Centuries of scientific and popular observations has given us a body of anecdotes peppered with actual physical measurements and recordings of anomalous phenomena occurring prior to large earthquakes. This fact is not in doubt. It has become well known that animals and the atmosphere exhibit odd behaviors that appears to be related to the coming quake. The failure of these signals to become a practical means by which we predict earthquakes in the short term results from 1.) unreliability of the phenomena, 2.) irreproducibility of the phenomena, 3.) inadequate explanation for the phenomena (which follows from the first two). We might add in there a hesitation to divert from the known path in science but that excuse is invalid. There is nothing wrong with holding to a path that has taken you quite far in the correct direction, but sticking to the path can’t hold forever because science progresses, perhaps in a different direction. In terms of earthquake prediction, the time may be now to diverge from the path.
When we consider anomalous earthquake-related phenomena (which I’ll dub ‘AERP’ just to save on typing), we don’t get far by just collecting the stories unless we proceed on to analyze, interpret and explain them. Then, to be of future value, they must be used to predict. Here are some of the theories that have been developed to explain pre-quake AERP such as animal behavior, EQLs (lights), EQCs (clouds), among other, stranger observations.
What happens before a quake
As discussed previously, squeezing, stretching and (micro)fracturing of the rock is inherent in the faulted area. Many AERP appear to have an electrical explanation – related to charged particles, currents and voltage. I do get a little technical in the descriptions that follow. It helps to have a primer in chemistry and physics (which I do have) but I’ll admit I had to look up some reference material in order to make sense out of it. (A great way to learn is to follow through completely on a topic of interest – it takes you to new places.)
First, an obvious consequence of stress and friction is heat. Do fault zones give off heat? The reports of hot, sticky weather preceding a quake suggests that but I had found no evidence to support it. Then, in 2006, Indian scientists report a bloom of plankton offshore prior to a quake may be the result of a release of thermal energy causing the local sea temperature to rise.
It is well known that gases escape from the ground before and during an earthquake. Water vapor, methane and other gases resulting from decomposing organic material may be released. Gas release may serve to explain the (nonelectrical) reports of putrid smells or atmospheric lens effects like observation of an elongated sun or moon before a quake. But, radon, a common, radioactive gas, trapped within rock, is released when the rock forms tiny cracks (microfractures). Increased radon in the air and groundwater has been measured numerous times prior to earthquakes.
Near the epicenter of the 1995 Kobe quake, a mineral water bottling plant noted that the gas content varied in the water prior to the earthquake. In Iceland, a similar observation was noted in 2002 when chemical constituents of a hot spring increased enormously within a 10 week span prior to a 5.8 magnitude quake.
Due to its radioactivity, radon can ionize air. Ionization of the air (by radon or other means) can creates particles known as aerosols. Ions are electrically charged and serve as nuclei for condensation. Aerosol particles can carry soot, dust, droplets, crystals (esp. salt), pollen, even virus and bacteria. The formation and collection of aerosols generated by underground processes would vary depending on the current weather conditions, the geology of the rock and other aerosols in the air. For example, wind and rain will quickly dissipate the particles. Aerosols may only last a few minutes after which they decay. However, if the mechanism that is creating them persists, a new supply can continually be formed.
The aerosol hypothesis, put forth by Tributsch (as an explanation for the various AERP, posited that the coming quake influenced the near-surface atmosphere to such a degree that unusual meteorological events occurred. But, that may be the beginning. There have been reports of metals in gaseous form being expelled in tectonically active areas. These metallic aerosols may play a specific role in the mechanisms that relate seismic activity to anomalies in the upper reaches of the atmosphere (a theory called “lithospheric-ionospheric [or seismo-ionospheric] coupling”).
Animals and people respond to ionized air. Positively charged ions in the air may affect serotonin levels. Serotonin is a hormone that regulates several physiological aspects in humans, such as mood, appetite, and a condition of feeling unwell.
The idea of earth currents was discussed in 1890 by Milne. In the days of telegraph lines, the wires transmitted signals and static by themselves in response to seismic activity. There are other cases of natural electrical phenomena from the earth other than the usual lightning bolts. Glowing patches on mountains have been observed as the electricity is dissipated into the sky across a broad area. St. Elmo’s fire was observed on the high masts of boats and even occurs on high pointed structures on land with enough regularity for it to have been studied. Earthquake lights may be in the same category as anomalous ball lightning, flickering ground lights or “spooklights” and perhaps even some UFOs in that they result from static electricity generated from the ground surface. The theory that an intense electrical field and electromagnetic pulses are generated at a fault zone can potentially explain the various AERP discussed. Dr. Michael Persinger proposed the tectonic strain theory in 1975 relating light phenomena at fault zones to what eyewitnesses report as UFOs.
