Earth's magnetic field has flipped hundreds of times, the last reversal was around 780,000 years ago, and the field has been weakening for the past two centuries

Earth’s magnetic field is not fixed. Over the planet’s history it has reversed polarity hundreds of times, swapping magnetic north and south. The most recent full reversal was around 780,000 years ago. And for the past two centuries, since people began measuring the field precisely, it has been slowly getting weaker.

All three of those statements are true, and they tend to get bundled into a single alarming one: that the field is about to flip. That last step does not follow. The weakening is real and worth watching, but it is not a countdown.

How we know the field has flipped

The evidence is written into rock. When molten material cools, magnetic minerals inside it lock into the direction of the field at that moment, like tiny frozen compass needles. Lava flows, and the seafloor spreading out from mid-ocean ridges, preserve a running record of which way the field pointed as they formed.

Read across the seafloor, that record shows a pattern of stripes, alternating normal and reversed polarity, that matches on both sides of a ridge. Piecing those stripes together gives a timeline of reversals stretching back tens of millions of years, and it contains several hundred of them. The switches do not come on a schedule. Sometimes the field holds one polarity for tens of millions of years, sometimes it flips again within a few hundred thousand.

The last complete reversal, from the earlier Matuyama period of reversed polarity to the present Brunhes period, is the reference point. It is usually dated to about 780,000 years ago, though recent work using a global compilation of rock records, published by Mahgoub and colleagues in the Journal of Geophysical Research in 2023, places it closer to 773,000 years ago.

What a reversal actually is

A flip is not a switch being thrown. In the rock record a reversal is a drawn-out affair, with the field growing weak and disorganised, the poles wandering, and more than two magnetic poles sometimes appearing at once, before the field settles into the opposite orientation.

Estimates of how long that takes vary a great deal, from perhaps a few thousand years to more than twenty thousand, and the apparent duration depends on where on Earth you measure it. What the record does not show is a clean overnight swap. Whatever happened 780,000 years ago unfolded over a span far longer than all of recorded human history.

The weakening we can measure

Direct measurement of the field’s strength goes back to the 1830s, when Carl Friedrich Gauss set up a way to record it in absolute terms. Since then the global average strength of the field has fallen by roughly nine per cent.

Most of that loss is concentrated in one place. A broad region stretching between South America and southern Africa, called the South Atlantic Anomaly, is markedly weaker than the rest of the field, and it has been growing. According to the European Space Agency, whose three Swarm satellites have tracked the field from orbit since 2013, the minimum strength in the anomaly dropped from about 24,000 nanoteslas to around 22,000 between 1970 and 2020, while the affected area widened and drifted west at roughly 20 kilometres a year. A second weak centre has been developing southwest of Africa, raising the possibility that the anomaly is splitting in two. More recent Swarm analysis, published in 2025, finds the anomaly has kept expanding.

The anomaly is not just a curiosity. It is where the field offers satellites the least shielding from charged particles, and spacecraft passing through it face a higher rate of radiation glitches, which is one practical reason the region is watched so closely.

Why weaker does not mean about to flip

Here the caution has to be plain, because this is where the science is most often overstated.

A weakening field is one of the things that happens before a reversal. It is also something that happens often without any reversal following. The field has dipped and recovered many times, including a sharp, short-lived wobble known as the Laschamp excursion about 41,000 years ago, when the field weakened severely and then bounced back to its normal orientation rather than flipping. The present strength, even after two centuries of decline, still sits within the range the field has occupied over the recent geological past.

There is also no reliable way to forecast a reversal from where we stand. The processes that drive the field, the churning of molten iron in Earth’s outer core, are chaotic, and the record shows no regular interval to count down from. The often-repeated line that we are overdue misreads an average as a schedule.

It is worth adding that reversals in the deep past are not clearly tied to mass extinctions or to any sudden catastrophe for life. The field would not vanish during a reversal; it would grow weaker and more tangled for a time, then reorganise.

What the next years of Swarm data should clarify is not when a flip will come, but how the South Atlantic Anomaly behaves, whether it keeps growing, and whether it settles into two cells or one. That is the measurable question. The rest, for now, stays in the rock.