Sunday morning's M6 South Napa Earthquake jolted thousands of Bay Area residents out of their sleep.
It also proved an ongoing experiment correct: earthquake warning systems work.
Just What the Heck Is An Earthquake Early Warning System? I thought They Couldn't Be Predicted!
Well, it's still true. Earthquakes cannot be predicted with any accuracy and only vaguely forecast at best. I seriously doubt this will change, ever. I'm aware there are a great many people who think otherwise and have formulated formulae and techniques of varying quality. If you want to read about some of them, I suggest Susan Hough's "Predicting the Unpredictable" (this article was expanded into a very good book too, go find it and read it.)
Twitter is already a warning system of sorts, as illustrated here in an xkcd cartoon, and then seen live on August 23, 2011 on the US East Coast where tweets (which once transmitted move at roughly the speed of light depending on network lag) outran the seismic waves of the quake. A great many people at distance on the US East Coast knew about the quake before they felt it. The USGS was so intrigued by the twitter response to the Central Virginia earthquake that they now use twitter reports to enhance their Did You Feel It data collection.
Earthquakes have properties that can warn people at a distance.
In short, an earthquake warning system is a system of accelerometers, computers and alarms that are devised for regional warnings. They work because earthquakes, when they happen, emit several types of seismic waves that move through the earth at different rates. Strong seismic waves, the ones that cause destructive damage (these are your S-waves), move through the earth at about two miles per second. P-waves, which rarely cause damage move much faster. Detection of the P-wave can then enable powerplant operators to go into standby, trains to stop (so they don't derail) and other automated systems to shut down.
Where Else In The World Are These Systems Used?
Not widely. Japan is the only nation with a national and comprehensive early warning system. When an earthquake occurs in Japan an alarm will sound on most televisions, with a map that displays the effected region of the country. Bullet trains will stop, automated systems will go into protected mode, and SMS texts will go out to cellphone users. Japan's early warning system is the most sophisticated in the world.
Tokyo received up to 90 seconds to two minutes of warning after the M9 quake in March of 2011 because of its distance from the point the earthquake nucleated. The first warnings went out across Japan 8 seconds after the first P-waves reached seismometers in Japan, or 31 seconds after the rupture began. It is very likely the early warnings that went out across Japan saved many lives despite the catastrophic death toll caused almost entirely by the tsunami.
Mexico City implemented a warning system after the catastropic September 1985 quake. Mexico City is far away from the subduction zone that periodically has M8+ earthquakes but is close enough to be effected by their shaking. It also sits on the bed of a lake basin, the lake being drained after the Nahautl capital was captured, leveled, and rebuilt by the Spanish. The sediments of the lakebed enhance the shaking of large, distant earthquakes. This led to the deaths of many thousands (a true count was never taken) in Mexico City.
The early warning system in Mexico City sounds an alarm across the city in most buildings, and in most media, and residents of the capital have up to a minute of warning. The system has been in operation since 1991.
Earthquakes in Romania largely are limited to the Vrancea region. However, there are periodic deep and large earthquakes of magnitude 7+ that have caused catastropic damage at a distance in the capital of Bucherest, most recently in 1977. A warning system is now in place. It provides about 25 to 30 seconds of warning to the capital city. There are plans to create more systems across Western Europe as the region is perhaps a bit more quake prone than one might think. They could be of great use in the Mediterranian region, for example.
But in the US? Outside of experimental systems, there are none currently in operation. California's experimental system did work though. In theory, were it a public system, residents of San Francisco and Oakland could have received ten seconds of warning.
Why Isn't This System Being Used If It Works?
Well, it sort of is being used. It already provides data to transit systems and a few test users in California. It isn't ready for widespread use because of that long-time issue: money.
The system needs money. It needs money to expand California's already dense seismic detection network. It needs money for outreach that will help users such as industry, business, school districts, universities, media, and the public know what to do with a warning. Most currently available messaging systems for SMS in the US are also actually much too slow for earthquake early warnings, unlike Japan. While California's legislature passed legislation for a statewide early warning system, it has not been funded. The system needs about $130 million over five years to get started. This is dirt cheap, in my book.
More info here.
Remember that this system works best with distance as well. For most large California earthquakes, the maximum time a location may receive from a warning before damaging shaking occurs is perhaps 50 seconds (the southern San Andreas and the city of Los Angeles is usually the example used.) The system wouldn't have done the residents of Napa any good, but 10 seconds is more than enough time to get under something sturdy or away from something that will fall on you (assuming it's daytime).
The Pacific Northwest is an area of the country for which this system is perfectly suited. A theoretical system could give, in the event of an M9 Cascadia event, the big cities of Portland, Seattle, and Vancouver (and San Francisco too---the southern end of Cascadia is at Cape Mendocino and it's theorized Cascadia M9s nucleate there and propagate northward) up to five minutes of warning depending on the rupture characteristics. It is likely that such an event could also cause effects far from the coast, including the failure of the electrical grid across multiple states and Canadian provinces. Warning systems would protect these crucial systems and get them up and running after a large earthquake. While there likely is some time here (the next Cascadia M9 quake could happen tomorrow, or several centuries from now), a system that's in place is still a very good idea.
There are a number of research projects underway in the Pacific Northwest. But again, they need money, and from where I sit, these systems are really inexpensive. We aren't that broke.