Hell to Pay: Why Aren’t We Fully Funding A Phone App to Warn Us of Earthquakes?

By Glen Martin

California wants to lay out some major cash for hyper-ambitious public works projects. For example, the Twin Tunnels, Jerry Brown’s retread of the peripheral canal that was defeated by voters in 1982 during his first go-round as governor. Depending on whom you talk to, this massive water conveyance scheme will cost between $25 and $67 billion.

Or how about that high-speed rail? When completed, California’s bullet train will shuttle citizens between L.A. and the Bay Area in less than three hours. Such impressive speed, of course, comes at a price; the latest guesstimate is that it’ll cost about $100 billion to build out the system.

So you’d think there might be a few million dollars in California’s petty cash drawer that could be put to a more immediate need: A seismic detection system that would greatly enhance the data set on California’s faults and quakes and could provide advance temblor warning to citizens. Enough warning time, in fact, to allow people to exit dicey structures before the first huge shockwaves hit.

That’s the dream, anyway. And UC Berkeley’s Seismology Laboratory and the U.S. Geological Survey are making dramatic progress in developing the systems that would vastly expand our knowledge of California seismicity. But like Richard III’s putative loss of the Battle of Bosworth Field due to the lack of a horse, the Lab stands to lose a whole lot of research momentum for the same reason—if instead of “horse” you understand “money,” that is.

Certainly, recent progress in seismological technology provides some reason for optimism. The digital seismographs that now represent the state of the art are ever so much more sensitive than the old analog devices of the 1970s and 1980s. Faster computers and sophisticated algorithms allow quick crunching of incoming numbers, providing highly accurate and rapid information on quake locations, scale, and likely ground shaking. New phone software—such as MyShake, an app developed by Berkeley graduate student Qingkai Kong—could potentially turn everyone with a cell phone into a walking seismograph, creating a powerful network that could provide ongoing data streams to first responders and government agencies. (Kong successfully recorded signals from last year’s Napa Valley quake in Berkeley using his app.)

Via the emerging Internet of things, you could link quake alerts to public infrastructure and health care systems: Firehouse doors could be programmed to open automatically, emergency care physicians could receive phone alerts to report to hospitals, trains could be stopped before they approach subpar tunnels and shaky bridges.

Jennifer Strauss, the external relations officer for Berkeley’s Seismological Lab, says the development and dissemination of apps such as MyShake could augment and enhance existing seismometers located throughout the state. Because when it comes to seismic detection, more is more: The denser the network, the better the incoming data.

“The established (government and University) seismometers throughout the state are supersensitive, recording not only large quakes but little tiny rumblings, all of which are necessary to do good seismic hazard assessment,” Strauss says. “But things get even more interesting once you start considering things like a MyShake network. Cellphone accelerometers aren’t as good as those on seismometers; they work a lot better for larger than smaller or even moderate quakes. But if you have a lot of (quake app–equipped) phones dispersed throughout an area, you should be able to get more detailed information than we’re getting now. The earth is not homogenous, and seismic waves propagate at different velocities depending on where you are. You should get information at a finer scale, a finer mesh, the more sensors you have out there.”

This isn’t to say phone quake apps are irrefutably ready for prime time. Significant refinement is desirable, allows Strauss. For one thing, the acceleration signals you get from an earthquake—even a big quake—are minor when compared to the kind of jostling a cell phone goes through during regular use: Think of all the rapid accelerations and stops, the abrupt shifts in pitch, yaw, and roll, that your phone endures when you jog, jump on a trampoline, gyrate at a rave, or swerve at 65 mph on a freeway.

“We’re working on (helping phones) untangle those kinds of everyday usage signals from seismic signals,” says Strauss. “You can’t just turn on the phone and assume the app will work. We’re working to ‘train’ the algorithms to understand the differences in signals—through shake table tests, human tests, and actual earthquakes. The data that we received from the Napa quake, for example, was extremely valuable.”

Another thing about quake apps, says Strauss: There’s more to them than simply getting them to do what you want. To be at all useful, they have to run 24/7, since earthquakes don’t make appointments. So you have to figure out ways to do that without interfering with other phone functions or constantly running down batteries.

“We also have to test for other issues that have nothing to do with scientific algorithms,” Strauss continues. “For example, is the app ‘sexy’? People won’t use it if it isn’t interesting. Also, is it secure? That’s a critical issue, obviously.”

And that brings us back to money, or rather, the lack of it. App development and testing, expansion of seismometer networks, new computers, and supporting the scientists, graduate students and technicians needed to do the work: It isn’t free. Meanwhile, the need for such research, it can be argued, has never been greater. The Napa quake was an unpleasant wake-up call about the parlous nature of our local and regional faults. And as California’s urban population continues to swell, concerns over what was long considered the granddaddy of all faults, the San Andreas, remain evergreen.

Nor is the San Andreas our biggest worry these days. The Cascadia subduction zone, a massive rift that extends from the tip of Vancouver Island to Northern California, makes the San Andreas Fault look like a pencil-necked pipsqueak. When—not if, but when—it rips loose, Seattle could be reduced to a smoldering rubble heap and tsunamis could inundate the North American littoral from Victoria to Eureka (or maybe San Francisco).

So what resources are we committing to counter this existential threat? Congress approved $5 million for earthquake early warning systems this year, which will be shared by Cal, Oregon State University, and the University of Washington. Of that, $4 million will be used for continued work on the “ShakeAlert” early warning system, which has been ongoing for the three continental Western coastal states since 2006. The remaining $1 million is earmarked for new seismometers.

That’s good as far as it goes, but it doesn’t go far enough. Earlier funding for the system from the Moore Foundation is winding down; no new funding source is apparent. But it wouldn’t take a massive infusion of taxpayer bucks to really bring things up to snuff—at least, not in comparison to trans-Delta water conveyance systems, bullet trains, or even a new F-22 jet fighter (current price tag: $412 million per plane).

“The total capital investment requirement for a West Coast early warning system is $38 million,” says Lind Gee, a supervisory geophysicist with the U.S. Geological Survey in Menlo Park. “That would cover new (sensor) instrumentation, processing computers—everything. Anticipated annual operational costs and maintenance is $16 million a year. That doesn’t replace ongoing investment in building retrofits. That’s equally important, and can’t be overlooked. But when compared to the costs incurred by recent earthquakes, these are small investments.”

Just from the existing prototype system, says Gee, San Francisco had a 9-second warning from the Napa quake.

“That’s significant, and we should be able to do better with more, and more modern, instrumentation. Berkeley and the USGS work as a partnership on this. Berkeley went totally digital on its instrumentation in the 1990s, but the USGS is still working with a lot of analog equipment from the 1970s and 1980s, particularly in Northern and Central California. Analog instrumentation is just a lot more limited in the detail of information it can provide. So we have a long way to go for those regions.”

Share this article:
Google+ Reddit

Add new comment