The Oklahoma Geological Survey (OGS) at OU developed a software using artificial intelligence to detect previously undetectable earthquakes.
According to an OU News press release, Oklahoma state seismologist Jake Walter and his team of researchers utilized the March lockdown caused by COVID-19 to develop “cutting-edge earthquake detection technology.”
Utilizing this software since early May, the OGS has been able to identify twice as many earthquakes “too small to be felt by people,” according to the release.
According to the release, Walter faced a problem in his early research when running his computer codes “each with unique approaches for detecting and locating earthquakes” in tandem. The data was “too much” for geoscientists to sift through.
“That’s when I realized we could leverage a machine-learning picker developed at the California Institute of Technology,” Walter said in the release.
The CalTech machine learning picker “learned” what an earthquake looks like using millions of seismogram datasets enabling the picker to identify earthquakes in a waveform, according to the release. With 90 OGS stations throughout Oklahoma, Walter and his team can feed their data into the picker to get picks from individual stations.
Adding various tools including event association onto the picker that can “output a full earthquake catalog for a provided dataset,” the result is a “fully operational software package that analyzes data and detects earthquakes” more sensitively than its predecessors, according to the release.
Although Walter’s new system has not replaced the current OGS “real-time earthquake detection system,” the old and the new systems are now running in tandem identifying smaller events due to the new software, according to the release.
According to the release, the open-source software package is available to other geoscientists and is believed to be the first “implementation of machine-learning technology by a regional seismic network for routine earthquake identification and alerting.”
“We’re already hearing from fellow researchers from around the world who are using our software on their projects,” Walter said in the release.
With a publication in geoscience journal Seismological Research Letters, Walter hopes his software will “continue the advancement” of understanding seismicity around the world, according to the release.