# 7. The Geodynamo: Core–Magnetic Field Interaction Earth's magnetic field originates entirely within the outer core, through a self-sustaining process called the geodynamo. Because this process is central to how the core governs conditions at the surface, this section treats it in depth: what physically must happen for a dynamo to exist, what powers it, how rotation shapes it, how it behaves at the surface, and what it reveals about Earth's deep past. ## 7.1 What a Dynamo Requires A magnetic dynamo is, at its core, a competition between two effects acting on a moving, electrically conductive fluid: field generation through fluid motion, and field decay through electrical resistance (ohmic diffusion). Motion that drags and twists magnetic field lines can amplify them; resistance in the fluid simultaneously bleeds that field away as heat. Whether a dynamo can sustain itself depends on which effect wins, a balance captured by a dimensionless quantity called the magnetic Reynolds number, Rm, which compares the rate of field generation by flow to the rate of field loss by diffusion. - If Rm is too small, diffusion wins and any magnetic field decays away — this is why laboratory-scale conducting fluids do not spontaneously magnetize. - Earth's outer core is large (thousands of kilometers) and flows fast enough that Rm is comfortably above the threshold required for self-sustaining field generation, which is why the geodynamo has persisted for billions of years. This same reasoning explains why a solid inner core, despite being conductive iron, cannot itself sustain a dynamo — it does not flow. The dynamo lives specifically in the liquid outer core. ## 7.2 Energy Sources Powering the Dynamo Sustaining convective motion against viscous and magnetic drag requires a continuous energy supply. Several sources contribute, though their relative importance has shifted as research has progressed: - **Secular cooling:** the core has been slowly losing primordial heat to the mantle since Earth's formation, and this cooling alone drives some thermal convection. - **Compositional (chemical) convection:** as the inner core freezes, it preferentially incorporates iron and nickel, rejecting lighter elements (sulfur, oxygen, silicon) into the overlying liquid. This buoyant, light-element-enriched fluid rises, and current models suggest this compositional buoyancy — not heat alone — is the dominant energy source for the modern dynamo. - **Radiogenic heating:** decay of radioactive isotopes (potassium-40 is the leading candidate) may contribute additional heat if present in the core in sufficient concentration, though this remains debated and is likely a minor contributor. - **Mechanical forcing:** Earth's precession, and to a lesser extent tidal forces, mechanically stir the outer core independent of buoyancy. This is thought to be a secondary contributor, though its magnitude is actively studied. A key implication follows from the compositional-convection mechanism: the dynamo's vigor is tied directly to the growth of the inner core. Before the inner core existed — for roughly 80–90% of Earth's history (the inner core is estimated to have begun forming only half a billion to a billion years ago, against Earth's full 4.5-billion-year age; Section 7.8) — the dynamo would have run on thermal convection alone, likely producing a weaker and possibly less stable field. ## 7.3 Rotation and the Organization of Flow Earth's rotation does not power the dynamo, but it decisively shapes it. The Coriolis force acting on convecting fluid in a rapidly rotating sphere forces flow into an organized pattern rather than disordered turbulence. - Convective motion is drawn into elongated, roughly cylindrical columns aligned parallel to the rotation axis — a tendency described by the Taylor–Proudman theorem, which shows that rotation strongly resists flow variation along the rotation axis. - These columns twist and drift as they convect, and it is this organized, helical motion — not random churning — that reliably stretches and folds magnetic field lines in a way that reinforces rather than cancels the field. - The ratio of inertial to Coriolis forces (the Rossby number) is very small in the outer core, confirming that rotation, not fluid inertia, dominates the flow's large-scale structure. This is why the resulting field is not a random tangle: rotation imposes enough order that the time-averaged field at the surface is dominated by a simple, axis-aligned dipole, closely resembling a bar magnet aligned near Earth's spin axis. ## 7.4 The Self-Sustaining Feedback Loop The core process is described mathematically by the magnetic induction equation, which states that the rate of change of the magnetic field equals a term for field generation by fluid motion minus a term for diffusive decay. Physically, the feedback loop runs as follows: - Convecting, conductive fluid moving across existing magnetic field lines induces electric currents (by the same physics as a wire moving through a magnetic field in a generator). - Those induced currents themselves generate additional magnetic field, according to Ampère's law. - If the fluid motion is organized appropriately — as rotation ensures it is — the newly generated field reinforces the pre-existing field rather than opposing it, so the field is stretched, twisted, and folded into a stronger configuration over time. - This growth continues until it is balanced by ohmic losses, at which point the dynamo settles into a statistically steady (though never perfectly constant) state. Dynamo theorists often decompose this process into an "alpha effect," in which small-scale helical turbulence twists field lines into loops, and an "omega effect," in which large-scale shear stretches field lines into stronger bands. Earth's dynamo