An AI system estimated a tsunami source from real observations within minutes

Tests using data from the 2024 Hyuga-nada earthquake suggest the approach could support faster tsunami forecasting.

Editorial illustration — not from the study.

The system uses a Fourier Neural Operator, an AI method trained to approximate how tsunami waves move under standard long-wave equations. It can predict future wave fields much faster than repeatedly running conventional tsunami simulations, and it calculates the information needed for source estimation without separately writing the usual adjoint equations.

The researchers used the system to estimate both the initial sea-surface height and earthquake fault parameters. They tested the methods with synthetic data and tsunami observations recorded by Japan’s N-net system after the magnitude 7.1 Hyuga-nada earthquake in 2024.

What the system estimated

The AI-based model accurately approximated tsunami propagation and enabled rapid source inversion, meaning it worked backward from wave observations to estimate the tsunami’s starting conditions. The researchers report that source estimates could be obtained within a few minutes in tests with both synthetic data and real observations from the 2024 Hyuga-nada earthquake. The approach estimated initial sea-surface height and, when combined with a model of earthquake-related crustal movement, fault parameters.

Why faster estimates matter

Tsunami forecasting depends on quickly determining the source of newly observed waves. Conventional backward-estimation methods require many forward and reverse simulations, which can be computationally expensive. By reducing the time needed for these calculations, the proposed approach could make source estimation more practical for future real-time tsunami forecasting systems.

Evidence and caveats

The study was validated with synthetic data and with real tsunami observations from the 2024 Hyuga-nada earthquake recorded by Japan’s N-net system. The abstract does not provide numerical accuracy measures, and it describes validation for this event rather than demonstrating performance across a broad range of earthquakes and tsunami settings. The model learns propagation governed by linear long-wave equations, so its performance in situations that require more complex wave physics is not established by the abstract.

A proposed experiment could make gravity appear to push instead of pull

A theoretical study proposes an experiment in which a mass placed in a quantum superposition could produce an effectively repulsive gravitational effect on another particle. The effect would arise from quantum interference and, if observed, would conflict with general relativity’s description of gravity as always attractive.

1 day ago·29 views

Physics & Space

Silicon quantum bits keep precise control when run together

Researchers report that silicon spin qubits can perform precise operations in parallel, overcoming a major source of error during simultaneous control. In a five-qubit array, gate fidelities stayed above 99.99% for up to three qubits and above 99.9% when all five operated together.

1 day ago·11 views

Physics & Space

Mercury’s magnetic shield may host recurring solar-wind outflows

A survey of MESSENGER measurements identified 116 candidate magnetic reconnection layers near Mercury’s magnetopause. The structures appear to recur under the planet’s typically low-Alfvén-Mach-number solar wind, but they cannot yet be confirmed as reconnection outflows.

6 days ago·51 views