1 Introduction

Marine observation and prediction are critical components of marine science, encompassing the monitoring and forecasting of various oceanic parameters such as sea surface temperature, ocean currents, and marine ecosystems. These activities are essential for understanding and managing the marine environment, which is increasingly impacted by human activities and climate change. Traditional methods of marine observation have relied heavily on physical models and in-situ measurements, but recent advancements in technology have introduced new tools and methodologies that enhance the accuracy and efficiency of these processes (Sarkar et al., 2020; Immas et al., 2021; Song et al., 2023).

The integration of key technologies such as artificial intelligence (AI), the Internet of Things (IoT), and advanced sensing technologies has revolutionized marine observation and prediction. AI, for instance, has been applied to various aspects of marine science, including the identification of ocean phenomena and the prediction of ocean components, significantly improving the accuracy and scope of marine forecasts (Sarkar et al., 2020; Song et al., 2023). IoT has facilitated real-time monitoring of marine environments by enabling the deployment of interconnected sensors that collect and transmit data continuously (Xu et al., 2019). Additionally, advancements in underwater sensing technologies have allowed for more precise and extensive data collection, which is crucial for accurate marine ecosystem modeling and forecasting (Capotondi et al., 2019; Sun et al., 2021).

This study aims to provide a comprehensive overview of the current state of research and technological advancements in marine observation and prediction, highlight the significant contributions of AI, IoT, and advanced sensing technologies in enhancing marine science, and identify the challenges and future directions in the application of these technologies for sustainable marine environment management. By synthesizing findings from multiple research studies, this study seeks to offer valuable insights into the evolving landscape of marine observation and prediction technologies and their potential to address pressing environmental challenges.

2 Key Technologies in Marine Observation

2.1 Remote sensing technologies

Remote sensing technologies have revolutionized marine observation by providing extensive spatial and temporal coverage of oceanic parameters. These technologies include satellite-based sensors that measure sea surface temperature, chlorophyll concentration, and sea level anomalies. The integration of remote sensing data with in-situ observations enhances the accuracy and comprehensiveness of marine environmental monitoring (Kim et al., 2020).

2.2 In-situ observation systems

In-situ observation systems are essential for obtaining high-resolution and accurate data on various oceanographic parameters. These systems include moored buoys, drifters, and underwater observatories equipped with sensors to measure temperature, salinity, currents, and other physical and chemical properties of seawater. The data collected from these systems are crucial for validating remote sensing observations and improving oceanographic models.

2.3 Autonomous underwater vehicles (AUVs)

Autonomous Underwater Vehicles (AUVs) have become indispensable tools in marine observation due to their ability to operate independently of a host vessel and access extreme environments. AUVs are equipped with advanced sensors such as multi-frequency acoustic imaging, side-scan sonars, and forward-looking sonars, which enable high-resolution mapping and monitoring of the seafloor and underwater features. They are used for various applications, including submarine volcanism studies, benthic habitat mapping, and seabed characterization (Sahoo et al., 2019; Zacchini et al., 2023).

2.4 Buoys and drifters

Buoys and drifters are critical components of the global ocean observing system. They provide real-time data on sea surface temperature, salinity, currents, and meteorological conditions. Drifters, which move with ocean currents, offer valuable insights into surface circulation patterns and help track the dispersion of pollutants and biological organisms. Moored buoys, on the other hand, provide long-term time series data at fixed locations, contributing to the understanding of oceanic processes and climate variability (Srinivasan et al., 2019).

2.5 Acoustic monitoring

Acoustic monitoring technologies play a vital role in marine observation by enabling the detection and characterization of underwater soundscapes. These technologies include hydrophones, acoustic Doppler current profilers (ADCPs), and passive acoustic monitoring systems. Acoustic monitoring is used for various purposes, such as studying marine mammal behavior, monitoring underwater volcanic activity, and mapping seafloor habitats. AUVs equipped with acoustic sensors have been particularly effective in conducting detailed seabed surveys and identifying buried objects (Lin et al., 2022).

2.6 Internet of things (IoT) in marine monitoring

The Internet of Things (IoT) has revolutionized marine monitoring by deploying various sensors in real-time environments to measure physical parameters. These sensors, however, rely on battery power, which can lead to interruptions in monitoring activities. Advanced prediction models using Principal Component Analysis (PCA) and Deep Neural Networks (DNN) have been developed to predict battery life and alert technologists, ensuring uninterrupted monitoring. These models have shown significant improvements in accuracy and time efficiency compared to traditional techniques (Siva et al., 2022).

