With the help of AI, 2 million previously “hidden” earthquakes in Southern California were identified, the Earthquake catalog expanded by a factor of 10, it was discovered there are about 495 earthquakes daily roughly three minutes apart (previous studies suggested a 30 min time lapse in between).

We are using machine learning to more reliably monitor earthquake activity in southern California. It is helping us to more precisely detect and locate earthquakes as part of aftershock sequences and swarms. These more detailed catalogs are used to better image fault zones beneath Earth’s surface and study how earthquakes interact and trigger other events. We also are using machine learning to accelerate simulations of seismic wave propagation.