For California, an Earthquake Early Warning System Is Up and Running

autospa writes "In California's Coachella Valley around Palm Springs, a state-of-the-art, first-in-the-world earthquake early warning system in now installed and operational. Twelve locations are now in place with 120 sites planned, all meant to detect an earthquake and give people a chance to get under a table, or in the case of a fire station, get the engines outside of the building."

Comments

Interesting idea, horrible article

[jlechem]

Let's see 3 paragraphs with no real info. What seismic level are they talking about? A 2, 3, 4, 5, or what? In Utah we got lots of 2 and 3s all the time. California is even worse. Who decides when it's time to hit the panic button? And if it's a person that means they have to have staff available 24x7. Still it seems pretty cool they're trying to solve this problem.

Re:Interesting idea, horrible article

[drerwk]

Without reading the article, it is a computer which then calls the fire department garage doors and they open. The FD will not collapse on the Engines but the door may jam, or not work for lack of electricity. Also some elevators may stop and open at the nearest floor. hospital generators may start. That sort of thing. I am not expecting a text that say duck. I was in Santa Cruz eating dinner for the '89 and it was terrifying. Even though the fire engines got out, the roads were choked and they could not get anywhere. After about 15 minutes, I could count about 6 fires in the distance. I had even heard it might give warning in surgery to pull out instruments and cover the patient to keep dust out.

Re:Interesting idea, horrible article

[geekoid]

"What seismic level are they talking about? A 2, 3, 4, 5, or what?"

what do you think? do you really think it would go off on a 2,3 or 4?

Re:Interesting idea, horrible article

[ColdWetDog]

I think you're correct (I read TFA, as usual it doesn't help).

The big question will be the false positive rate. If you're randomly opening up doors / turning on large, expensive generators and scrambling OR teams on a regular basis, it will get shut off like all of the OTHER alarm systems that cry wolf repeatedly. Presumably, this bit of wisdom has been considered by the engineering team and it's acceptable (if not dozens of Slashdot posts will helpfully remind them). Be nice to have more details.

Re:Interesting idea, horrible article

[iamhigh]

You should check out Nova Science Now on PBS. http://www.pbs.org/wgbh/nova/sciencenow/ They had a segment on this last night. I am sure they have an article about it, but what would /. be without indirect sources?

Yep... here it is.

http://www.pbs.org/wgbh/nova/earth/earthquake-detection.html

You would have to really trust it

[pudding7]

Meaning, no false alarms. Set the thresholds such that it doesn't go off for undetectable or even very minor quakes. With only seconds to act, people need to feel confident in reacting without a second thought. If the alarm sounds and the teacher says "Hmmm, let's see if it's really a big one before taking cover." then it's lost some of it's usefulness. Same with the automated stuff. It would be unfortunate to get to a point where the alarm goes off and doors roll up, gas is cut off, etc and people immediately think "Crap, not again. Now I have turn the gas back on and close the damn doors."

Re:You would have to really trust it

[natehoy]

As far as I can see, there's not much chance of false alarms unless someone drops something heavy right next to one of the seismometers or something. This is detecting actual earthquakes. The chances of actual false alarms are pretty low. The earth is shaking somewhere. In fact, this is data seismologists already gather as a routine.

The difference here is that it's propagating out an automated warning that can be responded to automatically to nearby locations. The key is "automatically". As in, people don't need to react. You'll only get a minute or two of warning at best - you want this to be automated.

Signal hits the fire station, and the fire station opens the doors immediately (so the quake can't jam them shut if power is lost or the doors get shaken out of track, for example). Alarm tells the firefighters to go get in the truck and pull it into the parking lot in case the building collapses. That's a bunch of fire engines and ambulances you've kept in service when they're likely to be needed very, very soon.

Signal hits a hospital, and they spin up their generator (so it's already running if the Big One hits and they lose power) and sound a tone in operating rooms telling doctors the floors might shake so starting a delicate cut around the brainstem is a bad idea for a few minutes.

