Can earthquakes be predicted? What the USGS and Japan's agencies say about prediction, probabilities and rumors

Every so often a date-specific "big earthquake" prediction spreads online, often aimed at Japan. Can anyone actually predict earthquakes? The U.S. Geological Survey (USGS) and Japan's own agencies give the same answer.

What "prediction" means

The USGS: "An earthquake prediction must define 3 elements: 1) the date and time, 2) the location, and 3) the magnitude." Japan's Headquarters for Earthquake Research Promotion uses the same idea: when, where, and how big, predicted scientifically before the quake.

The answer: no

• USGS: "Neither the USGS nor any other scientists have ever predicted a major earthquake."
• Japan's Headquarters for Earthquake Research Promotion says that, with current science, earthquake prediction is generally considered difficult, even for large quakes (our translation).
• The Japan Meteorological Agency (JMA) quotes a 2017 government panel: there is currently no scientifically established method for predicting the time, place and size of earthquakes with high accuracy (our translation). JMA stopped issuing information that assumed the Tokai earthquake could be predicted.

What science can do instead

• Probabilities. Japan publishes 30-year probabilities for faults and subduction zones. A low number doesn't mean safe: the fault behind the 2016 Kumamoto earthquakes had a probability of roughly 0% to 0.9% just before it ruptured, according to the Headquarters.
• "Relatively higher" chances. The Headquarters says it is considered possible to assess that the chance of a quake has risen relative to normal. Japan's megaquake advisories work this way; see "Japan's megaquake advisories explained."
• Early warning is not prediction. USGS: "Early warning is a notification that is issued after an earthquake starts."

Clouds, animals and prophecies

The USGS says earthquakes "have nothing to do with clouds, bodily aches and pains, or slugs," and that most claimed precursors "frequently occur without being followed by an earthquake." A prophecy with a date fails the scientific test before it starts. And a failed prophecy doesn't make the next day safe (our view).

What to do instead (our suggestions)

• Before sharing a prediction, check whether JMA, the USGS or local authorities say the same thing.
• Spend the energy on things that help on any day: securing furniture and knowing what to do when shaking starts. See "Earthquakes in Japan."

A Japanese version is available: "地震の予知はできるのか" (in Japanese).