Earth's Magnetic Field

The field that moves a compass needle is not fixed. It drifts, tilts and weakens on a timescale of decades, driven by convection in the liquid iron outer core 3,000 km down. These two scenes animate that drift from 1590 to 2025 on a World Grid globe.

Try it: download this example below (or the full examples set), then open Geomagnetic_Field_Example/Field_And_Poles/GeomagField_gv_node.csv — or drag its folder onto the window. Press Space to run the four centuries.

Field_And_Poles (2 MB) Intensity_Shell (2 MB)

The two scenes

Field_And_Poles/ — a lattice of 420 cone glyphs at 12° spacing, each aimed along the true local field: the cone's compass bearing is the declination, its dip is the inclination, its colour the field strength (purple 20 µT → yellow 68 µT). Both magnetic dip poles ride the same clock along trails of markers labelled by century.

Intensity_Shell/ — the same globe wearing a shell of 1,056 glyphs whose height above the surface is the field strength. Strong field over the poles stands high; weak field sags toward the ground.

What to look for

• The north dip pole's sprint. It crawled at roughly 6 km/yr through the 17th and 18th centuries and 10 km/yr through most of the 20th, then accelerated to about 43 km/yr since 1990, leaving Arctic Canada for Siberia. This is why the World Magnetic Model had to be updated off-schedule in 2019.
• The two poles are not opposite each other. Dip poles aren't antipodal, because the real field isn't a pure dipole — obvious the moment you can see both at once.
• Declination swinging over Europe. At London the model gives +11.6° east of north in 1590, −23.9° west in 1800, and +0.9° today. A 36° swing, and the reason every old British map carries a date beside its compass rose. Watch the cones over the North Atlantic rotate while the Pacific ones barely move.
• The South Atlantic Anomaly forms, deepens and marches west. In Intensity_Shell it is an enormous depression in the landscape — the region where Earth's field is weak enough that satellites take measurable radiation damage. From 31.2 µT off southern Africa in 1590 to 22.1 µT over Paraguay today: a 29% weakening and 74° of westward drift in four centuries, and very rarely rendered as anything but a contour map of one epoch.

The models

Model	Span	What it is
gufm1	1590–1990	Jackson, Jonkers & Walker (2000). Fitted to ~365,000 historical observations, most of them compass bearings logged by merchant and naval mariners — four centuries of the field reconstructed largely from ships' logbooks.
IGRF-14	1900–2030	International Geomagnetic Reference Field, 14th generation (IAGA, 2024). Satellite-constrained over its later epochs.

Nothing is downloaded: both scenes are computed from bundled Gauss coefficients and are pure CSV once built. The example ships two independent verification suites — one checking the physics against a separately written IGRF implementation, the analytic axial-dipole limit, and observed dip-pole positions from magnetic surveys; the other pulling each cone's direction back out of the real GlyphViz transform engine and confirming the picture agrees with the model to 0.0005°.

Aiming a glyph on a globe

Worth knowing if you build your own globe scene, because it took a probe against the real transform engine to establish: a globe child's local +Z is radially outward, +Y is east, +X is south. In Heading/Tilt/Roll mode, a compass azimuth A measured clockwise from north is heading = A − 90, so a field with declination D and inclination I is rotate_x = 90 + I, rotate_y = D − 90.