Geomagnetic Reversal — When Earth's Poles Flip

Roughly 42,000 years ago, for about 300 years, Earth lost most of its magnetic shield. The field collapsed to 5–10% of normal strength, the poles wandered to the equator, and aurorae appeared over the tropics. Compasses, had any existed, would have spun. The planet has done this — partially or fully — at least 183 times in the last 83 million years. The last full polarity flip was 780,000 years ago. Magnetic north is now drifting toward Siberia at 50 km per year, the fastest rate since measurements began in 1831, and the field has lost about 5% of its strength per century for the last few hundred years. Whether this is the early movement of another excursion or noise inside normal variability is one of the more consequential open questions in Earth science.

How the field generates itself

The field comes from the geodynamo: convecting liquid iron in the outer core, between 2,900 and 5,100 km depth, drives electrical currents, which sustain a magnetic field, which shapes the flow that creates it. At large scales the result is dipolar — a tilted bar magnet about 11° off the rotation axis — with messy local structure layered on top.

The system is chaotic in the strict mathematical sense. Tiny perturbations in core flow can amplify into large-scale reorganizations. That is why reversals follow no clock. Intervals between them range from tens of thousands to tens of millions of years. The planet is not "overdue" for one in any statistical sense — the process has no memory.

Key numbers

Parameter	Value
Last full reversal	780,000 years ago (Brunhes–Matuyama boundary)
Reversals in the last 83 million years	At least 183
Typical duration of a reversal	1,000–10,000 years
Current field weakening	~5% per century
Magnetic north drift rate	~50 km/year toward Siberia
Field strength during Laschamps	5–10% of normal
South Atlantic Anomaly field strength	~22,000 nT vs. global average ~50,000 nT

Excursions vs. full reversals

A full reversal persists for over 100,000 years and is recorded in rocks worldwide. An excursion is a near-reversal that recovers: the field weakens, the poles wander, sometimes briefly flip, then the original polarity returns. Excursions are more frequent and shorter — but the planet sees the same physics during them.

The major recent excursions:

Norwegian–Greenland Sea event (~64,500 years ago)
Laschamps (~42,200–41,500 years ago)
Mono Lake (~34,500 years ago)
Gothenburg (~11,000 years ago)

Laschamps: the best-studied collapse

Laschamps is the one we know best because it left fingerprints across the cryosphere, the trees, and the cave sediments. During roughly 300 years around 42,000 BP, atmospheric ¹⁴C and ¹⁰Be spiked sharply — both produced when cosmic rays hit the upper atmosphere. Ice cores from Greenland and Antarctica register the same event. Modelling work published in Science in 2021 (Cooper et al.) reconstructed the consequences as a global environmental shock: ozone depletion of a few percent globally with larger polar losses, elevated surface UV-B, and aurorae visible at equatorial latitudes.

A 2025 reconstruction in Science Advances mapped the auroral oval's migration during the event and supports the equatorial-aurora picture. The same window of years coincides with the final disappearance of Neanderthals from Europe (~40,000 BP) and with cave-art production at Chauvet and elsewhere — possibly evidence of behavioural adaptation (more time underground, ochre as a UV shield) by Homo sapiens.

What's contested

The Laschamps–Neanderthal link is the headline claim and the most disputed part of the literature. The objections are serious. The atmosphere absorbs most cosmic rays even at 5–10% field strength, so the dose increase at ground level is modest in absolute terms. A 2022 search for a population bottleneck signature in Neanderthal ancient-DNA timed to Laschamps found nothing crisp; the extinction trajectory looks gradual on either side of the event. Concurrent climate stress — Heinrich Event 4, Grand Solar Minimum, glacial advance — provides plenty of independent explanations.

The Gothenburg excursion sits on the same kind of edge. Its isotopic spike overlaps the late-Pleistocene megafauna extinction, when ~72% of large-bodied mammal species disappeared from North America and Australia. The temporal coincidence is real. Causation versus the Younger Dryas cold snap versus human hunting pressure remains unresolved.

The deeper point: across 183 reversals, no mass extinction in the geological record has been firmly tied to a geomagnetic event. Life has soaked them up. If reversals were catastrophic, the Phanerozoic would look very different.

What a reversal would do to a wired planet

The process unfolds over millennia, not weeks. That is good news for biology and difficult news for technology, because the stress is sustained.

During the weakening phase, satellites already feel it — the South Atlantic Anomaly, a regional patch where field strength has dropped to roughly 22,000 nT, forces operators to power down sensitive electronics on every orbital pass through it. The Hubble Space Telescope does not collect data inside the anomaly. Geomagnetically induced currents in long transmission lines grow more common; the 1989 Quebec blackout, caused by a solar storm hitting a weak-shield grid, is the closest analog for what a weaker baseline field would make routine. Aviation crews and passengers absorb more cosmic radiation at altitude. GPS accuracy degrades. Auroral displays migrate equatorward.

If the field falls to Laschamps levels, ozone losses of a few percent globally and larger losses at the poles become plausible. UV-B rises. The biosphere has handled this before. A grid full of transformers and a constellation of satellites costing hundreds of billions of dollars has not.

Why this has to do with other realms

A geomagnetic reversal is the closest natural experiment Earth runs on the radiation environment of concept-relativistic-travel. Mars-bound and interstellar-precursor crews leave the magnetosphere entirely; what they face for years, the surface biosphere faced for a few centuries at Laschamps. The fossil record during reversals — what survived, what bred, what migrated where — is the largest dataset humans have on how complex life copes with sustained ionizing radiation. The organisms that shrug it off, like concept-tardigrades with their Dsup proteins, did not evolve those tricks for human convenience; they may have been refined, in part, during exactly these field-collapse windows. Radiation biology and paleomagnetism, usually treated as separate fields, share the same archive.

An open question

Whales, sea turtles, and migratory birds navigate by the field. The magnetic north pole has drifted faster in the last forty years than at any time on record. Is there already a detectable signal — strandings, mistimed migrations, breeding-ground misses — and if not, what does that tell us about the redundancy of biological magnetoreception?

Key sources

Cooper, A. et al. (2021) "A global environmental crisis 42,000 years ago," Science — the Laschamps reconstruction paper that triggered the modern debate.
Singer, B. S. (2014) "A Quaternary geomagnetic instability time scale," Quaternary Geochronology — the inventory of excursions and full reversals across the last 2.6 million years.
to verify: 2025 Science Advances paper reconstructing auroral oval migration during Laschamps.
to verify: 2024 Nature Scientific Reports paper proposing core–mantle boundary triggering for regional excursions.
Glatzmaier, G. & Roberts, P. H. (1995) "A three-dimensional self-consistent computer simulation of a geomagnetic field reversal," Nature — the first numerical geodynamo to spontaneously reverse, still the canonical reference.

Abhishek