Striving to shed light on the mechanism of major earthquakes

The Great Hanshin-Awaji Earthquake of 1995 prompted the government to establish a nationwide network to monitor seismic activity, and this unparalleled and sophisticated observation system is making it possible to instantaneously detect the changes occurring underground. Professor YOSHIOKA Shoichi of the Research Center for Urban Safety and Security is an expert in seismology and analyzes the data obtained from the network as he takes on the challenge of shedding light on the mechanism of major earthquakes. We asked him about his research which aims to unravel the mystery of how earthquakes occur by focusing on subsurface temperature structures and slow slip events that happen at the boundary between tectonic plates.

The Great Hanshin-Awaji Earthquake prompted efforts to organize an observation network to monitor seismic activity

Have seismological approaches changed since the Great Hanshin-Awaji Earthquake?

Yoshioka: After the Great Hanshin-Awaji Earthquake, we saw the enactment of the “Special measure law on earthquake disaster prevention” in July 1995, and the Headquarters for Earthquake Research Promotion was established as a governmental organization to centralize the results of earthquake research, which they would then communicate to the general public. Before this was created, the main universities of each region were responsible for monitoring seismic activity. But 1995 was when everything changed. Efforts were made to organize a nationwide earthquake observation network to gather all of the data in one place and communicate earthquake information to the public.

As for the specific systems with the earthquake observation network, the National Research Institute for Earth Science and Disaster Resilience (NIED) operates the High Sensitivity Seismograph Network, or Hi-net, which has 1,000 observation stations at intervals of 20 km all over Japan. These observation stations are located underground, which reduces noise from cars and trains and makes it possible for observations to be more accurate. The installation of the strong-motion seismograph networks K-NET and KiK-net has also made it possible to monitor strong ground motion more precisely.

Meanwhile, the Geospatial Information Authority of Japan (GSI) has taken the lead in setting up a continuous GNSS (Global Navigation Satellite System) observation network, which is basically a type of GPS, and this has approximately 1,300 locations nationwide. The development of this system has enabled us to observe phenomena including slow slips, where tectonic plates slip slowly past each other.

These are all land observation networks, but other systems have also been developed after the Great East Japan Earthquake to make it possible to observe tsunamis by laying cables on the seafloor. One of them is the S-net (Seafloor observation network for earthquakes and tsunamis along the Japan Trench), which has 150 observation devices consisting of seismometers and water pressure gauges on the Pacific Ocean floor from off the coast of Hokkaido Prefecture to the Boso Peninsula of Chiba Prefecture. Earthquake and tsunami observation systems DONET-1 and DONET-2 have also been established to the south of Kumano-nada and the Kii Channel to observe earthquakes and tsunamis that originate in the Nankai Trough. All of these systems are managed and operated by NIED.

This earthquake observation network is the densest observation system in the world, and has enabled us to detect underground phenomena instantaneously.

Progress being made in assessing active faults

How about research into predicting earthquakes? How much progress has been made in that area?

Yoshioka: It is said that Japan has roughly 2,000 active faults, and the Headquarters for Earthquake Research Promotion is working on assessments for the 114 major active faults.

For each active fault, they examine when earthquakes have occurred in the past, assess how strong a potential earthquake could be in the future, and announce the probability of a megaquake happening in the next 30 years. They have more or less completed examining the active faults on land, and are now applying the same assessment process to active faults on the seafloor.

Earthquake prediction is about forecasting the timing, location, and magnitude of a future earthquake, but current earthquake research is yet to reach the level where we are able to make actual predictions. However, the development of the earthquake observation network does mean we are making progress in creating a system that will detect and notify us of any movement.

For example, the Hyuga-nada earthquake of August 2024 caused the first-ever “Nankai Trough earthquake extra information (major earthquake warning)” to be issued. The probability of a major earthquake occurring within 30 years had previously been estimated at 70%-80%, so I believe that this is exactly what happened in this instance.

What specific knowledge has been gained with the dense observation network?

