As the 50th anniversary of the 1964 Great Alaska Earthquake was acknowledged this spring, a smaller earthquake in California served as a real-time test for an effective early-warning system. The next step is a larger-scale prototype. U.S. Geological Survey
Researchers are ready to create a production prototype of an earthquake early-warning system that has proved itself on the West Coast.
April 8, 2014—When a 5.1 magnitude earthquake shook the Los Angeles area on March 28, it not only rattled homes and broke dishes, it also served as a robust test of a demonstration system that can provide advanced warning of an impending earthquakes. The system is being developed by the U.S. Geological Survey (USGS), Caltech Seismological Lab, the University of California at Berkeley Seismological Lab, and the University of Washington.
“The system is behaving and performing as designed,” says Douglas Given, the coordinator of the earthquake early warning system, dubbed ShakeAlert. The earthquake and an earlier one of 4.4 magnitude have provided valuable information to researchers, who are about to move from the demonstration phase to a production prototype.
“They are confirmation that the system is working as we expect it to. It’s a little difficult to test a system like this. You can run test data through it and simulate a real-time environment. But of course, you can’t simulate every potential earthquake in every potential location,” Given says.
The system currently monitors an array of 400 sensor sites located throughout California. Each site contains two sensors—a highly sensitive one that is especially effective in measuring weak ground movements and an accelerometer that is more attuned to strong motions. Global Positioning System devices are also employed at the sites. Together the sensors provide a three-dimensional picture of ground movement at the site. Networked together, they can follow an earthquake’s movements along a fault.
Data from the sensors is sent to three central processing locations, one in Southern California, one in northern California, and one in the Pacific Northwest. Computers at these processing sites analyze the data to determine if an earthquake is occurring.
“A lot of the effort and the science went into developing the algorithms,” Given notes. “How do you scan the ground motion data and recognize that there is a large earthquake under way? How do you tell the difference between a small earthquake and a large earthquake, very, very quickly? How do you tell the difference between an earthquake and a sonic boom or thunderclap? That’s where a significant amount of the development effort has been focused.”
The system provides a variable warning time depending on a location’s distance from the earthquake. Residents could have as much as a one minute warning of an earthquake about 140 km away. The warning would be just 20 seconds for an earthquake 60 km away. But that is enough time for high speed rail systems to slow trains, manufacturing facilities to stop production lines, pipeline operators to close sensitive valves, emergency responders to open garage bays, buildings to stop elevators at the closest floor, and physicians performing surgery to remove scalpels from patients, for example.
To provide this level of alert, the final system will need to provide data via direct lines. Given says that emergency responders and sensitive infrastructure will eventually need to receive streaming data during an earthquake that can be acted upon based on predetermined risk-management calculations.
The project team is examining all avenues for sending the warnings to the general public, including the Federal Emergency Management Agency’s Integrated Public Alert & Warning System, which would broadcast alerts to cellphones. Public communication efforts will require a significant education and training component, and the team has engaged social science researchers to aid this process.
“It doesn’t do you any good to send somebody a message if they have no idea what it is, where it came from, what it means, or what to do about it. Obviously, seconds count here. In the ideal, we would like people to not have to pull out their cell phone, open it up, find the message, and read it,” Given says.
“Keep in mind this is a very infrequent event. [People] are not always thinking about earthquakes. To [send] an alert about an earthquake and expect people to respond in a few seconds is a challenge,” he adds. “What is most effective is not to tell them what the problem is but to tell them what the solution is. What action they should take.”
Another challenge is funding. The system ideally would have twice as many sensors as are currently deployed. California recently enacted legislation that prohibits spending its state’s general fund resources on such a warning system, so other sources are being sought.
With research suggesting that a large earthquake on the San Andreas Fault and the Cascadia subduction zone are likely in the coming decades, Given says the team feels compelled to push the project forward. “There is a sense of urgency and, in fact, a bit of frustration that we have demonstrated the capabilities of the system and the only thing really preventing us from rolling it out to the public is the funding to make it reliable and robust and operable into the future,” he says. Mexico and Japan already have early-warning systems for earthquakes, he adds.
“We are just now commemorating the 50th anniversary of the Alaskan earthquake,” Given adds. “We’d like to get the system to Alaska before they have another one of those.”