High-resolution hydrodynamic data are essential for coastal and estuarine management. However, traditional downscaling methods based on numerical modeling remain computationally expensive, limiting their applicability for long-term hindcasts and operational forecasting systems. This study evaluates the use of machine learning for the reconstruction of sea surface height and surface currents in a semi-enclosed estuary, using Santander Bay as a case study. Three techniques spanning increasing model complexity are analyzed: K-nearest neighbors, Adaptive Boosting, and long short-term memory networks. The models are trained to emulate high-resolution hydrodynamic-model outputs using a comprehensive set of tidal, meteorological, and fluvial forcings. Performance is assessed through spatial validation, cluster-based analysis, representative-point time series, and independent comparison against in-situ observations. Results show that all techniques successfully reproduce the main hydrodynamic patterns, with accuracy increasing with model complexity. Long short-term memory networks achieve the highest skill in tidally energetic regions, while Adaptive Boosting provides more stable performance in low-energy and shoreline areas. Computational cost analysis demonstrates that all machine-learning approaches achieve speedups of several orders of magnitude relative to numerical modelling, with inference costs that are negligible at both point and domain scales. These findings demonstrate the potential of machine learning as a computationally efficient approach for high-resolution modelling of coastal hydrodynamics, with important implications for operational forecasting and coastal management applications.

Following an international workshop on the topic of research needs in oil spill modeling, arranged by CRRC in Ann Arbor, MI, in September 2024, we present a review of what we find to be the most important and timely directions for oil spill trajectory and fate modeling research.We discuss potential advances in the areas of transport (e.g., advection, diffusion, entrainment), fate (e.g., emulsification, shoreline interaction, countermeasures), uncertainty and calibration (e.g., ensemble simulations, dynamical systems tools, comparison to drifter and historical spills), as well as standardization efforts for oil spill model input and output. We argue that research in these topics will improve the quality and usefulness of oil spill models, as well as give us a better understanding of uncertainty and other limitations.Research projects related to several topics are ongoing and we believe the others should be prioritized. Many of the topics can only be addressed through cross-disciplinary work with researchers outside the oil spill modeling community, ideally through international, collaborative efforts.

On December 8, 2023, the cargo ship Toconao was involved in a maritime incident off the coast of northern Portugal that resulted in the loss of 1,000 bags of buoyant plastic pellets, or nurdles. The pellets fell into the sea and reached the Bay of Biscay, posing a potential threat to the coasts of Spain and France. In the weeks following the incident, pellets were found washing up on beaches throughout northern Spain, particularly in Galicia, Asturias, Cantabria and the Basque Country. In response to this environmental emergency, several contingency plans were activated along the Spanish coast, requiring the use of scientific tools such as numerical modeling to develop effective responses. Nevertheless, a lack of detailed knowledge of the factors governing the transport and dispersion of these pellets made accurate predictions difficult, posing a significant challenge to effective management and mitigation of the spill.For this reason, a set of laboratory experiments was conducted to study the dispersion of pellets under a range of hydrodynamic and wind conditions in two different physical settings. The behavior of the pellets collected from the Toconao spill, which showed a density of 900 kg/m³, was assessed in two different sections of a hydraulic flume (Z1 and Z2 zones) under different combinations of water level, current, and wind. Zone Z1, located within the flume itself (2 m x 0.35 m), exhibits a unidirectional flow pattern. In contrast, zone Z2, located in the expanded section of the flume (3 m x 1.5 m), shows three-dimensional asymmetric flow patterns. Hydrodynamic conditions were defined by combining one water level in each zone (30 and 40 cm in Z1 and Z2, respectively) and three flow rates (30, 40 and 50 l/s in both zones). In addition, the effect of four wind conditions was tested for the average flow rate of 40 l/s (in Z1: no wind, 0.5, 1.0, and 1.5 m/s; in Z2: no wind, 0.3, 0.7, and 1.1 m/s). The results provide valuable information on the effects of wind and ocean currents on the dispersion of a group of plastic pellets and demonstrate the importance of surface currents in this process. In addition, these findings provide a comprehensive database for validating numerical transport models, which will improve their predictive ability and usefulness in future emergencies.

