Although it’s been a while since California experienced a major earthquake, the team behind QuakeFinder hopes they’ll see the next big one before it comes.

The QuakeFinder project is currently in its fifteenth year working on mid-range (two week prior) earthquake predictions.

As project Director Dan Coughlin explained in an interview with The Sheet, QuakeFinder was developed by its Chief Technology Officer Tom Bleier in 2000 in the hopes of using the strange phenomena that occur before earthquakes to learn how to see an earthquake coming before it arrives.

“In 1989, Tom was aware of research being done by Stanford professor Dr. Tony Fraser Smith, who detected these large magnetic signals about two weeks and again three hours before the 1989 Loma Prieta earthquake (Magnitude 7.0),” said Coughlin.

As Bleier explained it, electromagnetic energy relates to energy that is non-mechanical but has both electrical elements (like lightning) and magnetism elements (like a refrigerator magnet), and moves from one place to another, like a light beam.

In the case of earthquakes, Bleier said QuakeFinder is trying to detect “underground lightning,” or large currents that move quickly from the future rupture zone to the surface of the ground.

These large currents generate magnetic pulses that can be detected on the ground level with sensitive magnetometers.

Electromagnetic energy two weeks prior to the Loma Prieta earthquake was about 20 times the regular levels; that energy “shot up to about 60 times normal a few hours before the earthquake,” Coughlin said.

At the time, scientists and researchers hadn’t yet come together to study these signs and what they might mean for earthquake detection. Bleier convinced Celeste Ford, CEO of Palo Alto-based company Stellar Solutions, that this and other pre-earthquake phenomena were worth studying.

“I thought gee, you know, there’s this fractionated report about earthquake lights and animals behaving funny, and water wells dropping dramatically before earthquakes, and I was wondering if there was some way of actually creating an instrument that could pick up these things,” said Bleier.

Stellar Solutions, which specializes in providing engineering services to government customers like the Department of Defense and NASA, has been the primary source of QuakeFinder’s $10 million in funding.

Bleier said he decided to use a magnetometer, an instrument that can detect very small magnetic field changes. He initially enlisted the help of students, creating an outreach project that gave 10 high schools kits with printed circuit boards, computers, and magnetometers, which students set up in their back yards. In 2002, the State of California offered the funding to create 20 more kits.

“We built those and started collecting data, but there really weren’t any earthquakes going on,” Bleier said. “You have a party and nobody comes,” he laughed.

QuakeFinder decided to up its game by building a satellite, which it launched out of Russia. “That gave us about two gigabytes of data, and we began seeing these funny looking signals as the satellite flew over California and other areas where large earthquakes were happening,” Bleier said.

The team went back to ground stations and began building sensor units professionally, placing 40 at sites along the San Andreas, Calaveras, San Jacinto, and Hayward faults. In 2014, units came to the Eastern Sierra, and are now dotted from Independence to Bridgeport. QuakeFinder now has 125 sites in the United States, as well as 15 sites in Taiwan, 10 in Peru, four in Chile, seven in Greece, and four in Sumatra.

“If you were to drive by, what you would see is a couple white stakes in the ground with two boxes, with a solar panel on it,” Coughlin said. The units “use a cellphone to get 60 megabytes of data per day per site to Palo Alto, where we process it.”

QuakeFinder uses special algorithms that look for strange-shaped pulses above a threshold limit, and then extract local noise generated by cloud lightning, cars, trucks, solar storms, and man-made magnetic noise.

“We’re trying to put these [sensor units] every 20 miles in California, so we’ll never be more than 10 miles from an earthquake epicenter,” Bleier said.

QuakeFinder had its first success in 2007, when it went back to its data after the magnitude 6.5 Alum Rock quake in San Jose and detected changes in electromagnetic activity two weeks prior.

“We also saw an air conductivity change,” Bleier said.  “Think of it like a thunderstorm going over; there’s lots of lightning crackling, and sometimes you can actually feel it in your hair, only in this case there aren’t any clouds overhead, instead it’s coming up from the ground.”

QuakeFinder discovered another sign of imminent earthquakes through a Jet Propulsion Laboratory weather satellite. The satellite was equipped with an infrared camera looking at the temperature of the earth, and found that the temperature of the ground rose during the nighttime, rather than going down, about a day before several earthquakes.

If electromagnetic energy, air conductivity, and ground temperature all change at the same time, “We’re pretty sure there’s going to be an earthquake within 24-48 hours,” Bleier said.

QuakeFinder has now detected similar signals before about six to eight other earthquakes in California and some other countries, all starting about two weeks prior to the quakes.

So far, Coughlin added, “We know that it will be a magnitude 5 or greater.”

The QuakeFinder project is working on peeling away extraneous data—the electromagnetic energy of a car driving by a sensor unit; nearby lightning storms; clouds—to begin detecting evidence of quakes below magnitude 5.

“We don’t see [signs] ahead of time, yet,” Bleier said. “When an earthquake happens, we go back [through the data] … Ultimately our goal is to be able to pick out these signals, these precursors, and create an early warning system.”

Although the U.S. Geological Survey (USGS) recently released a new earthquake forecast model for California earthquakes, projecting likely earthquakes over the next 30 years to inform building codes and insurance rates, QuakeFinder could be the key to a more accurate early warning system.

“In the future, you might see something on the evening news, or your phone, that says hey, there’s a 50 percent chance of rain tomorrow, and we’re starting to get some indications that an earthquake may be forming; stay tuned,” Coughlin said.

Bleier and Coughlin said they hope to transition the QuakeFinder project from research and development into an operational state in the next five years.

But in order to learn how to predict earthquakes, QuakeFinder needs more of them to occur.

“We’re the only company that wants more earthquakes in California,” Bleier concluded with a laugh.

For more information on the QuakeFinder project, or to look at data collected at sensor sites, visit https://www.quakefinder.com/.