# Why Earth’s magnetic field is reversing its direction every few million years? [duplicate]

Just now I was studying the Magnetism and Earth's magnetic field source. when I read that the earth's magnetic poles are not fixed and keep changing ( the north becomes south and the inverse) every few million year? can someone tell me why does that happen?

You are asking the most difficult question of magnetic dynamo theory. It is a mystery why the Sun’s field reverses at more-or-less regular intervals of 11 years, while the Earth’s field reverses at erratic intervals of 10^5 to 10^6 years. But ignorance won’t stop me from trying to answer.

We know more about the Sun than the Earth, because we can see what’s going on at the surface. We know where the field is strongest from studying sunspots, and we can see that the Sun is rotating faster at the equator than at the poles.

Somehow or other, turbulent convection in an electrically conductive fluid (molten iron or plasma) sustains the magnetic field, and the rotation of the body plays an indirect role in orienting it. Since turbulence is inherently erratic, it is no surprise that the Earth’s field flips erratically. It is surprising that the Sun’s flips so regularly.

Absent a dynamo mechanism, inductive decay would dissipate the Earth’s field in roughly 20,000 years, so let’s try to follow the energy flows. Some power source (probably settling of heavier elements in the Earth’s molten outer core, but simply heat generated by fusion in the Sun’s center and transported through the outer layers) drives turbulent convection, but some of the fluid’s kinetic energy gets transferred to the magnetic field. Some magnetic energy is dissipated by electrical resistance, and some kinetic energy is dissipated by viscosity.

The equations of fluid flow and magnetic induction are well known, so we understand the energy transfers locally, but not globally. Dynamo phenomena are being simulated on supercomputers, but insight is slow in coming.

It is thought that most of the magnetic field energy is contained in two main patterns: (1) the dipole field that we see emerging from the poles, and (2) an internal toroidal field comprising east-west belts at mid-latitude. Some more energy is contained in higher multipole patterns.

As for reversals ... Where does the energy in the dipole field go when the dipole field changes sign? Much of it goes into the toroidal field. There is evidence that this is what happens in the Sun, where the two main field amplitudes oscillate about 45° out of phase. The oscillation would keep the total energy constant if it went like ${{\cos }^{2}}(\omega t)+{{\sin }^{2}}(\omega t)$, but the phase difference is not 90°. Some energy is going elsewhere. Regular oscillation would be expected if the system could be modeled with coupled differential equations something like this:

• Hairpin formation ... $\tfrac{d}{dt}{{B}_{toroidal}}=+\ldots {{B}_{dipole}}$

• Regeneration ........... $\tfrac{d}{dt}{{B}_{dipole}}=-\ldots {{B}_{toroidal}}$

We understand hairpin formation under the influence of the Sun’s differential rotation, but there are many competing theories about regeneration. I won’t tire you with my own.

The magnetic oscillation is both damped and driven, but the driving terms probably represent erratic energy transfers from higher multipole patterns. (We are talking about turbulence, after all.) If the system is sufficiently damped, the oscillations would become very erratic, enough to obscure their inherent periodicity.