Giant earthquakes: not as random as thought

January 30, 2018, University of Innsbruck
Researchers taking sediment cores on Chilean lake Calafquén (with Villarrica Volcano in the background). Credit: Maarten Van Daele

By analyzing sediment cores from Chilean lakes, an international team of scientists discovered that giant earthquakes reoccur with relatively regular intervals. When also taking into account smaller earthquakes, the repeat interval becomes increasingly more irregular to a level where earthquakes happen randomly in time.

"In 1960, South-Central Chile was hit by the largest known quake on earth with a magnitude of 9.5. Its tsunami was so massive that –in addition to inundating the Chilean coastline– it travelled across the Pacific Ocean and even killed about 200 persons in Japan," says Jasper Moernaut, an assistant professor at the University of Innsbruck, Austria, and lead author of the study. "Understanding when and where such devastating may occur in the future is a crucial task for the geoscientific community".

It is generally believed that giant earthquakes release so much energy that several centuries of stress accumulation are needed to produce a new big one. Therefore, seismological data or historical documents simply do not go back far enough in time to reveal the patterns of their recurrence. "It is an ongoing topic of very vivid debate whether we should model large earthquake recurrence as a quasi-regular or random process in time. Of course, the model choice has very large repercussions on how we evaluate the actual seismic hazard in Chile for the coming decades to centuries."

In their recent paper in Earth and Planetary Science Letters, Moernaut`s team of Belgian, Chilean and Swiss researchers presented a new approach to tackle the problem of large earthquake recurrence. By analyzing sediments on the bottom of two Chilean lakes, they recognized that each strong earthquake produces underwater landslides which get preserved in the sedimentary layers accumulating on the lake floor.  By sampling these layers in up to 8 m long , they retrieved the complete earthquake history over the last 5000 years, including up to 35 great earthquakes of a magnitude larger than 7.7.

Jasper Moernaut holding a core sample. Credit: Maarten Van Daele
"What is truly exceptional is the fact that in one lake the underwater landslides only happen during the strongest shaking events (like a M9 earthquake), whereas the other also reacted to "smaller" M8 earthquakes," says Maarten Van Daele from Ghent University, Belgium. "In this way we were able to compare the patterns in which earthquakes of different magnitudes take place. We did not have to guess which model is the best, we could just derive it from our data."

With this approach, the team found that giant earthquakes (like the one in 1960) reoccur every 292 ±93 years and thus the probability for such giant events remains very low in the next 50-100 years. However, the "smaller" (~M8) earthquakes took place every 139 ±69 years and there is a 29.5% chance that such an event may occur in the next 50 years.  Since 1960, the area has been seismically very quiet, but a recent M7.6 earthquake (on 25 DEC 2016) near Chiloé Island suggests a reawakening of great earthquakes in South-Central Chile.

"These Chilean lakes form a fantastic opportunity to study recurrence," says Moernaut. "Glacial erosion during the last Ice Age resulted in a chain of large and deep lakes above the subduction zone, where the most are getting generated. We hope to extend our approach along South America, which may allow us to discover whether e.g. earthquakes always rupture in the same segments, or whether other areas in the country are capable of producing giant M9+ earthquakes."

"In the meanwhile, we already initiated similar studies on Alaskan, Sumatran and Japanese lakes," says Marc De Batist from Ghent University. "We are looking forward to some exciting comparisons between the data from these settings, and see if the Chilean patterns hold for other areas that have experienced giant M9+ earthquakes in the past."

Explore further: Residual strain despite mega earthquake

More information: J. Moernaut et al. Larger earthquakes recur more periodically: New insights in the megathrust earthquake cycle from lacustrine turbidite records in south-central Chile, Earth and Planetary Science Letters (2017). DOI: 10.1016/j.epsl.2017.10.016

Related Stories

Residual strain despite mega earthquake

December 13, 2017

On Christmas Day 2016, the earth trembled in southern Chile. In the same region, the strongest earthquake ever measured occurred in 1960. A comparison of data from seismic and geodetic measurements during and after both earthquakes ...

Geological evidence for past earthquakes in Tokyo region

January 31, 2012

In 1923, a magnitude 7.9 earthquake devastated the Tokyo area, resulting in more than 100,000 deaths. About 200 years earlier, in 1703, a magnitude 8.2 earthquake struck the same region, causing more than 10,000 deaths.

Global map to predict giant earthquakes

December 12, 2013

A team of international researchers, led by Monash University's Associate Professor Wouter Schellart, have developed a new global map of subduction zones, illustrating which ones are predicted to be capable of generating ...

Recommended for you

Sierra snowpack could drop significantly by end of century

December 11, 2018

A future warmer world will almost certainly feature a decline in fresh water from the Sierra Nevada mountain snowpack. Now a new study by the Department of Energy's Lawrence Berkeley National Laboratory (Berkeley Lab) that ...

A glimpse into future oceans

December 11, 2018

Something peculiar is happening in the azure waters off the rocky cliffs of Ischia, Italy. There, streams of gas-filled volcanic bubbles rising up to the surface are radically changing life around them by making seawater ...


Adjust slider to filter visible comments by rank

Display comments: newest first

not rated yet Jan 31, 2018
I wonder what USGS honcho Susan Hough has to say about this finding? She emphatically said No! to any tidal/quake (or for that matter any periodic) connection just last month. A known repeat cycle of anomalistic/draconic/synodic/annual lunisolar periodic variations is 353 years. Within that span, there is a single maximal gravitational forcing that the moon and sun apply to the earth.
not rated yet Jan 31, 2018
"It is generally believed that giant earthquakes release so much energy that several centuries of stress accumulation are needed to produce a new big one"

- Which may explain the apparent escalation of activity in tyhe western part of the US as far inland as montana.

There is an arc of increased hazard shown on the map

-stretching from the Los Angeles area through las vegas, salt lake city, yellowstone, and farther north. Activity all along this arc has increased dramatically in the last few months, as well as throughout CA and NV.

Los Angeles is the area where the Pacific plate is currently hung up on its travel northward, and it just looks like this stress is propagating along this entire arc.

Almost as if this entire area wants to tear itself away from the rest of the continent

Perhaps the 1906 San Francisco magnitude 7.8 earthquake was a foreshock.
not rated yet Jan 31, 2018
" 'It is generally believed that giant earthquakes release so much energy that several centuries of stress accumulation are needed to produce a new big one'

- Which may explain the apparent escalation of activity in the western part of the US as far inland as montana."

The accumulation idea may be modeled as a semi-Markov process. If the trigger delay duration of a quake is proportional to its magnitude, there could be some self-selective filtering occurring. This would censor a monotonically declining histogram into a recurrence time that peaks away from zero.

This has to assume a non-dispersive accumulation rate though. If the accumulation rate of strain buildup varies, i.e some slow, some fast, then this would again smear the off-zero recurrence time peak much more than they report.

not rated yet Jan 31, 2018
Dr prepuk

I don't understand most of your post but I still think I'm right.

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.