Seafloor spreading, magnetic reversals, and plate tectonics

Doing science consists of formulating refutable hypotheses, this is, new interpretations based on former experience that lead to predictions that can be either confirmed or falsified (by future research).

Timing of the last reversals of the Earth's
magnetic field. Time goes from 5 Million 
years ago (bottom) to present (top).

Periods in black match today's polarity;
periods in white underwent reversed polarity. 

Source: Wikimedia Commons. A more complete scale here.

A case history in Earth science is the confirmation of the plate tectonics theory during the 60's. Back in 1912, this theory was just a hypothesis known as continental drift and put forward most remarkably by Alfred Wegener, based on observations of the fossil fauna matching across different continents. Well, in addition to the matching coastlines of continents pointed out by Abraham Ortelius as early as in the... 16th century!

One implication of the continental drift idea was that the oceans laying between continents that drifted away from each other should have gradually spread apart. This is known as the seafloor spreading hypothesis. But how to prove it?

Much earlier than that, the Earth's magnetic field had been studied scientifically since the beginning of the Spanish and Portuguese explorations of the Americas (Alvarez & Leitao, 2010, Geology, The neglected early history of geoscience). By the 17th century, maritime trading was dependent on the accurate mapping of magnetic intensity across the Atlantic Ocean. These studies culminated by the 19th century during the so-called Magnetic Crusade, leading to the realisation that the magnetic poles migrate significantly over historical time periods. And in fact, these rapid changes of the magnetic field soon became one of the theories proposed to explain why the magnetic orientation recorded in rocks depends on their geological age.

North Magnetic pole wander from 1590 to 2015. Click on the pins to see the year. From the GUFM and IGRF models. Via NOAA.

Today we know that historical magnetic changes are normal in periods of stable magnetic polarity, and that although the polarity flips recorded in rocks take just a few thousand years, they occur only over geological time-scales (millions of years).

Computer model based on Glatzmaier & Roberts. Magnetic field lines are in blue
when the field points towards the center and yellow when pointing away from it. The
rotation axis of the Earth is centered and vertical. The dense clusters of lines are
within the Earth's core

Back in 1957, Marie Tharp found enigmatic alignments in the shape of the seafloor around the center of the Atlantic Ocean, roughly where seismicity was being detected. In 1963, both the geophysicist Frederick J. Vine and the geologist Lawrence W. Morley independently realized that if the seafloor spreading theory was correct, then the rocks surrounding mid-oceanic ridges should show symmetric patterns of magnetization reversals, recording the changes of the Earth's magnetic field in the volcanic rocks at the time when these erupted and cooled down at the mid-ocean ridges. This is  now known as the Vine–Matthews–Morley hypothesis, and became a validation test for the seafloor spreading, and for the plate tectonics theory in general.

Seafloor spreading at a mid-ocean ridge, recording time-changes of geomagnetic 
field polarity. Source: Wikimedia Commons.

Morley's letters to Nature (February 1963) and to the Journal of Geophysical Research (April 1963) were both rejected, so Vine and his advisor Matthews were first to publish the hypothesis on the same year. The patterns of ancient reversals of the Earth's magnetic field have been found thereafter in hundreds of paleomagnetic surveys, providing a robust validation of their hypothesis. In fact, a vast later work of age calibration of these magnetic reversals allowed for the detailed maps of the age of the oceanic floor that we have nowadays:

Map of the age of the seafloor based on the reversal of the magnetic field
recorded in the oceanic crust during its formation at mid-oceanic ridges.
Red indicates a young seafloor, whereas blue is used for the oldest oceanic crust.
(Source: National Geophysical Data Center)

Magnetic reversals are still today one of the key methods allowing rock dating (don't miss the name for it: magnetochronostratigraphy). But we know very little about the mechanisms responsible for these magnetic field changes. Computer simulations suggest that it is a natural result of feedback forces between the magnetic field and the flow in the Earth's core (see the reference to Glatzmaiers' below), similar to dynamo going tilted by its own magnetic field. It has been recently shown in this article in GRL a correlation between the distribution of tectonic plates and the frequency of magnetic reversals over geological time ("geological intervals characterized by an asymmetrical distribution of the continents with respect to the equator are followed by intervals of high reversal frequency"), suggesting a mechanical coupling between both phenomena. But the specific mechanism behind magnetic reversals and the additional information they may contain about the interior and the past of our planet remain, so far, a challenge (yet another Reto Terrícola!).

Update (2015-09): A Science News article on a recent study on core convection and the magnetic field.

Anyone thinking of the Earth as a static thing? Here is the global magnetic declination in 1492 and today: pic.twitter.com/0WjJiwDpmx

— ∆ Garcia-Castellanos (@danigeos) July 6, 2015

The origin of the Earth's magnetic field 

explained in 9 minutes.

References:

Vine, F., & Matthews, D. (1963). Magnetic Anomalies Over Oceanic Ridges Nature, 199 (4897), 947-949 DOI: 10.1038/199947a0

Pétrélis, F., Besse, J., & Valet, J. (2011). Plate tectonics may control geomagnetic reversal frequency Geophysical Research Letters, 38 (19) DOI: 10.1029/2011GL048784

Glatzmaiers, G., & Roberts, P. (1995). A three-dimensional self-consistent computer simulation of a geomagnetic field reversal Nature, 377 (6546), 203-209 DOI: 10.1038/377203a0