Previous: Space News From The Future!
Next: Warp Drives!



View count:296,105
Last sync:2023-01-10 01:30
Hank talks about why it is so difficult for scientists to predict earthquakes in the short term.

Like SciShow?
Follow SciShow!
T*mbl SciShow!

References for this episode can be found in the Google document:

On April 6th, 2009, an earthquake struck new the Italian city of L'Aquila, killing nearly three hundred people. It was a terrible disaster, and you've probably heard of it because six Italian scientists and one government employee were later tried and convicted on criminal manslaughter charges for providing inaccurate, incomplete, and contradicting information about the earthquake danger.

They were sentenced to six years in jail, basically for failing to predict an earthquake. This has riled up a lot of folks in the science world, and for good reason, because while we know a lot more than ever about earthquakes, there is still no reliable way to predict them. Now, can we forecast where earthquakes are likely to occur over a long period of time? Yes, of course we can. For example, scientists know that there have been twenty major earthquakes of magnitude nine or higher in the Cascadian region between Vancouver and Central California over the past ten thousand years. The last one took place in 1700 and there will be another in the future. Sometime.

But can seismologists predict the time, epicenter location, and magnitude of the next Cascadia earthquake? They cannot. There are lots of reasons for this, but it's mainly because it's so hard to study tectonic plates -- the giant slabs of the earth's crust that are always moving around, since they're usually at least one hundred kilometers thick, so while scientists believe that quakes start when two or more plates meet, the process that occurs before they start is subtle and almost impossible to observe.

That doesn't mean that there aren't potential indicators that can be studied or loads of intriguing anecdotal evidence that some people believe can help predict some earthquakes. For instance, many scientists think that rising levels of radon are a good indicator of an impending earthquake. A radioactive gas produced by the natural decay of uranium, radon builds up in cavities beneath the surface of the earth and may be released by shifts in the ground preceding a quake. But, while radon leaks have preceded some quakes in the past, in many instances, they haven't.

Others believe that changes in electromagnetic fields near fault lines indicate an impending quake. A Silicon Valley company called "Quake Finder" has installed two hundred sensors to measure changes in magnetic fields below the ground and to detect charged particles in the air, but the US Geological Survey doesn't support the project, calling it a scientific dead end.

Finally, almost everybody has heard of examples of animals predicting earthquakes. Examples go as far back as 373 B.C. Back then, as the story goes, pretty much every critter high tailed it out of the Greek city of Helike just five days before it was destroyed by a massive quake. And just seventy-four kilometers from L'Aquila, biologists studying toads reported that ninety-six percent of the amphibians abandoned their breeding site, again, five days before the 2009 disaster. Alas, as intriguing as they are, you never hear any of these stories before an earthquake strikes. While there might be something to this, strange animal behavior, like other precursors to earthquakes, is generally only observed in hindsight.

Plus, it's worth pointing out that other studies have found that animals do not respond to the threat of earthquakes at all. A group of biologists was studying ant colony behavior in the Mohave desert in 1992 when a 7.4 magnitude earthquake struck the region. The scientists reported that "contrary to anecdotal reports of earthquake effects on ant behavior", the quake had no effect on the insects.

So in order to predict an earthquake, and in turn, send out warnings of imminent danger to the public, you need reliable and repeatable indicators. And those just don't exist, yet, at least. Until they do, you should probably throwing scientists in jail for failing to do something that's impossible.

Thank you for watching this episode of SciShow. If you have any questions or comments or ideas, we're on Facebook and Twitter, and of course down in the comments below. And if you wanna keep getting smarter than us, you can go to and subscribe.