Earthquake prediction isn’t an art: it’s a fantasy. In a bitter blow to geologists, it turns out the foreshocks preceding a huge quake in Turkey nearly 20 years ago were not harbingers of the big one that followed – and future small quakes can’t be used to predict major temblors, as had been theorized.
In 1999, a quake of magnitude 7.6 killed more than 17,000 people near Izmit, northwestern Turkey. That jolt had been preceded by multiple small quakes that were interpreted, in a 2011 paper in Science, as “slow-slip foreshocks” – all originating in the same place. In their paper optimistically titled “Early Warning Signals Detected For the First Time,” the scientists then postulated that detecting foreshocks in other cases could serve to predict a big quake to come.
No it can’t, concludes the paper published Monday in Nature Geoscience by scientists at Stanford University and Istanbul’s Bogazici University. Using technologies not available in 2011, they revisited the Izmit disaster and concluded that the “foreshocks” were unrelated small quakes and could not serve to predict a thing.
“We’d all like to find a scientifically valid way to warn the public before an earthquake begins,” says co-author Fatih Bulut of Bogazici. “Unfortunately, our study doesn’t lead to new optimism about the science of earthquake prediction.”
That is a shame for Turkey, which is known for seismic activity. Most of Turkey lies on the small Anatolian Plate, which has a fault on either side of it. The rest of the country is not seismically tranquil, being affected by the tectonics in the Aegean seabed.
It isn't that foreshocks mean nothing. About half of all major quakes on the planet are preceded by small quakes, and it is perfectly clear that a flurry of small quakes could mean a big one is coming ... or it could mean nothing is coming.
There are two models for foreshocks: the slow-slip model and the cascade model, the scientists explain.
The cascade model posits that a minor quake along a given fault triggers another nearby; that triggers another; and the series may randomly trigger a major temblor. Or the series might just as easily peter out and the worms can settle back down to business.
“It’s a bit like dominos,” Bulut said. “If you put dominos on a table at random and knock one over, it might trigger a second or third one to fall down, but the chain may stop. Sometimes, you hit that magic one that causes the whole row to fall.”
In other words, in the cascade model, foreshocks may or may not mean a thing.
Izmit is smack on the North Anatolian Fault, a vast crack in the Earth stretching 1,500 kilometers (930 miles) across Turkey. In its case, the theory had been that the smaller quakes preceding the monster were caused by gradual, slow slippage along that whopping great fault. The 18 foreshocks measured in the 1999 Izmit quake were thought to have been small slip ruptures at the same part of the fault, ultimately triggering the big earthquake.
When a theory falls in the forest...
The 2011 paper argued that Izmit’s slow-slip foreshocks could have been detected if only we had the right equipment – and if they had been detected, we could start predicting major earthquakes. The 2018 paper, based on a lot more data, says they weren’t slow-slip quakes but cascade quakes. And, in either case, we still can’t predict big ones. Or any quakes.
“There was nothing diagnostic in their occurrence that would suggest that a major earthquake was about to happen,” concludes lead author William Ellsworth of Stanford.
In Izmit, the data for the 2011 paper originated in a single seismic station miles from the earthquake epicenter: it recorded 18 foreshocks occurring about 14 kilometers below the surface, near where the larger earthquake began. The quakes also seemed to have similar waveforms. So the researchers concluded that the foreshocks all happened in the same place, breaking the same spot along the rift, and that the slow slip ultimately triggered the major earthquake.
So, they postulated, a future series of small quakes on that fault could presage a future big one.
But the 2018 paper checked data on that 1999 quake from 10 seismic stations, not one. From the new data, they deduced that there had been a total of 26 foreshocks.
These small quakes were not identical and, if anything, were not co-located but moved along the fault line. Ergo, Izmit hadn’t been a slow-slip situation at all but apparently a cascade – one quake leads to another, but they don’t have to lead to a major quake. Maybe they did. Maybe it happened on its own in 1999.
The thing is that geologists have no theory for why some cascades wind up triggering huge, deadly quakes and others peter out. So merely noticing a series of quakes means, at this point, zip for earthquake prediction.
In short, quake prediction will have to wait for better equipment and better theories. Stay tuned.
“We’re not giving up on foreshocks just because we currently can’t tell them apart from other earthquakes,” Ellsworth stated for the record. “We want to understand if they have predictive value – and if not, why not. To answer that question will require observations made close to the action, deep in the heart of the earthquake machine, not as we currently do from the surface where we’re blind to changes deep underground.”