Signal hits a large commercial building, and the elevators all go to the nearest floor, open their doors, engage all friction locking mechanisms, and tell everyone to get out of the elevator right now.

Bridges might drop gates to keep people who are not on them yet off them. Water and gas mains might close some containment valves. Traffic lights might all turn red so cars stop. Bell goes off at the school telling the kids to get near a reinforced wall.

Nothing that people need to take conscious effort to react to, just automated stuff that makes the incoming quake a little easier to deal with. Also nothing that would cause all life to come to a complete stop. There'll still be enough gas and water pressure in the systems that most people wouldn't even notice the outage. Traffic would be stopped for a few minutes. The elevator alarm will shut off and people will get back in. And so on...

This is pretty useless if you're at the epicenter, but gives you increasing amounts of warning as you get further away. It also lets emergency personnel outside the quake zone know that they'd better start getting ready to head toward the epicenter, because they'll be needed very soon.

If The Big One ever hits, this might save a lot of lives and damage to a lot of useful rescue equipment miles from the epicenter.

Re:How do they know it works?

[Anonymous Coward]

So, barring an actual earthquake, how do they know this thing works?

They just pick up the sensor, shake it really hard, and listen for alarms.

Re:How do they know it works?

[intellitech]

It's little more than a fancy network of seismometers. Why wouldn't it?

Re:How do they know it works?

[geekoid]

It detects an earthquake is on it's why. Remember earthquakes travel down the fault. So when i happens at point A, Point Be might be far enough to get 30 seconds of warning.

Re:And

[olsmeister]

I'm sure they understand that Windows are one of the first things to go during an earthquake, and you don't want to be anywhere around them.

Re:How do they know it works?

[toastar]

Who would of thought the speed of light through a wire is faster then then the speed of sound through rock.

Re:How do they know it works?

[pyroclast]

Assuming this is the same QuakeGuard as mentioned in the article, here is their technical explanation http://www.seismicwarning.com/technology/waveseparation.php

Re:Not impressed.

[blair1q]

Was it even earthquake season?

It uses video cameras and cats

[billstewart]

If the video cameras detect the cats acting weird, then that means there's going to be an earthquake soon. It was easy to verify its accuracy - small earthquakes happen all the time in various parts of California, and they checked the video recordings and the cats had been acting weird just before the quakes.

Re:It uses video cameras and cats

[Dunbal]

That's what Schroedinger's box is for. The really weird ones don't survive.

Re:It uses video cameras and cats

[Transkaren]

Nothing is earthquake-proof. You can build an *incredibly* rigid structure. But when you do, the quake will be just that much stronger, and knock it down anyway. Earthquake engineering isn't designing structures to not take damage. It's designing structures to take the minimum amount of damage. In some cases, entire sections of floor might be considered sacrificial - beams are designed to bend side to side (by cutting the top and bottom off the I) instead of passing the force to the column (causing the column, and everything above it, to fail). The reason we have so few fatalities here isn't because we build our buildings strong. It's because we build our buildings flexible. Most areas of the world used brick, mortar and other rigid stone-like materials for hundreds of years. California is just plain newer, so while we have failures they tend to be less catastrophic because of the amount of steel (a ductile metal) and wood used in construction.

Re:It uses video cameras and cats

[geekoid]

The really weird both survive AND don't survive.

Re:It uses video cameras and cats

[pixelpusher220]

Your argument is flawed, but I understand your confusion :)

Events that are unrelated to one another, such as games of Blackjack from independent decks of cards, correctly do not influence the other events.

However, the CA quakes are not unrelated events. They occur because one tectonic plate is slipping past another. The longer that slip does not happen, the greater likelihood that it will happen in the future. The slip *will* happen. When it does, it will depend on how much force is built up.

Considering the force behind a moving tectonic plate is massive, the longer it is pent up without slippage means that energy is being stored up until a failure at some point along the fault and it breaks free. It is possible that we will see a series of smaller quakes rather than a big one, but simple physics dictate that once the object (plate) starts moving it's going to keep on moving unless equivalent force is applied to stop it.


