Here's something you might not know about the 6.4 earthquake epicentered near the Pacific Coast of Mexico on May 8: By the time it hit Mexico City, 170 miles away, people there already knew it was coming. Even before the shaking started, they had time to move to safety. They were ready—thanks to their advanced warning system.
After a devastating 8.1 earthquake in 1985 which may have killed upwards of 10,000 people, Mexico's government was determined to find a way to prevent this kind of loss of life again. Mexico City residents knew that this 6.4 earthquake was rippling towards their homes because, in 1992, they launched an effective system that's able to tell millions of people that an earthquake is on the way. It's simple, it's relatively affordable to implement, and it works.
In fact, Mexico is one a handful of seismically active countries that have a early warning systems. Taiwan, Turkey, and Romania each have one, too. After the 1995 Kobe quake killed 6,500 people, Japan employed its own early warning system, eventually becoming the first to take advantage of nascent cellphone technology. Following the Tohoku earthquake in 2011, over one million people downloaded a new app which helped prepare them—both physically and mentally—for the dozens of aftershocks that rocked the country.
A similar system, properly implemented, could give cities the time to stop trains and freeze elevators. It would give doctors time to halt surgical procedures in hospitals. It would allow police and firefighters to strategize in the case of a power or communications failure. Just a few seconds of warning could prevent millions in financial losses or even an environmental disaster like Fukushima—and it could potentially save thousands of lives.
So why doesn't the U.S., which bears a large amount of the global earthquake risk, have any kind of system in place?
Unbelievably, a prototype has already been built in California. The latest version is currently being tested by the United States Geological Survey (USGS) and several other partners. In fact, it's almost finished. The science and technology is sound. We know it can save money and lives. Perplexingly, there's just no plan in place to roll it out yet.
We can't—and shouldn't—afford to wait. "Most of the systems you find today are the result of some big killer earthquake," says Douglas Given, the Earthquake Early Warning Coordinator for the USGS. "So one of our goals is to build our system before the killer earthquake—rather than after."
Like Twitter, But Faster
If you've been in an earthquake, you can understand how an early warning system might work by watching a rudimentary version of it that gets used every single time: Twitter. When the shaking starts, our first instinct is to reach for our phones and post our take on the situation. At least this is what happens in my house.
But if you're looking at Twitter at just the right moment, you might even see those tweets from across town—"Earthquake?" "EARTHQUAKE" "Shaking right now in Hollywood!!!!"—moments before the shaking starts at your house.
@awalkerinla just ended here
— Dylan C. Lathrop (@DylanLathrop) March 29, 2014
That's a perfect illustration of the way information travels faster than seismic waves, Given tells me at the Caltech Seismographical Lab in Pasadena. Earthquakes move through the ground at roughly the speed of sound; data can travel at the speed of light. So an earthquake can be detected and reported long before the shaking starts at your house, he says. However, for an early warning system to work, the whole reporting process must be automated—no humans. "We'd hope to detect the earthquake and tell you about it faster than somebody who types, 'Hey, I just felt an earthquake.'"
Part of an earthquake sensor system on display in Caltech's lab
Across the country there's a network of sensors, each managed and operated by different groups which are part of the Advanced National Seismic System (ANSS). In California, for example, the USGS operates the seismic network in cooperation with Caltech and UC Berkeley. The ANSS consists of monitoring systems that are installed in the ground like vertical pipes—a large concrete-anchored cylinder that holds two different ways to measure movement: a broadband seismometer and an accelerometer. Above ground is a solar panel to power the unit, an antenna to transmit the data, and a high-precision GPS which can measure ground motion in real time to a preciseness of two centimeters. This is how seismologists can "see" that the Earth is shaking.
In a hallway at Caltech's Seismographical Lab, dozens of these stations, each with a lettered code like an airport, report their seismic data in real time. None of them are completely quiet. We see a small earthquake blip onto the screen in Mammoth Lakes, in the Northern Sierras. There's also another rhythmic pulse that shows up—it's the surf pounding the shoreline, and it can be seen on stations hundreds of miles inland. That's how sensitive the system is.
Modeling Every Quake, Before it Happens
At the Southern California Earthquake Center, a lab on the Caltech campus, seismologists study earthquake behavior—looking at the overall effect on Southern California according to an earthquake's size, depth, and location. And they're able to create models which can predict how quickly and how powerfully those variables will reach specific areas during a major earthquake.