3 Data Integration and Management

3.1 Big data analytics

The advent of big data analytics has revolutionized marine observation and prediction by enabling the processing and analysis of vast amounts of data collected from various sources. The integration of big data analytics allows for the extraction of meaningful patterns and insights from complex datasets, which is crucial for understanding and predicting oceanic phenomena. For instance, the use of high-performance computing and advanced algorithms facilitates the analysis of long-term datasets, enhancing the accuracy of marine models and predictions (Liu et al., 2017; Vance et al., 2019).

3.2 Data assimilation techniques

Data assimilation techniques are essential for improving the accuracy of marine models by integrating observational data with model predictions. Various methods, such as variational and sequential approaches, have been developed to address the challenges of representing physical, chemical, and biological properties in the ocean. Recent advances include ensemble and four-dimensional variational methods, which have shown promise in regional ocean systems and biogeochemical applications (Edwards et al., 2015; Zalesny et al., 2020). These techniques help in producing more accurate estimates of the ocean state by combining observations and model dynamics.

3.3 Cloud computing and storage solutions

Cloud computing offers significant opportunities for managing and analyzing marine data. The shift to cloud-based platforms allows for the development of shared data processing workflows and the utilization of adaptable software for data ingestion and storage. Cloud computing enables high-performance mass storage of observational data and on-demand computing for running model simulations, which can be done in close proximity to the data. This approach facilitates a more flexible and adaptable observation and prediction computing architecture, allowing researchers to access and analyze data more efficiently (Vance et al., 2019). Additionally, cloud platforms provide tools to manage workflows and frameworks for collaboration, making it easier to create, analyze, and distribute products derived from long-term datasets.

3.4 Interoperability standards

Interoperability standards are crucial for ensuring that different marine observation and prediction systems can work together seamlessly. These standards facilitate the integration of data from various sources and disciplines, enabling a more comprehensive understanding of oceanic processes. The development and adoption of common standards for data formats, metadata, and communication protocols are essential for achieving interoperability. This allows for the efficient sharing and utilization of data across different platforms and systems, enhancing the overall effectiveness of marine observation and prediction efforts (Vance et al., 2019).

By leveraging big data analytics, advanced data assimilation techniques, cloud computing, and interoperability standards, the field of marine observation and prediction can achieve significant advancements in understanding and forecasting oceanic phenomena. These technologies and methodologies provide the foundation for more accurate and reliable marine models, ultimately contributing to better management and protection of marine environments.

4 Key Technologies in Marine Prediction

4.1 Numerical modeling and simulation

Numerical modeling and simulation are foundational techniques in marine prediction, enabling the representation of complex oceanic processes through mathematical formulations. These models are essential for predicting various marine phenomena, including sea surface temperatures (SST), wave heights, and ocean currents. For instance, traditional numerical models based on physics-based assumptions are widely used for SST prediction, although they are often complemented by machine learning techniques to enhance accuracy and reduce computational demands (Sarkar et al., 2020; Ali et al., 2021). Additionally, the integration of high-resolution observational data, such as satellite and in-situ measurements, is crucial for improving the skill of these models, particularly in predicting mesoscale ocean features like eddies (Jacobs et al., 2021).

4.2 Machine learning and AI applications

The application of machine learning (ML) and artificial intelligence (AI) in marine prediction has seen significant growth, offering new methodologies to complement traditional numerical models. AI techniques, such as deep learning neural networks and ensemble machine learning models, have been successfully applied to predict SST, wave heights, and other oceanic parameters with high accuracy (Chen et al., 2021; Panda et al., 2021; Song et al., 2023). These models leverage large datasets from ocean observations and numerical simulations to train predictive algorithms, which can outperform traditional statistical models in terms of accuracy and computational efficiency (Ali et al., 2021). AI is also used for identifying and forecasting ocean phenomena like internal waves, El Niño-Southern Oscillation (ENSO), and sea ice, demonstrating its versatility and potential in marine science (Dong et al., 2022; Song et al., 2023).

