The geodynamo is the self-sustaining process by which Earth generates its magnetic field, driven by the churning motion of liquid iron in the planet’s outer core. It works a bit like a giant electrical generator, except instead of spinning copper wire through a magnet, convecting streams of molten metal create and maintain the very magnetic field that, in turn, shapes how the metal flows. The result is the invisible shield that deflects most of the solar wind, makes compass needles point north, and has done so for billions of years. How it actually works, what keeps it going, and why it occasionally flips have kept geophysicists busy for decades, and the answers are still sharpening.
How Liquid Iron Becomes a Magnet
Earth’s outer core is a roughly 2,200-kilometer-thick shell of liquid iron alloy sitting between the solid inner core and the rocky mantle above. That liquid is an excellent electrical conductor and it is in constant motion, driven upward by heat and by lighter elements released as the inner core freezes. The motion of an electrically conducting fluid through even a weak seed magnetic field generates electric currents, and those currents produce their own magnetic field, which then influences the fluid’s motion, which generates more current, and so on. When conditions are right, this feedback loop becomes self-sustaining: the field regenerates itself faster than it decays. That is the dynamo mechanism in a nutshell.
Two forces dominate the physics inside the core. The Coriolis effect, a consequence of Earth’s rotation, organizes the convection into tall columns roughly aligned with the spin axis. The Lorentz force, the push that a magnetic field exerts on a moving electrical conductor, would separately tend to resist the flow. But here is the counterintuitive part: acting together, the magnetic field actually relaxes the rigid constraint that rapid rotation imposes on the fluid, allowing larger-scale flows to develop, which in turn regenerate the field more effectively.1Annual Review of Fluid Mechanics. Rapidly Rotating Magnetohydrodynamics and the Geodynamo The rotation and the magnetism need each other. Without spin, the convection would be turbulent but disorganized; without a conducting fluid, there would be nothing to generate currents. The interplay is what makes Earth special among the rocky planets.
What Keeps the Dynamo Running
A self-sustaining dynamo still needs energy. Two main sources feed the outer core’s convection. The first is heat flowing outward from the core into the mantle, some of it leftover from planetary formation and some generated by the slow crystallization of the inner core. The second, and today the more important source, is compositional buoyancy: as the inner core grows, it preferentially incorporates iron and rejects lighter elements like silicon, oxygen, and sulfur into the surrounding liquid. Those lighter parcels of fluid are buoyant and rise, stirring the outer core and doing the work of dynamo generation.
The balance between these two energy sources depends on a property that has been surprisingly hard to pin down: the thermal conductivity of the core alloy. If conductivity is very high, heat escapes by simple conduction rather than by convection, leaving less energy available for the dynamo. Recent experiments on an iron-nickel-silicon alloy at core-like pressures measured a moderate thermal conductivity of about 73 watts per meter per kelvin. That value implies the conductive heat loss at the core-mantle boundary is roughly 11 terawatts, which is comfortably below the estimated total heat flow of about 15 terawatts leaving the core today. The difference means there is enough excess heat to drive thermal convection on top of the compositional stirring, so both engines are running at once.2Geophysical Research Letters. Moderate Thermal Conductivity of Fe‐Ni‐Si Alloy at Earth’s Core Conditions: Implications for Core Thermal Evolution and Geodynamo
One long-running question was how much total power the dynamo dissipates. Estimates from numerical simulations and laboratory experiments suggest the ohmic dissipation, the energy lost as electrical resistance heats the core fluid, falls in the range of roughly 0.2 to 0.5 terawatts.3PubMed. Power requirement of the geodynamo from ohmic losses in numerical and laboratory dynamos That is modest enough that the dynamo does not require large amounts of radioactive potassium or other heat-producing elements dissolved in the core, a scenario that had been proposed but never confirmed geochemically.