Given shows me one scenario they call the ShakeOut scenario, a hypothetical 7.8 earthquake that would strike near the Salton Sea, about 150 miles away from Los Angeles (about the same distance that last month's quake was from Mexico City).
The time it would take for shaking to reach L.A. after a 7.8 quake 150 miles away
The first model he shows me illustrates the shaking intensity for Downtown L.A. for that 7.8. quake. Although we use the magnitude of an earthquake to communicate its size, this is a better way to measure the local impact because it shows how strong the shaking is in one particular area—it's the best way to show what it will feel like on the ground. As he starts the simulation, I watch as the L.A. basin gets dyed like a watercolor, turning blue to yellow to red—signifying "very strong shaking."
Ground velocity shows how the quake ruptures along the San Andreas Fault
But earthquakes don't simply ripple out from a central point like a stone dropped in a pond. The dashed white line in the animation above is the San Andreas Fault. You can watch how the earthquake ruptures along the fault, which helps the waves to travel even further in a certain direction. This is why seismologists also have to look at ground velocity, or ground motion, which takes those ruptures into consideration.
The early warning system system can not only detect an earthquake, it can look at both of those factors and can adjust the area and scale of impact—estimating heavier or lighter shaking in certain places based on where it happens and the way it travels along a fault. So the system can also update in real time due to ruptures, or a seismic re-evaluation of the original event.
Right now, the USGS has an early warning prototype that many of the scientists have on their computers and phones. He turns on the demo, and I watch—and hear—the earthquake start to ripple towards my house.
The yellow circle is the P-wave, a fast wave that doesn't shake very hard. What you want to prepare for is the red circle, the S-wave, which causes the most destructive shaking
These graphics or sounds are not final, of course—this is just the data they're passing along to developers, who will morph it into a more palatable message. In this form, with the honking Area 51-esque alarm, the whole thing seems a bit terrifying. The countdown is especially unnerving, like some kind of twisted New Year's Eve Ball Drop of Doom. But having the information, clearly plotted on a map updating in real time, is also completely fascinating—and oddly reassuring.
Of course, this is just one scenario on one fault that's far away from Los Angeles—in actuality there are hundreds of faults underneath L.A. and we may have little to no warning at all. And if you lived near the Salton Sea during this particular quake, you'd know about the shaking the old fashioned way—long before any alert arrived.
A beta version of the warning prototype installed on Given's phone
But what I was more surprised about is how long it took for the shaking from such a large earthquake to travel to my home. Using the ShakeOut scenario, I would have 81 seconds to prepare for the earthquake. That seemed like a lot to me.
What would I do with those 81 seconds?
"Alert! Alert! Alert!"
The challenge to create Mexico's early warning system was a relatively easy proposition, geologically. The major faults in Mexico are located off the Pacific coast, so a large earthquake can really only be generated offshore. The most devastation from earthquakes, historically, has occurred in Mexico City, which is not only home to the largest population in the country, but also carries the most risk as a large portion of the historic city center is built on a filled-in lake. So sensors installed along the coastline can offer up to a minute warning time to people living in Mexico City, saving the most lives.
In 1992, when the system was first put into place, the warning consisted of sirens, radio and television messages, and some proprietary alerts that went to schools and government buildings. Residents can also buy FM radios, a little desktop device which can receive a wireless signal, or a USB-like stick that's triggered by satellite. A few years ago, the government released an iPhone app as well, called Alerta Sísmica.
Even if you don't live in Mexico, you can download the app (it costs $1.99), which I did on the day after the 6.4 quake. I was able to personalize all the settings, requesting that my phone vibrate and ring with the same alarm I use to wake up.
I was able to experience the warning system much sooner than I expected. At 2:36 a.m. Mexico time on that very night—12:36 my time—I was awakened by the alarm and a message on my phone: ALERTA SÍSMICA EN PROGRESO.
Screenshots from my iPhone of the Alerta Sísmica app after the May 10 aftershock
In the dark, I quickly flipped through the app to the map locating the epicenter. The 6.0 was relatively close to the other quake—likely an aftershock. If I was in Mexico City, I'd have about 60 seconds to prepare. I thought about where I was at the moment—in a bed on the second floor of my brother's home—and what I would do if I had 60 seconds. I wondered what the hundreds of Mexico City residents who had gotten the same alert might be doing.
