Check your email, play a song, or make a cup of coffee—in that amount of time, Southern California will have probably experienced an earthquake.
Every three minutes, a tremor ripples through the western state, according to nearly a decade’s worth of seismic data crunched by scientists at the California Institute of Technology (Caltech).
More than 1.81 million small earthquakes hit the region between 2008 and 2017, ten times more than previous estimates suggested, the researchers concluded in a study published in Science on Thursday.
Scientists knew these quakes were happening, but until now they’ve been virtually invisible, Zachary Ross, lead author of the study and soon-to-be assistant professor of geophysics at Caltech, told Motherboard.
“These tiny earthquakes are happening all the time,” Ross explained, “we simply lacked the capabilities to observe them over extended periods of time.”
Ross and three colleagues looked at historical earthquake data from the Southern California Seismic Network, a cooperative project between Caltech and the US Geological Survey that relies on hundreds of monitoring stations.
Instead of hunting for the “big one,” the team pored over catalogues for tiny temblors, or quakes between negative magnitude 2.0 and 1.7. These events are so small, they can go undetected amid traffic, construction, and other background noise.
A not-so-new technique called “template matching” was used to pull these signals from all the static. Template matching uses larger quakes, which are more distinct, as models for what a smaller quake might look like in a given area. That’s because the “wiggles” or waveforms of earthquakes appear similar when they happen in the same spot, National Geographic explained. Scientists have tapped this technique before, but the Caltech team was the first to apply it on a large scale.
From the data, they unearthed 495 quakes per day in Southern California, roughly three minutes apart.
Template matching can be useful in other seismic hotspots, too, though it relies on a network of seismometers that are closely spaced together—able to match quakes within a one to two mile radius, the study notes.
The technique could “be applied to the datasets collected by any of the regional seismic networks in the United States, including Northern California, Alaska, and Nevada,” Ross said, adding that Japan—where up to 2,000 earthquakes can be felt each year—is another viable candidate.
“The new catalog of 1.8 million earthquakes provides new information on the locations and geometry of faults deep below the surface,” Ross told Motherboard in an email. Filling these gaps could “help us to understand better how sequences of earthquakes evolve in space and time.”
In the future, Ross hopes their technique can shed new light on how earthquake sequences begin, and how they might interact with or trigger others.