Rock can actually produce electricity thanks to the “piezometric” effect. It is the result of the ability of crystals, especially quartz, to generate a voltage in response to applied stress. The act of squeezing a quartz crystal induces a polarity to the crystal (one end positive, the other negative). This allows for a current to travel across the crystal. Piezoelectricity can free electrical charges from atom or crystals. These charged particles are ions that may contribute to the generation of a strong electrical field above the ground surface (or aerosols). Quartz is abundant in the earth’s crust, especially in granite, the common foundation rock for continents. Curiously, in an assessment of areas with and without defined precursors, those zones with quartz-poor rock, such as New Zealand, have fewer observable AERP.
Light can result when bonds are broken in a crystal when it is rubbed or cracked. This is called triboluminescence. The broken bond creates a positive and negative charge that recombine as a spark. It can easily be demonstrated by cracking a LifeSaver candy. It has also been observed while cutting diamonds. Scientists are still far from understanding this effect since some substances exhibit this property while others do not.
If enormous electrical currents are being generated, through known or as yet unknown mechanisms, could they serve as the signal we can measure to predict quakes. Do these mechanisms result in observable surface manifestations (AERP)? The buildup of stress in the rock, and release of electrical energy can feasibly result in a release of light and measurable electricity all beginning with the rock fracturing at microscopic scale. Electrons in the atmosphere are accelerated at the fault zone and they produce light when they strike other atmospheric molecules. The manifestation of different kinds of EQLs could be a result of the difference in charge distribution and the uneven field across different areas.
Laboratory experiments on rock samples, such as granite, subjected to high pressure, show that the electrical resistance of water-saturated rock changes just before it shatters. From experiments, the intensity of the electric field generated was greater through the process of microfracturing the rock than at actual breakage.
A stream of charged particles is called plasma. Examples of plasmas are lightning bolts or candle flames. What observers see as various forms of EQLs may be, in fact, plasmas, a stream of electrons from the ground that generates visible light.
We need to pause here to address why the piezoelectricity theory as a cause of earthquake precursor phenomena has been abandoned by some. It is hard to accept that the rocks can become conductive enough to generate an electrical pulse. The assumption made is that the random orientation of crystals in the rock would not allow for the effect to propagate and that the generated positive and negative charges would just cancel each other out. However, lab experiments have shown that if at least some of the crystals are oriented in the same direction, voltages can occur in rocks under stress. Even if only 1% of the quartz grains are aligned, considerable voltage can be produced.
So, if it is feasible that this can occur, where might that lead? The movement of charge in a rock will generate electromagnetic waves.
Electricity and magnetism
Electricity is related to magnetism. The passing of current generates electromagnetic waves. Prior to strong earthquakes in fair weather, scientists have observed anomalous electrical fields and electromagnetic pulses (in the Ultra Long Frequency range or ULF). Electromagnetic pulses can travel through the ground, air and water. Intense electric fields formed by the microfracturing of the underground rock would extent above the ground surface. Although, there will be regions on earth where the tectonic action is so deep, the ULF waves won’t reach the surface and no precursors would occur.
One explanation for AERP we might discount is a change in the earth’s magnetic field before an earthquake. It appears the change is so small, it is negligible. There is a story of a magnet that hung on the wall in Tokyo. In 1855, nails held by the magnet suddenly fell as if the magnet had lost its power. It may seem that the magnetic field was disturbed but it may very well have been that the electrical charge appearing from the ground overwhelmed the magnet’s strength, causing the nails to sway and be attracted to the ground. We are exposed to weak electromagnetic fields (EMFs) all the time. Many experiments have shown that we generally aren’t affected by them. Animals experience small but unimpressive magnetic field changes periodically and do not frequently act unusual. The change in magnetic field can influence some animals but the field is constantly changing especially during solar wind storms. The change in the field appears to be relatively minor as a result of seismic activity compared to these other influences.
Long wave electromagnetic radiation appears whenever electrical charges are generated or neutralized. Electric charge and electromagnetic signals are not detected by seismographs because there is no vibration. Whereas, a radio receiver is a good sensor to detect electromagnetic waves. AM bands on a radio will transmit the EM noise generated from nearby thunderstorms. Radio interference has been mentioned as a possible precursor to quakes.