is generally modeled as some combination of both mechanisms operating together in the outer core. ## 7.5 Field Morphology and Behavior at the Surface The field produced by this process, once it emerges from the core and passes through the (magnetically passive) mantle and crust, has several well-documented characteristics: - **Dipole dominance:** roughly 90% of the field measured at Earth's surface can be described by a simple dipole, tilted about 10° from the rotation axis — which is why compass north and true geographic north are close, but not identical. - **Secular variation:** the field's intensity and direction drift slowly over years to centuries as the underlying flow pattern in the core evolves. Magnetic north has moved substantially over the last century. - **Westward drift:** large-scale features of the field tend to drift westward relative to the crust at roughly 0.2° of longitude per year, reflecting the pattern of core flow beneath the CMB. - **Regional weak spots:** the South Atlantic Anomaly, a region of unusually low field strength over South America and the South Atlantic, is a present-day expression of more complex, non-dipole structure in the core's output, and is monitored closely because of its effect on satellites passing through the region. ## 7.6 Geomagnetic Reversals and Paleomagnetism Perhaps the most striking behavior of the geodynamo is that it periodically reverses polarity — magnetic north and south trade places — over geologically short intervals (thought to take roughly 1,000 to 10,000 years to complete), separated by much longer stable intervals lasting tens of thousands to millions of years. - The evidence comes from paleomagnetism: as molten rock cools and solidifies (notably basalt erupted at mid-ocean ridges), magnetic minerals within it lock in the direction of the ambient field at that moment, creating a permanent record. - Symmetric, stripe-like patterns of alternating magnetic polarity in oceanic crust on either side of mid-ocean ridges provided some of the strongest confirming evidence for both geomagnetic reversals and plate tectonics itself, since the stripes record seafloor spreading over time like a tape recorder. - The most recent full reversal, the Brunhes–Matuyama reversal, occurred approximately 780,000 years ago. The field has not fully reversed since, though its overall intensity has been declining measurably over the past two centuries — a fact monitored but not, on its own, proof that a reversal is imminent. - Reversals are reproduced in numerical dynamo simulations (notably by Glatzmaier and Roberts in the 1990s) as an emergent, chaotic feature of the underlying fluid dynamics, rather than something requiring an external trigger — supporting the view that reversals are an intrinsic, if irregular, behavior of the dynamo itself. ## 7.7 Magnetic Diffusion in the Core Everything in Sections 7.1–7.6 describes how motion *generates* field. The other half of the induction equation — diffusion — describes how the field decays, and it turns out to be just as important for understanding what the geodynamo actually looks like and why continuous regeneration is necessary at all. **The diffusion timescale.** Magnetic diffusivity, η, is a property of the conducting material equal to 1/(μ₀σ), where σ is electrical conductivity and μ₀ is the magnetic permeability of free space. For outer-core iron alloy, σ is roughly 10⁶ siemens per meter, giving η on the order of 1–2 m²/s. Diffusion time over a length scale L scales as τ ≈ L²/η. Applied to the outer core as a whole (L on the order of a few thousand kilometers), this gives a **free-decay time of roughly 10,000 to 100,000 years** for the large-scale dipole field, with commonly cited estimates clustering around 15,000–50,000 years. **Why this matters:** if all fluid motion in the outer core stopped, Earth's magnetic field would not vanish instantly — it would fade away exponentially over roughly this timescale as the electric currents sustaining it dissipated resistively into heat. Paleomagnetic evidence shows the field has existed continuously for at least 3.5 billion years, a span roughly 100,000 times longer than the free-decay time. That mismatch is one of the clearest pieces of evidence that a **fossil field left over from Earth****'****s formation cannot explain the present field** — something must be actively regenerating it, which is precisely the role of the dynamo described above. **Diffusion is scale-dependent — and this shapes the field we actually observe.** Because the diffusion timescale scales with the *square* of the length scale, small-scale magnetic structures decay far faster than large-scale ones. Fine, tangled field structure generated near the top of the outer core diffuses away in a small fraction of the time it takes the broad, large-scale dipole component to decay. This is a major reason the field measured at Earth's surface is so strongly dominated by simple, low-order structure (Section 7.5): the messy, small-scale detail present right at the core surface is systematically filtered out by diffusion before — and as — it propagates outward, leaving the smooth, large-scale dipole as the most persistent and visible signature. **The frozen-flux approximation.** On the shorter timescales relevant to observed year-to-century secular variation, core fluid moves quickly enough relative to the diffusion rate (equivalently, the magnetic Reynolds number is large) that magnetic field lines behave, to good approximation, as if they are "frozen into" and carried along by the moving fluid — diffusion is simply too slow to act meaningfully over those timescales. This frozen-flux approximation is not just a theoretical convenience: it is the basis for a real observational technique. By tracking how the field measured at Earth's surface (and by orbiting magnetic survey satellites) changes over time, geophysicists can mathematically infer the pattern of fluid flow just beneath the core–mantle boundary — one of the very few ways to "see" core dynamics without direct access. **A known modeling limitation.