Maps reveal earthquake swarm near Mammoth in California
Machine learning helps to pinpoint quakes that can harm critical infrastructure
Seismicity in California is common — every 174 seconds, the ground trembles as Earth's tectonic plates shudder past one another. Now researchers at California Institute of Technology have applied machine learning to years of seismic data from a region known as the Coso Volcanic Field, just south of the San Andreas fault. The goal was to identify those quakes that can cause the most damage - the ones that shake the ground for a long time - so that critical infrastructure, such as bridges and pipelines, can be strengthened to withstand them.
The team's results, published in the journal Science Advances, show that machine learning can accurately identify the strongest earthquakes as well as humans can. But the technique has one major advantage: it can do so in real time.
That means that in the future, as seismic networks become denser and more data becomes available, machine learning could be used to issue early warnings for damaging earthquakes.
Comments