A bent candle flame, or a candle that is hard to light or burns inefficiently has been noted as an AERP. Ikeya reproduced this effect by generating a charge on the ground that attracts the flame. Ikeya reproduced many other precursor phenomena in the lab by exploring the effects of electrical fields and EM waves. He produced very good evidence to suggest that these conditions are occurring as part of the earthquake progression and showed that the values that could be produced in nature are reasonable to show effects.
Disturbing the upper atmosphere
If we assume these mechanisms are at work in the stressed rock, large electrical fields can occur hours, days or even more than a month before the seismic release. The seismo-ionospheric theory, in development by Russian scientists for decades, suggests these fields reach so high above the earth’s surface that they can affect the upper reaches of the atmosphere and interact with the earth’s global electrical circuit. This area of the atmosphere where the interaction is seen is called the ionosphere. It is a zone 50-1000 kilometers above the surface. Soviet military satellites first recorded changes in the ionosphere in the days before large quakes.
The ionosphere starts to “feel” the zone of pending seismic activity from the preparatory mechanisms of a magnitude 5 event and above. U.S. scientists have not caught on to the foreign ideas but more experimentation and modeling has produced a viable theory that is being tested. First, the physical mechanisms must be studied and understood before any promise of prediction can be examined.
Low altitude satellites have recorded seismo-electromagnetic waves over earthquake-prone areas such as Armenia. The ionosphere disturbances over other seismically active locations have been recorded.
Curiously, the generation of these charges on the ground may not necessarily mean that a massive fracturing of the rock (an earthquake) will occur there. It can instead indicate that the fault movement is blocked. Thus, we can only say that the electromagnetic phenomena indicates rock fracturing with a possible earthquake to follow.
Can we explain EQLs?
Until recently, most scientists rejected the reality of earthquake lights because there was not a satisfactory means to account for their origin. The lights do not appear as regular alarm signals before a quake but are contingent upon whether the required large electrical energy has built up. This only happens when the subsequent quake is large. Lights appear to be evidence that an electrostatic charge is present.
According to calculations by Ikeya, the shape of the glow produced by an intense electrical field generated via underground fracturing would be a dome or ball shape. Relating EQLs to animal behavior, the concentration of air ions can be less in order to produce animal behavior anomalies than that which must occur to produce EQLs.
Can we explain EQSs?
Earthquake sounds may be generated by stress and local fracturing in massive rock. An example of this stress phenomena can be noted when you are near a metal or wood structure on a hot, sunny day. As the air warms or cools rapidly, the change in temperature results in stress in the material that gives a noticeable “crack” on occasion. The material is not visibly damaged but the stress was released. Ultrasound and infrasound might result from rock cracking. Perhaps only some people are sensitive to these frequencies that might be out of the range of hearing for the average person. Some animals may be sensitive to them as well. But sound as a precursor is not very reliable since the sound can be swamped by background noises or dampened within the rock. So, we have some idea about how EQSs might come about but no body of evidence.
Can we explain EQ weather?
The release of gases, formation of aerosols and electrified air might play a part in the formation of anomalous clouds and fogs reported as part of “earthquake weather”. The ionization process can explain a feeing of hot, oppressiveness that hangs over the land. While no particular type of weather causes earthquakes, there may be circumstances in which factors combine to signal changes happening in the earth below.
Can we explain animal behavior?
Animal reaction is most likely to be a combination of several factors. Not all animals are sensitive to the same environmental stimuli. Some are acutely sensitive to smell (dogs) and others are not (birds). Some can sense vibrations but others, such as domestic animals, are surrounded by vibrations and noise that cancel out subtle signals.
Animals can be very sensitive to electric fields. Some have organs specifically for navigating or catching prey using electrical signals. Sharks and catfish, in particular, have extraordinarily sensitive electrosensory systems used to capture hidden prey and for communication, orientation and navigation. Mammals have hair that acts as a sensor for electrical fields. Even feathers, whiskers or antennae may be receiving electrical signals from the environment.
An electric field induces current to flow in the body. The animals, plants, objects and atmosphere may all be responding to the seismo-electromagnetic signals from the epicentral area of the coming quake. The generated electrical fields are strong enough for their local discharges to generate high frequency EM waves. A great number of results show revealing background anomalies in EM emission levels right up to the moment of the quake that may even continue after.
