ABSTRACT The Segment Anything Model 2 (SAM 2), a general-purpose segmentation model based on deep learning that does not require prior training by individual users, was applied to analyze video images that captured the tsunami caused by the 2024 Noto Peninsula earthquake. Subsequently, a time series of water-surface elevations was estimated from the segmentation results obtained with SAM 2 for the analyzed case. A comparative analysis was conducted between the results obtained using SAM 2 and those from existing Canny edge detection methods. The values measured by humans were used as the reference standard. For the examined tsunami event, SAM 2 exhibited accuracy and detection performance comparable to or higher than those obtained using Canny edge detection with optimized parameters, while maintaining nearly identical processing speed. Furthermore, whereas edge detection necessitates a process of trial and error to optimize the parameters, SAM 2 does not require such adjustments. The results demonstrate that SAM 2 can robustly extract time series of water-surface elevations during the 2024 Noto Peninsula earthquake tsunami and may suggest its potential for application to real-time tsunami monitoring.

ABSTRACT To address the need for background temperature calculation in studies on the environmental impacts of thermal discharge from coastal power plants and the limitations of traditional methods in characterizing regional temperature distribution, this study proposes a novel method for calculating background temperature, namely the Offshore Thermal Gradient Exponential Fitting Method (OTGEM). Based on 47 remote sensing images (2015–2024), OTGEM establishes 30 m-interval SST gradient strips to fit offshore water temperature natural gradient curves, linking with in-situ SST via characteristic parameters and reference points for precise calculation. Applied to Fujian Hongshan Thermal Power Plant and Ningde Nuclear Power Plant’s distinct coastlines, results show obvious offshore temperature gradients (inner-outer differences ∼ 0.9°C and 1.5°C, respectively) and tidal-phase-related variations. Error analysis indicated maximum average error ≤ 1°C within 4 km offshore: errors > 0.5°C are confined to 0.24–0.5 km nearshore sea area, ≤ 0.4°C in 0.5–1 km, and <0.2°C beyond 1 km. In conclusion, OTGEM dynamically reflects flood and ebb tides’ real-time impact, delivering regional characteristics, highly accurate results with low data demands and simple operation, suitable for diverse coastal scenarios like open coast areas and semi-enclosed bays.

ABSTRACT The mean wave overtopping rate is an essential parameter to design coastal protections. Estimating it with a high precision is primordial to find a balance between a satisfactory safety level and a limited impact on the environment and construction costs. A series of laboratory experiments was conducted in a wave flume to estimate the wave overtopping discharge over a rock-armored breakwater in bimodal sea state conditions (combining swell and wind waves). Both simulated swell and wind wave systems were long-crested and colinear. Preliminary tests were performed on a smooth breakwater to validate the experimental set-up. Some trends in the results with the smooth slope can be characterized by the representative wave steepness. These trends are confirmed and amplified in the presence of the armor rubble slope. In that case, the measured wave overtopping rate can be significantly overestimated by existing prediction formulas, especially for sea state conditions with a high representative wave steepness, corresponding to a high wind-wave proportion in the sea state energy. We suggest two methods to take into account the effect of sea-state bimodality via the representative wave steepness to improve the wave overtopping rate estimations for smooth and armored rubble breakwaters.

ABSTRACT The Noto Peninsula Earthquake that occurred on January 1, 2024, triggered tsunamis in the Sea of Japan, creating coastal hazards along the coastline of the Hokuriku region of Japan. On the day of the earthquake, the region was affected by winter storm waves, though the waves were not severe during the tsunami arrival. In this study, numerical experiments were conducted to assess the potential hazards of this tsunami occurring under storm wave conditions using the XBeach non-hydrostatic model. The model was validated against field data measured 150 km southwest of the epicenter. Model results indicate that the combination of the tsunami with wind waves increases the wave runup heights under rough wave conditions, while the contribution of the tsunami to shoreline excursions is strongly affected by the beach topography. The tsunami water level changes shift the swash zone along the concave beach profile, indirectly enhancing or reducing wind wave runup depending on the varying slope in the swash zone. Moreover, the presence of a high beach berm effectively constrains the wave runup, protecting the coastal zone from possible inundation. These findings highlight the importance of interactions among tsunamis, wind waves, and coastal morphology in hazard assessments and disaster mitigation planning.

ABSTRACT Accurate sea-level rise projections require accounting for vertical land motion, which can significantly amplify or offset relative sea-level trends, especially in tectonically active regions. The South American Pacific Coast exhibits strong spatial vertical land motion variability due to subduction processes and major earthquakes. We correct sea level rise projections under +4K global warming scenario using three datasets: satellite altimetry and tide gauge single difference from 1993 to 2020, InSAR-derived deformation 2020–2025, and Couple Model Intercomparison Project phase 6 vertical land motion outputs, to compare how point-based and spatially continuous vertical land motion estimates may differ/coincide and evaluate their impact on sea-level rise corrections. Results show different vertical land motion rates for the three datasets, where the highest values, obtained by differencing between satellite and tide gauge stations, vary from -1.64 mm/yr (subsidence in Manta, Ecuador) to +6.75 mm/yr (uplift in Talcahuano, Chile), comparable to the global mean sea-level rise. Including these corrections alters 2100 regional sea level rise projections by up to ±0.5 m, with results of the Coupled Model Intercomparison Project phase 6 often overestimating uplift compared to InSAR in spatial variability.

