SCIENCE / MEDICINE : Can Animals Really Predict Earthquakes?


In January, 1975, Chinese fieldworkers and woodcutters near Haicheng, a city northeast of Beijing, were startled when hundreds of snakes suddenly appeared from their hibernation holes and quickly froze to death in the snow.

At the same time nearby farmers reported strange behavior from livestock. Cows and horses refused to enter barns and repeatedly attempted to break out of their corrals. Chickens ran around in a frenzy and hens did not lay eggs for a week. Dogs barked day and night, and cats bristled and would not stay indoors.

Scores of fish were seen leaping from rivers and lakes, and wild animals such as tigers, deer and birds were reportedly seen migrating from the area.


Residents also reported sizable changes in the water level of their backyard wells. Within a few days scientists began to chart telltale changes in ground tilt and the Earth’s magnetic field.

On Feb. 3 alarmed officials evacuated residents from Haicheng. A few hours later, a 7.3 magnitude earthquake rumbled through the city--destroying or damaging 90% of its buildings. Few human casualties were reported; thousands of lives had been saved by the timely warnings.

The Haicheng experience sent a tremor of excitement through the scientific world, which has long sought a way to predict earthquakes.

Interest in the link between abnormal animal behavior and earthquakes soared. The U.S. Geological Survey launched three projects to determine by scientific method what has been maintained in folklore and rumored for hundreds of years--that animals can sense impending temblors.

The three studies conducted in California in the late 1970s included observing captured mice and kangaroo rats in an active earthquake area; recruiting 1,700 volunteers in seismically active areas to report unusual animal behavior via a telephone hot line, and interviewing residents living near five earthquake epicenters. Although some unusual animal behavior was noted, all three studies proved inconclusive.

“An experiment has to be repeatable to be credible in science,” said Don Dupras, associate geologist for the California Department of Mines and Geology. “The problem is, how do you set up a controlled, repeatable experiment given the random nature of earthquakes?”


One of the attempts to assess animal behavior was conducted by researchers from UCLA, who built a simulated outdoor habitat for about 50 pocket mice in the Morongo Valley about 30 miles north of Palm Springs. Pocket mice were used because they do not drink water and were therefore easy to isolate. “We had one cage above ground and a connected cage buried under the ground,” said UCLA research biologist Robert Lindberg. “It was as close to a natural environment as we could get.” In addition, several kangaroo rats were monitored in indoor cages.

On March 15, 1979, a swarm of earthquakes struck the area, the three largest measuring magnitudes of 5.2, 4.9 and 4.8. More than 35 smaller quakes were also recorded. Researchers reported that the ground moved for as long as 20 seconds during the larger quakes.

“To our obvious delight, significant, and in some cases dramatic, activity anomalies were recorded from most animals in both the indoor and outdoor activities,” the UCLA researchers reported. However, the report concluded that the increased activity could have been attributed to human intrusion during feeding time or the increase in outdoor air temperature during the spring.

“My feeling about the whole thing is that we did not have good evidence of animals behaving differently,” Lindberg said. “Just prior to the March 15 swarm of quakes we did have one or two of the pocket mice run above ground at a time when they are usually in their underground cages. But that could have been attributed to the warmer weather we were experiencing. Unfortunately we couldn’t prove if the animals could or could not predict earthquakes.”

The second Geological Survey study was conducted by the Menlo Park-based SRI International (formerly the Stanford Research Institute). Some 1,700 volunteers in Los Angeles County, the San Francisco Bay Area and Humboldt County were recruited to report anomalous animal behavior on a toll-free hot line. The messages were recorded and logged by researchers.

“We threw out the calls that came in just after an earthquake was reported,” said Leon Otis, the principal investigator in the project. “We felt reports of unusual animal behavior would be more objective if they were reported prior to the earthquake.”

Over a four-year period, 13 earthquakes of a Richter scale magnitude of greater than 4.0 were recorded in the study area. In seven of the 13 earthquakes, researchers fielded a “significant” increase in the number of calls. (Some of the reports included fish jumping, dogs howling, ants swarming, general restlessness among livestock, and domestic birds biting and killing their young.)

However, in only one case were the results dramatic. That was the 1981 Fremont quake (magnitude 4.3) when 21 hot line calls were received within a 30-day period--the average number during that period was one call.

“We felt there was something there, and I still feel that way. But without confirmation, we can only treat the Fremont episode as suggestive,” Otis said.

The third Geological Survey study was conducted at University of California, Davis. Residents within fewer than 3 miles of the epicenters of five earthquakes were asked--within days afterward--if they observed unusual animal behavior before the quakes.

Four out of the five cases resulted in findings similar to those at Ovando, Mont., where only one of 35 persons noted any unusual behavior before an earthquake that measured 4.9. The one exception was at Willits in Northern California where 17 of 50 people interviewed observed weird animal behavior within a week of a 4.7 quake.

“Although they were inconclusive, there were some interesting findings in the three studies,” Otis said. “Unfortunately, federal funding for animal/earthquake experiments has dried up under the current Administration. As far as I know, there are no further studies planned.”

In the laboratory setting, researchers have been able to determine that many animals possess seismic sensory powers far greater than those of humans.

For example, fish have highly sensitive otolithic organs of the ear that can detect extremely weak seismic activity--up to three magnitude levels smaller than those detected by humans. This includes the tiny foreshocks too small to register on the Richter scale that frequently precede sizable earthquakes.

Many fish and amphibians also possess the ability to sense exceptionally small gradients of water displacement that can be caused by the foreshocks of submarine earthquakes. Catfish, sharks and rays may also be able to sense light electromagnetic variations created by early seismic activity. Ground-dwelling animals may also feel foreshocks long before humans.

Moreover, some scientists believe that many animals may have tiny crystals of magnetite in their brains that are susceptible to seismic changes in the Earth’s magnetic fields--precursors to most earthquakes.

By and large, however, scientists have virtually abandoned animals in their search for earthquakes predictors. “There are no more ongoing experiments with animals and earthquakes. Even China has just about given up on the animal idea,” said Dupras of the California Department of Mines and Geology.

An unscientific survey of the Los Angeles Zoo following the Oct. 1 shaker also turned up inconclusive results.

“None of our keepers reported any unusual behavior the day prior to the quake,” said Laura LaMarca, curator of education at the zoo. “After the big shock--and prior to the aftershocks--our gorillas wouldn’t go back into their holding cage. But I don’t think that was predictive behavior; it was just good sense.”