Explore the vast range of titles available.

The use of ships and boats as sea-state (SS) measurement platforms has the potential to expand ocean observations while providing actionable information for real-time operational decision-making at sea. Within the framework of the Wave Buoy Analogy (WBA), this work develops an inverse approach in which efficient simulations of wave-induced motions of an advancing vessel are used to train a neural network (NN) to predict SS parameters across a broad range of wave climates. We show that a reduced set of novel motion discriminant variables (MDVs)—computed from short time series of heave, roll, and pitch motions measured by an onboard inertial measurement unit (IMU), together with the vessel’s forward speed—provides sufficient and robust information for accurate, near-real-time SS estimation. The methodology targets small, barge-like tugboats whose operations are SS-limited and whose motions can become large and strongly nonlinear near their upper operating limits. To accurately model such responses and generate training data, an efficient nonlinear time-domain seakeeping model is developed that includes nonlinear hydrostatic and viscous damping terms and explicitly accounts for forward-speed effects. The model is experimentally validated using a scaled physical model in laboratory wave-tank tests, demonstrating the necessity of these nonlinear contributions for this class of vessels. The validated model is then used to generate large, high-fidelity datasets for NN training. When applied to independent numerically simulated motion time series, the trained NN predicts SS parameters with errors typically below 5%, with slightly larger errors for SS directionality under relatively high measurement noise. Application to experimentally measured vessel motions yields similarly small errors, confirming the robustness and practical applicability of the proposed framework. In operational settings, the trained NN can be deployed onboard a tugboat and driven by IMU measurements to provide real-time SS estimates. While results are presented for a specific vessel, the methodology is general and readily transferable to other ship geometries given appropriate hydrodynamic coefficients.

On Dec. 22, 2018, at approximately 20:55–57 local time, Anak Krakatau volcano, located in the Sunda Straits of Indonesia, experienced a major lateral collapse during a period of eruptive activity that began in June. The collapse discharged volcaniclastic material into the 250 m deep caldera southwest of the volcano, which generated a tsunami with runups of up to 13 m on the adjacent coasts of Sumatra and Java. The tsunami caused at least 437 fatalities, the greatest number from a volcanically-induced tsunami since the catastrophic explosive eruption of Krakatau in 1883 and the sector collapse of Ritter Island in 1888. For the first time in over 100 years, the 2018 Anak Krakatau event provides an opportunity to study a major volcanically-generated tsunami that caused widespread loss of life and significant damage. Here, we present numerical simulations of the tsunami, with state-of the-art numerical models, based on a combined landslide-source and bathymetric dataset. We constrain the geometry and magnitude of the landslide source through analyses of pre- and post-event satellite images and aerial photography, which demonstrate that the primary landslide scar bisected the Anak Krakatau volcano, cutting behind the central vent and removing 50% of its subaerial extent. Estimated submarine collapse geometries result in a primary landslide volume range of 0.22–0.30 km 3 , which is used to initialize a tsunami generation and propagation model with two different landslide rheologies (granular and fluid). Observations of a single tsunami, with no subsequent waves, are consistent with our interpretation of landslide failure in a rapid, single phase of movement rather than a more piecemeal process, generating a tsunami which reached nearby coastlines within ~30 minutes. Both modelled rheologies successfully reproduce observed tsunami characteristics from post-event field survey results, tide gauge records, and eyewitness reports, suggesting our estimated landslide volume range is appropriate. This event highlights the significant hazard posed by relatively small-scale lateral volcanic collapses, which can occur en-masse , without any precursory signals, and are an efficient and unpredictable tsunami source. Our successful simulations demonstrate that current numerical models can accurately forecast tsunami hazards from these events. In cases such as Anak Krakatau’s, the absence of precursory warning signals together with the short travel time following tsunami initiation present a major challenge for mitigating tsunami coastal impact.