The next day, I talked to Itzel Eguiluz, a Mexico City resident who's currently getting a Ph.D in sociology. It was clear that the societal effects of the deadly earthquake that was almost 30 years ago were still fresh in her mind. "In the 1985 earthquake I was only three years old, but I had a lot of memories of the stories and also I remember my parents getting scared," she told me.
The app was only for iPhone until recently, so when it came out for Android she installed it on her Nexus tablet. But before these latest earthquakes, it didn't work. That made her a little upset, she says, because especially during the one in the middle of the night, she would have much preferred being awakened by the shrill alert from her phone over being jolted awake. "I really hope they manage to arrange it, because I don't like to wake up at night with the movement. I think it is great to have the option."
I asked Eguiluz what she would have done if the warning had worked. What she said surprised me: "I would always go downstairs, alarm or not, because I think it is safer."
That's the exact opposite of what we're told to do in the U.S.
The Problem With Amber Alerts
For Americans, the closest thing to an earthquake early warning system that we've experienced is a relatively new feature courtesy of our smartphones: the Amber Alerts or warnings from the National Weather Service, which essentially blanket a geographic area with pertinent information.
I remember when I received my first Emergency Alert. I had just arrived in Vegas when my phone made a noise I had never heard before. I shrieked and ran over to it, where I saw a message that flash floods were coming and to avoid the area. I looked out the window, where torrential rains were just starting to whip up on the Strip, then back to my phone. Wait—how did it know I was here?
Two examples of emergency alerts you might have received recently on your phone
This is the Wireless Emergency Alert system, which takes the warning system you know best from those screeching tones on the radio and scrolling messages running across the bottom of your local television coverage, and shifts it to the cellular phone network. Cellphone carriers are now required to relay those messages to their customers, and all new smartphones are required to have this capability. Which is why, even though my phone is registered in L.A., I got the flash flooding alerts on my phone in Vegas—AT&T simply blanketed the geographic area with the warning.
There are three levels of messages for the Integrated Public Alert and Warning System (IPAWS): Amber Alerts, for a child abduction, hazard alerts for weather, and the presidential alert, which would be something like, there's a meteorite heading towards Earth and we're all going to die (this has obviously never been used). You can opt out of Amber and hazard; you can't opt out of the presidential alerts.
But the issue with IPAWS is that while it's fast for, say, Amber Alerts or inclement weather, it's not fast enough for earthquakes. Imagine that an earthquake starts now: The data is processed by the sensors, confirmed at the seismic center, sent to the carriers, and then broadcast out. For an early warning to be effective, all of that has to happen in about six seconds.
Plus, right now, those geotargeted messages can only be issued on a county level. In actuality, the seismologists would like to have a little more granular focus than that. They should be able to tell you, based on the GPS in your phone, exactly how many seconds you have to get ready. The warning would be completely personalized to your location.
But say you're at work, in a meeting, when your phone starts alerting you about an impending quake. Anyone who has ever been in a building when a fire alarm goes off knows how that works—people look around to see if there's a fire. Even in a tornado, sirens are commonplace in the Midwest, but they don't always motivate people to head to the basement; they look outside instead. "We know that when people hear an alert tone or see a bell or something like that, they don't act," says Given. "They look for confirmation."
Depending on how the message comes into your phone, people just might pay attention and do the right thing. In fact, come communications scientists are betting on it.
The Science of Why We Freak Out
The USGS's most important partners are located halfway across the country, at the University of Kentucky's College of Communication and Information. Here, specialists in risk and crisis communication are determining how exactly to share the warning with the masses. That research is headed up by Dr. Deanna Sellnow, the director of the school's Division of Instructional Communication. With her husband Dr. Timothy Sellnow, she is leading the investigation on how to translate that (slightly shocking!) information from the USGS into user-friendly app messaging.
After studying how people react in a crisis, the Sellnows have developed a theory they call the IDEA Model, which drives their strategy for what to include in this kind of warning communication. "We know that certain elements need to be in the message," says Sellnow. "Our model suggests and has been proven that in order to have internalization, you must have explanation, have action—what am I supposed to do?—and deliver it over the right channel." The key is to provide enough information and evidence for people to internalize the data but not overwhelm them with a message that's too long or complicated.