4.3 Ensemble forecasting methods

Ensemble forecasting methods involve generating multiple forecasts using different models or varying initial conditions to account for uncertainties in predictions. This approach is particularly useful in marine prediction, where the inherent variability of oceanic processes can lead to significant forecast errors. Ensemble methods have been applied to predict surface chloride concentration in marine concrete, demonstrating improved accuracy over standalone models (Cai et al., 2020). Additionally, ensemble forecasting is used to manage the gap between observation and model resolution, allowing for better prediction of constrained and unconstrained ocean features (Jacobs et al., 2021). By combining predictions from various models, ensemble methods provide a more robust and reliable forecast, essential for operational marine forecasting.

4.4 Real-time prediction systems

Real-time prediction systems are critical for providing timely and accurate forecasts of marine conditions, which are essential for navigation, fisheries, and disaster management. These systems integrate real-time observational data with numerical models and machine learning algorithms to deliver up-to-date predictions. For example, a phase-resolving wave-forecasting algorithm that assimilates marine radar data has been developed to provide real-time wave forecasts, demonstrating the potential of integrating observational data with numerical models for immediate applications (Simpson, 2020). Similarly, real-time SST forecasts have been enhanced by combining deep learning neural networks with traditional numerical estimators, offering precise location-specific predictions (Sarkar et al., 2020). The continuous advancement of real-time prediction systems is vital for improving the safety and efficiency of marine operations.

5 Applications of Marine Observation and Prediction

5.1 Climate change studies

Marine observation technologies play a crucial role in understanding and predicting climate change impacts on marine environments. Advances in remote sensing, environmental DNA (eDNA) assessments, and animal telemetry provide valuable data on ocean conditions and biodiversity, which are essential for climate models and predictions (Capotondi et al., 2019; Ruhl et al., 2021). These technologies help monitor changes in sea level, temperature, and ocean currents, which are influenced by climate variability and change (Capotondi et al., 2019). Additionally, artificial intelligence (AI) algorithms are increasingly used to analyze large datasets, improving the accuracy of climate predictions and identifying phenomena such as El Niño-Southern Oscillation (ENSO) and heatwaves (Song et al., 2023).

5.2 Marine resource management

Effective marine resource management relies on accurate and timely data on marine ecosystems. Innovations in monitoring systems, such as molecular approaches (e.g., qPCR, metabarcoding), optical sensing, and acoustic methods, enhance the ability to assess marine biodiversity and environmental status. These technologies support the implementation of frameworks like the Marine Strategy Framework Directive (MSFD) by providing cost-effective and efficient monitoring solutions. AI models also contribute to resource management by predicting ocean components and optimizing the use of marine resources (Jiang et al., 2022; Song et al., 2023).

5.3 Disaster prevention and mitigation

Marine observation and prediction technologies are vital for disaster prevention and mitigation. Satellite remote sensing and in situ monitoring instruments provide real-time data on ocean conditions, which are crucial for forecasting and responding to natural disasters such as tsunamis, hurricanes, and storm surges (Capotondi et al., 2019; Ruhl et al., 2021). AI algorithms enhance the predictive capabilities of these systems, allowing for more accurate and timely warnings (Jiang et al., 2022; Song et al., 2023). Improved observational networks and holistic approaches, such as those inspired by landscape ecology, support better understanding and forecasting of marine disasters (Capotondi et al., 2019).

5.4 Maritime navigation and safety

Ensuring maritime navigation and safety requires comprehensive monitoring of marine environments. Technologies such as underwater imaging devices, passive and active acoustic sensors, and satellite observations provide critical data on marine life, boat traffic, and environmental conditions (Ruhl et al., 2021). These observations help in planning safe navigation routes and avoiding hazards. AI-driven models further enhance navigation safety by predicting sea ice, tide levels, and other navigational challenges (Jiang et al., 2022; Song et al., 2023). The integration of various observational techniques, including hydroacoustics and video surveys, offers a cost-effective approach to monitoring and ensuring maritime safety.

5.5 Ecosystem and biodiversity monitoring

Monitoring marine ecosystems and biodiversity is essential for conservation and management efforts. The Marine Biodiversity Observation Network (MBON) and other collaborative initiatives facilitate the exchange of information on marine life, linking policy and management needs with observational data (Kavanaugh et al., 2021; Ruhl et al., 2021). Remote sensing technologies, combined with in situ observations, provide comprehensive data on biophysical interactions and biodiversity changes in coastal zones (Kavanaugh et al., 2021). AI algorithms also play a significant role in identifying and predicting changes in marine biodiversity, supporting efforts to protect and manage marine ecosystems (Jiang et al., 2022; Song et al., 2023).

