Magnitude 6.9 Misreported as 9.9—And Yet Japan's Earthquake Early Warning System Is Still Remarkable

Recently, Japan’s Earthquake Early Warning system caused a bit of a commotion.

An earthquake that was actually magnitude 6.9 was somehow assessed as magnitude 9.9, triggering an emergency alert warning people to prepare for an earthquake of the largest possible scale.

A magnitude 9.9 earthquake would far exceed the magnitude 9.0 Great East Japan Earthquake. In terms of energy, it would be about 22 times greater.

The mistake was later discovered and corrected, but it was certainly a spectacular error.

Yet when I saw the news, what struck me anew was not the false report itself, but how extraordinary the Earthquake Early Warning system is.

Mobilizing an Entire Society Before the Earthquake Arrives

When an earthquake is detected, the Earthquake Early Warning system uses the P-waves that arrive first to predict the stronger shaking that will follow.

It detects the earthquake, estimates its epicenter and magnitude, predicts the shaking in different locations, and distributes the information across the country.

It does all of this in just a few seconds.

When you stop to think about it, that is an astonishing technology.

And humans are not the only ones receiving the alert.

Trains automatically slow down or stop, and elevators stop at the nearest floor. Some factories also have systems that shut down dangerous equipment.

Of course, not every piece of equipment supports these functions, but automated safety measures operate throughout many parts of the country’s social infrastructure.

If people had to watch the news, make a decision, and call others, there would be nowhere near enough time.

With only a few seconds of warning, an enormous social system moves all at once.

I used to think of it as little more than an annoying alert that suddenly made my smartphone ring. Behind it, however, an extraordinary mechanism is at work.

Even When an Alert Is Correct, People Grow Accustomed to It

There is, however, a difficult problem here.

The problem is not false alerts, but the fact that people become accustomed to warnings even when the alert is correct.

The earthquake alert sounds, so you brace yourself, but there is only a little shaking. Or perhaps you barely feel anything where you are.

Experiences like that are not unusual.

Of course, the alert may have been appropriate when the entire target area is considered. Earthquake intensity varies greatly by location.

But that distinction means little to the person receiving it.

“The alert went off again, but it was no big deal.”

After this happens repeatedly, people gradually stop taking the warnings seriously.

Machines are different.

When they receive an alert, their safety systems activate according to predetermined conditions.

They do not decide to keep running today just because the shaking yesterday and the day before was minor.

Machines do not complain about previous false alarms. Humans, however, become more complacent with experience.

Even after an advanced warning system has been completed, the final problem that remains is the particularly troublesome one of human psychology.

Accuracy and Speed Cannot Be Fully Reconciled

Then should we simply make the system more accurate?

It is not that easy.

The Earthquake Early Warning system predicts shaking that has not yet arrived using limited observational data.

If it waits until enough information has been collected, the earthquake itself will arrive first.

Issue the alert quickly, and the margin of error grows. Proceed cautiously, and the warning arrives too late.

There is an unavoidable trade-off between speed and accuracy.

Moreover, just because someone who received the warning did not experience major shaking does not necessarily mean the prediction was wrong.

The system prioritizes warning people before dangerous shaking arrives, even if that means tolerating some false alarms.

As the design of a safety system, that is extremely rational.

The Unstable Component Left at the End: Humans

The Earthquake Early Warning system cannot prevent an earthquake itself.

All it can do is buy a few seconds before the strong shaking arrives.

But in those few seconds, a train can slow down. Dangerous machinery can be stopped. People can take action to protect themselves.

Viewed across society as a whole, the value of those seconds is incalculable.

The fact that such an advanced system operates as part of everyday life deserves greater recognition.

Still, no matter how far technology advances, human responses cannot be completely controlled.

Even when the alert sounds, people may assume that it will probably be nothing serious.

We have acquired the technology to predict an earthquake a few seconds in advance. Yet we still cannot predict whether the people who hear the warning will act.

It seems that the most difficult thing to control in an Earthquake Early Warning system is not the earthquake, but the human being.