The authors recently proposed a new method for detecting tsunamis using high-frequency (HF) radar observations, referred to as “time-correlation algorithm” (TCA; Grilli et al. Pure Appl Geophys 173(12):3895–3934, 2016a, 174(1): 3003–3028, 2017). Unlike standard algorithms that detect surface current patterns, the TCA is based on analyzing space-time correlations of radar signal time series in pairs of radar cells, which does not require inverting radial surface currents. This was done by calculating a contrast function, which quantifies the change in pattern of the mean correlation between pairs of neighboring cells upon tsunami arrival, with respect to a reference correlation computed in the recent past. In earlier work, the TCA was successfully validated based on realistic numerical simulations of both the radar signal and tsunami wave trains. Here, this algorithm is adapted to apply to actual data from a HF radar installed in Tofino, BC, for three test cases: (1) a simulated far-field tsunami generated in the Semidi Subduction Zone in the Aleutian Arc; (2) a simulated near-field tsunami from a submarine mass failure on the continental slope off of Tofino; and (3) an event believed to be a meteotsunami, which occurred on October 14th, 2016, off of the Pacific West Coast and was measured by the radar. In the first two cases, the synthetic tsunami signal is superimposed onto the radar signal by way of a current memory term; in the third case, the tsunami signature is present within the radar data. In light of these test cases, we develop a detection methodology based on the TCA, using a correlation contrast function, and show that in all three cases the algorithm is able to trigger a timely early warning.

Tsunami coastal hazard is modeled along the US East Coast (USEC), at a coarse regional (450 m) resolution, from coseismic sources located in the Açores Convergence Zone (ACZ) and the Puerto Rico Trench (PRT)/Caribbean Arc areas. While earlier work only considered probable maximum tsunamis, here we parameterize and simulate 18 coseismic sources, with magnitude M8-9 and return periods ∼70–2000 year, using seismo-tectonic and historical data. The largest sources in the ACZ are repeats of the 1755 M8.6-9 Lisbon earthquake and tsunami; other sources are hypothetical. In the ACZ, due to the limited data on faults, each source is parameterized with a single fault plane, while in the PRT, coseismic sources are parameterized based on fault segmentation established during a 2019 USGS workshop of experts, using 10–26 SIFT subfault planes (Gica et al. in NOAA Tech. Memo., OAR PMEL-139, 2008). Tsunamis are simulated for each source using the fully nonlinear and dispersive model FUNWAVE-TVD, in two levels of nested grids. At the considered scales, dispersion is shown to affect tsunami propagation. Coastal hazard is quantified by four metrics computed at many save points (∼20–30 thousand) defined along the 5-m isobath (due to the coarse resolution), i.e., maximum (1) surface elevation, (2) current, (3) momentum force; and (4) travel time, representing flooding, navigation, structural, and evacuation hazards. Overall, the first three metrics are larger, the larger the source magnitude, and their alongshore variation shows similar patterns of higher/lower values, due to the shelf bathymetric control (refraction). The fourth metric mostly differs between sources from each area, but less so among sources from the same area; its inverse quantifies evacuation hazard. A 1–5 score is given to results for each metric, based on five intensity classes representing low, medium low, medium, high, and highest tsunami hazard. A novel tsunami intensity index is computed as a weighted average of these scores, allowing both a comparison among sources and a quantification of tsunami hazard as a function of their estimated return periods. In the most impacted areas of the USEC, the highest tsunami hazard in the 250–500-year return period range is commensurate with that posed by 100-year category 3–5 tropical cyclones, taking into account the larger current velocities and forces caused by tsunami waves. Results of this work could serve as a basis for a future regional Probabilistic Tsunami Hazard Analysis for the USEC, considering additional source types such as underwater landslides, volcanic flank collapse, and meteotsunamis, that were studied elsewhere.

