First author

Across the subduction zones of Japan, where one tectonic plate dives beneath another, the propagation of seismic waves varies from place to place. Earth scientist Ikuo Katayama of Hiroshima University in Japan sought to find out why. By using a high-pressure device to squeeze serpentine, a hydrous mineral, to determine how its crystal orientation would develop, he and his colleagues simulated what might happen in the Japanese subduction zones (see page 1114). Katayama tells Nature more.

What inspired the study?

I realized that in some subduction zones, data on seismic anisotropy — the fact that seismic waves move at different speeds according to their direction — could not be explained by the classical idea that wave propagation depends on the crystal orientation of the mineral olivine, a major constituent of Earth's mantle. We wanted to test a newer idea that serpentine might contribute to seismic anisotropy. Serpentine can form in the mantle, where tectonic plates converge.

How did you conduct your experiment?

We simulated the conditions of the subducting-plate interface by cooking the materials thought to be present in Earth's mantle at high pressure and temperature. We then analysed the crystal orientation of the deformed minerals using scanning electron microscopy and computed the seismic anisotropy.

What did you learn?

We found that the crystal orientation that results when serpentine forms at such high temperature and pressure could explain the seismic anisotropy observed in subduction zones such as the Ryukyu arc, an island arc that connects southwest Japan and Taiwan.

What can your findings tell us about the subduction process more generally?

Water is transported into Earth's interior at subduction zones and has important roles in volcanism and seismic activity. Serpentine's presence in subduction zones suggests that water can be released at shallow depths in a warm subduction zone such as the Ryukyu arc. Serpentine can form in warm subduction zones as water is ejected from the descending plate and reacts with mantle rocks. It is not expected in dry conditions. In a cold subduction zone, such as that in northeast Japan, water is not supplied to the shallow mantle wedge. The heterogeneous distribution of water — or serpentine — that we propose may cause distinct geological processes in warm and cold subduction systems. This may be why seismic activity differs between these regions.