Can animals sense earthquakes?

For centuries, unusual animal behaviour before earthquakes has been reported worldwide. Livestock becoming restless, wildlife disappearing and snakes emerging from hibernation in the middle of winter. For a long time, scientists dismissed such observations as folklore.

In recent years, however, systematic research has begun to explore whether animals genuinely respond to environmental changes preceding major earthquakes. Although earthquakes are hard to predict even for humans, several studies suggest intriguing patterns in animal behaviour before seismic events.

As the world population increases, more people will be affected when earthquakes happen, making this research more important than ever.

My own research journey began with a serendipitous observation in Italy. I was studying the effects of moon phases on toad reproduction at San Ruffino Lake in 2009, when the toads disappeared for five days. They returned only after a magnitude 6.3 earthquake struck the city of L’Aquila, about 50 miles away.

This observation formed the basis of my 2010 study showing that 96% of common toads abandoned their breeding site five days before the 2009 L'Aquila earthquake. It was one of the first studies to quantify a shift in wild amphibian behaviour before seismic activity. Amphibians’ permeable skin makes them especially sensitive to changes in water chemistry which could make their behaviour a potential early warning of seismic activity.

I also conducted a multi-species study of Yanachaga National Park, Peru, before a major earthquake in 2011. A charity called Wildlife Insights (formerly Team Network) places cameras in many locations in national parks for conservation monitoring. I looked for parks where a large earthquake had occurred and analysed the charity’s photographs for Yanachaga National Park.

The motion-activated cameras recorded a sharp decline in animal activity in the weeks leading up to the quake. Daily counts fell from typical values of around five to 15 separate animal records per day to fewer than five, across all seven orders of vertebrates in the forest. In the final 24 hours before the quake, animal movements completely ceased.

I compared records from around the time of the earthquake to seismically quiet periods in the same season. I found that during less seismically active times, animal numbers stayed constant.

In Peru, the steep decline in activity was pronounced not only in small and medium sized rodents such as pacas and capybaras but also in bigger animals like long nosed armadillos. This “silencing” of the forest suggests that earthquake-related cues affect entire animal communities rather than just one species.

It’s not just wildlife

Research has shown that livestock around the world, particularly cows, also show signs of pre-seismic behavioural and physiological change.

There are numerous reports of cows panicking and wandering around in areas where they would not normally be seen. For example, stories that cows converged on San Francisco’s Chinatown in 1906 prior to a large earthquake which killed 3,000 people. In 2012, a blog post circulated on the internet showing photographs of cows entering a suburb of Malaysia’s capital city, Kuala Lumpur, and feeding in gardens, two days prior to a magnitude 8.6 earthquake off the coast of Sumatra.

Several Japanese studies have monitored dairy cows using automated milking and activity systems. These studies have reported modest but statistically significant reductions in milk yield and changes in rumination or restlessness in the days preceding some local earthquakes.

Pets seem to be affected too. In 2011, a massive magnitude 9.1 earthquake struck off the northeast coast of Honshu in Japan, generating a tsunami that disabled three nuclear reactors. Post earthquake questionnaires surveyed 1,259 dog owners and 703 cat owners about their pet’s behaviour before the earthquake. About 19% of dog owners and 16% of cat owners reported unusual behaviour. Restiveness was a dominant behaviour in both species, usually within one day prior to the quake. It’s important to note though, that post-event recollections are not considered as scientifically robust as data collected in real time.

What might animals be sensing?

The key question is not whether animals behave differently, but why.

One leading hypothesis, proposed by Friedemann Freund (a scientist for NASA), focuses on environmental changes caused by stress building up in rocks as tectonic plates shift, prior to large earthquakes, releasing electrically charged particles.

These particles can alter the properties of air and soil in the area by increasing the number of positive airborne ions (electrically charged molecules) and appear to affect stress levels and behaviour in animals (including humans). More research is needed but the phenomenon may help explain the changes in animal behaviour before the Italian and Peruvian earthquakes.

However there are many other cues which could contribute to unusual animal behaviour before earthquakes. For example vibrations, disturbances to the local electromagnetic field or sounds outside of human hearing range. We still don’t know exactly which signals, or combination of cues, explains the behaviour.

Despite growing evidence that animals can sense environmental changes preceding earthquakes, the scientific community remains cautious. Several studies have found unusual animal behaviour before earthquakes could later be explained by normal seasonal activity.

Then there’s the fact that earthquakes are rare, which makes the phenomenon difficult to study. I believe animals simply move away from unpleasant or unusual environmental changes, rather than “predicting” earthquakes.

Of ants and earthquakes

There are ongoing studies that may help us learn more about animal behaviour and earthquakes. A systematic trial called Animal Alerts is underway in Lima, Peru, an area with a high level of seismic activity. Researchers have fitted dogs with smart collars which record their heart rate, movement and other parameters in real time.

A 2013 study carried out long-term observations of red wood ant mounds on active faults (cracks in the Earth’s crust that have recently moved and may cause earthquakes). The researchers reported alterations in daily activity rhythms of the ants living on these fault lines. Building on this work, my postgraduate research student, Shanza, is studying earthquake precursors for her master’s degree. She aims to identify which animal species are most likely to respond to early earthquake signals such as positive ions or magnetic field fluctuations. She then plans to simulate some of these conditions in the lab, using ants as a model species.

Animal data alone are unlikely to give reliable earthquake warnings. But the more we can combine animal data with environmental measurements, the closer we will come to reliable forecasts of earthquake hazard risk.

Rachel Grant does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.