The IDEA Model for emergency messages: Internalization, Distribution, Explanation, Action
The "explanation" part of the IDEA Model has proved to be especially important. For a brief message like this, the seismologists actually have quite a bit of information to pass along to users: The epicenter of the earthquake that can be plotted on a map, the rupture in the fault, the magnitude, the expected intensity, and also the number of seconds before the strongest shaking arrives. Testing showed that people actually appreciated having more information, namely the map and the countdown, both of which provide that real-time, "this is really happening" feeling that would hopefully spur them to take action.
Working with students at Art Center College of Design and designers at Foxio, the USGS information has been sculpted into several different versions of the early warning messaging, each using different design hierarchy and language. Currently the Sellnows are testing four different versions of the "shake alert" messages to see how people respond to them. You can even take a survey to help them craft their messaging.
Getting the early warning out to a large population is one thing. What happens when those people see the early warning is another issue entirely. There's one big challenge that they're facing with the messaging in certain markets, says Sellnow. "When we are testing in Southern California there's this apathy: Earthquake? Whatever."
The response to such a warning varies widely by culture. Angelenos might shrug, but during the May 8 Mexico earthquake, which happened during the work day, people streamed into the streets, just like Eguiluz told me she would do. That's not surprising, seeing as more than 400 poorly constructed buildings collapsed in the city in 1985, and that's the way many people died. But that's not what the U.S. wants you to do in an earthquake.
In the U.S., contrary to many people's first reactions—stand in a doorway, run outside, do nothing—the safest thing to do is stay where you are, drop, cover, and hold on. The early warning messaging will, in fact, probably have to say this very specifically. "What we're trying to make sure that is they don't do the wrong thing: If you're inside, stay inside, drop, cover, and hold on," says Sellnow.
Testing two different ways to deliver the information, showing a map vs. shaking intensity, plus graphic instructions of what to do. Images by Art Center College of Design and Foxio
Currently the messaging is going through its final testing phase and will soon be ready for full-on app development. "The trick right now is the short time," says Sellnow, of the less-than-10-second turn around. "But I think were getting there, we're really close."
The Sellnows' vision is that the app would be free to download, and immediately available across all smartphone platforms. (In Japan, the early warning option is integrated into all iPhones.) Then, of course, they'd have to look at accessibility issues: reaching people who don't have smartphones, deciding on features like sounds and vibrations and whether or not they'd be universal. They're also looking at how to serve blind and deaf audiences.
The rollout would also have to include a hefty educational campaign teaching people how to use the app and respond appropriately. And all of this requires funding: Lots of funding. That piece of the puzzle is political, says Sellnow. "I hope that our evidence-based testing where we can show data on how this messaging works will convince political people to put their money behind it."
Our One Early Warning System That Works
It's difficult to study the impact of an early warning system because every earthquake is different—scientists can't really measure what the outcome would have been without it. Which makes it especially hard to convince government leaders that the investment in a warning system is indeed worth it.
Most stories are anecdotal, like a semiconductor plant in Sendai which not only uses early warning data from the Japanese Meteorological Agency, it also placed its own sensors in the factory which could detect movement and shut down the plant. Simply comparing the damage the factory experienced in an earthquake before the safety measures and after the safety measures, the company estimated it saved almost $15 million in damages.
There is one U.S. organization using its own earthquake early warning today: San Francisco's BART system. You may have been riding on a BART train, in fact, when it stopped for a moment in anticipation of a small quake rumbling through the Bay Area. Although you wouldn't have known it—the system does not announce when they are slowing and stopping trains in response to USGS ground shaking information.
When alerted to seismic activity, San Francisco's BART system will stop or slow trains or hold them at the station. Photo by ShakataGaNai
BART is a perfect candidate for helping to illustrate to city leaders the importance of an early warning: It's in a city prone to earthquakes, it's a critical part of the infrastructure that needs to keep running, and it has proven that a train that's not moving is less likely to be derailed. Plus the action it takes based on the warning—slowing or stopping a train—is low-cost and effective. It seems like a no-brainer that San Francisco would want to invest in something that would ensure that its transit system would escape major damage and might continue to run after an earthquake.
Or what about this reason: Perhaps the most positive impact that an early warning system could have citywide is to protect first responders during the aftershocks of a relatively large quake. Not only are structures weakened from the original quake at this point, but firefighters and police are in those structures looking for people. If an aftershock is coming, they could evacuate the building and resume efforts when it's safe.
