Because earthquakes result ultimately from the relative movement of lithospheric plates, and take the form of various kinds of ground motion it is easy to think of them just in mechanical terms.  However, such movements affect materials that respond in odd ways to motion and friction. One of the most obvious is the sound near a fault zone during an earthquake, which can range from a rumble to a piercing shriek, depending on the near-surface rocks being dragged past each other. There are other, more subtle effects.  For instance, if grains of quartz or dolomite are rubbed against one another they glow – a nice piece of natural magic for the dark days of winter.  There have been many reports of so-called ‘Earth lights’ along active fault zones before and during earthquakes, and they might result from this piezoluminescence. Rocks differ in their ability to conduct electricity, but Faraday’s laws of electromagnetism show that if a conductor is moved in a magnetic field, currents flow through it; the principle behind electricity generation.  In turn, motion in a magnetic field of a conductor in which electricity flows generates electromagnetic radiation, whose frequency depends on the rate of motion. Electromagnetic effects may also result from build-up of electrical charge derived from minerals in the crust, or from crushing of magnetic minerals. Along with even less well understood phenomena, such as the rise and fall of water levels and various gas discharges in wells, and animal behaviour, physical changes are potential means of earthquake warning, if they can be detected and properly understood, that could supplement and even supersede conventional approaches to early warning.

Minoru Tsutsui of the Kyoto Sangyo University in Japan has concentrated on the EM radiation known to precede earthquakes (Tsutsui, M. 2005. Identification of earthquake epicentre from measurements of electromagnetic pulses in the Earth. Geophys. Res. Lett., 32, L20303, doi:10.1029/2005GL023691). Previously published observations have been limited to noting EM pulses before major seismic events. These showed that in some cases nearby areas experienced increased EM noise up to a few months beforehand, to peak a few hours before events. The radiation is at very low frequencies, i.e. wavelengths are much longer than normal radio waves. Such ultra-low frequency (ULF) radiation passes extremely efficiently through rock, and ULF has been used for secret communications between submarines and their bases, as it passes through the whole Earth. In the context of seismic prediction, detecting ULF changes is not enough: the object is to predict the position of an earthquake’s focus as well as its timing. Tsutsui has developed a means of finding the direction in which ULF radiation moves, which has been calibrated using the ULF from lightning strikes and the position of the thunder clouds found using weather radar systems. A strong ULF EM pulse that accompanied a magnitude 5.5 earthquake, whose epicentre was known from studies of seismograph records, enabled the Kyoto team to try out their method.  It succeeded in accurately pinpointing the epicentre, thereby proving that ULF radiation is generated at the site of earth movements. But that is not sufficient to provide a warning system. The equipment and data analysis have to be refined and continually tested to detect and use ULF noise long before events, to see whether or not these preceding signals point to future epicentres.

As Charles Darwin noted in Voyage of the Beagle, following his experience of a major earthquake in Chile, nothing is more frightening than the unexpected movement of the ground on which one stands. Every victim of an earthquake suffers post-traumatic stress disorder, whether or not they are injured or lose people close to them – we all implicitly trust solidity. Yet many survive physically because they instinctively seek some kind of shelter; perhaps one advantage of panic in the face of such a sudden threat.  How much warning is needed in order to act according to a learned plan, in the manner of following a fire drill?  Would say 20 seconds be enough? With even such a short warning, automated shut-down mechanisms for gas supplies – much damage and fatality is caused by fires in the aftermath of earthquakes – and activation of road and rail warnings would be possible.  It would also enable people to escape from small buildings or to seek shelter in larger ones, given an ‘earthquake drill’, and an audible alert, such as a siren.

During research into the way in which faults rupture, based on seismograms of events of all detectable magnitudes, Erik Olson and Richard Allen of the University of Wisconsin, USA, made a potentially useful discovery (Olson, E.L. & Allen, R.M. 2005. The deterministic nature of earthquake rupture. Nature, v. 438, p, 212-215). Previously, the most widely held view was that the magnitude of an earthquake could not be calculated until all its energy had been released. Indeed, the magnitudes of the events that caused the 26 December 2004 Indian Ocean tsunamis and the massive loss of life in Kashmir and northern Pakistan in October 2005 were not calculated until hours afterwards. Olson and Allen found that the energy delivered by the first arrivals of fast seismic P waves correlated closely with the total energy of the full event, i.e. with its magnitude. The key to this finding was their analysis of the frequency of the early P waves, which show sufficiently good correlation with final magnitude for useful prediction of the most damaging events. P waves arrive around 20 to 30 seconds before the most energetic but slower surface waves, and they are rarely noticeable. If frequency analysis of the kind used by the authors were to be systematised at seismic stations, automatic warnings could be generated.  They would not be false alarms because they are based on actual seismicity, although imprecision might mean that some alarms were followed by smaller earthquakes than the theory predicts.

See also: Tata, P. 2005. Can Earth’s seismic radio help predict quakes? New Scientist, 19 November 2005, p.28-29.