The pole that won't hold still

Methods & sources

Every number here is published data — nothing is stylised.

Data

Rotational (spin) pole — IERS Earth-orientation series EOP C01 (polar motion), 1900–2025.

Magnetic (dip) pole — NOAA/NCEI: the gufm1 historical field model before 1900, IGRF from 1900. 1600–2025.

Stars — J2000 positions (RA/Dec → ecliptic), treated as fixed; proper motion not modelled.

How it’s drawn

Azimuthal-equidistant projection (radial distance ∝ angle from the pole). The sky view is centred on the north ecliptic pole. Pole speeds are great-circle (haversine), ~11-yr smoothed.

What a compass actually points to

A compass follows declination — the horizontal field — which points toward this region, not exactly at it, and becomes useless near the pole. The red dot is the dip pole (field vertical); the geomagnetic pole (best-fit dipole axis) is a third, different point.

Why the pole moves at all

Earth’s dipole is about 9% weaker than in 1840; the South Atlantic Anomaly — a growing weak patch — reflects the same core changes; the last full magnetic reversal was ~780,000 years ago (Brunhes–Matuyama). The recent sprint is the core geodynamo reorganising — not a forecast of a flip.