Animals also have been reported to act unusually before and during other catastrophes like storms, tsunami landfalls and during a house fire. These are also examples of where the precursors or early conditions may be perceived by animals but not by humans. Crocodiles in Japan behaved violently prior to an earthquake in the area and had similar behavior prior to approaching storms which may indicate they are responding to EM waves. Because we see parallels in behavior between coming storms and earthquakes, perhaps the underlying reason is also the same.
If we consider the process where the rock fractures on a small scale prior to breaking at the large scale and giving way to the quake, and that this creates EM pulses, then compression of rock in experiments should produce the desired effect. In fact, animal experiments have shown that mice become restless and show signs of fear and distress when in proximity to rock under pressure prior to bursting. Anecdotal evidence also exists for animals to sense rockslides and move away from the affected area days before an event.
Experiments showed that the electrical field values and their effects were consistent between that which might be generated in a seismic event and that which was recorded as affecting animal behavior. Ikeya posits that local stress changes in rock generates charges via the piezometric effect during microfracturing, frictional electricity or fluid flow electrokinetics. As an electric dipole collapses, it produces EM waves in pulses. The animals respond physiologically to the electro-phenomena. By using experimental results on the tolerances of animal behavior to electrical effects, he can estimate the electrical field strength needed to produce such effects. He estimates that a large earthquake, producing six billion watts (a small fraction of a large earthquake), can theoretically produce these effects.
Regarding EM pulses, some animals respond through surface contact and some can perceive it through the air. Ikeya’s experiments showed animals expressed distress when the applied voltage was effectively too low to actually hurt them. His experiments reproduced the reported behavior of animals before earthquakes by using generated EM waves. However, it does appear to depend on the particular species and individual sensitivity with some animals – like mice, rats and parrots – showing odd behavior at low currents. Animals attempted to move away from an electrical field. In addition, they tried to minimize the effect by avoiding water, rubbing or preening themselves in an attempt to relieve irritation, minimize contact with the ground, stay in contact with metal and aligning their body with or against the field. To produce a response, he notes that the earthquake must be greater than M4, the animal must be within 30 km of the epicenter and the intensity of the field must be greater than 1 volt/minute. The mechanism by which animals respond to EM waves is not clear.
The evidence does suggest…
While there are many, many questions that remain, experimental results have shown that the anomalies might be reproducible in a lab or a reasonable theory can be posited for them. In summary, those atmospheric changes can be accounted for if the electrical effects resulting from stressed rock conditions are occurring. There may also be some unknown mechanism at work underground that scientists have not yet measured or accounted for.
State of Prediction
It wasn’t until around 1800 that theories about the causes of earthquakes included the idea of precursors in the research. Precursors, such as water level changes, were just “curiosities of nature”. In the early 1900’s, an instrument, called a “coherer” was used in Italy to detect electromagnetic emissions, probably the first attempt to produce a practical device to recognize precursors before a quake.
Decades ago, the theory of rock dilatancy prior to a quake was tested. Dilatancy is when the rock develops cracks (or ‘dilates’) due to stress. This process might be measured through observance of the following: a lowered velocity of artificially (or naturally) generated seismic waves, ground uplift or tilt, increased radon emission, lowering of electrical resistivity through the rock. After the initial dilatancy (increasing the volume of the rock), there was presumed to be an influx of water into the fault zone as a result. Consequently, the seismic pressure wave speed would return to normal, the electrical resistivity would continue to lower and there would be an increase in the number of small local quakes just before the fault ruptured. Result of employing this theory for prediction were less than stellar and it was, more or less, abandoned, though not invalidated, as seismologists pursued the idea of foreshocks to predict the main shock.
In the last 20 years, the study of changes in electrical fields before earthquakes has made progress first begun in Greece, Japan and France. Seismologists were skeptical. But, the results were valid. Changes prior to earthquakes have been measured and electromagnetic anomalies have been documented. It is still not clear how the measured anomalies are linked with the quake itself, what they mean, and how they can be potential used as a predictive tool.
There are difficulties in measuring electromagnetic changes associated with earthquakes because of all the other sources for these waves – lightning, magnetic storms, artifacts from culture machinery. To eliminate the noise, the best locations to monitor appear to be deep boreholes or the sea floor. That’s not too practical. There are regular variations (hourly, daily, seasonally) in addition to a noisy background in the atmospheric electric parameters from storms, precipitation, winds, dust, etc. These factors complicate the processing of data to determine if a seismic-generated signal is within it.
