Skip to Main Content
Skip Nav Destination

Hidden water puts the brakes on earthquakes

11 August 2021

When extinct undersea volcanoes are shoved beneath adjacent tectonic plates, they can carry trapped seawater with them.

Two people lowering a device into the ocean from a ship
Christine Chesley (left) and technician Jacob Perez deploy an electromagnetic instrument off the coast of New Zealand in December 2018. Credit: Samer Naif

All around the world, tectonic plates are doing the same two-step dance. As adjoining plates gradually shift relative to each other, they slowly build up stress at their interface until suddenly, the deformation is too much for the interface to bear, the rock fractures, and the pent-up energy is released all at once: an earthquake.

But sometimes—and at some plate boundaries more than others—the energy release happens over a leisurely period of weeks to months rather than seconds to minutes. (See the article by John Vidale and Heidi Houston, Physics Today, January 2012, page 38.) Understanding how and why those so-called slow-slip events occur would be a big step toward being able to forecast the faster, more destructive earthquakes. But despite two decades of research, the inner workings of a slow-slip event remain unclear.

One pressing open question is what happens when an oceanic plate that’s pockmarked with bumps, called seamounts, subducts beneath an adjacent plate. Do the seamounts get caught on the underside of the upper plate, lock the plates together more firmly, and lead to more violent earthquakes? Or do they crunch their way through the upper plate and break it up into a mushier system that’s more prone to slow slip? Evidence abounds to support both theories, so the matter is far from settled.

So far, most researchers have approached the question by studying the seamounts’ topography alone. But work by Columbia University’s Christine Chesley and colleagues now suggests that an important part of the picture is the seamounts’ internal structure, which stems from their history as underwater volcanoes. The long-ago eruptions trapped large amounts of seawater inside the volcanic rocks, where it remained for many millions of years. The researchers find that when a seamount gets subducted, the water is released, lubricates the plate boundary, and leads to slow slip.

To visualize the deeply buried water, Chesley and colleagues dropped electromagnetic sensors, as shown in the photo, onto the seafloor around the Hikurangi Margin, a subduction zone just east of New Zealand’s North Island that’s known for its slow slip. (See Physics Today, July 2016, page 18.) Seawater is a far better conductor of electricity than solid rock is, so low-resistivity regions indicate the presence of trapped water.

A graph showing resistivity as a function of depth and distance from the Hikurangi deformation front
Credit: Adapted from C. Chesley et al., Nature 595, 255 (2021)

Their measurements, shown in the figure, reveal a vertical slice through two seamounts: one on the left that’s already been subducted beneath the overlying plate (the plate boundary is marked by the solid black line) and one on the right that hasn’t. The unsubducted seamount has the expected structure, with a core of solid rock (R1p, on the right) topped with layers of water-laden volcanic rocks (C1p and C2p) and capped with another layer of less porous rock (R2p).

The subducted seamount has a similar resistive core (R1f, on the left), but some of its water may have been released into cracks in the overlying plate (C1f and C2f). Furthermore, one of the water-rich regions, C2f, coincides exactly with observed areas of recent slow-slip seismicity (white circles and stars).

An analysis of just one subducting seamount doesn’t provide the final answer for how hidden water affects seismic activity in general. To test their interpretation, the researchers’ next step is to analyze the measurements they’ve taken elsewhere along the Hikurangi Margin, in regions with no subducted seamounts in contact with the overlying plate. (C. Chesley et al., Nature 595, 255, 2021.)

Subscribe to Physics Today
Get our newsletters
Close Modal

or Create an Account

Close Modal
Close Modal