Dredging and dumping in-situ sediments is a fundamental operation for most coastal engineering projects and coastal defense projects, such as the construction of breakwaters, beach nourishment and land reclamation. Future projections in terms of coastal hazard suggest that coastal protection and land reclamation project will be more and more frequent. In this context, the assessment of the environmental and socio-economic impact of the risk induced by dredging is a fundamental step during both the design stage and the operational management. Most of the standard practices and available risk assessment frameworks rely on the numerical prediction of the sediment plume in the large field driven by coastal circulations forced by tides, winds and waves. In this study, we formulated a new risk assessment framework based on an unsupervised machine learning clustering algorithm, K-means clustering, for generating representative meteocean scenarios subsequently used to force a regional circulation model. Moreover, we introduced three criteria of hazard/risk based on the spatial and temporal evolution of the suspended sediment concentration that explained different environmental impacts and two new methods to synthetically present the risk values. The major improvement of the present framework is that the final probability of risk fully describes the statistics in terms of hydrodynamic and dredging conditions.This framework presents the probability analysis of risk spatial distribution based on representative hydrodynamic conditions and dredging scenarios, which is a major improvement of this study compared with previous risk assessment strategies that were unable to predict quantified dredging risk before field construction. Finally, to demonstrate the potentiality of the risk assessment framework, we applied this methodology to the Hong Kong Water and Pearl River Estuary (China) as a pilot case.

This paper presents a novel method to select the optimal combination of grid resolution and number of Lagrangian elements (LEs) required in numerical modelling of oil concentrations at sea. A sensitivity analysis in terms of grid resolution and the number of LEs, was carried out to understand the uncertainty that these user-dependent parameters introduce in the numerical results. A dataset of 211,200 simulations performed under 400 metocean patterns, 6 initial volumes, 11 grid resolutions, and different numbers of LEs (100 to 500,000), was used to analyze the sensitivity of the model along different Thresholds of Concern. Results show the importance of a correct selection of the number of LEs and the grid resolution in Lagrangian modelling of surface oil concentrations. The method proposed will allow selecting the optimal combination of these parameters to find an optimal balance between the accuracy and the computational cost of the simulation.

<p>Plastic debris is currently a significant threat to marine and coastal ecosystems. Most previous research focused on the behavior of drifting macro and mesoplastics on global and regional ocean scales. Furthermore, a few more recent studies provide some first insights into the microplastic dispersion in coastal areas. These studies found that waves and wind, as well as the density, size, and shape of microplastics, drive their transport and dispersion in coastal areas; however, they point to the need for a more extensive characterization. This laboratory study assesses the effect of waves and wave-induced currents on the input rate from land to sea and on the cross-shore transport and dispersion of different types of plastic debris, including the macro and mesosizes, in addition to microplastics. A total of 15 types of plastic debris characterized by different sizes, shapes, and densities, including face masks, were analyzed under regular and irregular wave conditions. The results show that the input rates and transport of plastic debris in the marine environment depend on the position they acquire in the water column, which is related to the terminal velocities and the wave steepness. A higher input rate from the beach was found for plastic materials moving closer to the sea bottom and under less steep wave conditions, as these conditions allow items to escape from coastal entrapment. Furthermore, greater onshore transport was observed for plastic debris that showed greater buoyancy under steeper wave conditions. Regarding the cross-shore distribution, the heaviest plastic debris that managed to be transported accumulated in the breaking zone, while the buoyant elements showed a predominant accumulation closer to the shoreline.</p>

Although rivers contribute to the flux of litter to the marine environment, estimates of riverine litter amounts and detailed studies on floating riverine litter behaviour once it has reached the sea are still scarce. This paper provides an analysis of the seasonal behaviour of floating marine litter released by rivers within the south-eastern Bay of Biscay based on riverine litter characterizations, drifters, and high-frequency radar observations and Lagrangian simulations. Virtual particles were released in the coastal area as a proxy of the floating fraction of riverine litter entering from rivers and reaching the open waters. Particles were parameterized with a wind drag coefficient (Cd) to represent their trajectories and fate according to the buoyancy of the litter items. They were forced with numerical winds and measured currents provided by high-frequency radars covering selected seasonal week-long periods between 2009 and 2021. To gain a better insight into the type and buoyancy of the items, samples collected from a barrier placed at the Deba River (Spain) were characterized at the laboratory. Items were grouped into two categories: low-buoyancy items (objects not exposed to wind forcing, e.g. plastic bags) and highly buoyant items (objects highly exposed to wind forcing, e.g. bottles). Overall, low-buoyancy items encompassed almost 90 % by number and 68 % by weight. Weakly buoyant items were parameterized with Cd = 0 % and highly buoyant items with Cd = 4 %; this latter value is the result of the joint analysis of modelled and observed trajectories of four satellite drifting buoys released at the Adour (France), Deba (Spain), and Oria (Spain) river mouths. Particles parameterized with Cd = 4 % drifted faster towards the coast through the wind, notably during the first 24 h. In summer, over 97 % of particles beached after 1 week of simulation. In autumn this value fell to 54 %. In contrast, low-buoyancy items took longer to arrive at the shoreline, particularly during spring with fewer than 25 % of particles beached by the end of the simulations. The highest concentrations (>200 particles km−1) were recorded during summer for Cd = 4 % in the French region of Pyrénées-Atlantiques. Results showed that the regions in the study area were highly affected by rivers within or nearby the region itself. These results couple observations and a river-by-river modelling approach and can assist decision-makers on setting emergency responses to high fluxes of floating riverine litter and on defining future monitoring strategies for heavily polluted regions within the south-eastern Bay of Biscay.

