About 13–15 minutes after the magnitude 9.0 Great East Japan Earthquake on Mar. 11, 2011, the Japanese archipelago shifted ever so slightly east again. The motion, only up to 5–6 mm, was too small for people to feel, but it was clearly recorded in Japan's GPS data and on seismometers.
A recent study found that this motion was caused by seismic waves that returned from deep within Earth. A team led by Park Sun-young, a professor at the University of Chicago, reported on the 19th in the international journal Science that S-waves (shear waves) generated by the Great East Japan Earthquake traveled thousands of kilometers down to near the core, reflected, rose back to the surface, and simultaneously disturbed plate boundaries around Japan. This is the first confirmed case of returning deep-Earth seismic waves reactivating plate boundaries.
Park, who led the study, graduated from the Department of Energy Resources Engineering at Seoul National University and joined the University of Chicago in 2021. Park drew attention at the time as the first seismology professor in the university's history. Park has studied how seismic waves travel through Earth's interior and how that process affects the surface and plate boundaries.
What caught Park's attention in this study was a small signal across Japan more than 10 minutes after the main quake ended. Additional motion after a large earthquake is usually related to aftershocks or slip on nearby faults. But this signal was hard to explain by aftershocks or additional ruptures alone.
"Additional motion observed across Japan is generally associated with another earthquake, but there was no quake that matched that timing," Park said. "I started this study wondering what caused this movement."
The team focused on ScS waves. ScS waves are seismic waves in which S-waves reflect near the core–mantle boundary after traveling through Earth's interior and then return to the surface. The team reasoned that because the Great East Japan Earthquake was so powerful, the ScS waves reflected near the core would have retained some energy on their return.
Observations backed up the team's interpretation. ScS waves were clearly identified at observation stations across Japan, followed by step-like eastward changes in GPS data. The fact that the signal did not stay in one region but was recorded at multiple stations around the same time added weight to the ScS-wave explanation.
"This ScS wave traveled about 5,800 km, almost Earth's radius," Park said. "Seismic waves that go down near the core and come back are generally thought not to have much energy, making it hard for them to trigger other quakes, and no such case had been reported."
The team concluded that the phenomenon likely occurred as friction on plate boundaries weakened after the main shock. The strong main shock destabilized the plate boundaries, and ScS waves reflected near the core acted as an additional stimulus, causing small, simultaneous slips along plate boundaries around Japan.
The researchers also examined other causes besides ScS waves. They considered whether the rupture process of the main shock itself belatedly produced surface motion, or whether a massive submarine landslide that can occur after a great earthquake played a role. But while those causes could explain movements near the main shock area, they struggled to account for the consistent eastward motion across all of Japan.
They also considered whether the signal might be an artificial error introduced during GPS data processing, but the pattern differed from known error signatures. Above all, an independent match in seismometer records at the same time made it hard to attribute the signal to a simple GPS error.
The study also affects how we assess risks that can arise after a great earthquake. While attention typically focuses on aftershocks after the main shock ends, the study shows that seismic waves returning through Earth's interior can move plate boundaries again 10 or 20 minutes later.
The team said, "The range of the triggered slip is unusual relative to the amount of motion," adding, "its total length is about 3,000 km, similar to the length of Japan's main islands, 6–7 times longer than the rupture length of the Great East Japan Earthquake's main shock, and more than twice as extensive as during the 2004 Sumatra earthquake."
On this, Park added, "We should treat this type of triggering as a new seismic hazard factor," noting, "It applies not only to Japan but also to other regions prone to giant subduction earthquakes, such as Alaska, South America, and Indonesia."
References
Science (2026), DOI: https://doi.org/10.1126/science.aec4190