6 Case Studies

6.1 Successful implementations of key technologies

6.1.1 Drifting buoys

Drifting buoys, part of the global Argo program, have significantly contributed to the observation of ocean temperature and salinity profiles. These buoys float with the ocean currents, collecting data that improves weather forecasts and climate models. The widespread deployment of Argo buoys has created a comprehensive dataset that is invaluable for oceanographic research. Li et al. (2023) systematically analyzed the research status of salinity optic fiber sensors (OFSs) on the drifting buoys for seawater salinity in marine environmental monitoring, and summarized the sensing mechanisms (Figure 1), research progress, and measurement performance indicators of various existing salinity optic fiber sensors (OFSs), in response to the actual measurement needs of seawater salinity.

Figure 1 Sensing mechanisms of the salinity optic fiber sensors (OFSs) (Adopted from Li et al., 2023) Image caption: (a) Structure of the composite reflection probe; (b) Structure and sensing mechanism of the two-channel SPR sensor based on HCF; (c) Schematic of the EC-MOF sensor; (d) Schematic diagram of the SPR-based optical fiber sensor (Adopted from Li et al., 2023)

6.1.2 Satellite remote sensing

Satellite remote sensing has been successfully employed to monitor sea surface temperatures, chlorophyll concentrations, and ocean currents. The MODIS (Moderate Resolution Imaging Spectroradiometer) instrument aboard NASA's Aqua satellite has been particularly effective in tracking algal blooms and assessing their impact on marine ecosystems. This technology has provided critical data for understanding the dynamics of large-scale oceanographic phenomena. Studies such as Nielsen-Englyst et al. (2018) have shown the effectiveness of satellite remote sensing in monitoring sea surface temperatures, which is essential for climate research and marine biology.

6.1.3 High-frequency radar (HFR)

High-Frequency Radar (HFR) technology is widely regarded as a cost-effective tool for monitoring coastal areas and has been employed in coastal monitoring around the world. Globally, the number of HFR stations is steadily increasing. Mantovani et al. (2020) studied the best practices for the deployment and operation of high-frequency radar for ocean current measurements (Figure 2).

Figure 2 A typical setup of a beam forming system (Adopted from Mantovani et al., 2020) Image caption: The photos show the antennas placement of a WERA system on the island of Wangerooge at the German coast of the North Sea. The receive antenna array (Rx) is shown on the left-hand side and the transmit array (Tx) on the right hand side, respectively. The bottom panel shows a typical direction finding HF radar system setup for the low frequency (4, 5, and 9 MHz) bands (left hand side). To the right, the SeaSonde installed at Matxitxako Cape (northern coast of Spain) (Adopted from Mantovani et al., 2020)

6.1.4 Autonomous underwater vehicles (AUVs)

AUVs have revolutionized marine observation by enabling high-resolution mapping of the seafloor and monitoring of oceanographic parameters. A notable example is the use of AUVs in the exploration of the Mariana Trench, where they provided unprecedented data on the trench’s depth and marine life. The AUVs deployed were equipped with advanced sonar systems, cameras, and sensors, allowing for detailed geological and biological surveys (Crawford et al., 2022).

6.1.5 Ocean gliders

Ocean gliders, such as those used in the RAPID project, have been instrumental in monitoring the Atlantic Meridional Overturning Circulation (AMOC). These gliders, equipped with CTD (conductivity, temperature, depth) sensors, provide continuous measurements of water column properties, helping to improve predictions of climate variability and change. Gentil et al. (2020) discuss the integration of ADCPs onto gliders for monitoring currents and turbidity in the coastal zone, highlighting their use in the Rhone River region (Figure 3).

Figure 3 Observation results of ocean gliders (Adopted from Gentil et al., 2020) Image caption: Hydrological variables: (a, b) temperature, (c, d) absolute salinity, (e, f) density anomalies, and (g, h) the Brunt–Väisälä frequency. The isopycnals are superimposed on all plots and indicated by black or white lines. The black arrow at the top of each panel indicates the direction of the glider’s motion (Adopted from Gentil et al., 2020)

6.2 Lessons learned from field applications

Drifting buoys face challenges such as loss due to harsh ocean conditions and limitations in real-time data transmission. Enhancing the robustness of buoys and improving satellite communication technologies are vital for the reliability of these systems.

The primary challenge with satellite remote sensing is the resolution and coverage limitations. Cloud cover and atmospheric conditions can impede data quality. Integrating satellite data with in-situ observations can mitigate these limitations and provide more accurate and comprehensive datasets.

