16 Jan 2018

Ants have an(other) unexpected superpower

In 2013, Gabriele Berberich’s seminar on earthquake prediction at the European Geosciences Union annual meeting, in Vienna, caused shock waves among the audience. Berberich, a geologist at University Duisburg-Essen, in Germany, presented tantalizing results from her fieldwork showing that red wood ants (Formica rufa-group) change their behaviour hours before an earthquake. Shortly after Berberich gave her presentation, headlines in the mainstream media were boldly claiming that ants sense earthquakes several hours before they strike. This was huge. Despite decades of research, geologists continue struggling to predict earthquakes more than a few minutes in advance, which is obviously not long enough to evacuate people from affected areas. Could these ants come to save the day?

Red wood ant (Formica rufa)
Schreiber discovered that red wood ants make their nests along a particular type of faults, called strike-slip faults. These are vertical cracks in the rock that can act as a chimney for liquids or gases ascending from the Earth’s crust or mantle. These faults don’t usually occur as a single, clean fracture, but are rather found in complex networks of cracks and deformations in the rock. For this reason, geologists can struggle to accurately map them. Typically laborious soil gas measurements and structural geological mappings are necessary, which are time-consuming and costly. And this is why Schreiber’s discovery was so exciting.

Tales of abnormal animal behaviours before an earthquake have been around for centuries. In 373 B.C., Diodorus, a Greek historian, reported that rats, snakes, weasels, centipedes, worms, and beetles, migrated in droves a few days before a violent earthquake hit the city of Helice, in Greece. Accounts of animals in panic seconds to hours before an earthquake are the most common, for example, dogs barking or whining, nervous cats (sometimes jumping out of windows), bees leaving their hives, cows producing less milk, to name a few.

China and Japan are by far the countries that have invested more resources into studying a connection between animal behaviour and earthquakes. In 1975, an earthquake of magnitude 7.3 was predicted months in advance in China based on accounts of weird animal behaviours, including snakes coming out of hibernation and freezing on the ground surface and rats leaving their holes en masse. The city of Haicheng, with a population of a million people, was evacuated days before the earthquake, saving thousands of lives. However, this earthquake was preceded by many smaller temblors (called foreshocks) over several months, which was probably what really convinced Chinese officials to order the evacuation. So, despite nearly four decades of research, there is still no evidence that abnormal animal behaviour can be predictive of earthquakes, well, at least not reliably.

Scientists remain skeptical because reports of this kind are typically collected after the quake, so people are biased to recall behaviours that might have always been there—they just didn’t notice them. The other reason is that while it is easy to explain why animals sense earthquakes seconds before they start (they can feel a type of weak seismic wave that reaches the Earth’s surface seconds before stronger, more noticeable waves), it’s more difficult to understand how they may do this days or weeks in advance. Some scientists have suggested these animals detect electromagnetic changes or gases released from the ground before earthquakes, but most animals definitely can’t do this. For ants though, it’s a different story.

Can ant behaviour predict earthquakes?
Red wood ants are remarkably sensitive to their environment. Besides detecting the tiniest changes in temperature, ants can respond to electromagnetic  fields and have protein receptors that sense changes in carbon dioxide (CO2) concentrations. So could these ants sense earthquakes long before they start? Berberich and Ulrich Schreiber, her then PhD supervisor at University Duisburg-Essen, thought it was worth having a look. 

During a 3-year survey, from 2009 to 2012, Berberich, Schreiber and colleagues filmed two red wood ant mounds that were on top of a seismically active tectonic fault. They used high-resolution cameras in colour and infrared to monitor the ants’ activity 24/7. After processing over 45,000 hours of video streams with an automated software that tracked the tireless ants, the team discovered that these hard-working insects have a daily routine that resembles our own: during the day they are franticly busy at work, and during the night they rest inside their nests. Initially the researchers thought this activity pattern was only perturbed when unwelcomed guests visited their nest, such as birds or badgers looking for a snack. But gradually it became clear that something else was occasionally upsetting the ants: earthquakes.

Berberich, Schreiber and colleagues found that hours before an earthquake, the ants stopped their daily routine and just hung out outside the nests until the next day. They observed these behavioural changes only before earthquake events with magnitudes between 2 and 3 (smaller temblors didn’t seem to bother the ants), but they still don’t know how the ants would react to stronger earthquakes.

“Our observations […] took place in a region with low seismic activity,” says Schreiber. “It would be better to make this research in Italy or other regions with higher seismic activity but we got no funding for this type of research.”

Another problem was that during the winter the ants remained mostly inside the nests, and it’s more difficult, if not impossible, to detect their activity over the cold months. So, whether red wood ants can be used to predict earthquakes remains an open question, though at present that seems rather unlikely.

“[…] Red wood ants cannot predict earthquakes. We never stated that in our publication” said Berberich in an email to LabTimes. “Earthquake prediction would mean providing information on the exact date, time, location and magnitude. That is definitely not possible.”

Nonetheless, these tiny creatures turned out to have yet another superpower that could be incredibly useful for humans.

Nesting on active tectonic faults
Shortly before Berberich started her research on red wood ants and earthquake prediction, Schreiber had discovered that these insects have a strange preference for building their nests on top of active tectonic faults.

