Why earthquake predictions are usually wrong

Brent Dmitruk calls himself an earthquake predictor.

In mid-October, he told his tens of thousands of social media followers that an earthquake would soon hit at the westernmost point of California, south of the small coastal city of Eureka.

Two months later, a magnitude 7.3 struck the site in northern California - putting millions under a tsunami warning and growing Mr Dmitruk's following online as they turned to him to forecast the next one.

"So to people who dismiss what I do, how can you argue it's just a coincidence. It requires serious skill to figure out where earthquakes will go," he said on New Year's Eve.

But there's one problem: earthquakes can't be predicted, scientists who study them say.

It's exactly that unpredictability that makes them so unsettling. Millions of people living on the west coast of North America fear that "the big one" could strike at any moment, altering landscapes and countless lives.

Lucy Jones, a seismologist who worked for the US Geological Survey (USGS) for more than three decades and authored a book called The Big Ones, has focused much of her research on earthquake probabilities and improving resiliency to withstand such cataclysmic events.

For as long as she has studied earthquakes, Dr Jones said there have been people wanting an answer to when "the big one" - which means different things in different regions - will happen and claiming to have cracked the code.

"The human need to make a pattern in the face of danger is extremely strong, it is a very normal human response to being afraid," she told the BBC. "It doesn't have any predictive power, though."

With some 100,000 earthquakes felt worldwide each year, according to the US Geological Survey (USGS), it is understandable that people want to have warning.

The Eureka area - a coastal city 270 miles (434km) north of San Francisco, where December's earthquake occurred, has felt more than 700 earthquakes within the last year alone - including more than 10 in just the last week, data shows.

The region, which is where Mr Dmitruk guessed correctly that a quake would occur, is one of the most "seismically active areas" of the US, according to the USGS. Its volatility is due to three tectonic plates meeting, an area known as the Mendocino Triple Junction.

It is the movement of plates in relation to each other - whether above, below or alongside - that causes stress to build up. When the stress is released, an earthquake can occur.

Guessing that an earthquake would happen here is an easy bet, Dr Jones said, although a strong magnitude seven is quite rare.

The USGS notes there have been only 11 such quakes or stronger since 1900. Five, including the one Mr Dmitruk promoted on social media, happened in that same region.

While the guess was correct, Dr Jones told the BBC that it's unlikely any earthquake - including the largest, society-destroying types - will ever be able to be forecasted with any accuracy.

There is a complex and "dynamic" set of geological factors that lead to an earthquake, Dr Jones said.

The magnitude of an earthquake is likely formed as the event is happening, she said, using ripping a piece of paper as an analogy: the rip will continue unless there's something that stops it or slows it - such as a water marks that leave the paper wet.

Scientists know why an earthquake occurs - sudden movements along fault lines - but predicting such an event is something the USGS says cannot be done and something "we do not expect to know how any time in the foreseeable future".

The agency notes it can calculate earthquake probability in a particular region within a certain number of years - but that's as close as they can come.

Geological records show that some of the largest types of earthquakes, known as "the big one" to locals, do happen with some amount of regularity. The Cascadia subduction zone is known to slip every 300 to 500 years, regularly upending the Pacific north-western coast with 100-ft (30.5 metres) tall mega-tsunamis.

While the San Andreas fault in Southern California is also the source of another potential "big one", with bone-rattling earthquakes happening there every 200-300 years. Experts have said the "big one" could happen at any moment in either region.

Dr Jones says over her career, she's had several thousand people alert her to such predictions of a big earthquake - including people in the 1990s who would send faxes to her office in hopes of alerting them.

"When you get a prediction every week, somebody's going to be lucky, right":[]}