(This new formatting system is sucky sucky. I'm using HTML formatting but it's still making the lines massively separated :( )

Re:It uses video cameras and cats

[VanGarrett]

The regions of California which are prone to earthquakes are all along the coast. The weather in these regions tends to be quite wet and humid (the fog in San Francisco borders on legendary), and as such, rust is a substantial problem for vehicles. Keeping the trucks harbored inside makes them much easier to maintain. Granted, Southern California earthquake regions tend to be drier, but again, it's easier for fire services to maintain their vehicles when they're kept inside-- not to mention that it's easier to get to them in an emergency, when they're kept in a location right next to the equipment.

A practical system with a good track record

[mjpvirtual]

Portions of the system have been in operation since 2001. There have been several moderate and many small events. The system has produced no false positives or negatives, so far. It works by detecting the P-wave (6.2km/s), analyzes it to estimate the intensity of the coming S-waves (3.6km/s), and automatically triggers protective measures if the intensity is expected to exceed MMI V. It does not estimate earthquake magnitude, since that tells you nothing about the intensity at your location. The P-waves convey about 6% of the earthquake's energy; the rest is conveyed by the S-waves. The P-waves provide a natural warning that you're about to experience strong shaking.

The warning time varies from 0 (at the epicenter) to many seconds farther away. A networked system provides up to 1 second of warning for every 3.6 km from the epicenter. This is enough time to protect equipment and give people a chance to prepare themselves.

Re:A practical system with a good track record

[hedwards]

The problem though is that there's a significant drop off in intensity related to the distance. Which means that by the time you're far enough away from the epicenter to be able to actually use the information, chances are that the intensity is low enough that it's more of a minor nuisance.

Re:A practical system with a good track record

[mjpvirtual]

You're assuming that nothing useful can be accomplished in a few seconds. The applications in Coachella all complete in less than 10 seconds. A few seconds is plenty of time to duck under a table.

As for epicentral distance, the intensity drops 90%/100km after the first 100km. If you're 200 km from a magnitude 7 event, the intensity may be low enough to not matter, unless you're up-rupture, or on alluvial soil, or at convergence zone for shockwave reflections, etc. If you're in the SF Bay Area you're less than 100 km from likely epicenters, so it's a few seconds of warning or nothing.

Large events mean large rupture. If you're up-rupture you'll see more intense shaking than at the epicenter, not less. Simple metrics like distance from the epicenter provide no guidance about intensity.

Re:Early warning? I'll give you one.

[Dunbal]

Everyone always underestimates the volcano...

Re:How do they know it works?

[pyroclast]

From QuakeGuard technical page: "The QuakeGuard technology detects the non-destructive P-waves while filtering other sources of vibrations that can lead to false alarms. The elimination of false warnings is a result of QuakeGuard's patented DSP algorithms that filter detected vibrations to isolate the signature waveforms of a seismic event that has just occurred. Depending on the geological composition of the terrain and the distance from the epicenter of the seismic event, a warning of 10 to 60 seconds is possible."

Re:How do they know it works?

[MobileTatsu-NJG]

So, barring an actual earthquake, how do they know this thing works?

I assume this isn't predicting, but setting off alarm bells as soon as possible?

Quick to ask but slow to RTFA? Heh.

An earthquake creates two waves. The first one triggers the alarm before the second one reaches you. And yes, it'll be clear that it works after the first quake hits. However, they're already tracking the seismic data reliably anyway.

Re:Early warning? I'll give you one.

[hedwards]

If you look at the US, it's hard to find anywhere that doesn't have earthquakes or some other natural disaster. Personally, I think it's more reasonable to build in earthquake country than in areas where it floods or gets hurricanes. At some point you hit the point where it's just not realistic to avoid the danger and have to focus on mitigating it. There have been some pretty substantial earthquakes back east, it's just that people forget about the 1812 New Madrid earthquake

Re:Installed and Operational

[geekoid]

The death star doesn't need to be completely built in order to be fully operational.

Have you learned nothing from our lord and master, the Emperor Palpatine?