You would think that something so simple, which could protect transit infrastructure and emergency crews, would be enough to motivate the country to implement the warning system. The idea may be gaining traction. Last week, the House Committee on Natural Resources held hearings on the importance of establishing the warning system. "We're hopeful that there will be some movement in Washington D.C. to fund it through federal dollars," says Given. "There's a little bit of movement at the state level." Last year, California passed a law requiring that the state create an early warning system. The bad news is that it had no dollars attached to execute it.
On the federal level, there's also the issue that not all parts of the U.S. are geologically equal: Although there are pockets of seismic activity all over the country, this is absolutely more of a priority for the West Coast, which carries 77 percent of the country's earthquake risk. So what will likely happen is that an app might roll out for California first, then be available for the rest of the country.
Luckily, the ground has been nudging earthquake safety back into the national consciousness, which is a very good thing, according to Given. "When earthquakes are in the news, people start to think about them again and it seems like, oh, maybe this is a good investment."
It's true: This has been an especially shaky year—April, for example, was a record-setting month for large earthquakes globally. Los Angeles has had five 4.0 and higher quakes this year, the most since 1994.
That includes a 4.4 rattler that struck just after I had gotten up one morning and sent me yelping curses and leaping back into my bed. Even though I knew exactly what to do—heck, I was researching a story on earthquakes!—I panicked when the ground started to move. After the shaking stopped, I wondered: Would a warning help make me calmer or more freaked out?
What Would You Do?
I posed the question on my Facebook page: "What would you do differently if you had 60 seconds to prepare for an earthquake?" Popular answers included putting on shoes, pulling out the earthquake kit, gathering pets, and filling up water bottles. But just as many people said that if they knew they had the time, their first priority would be to run out of their house. And down the street. Screaming.
Every earthquake is different. Some buck so roughly that the house groans in agony and you're convinced the world is ending. Others almost purr in a way that feels like the Earth is giving you a gentle foot massage. When they start, it's very difficult to get a sense of what's about to happen, so I always mentally prepare for the worst, which often means making bad decisions. If I had more time and information, I would probably think more rationally about my specific situation. Right?
Official Drop, Cover and Hold On messaging from the Southern California Earthquake Center
Even the "drop, cover, and hold on" rule is a reactionary response—it's what we're taught to do once we start to feel the ground shake. The next phase of a warning system, and how it's implemented, could change that, too. Using a sophisticated early warning system, the advice might even be situational. If your phone knew you were driving, for example, it could instruct you not to drive under the overpass you were about to drive beneath. "We actually think that there will be a very important place for, say, a cell phone app that knows where you are and tells you what to do," says Given. But that's not necessarily the role of the USGS, that's the role of whoever ends up developing the app. "We intend to produce the stream of information and let the market do what the market does and make those things that people think are valuable."
With the technology in place as a kind of Shake Alert API, one can only imagine the next generation of devices that will help keep us safe in a quake. Maybe it's a tiny Nest-like appliance in your house that gets pinged at a certain magnitude to turn off your gas. Or the navigation device in your autonomous vehicle that pulls to the side of the road and shuts off the ignition.
Japan's early warning system broadcasts a message on TV before a strong quake
But before all that happens, I do think there's something to be said about the awareness and education that an early warning app could provide—before it even notifies you about a quake.
Just having the Alerta Sísmica app on my phone—which has made me jump a few times since, warning me of earthquakes happening hundreds of miles away—helps me to be more conscious of the frequency and size of earthquakes throughout the country. To try to monitor what was happening here, I followed as many quake-tracking Twitter accounts as I could find for L.A. and California. While not technically warning me, they did make me feel more connected to local seismic patterns.
I've now become a bit obsessive about earthquake preparedness in general. I find myself pausing during different parts of the day, rehearsing what I would do in each room of my house, on the street, in a bus, giving myself a silent countdown until the strongest shaking begins. Dashing up to the bedroom and putting on my shoes takes longer than you'd think. So does finding my phone, in most cases. It's always tempting to think I could bolt out the door, outrunning the earthquake in my freshly laced Nikes, but I know I'm safer under the kitchen table.
Just thinking about the potential of early warnings has caused me to rehearse my behavior, which I think is one of the key benefits of such a system. It forces us all to run scenarios in our own heads, comparing what it will take to get to a safe and secure place for that first jolt instead of the stomach-churning alternative.
Give me those 80 seconds or 40 seconds or even just 10 seconds: I can't wait for the early warning system to become a reality. But for now, it helps to know that because I've been practicing, I'll be far more calm next time an earthquake takes me by surprise. I'll be ready.