Does Earth's N and S pole really swap?
― A cliff in Chiba records a swap from 770,000 years ago
It's true. Earth, as a giant magnet, has swapped its north-south direction hundreds of times over its long history. The last swap happened about 770,000 years ago. If you'd held a compass back then, its N end would have pointed south. That's because Earth's magnetism isn't made by a solid iron bar — it's made by liquid iron that keeps flowing deep underground.
You've probably held a compass in your palm while hiking or in a map class. The needle wobbles a little, then settles in the same direction every time.
We treat that direction as something fixed. Ships and migrating birds alike have relied on it for thousands of years.
But rewind the clock a few hundred thousand years, and you find many eras when the needle pointed exactly the opposite way from today.
Only two facts explain both the flip and how we know about it
Between roughly 2,900 and 5,100 km underground lies a layer of molten iron. As this iron slowly swirls and flows, it generates electric currents, forming a giant electromagnet. When the flow gets disturbed, the magnet's direction can flip too.
When lava cools and solidifies, the tiny iron grains inside it line up with Earth's magnetic field at that moment and freeze in place. The rock keeps recording the "north" of its formation era for millions of years.
The first fact answers "why does it flip?" The second answers "how do we know about the past?" Let's look at each in turn.
Why does Earth's magnet flip?
Science textbooks often show a picture of a giant bar magnet inside Earth. But near Earth's center, temperatures reach several thousand degrees Celsius. An iron bar magnet loses its magnetism when heated. So there isn't a solid magnet sitting inside Earth.
What works instead is the flow of liquid iron. When something that conducts electricity moves through a magnetic field, it generates a current. That current then creates its own magnetic field. Earth's magnetic field is thought to be sustained by this kind of self-perpetuating "generator."
The flow keeps shifting bit by bit, much like weather patterns. It's usually aligned in one direction, but sometimes it gets badly disturbed and the field weakens. When the disturbance settles, the flow can realign in the opposite direction. That's a geomagnetic reversal.
Move the slider in Figure 1 to see the field's direction over the past 2.6 million years. You'll see today's direction and the reversed direction mixed together in a patchwork.
The intervals between flips are far from regular. Some reversed within tens of thousands of years, while other eras reportedly kept the same direction for tens of millions of years. Averaged over the last few million years, it's roughly once every few hundred thousand years.
How do we know "north" from hundreds of thousands of years ago?
The clue is a mineral called magnetite found in rock. Inside hot lava, magnetite grains have lost their magnetic properties. Once it cools below about 580°C, the grains align with Earth's magnetic field at that moment and become magnetized. Once set, that direction stays locked in for millions of years.
One of the first people to notice this "memory in rock" was Motonori Matuyama of Kyoto Imperial University. In the late 1920s, he examined rocks such as those at Genbudo in Hyogo Prefecture and reported finding rocks magnetized in the opposite direction from today. Many doubted him at the time, but the reversed era is now named "Matuyama" in his honor.
The clincher came from the seafloor. On the ocean floor there are mountain ranges called mid-ocean ridges, where lava wells up from Earth's interior and spreads outward on both sides. The emerging lava records the direction at the time and is carried outward on both sides like a conveyor belt. As in Figure 2, this creates a stripe pattern that's symmetric on either side of the ridge.
In the 1960s, this symmetric pattern was actually found on the real seafloor. It was the moment that confirmed, at once, both that geomagnetic reversals had occurred and that the seafloor was spreading.
Along the Yoro River in Ichihara City, Chiba Prefecture, a cliff preserves the flip from about 770,000 years ago exceptionally well within its rock layers. In 2020, this layer was selected as an international reference point, and the era from roughly 770,000 to 130,000 years ago was named the "Chibanian" (meaning "Chiba era"). It was the first time a geologic era was named after a place in Japan.
During a flip, the magnetic field is thought to weaken over roughly a few thousand years. Particles arriving from space would increase, which might affect satellites and power grids. On the other hand, our human ancestors have already lived through many flips, and no evidence has been found linking any of them to mass extinctions.
Summary
Earth's magnetism is a constantly moving magnet, sustained by the flow of liquid iron deep underground. That's why its direction occasionally flips — the last time was about 770,000 years ago. The record of this gets locked into rock when lava cools and solidifies, and we can read it both on the seafloor and in the cliffs of Chiba. A compass pointing north is, in the long history of Earth, simply "how things happen to be right now."
The compass's "north" isn't an eternal rule.
It just reflects the current state of an Earth that keeps on flowing.
You can read more about how a compass needle points north in "Why does a compass needle point north?", and about Earth's core, which generates the field, in "Earth's center is as hot as the Sun's surface — so why isn't it molten?" The gap between map north and magnetic north is covered in "Why does a compass's north differ from the map's north?"
- Stroke a sewing needle about 20 times in one direction, from base to tip, using the same pole of a magnet. Place the needle on a cork floating in water, and it will point north-south like a compass.
- Mark which end pointed north.
- Now stroke it again in one direction using the opposite pole of the magnet, then float it again. If the opposite end now points north, the magnetic direction inside the needle has flipped.
The way the tiny magnets inside the needle align or flip is similar to how magnetite in rock records its direction. Be careful not to prick your fingers when handling the needle.
For those who want to know more ― terms, formulas, and textbook linksShows clearly which level each part belongs to, from junior-high science to university specialist courses
- JHSCovered in junior high school science
- HSCovered in high school "Physics" or "Earth Science"
- HS+Advanced high school content, or textbook sidebar material
- Univ.Not taught in high school — university-level specialist content (geophysics, paleomagnetism)
- ResearchNot even settled as "established theory" at university — an active research question
JHSTerminology: this phenomenon has names
- Geomagnetism: the magnetic field Earth produces. It's why a compass needle points north-south.