Four Billion Years Old, at Least
The geodynamo is ancient. Paleomagnetic evidence from tiny magnetic mineral inclusions trapped inside zircon crystals from the Jack Hills in Western Australia indicates that a geomagnetic field existed at least 4.2 billion years ago, only a few hundred million years after the Moon-forming impact.4PubMed Central. Earth’s magnetic field and its relationship to the origin of life, evolution and planetary habitability Those zircons lock in a snapshot of the ambient magnetic field when they crystallize, and detailed analyses confirm that the magnetization in many of them is primary, meaning it was recorded when the crystal formed rather than overprinted later.5PubMed Central. Paleomagnetism indicates that primary magnetite in zircon records a strong Hadean geodynamo
Measurements from zircons spanning 3.3 to 4.2 billion years old show field strengths ranging between about 12 percent and 100 percent of today’s equatorial field, suggesting the early dynamo was not a feeble affair but sometimes comparable to the present one.6PubMed. A Hadean to Paleoarchean geodynamo recorded by single zircon crystals A period of particularly strong field around 4.1 to 4.0 billion years ago may reflect vigorous convection in the early core, possibly driven by chemical precipitation in a liquid iron core that had not yet begun to solidify at the center.7PubMed Central. Paleomagnetism indicates that primary magnetite in zircon records a strong Hadean geodynamo This is a striking finding. It means Earth had a magnetic shield almost from the start, which has implications for how the early atmosphere was retained and how conditions became hospitable for life.
When the Inner Core Was Born
The solid inner core has not always existed. It nucleated at some point as the core cooled enough for iron to start freezing at the center, where pressure is highest. Pinning down when that happened is one of the more active puzzles in deep-Earth science, with estimates ranging from under a billion years ago to over two billion years.
The moderate thermal conductivity measurements mentioned earlier point to a maximum inner core age of roughly 0.8 to 1.3 billion years.8Geophysical Research Letters. Moderate Thermal Conductivity of Fe‐Ni‐Si Alloy at Earth’s Core Conditions: Implications for Core Thermal Evolution and Geodynamo That bracket is consistent with a separate line of evidence from paleomagnetic records that show an extremely weak geomagnetic field during the Ediacaran period, around 565 million years ago, followed by a rapid recovery to something close to modern field strength within about 30 million years. Researchers interpret that recovery as the signature of inner core nucleation switching on the powerful new compositional energy source. Paleointensity data place the onset at about 550 million years ago, which, given certain thermal conductivity assumptions, fits models with conductivities of 89 to 121 watts per meter per kelvin.9Nature Communications. Early Cambrian renewal of the geodynamo and the origin of inner core structure
Numerical simulations that evolve the dynamo over billions of years sketch a plausible narrative: a multipolar field that shifted to a dipole-dominated state around 1.7 billion years ago, weakened again around 1 billion years ago, then recovered to a strong dipole after inner core nucleation around 650 million years ago, a state maintained to the present day.10Geophysical Research Letters. Simulating 2 Ga of geodynamo history Not everyone agrees on the timeline. Some modeling work suggests the dynamo stayed dipole-dominated throughout, with inner core nucleation leaving only a subtle imprint, such as a gradual change in the shape of the field rather than a dramatic jump in strength.11Earth and Planetary Science Letters. The signature of inner-core nucleation on the geodynamo The disagreement comes down to assumptions about how efficiently the early core convected before the inner core existed, which feeds back into those stubbornly uncertain thermal conductivity values.
Why the Field Flips
The geomagnetic field has reversed its polarity hundreds of times over the past several hundred million years. North becomes south and south becomes north, usually over a transition period of a few thousand years, after which the new orientation persists for anywhere from tens of thousands to tens of millions of years. The pattern looks random: there is no strict periodicity, and the lengths of stable intervals follow a roughly statistical distribution.
Numerical dynamo models have shed light on what triggers a flip. In simulations, patches of reversed magnetic flux are generated locally in rising convective plumes inside the core. When enough of these reversed-flux patches accumulate and get spread across the core-mantle boundary by the large-scale meridional circulation, the overall dipole can tip over. The actual polarity switch, as seen from the surface, takes roughly as long as the meridional transport time, on the order of a few thousand years.12Geochemistry, Geophysics, Geosystems. A detailed study of the polarity reversal mechanism in a numerical dynamo model One surprising result from these models is that the reversals are not necessarily triggered by sudden changes in the flow. They can arise as a purely magnetic induction effect: the balance between normal-polarity and reversed-polarity field generation drifts until the reversed component wins.