ABSTRACT Coastal regions face increasing disaster risk from tropical cyclones (TCs), intensified by climate change. Accurate and rapid prediction of peak surge height (PSH) is essential for coastal disaster prevention and sustainable development. However, numerical models are computationally intensive, while empirical regression and machine learning approaches require large datasets and are often site-specific, limiting their applicability for real-time prediction. This study, considering a wide range of tropical cyclone scenarios under idealized coastal bathymetry, aims to address limitations by developing a statistical formula to predict PSH in nearshore areas. A total of 1080 numerical experiments were conducted using a dynamic storm surge model at a simplified bathymetry under TC parameters of central pressure deficit, maximum wind radius, and translation speed, as well as additional parameters of incidence angle, distance from the typhoon center, and initial water depth. Based on these results, a PSH prediction formula was derived using log-transformed linear regression analysis for sites below 23 m water depth. The study showed that the proposed formula provided reliable PSH predictions with R 2 of 0.91. The formula demonstrates potential for computational efficiency and accuracy within the confines of the idealized conditions, providing a foundational approach to address the increasing risks of coastal disasters.

ABSTRACT Ray tracing of tsunamis generated at the Sea of Japan (SOJ) was performed to investigate far-field tsunami propagation in a semi-enclosed sea with complicated bottom features. This study focused on analyzing the propagation of the tsunami generated by the 2024 Noto Peninsula Earthquake that emerged in the southern SOJ, which differed from previously recorded large tsunami events that originated in the northern part of the SOJ. The tsunami propagation path and traveling time were estimated by predicting the wave refraction according to Snell’s Law. The numerical runs were made under the setting of four major tsunami events that occurred in the SOJ, and the calculated arrival time of the initial wave at the coast agreed well with the observed values. The trajectories of tsunami waves in the SOJ depend greatly on the meridional position of the wave source. In the case of the Noto Earthquake, the tsunami generated at the western and eastern parts of the rupture zone behaved differently, mainly due to the presence of the deep Toyama Submarine Canyon situated on the eastern side of the tsunami source. The findings of this study may aid in interpreting the observed or simulated results regarding tsunami propagation in the SOJ.

ABSTRACT Field surveys of tsunami heights caused by the January 1, 2024, Noto Peninsula earthquake (M 7.6, Japan Meteorological Agency) were conducted along approximately 700 km of coast from Yamagata Prefecture to Shimane Prefecture and on Sado Island, Japan. The region around the source area, which was severely damaged, was excluded. We measured tsunami heights of about 2–3 m at sites that were relatively close to the wave source area. The tsunami heights tended to be smaller or attenuated to an indistinguishable level at measurement points further away from the source area; however, a 1.4-m tsunami, which was higher than the tsunami heights in the surrounding areas, was found to have hit the port of Hyogo Prefecture, approximately 300 km from the wave source. To determine the overall magnitude of this tsunami, we estimated the tsunami magnitude m to be 1.5 from our measurement data and other published data. According to the relationship between the earthquake magnitude M and tsunami magnitude m of past earthquakes on the eastern margin of the Japan Sea, the m value of the 2024 tsunami was smaller than the value that would be expected from the earthquake magnitude.

ABSTRACT Understanding the spatial distribution of wave overtopping behind coastal protection structures is critical for an accurate safety assessment of the protected regions. This paper presents a systematic review of the current state of research in this field, summarizing the key mechanisms and predictive methodologies. While physical modeling has traditionally been the primary method for investigating spatial overtopping, its outcomes are often simplified into empirical formulas. However, their predictive accuracy of these formulas is compromised when applied to diverse structural geometries, a limitation stemming from their calibration against a narrow range of configurations. To address this limitation, this study proposes a novel semi-analytical model that, by refining the calculation of overtopping velocity at the structure’s crest, achieves a more accurate prediction of the spatial distribution of overtopping on coastal structures with diverse slopes. The model results show good agreement with numerical simulations and multiple sets of experimental data. Additionally, researchers still face challenges in addressing uncertainties from scale effects, quantifying wind effects in laboratory settings and translating them into practical guidance, predicting overtopping distribution under extreme climatic conditions, and assessing the effects of oblique waves and 3D effects on the spatial distribution of overtopping.

ABSTRACT This study proposes a new model using Light Gradient Boosting Machine (LightGBM) to predict breaking wave height based on input wave parameters. To determine optimal hyperparameter, Optuna is employed and conducts 100 independent runs with 10-fold cross-validation. Additionally, SHapley Additive exPlanations (SHAP) analysis is applied to investigate behavior of model. Results show that LightGBM model optimized with Optuna shows excellent performance for estimating breaker height. Root mean square error of model is 1.861 cm (for training dataset) and 3.518 cm (for testing dataset). Coefficients of determination are also high with 0.998 and 0.992 for training and testing datasets, respectively. This accuracy is remarkably higher than previous existing breaking wave height models. Besides, SHAP analysis highlights that deep-water wave height and water depth have the greatest impact on breaker height prediction. The results demonstrate that combination of Optuna and LightGBM enhances robustness and generalization of model for predicting breaking wave height.