Radon monitoring is being used to look for a characteristic rise and decrease in radon just before a quake. Observation of water levels is an inexpensively measurable precursor but gives us little information as to when and where the quake might occur.
The most cited example of animals aiding earthquake prediction was during the 1974-1975 lead up to the Haicheng, China quake. Along with the strange animal behavior observed by everyday folk, other ground indicators suggested the quake was near and prompted the government to act. A 7.3 magnitude earthquake occurred, 50% of the buildings were destroyed around the epicenter but there were few human victims. But, animals don’t always react reliably before a quake. Their behavior is not consistently recognizable as odd or indicative of a coming quake. There may be several alternate reasons why animals behave differently than normal. Therefore, animal behavior isn’t the best gage to use to predict quakes.
Satellites have provided us with unique views of our world. Remote sensing equipment that measures changes in the ionosphere is proving to be worthwhile tool to help judge where the next epicenter will be. Ionospheric precursors give a quite reasonable and useful expectation time of 1-5 days. A statistical study done by Chen in 1999 showed that ionospheric precursors occurred within 5 days of a magnitude 5 event 73% of the time but 100% of the time for magnitude 6 quakes. The one- to five-day interval has been well established for ionospheric anomalies. There are complex electrodynamical, meteorological and chemical processes involved in producing an ionospheric disturbance. But satellite studies have clearly indicated the region of the future quake.
A New Science
An earthquake of magnitude 5.7 occurred near Coyote Lake California in August 1979. The area was loaded with geophysical instruments. Not a single precursor was identified via these instruments. However, the local spring experienced a change in water level and some abnormal animal behavior was reported. Along with the Parkfield experiment to capture an earthquake that finally occurred in 2004 (with no obvious precursors), hopes for prediction waned. What are the precise conditions under which an earthquake preparation area exhibits precursory activity? Not only are these conditions unknown, but the actual occurrence of precursors is still widely doubted by seismologists.
If charged particles are indeed released from the ground prior to a large quake, what might be measured to help in prediction? Release of gases, positive ions near the ground surface, change in the atmospheric electrical fields above the fault area, change in the vertical stream of ions into the atmosphere, increase in EM radiation, appearance of electrical earth currents, changes in the electrical potential of groundwater or surface waters. The trick is what to measure and how to do it.
Much has been learned about the earth signals before a quake. The most important may be the electrical effect. Ikeya hopes that recent progress will spawn a new discipline called “electromagnetic seismology”. The most interesting aspect is how wide-ranging the effect may be. It was previously assumed that any changes in the ionosphere were caused by environmental variability, geomagnetic storms and the like. Now, the thought is that seismic activity around the globe may play an important role in its variability.
Lithosphere-atmosphere-ionosphere coupling is a very complicated subject involving an array of physical effects and interactions on all levels from underground to the earth’s magnetosphere. The volume of knowledge is so large, it is hard to research the topic in all directions. Starting in the 1930’s, from observation of seismogenic electric fields, the idea of connecting the lithosphere effects with the atmosphere has been a zone of conflict. Because of the interdisciplinary aspects, the field is off limits to many scientists. The theory requires knowledge of tectonics, seismology, atmospheric and ionospheric physics, chemistry and electricity. Discussions between experts in these various groups ends up in complications and disagreement.
For short-term prediction and accuracy, we are hardly farther along than the ancient observers were. However, the new theory of seismo-ionospheric coupling is very promising. Russian scientists, such as S.A. Pulinets have called for a satellite system with ground-based measurements to analyze the anomalies and possibly turn them into a predictive method. Measuring only one parameter will not give enough confidence for prediction.
U.S. Scientists are now examining this idea. An ionospheric perturbation was produced by the Coalinga, California earthquake of May 2, 1983, detected by a network of high-frequency radio links in northern California. If we look for more, it seems likely that we will find more evidence to support this phenomenon.
