It has long been suggested that animals ‘know’ when an earthquake is about to occur: changes in behaviour have been noted in laboratory mice, daily rhythms of ants have reportedly been disrupted, and cows have been observed to behave unusually (in one case an entire herd of cows was witnessed lying down in unison before an earthquake struck). There were reports of elephants and flamingos heading to higher ground before the 2004 Boxing Day earthquake and tsunami, and more recently of zoo animals acting strangely before an earthquake that struck Washington, D.C. One of the earliest reports of animal behaviour predicting earthquakes is from Greece in 373 BC, when rats, weasels, snakes, and centipedes were said to have left their usual homes several days before it struck.
Skeptics on the other hand have suggested that these reports can be dismissed as examples of confirmation bias, where incidental correlations between animal behaviour and earthquakes are remembered and falsely attributed to the existing folk belief, while the many non-manifestations of such behaviour are forgotten.
A particular skeptical mantra is that ‘the plural of anecdote is not data’ – a comment on the non-evidential value of anecdotal reports. When it comes to animal behaviour before earthquakes, however, a new study has gathered actual data on the topic, and perhaps may have even contributed to substantiating the long held belief that animals can sense an earthquake coming.
The researchers took advantage of nine ‘camera traps’ being used in Yanachaga National Park in Peru to track the movements of rarely seen animals. Each time these traps’ motion sensors are triggered, they take a picture, capturing an image of the animal that is moving past the field of view.
Analysing the images over a 30 day period leading up to the 2011 magnitude 7.0 Contamana earthquake, (and comparing with a ‘control’ period not associated with an impending earthquake), they found that the camera traps usually ‘captured’ up to 18 animals a day. However, this number dropped off to much lower numbers consistently around 23 days before the earthquake, and then reduced further just over a week from it striking. In fact, only three animals in total were photographed in the last six days before the earthquake struck, with rodents – the most abundant animal in the forest environment – almost completely disappearing.
In the paper, the researchers theorise that the changes in behaviour might be caused by the sub-surface grinding of rocks in the lead-up to an earthquake, creating an electric charge that has a number of effects which might be sensed by animals:
Emission of ultra-low frequency (ULF) electromagnetic waves that may affect biochemical reactions and disrupt circadian rhythms.
Oxidisation of soil organics, creating toxic and/or irritating trace gases, such as carbon monoxide.
Ionisation of air molecules – which has been reported to cause blood serotonin levels to increase in animals and humans.
The researchers concluded:
An enhanced air ionisation at the ridge prior to the magnitude 7 Contamana earthquake may have caused the animals to escape to lower altitudes, where they would have been exposed to fewer positive airborne ions. The pre-earthquake anxiety, restlessness and escape reactions of domestic or captive animals, reported anecdotally for many decades, even centuries, may simply be due to the fact that confined animals tend to panic when they are unable to move away from aversive stimuli in their environment. If this correlation can be substantiated by systematically monitoring a wider range of reported pre-earthquake phenomena, this would lead to a better understanding of the premonitory abilities of animals.