HFR systems can be affected by electromagnetic interference and require substantial infrastructure. Strategic placement and regular calibration of HFR systems are necessary to maintain data accuracy and reliability.

While AUVs have proven to be invaluable tools, challenges such as battery life, data storage capacity, and navigation in complex terrains have been noted. Continuous improvements in power management and autonomous decision-making algorithms are essential for enhancing their operational efficiency.

Operational challenges for gliders include biofouling, which affects sensor accuracy, and difficulties in retrieving data from remote locations. Regular maintenance and the development of antifouling technologies are critical for long-term deployments.

6.3 Comparative analysis of different approaches

6.3.1 Data coverage and resolution

Satellite remote sensing provides broad spatial coverage but may lack the resolution of in-situ measurements like those from AUVs and gliders. Combining these technologies can optimize data collection, offering both extensive coverage and high-resolution insights (Chai et al., 2020; Loveday et al., 2022).

6.3.2 Operational cost and maintenance

AUVs and gliders involve higher operational costs and maintenance compared to drifting buoys and satellite remote sensing. However, they provide more detailed and specific data. Budget considerations and specific research goals should guide the choice of technology (Whitt et al., 2020).

6.3.3 Real-time data availability

Drifting buoys and satellite systems generally offer real-time data transmission, which is critical for immediate decision-making. In contrast, AUVs and gliders may have delays in data retrieval. Ensuring a balance between real-time data needs and detailed measurements is crucial for effective marine observation (Vicen-Bueno et al., 2019).

6.3.4 Environmental impact

The environmental impact of deploying these technologies varies. AUVs and gliders are less intrusive but may disturb local marine life temporarily. Drifting buoys and HFR have minimal direct impact but require careful consideration of their long-term deployment on marine ecosystems (Aniceto et al., 2020).

6.3.5 Integration and interoperability

The integration of different observation technologies enhances data robustness and reliability. Projects like the Global Ocean Observing System (GOOS) exemplify the benefits of combining satellite, buoy, and in-situ data. Standardizing protocols and improving interoperability are key to maximizing the utility of diverse marine observation systems (Whitt et al., 2020).

7 Challenges and Future Directions

7.1 Technical and methodological challenges

The integration of advanced technologies such as artificial intelligence (AI) and machine learning (ML) into marine observation and prediction systems presents several technical and methodological challenges. One significant challenge is the discrepancy between the volume of data collected and our capacity for data analysis. While sensor technology and autonomous platforms have advanced, enabling the collection of vast amounts of data, the analytical capabilities have not kept pace, creating a bottleneck in effective data utilization (Malde et al., 2020). Additionally, the formulation and accuracy of data assimilative models are highly dependent on the quality and quantity of interdisciplinary observational data, which remains a challenge due to the complexity and variability of oceanographic processes.

7.2 Knowledge gaps and research needs

There are notable knowledge gaps in understanding the physical and biological processes that influence marine ecosystems, which limit prediction capabilities. For instance, the lack of sufficient observations for forecast initialization and verification hampers the development of accurate models (Capotondi et al., 2019). Furthermore, the influence of climate change on ocean conditions, such as sea level rise and increased stratification, adds another layer of complexity to marine forecasting (Capotondi et al., 2019). Research is needed to develop more comprehensive observational networks and to improve the understanding of the interactions between different oceanographic variables and processes.

7.3 Emerging trends and innovations

Emerging trends in marine observation and prediction include the increasing application of AI and ML to enhance data analysis and forecasting capabilities. AI algorithms are being used to identify and predict ocean phenomena such as internal waves, heatwaves, and the El Niño-Southern Oscillation (ENSO) (Song et al., 2023). The integration of AI with traditional physical models is also a growing trend, offering new ways to improve the accuracy and efficiency of marine predictions (Song et al., 2023). Additionally, new holistic observational approaches, inspired by initiatives like Tara Oceans, are being developed to support and expand ecosystem modeling and forecasting by bridging global and local observations (Capotondi et al., 2019). These innovations hold promise for addressing current challenges and advancing the field of marine science.

