Geomagnetic Reversal: When Earth's Magnetic Poles Flip

What actually happens: a compass gone rogue

Strip away the mystique and a geomagnetic reversal is deceptively simple to state: the magnetic north pole and magnetic south pole trade places. A compass that pointed toward the Arctic before a reversal would, afterward, point toward Antarctica. What makes it dramatic is how it happens. The field does not smoothly pivot like a spinning top tipping over. Instead, the strong, orderly dipole field — the simple bar-magnet pattern we live under — first decays away to a small fraction of its normal strength, breaks up into a messy tangle of weaker poles scattered across the globe, and only then reassembles in the opposite orientation.

During the low point, you might have found not two poles but four, five, or more, wandering and short-lived, so that a compass at any given spot would swing wildly over human lifetimes. The best-studied case, the Brunhes–Matuyama reversal ~780,000 years ago, is named for the boundary between the current "Brunhes" normal-polarity epoch and the earlier "Matuyama" reversed epoch. A 2019 study of a Chinese sediment record argued the full field-instability process — including a long pre-reversal weakening — spanned roughly 22,000 years, though at any single location the actual sign-flip can appear to take just 1,000–10,000 years. The duration remains genuinely debated because different rock and sediment archives record the event at different resolutions.

The crucial, often-missed point: reversals are not scheduled or periodic. They happen at wildly irregular intervals, from a few thousand years apart during hyperactive stretches to tens of millions of years apart during quiet ones. There is no clock ticking down to the next one.

The engine: a molten iron dynamo 2,900 km down

Earth's magnetic field is generated in the outer core, a shell of liquid iron and nickel beginning about 2,900 km beneath your feet, at temperatures near 4,000–5,400 °C in the outer core (climbing toward ~6,000 °C in the inner core) — comparable to the surface of the Sun. This ocean of molten metal is churned by heat escaping from the inner core, by lighter elements floating up as the inner core slowly freezes, and by Earth's rotation twisting the flow through the Coriolis effect. Moving electrically conducting fluid drags magnetic field lines with it, and under the right conditions the flow reinforces its own field. This self-sustaining loop is the geodynamo.

A dynamo like this has no preference for which way it points — the physics works equally well in either polarity. That symmetry is exactly why reversals are possible. The field inside the core is roughly 50 times stronger than what leaks out to the surface, and it is constantly jostled by turbulence. Most of the time the dominant dipole is robust and stable. But occasionally the swirling flow produces large patches of "reversed flux" that grow and undermine the main dipole. If they grow enough, the dipole collapses and the field must find a new stable configuration — which may be the old polarity again (a failed attempt, or excursion) or the opposite one (a completed reversal).

At the surface today, the field ranges from about 25 µT near the equator to ~65 µT near the poles — for comparison, a refrigerator magnet is roughly 5,000 µT up close, so Earth's field is genuinely feeble, yet it deflects the solar wind across a region tens of thousands of kilometres across. Supercomputer simulations of the geodynamo, pioneered by Gary Glatzmaier and Paul Roberts in the mid-1990s, spontaneously produced reversals for the first time — strong confirmation that no external trigger is needed. The core does this on its own.

How we know: stripes on the seafloor

The evidence that reversals are real and repeated is written into the rock itself. When lava or magma cools below the Curie temperature (about 580 °C for magnetite), its iron-bearing minerals lock in the direction of the ambient magnetic field like a frozen compass. Read enough of these fossil compasses and you can reconstruct the field's history — a science called paleomagnetism.

The decisive discovery came from the ocean floor. In 1963, Cambridge geologists Frederick Vine and Drummond Matthews, and independently the Canadian geologist Lawrence Morley, realized that the puzzling symmetrical stripes of alternating magnetic anomaly mapped across mid-ocean ridges made sense if two things were true: new crust is continuously created at the ridge and spreads outward (seafloor spreading), and Earth's field periodically reverses. Each stripe is a band of ocean crust that froze during one polarity epoch; the mirror-image symmetry on both sides of the ridge is the tape recording playing out in both directions. This is now called the Vine–Matthews–Morley hypothesis, and it turned seafloor spreading from a bold idea into an established fact — a cornerstone of plate tectonics.

• Lava flows give sharp, datable snapshots of the field at single moments — the Laschamp lava flows in France's Massif Central first revealed a reversed direction in the 1960s.
• Deep-sea and lake sediments accumulate slowly and continuously, capturing the gradual weakening and directional swings across a whole transition.
• Ice cores and tree rings record indirect fingerprints: when the field weakens, more cosmic rays hit the atmosphere and produce spikes of cosmogenic isotopes like ¹⁰Be and ¹⁴C.

The numbers: how often, and the great quiet spells

Over the last ~83 million years the geomagnetic timescale contains more than 180 reversals. But that average hides enormous variation. In the most recent stretch — the last ~31 million years — reversals have come at roughly 4 per million years, meaning a typical polarity interval (a chron) lasts on the order of a few hundred thousand years. Some chrons last only ~30,000 years; a few stretch beyond 10 million.