** Reproducing the true balance between induction and diffusion in numerical dynamo simulations is currently impossible at the real physical parameters of the core. The relevant ratio — the magnetic Prandtl number, comparing viscous diffusion to magnetic diffusion — is roughly one-millionth in the real outer core, meaning viscosity is essentially negligible next to magnetic diffusion at core-relevant scales. No computer simulation can resolve motion at that real ratio, so all current dynamo models use artificially inflated viscosity as a workaround. This is a well-recognized limitation in the field, and it is one reason dynamo simulations are validated against paleomagnetic reversal records and satellite field measurements rather than taken as literal reconstructions of core flow. ## 7.8 Open Questions - **Inner core rotation:** seismic studies suggest the solid inner core may rotate at a very slightly different rate than the mantle and crust ("super-rotation"), and some recent studies argue this differential rotation oscillates over decades. The precise behavior and its coupling to the outer-core dynamo remain under active debate. - **Balance of energy sources:** the relative contributions of thermal versus compositional convection to the present-day dynamo, and how that balance has shifted since the inner core began solidifying (estimated at roughly half a billion to a billion years ago), are not fully settled. - **Age of the field:** paleomagnetic evidence in ancient zircon crystals suggests a magnetic field may have existed as far back as 3.5–4.2 billion years ago, before the inner core had formed at all — implying an early dynamo powered differently than today's, a topic of ongoing research. ## 7.9 Why This Matters at the Surface The resulting field extends far beyond the crust into space as the magnetosphere, which deflects most solar wind and the charged particles that would otherwise erode the atmosphere and elevate surface radiation. This is perhaps the most striking example of core–crust interaction: a fluid-dynamics process occurring 2,900+ km below the surface, powered by the freezing of the inner core, directly determines the radiation environment at the surface, the reliability of satellite navigation and communications, and — over geological timescales — whether an atmosphere is retained at all. ## 7.10 How the Dynamo Actually Switches On: Nonlinear Transition Theory Everything above describes the geodynamo once it is already running. A separate and more mathematically demanding question — currently active research, not settled textbook material — is *how* a dynamo turns on in the first place, and why the transition looks the way it does. The following is drawn directly from Skene, Marcotte & Tobias (2025, *Journal of Fluid Mechanics*, 1021:A37) — verified against the paper's own text, not independent replication, since the paper is only months old as of this writing and has not yet had the years of follow-up work that most other findings in this document rest on. - **The setup:** the study models a rotating spherical shell of electrically conducting fluid under the Boussinesq approximation — the same governing physics as Sections 7.1–7.4, but studied at an idealized benchmark parameter point rather than at realistic Earth values (the real core's Ekman number and magnetic Prandtl number are many orders of magnitude away from anything a simulation can reach directly — the companion TW2 research notes document covers this parameter gap in full). - **Three key solution states, found by direct numerical continuation at a standard benchmark parameter point (Ekman number 10⁻³, magnetic Prandtl number 5):** - **TW0** — a purely hydrodynamic (non-magnetic) convective state, linearly stable to both hydrodynamic and magnetic perturbations at this point. - **TW1** — a self-sustaining magnetic dynamo state, stable, mainly dipolar with a quadrupolar toroidal component — structurally the same kind of solution as the dipole-dominated field described in Section 7.5. - **TW2** — an unstable magnetic state that is neither of the above: it acts as an **"****edge state,****"** a saddle point in the space of possible solutions sitting directly between TW0 and TW1, confirmed by direct linear stability calculation to possess exactly **one unstable direction** — the single perturbation direction along which trajectories are repelled from the saddle toward either TW0 or TW1. - **The mechanism this reveals:** TW1 and TW2 are born together at a single saddle-node bifurcation (near a Rayleigh-number value of roughly 98 in this benchmark) as the convective driving is increased, and which one a given magnetic seed field grows into — a working dynamo (TW1) or decay back to no field at all (TW0) — depends entirely on which side of the unstable TW2 edge the seed happens to land on. This is a **subcritical** transition: a dynamo can exist at driving strengths where the non-magnetic state is *also* still stable, so the outcome depends on the history and size of the perturbation, not on driving strength alone. This directly extends the weak-field/strong-field distinction discussed conceptually in Section 7.1 (whether the Lorentz force enters the momentum balance) into an explicit, tested dynamical map of *how* a system arrives at one regime or the other. - **Why this matters for the ****"****does the geodynamo need a large trigger****"**** question:** because TW0 is linearly stable to magnetic perturbations, a small, infinitesimal magnetic field will *not* spontaneously grow into a dynamo — a standard linear stability analysis of TW0 alone would misleadingly conclude no transition is possible at all. Only a sufficiently large, correctly-structured finite perturbation can push the system across the TW2 edge into the self-sustaining TW1 state. The study locates this perturbation directly using adjoint-based optimal control, computing the **minimal seed** — the smallest-energy magnetic field addition to TW0 that successfully triggers the dynamo. This offers a concrete, quantifiable version of a long-standing qualitative idea in geodynamo theory (Roberts, 1978) that the real geodynamo may itself have originated subcritically — needing an initial seed field of sufficient size and structure, rather than switching on gradually from zero. - **Where this fits relative to the rest of Section 7:** Sections 7.1–7.9 describe the confirmed physics of the *running* geodynamo — energy sources, rotation's organizing role, the induction feedback loop, diffusion, and the resulting field's observed behavior. This subsection describes the separate, still-developing mathematical theory of dynamo *onset* — an idealized model system, not a direct simulation of Earth's core, but one built on the same governing equations and offering the most rigorous current answer to how a subcritical dynamo transition actually works. A dedicated companion document (*TW2 Nonlinear Transition — Living Research Notes*) goes considerably deeper into this specific topic — the full governing equations, non-dimensional parameters, boundary-condition choices, and open research questions around interpreting this class of transition. ## 7.11 The Magnetic Poles: Geographic, Geomagnetic, and Dip Poles Sections 7.5 and 7.6 mentioned the magnetic poles in passing — this subsection makes the terminology precise, since three genuinely different "north poles" are routinely conflated in casual usage. - **The geographic (rotation) pole:** the fixed point where Earth's spin axis intersects its surface. This is what maps and GPS coordinates use as "true north," and it has no direct connection to magnetism at all. - **The geomagnetic pole:** the point where the axis of the *best-fit dipole* approximation to the field (Section 7.5) intersects the surface. Because that dipole is tilted roughly 10° from the rotation axis, the geomagnetic pole sits a measurable distance from the geographic pole — currently in the vicinity of northwest Greenland, not at the rotation pole. - **The magnetic dip pole:** the point where the *actual, full* field (dipole plus all the smaller non-dipole components, Section 7.5) points straight down — the location a real dip-needle compass would indicate. Because the real field is messier than a pure dipole, the dip pole sits somewhere different again from the geomagnetic pole, and it is the dip pole, not the geomagnetic pole, that is popularly reported as "the magnetic north pole" in the news. - **The dip pole is genuinely, measurably moving:** historical surveys tracked the north magnetic dip pole wandering slowly around the Canadian Arctic through most of the 20th century, at roughly 10–15 km per year. Starting in the 1990s, its drift rate and direction changed markedly, accelerating to as much as 50–60 km per year at times during the 2000s and 2010s and shifting onto a trajectory heading toward Siberia — a real, ongoing, closely monitored expression of the secular variation discussed in Section 7.5, reflecting genuine changes in the pattern of flow in the outer core beneath the pole's projected location. - **Not the same thing as true polar wander:** Section 6.5's mention of "true polar wander" refers to the entire solid Earth reorienting relative to the fixed rotation axis in response to mass redistribution (ice ages, mantle convection) — a completely different, far slower phenomenon from magnetic dip-pole wander, which reflects core fluid motion rather than any reorientation of the planet itself. The similar names are a common source of confusion between two genuinely unrelated processes. - **Paleomagnetic poles and the evidence for continental drift:** because cooling volcanic rock locks in the direction to the magnetic pole at the time it solidified (Section 7.6), geologists can reconstruct a rock sample's "paleomagnetic pole" position at the time it formed. Plotting these reconstructed pole positions from progressively older rocks on a single continent traces out an "apparent polar wander path" that appears to show the pole itself wandering over hundreds of millions of years — but because the geomagnetic pole averages close to the rotation axis over any few-thousand-year window (Section 7.5), the actual, more sensible interpretation is that the *continent* has moved beneath a comparatively stable pole. Distinct continents show different, non-matching apparent polar wander paths for the same time periods, which was in fact one of the pivotal pieces of quantitative evidence supporting continental drift and, eventually, plate tectonics (Section 3.4) in the mid-20th century. - **Is a pole reversal imminent?** the field's overall dipole strength has been declining measurably for roughly two centuries, and the South Atlantic Anomaly (Section 7.5) has been slowly growing — both facts often cited as signs a reversal (Section 7.6) might be approaching. Current scientific consensus treats this as a monitored, genuinely uncertain possibility rather than a predictable near-term event: reversals take on the order of 1,000–10,000 years to complete once begun, comparable declines in field strength have occurred before without leading to a reversal, and there is presently no reliable way to forecast if or when the next one will actually begin. where our gaseous envelope thins out and merges with outer space.

Back