8 Concluding Remarks

The development and application of key technologies in marine observation and prediction have significantly advanced our understanding of oceanographic processes and improved the accuracy of marine forecasts. Drifting buoys, part of the Argo program, have provided invaluable data on ocean temperature and salinity profiles, significantly impacting weather forecasts and climate models. Satellite remote sensing, particularly through instruments like MODIS, has enabled extensive monitoring of sea surface temperatures, chlorophyll concentrations, and ocean currents, playing a crucial role in tracking large-scale oceanographic phenomena. High-Frequency Radar (HFR) systems have been effective in measuring coastal surface currents. Autonomous Underwater Vehicles (AUVs) have revolutionized seafloor mapping and deep-sea exploration, providing high-resolution geological and biological data. Ocean gliders has enhanced the monitoring of water column properties and current velocities, contributing to climate variability predictions and navigation support.

The rapid advancements in marine observation technologies underscore the importance of continued innovation and research. As environmental challenges and climate change impacts intensify, there is an increasing need for precise and comprehensive marine data. The integration of artificial intelligence (AI) and machine learning (ML) into marine observation and prediction systems holds great potential to enhance data analysis capabilities and improve the accuracy of ocean forecasts. However, the challenges of data volume, analytical capacity, and model accuracy necessitate ongoing research and development. Interdisciplinary observational networks and improved understanding of oceanographic processes are critical to addressing these challenges and advancing the field.

To maximize the potential of marine observation technologies, several recommendations are proposed. Firstly, enhancing data integration and interoperability by developing standardized protocols for data collection, processing, and sharing across different technologies and platforms, as well as fostering collaborations between international marine observation programs to create a unified global observation network. Meanwhile, investing in AI and ML technologies by focusing on developing algorithms tailored to marine data analysis and prediction and integrating AI with traditional physical models to enhance the accuracy and efficiency of marine forecasts. Additionally, expanding observational networks by increasing the deployment of drifting buoys, gliders, and HFR systems to improve spatial and temporal coverage of oceanographic data and establishing more comprehensive observational networks in underrepresented regions, particularly in the Southern Hemisphere and polar areas. Lastly, addressing knowledge gaps through targeted research by conducting focused studies on the physical and biological processes that influence marine ecosystems to improve model formulations and studying the impacts of climate change on ocean conditions to enhance predictive capabilities.

Continued innovation, research, and collaboration are essential to advancing marine observation technologies and improving our understanding and management of the oceans. These efforts will ultimately contribute to better environmental stewardship and more effective responses to global marine challenges.

Conflict of Interest Disclosure

The authors affirm that this research was conducted without any commercial or financial relationships that could be construed as a potential conflict of interest.

References

Ali A., Fathalla A., Salah A., Bekhit M., and Eldesouky E., 2021, Marine data prediction: an evaluation of machine learning deep learning and statistical predictive models, Computational Intelligence and Neuroscience, 2021(1): 8551167.

https://doi.org/10.1155/2021/8551167

Aniceto A., Pedersen G., Primicerio R., Biuw M., Lindstrøm U., and Camus L., 2020, Arctic marine data collection using oceanic gliders: providing ecological context to cetacean vocalizations, Frontiers in Marine Science, 7: 585754.

https://doi.org/10.3389/fmars.2020.585754

Cai R., Han T., Liao W., Huang J., Li D., Kumar A., and Ma H., 2020, Prediction of surface chloride concentration of marine concrete using ensemble machine learning, Cement and Concrete Research, 136: 106164.

https://doi.org/10.1016/j.cemconres.2020.106164

Capotondi A., Jacox M., Bowler C., Kavanaugh M., Lehodey P., Barrie D., Brodie S., Chaffron S., Cheng W., Dias D., Eveillard D., Guidi L., Iudicone D., Lovenduski N., Nye J., Ortiz I., Pirhalla D., Buil M., Saba V., Sheridan S., Siedlecki S., Subramanian A., Vargas C., Lorenzo E., Doney S., Hermann A., Joyce T., Merrifield M., Miller A., Not F., and Pesant S., 2019, Observational needs supporting marine ecosystems modeling and forecasting: from the global ocean to regional and coastal systems, Frontiers in Marine Science, 6: 623.

https://doi.org/10.3389/fmars.2019.00623

Chai F., Johnson K., Claustre H., Xing X., Wang Y., Boss E., Riser S., Fennel K., Schofield O., and Sutton A., 2020, Monitoring ocean biogeochemistry with autonomous platforms, Nature Reviews Earth and Environment, 1: 315-326.

https://doi.org/10.1038/s43017-020-0053-y.

