Tikalon Header Blog Logo

Earthquake Prediction

February 18, 2011

At times, science seems infallible. When you have a mystery, you follow the scientific method, and if your instrumentation is capable enough, you end up with an answer. Sometimes, all you need is a little more money, or a little more time. We can predict what will happen when we mix certain chemicals together, what will happen when we "knock-out" certain genes in an organism, and where a particular asteroid will be on any given day next year. What we can't predict is the time of the next earthquake.

At first, earthquake prediction didn't appear to be that difficult a problem. From antiquity we had reports of earthquake precursors that included changing water levels in wells and unusual animal behavior. There's evidence that common toads may have sensed the 6.3-magnitude L'Aquila, Italy, earthquake of April 6, 2009, three days in advance.[1] Certainly we can measure water level very accurately; and the animal behavior might be caused by micro-tremblers, or subtle changes in electric field and magnetic field strength. We've got great instruments to measure these, too, with probably more sensitivity than a toad. Need I mention that none of these seem to work.

Major earthquakes in Morris County, New Jersey

Not even safe in my backyard.
Major earthquake epicenters in Morris County, New Jersey.

So, science stepped up a notch, and started to examine other things that may be related to underground stress, such as radon dissolved in well water, and changes in the electrical noise measured at ground level. The premise of the later is that materials under stress will emit voltage signals. A variation of this is fractoluminescence, in which the electric fields caused by separation of mineral crystal planes generate light. I wrote about fractoluminescence in a previous article (Sound-Assisted Reactions, May 18, 2007). One useful predictor is the presence of medium-sized earthquakes, called foreshocks that precede major quakes. Unfortunately, the "fore" appellation is misleading, since about 90% of these are just an earthquake itself, and not a precursor to something larger.

Many things in Nature are regular, so it might be useful to look for patterns in earthquake occurrence. A paper published several years ago in Physical Review Letters [2] found that the distribution of recurrence times for earthquakes depends strongly on the previous recurrence time. In effect, small and large recurrence times cluster in time, and the possibility of having an earthquake within a certain time span after the last earthquake depends on history.

With the failure of every other approach, considerable effort has been applied to measuring the changes in local electric and magnetic fields. Observations by the Demeter (Detection of Electro-Magnetic Emissions Transmitted from Earthquake Regions) microsatellite have shown a strong correlation of low frequency electromagnetic activity and seismic activity. Indeed, there have been several notable coincidences. [3-5] Ultra-Low Frequency (ULF) electromagnetic waves were detected by Demeter as it passed over Haiti before and after the magnitude-7.0 earthquake that struck Haiti on January 12, 2010. A significant increase in ULF energy was observed in the 30 day interval before the strong earthquake, and changes in energy were also observed 25 days after, coincident with strong aftershock activity.[6-7] The instruments aboard Demeter can detect electric field from DC up to 3.5 MHz, and magnetic field from a few Hz up to 18 kHz.

Ultra-Low Frequency seems to me to be the most exciting area for earthquake prediction research. It all started when Antony Fraser-Smith, an electrical engineer at Stanford University, detected high intensity ULF radio signals in the days leading up to the Loma Prieta earthquake in 1989. His instrument was located seven kilometers from the epicenter. As is typical in science, this was an accidental discovery. He and a graduate student were working on a US Navy program studying how natural noise affects satellite communication. They were most interested in ULF radio waves produced when the solar wind interacts with Earth's magnetosphere. Twelve days before the Loma Prieta earthquake, they recorded a large signal; and then, three hours before the earthquake, the signal went off scale. It was 20-30 times larger than the typical signal level.[8]

Radio propagation, as well as radio emission, has been studied. There were reported anomalies in very low frequency (VLF) signal propagation associated with the Abruzzo, Italy, (L'Aquila) magnitude 6.3 earthquake of April 6, 2009. Radio signals that crossed the epicenter showed differences in signal strength compared to controls 2-8 days prior to the earthquake.[9] The signals were in the 19.58 kHz to 45.9 kHz frequency range.

I could probably conduct some ULF radio experiments in my own backyard. On February 2, 2009, a magnitude-3.0 earthquake occurred with an epicenter just seven miles from my house.[10]


  1. Matt Walker, "Toads can 'predict earthquakes' and seismic activity," BBC News, March 31, 2010.
  2. Valerie N. Livina, Shlomo Havlin and Armin Bunde, "Memory in the Occurrence of Earthquakes," Phys. Rev. Lett., vol. 95, no. 20 (November 10, 2005), document 208501 (DOI: 10.1103/PhysRevLett.95.208501).
  3. F. Muto, M. Yoshida, T. Horie, M. Hayakawa, M. Parrot, and O. A. Molchanov, "Detection of ionospheric perturbations associated with Japanese earthquakes on the basis of reception of LF transmitter signals on the satellite DEMETER," Natural Hazards and Earth System Sciences, vol. 8, no. 1 (February 26, 2008), pp. 135-141.
  4. A. Rozhnoi, M. Solovieva, O. Molchanov, P.-F. Biagi, M. Hayakawa, K. Schwingenschuh, M. Boudjada, and M. Parrot, "Variations of VLF/LF signals observed on the ground and satellite during a seismic activity in Japan region in May-June 2008," Natural Hazards and Earth System Sciences, vol. 10, no. 3 (March 16, 2010), pp. 529-534.
  5. Demeter Data Server
  6. M. Athanasiou, G. Anagnostopoulos, A. Iliopoulos, G. Pavlos and K. David, "Enhanced ULF radiation observed by DEMETER two months around the strong 2010 Haiti earthquake," arXiv Preprint Server, December 7, 2010
  7. kfc, "Spacecraft Saw ULF Radio Emissions over Haiti before January Quake," arXiv Blog, December 9, 2010.
  8. Scientists debate new evidence for electromagnetic earthquake predictors, Stanford University News Service, December 31, 1991.
  9. A. Rozhnoi, M. Solovieva, O. Molchanov, K. Schwingenschuh, M. Boudjada, P. F. Biagi, T. Maggipinto, L. Castellana, A. Ermini and M. Hayakawa, "Anomalies in VLF radio signals prior the Abruzzo earthquake (M=6.3) on 6 April 2009," Natural Hazards and Earth System Sciences, vol. 9, no. 5 (21 October 2009), pp. 1727-1732.
  10. Trymaine Lee, "Small Earthquake Rattles New Jersey," New York Times, February 3, 2009.

Permanent Link to this article

Linked Keywords: Science; scientific method; instrumentation; chemicals; knockout mouse; gene; asteroid; earthquake; earthquake prediction; earthquake precursor; common toads; moment magnitude scale; 2009 L'Aquila earthquake; Italy; electric field; magnetic field; earthquake epicenters in Morris County, New Jersey; stress; radon; VAN method; electrical noise; piezoelectricity; triboluminescence; fractoluminescence; Nature; pattern; Physical Review Letters; probability distribution; Demeter satellite; electromagnetic field; Ultra-Low Frequency; Haiti; 2010 Haiti earthquake; direct current; DC; MHz; Hz; kHz; electrical engineer; Stanford University; 1989 Loma Prieta earthquake; kilometer; epicenter; serendipity; US Navy; natural noise; satellite communications; solar wind; Earth's magnetosphere; radio propagation; Abruzzo, Italy.