Stanford seismologists are now using technology inspired by popular music apps to find earthquakes never detected before.
“Feel Good Inc.” by Gorillaz is the song that rocked geophysicist Greg Beroza’s world. Several years ago, the Stanford Professor was perusing an electronics store when the tune came on.
“I used Shazam, and it ‘listened,’ connected to the cloud, made the match and tried to sell it to me – all in about 10 seconds,” Beroza said.
The professor and group of graduate students used the song-matching concept to develop similar software called “Fingerprint and Similarity Thresholding” or FAST. The technology mines earthquake databases and matches waveforms, allowing scientists to detect “microquakes.”
“These earthquakes are so small you can’t feel them at all. You can’t hear them,” graduate student Clara Yoon said.
Stanford researchers examined one week’s worth of 2011 data from the Calaveras fault, which runs through Santa Clara, Alameda and Contra Costa counties. They’ve pinpointed hundreds of small earthquakes previously undetected.
“They in themselves are not a threat, but they tell us about the same faults that could host much larger earthquakes,” Beroza said.
Beroza and his team now have their eyes on earthquakes in Oklahoma, Kansas and Arkansas because they believe they are man-made. It’s common practice for oil and gas operators to inject wastewater back into the ground, but if they inject it into an area where there may be a fault it could cause earthquakes.
“Most of the wells aren’t a problem, but some of them are. So if you see a bunch of little earthquakes sort of lining up and illuminating a larger fault near an injection well, you might say, ‘hey, let’s not inject into this well anymore. Let’s do it somewhere else,’” Beroza said.
For now, FAST is simply a detection tool. And as for predicting the next “big one,” Beroza says the technology isn’t there yet:
“That’s sort of the Holy Grail, but it’s not yet been shown possible.”