High-resolution source inversion of 2024 Noto Peninsula earthquake tsunami with modeling error corrections.

On New Year's Day, January 1, 2024, a significant earthquake struck the Noto Peninsula, Japan, triggering a tsunami with a maximum height of exceeding 5 m along the coast. This rare and destructive event is associated with several active submarine faults in the area, which are inferred to have ruptured simultaneously. However, due to the lack of direct observation from the ocean, the precise nature of the initial disturbance remains uncertain. In this study, we utilized the latest high-resolution adjoint synthesis inversion technique to analyze the tsunami's initial state based on observed water level and velocity fluctuations. Furthermore, we propose a method to correct modeling errors in amplitude and phase, thereby reducing misfits in trace heights: the geometric mean [Formula: see text] improves from 1.16 to 0.99, and the geometric standard deviation κ decreases from 1.34 to 1.29. Our estimated tsunami source model identified three distinct uplift peaks, each approximately 3 m in height, in the offshore region. The model accurately reproduced both offshore and onshore observation data. This research provides a robust framework for understanding tsunami generation mechanisms and improving hazard assessment models. By offering accurate initial conditions for tsunami simulations, our findings contribute to better disaster preparedness and mitigation strategies, particularly for coastal regions prone to similar events.