Why the North Magnetic Pole Is Racing to Siberia: A 2,890-km-Deep Tug-of-War Written on Every Compass

What you'd actually see: a compass that keeps changing its mind

Take a good magnetic compass to almost anywhere on Earth and it will not point at the geographic North Pole — the fixed point where the planet's spin axis pierces the surface. It points, more or less, toward the north magnetic dip pole: the single spot in the Arctic where the field is aimed straight down into the ground, so a freely suspended needle stands vertical. The angular gap between where your compass points and true north is called magnetic declination, and in some places it exceeds 20°. Every aviation chart, marine chart, and smartphone maps app silently corrects for it.

Here is the unsettling part: that correction has an expiration date. The dip pole is not anchored. Over the past century it has slid out of Arctic Canada, wandered across the top of the Arctic Ocean, crossed the geographic pole's longitude, and is now closer to Siberia than to Canada — a milestone reached after more than 190 years of tracked motion. Because the pole moves, declination at any given city changes year by year. Airports occasionally have to rename their runways: a runway labeled "27" (270° magnetic) can drift enough that it must be repainted "28" to stay honest.

What you would not see is anything dramatic in the sky or underfoot. There is no rumble, no visible aurora tied to the drift, no sudden lurch. The whole phenomenon is invisible except through instruments — a needle, a satellite magnetometer, or a global model. That quietness is exactly why it took centuries of patient measurement to notice that the pole was accelerating.

The mechanism: a tug-of-war 2,890 kilometers down

Earth's magnetic field is generated by the geodynamo — churning, electrically conducting liquid iron and nickel in the outer core, a shell that begins about 2,890 km beneath your feet at the core-mantle boundary and runs down to roughly 5,150 km. Convection and Earth's rotation twist this molten metal into swirling flows that sustain a self-exciting electric current, and that current produces the field. Roughly 80% of the surface field looks like a simple bar magnet (a dipole) tilted about 9.5° from the spin axis (and currently closer to ~9°), but the remaining 20% is a lumpy, ever-shifting patchwork.

The north magnetic pole's position is largely set by a balance between two big patches of intense downward magnetic flux frozen into the top of the core — one under Canada and one under Siberia. Think of them as two magnets pulling the pole toward themselves in a tug-of-war. For most of recorded history the Canadian lobe won, which is why the pole sat in Arctic Canada. Then, in work led by Phil Livermore (University of Leeds) and colleagues, published in Nature Geoscience in 2020, researchers showed that a change in the pattern of core flow between roughly 1970 and 1999 stretched the Canadian lobe out, splitting and weakening its surface signature.

• The Canadian lobe weakened: between 2014 and 2025 its peak intensity dropped from about 58,832 nT to 58,031 nT.
• The Siberian lobe strengthened: over the same span it rose from about 61,359 nT to 61,619 nT.

Once Canada's grip loosened, Siberia won the rope, and the pole lurched toward Russia — fast. This is a beautiful example of secular variation: slow changes in the field driven by fluid motion deep inside the planet, not by anything happening at the surface or in space.

The numbers: from a crawl to a sprint and back to a jog

The drift history reads like an acceleration curve. From about 1600 to 1990, the north magnetic pole ambled along at roughly 10 km per year — slow enough that generations of navigators barely noticed. Then it took off:

• 1990s–2000s: speed climbed steeply, reaching a record of about 55 km per year.
• 2019: the drift was so fast and had pushed the pole so far from where the 2015 model predicted that NOAA and the British Geological Survey issued an out-of-cycle World Magnetic Model update in February — a full year early.
• 2020–2025: the pole unexpectedly decelerated to about 35 km per year, which the 2025 model flagged as the largest slowdown in pole speed on record.

For contrast, the south magnetic pole is a homebody, drifting at only about 9 km per year — the two poles are entirely independent because they respond to different flow patterns in the core. And to be clear about scale: 35 km per year is fast for a geophysical pole, but it is a walking-pace 0.004 km/h. "Racing" is relative to a process that normally unfolds over centuries.

The current slowdown was not predicted and is not fully explained; it appears the two lobes have shifted back toward a new balance. That honesty matters: the geodynamo is chaotic, and even the best models extrapolate the field forward only about five years before they must be re-fit to fresh satellite data.

A worked comparison: why the World Magnetic Model exists

To feel why a drifting pole is a practical headache, follow the chain from the core to your pocket. The World Magnetic Model (WMM) is a mathematical description of the entire global field, produced jointly by NOAA's National Centers for Environmental Information and the British Geological Survey and updated on a five-year cycle (2015, 2020, 2025…). It is the reference that hundreds of millions of devices use to convert a raw compass reading into a true-north bearing.