Chen J., Pillai A., Johanning L., and Ashton I., 2021, Using machine learning to derive spatial wave data: a case study for a marine energy site, Environ., Model., Softw., 142: 105066.

https://doi.org/10.1016/J.ENVSOFT.2021.105066

Crawford A., Shore J., and Shan S., 2022, Measurement of tidal currents using an autonomous underwater vehicle, IEEE Journal of Oceanic Engineering, 47: 282-294.

https://doi.org/10.1109/joe.2021.3112807

Dong C.G., Xu G.J., Han G.Q., Bethel B.D., Xie W.H., and Zhou S.Y., 2022, Recent developments in artificial intelligence in oceanography, Ocean-Land-Atmosphere Research, 6: 623.

https://doi.org/10.34133/2022/9870950

Edwards C., Moore A., Hoteit I., and Cornuelle B., 2015, Regional ocean data assimilation, Annual Review of Marine Science, 7: 21-42.

https://doi.org/10.1146/annurev-marine-010814-015821

Gentil M., Many G., Madron X., Cauchy P., Pairaud I., Testor P., Verney R., and Bourrin F., 2020, Glider-based active acoustic monitoring of currents and turbidity in the coastal zone, Remote., Sens., 12: 2875.

https://doi.org/10.3390/rs12182875

Immas A., Do N., and Alam M., 2021, Real-time in situ prediction of ocean currents, Ocean Engineering, 228: 108922.

https://doi.org/10.1016/J.OCEANENG.2021.108922

Jacobs G., D’Addezio J., Ngodock H., and Souopgui I., 2021, Observation and model resolution implications to ocean prediction, Ocean Modelling, 159: 101760.

https://doi.org/10.1016/J.OCEMOD.2021.101760

Jiang M., and Zhu Z.Y., 2022, The role of artificial intelligence algorithms in marine scientific research, Frontiers in Marine Science, 9: 920994.

https://doi.org/10.3389/fmars.2022.920994

Kavanaugh M., Bell T., Catlett D., Cimino M., Doney S., Klajbor W., Messié M., Montes E., Karger F., Otis D., Santora J., Schroeder I., Trinanes J., and Siegel D., 2021, Satellite remote sensing and the marine biodiversity observation network: current science and future steps, Oceanography, 34(2): 62-79.

https://doi.org/10.5670/oceanog.2021.215

Kim K., Kim E., Choi J., Shin J., Kim W., Lee K., Son Y., and Ryu J., 2020, Simulation approach for the tracing the marine pollution using multi-remote sensing data, Journal of remote sensing, 36: 249-261.

https://doi.org/10.7780/KJRS.2020.36.2.2.3

Li G.C., Wang Y.J., Shi A.C., Liu Y.H., and Li F., 2023, Review of seawater fiber optic salinity sensors based on the refractive index detection principle, Sensors Basel Switzerland, 23(4): 2187.

https://doi.org/10.3390/s23042187

Lin Y.T., Chaytor J., Packard G., Chen T.T., and Kukshtel N., 2022, Acoustic seabed characterization with autonomous underwater vehicles, The Journal of the Acoustical Society of America, 152(4_Supplement): A102-A102.

https://doi.org/10.1121/10.0015684

Liu Y.J., Qiu M., Liu C., and Guo Z.W., 2017, Big data challenges in ocean observation: a survey, Personal and Ubiquitous Computing, 21: 55-65.

https://doi.org/10.1007/s00779-016-0980-2

Loveday B., Smyth T., Akpınar A., Hull T., Inall M., Kaiser J., Queste B., Tobermann M., Williams C., and Palmer M., 2022, Application of a new net primary production methodology: a daily to annual-scale data set for the North Sea derived from autonomous underwater gliders and satellite Earth observation, Earth System Science Data, 14(9): 3997-4016.

https://doi.org/10.5194/essd-14-3997-2022

Malde K., Handegard N., Eikvil L., and Salberg A., 2020, Machine intelligence and the data-driven future of marine science, ICES Journal of Marine Science, 77(4): 1274-1285.

https://doi.org/10.1093/icesjms/fsz057

Mantovani C., Corgnati L., Horstmann J., Rubió A., Reyes E., Quentin C., Cosoli S., Asensio J., Mader J., and Griffa A., 2020, Best practices on high frequency radar deployment and operation for ocean current measurement, Frontiers in Marine Science, 7: 210.

https://doi.org/10.3389/fmars.2020.00210

Panda J., 2021, Machine learning for naval architecture ocean and marine engineering, Journal of Marine Science and Technology, 28: 1-26.