Faults are fractures in the rock that can extend into the Earth’s crust or mantle. Because the rocks on either side of these cracks cannot slide against each other easily, tension gradually builds up until the rocks suddenly break (and slide). The energy released propagates to the Earth’s surface in seismic waves—and this is what causes earthquakes.

Schreiber first noticed that the ant mounds lined up along tectonic faults back in 2003 during fieldwork in the Rhenish Massif, in Germany, but because this was so unusual, he spend a few more years collecting data in different sites in Germany, Austria and Southern Scandinavia. And yet, despite counting over 1000 ant mounds, “it needed a long time to convince biologists, reviewers and colleagues,” Schreiber says.

Nest of red wood ant

After these findings were finally published in 2009, Schreiber and Berberich, who in the meantime had joined his lab, continued collecting data in other sites in Germany. But even after mapping more than 3000 ant nests and consistently finding a clear statistical correlation between nest position and the active faults, the scientific community remained unconvinced.

“One of the challenges of this work has been convincing ecologists and myrmecologists [people who study ants] that there actually is a spatial relationship between ants and fault lines,” says Aaron Ellison, an ecologist from Harvard University, USA, who has collaborated with Berberich for a number of years. “The major critique is that we don’t really have a good biological mechanism for why ants should be related to faults.”

“Ants are easier to sample and identify than geological features,” Ellison says. “Ants may also be more sensitive at identifying very small or newly opened fault systems, and more quickly, than we might get with a geological survey.”

Reliable bioindicators
Red wood ants are well-known bioindicators (organisms used to monitor changes in the environment). In Australia, these ants have been used successfully to check pesticide contamination in cotton wool-producing areas, for example, and to measure the environmental impact of mining, deforestation and logging, and urbanisation.

To convince their colleagues that red wood ants can be used as reliable bioindicators for active faults, Berberich and colleagues decided to carry out a double-blind study. This idea was suggested by Ellison, who at first was one of Berberich’s strongest critics. When they met at the annual meeting of the Entomological Society of America in 2012, Ellison pointed out that not enough sampling had been done in localities where there wasn’t tectonic activity, so the sampling could have been biased.

“It’s easy to say that, there’s a fault here and there are ants here, so they’re related to one another, but to actually do a very careful and intensive spatial sampling and see if the samples are not biased is harder to do,” says Ellison.

Berberich and Ellison began their collaboration and applied for funding to do the double-blind study. They finally started the fieldwork in 2016 with a grant from the Volkswagen Foundation. Ellison recruited several students, who hadn’t previously been involved in the research, to sample ant mounds over an area of about 1400 km2 in Denmark. But there was a twist.

“[…] We didn’t tell them what they were sampling, or why they were sampling it,” says Ellison.

And not only were the students not aware of the study’s purpose, but also none of the researchers knew where the active faults were located. After mapping nearly 300 ant mounds, the team got hold of the detailed geological maps of the region. A careful statistical analysis confirmed that the ant mounds really are closer to active faults than what you would expect by chance alone, and these results were recently published in the journal PeerJ.

Double-blind experiments are designed to avoid conscious and unconscious biases in research. Although this robust method is widespread in psychology and clinical research, it is virtually absent from ecological studies. Recent alarming numbers in retractions and fraud cases resulting from the ‘publish-or-perish’ culture, however, are raising awareness for the need to improve (and test) reproducibility in science.

A recent survey in the journal Nature showed that over 70% of researchers failed to reproduce someone else’s experiments. Perhaps even more worrying, more than half of the 1576 researchers who took part in the survey said they couldn’t reproduce their own experiments. Two notorious studies tackling this problem showed that the scientific literature in psychology and cancer research is, respectively, only 40% and 10% reproducible. But despite this bleak picture, studies reproducing or failing to reproduce experiments are rare and often difficult to publish.

“We keep talking in science about how reproducibility and replication is so important, so here was an opportunity to really put that to the test,” Ellison remarks. “By doing a double-blind experiment, we were able to show that there really is a relationship here in need of an explanation and further work.”

A brilli(ant) gas sensor
The million-dollar question now is why red wood ants fancy building their nests on active strike-slip faults.

“I think that it’s a combination of different aspects: metabolism, defence against enemies due to increase the CO2 concentration in nests, CO2 content necessary for beneficial bacteria, or others,” says Schreiber. “We don’t know exactly and need more time for research.”

Red wood ant nests release large amounts of CO2 and are warmer than the surrounding ground, but it’s not known why. Berberich and Schreiber’s findings open the possibility that these ants choose to build their nests where there’s more CO2 available in the ground, like on top of strike-slip gas-permeable faults, for example. Schreiber is currently following up preliminary observations that support this hypothesis, while Ellison and Berberich, who is now working at the Technical University of Dortmund, are applying for funding to study whether red wood ants sense temperature and gases (radon, methane, helium…) released from faults.

“We really don’t know why [ants build their nests on active faults] but we’re working on promising hypotheses, and all of them will be improved by doing double-blind experiments,” Ellison concludes.


Reference: Toro​, Israel Del, et al. “Nests of red wood ants (Formica rufa-Group) are positively associated with tectonic faults: a double-Blind test.” PeerJ, PeerJ Inc., 12 Oct. 2017, peerj.com/articles/3903/.

This article was published in the printed issue of Lab Times on the 29th November 2017.