Looking at 180 scientific publications, many of which came from the 2010 Canterbury earthquake in Darfield, researchers categorised the findings as best as they could. Sorting by factors including how far the animal was from the epicentre, the magnitude of the earthquake and the quality of the observations made, the study was able to sort through more than 700 records from 160 earthquakes and more than 130 species of animal.
The data included case studies that recorded ants behaving oddly, toads changing their migration pathways and elephants moving to higher ground.
Some animals were found to react close to the epicentre seconds before an earthquake struck, whereas others seemed to alter their behaviour weeks before and hundreds of kilometres away from where an earthquake eventually hit.
The randomness and lack of consistency within the data collected highlights the challenge around studies on animals and earthquakes. Without being able to know where, when and how big the next earthquake is going to be, there is currently no way to set up a controlled experiment that can place different animals close to and far from the epicentre of an earthquake to monitor their behaviour.
Because of this, at the end of the analysis, the only thing the study was able to positively conclude was that the data wasn't really reliable enough to make a solid conclusion.
What we do know about animals, though, is that many species could possibly pick up earthquake warnings that humans can't.
For example, some species such as catfish can detect changes in electromagnetic fields caused by seismic activity, which can occur days before an earthquake.
Other animals such as dogs can hear very low sound frequencies that may allow them to pick up the initial shifting of the Earth's plates. Even elephants have the ability to pick up subtle ground vibrations through their feet, which may occur as primary seismic waves up to 90 seconds before the shaking starts.
Another possibility is that animals might be detecting the secondary effects triggered by foreshocks such as changes in groundwater levels or the release of gases from the ground.
Whichever it is, if any, we just don't have enough continuous, long-term observational data of animals experiencing earthquakes.
There is still no way to scientifically tell if the anecdotical observations seen with animals are due to an earthquake or some other environmental or health change affecting the animal.
As a shaky nation, we are always looking for ways to help to predict an oncoming earthquake, but for now, using your four-legged friend is probably not the most reliable method out there.