This work is part of a tsunami inundation mapping activity carried out along the US East Coast since 2010, under the auspice of the National Tsunami Hazard Mitigation program (NTHMP). The US East Coast features two main estuaries with significant tidal forcing, which are bordered by numerous critical facilities (power plants, major harbors,...) as well as densely built low-level areas: Chesapeake Bay and the Hudson River Estuary (HRE). HRE is the object of this work, with specific focus on assessing tsunami hazard in Manhattan, the Hudson and East River areas. In the NTHMP work, inundation maps are computed as envelopes of maximum surface elevation along the coast and inland, by simulating the impact of selected probable maximum tsunamis (PMT) in the Atlantic ocean margin and basin. At present, such simulations assume a static reference level near shore equal to the local mean high water (MHW) level. Here, instead we simulate maximum inundation in the HRE resulting from dynamic interactions between the incident PMTs and a tide, which is calibrated to achieve MHW at its maximum level. To identify conditions leading to maximum tsunami inundation, each PMT is simulated for four different phases of the tide and results are compared to those obtained for a static reference level. We first separately simulate the tide and the three PMTs that were found to be most significant for the HRE. These are caused by: (1) a flank collapse of the Cumbre Vieja Volcano (CVV) in the Canary Islands (with a 80 km 3 volume representing the most likely extreme scenario); (2) an M9 coseismic source in the Puerto Rico Trench (PRT); and (3) a large submarine mass failure (SMF) in the Hudson River canyon of parameters similar to the 165 km 3 historical Currituck slide, which is used as a local proxy for the maximum possible SMF. Simulations are performed with the nonlinear and dispersive long wave model FUNWAVE-TVD, in a series of nested grids of increasing resolution towards the coast, by one-way coupling. Four levels of nested grids are used, from a 1 arc-min spherical coordinate grid in the deep ocean down to a 39-m Cartesian grid in the HRE. Bottom friction coefficients in the finer grids are calibrated for the tide to achieve the local spatially averaged MHW level at high tide in the HRE. Combined tsunami–tide simulations are then performed for four phases of the tide corresponding to each tsunami arriving at Sandy Hook (NJ): 1.5 h ahead, concurrent with, 1.5 h after, and 3 h after the local high tide. These simulations are forced along the offshore boundary of the third-level grid by linearly superposing time series of surface elevation and horizontal currents of the calibrated tide and each tsunami wave train; this is done in deep enough water for a linear superposition to be accurate. Combined tsunami–tide simulations are then performed with FUNWAVE-TVD in this and the finest nested grids. Results show that, for the 3 PMTs, depending on the tide phase, the dynamic simulations lead to no or to a slightly increased inundation in the HRE (by up to 0.15 m depending on location), and to larger currents than for the simulations over a static level; the CRT SMF proxy tsunami is the PMT leading to maximum inundation in the HRE. For all tide phases, nonlinear interactions between tide and tsunami currents modify the elevation, current, and celerity of tsunami wave trains, mostly in the shallower water areas of the HRE where bottom friction dominates, as compared to a linear superposition of wave elevations and currents. We note that, while dynamic simulations predict a slight increase in inundation, this increase may be on the same order as, or even less than sources of uncertainty in the modeling of tsunami sources, such as their initial water elevation, and in bottom friction and bathymetry used in tsunami grids. Nevertheless, results in this paper provide insight into the magnitude and spatial variability of tsunami propagation and impact in the complex inland waterways surrounding New York City, and of their modification by dynamic tidal effects. We conclude that changes in inundation resulting from the inclusion of a dynamic tide in the specific case of the HRE, although of scientific interest, are not significant for tsunami hazard assessment and that the standard approach of specifying a static reference level equal to MHW is conservative. However, in other estuaries with similarly complex bathymetry/topography and stronger tidal currents, a simplified static approach might not be appropriate.

Sandy barrier systems are highly dynamic, with the most significant natural morphological changes to these systems occurring during high-energy storm conditions. These systems provide a range of economic and ecosystem benefits and protect inland areas from flooding and storm impacts, but the persistence of many coastal barriers is threatened by storms and sea-level rise (SLR). This study employed observations and modeling to examine recent and potential future influences of storms on a sandy coastal barrier system in Nauset Beach, MA. Drone-derived imagery and digital elevation models (DEMs) of the study area collected throughout the 2023–2024 winter revealed significant alongshore variability in the geomorphic response to storms. Severe, highly localized erosion (i.e., an erosional “hotspot”) occurred immediately south of the Nauset Bay spit as the result of a group of storms in December and January. Modeling results demonstrated that the location of the hotspot was largely controlled by the location of a break in a nearshore sandbar system, which induced larger waves and stronger currents that affected the foreshore, backshore and dune. Additionally, model simulations of the December and January storms assuming 0.3 m (1 ft) of SLR showed the system to be relatively resistant to major geomorphic changes in response to an isolated storm event, but more susceptible to significant overwash and breaching in response to consecutive storms. This research suggests that both very strong isolated storm events and sequential moderate storms pose an enhanced risk of major overwash, breaching, and possibly inlet formation today and into the future, raising concern for adjacent communities and resource managers.