Yoshioka: Hi-net manages to detect the slight movements of deep low-frequency earthquakes; and detailed examination has revealed that tremors are happening deep down where the tectonic plates meet, from the northwestern part of Shikoku all along the southeastern part of the Kii Peninsula to the Tokai region. This is the zone where the Nankai Trough megaquake is predicted to occur.

As I mentioned earlier, the continuous GNSS observation network has also made it possible to observe the slow slip phenomenon. In a regular earthquake, tectonic plates move, releasing the strain energy accumulated underground, and the fault slips at high speed, causing the ground to shake. Meanwhile, a slow slip event is a phenomenon where a plate boundary fault shifts slowly, and the movement can only be observed by the GNSS continuous observation network. It was first confirmed in 1997 with the subduction of the oceanic plate beneath the Bungo Channel between Kyushu and Shikoku. Since then, slow slip events have also been confirmed in other parts of the Pacific Rim, including the United States, Canada, New Zealand, and Alaska.

Slow slip events can be long-term changes that occur over several months or years, or short-term changes that happen over two or three weeks. Short-term slow slip events occur at a depth of around 30-40 kilometers, which is practically where deep low-frequency tremors occur. So this has led to the idea that perhaps short-term slow slip events cause low-frequency tremors.

We are hoping that computer simulations of deep low-frequency tremors and slow slip events will help us understand the mechanism of major earthquakes and perhaps make it possible to have some idea of when they will happen.

Analyzing slow slip events from subsurface temperature structures

What motivated you to study slow slip events?

Yoshioka: I became interested in slow slip events when I was studying abroad at a research institute in Canada and learned that slow slip events were also being confirmed on the west coast of Canada. That was when I began analyzing the data.

The destruction caused by a slow slip event may lack speed, but we are now discovering that the energy released is quite large. When a slow slip event occurs, energy accumulates in the shallow part of that plate, and some people believe that this may be what causes a major earthquake such as the Nankai Trough earthquake.

In the case of the Great East Japan Earthquake (2011 Tohoku Earthquake and Tsunami), we later found that an unusual slow slip event had started about a month before March 11 when the earthquake struck. As such, abnormal slow slip events are now one of the criteria assessed before issuing a Nankai Trough Earthquake Extra Information warning, which is provided when an abnormal phenomenon is observed along the Nankai Trough, or when the possibility of an earthquake is evaluated as relatively high.

What kind of research are you currently focused on?

Yoshioka: Besides my research into where and how slow slip events occur, I am also working on using computer simulations to analyze the movement of tectonic plates in three-dimensional space, simulating for example the direction and depth of plate subductions since 15 million years ago. I also analyze Hi-net data such as heat flows from the Earth’s interior to study models of subsurface thermal structures.

As a result, we have reached the conclusion that perhaps slow slip events, which are traditionally believed to happen at a depth of 30-40 kilometers underground, actually occur in places with temperatures of around 350 ºC, rather than areas defined by depth. How close will we get to actually clarifying the mechanism? I do not know. But I hope to continue with this dual approach of studying slow slip events and thermal structure models to further my research.

What are your thoughts on the future of earthquake research, including the ability to predict earthquakes?

Yoshioka: The probability of a Nankai Trough megaquake occurring within the next 30 years is estimated at 70%-80%, and this is based on an examination of past earthquakes in the area. Now that we have developed an advanced network to observe seismic activity, I hope we can scientifically improve the accuracy of our predictions by leveraging these infrastructures.

That said, there are still many underground phenomena that we do not understand, just like what is happening in outer space. It would be wonderful if we could identify what kind of underground phenomena causes earthquakes as if we were solving an equation. But unfortunately, we are still yet to learn if they even follow a pattern, or are simply chaotic random events. But I am determined to carry on with my research to close in on the secrets of earthquakes.

Resume

Graduated from Kobe University Faculty of Science in 1985, and received his Ph.D. in geophysics from Kyoto University Graduate School of Science in 1990. Became a research associate at Ehime University Faculty of Science and an associate professor at Kyushu University Graduate School of Science, before becoming a professor at Kobe University Research Center for Urban Safety and Security and Graduate School of Science in 2009.