometer channels.

P1 VERIFIEDP2 TRUED2b FAILED

Question

Two-way packet timing cannot identify asymmetric one-way delay from timestamps alone. The experiment asks whether a separately measured, naturally shared field adds an independent observable.

Method

Four magnetometer channels, 21,600 seconds, 128 Hz after decimation, twelve 30-minute windows. The analysis separates physical coherence from propagation-lag stability and uses surrogate nulls that destroy inter-site timing correspondence while preserving local spectral structure.

Result

P1 passes strongly: the primary canonical coherence at the fundamental is about 0.3291 with z≈24.54. P2 is true. D1 gives a sub-millisecond record-length information bound.

Failure retained

D2b does not pass. Eleven of twelve windows are significant and the median lag is within the geometric bound, but between-window scatter is about 4.953 ms against a frozen ≈0.796 ms threshold — a miss of roughly 6.2×.

Provenance

The frozen v4 source and v4.1 execution lineage are preserved. Byte comparison establishes the correction as exactly one serialization reference, downstream of scored calculations. The in-tree package includes sources, preservation code, scored results, spectra, checksums, and reproduction instructions; large capture/state binaries remain external with recorded hashes.

Limits

This is not a positioning system, timing service, GNSS replacement, or demonstration of a stable propagation-corrected offset at the preregistered precision.