In the last few weeks, interest in a YouTube video purporting to show bison fleeing Yellowstone National Park has rocketed around the Internet. The video, which has racked up nearly two million views since publishing on March 20, is titled "ALERT! Yellowstone Buffalo Running for Their Lives!"
The animals are supposedly evacuating themselves in anticipation of an eruption at the park, which sits on a huge volcanic system. Curiosity about the claim was surely fueled by the recent magnitude 4.8 earthquake last month, the largest at the park in 30 years.
But the warning turned out to be a hoax: Yellowstone released its own counter-video with an affable staffer explaining that it’s not unusual for wildlife to leave the park during the winter to seek food. Plus, one spokesman pointed out that the bison featured in the video are actually running into the park—that is, towards the volcano.
For millennia, humans have been obsessed with the possibility that animals can anticipate natural disasters, especially phenomena like earthquakes. It’s often dismissed outright as myth or misunderstanding, certainly a safe conclusion in the lack of clear evidence. As the U.S. Geological Survey’s brief, informative webpage on the concept notes, an animal’s reaction in the moments before an earthquake may simply be the result of a finer sensitivity to earthquake waves. The USGS is appropriately skeptical—but it doesn’t close the door on the possibility.
There have been studies on the animal-earthquake connection for decades, particularly in the years following the 1975 Haicheng earthquake in China. In the hours before the quake, some local authorities ordered evacuations that proved effective in saving lives. It’s considered the only “predicted” earthquake, even if the details themselves—exact time and magnitude—were not themselves predicted.
A variety of possible signals were observed leading up to the Haicheng quake, among them a series of strange animal behaviors, such as hibernating snakes leaving their burrows only to freeze aboveground, or a group of mice that seemed dazed and unafraid of humans (even as somebody started killing them).
Whether these observations actually meant anything substantial is debatable. In Haicheng, amateurs reported a number of animal observations after the quake, sometimes in answer to researchers who asked them if they noticed anything odd. How much of that evidence, like any post-quake recollection, was the result of suggestion, a sort of “Now that you mention it…” retroactive observation that fits the narrative? After all, the area also experienced a series of what are now considered foreshocks—that is, smaller earthquakes—before the big 7.3 earthquake. Seismologists have long concluded that those are what really triggered the predictions.
Despite a lack of direct evidence connecting animals and earthquakes, researchers continue to explore the connection. In the late ‘70s, the U.S. Geological Survey put on a pair of conferences to discuss the issue. More recently, researchers have published papers describing behavioral changes in lab mice before the 2008 Wenchuan earthquake in China, and in toads who abandoned their spawning ground before the 2009 L’Aquila earthquake in Italy. One German group is analyzing 45,000 hours of video data as part of its effort to find any correlation between earthquakes and behavioral changes in ant colonies.
But these examples are unusual. The toads, for example, were observed regularly as part of a multi-year biological study, and the L’Aquila quake happened to occur during that time. Those researchers were already keeping a good record that could be compared pre- and post-earthquake. Most of us can’t sit around staring at Rover, taking careful notes, just in case an earthquake strikes.
And ultimately, these observations are anecdotal. How can we be sure that oads are good earthquake predictors? Is it all toads? Or just these toads? Or was it something about the earthquake or the local environment? As much as we want to believe there is a connection, when it boils down to reliable, specific signals, we don’t necessarily know what we’re looking for, both in terms of the behavior—what constitutes abnormal?—as well as what, if any, clues emanate from the earth before a quake that could trigger these behaviors (such as changes in the terrain, moisture, water chemistry, or electromagnetism, subtle factors for which animals might have developed a sensitivity).
And if we if we could isolate those natural signals, might it be more efficient to develop our own electronic detectors than to start stalking whatever creatures respond to it?
Next year, a consortium of international researchers plans to begin a large project to monitor animal movements from space. Among the data on migration and ecosystems, the group notes that it may also turn up information on connections between behavior and disaster forecasts.
After all, any improvement in understanding when such an event might occur, no matter how strange or unlikely, could yield a lifesaving benefit. And so we keep looking for signals in the noise.