Certainly scientific understanding of tectonic plate movement has evolved since the days when people relied on weather patterns, barking dogs and the laying habits of their prized hens to predict earthquakes.
Scientists, for example, predicted in 2003 that there is a 62 percent chance that a Big One will erupt on the Hayward Fault sometime between then and 2032. But the ability to pinpoint with any degree of certainty and specificity when that Big One will hit continues to elude us.
"Earthquake prediction is obviously the holy grail of seismology," said Richard Allen, associate director of the Berkeley Seismological Laboratory at UC Berkeley. "If we could predict earthquakes that's certainly what a lot of people want. ... It would be a big advantage if we could do it. However, I think the reality is that most seismologists do not think we can have earthquake prediction in the foreseeable future. The reason for that is the complexity of the earthquake process."
Yet Allen is developing a system that could give folks a heads up that a quake is coming — not weeks or months of warning, mind you, but a few seconds to get out of harm's way.
Life-saving seconds
Would a 10-second warning be worth the effort? Allen believes even seconds would be enough time to dive under a sturdy table to avoid flying glass and bookcases, enough time to keep an airplane from trying to land and enough time for a train to stop or a factory to shut down."Earthquake early warning is the next generation of real-time earthquake information," Allen said. "It's very rapidly detecting the beginnings of an earthquake, assessing the hazards associated with that quake and then providing people with a warning just a few seconds before the ground shakes."
Allen has just completed the first three-year phase of the ElarmS system, a research project run through the California Integrated Seismic Network, a consortium of UC Berkeley, the California Institute of Technology, the U.S. Geological Survey, the California Geological Survey and the California Governor's Office of Emergency Services. UC Berkeley runs a network of about 50 monitoring stations. There are hundreds more throughout California that ElarmS uses.
Deep in the vault
In a cement vault built into a hillside above the UC Berkeley campus, the Berkeley Seismological Lab runs one of the sensor hubs in the network. Built in 1965, the vault originally housed a network to detect nuclear testing. Now it is home to a sophisticated sensor system for earthquake monitoring. The sensors pick up many small quakes each day that no one feels. It's the larger, potentially destructive quakes that define the early warning system's purpose.This is how the alarm system works: Sensors detect the first wave of the earthquake, the P-wave which travels fastest but has little energy. The sensors use the P-wave to assess the hazards associated with the quake before the secondary wave is felt. The S-wave travels more slowly but carries more energy and does the most damage. The farther you are from the quake epicenter, the more warning time you will get between the two waves.
The measuring instrument consists of three sensors which measure vertical ground motion, east-west motion and north-south motion. The instrument streams the data through a cable to a data logger in the vault which translates the analog signals into digital ones that could then be sent over the Internet and cell phone networks to deliver an early warning.
It really works
The first half of the system — the detection part — has successfully performed during two 5.4 magnitude quakes: the 2007 Alum Rock earthquake in Northern California and the 2008 Chino Hills temblor in Southern California."So that was our proof of concept," said Allen. "So now we're moving into phase two of the project. Another three-year effort to build a prototype system to provide warning to a small group of end users."
Allen is looking for about 10 users among school districts, chemical or manufacturing plants, transportation systems or airports to test the second half of the system, the delivery of the warning through e-mail alerts or cell phone calls or text messages.
Next steps
Such systems already exist in Japan, Turkey and Mexico, and Allen's next steps are to determine if such systems will work in California. At the same time he cautions that even if scientists could get to a point of giving weeks of warning that a quake is coming, quake prediction is no panacea. Evacuating an entire city before a quake would cause major disruption of people's lives and economic stability. A better approach is the one California employs now, he said: building an infrastructure that is resilient to the effects of earthquakes."We can build buildings that will not fall down and which may be largely operable after quakes," Allen said. "And that's a much better approach than pinning all of our hopes on earthquake prediction."
Donna Hemmila is managing editor at the UC Office of the President's Integrated Communications group.