In their pioneering work, Ward and Day suggested that a large scale flank collapse of the Cumbre Vieja Volcano (CVV) on La Palma (Canary Islands) could trigger a mega-tsunami throughout the North Atlantic Ocean basin, causing major coastal impact in the far-field. While more recent studies indicate that near-field waves from such a collapse would be more moderate than originally predicted by Ward and Day [Løvholt et al. (J Geophy Res 113:C09026, 2008 ); Abadie et al. (J Geophy Res 117:C05030, 2012 )], these would still be formidable and devastate the Canary Island, while causing major impact in the far-field at many locations along the western European, African, and the US east coasts. Abadie et al. (J Geophy Res 117:C05030, 2012 ) simulated tsunami generation and near-field tsunami impact from a few CVV subaerial slide scenarios, with volumes ranging from 20 to 450 km 3 ; the latter representing the most extreme scenario proposed by Ward and Day. They modeled tsunami generation, i.e., the tsunami source, using THETIS, a 3D Navier-Stokes (NS) multi-fluid VOF model, in which slide material was considered as a nearly inviscid heavy fluid. Near-field tsunami impact was then simulated for each source using FUNWAVE-TVD, a dispersive and fully nonlinear long wave Boussinesq model [ Shi et al. (Ocean Modell 43–44:36–51, 2012 ); Kirby et al. (Ocean Modeling, 62:39–55, 2013 )]. Here, using FUNWAVE-TVD for a series of nested grids of increasingly fine resolution, we model and analyze far-field tsunami impact from two of Abadie et al. ’s extreme CVV flank collapse scenarios: (i) that deemed the most “credible worst case scenario” based on a slope stability analysis, with a 80 km 3 volume; and (ii) the most extreme scenario, similar to Ward and Day’s, with a 450 km 3 volume. Simulations are performed using a one-way coupling scheme in between two given levels of nested grids. Based on the simulation results, the overall tsunami impact is first assessed in terms of maximum surface elevation computed along the western European and African, and US east coasts (USEC). Strong wave elevation decay is predicted over the wide USEC shelf, which is shown to be essentially due to bottom friction effects. We then show more detailed results for the USEC, which is the object of high-resolution tsunami inundation mapping under the auspices of the US National Tsunami Hazard Mitigation Program. In this context, we compare the maximum surface elevation predicted along the coastline for each CVV scenario and show that, besides the initial directionality of the sources, coastal impact is mostly controlled by focusing/defocusing effects resulting from the shelf bathymetric features. A simplified ray-tracing analysis confirms this controlling effect of the wide USEC shelf for incident long waves. Finally, we perform high-resolution (10 m) inundation mapping for the most extreme CVV scenario and show results at one of the most vulnerable and exposed communities in the mid-Atlantic US states, in and around Ocean City, Maryland. Such maps are being generated for all exposed areas of the USEC, to be used in tsunami hazard assessment and mitigation work.

Since the devastating earthquake of 2010 in Haiti, significant efforts have been devoted to estimating future seismic and tsunami hazard in Hispaniola. In 2013, following a workshop of experts, UNESCO commissioned an initial modeling study to assess tsunami hazard, essentially from seismic sources, along the North shore of Hispaniola (NSOH), which is shared by the Republic of Haiti (RH) and the Dominican Republic (DR). The scope of this study included detailed tsunami inundation mapping for two selected critical sites, Cap Haitien in RH and Puerto Plata in DR. Results of this effort are reported here, and, although still limited in scope, they are within the framework and contribute to the advancement of the UNESCO IOC Tsunami and other Coastal Hazards Warning System for the Caribbean and Adjacent Regions (CARIBE EWS; von Hillebrandt-Andrade in Science 341:966–968, 2013 ). In similar work done for critical areas of the US east coast (under the auspice of the US National Tsunami Hazard Mitigation Program), the authors have modeled the most extreme far-field tsunami sources in the Atlantic Ocean basin, including: (1) a hypothetical M w 9 seismic event in the Puerto Rico Trench (PRT); (2) a repeat of the historical 1755 M w 9 earthquake in the Azores convergence zone (LSB); and (3) a hypothetical extreme 450 km 3 flank collapse of the Cumbre Vieja Volcano (CVV) in the Canary Archipelago. Here, tsunami hazard assessment is performed along the NSOH for these three sources, plus two additional near-field coseismic tsunami sources: (1) a M w 8 earthquake in the western segments of the nearshore Septentrional fault (SF), as a repeat of the 1842 event; and (2) a M w 8.7 earthquake occurring in selected segments of the North Hispaniola Thrust Fault (NHTF). Initial tsunami elevations are modeled based on each source’s parameters and propagated with FUNWAVE-TVD (a nonlinear and dispersive long-wave Boussinesq model) in a series of increasingly fine-resolution nested grids (from 1 arc-min to 205 m) using a one-way coupling methodology. For the two selected sites, coastal inundation is computed with TELEMAC (a Nonlinear Shallow Water wave model), in finer-resolution (12–30 m) unstructured nested grids. While for the EC, PRT is a far-field source, for RH and DR, this would be local source as some of the NSOH would be affected within 1 h or is within 200 km of the PRT. This is per definitions of UNESCO IOC. Regional goes from 200 to 1000 km and within 1 and 3 h, and distant is greater than 3 h and more than 1000 km. We find that among the far-field sources CVV causes the largest impact, with up to 20-m runup at the critical sites while PRT, which is a local source for the NSOH, only causes up to 4-m runup due to its directionality; PRT, however, has both a much shorter return period and would impact the NSOH within 30 min of the earthquake. Among near-field sources, the SF event, as could be expected from a strike-slip fault, only causes a small tsunami, but the NHTF event causes up to 12-m runup in the critical sites, with the tsunami arriving within minutes of the earthquake. Hence, the latter event can be considered as the “Probable Maximum Tsunami” (PMT; following, e.g., the US Nuclear Regulatory Commission terminology) for the NSOH. Results of detailed coastal modeling for this PMT can be used to develop maps of vulnerability for the critical sites and prepare for mitigating measures and evacuation; a few examples of such maps are given in the paper. Although a number of earlier studies have dealt with each of the far-field tsunami sources, the modeling of their impact on the NSOH and that of the near-field sources, presented here as part of a comprehensive tsunami hazard assessment study, are novel. Future work should model additional coastal sites and may consider effects of tsunamis generated by near-field submarine mass failures.

