The 2025 Kamchatka tsunami kept the Pacific ringing for more than ten hours. Its rhythm could help tsunami warnings

On July 29, 2025, at 23:24 Universal Time, a long stretch of seafloor off Russia's Kamchatka Peninsula was thrust upward. The magnitude 8.8 earthquake was the largest on Earth since the Tohoku earthquake of 2011, according to the U.S. Geological Survey, and within minutes a tsunami was spreading across the Pacific. Warnings went out from Japan to Chile. On Shumshu, in the Kuril Islands, the water splashed up to about 19 meters (62 feet) high.

Tsunamis are where my own research began, with a doctoral thesis on their mathematics, and since 2018 I have studied the tsunami hazard of real coastlines with Amin Rashidi of the University of Tehran, who led this study. When this tsunami crossed the ocean, we asked a simple question: How much of what the ocean recorded can the earthquake alone explain? With Mohammad Mokhtari, we give the answer in the journal Natural Hazards.

Six buoys in the deep ocean

Our judges were six stations of the DART network, which the U.S. National Oceanic and Atmospheric Administration keeps in the deep Pacific to detect tsunamis. Each station has a pressure sensor on the seafloor that feels the weight of the water above it and, therefore, the height of the sea surface. A buoy at the surface relays the readings by satellite. When a tsunami passes, the station switches from one reading every 15 minutes to one every minute or faster.

The station nearest the source, whose sensor sits 5.8 kilometers (3.6 miles) down, saw the first waves 27 minutes after the earthquake and then a crest of 85 centimeters (33 inches), the second-largest reading in the history of the network after the Tohoku tsunami of 2011. The farthest station, southeast of Japan, saw the waves arrive after 144 minutes, only 8 centimeters (3 inches) high.

Starting from the earthquake alone

Nothing in our simulation was adjusted to the buoy records. We started from the rupture model that the U.S. Geological Survey built from seismic waves recorded around the world: a fault surface about 690 kilometers (430 miles) long, cut into 345 patches, each with its own slip, which exceeds 10 meters (33 feet) in places. From it we computed how the seafloor rose and sank, and we let this motion unfold over the roughly 200 seconds of the rupture instead of all at once. How to turn such rupture models into a seafloor that moves in time is a question I have worked on since my thesis years.

The equations of long waves then carried the tsunami across the ocean. In the deep Pacific, a tsunami is hundreds of kilometers long while the ocean is only about 4 kilometers (2.5 miles) deep, so the whole column of water moves together, and the waves run at about 700 kilometers per hour (435 miles per hour), almost as fast as a jet airliner.

The agreement is close. At the six buoys, the simulated waves arrived on average within about a minute of the real ones, for arrival times between 27 and 144 minutes. The highest waves, from 7 to 85 centimeters (3 to 33 inches), were matched within 6 centimeters (2.4 inches) on average.

Along the coasts, from Australia to Chile, the simulation agreed with 21 reported wave heights within about 16% on average. At Baikovo on Shumshu, it gave a flow 6.2 meters (20 feet) deep reaching about 600 meters (2,000 feet) inland, where 6 meters (20 feet) and 650 meters (2,100 feet) were reported. No underwater landslide or second fault is needed to explain what the buoys saw: The earthquake alone accounts for it.

The ocean kept ringing

The buoy records hold a second story. For the first hour, seismic waves from the earthquake shook the sensors, and the records jitter at periods of a few minutes. Then the tsunami arrives, and its energy settles into a band of long periods, between about 30 and 80 minutes. It stays there. More than 10 hours after the first waves, the sea at every buoy was still oscillating in this band, far above its quiet level. The dominant period differed from one buoy to the next, most likely because each station hears the tsunami after its own journey over ridges, trenches and islands.

Physicists would say that the Pacific behaved like a bell with little damping: Struck once, it rang for a long time. For me, this ringing is the most beautiful part of the study because it shows how long the ocean keeps the memory of a single earthquake.

Why a rhythm matters

The rhythm is also useful. Bays and harbors have natural periods of their own, set by their size and depth, much like the pitch of a glass. When the incoming waves share that period, the water inside can build up far beyond the height of the waves outside, and this is resonance. The buoys register the rhythm of a tsunami hours before it reaches distant shores. Comparing that rhythm with the known natural periods of the harbors along a coast could help warning centers foresee where the waves will grow, and our study suggests that the period content of a tsunami deserves a place in warnings, next to arrival times and heights.

Seventy-three years after the magnitude 9.0 earthquake of 1952 in the same region, Kamchatka has reminded us that it is one of the most efficient tsunami sources in the Pacific. What I take from this study is encouraging: A careful model, fed only with what seismologists learned from the earthquake itself, can follow a tsunami across an ocean within about a minute. The next step is to test the same approach, and the same attention to the rhythm of the waves, on other great tsunamis.

This story is part of Science X Dialog, where researchers can report findings from their published research articles. Visit this page for information about Science X Dialog and how to participate.

Denys Dutykh is an applied mathematician and Associate Professor of Mathematics at Khalifa University of Science and Technology in Abu Dhabi, United Arab Emirates, where he is also Acting Associate Dean of Graduate Studies of the College of Computing and Mathematical Sciences. Tsunamis are where his research began: his doctoral thesis, defended in 2007 at the École Normale Supérieure de Cachan under Frédéric Dias, was devoted to the mathematical modeling of tsunami waves. He then spent 14 years as a researcher of the French National Center for Scientific Research (CNRS) at the University Savoie Mont Blanc, before joining Khalifa University in 2022. His research develops numerical methods for waves, from tsunamis in the ocean and in lakes to the vibrations of black holes. Since 2018, he has studied the tsunami hazard of real coastlines with Amin Rashidi, from the Makran coast of Iran and Pakistan and the Caribbean to Kamchatka.