The Digital Twins of the Ocean (DTOs) represent an emerging framework for monitoring, simulating, and predicting ocean dynamics, supporting a range of applications relevant to understanding and responding to the global climate system. By integrating large-scale, multi-sourced datasets with advanced numerical models, DTOs provide a powerful tool for climate science. This review examines the role of machine learning (ML) in advancing DTOs applications, addressing the limitations of traditional methodologies under current conditions of increasing data availability from satellites, in situ sensors, and high-resolution numerical models. We highlight how ML serves as a versatile tool for enhancing DTOs capabilities, including real-time forecasting, correcting model biases, and filling data gaps where conventional approaches fall short. Furthermore, we review surrogate models that aim to complement or replace traditional physical models, offering increasing accuracy and the appeal of much faster inference for forecasts, and the insertion of hybrid models, which couple physics-based simulations with ML algorithms and are proving to be continuously improving in accuracy for complex oceanographic tasks as bigger datasets become available and methodologies evolve. This paper provides a comprehensive review of ML applications within DTOs, focusing on key areas such as water quality and marine biodiversity, ports, marine pollution, fisheries, and renewable energy. The review concludes with a discussion of future research directions and the potential of ML to foster more robust and practical DTOs, ultimately supporting informed decision-making for sustainable ocean management.

This study investigated advancements in atmospheric forecasting by integrating real-time observational data into the Weather Research and Forecasting (WRF) model through the WRF-Data Assimilation (WRF-DA) framework. By refining atmospheric models, we aimed to improve regional high-resolution wave and hydrodynamic forecasts essential for environmental management. Focused on southern Greece, including Crete, the study applied a 3D-Var assimilation technique within WRF, downscaling forecasting data from the Global Forecast System (GFS) to resolutions of 9 km and 3 km. The results showed a 4.7% improvement in wind speed predictions, with significant gains during forecast hours 26–72, enhancing model accuracy across METAR validation locations. These results underscore the positive impact of the integration of additional observational data on model accuracy. This study also highlights the utility of refined atmospheric models for real-world applications through their use in forcing ocean circulation and wave models and subsequent Digital Twin of the Ocean applications. Two such applications—optimal ship routing to minimize CO₂ emissions and oil spill trajectory forecasting to mitigate marine pollution—demonstrate the practical utility of improved models through what-if scenarios in easily deployable, containerized formats.

The world’s oceans have been studied and monitored for many decades to enhance our understanding. In today’s world, with the explosion of new data provided by many different Earth observation sources and the availability of advanced computing infrastructures (cloud computing, HPC, IoT, Big Data), creating a digital representation of the ocean is becoming a reality. The EC recently funded the H2020 ILIAD project, which aims at establishing an interoperable, data-intensive Digital Twin of the Ocean (DTO). The ILIAD DTO will integrate real-time sensing of ocean variables, state-of-the-art high-resolution models, modern data analytics and digital infrastructures to create virtual representations of physical processes and understand their behaviour, anticipating and predicting their response to simulated events and future changes. ILIAD will enable an ecosystem of interoperable DTOs, integrating the plethora of existing EU Earth Observing and Modelling Digital Infrastructures. It will fuse a large volume of diverse data and will enhance ocean data infrastructures with additional observation technologies and citizen science. ILIAD will provide a virtual environment representing the ocean, capable of running predictive management scenarios and will utilize Big Data analytics for forecasting of spatiotemporal events and pattern recognition. Several DT pilots will be undertaken in several key thematic areas such as offshore wind energy, wave and tidal energy, biodiversity assessments, marine pollution and more. The current work presents ongoing activities for a coastal, high-resolution Digital Twin pilot for Cretan Sea, to be demonstrated in the frame of ILIAD project. The pilot focuses on oil spill pollution monitoring and forecasting. The DT environment combines high-resolution forecasting services based on numerical weather (WRF), hydrodynamic (NEMO), sea state (WAVEWATCH III) and particle tracking models (MEDSLIK-II), enhanced by the integration of Sentinel data and real-time observations from novel, low cost current and waves meters, drifting trackers, as well as citizen science. WRF model is applied for forecasting of meteorological variables at ̴ 3 km resolution by dynamic downscaling of coarser resolution climatic modelling forecast data (NOAA’s GFS). This way, higher computational accuracy is achieved over Cretan Sea, thus revealing finer wind scales phenomena. The downscaled weather forecasting data are used to force NEMO and WAVEWATCH III, to obtain high-resolution forecasts of important marine parameters, such as sea currents, temperature, salinity and waves over a fine grid of ̴ 1 km for the coastal area of Crete. For oil spills, the DT of Cretan Sea will integrate operational analysis of Sentine-1 images, triggering MEDSLIK-II oil spill model once an oil spill event or anomaly is identified. Adjusting forecasts to observations by reinitialising the model with updated observational patterns will contribute to the forecast error growth being implicitly accounted for and minimized. The pilot DT virtual environment will allow on-demand simulations of predictive scenarios of oil spill events and response strategies. The aim of the Digital Twin is to aid the immediate response in case of accidental oil releases, minimize the damage and reduce the time for environmental recovery.Acknowledgement: This research has received funding from the European Union’s H2020 RIA programme under GA No 101037643.