Conventional tsunami warning systems for local and far‐field areas utilize uniform slip models to predict tsunami waves. The reasons are twofold. First, it is challenging to develop accurate finite‐fault slip models in a short time after an earthquake. Second, tsunami waves are long waves, and hence, the main feature may be predicted without knowing earthquake rupture details. Still, there have been few studies that quantitatively analyze the errors caused by uniform slip models. In this paper, we evaluate if and how such models may be applied for tsunami warnings. For the 2011 Tohoku, 2014 Iquique, and 2015 Illapel tsunamis, we first construct optimum uniform slip models with minimum tsunami waveform misfit and then compare the synthetic tsunami waves with finite‐fault model predictions. Predictions from both type of models match the tsunami data very well, indicating that the prediction errors caused by neglecting slip heterogeneity are insignificant. Further, we derive a common relation between the rupture area and earthquake magnitude. Additionally, the optimum rupture length and width ratio in terms of predicting tsunami waves is determined to be 1 for the three earthquakes. Lastly, we find that moving the uniform slip model to the center of global Centroid Moment Tensor solution produces reasonably small errors in the predicted waveforms. Applying the methodology to three more historic tsunamis shows that uniform slip models can well recover the Deep‐ocean Assessment and Reporting of Tsunamis system recordings, but the rupture center can differ from the global Centroid Moment Tensor solution. Our findings can potentially prompt more reliable tsunami warning strategies for future events.

The Herschel-Bulkley rheological theory is used to describe the viscoplastic debris landslide flow. The shallow water equations considering the time-dependent deformation of the seafloors are adopted to simulate the generation, propagation, and run-up of the landslide induced tsunami. The one-way coupled method of the landslide induced tsunami is implemented through satisfying the kinematic bottom boundary condition. The 1998 Papua New Guinea landslide tsunami is simulated to validate the numerical model by comparing with measurements. We found that the mechanism of the 1992 Hainan Island tsunami in the South China Sea is due to a submarine landslide by comparing the numerical results between earthquake and landslide. With respect of the Baiyun slide, the effects of remolding rate, initial, and residual yield strength on landslide and tsunami are studied numerically. To distinguish the potential landslide tsunami hazard in the South China Sea, the scenarios of the landslides with the volume of 10, 50, 100, and 200 km³ in the Baiyun slide and 1200 km³ in the Brunei slide are presented. Comparison with the nondeformation model in the near-field illustrates the crucial role of rheological property in the landslide tsunami modeling. Furthermore, the characteristics of the propagation of the landslide tsunami in the South China Sea and coastal hazards are analyzed.

Employing a linear shallow water equation (LSWE) model in the spherical coordinates, this paper investigates the tsunami waves generated by the atmospheric pressure shock waves due to the explosion of the submarine volcano Hunga Tonga–Hunga Ha'apai on 15 January 2022. Using the selected 59 atmospheric pressure records in the Pacific Ocean, an empirical atmospheric pressure model is first constructed. Applying the atmospheric pressure model and realistic bathymetric data in the LSWE model, tsunami generation and propagation are simulated in the Pacific Ocean. The numerical results show clearly the co‐existence of the leading locked waves, propagating with the speed of the atmospheric pressure waves (∼1,100 km/hr), and the trailing free waves, propagating with long gravity ocean wave celerity (∼750 km/hr). During the event, tsunamis were reported by 41 Deep‐ocean Assessment and Reporting of Tsunamis (DART) buoys in the Pacific Ocean, which require corrections because of the occurrence of atmospheric pressure waves. The numerically simulated tsunami arrival time and the amplitudes of the wave crest and trough of the leading locked waves compare reasonably well with the corrected DART measurements. The comparisons for the trailing waves are less satisfactory, since free waves could also have been generated by other tsunami generation mechanisms, which have not been considered in the present model, and by the scattering of locked waves over changing bathymetry. In this regard, the numerical results show clearly that the deep Tonga trench (∼10 km) amplifies the trailing waves in the Southeast part of the Pacific Ocean via the Proudman resonance condition.

On 15 January 2022, an exceptional eruption of the Hunga Tonga–Hunga Ha'apai volcano generated atmospheric and tsunami waves that were widely observed in the oceans globally, gaining remarkable attention from scientists in related fields. The tsunamigenic mechanism of this rare event remains enigmatic due to its complexity and lack of direct underwater observations. Here, to explore the tsunamigenic mechanisms of this volcanic tsunami event and its hydrodynamic processes in the Pacific Ocean, we conduct statistical analysis and spectral analysis of the tsunami recordings at 116 coastal gauges and 38 deep-ocean buoys across the Pacific Ocean. Combined with the constraints of some representative barometers, we obtain the plausible tsunamigenic origins of the volcano activity. We identify four distinct tsunami wave components generated by air–sea coupling and seafloor crustal deformation. Those tsunami components are differentiated by their different propagating speeds or period bands. The first-arriving tsunami component with an ∼ 80–100 min period was from shock waves spreading at a velocity of ∼ 1000 m s⁻¹ in the vicinity of the eruption. The second component with extraordinary tsunami amplitude in the deep ocean was from Lamb waves. The Lamb wave with a ∼ 30–40 min period radically propagated outward from the eruption site with spatially decreasing propagation velocities from ∼ 340 to ∼ 315m s⁻¹. The third component with a ∼ 10–30 min period was probably from some atmospheric-gravity-wave modes propagating faster than 200 m s⁻¹ but slower than Lamb waves. The last component with a ∼ 3–5 min period originated from partial caldera collapse with dimension of ∼ 0.8–1.8 km. Surprisingly, the 2022 Tonga volcanic tsunami produced long oscillation in the Pacific Ocean which is comparable with that of the 2011 Tohoku tsunami. We point out that the long oscillation is associated not only with the resonance effect with the atmospheric acoustic-gravity waves but more importantly with their interactions with local bathymetry. This rare event also calls for more attention to the tsunami hazards produced by an atypical tsunamigenic source, e.g. volcanic eruption.

Earthquakes occurring in the Ryukyu Trench, Nankai Trough, and Japan Trench can trigger tsunamis, affecting Hangzhou Bay, China. The 2011 Tohoku tsunami reached the east China coast and resulted in prolonged oscillations in Hangzhou Bay. We used this event as an example and studied the resonance behavior in Hangzhou Bay by spectral and modal analyses. We simulated the wavefield of the 2011 Tohoku tsunami and validated the modeling results using the tide gauge records. The spectral analysis based on the fast Fourier transform of the entire bay area indicated that the resonance oscillations with periods of 66 and 54 min had the highest energy amplification, and the mode of 120 min had a large amplitude near Zhenhai and Fengxian. Additionally, we conducted a modal analysis that relied solely on bathymetry data to analyze the eigenmodes of natural oscillations inside the bay. The regional bay-scale oscillations had periods of up to 286 min. The energy amplification area was generally consistent with the spectral analysis. These findings have implications for tsunami hazard assessments in coastal communities.