Sea-based sources account for 32-50 % of total marine litter found at the European basins with the fisheries sector comprising almost 65 % of litter releases. In the south-east coastal waters of the Bay of Biscay this figure approaches the contribution of just the floating marine litter fraction. This study seeks to enhance knowledge on the distribution patterns of floating marine litter generated by the fisheries sector within the Bay of Biscay and in particular on target priority Marine Protected Areas (MPAs) to reinforce marine litter prevention and mitigation policies. This objective is reached by combining the data on geographical distribution and intensity of fishing activity, long-term historical met-ocean databases, Monte Carlo simulations and Lagrangian modelling with floating marine litter source and abundance estimates for the Bay of Biscay. Results represent trajectories for two groups of fishing-related items considering their exposure to wind; they also provide their concentration within 34 MPAs. Zero windage coefficient is applied for low buoyant items not subjected to wind effect. Highly buoyant items, strongly driven by winds, are forced by currents and winds, using a windage coefficient of 4 %. Results show a high temporal variability on the distribution for both groups consistent with the met-ocean conditions in the area. Fishing-related items driven by a high windage coefficient rapidly beach, mainly in summer, and are almost non-existent on the sea surface after 90 days from releasing. This underlines the importance of windage effect on the coastal accumulation for the Bay of Biscay. Only around 20 % of particles escaped through the boundaries for both groups which gives added strength to the notion that the Bay of Biscay acts as accumulation region for marine litter. MPAs located over the French continental shelf experienced the highest concentrations (>75 particles/km2) suggesting their vulnerability and need for additional protection measures.

Oil spill risk assessments are important tools for the offshore oil and gas industries to minimize the consequences of deep spills. The stochastic modeling required in this kind of studies, is generally centered on surface transport and based on a Monte Carlo selection of hundreds or thousands of met-ocean scenarios from reanalysis databases, to create an ensemble of spill simulations. We propose a new integrated stochastic modeling methodology including both surface and subsurface transport, based on the specific selection of the most relevant environmental conditions through data-mining techniques. The methodology was applied to evaluate oil contamination probability as a consequence of a simulated deep release in the North Sea. Our results show the effectiveness of the proposed methodology to select representative evolutions of met-ocean conditions and to obtain pollution probabilities from an integrated subsurface and surface oil spill stochastic modeling, while assuring a manageable computational effort.

Marine litter is one of the main threats for the marine environment, causing significant damage at ecological, economic and social levels. Approximately, 80% of the marine litter comes from land-based sources, mainly from rivers. A significant percentage of this litter reaches the open oceans and the rest are retained inside the estuaries. The mechanisms that favor the marine litter accumulation inside the estuaries are determined by the interactions between the tide, the river flow, the waves, and the wind, as well as by the physical features of the estuary. In addition, tides and waves can contribute to the introduction of litter from marine sources. Therefore, estuaries frequently act as sinks for marine litter. In this study, a methodology to identify the most probable areas of marine litter accumulation inside estuaries is developed. The methodology is based on numerical modeling and statistical analysis and consists of 4 fundamental steps. First, a series of metocean scenarios (tidal conditions, river flow, waves, and wind), statistically representative of the studied estuary, are identified using the clustering algorithm K-means. Second, these conditions are used as forcing and boundary conditions of a hydrodynamic model to obtain the high spatial resolution currents and waves that determine the transport of litter. Third, with these high-resolution drivers, a Lagrangian transport model is fed to generate a database of potential marine litter trajectories from different litter-sources, where each litter-source is carefully selected according to the activity developed in the area. Finally, the statistical analysis of these trajectories allows identifying the most probable areas of marine litter accumulation. The efficacy of this methodology is demonstrated by its application to the Pas estuary (northern coast of Spain) by comparing the numerical results with field data. Results show that the greatest accumulations of marine litter take place in the curves that imply important changes in the direction of the flow, where the probabilities range between 20 and 30%, and that the distance to the litter-sources also plays a fundamental role.

