20 Seconds Into An Earthquake We Can Tell How Big It Will Become


Stephen Luntz

Stephen has a science degree with a major in physics, an arts degree with majors in English Literature and History and Philosophy of Science and a Graduate Diploma in Science Communication.

Freelance Writer

It takes about 15 seconds into an earthquake to reveal whether it will be a modest tremor, or the sort of major quake that leaves a legacy like this. Prometheus72/Shutterstock

Seismologists are claiming to have found one of the holy grails of disaster prevention, a signal that can be used as an early warning system for the most destructive earthquakes. Unfortunately, the warning time is less than a minute before peak impact, but the team who found it still think this may be enough to save some lives.

Professor Diego Melgar studied US databases of earthquake events. By looking at the rate of acceleration, rather than of movement, over the first 20 seconds, he was able to distinguish events that would become large enough to do serious damage from those that never subsequently grew so large. “On average, earthquakes with a larger final magnitude grow faster early on,” Melgar writes in Science Advances


Melgar then sought to confirm his conclusions using European and Chinese records. "To me, the surprise was that the pattern was so consistent," he said in a statement. "These databases are made [in] different ways, so it was really nice to see similar patterns across them."

Melgar analyzed more than 3,000 earthquakes of magnitude 6 or larger. Among these were 12 major events between 2003-2016, all of which could be identified as on the way to something big by around the 15-second mark. The time from there to the peak varied between 10 seconds and a minute. 

Until around the 10-second mark, earthquakes look alike, but they diverge between 10 and 20 seconds after beginning, revealing their ultimate size. University of Oregon

Melgar has already provided evidence that GPS data from the sea floor could add a life-saving 20 minutes to tsunami warning times. It’s easy to see how something like that could prevent most of the direct loss of life from such events. It will be more challenging to make similar use of his latest discovery, perhaps through an automatic shutdown of fire hazards.

Some seismologists consider the Cascadia subduction zone to pose the greatest earthquake threat to America. Archaeological evidence and indigenous stories indicate the zone has produced ruptures in the past on the scale of those that produced the 2004 Boxing Day tsunami and Japan's 2011 Tohoku earthquake. The fact there has been no large movement for three centuries means such an event is more, rather than less, likely to happen soon.


Based at the University of Oregon, Melgar is naturally particularly concerned about an event that could devastate his home state. He suggests placing GPS sensors on the sea floor along the fault line to provide alerts, as is done for fault lines on land. Relying on coastal stations alone is insufficient, he argues, since by the time the earth movement reaches there it will be too late to provide useful warnings. However, he acknowledges the cost of such a system of sensors would be substantial.