Attuned to temblors: How well can scientists forecast massive earthquakes?

Headlines portending a coming spike in devastating earthquakes ignited concerns this week. But seismologists say such forecasts work best when they spark action, not fear.

It usually begins with a normal day. People go about their business: running errands, attending class, or going to work. Birds sing, dogs bark, and trees blow in the wind. Then suddenly the ground lurches, buildings quiver on their foundations, unstable structures crack and crumble. Within seconds, the humdrum of daily life has been shaken and irrevocable damage has been done. Lives can be lost, sometimes in the thousands.

Earthquakes surprise their victims. And any attempts to predict them remain on shaky ground. Pinpointing precisely when, where, and with how much magnitude a specific earthquake will occur is impossible, most seismologists agree.

But that doesn’t mean that there’s nothing scientists can do to help communities anticipate earthquakes. Forecasting efforts that calculate the probability that an earthquake will occur in the near future can prompt municipalities to revisit emergency and preparedness plans. And more immediately, early warning systems can buy communities valuable seconds to brace for incoming seismic ripples.

One attempt to provide a long-term forecast came last month when a pair of seismologists suggested a possible earthquake forecast at the Geological Society of America annual meeting. The team spotted a pattern in historical data that suggests that 2018 and the following four years could bring an increase of earthquakes over magnitude 7 globally.

This forecast isn’t meant to scare anyone, says Rebecca Bendick, a geoscientist at the University of Montana in Missoula and one of the researchers behind the latest forecast. “We actually would like to empower communities to be safer and more responsive to their context, not to terrify people for these years.” 

Getting a handle on the big picture

The tectonic plates that make up the Earth’s crust are constantly shifting atop the planet’s liquid mantle. But the plates don’t always slide past each other with ease. Instead, their edges stick together, much like zippers or seams. The movement of the plates puts pressure on the crust around these seams, forcing it to bend elastically. But that stress builds up over time, and the seams can only take so much pressure before they give in, letting loose that stress in the form of an earthquake.

Unlike forecasting a hurricane, for example, by directly observing atmospheric pressures in a region, scientists cannot directly measure how much stress the crust is under. Faults are generally too deep beneath the Earth’s surface. But researchers can use an indirect approach. Satellite data, for example, can show how tectonic plates are moving. Scientists can use that data to calculate how much pressure has built up at a fault since its last temblor.

“We know something’s going to happen because we can see this stress building up on the fault plane” in the model, explains Richard Allen, director of the Berkeley Seismology Lab at the University of California, Berkeley. But this doesn’t give scientists a specific day, week, or even month when that pressure might erupt. Instead, it simply hints that an earthquake might be impending over the next handful of decades.

The new hypothesis put forth by Professor Bendick and her colleague, Roger Bilham, at the University of Colorado at Boulder, looks at forecasting from a global perspective rather than forecasting for individual faults and considers whether there might be a calculable cycle of global upswings in large quakes.

Bendick and Professor Bilham looked at the history of earthquakes over a magnitude of 7 (like September’s Mexico City quake) over the past century, and they found an interesting pattern, correlated with another planetary cycle: variation in the Earth's rotation.

The Earth's rotation is known to undergo a slight change (just a few milliseconds per day) in its rotation rate over a multi-decade cycle. This shift doesn’t affect our daily lives, but it does slightly change the shape of the Earth. As the Earth’s rotation speeds up, the equator expands ever so slightly. As the Earth slows, the equator shrinks. The scientists think the added stress of that distortion might add an extra little nudge of pressure on tectonic plates that pushes faults that are already on the verge of an earthquake over the edge. Alternately, they propose, the mantle and crust of the Earth respond to the slowdown differently, and that distinction adds a similar nudge of pressure.

When the scientists compared the timing of this cyclical slowdown with historic increases in large earthquakes, they noticed that there tended to be more temblors near the time of the Earth's slowest rotation rate. The planet is currently in the midst of such a rotational slow down, prompting Bendick and Bilham to forecast a corresponding uptick in earthquakes.

According the scientists’ calculations, the likelihood of magnitude 7 or greater earthquakes is roughly 25 to 30 percent higher during the next five years, particularly around the equator. Their hypothesis has yet to be published and is undergoing peer review, but if it holds up, it could be another tool in forecasters’ toolboxes. They hope their research will help earthquake-prone regions near the equator prepare their infrastructure and emergency response systems to be more resilient in the event of a quake.

Buying critical seconds

On a shorter timescale, researchers and governments are working together to detect the first grumblings of an earthquake and alert communities that might be affected. Earthquake early warning systems use seismometers (instruments that measure motion in the ground) to pick up initial ground shaking. That signal is then processed to confirm that it is an earthquake, before triggering an alert. The whole process can take four or five seconds, or even fewer.

Seismic waves do not instantly affect everyone when a quake occurs, they propagate outward like the waves created by a stone dropped in a still lake. So for communities away from the epicenter of an earthquake, early warning systems could add valuable seconds or even up to a minute of time to prepare for the incoming quake.

But what can a few mere seconds do in such an immediate crisis situation?

Actually, a lot, says Bob de Groot of the U.S. Geological Survey. For individuals, even a few seconds of warning can buy time to move to a safe location, duck, cover their bodies, and find something to hold onto.

Early warning systems can also be connected to infrastructure control systems to trigger automated safeguards. Institutions like hospitals or public transit authorities can program an automated action in response to such an alert. At a hospital, for example, generators could kick on even before ground shaking reached the area, saving valuable time keeping key machines running. Or trains could be programmed to slow to a stop with such an alert, so they're not hurtling along tracks that might be disrupted by a quake.

On the earthquake-prone West Coast of the United States, the USGS is currently piloting just such an early warning system, called ShakeAlert. Efforts are underway to expand the seismometer network there for more accurate and expedient predictions.

Countries like Japan and Mexico have had earthquake early warning systems in place for more than a decade. And during Mexico City's devastating earthquake in September, the early warning system bought residents valuable seconds to get outside.

“The system has great potential, as has been demonstrated around the world, of saving lives and reducing damage,” Dr. de Groot says.