When Earth got its magnetic field

For a planet to produce life it needs various attributes. Exoplanet hunters tend to focus on the ‘Goldilocks’ Zone’ where solar heating is neither so extreme nor so little that liquid water is unstable on a planet’s surface. It also needs an atmosphere that retains water. Ultraviolet radiation emitted by a planet’s star dissociates water vapour to hydrogen and oxygen and the hydrogen escapes to space. The reason Earth has not lost water in this way is that little water vapour reaches the stratosphere because it is condensed or frozen out of the air as the lower atmosphere becomes cooler with altitude. Given moist conditions survivability to the extent that exists on Earth still needs another planetary parameter: the charged particles emitted as an interplanetary ‘wind ‘by stars must not reach the surface. If they did, their potential to break complex molecules would hinder life’s formation or wipe it out if it ventured onto land. A moving current of electrical charge, which is what a stellar ‘wind’ amounts to, can be deflected by a magnetic field. This is what happens on Earth, whose magnetic field is a good reason why our planet has supported life and its continual evolution since at least about 3.5 billion years ago.

Artist's rendition of Earth's magnetosphere.
Deflection of the solar ‘wind’ by Earth’s Earth’s magnetosphere. (credit: Wikipedia)

Direct proof of the existence of a geomagnetic field is the presence of aligned particles of magnetic minerals in rocks, for instance in a lava flow, caused by their acquiring magnetisation in a prevailing magnetic field once they cooled sufficiently. The earliest such remanent magnetism was found in igneous rocks from north-eastern South Africa dated at between 3.2 to 3.45 billion years. All older rocks do not show such a feature dating back to their formation because of thermal metamorphism that resets any remanent magnetism to match the geomagnetic field prevailing at the time of reheating. There are, however, materials that formed further back in time and are also known to resist thermal resetting of any alignments of magnetic inclusion. They are zircons (ZrSiO4), originally crystallised from igneous magmas, which may have locked in minute magnetic inclusions. Zircons are among the most change-resistant materials and they can also be dated with great precision, with the advantage that the U-Pb method used can distinguish between age of formation and that of any later heating. Famously, individual grains of zircon that had accumulated in an early Archaean conglomerate outcropping in the Jack Hills of Western Australia yielded ages going back from 3.2 to 4.4 billion years; far beyond the age of any tangible rock and close to the formation age of the Earth. Quite a target for palaeomagnetic investigations once a suitable technique had been developed.

Western Australia's Jack Hills
Western Australia’s Jack Hills from Landsat (credit NASA Earth Observatory)

John Tarduno and colleagues from the Universities of Rochester and California USA and the Geological Survey of Canada report the magnetic properties of the Jack Hills zircons (Tarduno, J.A. et al. 2015. A Hadean to Paleoarchean geodynamo recorded by single zircon crystals. Science, v. 349, p. 521-524). All of the grains analysed record magnetisation spanning the period 3.2 to 4.2 billion years that indicate geomagnetic field strengths ranging from that found today at the Equator to about an eighth of the modern value. So from 4.2 Ga onwards geomagnetism probably deflected the solar wind: the early Earth was set for living processes from its earliest days. The discovery also supports the likelihood of functioning plate tectonics during the Hadean.

Core’s comfort blanket and stable magnetic fields

Pangea animation
Pangaea and its break-up. Image via Wikipedia

The record of the Earth’s magnetic field for the most part bears more than a passing resemblance to a bar-code mark, by convention black representing normal polarity, i.e. like that at the present, and white signifies reversed polarity. The bar-code resemblance stems from long periods when the geomagnetic poles flipped on a regular, short-term basis, by geological standards. The black and white divisions subdivide time as represented by geomagnetic into chrons of the order of a million-years and subchrons that are somewhat shorter intervals. Stemming from changes in the Earth’s core, magnetostratigraphic divisions potentially occur in any sequence of sedimentary or volcanic igneous rocks anywhere on the planet and so can be used as reliable time markers; that is, if they can be defined by measurements of the remanent magnetism preserved in rock, which is not universally achievable. Yet this method of chronometry is extremely useful, for most of the Phanerozoic. However, there were periods when the geomagnetic field became unusually stable for tens of million years so the method is not so good. These have become known as superchrons, of which three occur during Phanerozoic times: the Cretaceous Normal Superchron when the field remained as it is nowadays from 120 to 83 Ma; a 50 Ma long period of stable reversed polarity (Kiaman Reverse Superchron) from 312 to 262 Ma in the Late Carboniferous and Early Permian; the Ordovician Moyero Reverse Superchron from 485 to 463 Ma.