The topic of earthquake prediction highlights the value of reported anomalies. We have seen how many anomalous observations by average citizens and seasoned experts formed the basis for learning valuable lessons about earth. A unifying theory to explain the reports gives them credibility. While not all the anomalies can be adequately explained, it is the hope of those who report and study them that one day they will fit within a scientific explanation. When scientists, such as Tributsch and Ikeya, proceeded with their research and publication, they were met with rejection from other professionals who did not judge citizen observation worthy of scientific research. Delving into these topics mean grants are hard to come by, your professional reputation can become tarnished. But, the public and mass media can be your strongest support. They expect science to get to the root of these stories. Scientists are reluctant to leave the safe environment of their practice. That attitude undermines the strong public interest in the phenomenon. Curiously, cultural differences may play a role with the western scientists less open to these ideas that may taste of superstition, while other cultures have different thoughts.
The public expects progress in science to develop ways to make them safer. Science progresses in pulses. The previous failures in EQ prediction do not necessarily mean that it can’t be done, it means we may be looking in the wrong places for answers. It does appear that there were many cases where we had adequate data prior to the quake but it was not properly used to save lives. As we saw in the Asian tsunami disaster, a coordinated effort is critical to success.
As a scientist, a geologist, I am admittedly out of my range of expertise when it comes to understanding the concepts and theories associated with these emerging ideas about the lithospheric-atmospheric connection. But, after much musing, it makes sense that processes on earth are interconnected. It does give me hope that the eyewitnesses and experimenters ridiculed by science and the anomalous observations once rejected, are now being accepted as valid. It is heartening to see that we may be on a path now to understand how the earth alerts us to catastrophic events and how we can use the signals along with personal precautions to minimize or eliminate the associated suffering and death. That’s the ultimate purpose of science.
BBC News.(2006) “Plankton Blooms linked to quakes”, May 9, 2006 from http://news.bbc.co.uk/go/pr/fr/-/1/hi/sci/tech/4750557.stm accessed on May 12, 2006.
Bolt, Bruce A.(1993) Earthquakes, W. H. Freeman and Company: New York.
Corliss, William R. (1983) Earthquakes, Tides, Unidentified Sounds and Related Phenomena, The Sourcebook Project: Glen Arm, MD
Corliss, William R. (1995) Handbook of Unusual Natural Phenomena: Eyewitness Accounts of Nature’s Greatest Mysteries, Gramercy Books/Random House
Geller, Robert J. et al. (1996) “Earthquakes Cannot Be Predicted”, Science 275 (5306): 1616.
Gokhberg, Morgounov, and Pokhotelov (1995) Earthquake Prediction – Seismo-electromagnetic Phenomena, Gordon & Breach Publishers.
Hough, Susan E.(2005) “Earthquakes: Predicting the Unpredictable?”, Geotimes, March 2005.
Ikeya, Motiji (2004) Earthquakes and Animals: From Folk Legends to Science, World Scientific Publishing Co., Pte. Ltd.: Singapore.
Martinelli, Giovanni.(1998) “Earthquakes, Prediction” in Sciences of the Earth: Volume 1, Edited by G.A. Good. Garland Publishing, New York.
O’Hanlon, Larry (2003) “Earthquake Warnings in Ionosphere?” Discovery News (Discovery Channel), March 27, 2003.
Pulinets, S.A. (1998) “Strong Earthquake Prediction Possibility with the Help of Topside Sounding from Satellites”, Advances in Space Research, Vol 21 No 3.
Pulinets, S.A. (1998) “Seismic activity as a source of the ionospheric variability” Advances in Space Research, Volume 22, No 6. http://www.izmiran.rssi.ru/~pulse/COST%20SEISMO.pdf
Pulinets, Sergey and Kirill Boyarchuk (2004) Ionospheric Precursors of Earthquakes, Springer-Verlag: Berlin Heidelberg.
Serebryakova, O. N., S. V. Bilichenko, V. M. Chmyrev, M. Parrot, J. L. Rauch, F. Lefeuvre, and O. A. Pokhotelov (1992), “Electromagnetic elf radiation from earthquake regions as observed by low-altitude satellites”, Geophys. Res. Lett., 19(2), 91–94.
Tributsch, Helmut (1982) When the Snakes Awake: Animals and Earthquake Prediction, MIT Press: Cambridge MA.
U.S. Geological Survey http://earthquake.usgs.gov
Veysey, John (2004) “Icelandic Water gives clues about quakes”, Milwaukee Journal Sentinel, Aug 8, 2004 http://www.jsonline.com/alive/news/aug04/249223.asp accessed Aug. 10, 2004.
Wu, C.(1997) “Impurities give crystals that special glow” Science News Online, May 17, 1997 http://www.sciencenews.org/pages/sn_arc97/5_17_97/fob2.htm.