Whether reversal frequency changes over geological time is another open question. Models that incorporate the slow evolution of the core, including the growth of the inner core and the gradual deceleration of Earth’s rotation due to tidal friction, show that inner core growth rate can influence how often the field flips. A simulation starting in a non-reversing state and subjected to an increasing inner core growth rate evolved over 20 million model years into a reversing state, with substantial trends in dipole intensity and the length of stable polarity intervals.13Earth and Planetary Science Letters. Polarity reversals in geodynamo models with core evolution
Excursions and the Inner Core’s Magnetic Memory
Not every wobble of the field amounts to a full reversal. Geomagnetic excursions are events in which the field direction swings dramatically, sometimes briefly reaching the opposite polarity, but then snaps back to its original orientation. These are more common than complete reversals and tend to be shorter-lived.
A compelling explanation involves the solid inner core acting as a kind of magnetic flywheel. Because the inner core is also an electrical conductor, it carries its own slowly changing magnetic field, which can only evolve by diffusion on a timescale of about 3,000 years. The liquid outer core, in contrast, can change its field in 500 years or less. During an excursion, the outer core’s field may briefly reverse, but the inner core’s field has not had time to follow. That “magnetic inertia” from the inner core tends to pull the outer core field back to its original polarity before a complete reversal can take hold.14Geophysical Journal International. The distinction between geomagnetic excursions and reversals A full reversal, then, requires the outer core to sustain the reversed state long enough for the inner core’s field to catch up by diffusion. Excursions are failed reversals, in a sense, aborted by the stabilizing influence of the solid center.
Jerks, Waves, and What We Can See From the Surface
The geomagnetic field does not change smoothly between reversals. Ground observatories and satellites detect abrupt kinks in the rate of change of the field, called geomagnetic jerks, that happen every few years. These are not slow drifts but sudden shifts in the field’s acceleration that show up clearly in time series of observatory data.
Recent work attributes these jerks to rapid waves propagating through the outer core. When the leading-order force balance in the core is locally disrupted, the system compensates by emitting fast hydromagnetic waves, a type of wave that travels through the magnetized fluid. These waves carry the disturbance outward and produce the jerk signal observed at the surface. The evidence points to a single underlying physical mechanism for all observed jerks: disruption of the dominant force balance followed by the emission of magneto-inertial waves.15Geophysical Journal International. A taxonomy of simulated geomagnetic jerks
Satellite magnetic data have also revealed wavelike patterns at the core surface with interannual periods near seven years, amplitudes of about 3 kilometers per year, and coherent westward drift at phase speeds around 1,500 kilometers per year.16Proceedings of the National Academy of Sciences (PNAS). Satellite magnetic data reveal interannual waves in Earth’s core These are interpreted as Magneto-Coriolis eigenmodes: oscillations governed jointly by the magnetic field and Earth’s rotation. Detecting these waves is a breakthrough because their properties carry information about the field strength and flow structure deep inside the core, regions we cannot sample directly. Simulations that push toward realistic core conditions confirm that such waves should be detectable in geomagnetic observations, lending confidence that these signals are genuine glimpses of core dynamics rather than artifacts.17Geophysical Journal International. The interplay of fast waves and slow convection in geodynamo simulations nearing Earth’s core conditions
The Challenge of Simulating the Real Core
Numerical simulations of the geodynamo have come a long way since the first self-sustaining models appeared in the mid-1990s.18PubMed. Earth’s core and the geodynamo Yet a fundamental gap persists. Earth’s outer core is an extraordinarily low-viscosity fluid: the ratio of viscous forces to rotational forces is fantastically small, far smaller than any computer can currently resolve. As a result, every simulation inflates the viscosity by many orders of magnitude to keep the calculation tractable.
This has real consequences. In typical models, viscous and inertial forces end up being almost as important as the Coriolis force, and the magnetic Lorentz force plays a smaller role than it should. That raised a legitimate concern: maybe these simulations are controlled by their artificially high viscosity rather than by the physics that actually governs the core. Progress has come from running models at increasingly extreme parameters, steadily pushing toward a force balance where magnetic, rotational, and buoyancy forces dominate while viscosity becomes negligible, which is the regime Earth’s core is thought to occupy.19PubMed Central. Approaching a realistic force balance in geodynamo simulations The most recent generation of models is getting close, but “close” still means many orders of magnitude away from true core values. Every finding from these simulations comes with that caveat.
