Haicheng, China, 1975. A massive earthquake hits. Buildings are demolished, roads destroyed, but thanks to an evacuation several hours earlier, thousands, possibly tens of thousands of human lives are saved. The Chinese claimed they’d predicted an earthquake within hours of its occurrence. Their forecasting system: animals.

This success sparked the interest of the U.S. Geological Survey. What were animals cuing in on? How did their detection systems work? Could answers to these questions lead to the development of a high-tech earthquake forecasting system?

Dr. Benjamin Hart, a veterinary behaviorist at the University of California, Davis, School of Veterinary Medicine, was on the front lines. “I received a call from the USGS,” says Hart. ” `Can you tell us what it is that animals respond to in an earthquake?’ the caller asked, almost expecting that I knew the answer already. I told him that I hadn’t the slightest idea and that I was sure none of my colleagues knew either. First tell me what the stimuli are.”

Early Warning Signs

The caller rattled off different types of stimuli—radon gas, magnetic lines, electron particles, seismic foreshocks. While animals have remarkable senses, Hart knew of no examples of animals responding specifically to earthquake stimuli, so he suggested a meeting of geologists and experts in sensory systems in animals. And so it was. A think tank of scientists ranging from seismologists to experts in pigeon homing and hearing in lizards convened to ponder how or even whether animals might detect earthquakes.

Says Dale Lott, professor of conservation biology at UC Davis, “I was pretty skeptical about animals having a specific earthquake detection ability. In evolutionary terms, why should animals care if there’s an earthquake? So the Earth shakes a little. The food still grows, and the streams still flow.”

But Lott and the others did feel that animals might detect precursors as something odd without knowing specifically that an earthquake was coming. In this case, you would expect to see a species-specific anxiety or fear reaction.

The symposium brought forth many ideas and a call for research. Hart and Lott teamed up with geologist Ken Verosub and proposed several projects, one of which received funding—an interview of earthquake victims just after an earthquake.

Their first research opportunity was the 1977 earthquake in Willits (Mendocino County).

Posing as an earthquake survey team so they wouldn’t bias the interviewees, they questioned victims about damage to their homes first. Then, well into the interview, they asked, “Did you have any idea this was going to happen? Did anything unusual occur before the earthquake?”

Usually if they were going to get anything about animals, they’d get it here. The results were quite interesting.

“It was a melange,” says Lott. “Ben would do some interviews and get a lot of positives, and then I would do some and get nothing and vice versa.”

Positives were marked: A cat that normally entered the house about 7:30 a.m. to eat and sleep continually paced, entering and leaving the house repeatedly. The owners had discussed the unusual behavior before the earthquake.

A 2-year-old Doberman that normally slept in the morning shadowed her owner continually from 8 a.m. until the earthquake, sometimes whining and pacing. The owner wondered whether the dog needed a tranquilizer.

Clusters of Awareness

When the researchers finally were informed of the exact epicenter, they examined the mapped interview locations and found that the positives were clustered around the epicenter. An incredible 50 percent of the households around the epicenter reported strange animal behavior, whereas the baseline for positive answers far away from the epicenter was only 10 percent.

Both Lott and Hart were impressed with the pattern. Maybe they were onto something. But the excitement ended there.

Disappointing Discoveries

Over the next year, they studied six more earthquakes spanning North, Central and South America and came up with nothing.

Their conclusion was that animal earthquake detection works sometimes but not often. Some earthquakes are preceded by cues that some animals can detect, but since this phenomenon is not consistent, it’s not reliable enough to be useful in predicting earthquakes.

While the findings may have been disappointing, they were still valuable. Says Lott, “It stopped the folklore about animals having specific earthquake-predicting abilities and allowed the geologists to go on to study other forms of earthquake prediction. It was an idea that had to be rigorously tested.”

And what of the Haicheng earthquake? Turns out that it was a propaganda act by the Chinese government during the Cultural Revolution madness. The government never produced scientific evidence supporting its claim. Furthermore, shortly before the Haicheng earthquake, there had been many smaller earthquakes in the area. General warnings were issued many times, and people were encouraged to stay outdoors. One such warning happened to coincide with the Haicheng earthquake.

A more recent earthquake in southwest China, in 2008, left 87,000 people dead or missing.

© Sophia Yin, DVM. Originally appeared in the San Francisco Chronicle in 2000.