In this study, a general methodology that is based on numerical models and statistical analysis is developed to assist in the definition of marine litter cleanup and mitigation strategies at an estuarine scale. The methodology includes four main steps: k-means clustering to identify representative metocean scenarios; dynamic down scaling to obtain high-resolution drivers with which to force a transport model; numerical transport modelling to generate a database of potential litter trajectories; and a statistical analysis of this database to obtain probabilities of litter accumulation. The efficacy of this methodology is demonstrated by its application to an estuary along the northern coast of Spain by comparing the numerical results with field data. The necessary criteria to ensure its applicability to any other estuary were provided. As the main conclusion, the developed methodology successfully assesses the litter distribution in estuaries with minimum computational effort.

Rivers and estuaries are among the main entrances of litter to the marine environment. This study characterizes marine litter deposits in three estuaries of the Gulf of Biscay, assesses its potential impact in estuarine habitats based on expert elucidation, and develops a methodology to estimate the associated environmental risk. Litter was ubiquitous in the estuaries of study, mostly represented by plastic debris and sanitary waste. High marsh communities acted as litter traps, showing significantly higher litter densities than adjacent habitats. The expected impact was valued to be low but different across habitats and possible litter-habitat interactions. The estimated risk was low but different across habitats and estuaries, determined by the probability of encounter and the expected impact. This study contributes to increase the scarce knowledge available on the threat that marine litter poses in estuarine environments and presents a methodology to help identify those habitats under a higher risk.

Past major oil spill disasters, such as the Prestige or the Deepwater Horizon accidents, have shown that spilled oil may drift across the ocean for months before being controlled or reaching the coast. However, existing oil spill modelling systems can only provide short-term trajectory simulations, being limited by the typical met-ocean forecast time coverage. In this paper, we propose a methodology for mid-long term (1-6 months) probabilistic predictions of oil spill trajectories, based on a combination of data mining techniques, statistical pattern modelling and probabilistic Lagrangian simulations. Its main features are logistic regression modelling of wind and current patterns and a probabilistic trajectory map simulation. The proposed technique is applied to simulate the trajectory of drifting buoys deployed during the Prestige accident in the Bay of Biscay. The benefits of the proposed methodology with respect to existing oil spill statistical simulation techniques are analysed.

In this work, we develop a high-resolution oceanographic database, suitable for marine renewable energy (MRE) applications. We apply downscaling techniques, using ROMS numerical model, to generate a 29-year (1985-2013), high-resolution (325 m), hourly sea level and currents (10 vertical levels) hindcast that reproduces storm surge and astronomical tide dynamics at a MRE test site (BiMEP) in the Bay of Biscay. The results are fully validated using instrumental data. We also analyze the result sensitivity to the open boundary conditions. We find that using boundary conditions from a baroclinic reanalysis, reproducing a wide range of physical processes, does not give more accurate results than a barotropic reanalysis, reproducing storm surge and astronomical tides, the main forcings at the study area. This analysis also leads us to address the importance of the temporal extent in the databases used for MRE applications. Finally, the generated hindcast is used to analyze normal and extreme currents and sea level regimes in BiMEP, following the internationally accepted standards for the design of MRE converters. The generated database offers MRE converter developers an alternative to the often-proposed empirical formulae for the definition of the design cases to consider in the loads and safety analysis of their devices. (C) 2018 Elsevier Ltd. All rights reserved.

Autoregressive logistic regression models have been demonstrated to be a powerful tool for statistical simulation of spatial patterns in climate and meteorology fields. In this paper we introduce a statistical framework for the simulation of ocean current patterns based on the autoregressive logistic regression models, and apply it to the Gulf of Mexico Loop Current. The statistical model is forced by three autoregressive terms, the wind stress curl in the Gulf of Mexico and in the Caribbean Sea, and the sea level pressure anomalies over the North Atlantic. It is used to replicate the bi-weekly historical sequence of 8 Loop Current patterns, obtained from a 24-year altimetry derived dataset. The model reproduces the inter-annual and intra-annual variability of the original time series, showing notable fitting capacity. A point-by-point comparison between the actual and simulated pattern series confirms the capability of the model in analysing the evolution of ocean current patterns. The predictive skill of the model is also explored, and the preliminary forecast (up to 3 months) results are encouraging. The presented statistical framework may find more practical applications in the future, such as the generation of statistically sound climate-based oceanographic scenarios for risk analyses, and the mid-term probabilistic prediction of ocean current patterns.

