Earthquakes don’t wait for time or tide


US research contradicts early studies that found quake frequency is affected by season. Andrew Masterson reports.


Springing into action: research reveals no link between season and earthquakes.
Springing into action: research reveals no link between season and earthquakes.
Ben Wels / Getty Images

Contrary to the baffling and counter-intuitive results of several earlier studies, earthquake frequency is not affected by the season, nor by tides.

That’s the central finding of a paper in the journal Seismological Research Letters, written by Susan Hough of the U.S. Geological Survey.

The idea that earthquake frequency is affected by the time of year was first mooted in the 1930s. Since then, there has been a number of studies which seem to show that the notion can be backed up with evidence.

In 2006, for instance, a report in the journal Science found “strong seasonality” in Greenland quakes linked to the movement of glaciers.

In 1999, Betim Muco of the Albanian Seismological Institute found that variations in the underground water supply, fed by rainfall, could be influencing the seasonality of earthquakes in the Balkans.

A 1984 paper in the journal Nature revealed that quakes along the San Andreas Fault in the 25 years before the big one in San Francisco in 1904 were most common during spring.

Other studies have identified seasonal weighting for quake numbers in several other countries, including Japan, India and Iran.

The latest research, however, suggests that these findings are perhaps all products of statistical error or the product of randomness.

To exclude as much as possible the presence of clustered quakes – when a main event is followed by several aftershocks – Hough decided to look for patterns only in big quakes above magnitude eight.

To do this, she identified 204 candidate events stretching back to the Seventeenth Century and logged them against season. She also logged them against lunar phases information to test the suggestion that quake rates might be influenced by tides.

To control for any coincidences, such as a few quakes occurring on a single day in different locations, she randomised the dates and the data to see what, if any, patterns would show up.

Correlation between quake numbers and season turned out to be flat. A couple of apparent conjunctions between time and tide did appear, but their relationship, it turned out, was more apparent than real. For instance, she found 16 earthquakes occurred on a single day, which fell seven days after a new moon.

This, she says, is just a blip. The total is not statistically significant, and, anyway, on that day the lunar tides would have been at their minimum, thus excluding them as an influencing force.

“When you have random data, you can get all sorts of apparent signals, just like when you flip a coin, you sometimes end up with five heads in a row,” she says.

Despite the solidity of her findings, Hough has no illusions that her paper will put an end to the periodic rise and fall of these two bits of earthquake lore.

“Sooner or later there is going to be another big earthquake on a full moon, and the lore will pop back up,” she says.

“The hope is that this will give people a solid study to point to, to show that over time, there isn't a track record of big earthquakes happening on a full moon.”

Contrib andrewmasterson.jpg?ixlib=rails 2.1
Andrew Masterson is news editor of Cosmos.
  1. http://dx.doi.org/10.1785/0220170154
  2. http://science.sciencemag.org/content/311/5768/1756
  3. http://science.sciencemag.org/content/311/5768/1756
  4. http://www.sciencedirect.com/science/article/pii/S0031920199000515
  5. ftp://ftp.gps.caltech.edu/pub/avouac/Ge277-2007-spring/McClellan_1984.pdf
  6. ftp://ftp.gps.caltech.edu/pub/avouac/Ge277-2007-spring/McClellan_1984.pdf
Latest Stories
MoreMore Articles