Wrangling over an early warning system for earthquakes

Margaret Vinci from Caltech briefs participants ahead of a functional exercise for first responders in a simulated 7.8 magnitude earthquake drill at the Office of Emergency Management in Los Angeles. This year's exercise featured the California Integrated Seismic Network's Earthquake Early Warning Demonstration System, as seen on screens pictured here.
(Frederic J. Brown / AFP/Getty Images)

This post has been updated, as explained below.

Two different approaches to alerting the public about earthquakes are vying for the state’s support, one based on a network of government-operated sensors, the other built around a private company’s equipment. In a move that may shed some light for policymakers, the California Emergency Management Agency has pulled together a group of experts to make recommendations on how to implement an earthquake early warning system.

Then again, maybe it won’t. The Cal EMA working group includes representatives of both factions, and their diverging views may prove impossible to reconcile.


On the one hand you’ve got the early warning system being developed by the U.S. Geological Survey, the California Institute of Technology and UC Berkeley, which piggybacks on the sensors used in the earthquake-detecting California Integrated Seismic Network. On the other hand is Seismic Warning Systems, a Scotts Valley company that uses its own sensors to alert customers -- typically large businesses, institutions and first responders -- about quakes.

The idea behind an earthquake early warning system is to take advantage of the time lag between a quake’s initial, relatively mild shockwaves and the later ones that inflict all the damage. Seismic sensors can detect the first waves a few seconds or more before the severe shaking begins. The challenge, though, is to determine from the first waves whether a) it’s really an earthquake and not just a big truck rolling by, and b) how intense the shaking may be.

Patterned after Japan’s early warning system, the CISN approach typically takes readings from several sensors in the state’s network to determine whether the vibrations are coming from a quake large enough to cause damage. (It could use the readings from a single sensor, but some scientists in the program worry that this approach would generate too many false alarms.) The plan is to use that data to trigger alerts that could be publicized by broadcasters and phone companies, while also being transmitted to transit systems, manufacturers and other entities that could program their control systems to respond automatically.

The March 11 temblor in Riverside County gave the CISN team the chance to demonstrate the capabilities of its system, which is still in its formative stages. It alerted seismologists in Pasadena 30 seconds before they were hit by the stronger waves of the quake, which measured a moderate 4.7 magnitude. By the time the warning went out, however, those waves had already been propagating for 12 seconds, extending about 25 miles from the epicenter.

[Updated, Aug. 28, 9:07 a.m.: Douglas Given, earthquake early warning coordinator for the U.S. Geological Survey, said in an email that the actual performance of the system was better than was first reported by the news media. Seismologists in Pasadena received an alert 40 seconds before the stronger waves arrived, and the warning went out 7.5 seconds after those waves began propagating. That reduced the blind zone to an area extending about 14 miles from the epicenter, which means that all of the cities in the Coachella Valley would have gotten at least some advance warning before the stronger jolts hit. “The performance of the system will get better with time as we improve algorithms and add more sensors,” Given wrote in an email.]

That’s not necessarily a problem for earthquakes that start in a remote stretch of the high desert. But a delay that long for a quake in Northridge would leave millions of people and much of Los Angeles in the warning-free “blind zone.”


Seismic Warning Systems has an array of sensors in the Coachella Valley, not far from where the March 11 quake struck. In an interview this week, Michael Price, the company’s chief technologist, said its alerts went out before the more damaging waves had even hit the surface. The first stations it’s working with in the valley received an alert three to five seconds ahead of the shaking, Price said; by contrast, the entire area was in the CISN system’s blind zone.

The company’s speed advantage stems from its ability to use two sensors at a single site to calculate a quake’s intensity and trigger a warning, rather than having to pool data from sensors at multiple sites. It also contends that it could partner with Sacramento on a statewide warning system that would cost far less than what the CISN’s backers have proposed.

Sen. Alex Padilla (D-Pacoima) has introduced a bill that would have the CISN develop a “comprehensive statewide earthquake early warning system.” There’s no cost estimate in the bill, but supporters say it could cost $80 million to finish developing and deploying the system.

Price said Seismic Warning plans to build its own statewide sensor array for the sake of selling its alert services to manufacturers and institutions. It has offered to operate and maintain an alert system at its own expense if the government picks up the tab for the equipment needed to deliver the warnings effectively to schools, fire stations and other public facilities.

Doug Given, national coordinator of the U.S. Geological Survey’s earthquake early warning efforts, said the agency intends to build a warning system for the West Coast, and is looking for more help from the private sector to accomplish that. So is the state of California, which is “under a lot of pressure to not spend any new money,” Given said.

So what’s not to like about Seismic Warning’s proposal? Given said his team is trying to get some independent validation of the company’s claims about its technology. So there’s that. But there’s also some resistance on the CISN team to entrusting the crucial function of triggering a public alert to a private company. You can only imagine the liability issues involved.

Seismic Warning executives also have expressed skepticism about sending alerts to the general public, arguing that it would do more harm than good because many people wouldn’t know the safe way to respond. But they softened their stance on that issue this week. “We have taken a step back a little,” said Scott Nebenzahl, the company’s director of government affairs. When it comes to warning the general public, he said, “we will work with the state to achieve its goals and objectives.”

In addition to the USGS and Seismic Warning Systems, the Cal EMA working group includes representatives from the California Geological Survey, the state Seismic Safety Commission, the Federal Emergency Management Agency, UC Berkeley, Caltech and the Gordon and Betty Moore Foundation, which has helped fund development of the USGS system. Greg Renick, a spokesman for Cal EMA, said the group’s assignment is to develop “recommendations regarding the implementation of an earthquake early warning system” based on the “technological, legislative, organizational and managerial” issues involved.

The group has held one meeting so far, Renick said. Nebenzahl said that one more meeting is expected late next month, then “it will be completing a policy paper and a path for the governor and the Legislature.”

Assuming, of course, it can agree on what that path might be.


Unmaking Europe

Bump in the road for healthcare law

Gay marriage: What’s most offensive about DOMA

Follow Jon Healey on Twitter @jcahealey