Detection of small earthquakes in Southern California: NPR

The Northridge earthquake in southern California in 1994, which killed dozens of people and caused the collapse of the crossing, was a magnitude of 6.7. Scientists say improvements in the detection of earthquakes reveal that hundreds of small earthquakes occur every day in the region.

Douglas C. Pizac / AP

hide legend

activate the legend

Douglas C. Pizac / AP

The Northridge earthquake in southern California in 1994, which killed dozens of people and caused the collapse of the crossing, was a magnitude of 6.7. Scientists say improvements in the detection of earthquakes reveal that hundreds of small earthquakes occur every day in the region.

Douglas C. Pizac / AP

Detecting very small earthquakes is notoriously difficult. The roar of the ocean, the passage of a car or even the wind can give the impression of a minor earthquake to the sensors that cover seismically active parts of the United States.

This is a problem for scientists who rely on data relating to all earthquakes in a region to study what triggers the most important and destructive ones.

A team of scientists said they have found a way to accurately detect tiny earthquakes and have released a new, more complete list of earthquakes in the last 10 years in southern California. The work was published today in the journal Science.

The team relied on data from a network of approximately 400 seismic sensors in California, distributed between the US-Mexico border and the southern part of the state. These sensors continuously measure movements in the earth's crust, looking for traces of earthquakes. During the 2008-2017 decade, scientists have already identified 180,000 earthquakes in the region.

"They have a robust seismic network in Southern California," says Daniel Trugman, seismologist at Los Alamos National Laboratory and author of the study. But while 180,000 could appear to be a large number of earthquakes, there were many more who were hiding without being detected in the data.

When Trugman and his colleagues reanalyzed the data with the help of a powerful set of computer processors, they found evidence of a number of earthquakes ten times larger: 1.81 million earthquakes in a decade, about a small earthquake every 3 minutes or so.

"You do not feel it happening all the time," says Trugman. But "they happen all the time."

Most of the earthquakes detected in the study are so small that their magnitude falls below zero. It is not impossible for humans to feel such a subtle tremor in the rock beneath our feet, but it is unlikely. Trugman likens this to a table that's overturned in your kitchen. If you find yourself in the kitchen when this happens, you will notice that the table has hit the floor. But if you are on the street when this happens, or even just outside, you will probably miss it.

To detect tiny tremors without confusing them with vibrations without trembling (like a passing truck), Trugman and scientists from the California Institute of Technology and the University of California at San Diego used computers to do some research over a decade looked like known earthquakes.

The analysis has been made possible by advances in computer processors over the past decade or so. Still, it took tens of thousands of hours for a group of 200 graphics processors hosted at Caltech – essentially enlarged versions of the laptops' graphics cards – to scan all the data and spot potential earthquakes, as well as hundreds of thousands of overtime for other computers to complete the analysis.

The ability to measure more small earthquakes will hopefully help scientists answer some of the most intriguing questions about how, where and why earthquakes occur.

In California, many communities rely on outage maps indicating the areas most likely to be earthquake-prone to assist in decision-making regarding infrastructure, building codes, and building codes. emergency plans. Having more comprehensive information on earthquakes in the region could make these maps more complete and help identify so-called blind faults that are not visible on the surface, but are likely to go underground.

The Northridge earthquake in southern California in 1994, which killed more than 50 people, injured thousands more and caused billions of dollars in damage, is responsible for blind fault. In fact, the seismic sensor network that made the new study possible was put in place following the disaster.

It would also be possible to use data from similar sensor networks located in other parts of the United States – for example, in the Pacific Northwest – to create more comprehensive catalogs of earthquakes in these areas.

The authors of the study hope to use the data to determine the degree of onset of large earthquakes and the role small earthquakes play in this process. Understanding this complex relationship could eventually help seismologists predict earthquakes.

"We are going to look at a lot of these issues in a much more detailed way," said lead author Zachary Ross, a geophysicist at Caltech.

"The holy grail of seismic seismology has always been a prediction," says Trugman. In recent years, for example, the US government has deployed a seismic warning system along the west coast, which uses the same sensor networks as those studied by scientists. Having a deeper understanding of the seismic information provided to this early warning system could help make them more accurate.

"I am cautiously optimistic that we will advance the prediction of earthquakes," Trugman said.