Laboratory experiments offer a complementary approach. The von Kármán sodium experiment demonstrated dynamo action in a turbulent flow of liquid sodium, producing a self-sustaining magnetic field from fluid motion alone.20PubMed. Generation of a magnetic field by dynamo action in a turbulent flow of liquid sodium Liquid sodium is the best laboratory analogue for liquid iron because it is both an excellent electrical conductor and a fluid that can be driven to high turbulence. These experiments cannot replicate core conditions, but they validate the basic dynamo mechanism in a physical system rather than a purely computational one, which matters for testing whether the simulations are capturing the right physics.
The South Atlantic Anomaly
If you have ever wondered why satellites sometimes glitch over Brazil, the answer traces back to the geodynamo. The South Atlantic Anomaly is a region where Earth’s surface magnetic field is unusually weak, roughly centered over South America and the southern Atlantic Ocean. Charged particles from the solar wind dip closer to the surface there, increasing radiation exposure for spacecraft and occasionally disrupting electronics on low-orbit satellites.
Dynamo simulations can reproduce the longitude of the anomaly fairly well when they impose a heat-flux pattern at the core-mantle boundary that mirrors large-scale structures inferred from seismic imaging of the lowermost mantle. But getting the latitude right has proven harder. Models show that equatorially asymmetric heat-flux patterns are needed to push the surface intensity minimum to higher latitudes matching the observed anomaly. The catch is that these asymmetric patterns tend to produce dynamos that reverse more often and lose their dipole dominance.21Journal of Geophysical Research: Solid Earth. Necessary Core‐Mantle Boundary Heat Flux Patterns for Recovering the Latitude of the South Atlantic Anomaly This is a recurring theme in geodynamo research: the mantle’s thermal structure influences what the dynamo can do, because the mantle controls where heat leaves the core. The relationship between mantle convection and the magnetic field links plate tectonics to geomagnetic behavior over timescales of hundreds of millions of years.22Geophysical Journal International. Changes in core–mantle boundary heat flux patterns throughout the supercontinent cycle
Dynamos Across the Solar System
Earth is not the only body running a dynamo. Jupiter, Saturn, and the other gas giants all have active magnetic fields, generated not in liquid iron but in layers of metallic hydrogen deep inside the planet. Jupiter’s field is the strongest in the solar system, and simulations suggest it is shaped partly by a stably stratified layer in the planet’s upper interior, between about 0.9 and 0.95 Jupiter radii. That layer acts as a filter, smoothing out smaller-scale field features and helping to explain why Jupiter’s field appears strongly dipolar at the surface despite the complex convection underneath.23Journal of Geophysical Research: Planets. Dynamo Simulations of Jupiter’s Magnetic Field: The Role of Stable Stratification and a Dilute Core
Among the rocky planets, the situation is more varied. Mercury, despite being small enough that you might expect its core to have cooled and solidified long ago, retains a weak but active dynamo. Venus, nearly Earth’s twin in size, has no detectable global magnetic field at all. The likeliest explanation is that Venus lacks the plate tectonics that help Earth’s mantle extract heat efficiently from the core. Without that heat extraction, the core may not convect vigorously enough to sustain a dynamo, or it may never have developed an inner core whose crystallization could drive compositional convection.24Advances in Space Research. The magnetic fields of Mercury, Venus and Mars Mars once had a dynamo, as crustal magnetism in its ancient southern highlands attests, but it shut down roughly four billion years ago. The comparison between these worlds is a reminder that having a liquid iron core is necessary but not sufficient: you also need the right thermal and compositional conditions to keep the fluid moving.
Listening for Exoplanetary Dynamos
Beyond our solar system, detecting a magnetic field around an exoplanet would be a significant find because it would reveal information about the planet’s interior structure and its potential to retain an atmosphere over geological time. The most promising detection method involves auroral radio emissions. When a planet with a magnetic field interacts with its host star’s wind, charged particles are funneled toward the magnetic poles and emit radio waves at specific frequencies tied to the field strength. Ground-based radio arrays are now sensitive enough to search for these signals, and tentative detections of circularly polarized radio bursts from systems like HD 189733 have been reported, though none has been confirmed beyond doubt.25Astronomy & Astrophysics. Tentative detection of circularly polarized bursty radio emissions from the HD 189733 exoplanetary system using NenuFAR beamformed observations If confirmed, such detections would extend geodynamo science from a handful of solar system bodies to potentially thousands of worlds, turning a problem once confined to one planet’s deep interior into a tool for comparative planetology on a galactic scale.