Aleksei V. Buinyi, Dias A. Irishev, Edvard E. Nikulin, Aleksandr A. Evdokimov, Polina G. Ilyushina, Natalia A. Sukhikh
{"title":"Optimizing data-driven arctic marine forecasting: a comparative analysis of MariNet, FourCastNet, and PhyDNet","authors":"Aleksei V. Buinyi, Dias A. Irishev, Edvard E. Nikulin, Aleksandr A. Evdokimov, Polina G. Ilyushina, Natalia A. Sukhikh","doi":"10.3389/fmars.2024.1456480","DOIUrl":null,"url":null,"abstract":"IntroductionMarine forecasts play a crucial role in ensuring safe navigation, efficient offshore operations, coastal management, and research, particularly in regions with challenging conditions like the Arctic Ocean. These forecasts necessitate precise predictions of ocean currents, wind-driven waves, and various other oceanic parameters. Although physics-based numerical models are highly accurate, they come with significant computational requirements. Therefore, data-driven approaches, which are less computationally intensive, may present a more effective solution for predicting sea conditions.MethodsThis study introduces a detailed analysis and comparison of three data-driven models: the newly developed convLSTM-based MariNet, FourCastNet, and PhydNet, a physics-informed model designed for video prediction. Through the utilization of metrics such as RMSE, Bias, and Correlation, we illustrate the areas in which our model outperforms well-known prediction models.ResultsOur model demonstrates enhanced accuracy in forecasting ocean dynamics when compared to FourCastNet and PhyDNet. Additionally, our findings reveal that our model demands significantly less training data and computational resources, ultimately resulting in lower carbon emissions.DiscussionThese findings indicate the potential for further exploration of data-driven models as a supplement to physics-based models in operational marine forecasting, as they have the capability to improve prediction accuracy and efficiency, thereby facilitating more responsive and cost-effective forecasting systems.","PeriodicalId":12479,"journal":{"name":"Frontiers in Marine Science","volume":"245 1","pages":""},"PeriodicalIF":2.8000,"publicationDate":"2024-11-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Frontiers in Marine Science","FirstCategoryId":"99","ListUrlMain":"https://doi.org/10.3389/fmars.2024.1456480","RegionNum":2,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"MARINE & FRESHWATER BIOLOGY","Score":null,"Total":0}
引用次数: 0
Abstract
IntroductionMarine forecasts play a crucial role in ensuring safe navigation, efficient offshore operations, coastal management, and research, particularly in regions with challenging conditions like the Arctic Ocean. These forecasts necessitate precise predictions of ocean currents, wind-driven waves, and various other oceanic parameters. Although physics-based numerical models are highly accurate, they come with significant computational requirements. Therefore, data-driven approaches, which are less computationally intensive, may present a more effective solution for predicting sea conditions.MethodsThis study introduces a detailed analysis and comparison of three data-driven models: the newly developed convLSTM-based MariNet, FourCastNet, and PhydNet, a physics-informed model designed for video prediction. Through the utilization of metrics such as RMSE, Bias, and Correlation, we illustrate the areas in which our model outperforms well-known prediction models.ResultsOur model demonstrates enhanced accuracy in forecasting ocean dynamics when compared to FourCastNet and PhyDNet. Additionally, our findings reveal that our model demands significantly less training data and computational resources, ultimately resulting in lower carbon emissions.DiscussionThese findings indicate the potential for further exploration of data-driven models as a supplement to physics-based models in operational marine forecasting, as they have the capability to improve prediction accuracy and efficiency, thereby facilitating more responsive and cost-effective forecasting systems.
期刊介绍:
Frontiers in Marine Science publishes rigorously peer-reviewed research that advances our understanding of all aspects of the environment, biology, ecosystem functioning and human interactions with the oceans. Field Chief Editor Carlos M. Duarte at King Abdullah University of Science and Technology Thuwal is supported by an outstanding Editorial Board of international researchers. This multidisciplinary open-access journal is at the forefront of disseminating and communicating scientific knowledge and impactful discoveries to researchers, academics, policy makers and the public worldwide.
With the human population predicted to reach 9 billion people by 2050, it is clear that traditional land resources will not suffice to meet the demand for food or energy, required to support high-quality livelihoods. As a result, the oceans are emerging as a source of untapped assets, with new innovative industries, such as aquaculture, marine biotechnology, marine energy and deep-sea mining growing rapidly under a new era characterized by rapid growth of a blue, ocean-based economy. The sustainability of the blue economy is closely dependent on our knowledge about how to mitigate the impacts of the multiple pressures on the ocean ecosystem associated with the increased scale and diversification of industry operations in the ocean and global human pressures on the environment. Therefore, Frontiers in Marine Science particularly welcomes the communication of research outcomes addressing ocean-based solutions for the emerging challenges, including improved forecasting and observational capacities, understanding biodiversity and ecosystem problems, locally and globally, effective management strategies to maintain ocean health, and an improved capacity to sustainably derive resources from the oceans.