The most striking feature of the record is that the rate is not constant. During the Cretaceous Normal Superchron (roughly 84–121 million years ago), the field held a single normal polarity for almost 40 million years with essentially no successful reversals — the age of the dinosaurs unfolded under a stubbornly steady compass. Go further back and you find the opposite regime: around 165 million years ago, in the Jurassic, some intervals show more than 10 reversals per million years. Why the geodynamo shifts between hyperactive and superchron modes is one of the field's open questions, with suspicion falling on how the mantle above the core regulates heat flow over tens of millions of years.

It is worth being precise about the last flip: the Brunhes–Matuyama boundary is dated by ⁴⁰Ar/³⁹Ar and astronomical tuning to about 773,000–780,000 years ago. That means humans, as a genus, have never lived through a full reversal — Homo sapiens arose long after it. We have, however, lived through excursions: brief, failed reversals where the field plunged and recovered. The Laschamp excursion ~41,000 years ago dropped the dipole to roughly 10% of its present strength — as low as ~5% at the deepest moment — for a few centuries before the field snapped back to its original polarity, so our ancestors did experience a badly weakened shield.

Is one starting now? The South Atlantic Anomaly

Two facts fuel the modern anxiety. First, the field's dipole strength has been declining ~5% per century since about 1840, and by roughly 9% over the last 150 years. Second, there is a large, deepening weak spot — the South Atlantic Anomaly (SAA) — stretching from South America across the South Atlantic to southern Africa, where the surface field is markedly weaker and reversed flux is welling up from the core. ESA's Swarm satellite trio, launched in 2013, has tracked the SAA growing and splitting, expanding by an area comparable to a large chunk of a continent in just the past decade. Because the SAA lets energetic particles dip closer to the surface, satellites crossing it — including the Hubble Space Telescope — routinely switch off sensitive instruments to avoid glitches, and it is a known radiation hazard at low orbit.

Does this mean a reversal is imminent? Most geophysicists say probably not — or at least, not on any timescale that should worry a person. A 2018 study led by Maxwell Brown and Monika Korte compared today's field to two past excursion-onset situations and concluded the present configuration does not resemble the run-up to a reversal; the field is more likely undergoing a normal fluctuation. Key context often lost in headlines:

• The current field, even weakened, is still stronger than the long-term average of the last few million years — we may simply be coming down from an unusually high peak.
• Weak spots like the SAA come and go; a decaying dipole does not commit the field to flip. Many excursions weaken and then recover the same polarity.
• Even if a reversal were beginning, it would unfold over thousands of years, not overnight — there is no cliff-edge "flip day."

What a reversal would (and wouldn't) do

Popular culture loves to pin catastrophe on pole flips — crust cracking, oceans boiling, mass extinction. The geological record is far more reassuring. Reversals have happened hundreds of times during the evolution of complex life, and no reversal has ever been convincingly tied to a mass extinction. Life, including us, has walked through many weakened-field intervals and come out the other side. Crucially, a reversal does not change Earth's rotation, its gravity, or the geographic North Pole — only the magnetic orientation changes.

That said, the effects would be real and worth taking seriously, mostly because of the weakened, disorganized field during the transition rather than the flip itself:

• More radiation reaching the atmosphere and surface. With the magnetic shield diminished, more solar and cosmic particles get through. The immediate ground-level health risk is modest — the atmosphere itself is a formidable shield equal to ~10 metres of water or a few metres of concrete — but high-altitude and high-latitude exposure would rise.
• Auroras everywhere. With multiple weak poles, the auroral ovals would migrate and spread; a 2025 reconstruction of the Laschamp excursion found the auroras had wandered far from the poles toward the equator.
• Technology under strain. Our real vulnerability is not biological but electronic. Satellites, GPS, and power grids are sensitive to space weather; a Laschamp-like field, combined with a major solar storm on the scale of the 1859 Carrington event, could be very costly for a technological civilization. This is a strong argument for hardening infrastructure, not for panic.

The honest bottom line: a geomagnetic reversal is a slow, natural, survivable process that Earth has performed routinely for billions of years. It reshuffles compasses and auroras, briefly thins our invisible shield, and leaves its signature in stone — but it is not the end of the world, and it is almost certainly not happening on a human timescale near you.

Reversal vs. excursion: two very different kinds of magnetic upheaval

Feature	Full reversal	Excursion
Outcome	Poles swap permanently (until the next flip)	Field wanders/weakens, then snaps back to original polarity
Duration	~1,000–10,000 yr locally; whole process may span ~10,000–20,000+ yr	Often ~1,000–2,000 yr, sometimes centuries
Field strength at minimum	Collapses to ~5–25% of normal, briefly multipolar	Also collapses deeply (Laschamp: ~10% of present strength, as low as ~5%)
Frequency (recent)	~4 per million years	Several per million years — more common than full flips
Best-known example	Brunhes–Matuyama, ~780,000 yr ago	Laschamp, ~41,000 yr ago
Recorded in	Seafloor stripes, lava flows, sediment cores	Sediments, ice-core ¹⁰Be/¹⁴C spikes, lava