Fitness Trackers Quantify Sunday’s Northern California Earthquake
Many of us wear fitness bands during the day (to track activity) and at night (to track sleep). But it turns out that the data they gather can do a lot more—like track an earthquake.
At 3:20 a.m. Sunday morning 24 August, an
6.0 on the Richter scale struck northern California near the city of Napa. Down in Palo Alto where I live, the shaking was minor: not enough to set off car alarms, but enough for my husband to wake me up to tell me that we might be having an earthquake and to cause something in another room to crash to the floor. (It turned out to be my son’s old Mr. Potato Head toy.) After that commotion, it took me a little while to fall back to sleep.
I often wear a fitness tracker; on Saturday night, it was in its charger, so it missed my earthquake-driven awakening. But enough people in Northern California were wearing their fitness trackers that night to enable tracker-maker Jawbone to create a snapshot of the earthquake’s intensity by analyzing sleep disruptions of users in the region.
Jawbone’s results weren’t surprising—people living closest to the epicenter were more likely to have been immediately awakened by the earthquake (93 percent, compared with 55 percent a little farther away) and to stay up longer—45 percent of Jawbone users living 24 kilometers or fewer from the epicenter stayed up all night. But it was a fast and powerful demonstration of a new way to use anonymized fitness tracker data.
Data about the amount of ground shaking produced by an earthquake–which differs from the earthquake’s magnitude or the distance from the epicenter because it is affected by the type of soil and other factors—is currently provided by the U.S. Geological Survey (USGS) in the form of ShakeMaps. The maps guide earthquake response and disaster planning. The USGS produces its ShakeMaps by combining measured ground motion from earthquake monitors with predicted motion based on geological features that fills in the gaps between the monitors. There are plenty of gaps, which is why data like that gathered by Jawbone could help improve accuracy.
Jawbone isn’t the only tech company testing its data as a supplement or alternative to ShakeMaps. Twitter has been working with researchers at Stanford University to determine if Tweets can be used to create accurate ShakeMaps. The company took a look at geo-tagged tweets sent in the first 10 minutes following a number of Japanese earthquakes in 2011 and 2012 and found that the ground-shaking estimates generated from those tweets were comparable to the official ShakeMaps.