The high energy demand and the threat of climate change have led to a remarkable development of renewable energies, initially through technologies applied to the terrestrial environment and, recently, through the awakening of marine renewable energies. However, the development of these types of projects is often hampered by failure to pass the corresponding environmental impact assessment process. The complexity of working in the marine environment and the uncertainties associated with assessing the impacts of such projects make it difficult to carry out objective and precise environmental impact assessments. AMBEMAR-DSS seeks to establish a basis for understanding and agreement between the different stakeholders (project developers, public administrations, environmental organizations and the public in general), in order to find solutions that allow the development of marine renewable energies, minimizing their environmental cost. For this purpose, a DSS is proposed which, based on cartographic information and using objective and quantifiable criteria, allows comparative assessments and analyses between different project alternatives. The analytical procedures used by the system include, among others, hydrodynamic modeling tools and visual impact simulators. In addition, impacts on marine species are assessed taking into account intrinsic ecological and biological aspects. The magnitude of the impacts is quantified by means of fuzzy logic operations and the integration of all the elements is carried out by an interactive multi-criteria analysis. The results are shown in tables, graphs and figures of easy interpretation and can be also visualized geographically by means of a cartographic viewer. The system identifies the main impacts generated in the different phases of the project and allows establishing adequate mitigation measures in search of optimized solutions. The establishment of the assessment criteria has been based on the abundant, but dispersed, scientific literature on the various elements of the system and having the opinion of experts in the various fields. Nevertheless, the DSS developed constitutes a preliminary basis on which to build and improve a system with the input of researchers, promoters and experts from different disciplines.

ABSTRACTThis paper presents two methodologies to provide short-term and medium-term forecast of oil spill trajectories at local and regional scales. For short-term predictions (within 48 hours), a high-resolution operational oil spill forecast system is developed in Belfast Lough (Northern Ireland). Hydrodynamics are based on a Delft3D model which uses daily boundary conditions and meteorological forcing obtained from Copernicus Marine Environment Monitoring Service (CMEMS) and from the UK Meteorological Office. Downscaled currents and meteorological forecasts are used to provide short-term oil spill fate and trajectory predictions in the Lough using the oil spill numerical model TESEO. The system is integrated in a user-friendly web application that allows end users to launch the oil spill model both in case of pollution threat and for training purposes.For mid-term predictions (15–60 days), a stochastic methodology to provide probabilistic oil spill forecasts is presented and applied to the Bay of Biscay (Northern Spain). The method encompasses the following steps: 1) Classification of representative atmospheric patterns using principal component analysis and the k-means technique; 2) Setup of an autoregressive logistic model taking into account seasonality, covariates, long-term trends and autoregressive terms. In the case of an accident, we sample the evolution of the metocean conditions using the autoregressive model, which provides us with possible evolution patterns for these conditions during the forecasting period. These results are used to force the oil spill transport model TESEO allowing the characterization of trajectories in probabilistic terms.Drifting buoys released in Belfast Lough and observations reported during the Prestige accident have been used to validate the operational system and the medium-term forecasting methodologies.

This paper presents a high-resolution operational forecast system for providing support to oil spill response in Belfast Lough. The system comprises an operational oceanographic module coupled to an oil spill forecast module that is integrated in a user-friendly web application. The oceanographic module is based on Delft3D model which uses daily boundary conditions and meteorological forcing obtained from COPERNICUS and from the UK Meteorological Office. Downscaled currents and meteorological forecasts are used to provide short-term oil spill fate and trajectory predictions at local scales. Both components of the system are calibrated and validated with observational data, including ADCP data, sea level, temperature and salinity measurements and drifting buoys released in the study area. The transport model is calibrated using a novel methodology to obtain the model coefficients that optimize the numerical simulations. The results obtained show the good performance of the system and its capability for oil spill forecast.

This paper presents a novel operational oil spill modelling system based on HF radar currents, implemented in a northwest European shelf sea. The system integrates Open Modal Analysis (OMA), Short Term Prediction algorithms (STPS) and an oil spill model to simulate oil spill trajectories. A set of 18 buoys was used to assess the accuracy of the system for trajectory forecast and to evaluate the benefits of HF radar data compared to the use of currents from a hydrodynamic model (HDM). The results showed that simulated trajectories using OMA currents were more accurate than those obtained using a HDM. After 48h the mean error was reduced by 40%. The forecast skill of the STPS method was valid up to 6h ahead. The analysis performed shows the benefits of HF radar data for operational oil spill modelling, which could be easily implemented in other regions with HF radar coverage.