- Geomagnetic reversal: when the geomagnetic field's direction flips north-south. An era matching today's direction is called a "normal polarity period," and a reversed one a "reversed polarity period."
- Mid-ocean ridge: a mountain range on the seafloor where lava wells up from Earth's interior and the seafloor spreads outward on both sides.
JHSHSCheck with a formula: how far has the seafloor spread since the last flip?
The white band in the center of Figure 2 is rock that formed at the ridge from about 770,000 years ago to today. Let's find the width of one side using the seafloor's spreading speed.
| In symbols | L ＝ v × t |
| In words | Stripe width ＝ speed of seafloor spreading on one side × duration the direction stayed the same |
| Where this comes from | Rock formed at the ridge is carried outward on both sides at roughly constant speed. It's just "distance = speed × time" from junior high. Since rock records the field direction the instant it forms, the length of a same-direction stretch becomes the width of one stripe |
| Symbol | Meaning and unit |
| L | Width of one side of a stripe (same-direction band) 〔cm〕 |
| v | Speed of seafloor spreading on one side 〔cm/year〕 |
| t | Duration the direction stayed the same 〔years〕 |
| Speed of one-side seafloor spreading (varies 1 to several cm by ridge; used here as an example) | about 2cm/year |
| Time since the last flip | about 770,000 years |
| Change in geomagnetic field strength (since observations began in the early 1800s) | reportedly weakened by just under 10% over about 200 years |
| Stripe width in cm | 2 × 770000 ＝ 1540000 cm |
| Converted to km (1km＝100000cm) | 1540000 ÷ 100000 ＝ 15.4 km |
| Time to walk across at 4km/h | 15.4 ÷ 4 ≒ 3.9 hours |
| Reference: if it dropped 10% each time, how many steps to reach zero? | 100 ÷ 10 ＝ 10 times |
| Reference: that many years (assuming today's pace continues) | 10 × 200 ＝ 2000 years |
Over 770,000 years, the seafloor spreads only about 15km. Even at a fingernail-growth pace, that adds up over hundreds of thousands of years into stripe patterns large enough to measure by ship. Note the last two rows are rough guides from "simply extrapolating" — actual geomagnetic strength repeatedly rises and falls, so they're not a prediction of the next reversal.
HSHS+Current creates a field, and motion within a field creates current
HSHigh school physics teaches that a current creates a magnetic field around it, and that a conductor moving through a magnetic field generates a voltage (electromagnetic induction). In Earth's outer core, liquid iron convects while being twisted by Earth's rotation, and both effects happen at once. In high school earth science, the seafloor's magnetic stripe pattern is treated as evidence for plate tectonics.
HS+Magnetite loses its magnetic properties above about 580°C (the Curie temperature), and becomes magnetized in the field's direction at the time as it cools. This is called thermoremanent magnetization. There's also depositional remanent magnetization, where fine settling grains align their direction — this is what the Chiba rock layer uses.
Univ.Dynamo theory and proof of seafloor spreading
The mechanism by which a flowing liquid metal sustains its own magnetic field is studied as dynamo theory. Computer simulations have reproduced spontaneous field reversals arising from the combination of outer-core convection and the Coriolis force. The field of reading past geomagnetism from rocks is called paleomagnetism, and the last reversal is called the Matuyama-Brunhes boundary. The explanation linking seafloor stripes to reversals is known as the 1963 Vine-Matthews hypothesis.
📖 For the derivation of the formulas and further reading: Dynamo theory (Japanese Wikipedia) ／ Geomagnetic reversal (Japanese Wikipedia)
ResearchWhat's still not fully understood
- When will the next flip happen? The geomagnetic field has been weakening for about 200 years, but whether this is the start of a reversal or just a fluctuation remains unknown.
- What happens during a flip? How weak the field gets, what shape the disturbance takes, and how many years it takes all vary across records, and research continues.
- Why are the intervals so irregular? The eras lasting tens of millions of years without a reversal may be related to temperature distribution deep inside Earth, but this isn't settled.
In other words, even this article describes things "as currently understood." No one has ever been to Earth's center — our only clues are the record in rocks and the field measured at the surface.
Links to textbooks (by level)
| Level | Subject/Unit | Where in this article |
|---|---|---|
| JHS | Science: magnetic fields, changes to the land | Compass, distance＝speed×time |
| HS | Physics: current and magnetic fields, electromagnetic induction／Earth Science: plate tectonics | The electromagnet made by liquid iron, seafloor stripe pattern |
| HS+ | Advanced earth science: rock magnetization | Curie temperature, thermoremanent magnetization |
| Univ. | Geophysics / Paleomagnetism | Dynamo theory, Matuyama-Brunhes boundary |
| Research | Research into Earth's deep interior | Timing of the next reversal, what happens mid-reversal |
| ― | Connection to daily life | Compass reliability, the Chibanian |
- Chibanian (Japanese Wikipedia)
- Vine, F. J. & Matthews, D. H. (1963) Magnetic Anomalies Over Oceanic Ridges. Nature 199
- Matuyama, M. (1929) On the Direction of Magnetisation of Basalt in Japan, Tyôsen and Manchuria. Proceedings of the Imperial Academy 5
- Japan Meteorological Agency, Kakioka Magnetic Observatory (気象庁 地磁気観測所)
※This article is a general-audience science explainer. The figures given are approximations meant to aid understanding. The dating of past reversals can vary by several thousand years depending on the study.