https://doi.org/10.1007/s00773-022-00914-5

Ruhl H.A., Brown J.A., Harper A.R., Hazen E.L., Dewitt L., Daniel P., DeVogelaere A., Kudela R.M., Ryan J.R., Fischer A.D., Muller‐Karger F.E., and Chavez F.P., 2021, Integrating biodiversity and environmental observations in support of national marine sanctuary and large marine ecosystem assessments, Oceanography, 34(2): 142-155.

https://doi.org/10.5670/oceanog.2021.221

Sahoo A., Dwivedy S., and Robi P., 2019, Advancements in the field of autonomous underwater vehicle, Ocean Engineering, 181: 145-160.

https://doi.org/10.1016/J.OCEANENG.2019.04.011

Sarkar P., Janardhan P., and Roy P., 2020, Prediction of sea surface temperatures using deep learning neural networks, SN Applied Sciences, 2(8): 1458.

https://doi.org/10.1007/s42452-020-03239-3

Simpson A., 2020, Wave-by-wave forecasting via assimilation of marine radar data, Journal of Atmospheric and Oceanic Technology, 37(7): 1269-1288.

https://doi.org/10.1175/JTECH-D-19-0127.1

Siva T., Benő A., Lanitha B., Yogalakshmi V., Manikandan M., Kumar S., Peroumal V., Nesakumar A., and Prasad V., 2022, Hybrid LSTM-PCA-powered renewable energy-based battery life prediction and management for IoT applications, Journal of Nanomaterials, 2022(1): 9807511.

https://doi.org/10.1155/2022/9807511

Song T., Pang C., Hou B.Y., Xu G.X., Xue J.Y., Sun H.D., and Meng F., 2023, A review of artificial intelligence in marine science, Frontiers in Earth Science, 11: 1090185.

https://doi.org/10.3389/feart.2023.1090185

Srinivasan R., Rajendran V., Zacharia S., and Sudhakar T., 2019, A study of ocean parameters in bay of bengal BoB using indigenised drifting buoys, Journal of Earth System Science, 128: 1-8.

https://doi.org/10.1007/s12040-019-1242-2

Sun K., Cui W., and Chen C., 2021, Review of underwater sensing technologies and applications, Sensors Basel Switzerland, 21(23): 7849.

https://doi.org/10.3390/s21237849

Vance T., Wengren M., Burger E., Hernandez D., Kearns T., Medina-López E., Merati N., O’brien K., O’Neil J., Potemra J., Signell R., and Wilcox K., 2019, From the oceans to the cloud: opportunities and challenges for data models computation and workflows, Frontiers in Marine Science, 6: 211.

https://doi.org/10.3389/fmars.2019.00211

Vicen-Bueno R., Cimino G., Cecchi D., and Garau B., 2019, Near real-time scientific data provision from unmanned underwater gliders to remote operational sites, OCEANS 2019-Marseille, 2019: 1-9.

https://doi.org/10.1109/OCEANSE.2019.8867305

Whitt C., Pearlman J., Polagye B., Caimi F., Muller‐Karger F., Copping A., Spence H., Madhusudhana S., Kirkwood W., Grosjean L., Fiaz B., Singh S., Singh S., Manalang D., Gupta A., Maguer A., Buck J., Marouchos A., Atmanand M., Venkatesan R., Narayanaswamy V., Testor P., Douglas E., Halleux S., and Khalsa S., 2020, Future vision for autonomous ocean observations, Frontiers in Marine Science, 7: 697.

https://doi.org/10.3389/fmars.2020.00697

Xu G.B., Shi Y.J., Sun X.Y., and Shen W.M., 2019, Internet of things in marine environment monitoring: a review, Sensors Basel Switzerland, 19(7): 1711.

https://doi.org/10.3390/s19071711

Zacchini L., Topini A., Franchi M., Secciani N., Manzari V., Bazzarello L., Stifani M., and Ridolfi A., 2023, Autonomous underwater environment perceiving and modeling: an experimental campaign with feelhippo auv for forward looking sonar-based automatic target recognition and data association, IEEE Journal of Oceanic Engineering, 48: 277-296.

https://doi.org/10.1109/JOE.2022.3209719

Zalesny V., Agoshkov V., Shutyaev V., Parmuzin E., and Zakharova N., 2020, Numerical modeling of marine circulation with 4D variational data assimilation, Journal of Marine Science and Engineering, 8(7): 503.

https://doi.org/10.3390/jmse8070503