One of the most fascinating spin-offs of detailed palaeontology is that the growth layers in corals and the carbonate shells of other organisms can record how many days there once were in a year. Records of shell growth can even chart variations in the lunar cycle, backed up by subtle features in cyclical sediments. Such data infer that the speed of the Earth’s rotation has changed (Ravilious, K. 2002. Wind up. New Scientist, 23 November 2002, p. 30-33). As well as the general slowing through the Phanerozoic, from a rate that gave 420, 21-hour days in a Cambrian year, there have also been times when the rate has strangely speeded up again. Such curious events occurred at 400 Ma and again around 180 Ma.
Planetary spin can be set in motion or changed by very large impacts, in the manner of whipping a spinning top. But there is little sign for such drama at those times. Another possibility is a change in the Earth’s moment of inertia by a shift of mass relative to the spin axis, in the manner of a skater speeding up a spin by pulling in her arms. What could induce such an effect at the scale of our planet? Cold, dense lithosphere continually sinks at subduction zones, but that is normal behaviour in balance with rotation. One possible trigger for sudden changes in moment of inertia is the breaking away of a substantial chunk of the mantle that lies above the discontinuity 670 km beneath the surface to sink to deeper levels. This dramatic suggestion stems from modelling by Philippe Machetel and Emilie Thomassot of the University of Montpellier in France (Machetel , P. & Thomassot , E. 2002. Cretaceous length of day perturbation by mantle avalanche. Earth and Planetary Science Letters, v. 202, p. 379-386). Their model focussed on the transition zone between lower and upper mantle around the 670 km discontinuity, and how it might respond to the fluid dynamics of Earth’s convective heat transfer, particularly that involving heat originating in the core. The transition, they claim, acts as a “lid” to efficient heat transfer between lower and upper mantle. Their model suggested that additional deep-mantle heat flow might destabilise the transition’s strength, so that it would no longer support the mass of cooler and more rigid mantle above it. Failure could then allow a massive slab of upper mantle literally to fall to the core-mantle boundary, spreading out to displace material there upwards as the precursor of a superplume.
The link to day-length comes from Machetel and Thomassot’s search for evidence that such collapses might have occurred, and they concentrated on the 180 Ma change (Mid Jurassic). Around 170 Ma the current round of continental drift began in earnest. In the Early Cretaceous (130 Ma) the geomagnetic field became locked into quiescence, remaining with the same polarity for an unprecedented 40 Ma during which the giant Ontong Java oceanic flood volcanism took place. Their explanation for both is that upper mantle avalanched, eventually to reach the core-mantle boundary. When the mass “bottomed out” it cooled the outer core, settling it into regular motion, so that the geomagnetic field became constant. Coincidence? I am reminded that when skaters wish to stop their spins, they throw out their arms. The law of conservation of angular momentum also demands that the Earth behaves in the same way. In fact it applies to the Earth-Moon system, so that the general slowing of Earth’s rotation has been accompanied by the Moon receding into ever more distant orbit, and gaining momentum.