Because the geomagnetic field is almost certainly generated by a self-exciting dynamo in the convecting  liquid metallic outer core, polarity flips mark sudden changes in how heat is transferred through the outer core to pass into the lower mantle. It follows that if there are no magnetic reversals then the outer core continued in a stable form of convection; the likely condition during superchrons. But why the shifts from repeated instability to long periods of quiescence? That is one of geoscience’s ‘hard’ questions, since no-one really knows how the core works at any one time, let alone over hundreds of million years. There is however a crude correlation with events much closer to the surface. The Kiaman superchron spans a time when Alfred Wegener’s supercontinent Pangaea had finished assembling so that all continental material was in one vast chunk. The Cretaceous superchron was at a time when sea-floor spreading and the break-up of Pangaea reached a maximum. The Ordovician, Moyero superchron coincides with the unification of what are now the northern continents into Laurasia and the continued existence of the southern continents lumped in Gondwana, so that the Earth had two supercontinents. Those empirical observations may have been due to chance, but at least they provide a possible clue to linkage between lithosphere and core, despite their separation by 2800 km of convecting mantle that transfers the core heat as well as that produced by the mantle itself to dissipate at the surface. Enter the modellers.

How part of the Earth transfers heat is, not unexpectedly, very complex, depending not only on what is happening at that point but on heat-transfer processes and heat inputs both above and below it. The surface heat flow is complex in its own right ranging from less than 20 to as much as 350 mW m-2, the largest amount being through zones of sea-floor spreading and the least  through continental lithosphere. Wherever heat is released in the core and mantle, willy-nilly the bulk of it leaves the solid Earth along what is today a complex series of lines; active oceanic ridge and rift systems such as the mid-Atlantic Ridge.  These lines weave between six drifting continental masses and many more sites of additional heat loss – hot spots and mantle plumes. The many heat escape routes today complicate the deeper convective processes and there are many possibilities for the core to shed heat, yet they continually change pace and position. When, inevitably, all continental lithosphere unites in a supercontinent, almost by definition, the sites of heat loss simplify too, the supercontinent acting like an efficient insulating blanket. In a qualitative sense, this kind of evolving scenario is what modellers try to mimic by putting in reasonable parameters for all the dynamic aspects involved.  Two physicists at the University of Colorado in Boulder, USA, Nan Zhang and Shije Zhong, have formulated 3-D spherical models of mantle convection with plate tectonics as a basis for whole Earth thermal evolution over that last 350 Ma (Zhang, N & Zhong, S. 2011.  Heat fluxes at the Earth’s surface and core–mantle boundary since Pangea formation. Earth and Planetary Science Letters, v. 306, p. 205-216). The acid test is whether the model can end with a close approximation to modern variations in heat flow and distribution of ages on the sea floor; it does. A probable key to stability in the means of transfer of heat from core to lower mantle – itself a key to a constant outer-core dynamo and geomagnetic polarity – is reduced heat flow at equatorial latitudes; a sort of equatorial downflow of convection with upflows in both northern and southern hemispheres. Zhang and Zhong’s model produced minimal core-to-mantle heat flow at  the Equator at 270 and 100 Ma, both within geomagnetic-field superchrons. Well, that is a good start. Superchrons seem also to have occurred from time to time during the Precambrian, one being documented at the Mesoproterozoic-Neoperoterozoic boundary about 1000 Ma ago. At that time, all continental lithosphere was assembled in a supercontinent dubbed Rodinia (‘homeland’ or ‘birthplace’ in Russian).