SAN FRANCISCO, April 10 (UPI) -- The next big earthquake, scientists say, could be predicted by data from smartphones and fitness trackers, and they've got the data to prove it.
Last year, in the wake of the 6.0-magnitude earthquake that rocked Napa Valley, sensors in the fitness trackers made by Jawbone were able to document seismic disturbances in Berkeley, Oakland and San Jose. Bloggers (doubling as data scientists) at the company shared the results, showing the sudden shift in sleep patterns in the wee hours of the morning.
But the question remained: could such data ever be more than fodder for viral infographics? Could data collected by sensors in smartphones and life trackers have predictive value?
Researchers at the U.S. Geological Survey suggest that it can. The GPS sensor in a smartphone, they say, could pick up a sudden jerk in one direction. If hundred or thousands of smartphone users were to register the same movement at once, it could be an early sign of a seismic shift.
"Imagine all of Portland was out at a cafe on a sunny day, and everyone's smartphones were sitting on the table when one of these great earthquakes happened," USGS scientists Benjamin Brooks told NewScientist. "The whole city would appear to move."
Scientists at the agency tested the idea using a computer model designed to predict what smartphone data might look like in the moments before, during and after a major earthquake. The model simulated a hypothetical 7.0-magnitude earthquake along San Francisco's Hayward Fault Zone, and also re-enacted Japan's devastating 2011 earthquake using real seismic data.
Their analysis suggested readings from as few as 5,000 smartphone users would be enough to identify the beginnings of a major earthquake, allowing officials to issue a five-second warning before the disaster struck other densely populated locations. While five seconds may not sound like much, researchers say it could be enough to sound alarms, shut off gas lines and even trigger the garage doors of fire stations.
Researchers have increasingly looked at crowdsourcing as a tool for disaster prediction. Previous studies suggested "reporting, collection, and analysis of individual experiences" during and after an earthquake could help scientists build more accurate predictive models.
But the latest study is one of the first to gauge the potential for smartphone users to identify the very beginnings of an earthquake. The study was published this week in the journal Science.