Your smartphone contains a tiny three-axis magnetometer. When a maps app shows which way you're facing, it reads that magnetometer, then applies the WMM's predicted declination for your GPS location and the current date to rotate the raw reading onto true north. The same model underlies the attitude-and-heading systems in aircraft, ships, submarines, drilling rigs, and the U.S. Department of Defense's navigation stack — which is why the model is co-produced for military use.

Now the problem is obvious. A five-year model assumes the field changes smoothly and predictably over those five years. When the pole sprinted at 55 km/yr, the real field outran the 2015 prediction near the Arctic, where declination errors matter most and where the geometry is most sensitive. By late 2018 the error had grown beyond the model's own accuracy specification at high latitudes, forcing the emergency February 2019 update. The lesson: the map of the field is only as good as our assumption that the core keeps doing what it just did — and the core does not sign contracts.

Misconceptions: this is not a pole reversal, and it isn't the apocalypse

Rapid drift routinely gets conflated with a full geomagnetic reversal, where the north and south magnetic poles swap. These are different phenomena. Drift is the everyday wandering of the poles; a reversal is a rare, global flip of the whole dipole that takes on the order of 1,000 to 10,000 years to complete. The last full reversal, the Brunhes–Matuyama, happened about 780,000 years ago, and reversals have occurred on average every couple hundred thousand years over the last few million.

It is tempting to add up "780,000 years since the last one" and "average ~200,000–300,000 years between them" and conclude we're overdue. Paleomagnetists caution against this. Reversals are irregular, not clockwork; the intervals vary enormously, and there have been stretches of tens of millions of years (like the Cretaceous Normal Superchron) with no reversals at all. Current evidence does not show the field entering a reversal.

Two facts that are real but often over-dramatized:

• The global dipole has weakened roughly 9% over the past ~180 years (since the first absolute intensity measurements of the 1830s–1840s) — measurable, but the field remains strong and this rate does not imply an imminent flip.
• The South Atlantic Anomaly, a region of unusually weak field over the South Atlantic and South America, lets radiation belts dip lower and forces satellites (including the Hubble Space Telescope) to power down sensitive instruments as they pass through. It is a genuine operational nuisance — but it is a regional weak spot, not proof of a global collapse.

Even during a real reversal, the field does not vanish. It weakens and becomes multipolar, but the planet retains meaningful protection; there is no credible evidence that past reversals caused mass extinctions.

History and observation: from Ross's dip circle to the Swarm satellites

The magnetic pole was first physically reached and measured on 1 June 1831 by James Clark Ross on the Boothia Peninsula in Arctic Canada. He confirmed it with a dip circle — an instrument that measures the field's inclination — reading a dip of 89°59′, essentially straight down. That single field observation is the anchor point from which the pole's entire 190-plus-year journey is traced.

For most of the intervening period, tracking the pole meant sporadic, arduous Arctic expeditions. Everything changed with space-based magnetometry. Denmark's Ørsted (launched 1999), Germany's CHAMP (2000), and above all the European Space Agency's three-satellite Swarm constellation (launched November 2013) now map the global field continuously and with enough precision to watch the flux lobes at the core-mantle boundary morph in near-real time. It was Swarm and CHAMP data that let Livermore's team catch the Canadian lobe stretching.

Satellites also do something ground stations never could: by combining data from many altitudes and separating the core's signal from crustal, oceanic, and space-weather contributions, they let geophysicists build models like the International Geomagnetic Reference Field (IGRF) and the WMM that resolve the field down to the core surface. The story of the racing pole is, in the end, a story about instruments finally becoming good enough to see a chaotic engine 2,890 km below the surface do something we never expected — speed up, then slow down — and to admit, honestly, that we cannot yet predict its next move.

Three different "north poles" that are routinely confused — they are in three different places and move for different reasons.

Feature	Geographic North Pole	Geomagnetic North Pole	North Magnetic (Dip) Pole
Definition	Where the spin axis meets the surface	North end of the best-fit dipole (~80% of the field)	Where field lines point straight down (dip = 90°)
Location (approx.)	Fixed at 90° N	NW Greenland, ~80.7° N	Arctic Ocean, drifting past ~86° N toward Siberia
What moves it	Essentially fixed (tiny polar wobble)	Slow changes in the overall dipole	Regional flux lobes in the liquid core
Speed	Negligible	A few km/yr	≈ 35–55 km/yr in recent decades
What it steers	Star charts, true north	Auroral oval, geomagnetic latitude	Your compass needle and phone declination