Within the perspective of siting wind farms offshore of Rhode Island, USA, the State and National Environmental Agencies had requested a local marine ecological assessment, which led to an ecological zoning of the area. In view of expanding this zoning outside its limit of the test area and filling gaps in ecological zones, an effort to model those ecological zones using marine landscape or abiotic features was carried out. This study tests the accuracy of selected machine learning algorithmic models, decision tree, and random forest, for relating marine landscapes features to ecological sub-regions. Both models show to be good predictive tools with accuracy after cross validation of the order of 5–3%. Key abiotic variables to provide an accurate model were investigated. The study demonstrates the importance of the distance to coast, the sediment characteristics (fraction of clay, median size of the sediments), the hydrodynamic features, in particular not only tidal current/drag force, but also wave drag force, and finally the oceanographic characteristics such as stratification and sea surface temperature to built a good predictive model. Those findings provide some insight on the pre-monitoring effort optimization.

In past work, tsunami detection algorithms (TDAs) have been proposed, and successfully applied to offline tsunami detection, based on analyzing tsunami currents inverted from high-frequency (HF) radar Doppler spectra. With this method, however, the detection of small and short-lived tsunami currents in the most distant radar ranges is challenging due to conflicting requirements on the Doppler spectra integration time and resolution. To circumvent this issue, in Part I of this work, we proposed an alternative TDA, referred to as time correlation (TC) TDA, that does not require inverting currents, but instead detects changes in patterns of correlations of radar signal time series measured in pairs of cells located along the main directions of tsunami propagation (predicted by geometric optics theory); such correlations can be maximized when one signal is time-shifted by the pre-computed long wave propagation time. We initially validated the TC-TDA based on numerical simulations of idealized tsunamis in a simplified geometry. Here, we further develop, extend, and apply the TC algorithm to more realistic tsunami case studies. These are performed in the area West of Vancouver Island, BC, where Ocean Networks Canada recently deployed a HF radar (in Tofino, BC), to detect tsunamis from far- and near-field sources, up to a 110 km range. Two case studies are considered, both simulated using long wave models (1) a far-field seismic, and (2) a near-field landslide, tsunami. Pending the availability of radar data, a radar signal simulator is parameterized for the Tofino HF radar characteristics, in particular its signal-to-noise ratio with range, and combined with the simulated tsunami currents to produce realistic time series of backscattered radar signal from a dense grid of cells. Numerical experiments show that the arrival of a tsunami causes a clear change in radar signal correlation patterns, even at the most distant ranges beyond the continental shelf, thus making an early tsunami detection possible with the TC-TDA. Based on these results, we discuss how the new algorithm could be combined with standard methods proposed earlier, based on a Doppler analysis, to develop a new tsunami detection system based on HF radar data, that could increase warning time. This will be the object of future work, which will be based on actual, rather than simulated, radar data.