New Radiocarbon Method Reveals Earthquakes Hidden in Japanese Seafloor Mud

A team of geoscientists at the University of Tokyo has developed a radiocarbon dating method sensitive enough to detect earthquakes that left no visible mark on the seafloor, uncovering several previously unrecognized disturbances buried in a sediment core from the Kumano Sea off central Japan. The findings, published this week in the journal Radiocarbon, describe sediment layers that looked identical to the undisturbed mud around them under every standard method of inspection, yet carried a chemical signature of having been violently reshuffled, most likely by earthquakes along the nearby Nankai Trough.

The core itself is unremarkable to the eye: 40 centimeters of fine, gray clayey silt, the kind of featureless "hemipelagic mud" that accumulates slowly on the seafloor over centuries. Researchers Akiko Omura, Yusuke Yokoyama, Shoko Hirabayashi, Yosuke Miyairi and Juichiro Ashi, all at the university's Atmosphere and Ocean Research Institute, measured radiocarbon in the organic carbon trapped within that mud at intervals of roughly a centimeter or finer, building a far more detailed age profile than earlier studies of similar cores attempted. The deepest, oldest sediment they examined dated to around A.D. 1440.

A Signal With No Visible Source

At several points in the core, the team found something that should not occur in an undisturbed mud column: organic carbon that measured older than the material directly beneath it. Normal seafloor deposition produces a steady march of ages, each layer slightly younger than the one below. An age reversal implies that older, previously buried carbon was torn loose and redeposited on top of younger sediment, which the researchers attribute to submarine slope failures, slumps of sediment dislodged and resettled, possibly triggered by shaking from an earthquake.

What makes the result notable is not the mechanism, which paleoseismologists have long used to read earthquake history from the seafloor, but the fact that these particular layers were undetectable by any other means the team tried. They did not show up as distinct bands to the naked eye, in X-ray imaging, or in CT scans, because the redeposited material is physically and visually the same stuff as the sediment around it. Only the carbon-14 clock, run at very fine resolution, could tell the two apart.

"We found several new geological layers that were not visible before," said Yusuke Yokoyama, a professor at the Atmosphere and Ocean Research Institute, in comments accompanying the university's announcement of the findings. "We used a new, ultrasensitive method to measure the trace amounts of carbon-14 … This revealed multiple events which were not found through previous visual observations."

To convert the raw radiocarbon measurements into calendar dates, the team built an age model using the shells of planktonic foraminifera, microscopic marine organisms whose remains are mixed into the same sediment and can be dated with fewer complicating assumptions than bulk organic matter. Cross-checked against that timeline, several of the disturbed layers lined up with periods when written Japanese records describe submarine earthquakes, a correspondence the authors say supports, though does not prove, an earthquake origin for the events. The historical record the team compared against stretches back to the late 15th century.

Why the Kumano Sea Matters

The Kumano Sea sits directly above a stretch of the Nankai Trough, the subduction zone running along Japan's southern coast where the Philippine Sea plate dives beneath the Eurasian plate. It is the source of Japan's most closely watched seismic hazard: a repeating sequence of magnitude-8-to-9 megathrust earthquakes that has struck the region roughly every 100 to 200 years for centuries, most recently in 1944 and 1946. Japan's Headquarters for Earthquake Research Promotion, the government body that issues official seismic forecasts, has put the 30-year probability of the next such event at roughly 70 to 80 percent in its standard assessments, a figure the panel has since revised upward using new statistical methods, underscoring how seriously the risk is treated and how much rides on understanding the fault's past behavior.

That is precisely the problem paleoseismologists face: written and instrumental records of Nankai Trough earthquakes extend back only a few centuries, far too short a window to say with confidence how the fault behaves over the longer term. Scientists have tried to extend that record by reading disturbed layers, known as turbidites, in cores and by studying coastal evidence such as uplifted marine terraces and tsunami deposits. But those methods generally catch only the most forceful events, the ones that leave an obvious physical scar. The Kumano Basin has been a focus of this work for years, including through international scientific ocean drilling expeditions that have cored deep into the area's sediment and fault structure as part of long-running efforts to understand the tectonic structure of the Nankai margin. A technique that can flag disturbances invisible to the eye promises a longer, more complete catalog of past earthquakes, which feeds directly into the statistical models used to estimate recurrence intervals and, ultimately, public hazard forecasts.

A Tool, Not Yet a Verdict

The authors are careful to frame their results as evidence of "possible" earthquake traces rather than a confirmed new chapter of seismic history. Correlating a disturbed sediment layer with a specific historical earthquake is an inference, not a direct measurement, and the researchers note that storms and flooding can also redistribute old organic material into younger sediment, so each event layer has to be weighed individually rather than assumed to be seismic in origin. The work grew out of a presentation at the 25th International Radiocarbon Conference, held in Kraków, Poland, in 2025, before being written up as a peer-reviewed paper.

Coastal communities along the Kii Peninsula, including the city of Shingu near where the core was collected, sit squarely within the zone expected to be hit hardest by a future Nankai Trough rupture, as do much larger population centers further along the coast, including Osaka and Nagoya. For the seismologists and disaster planners who build hazard models for that stretch of coastline, a finer-grained earthquake catalog, even one built from inference rather than certainty, narrows the uncertainty that currently surrounds how often the trough truly ruptures.

Yokoyama's team says the next step is to apply the same high-resolution dating approach to additional cores across a wider area of the Kumano Sea, and eventually to other seismically active coastlines, to see whether the hidden layers they found represent a general feature of subduction-zone sediment that earlier studies have simply been too coarse to detect. If the pattern holds across a broader sample, it would suggest that existing sediment archives around the world's megathrust faults may hold more earthquake history than anyone has yet managed to read out of them.