Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
4.4
2.7
Journals
JMSE
Special Issues
Ocean Climate: Deep Learning, Statistical Methods and Dynamical Modeling
share announcement

Ocean Climate: Deep Learning, Statistical Methods and Dynamical Modeling

A special issue of Journal of Marine Science and Engineering (ISSN 2077-1312). This special issue belongs to the section "Ocean and Global Climate".

Deadline for manuscript submissions: 5 July 2025 | Viewed by 3262

Special Issue Editors

Prof. Dr. Shaoqing Zhang

grade E-Mail Website
Guest Editor
Key Laboratory of Physical Oceanography, Ministry of Education, Ocean University of China, Qingdao, China
Interests: coupled modeling; coupled model data assimilation; weather-climate predictability; parameter estimation
Special Issues, Collections and Topics in MDPI journals
Prof. Dr. Yuxin Zhao

E-Mail Website
Guest Editor
College of Intelligent Systems Science and Engineering, Harbin Engineering University, Harbin, China
Interests: intelligent systems; information fusion; ocean information engineering
Special Issues, Collections and Topics in MDPI journals
Prof. Dr. Junyu Dong

E-Mail Website
Guest Editor
The College of Information Science and Engineering, Ocean University of China, Qingdao, China
Interests: ocean data science; machine learning; computer vision
Special Issues, Collections and Topics in MDPI journals
Dr. Hao Zuo

E-Mail Website
Guest Editor
European Centre for Medium-Range Weather Forecasts, Reading, UK
Interests: ocean data assimilation; ocean analysis; climate reanalysis
Special Issues, Collections and Topics in MDPI journals
Dr. Chang Liu

E-Mail Website
Guest Editor
The College of Marine Science and Engineering, Shandong University of Science and Technology, Qingdao, China
Interests: artificial intelligence oceanography; intelligent application of marine information

Special Issue Information

Dear Colleagues,

In this data blooming time, outstanding challenges are to be resolved: How to combine data-driven machine deep learning (MDL) with science-driven dynamical models to advance the science? How can MDL gain benefits as well as advance science and technology? Apparently, science-driven MDL is a heathy track. In fact, MDL did originate from our understanding of the natural world—mathematical modeling for the dynamics and physics and data sampling for the state—Bayes’ Theorem guiding combination of models and data.

MDL refers to the theory and development of a computer system that simulates the laws governing the evolution of the natural system. The essence is a programmed machine that learns from experiences for making decisions and forming new experiences. The core of this process is the statistical analysis of data and event-relevant condition judgments. The amount of data and analysis methods are the basis of successful MDL.

Nowadays, MDL is playing increasingly important roles in almost every field. Particularly, in fields such as marine science and engineering, which have a strong applied nature, MDL is a very powerful tool, and its development and application have become hot topics. On the one hand, the dynamical model-based reanalysis extends data sources for MDL. On the other hand, high-precision dynamical modeling needs MDL to break through the bottlenecks brought by nonlinear feedback in detailed scale interactions.

In this Special Issue, we call for papers that deal with recent advances in machine deep learning, statistical methods and dynamical modeling associated with research and development in marine science and engineering research and applications, including advanced machine learning algorithms, new ideas in statistical methods and dynamical models, data assimilation, MDL-induced physical parameterization, etc. We address the concept that the science-driven MDL development returns to help further our understanding on dynamics and physics, thus further advancing science and technology. Potential topics include, but are not limited to, the following:

  • Earth system modeling, data assimilation and parameter estimation;
  • Mesoscale and submesoscale ocean processes;
  • Bayes’ Theorem-based MDL algorithms;
  • Data assimilation-induced MDL algorithms;
  • MDL-induced new data assimilation algorithms;
  • MDL-induced parameterization and parameter estimation;
  • Advanced deep neural network algorithms and statistical methods;
  • MDL-induced climate and chemistry modeling;
  • Advanced multiscale MDL models;
  • MDL-driven cloud and micro-physics expressions;
  • Advanced dynamical modeling methods and schemes.

Prof. Dr. Shaoqing Zhang
Prof. Dr. Yuxin Zhao
Prof. Dr. Junyu Dong
Dr. Hao Zuo
Dr. Chang Liu
Guest Editors

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Journal of Marine Science and Engineering is an international peer-reviewed open access monthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2600 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.

Keywords

  • climate modeling
  • data assimilation
  • deep learning
  • artificial intelligence
  • ocean information engineering

Benefits of Publishing in a Special Issue

  • Ease of navigation: Grouping papers by topic helps scholars navigate broad scope journals more efficiently.
  • Greater discoverability: Special Issues support the reach and impact of scientific research. Articles in Special Issues are more discoverable and cited more frequently.
  • Expansion of research network: Special Issues facilitate connections among authors, fostering scientific collaborations.
  • External promotion: Articles in Special Issues are often promoted through the journal's social media, increasing their visibility.
  • e-Book format: Special Issues with more than 10 articles can be published as dedicated e-books, ensuring wide and rapid dissemination.

Further information on MDPI's Special Issue policies can be found here.

Published Papers (5 papers)

Download All Papers
Order results
Result details
Show export options expand_more Show export options expand_less
Select all
Export citation of selected articles as:

Research

20 pages, 5079 KiB  
Article
Research on the Wetland Vegetation Classification Method Based on Cross-Satellite Hyperspectral Images
by Min Yang, Jing Qin, Xiaodan Wang and Yanfeng Gu
J. Mar. Sci. Eng. 2025, 13(4), 801; https://doi.org/10.3390/jmse13040801 - 17 Apr 2025
Viewed by 128
Abstract
In recent years, the global commercial aerospace industry has flourished, witnessing a rapid surge in customized satellite services. Deep learning has emerged as a pivotal tool for accurately identifying wetland vegetation. However, hyperspectral remote sensing images are often plagued by varying degrees of [...] Read more.
In recent years, the global commercial aerospace industry has flourished, witnessing a rapid surge in customized satellite services. Deep learning has emerged as a pivotal tool for accurately identifying wetland vegetation. However, hyperspectral remote sensing images are often plagued by varying degrees of noise during acquisition, leading to subtle differences in spectral responses. Currently, vegetation classification models are tailored specifically for each hyperspectral sensor, making it challenging to generalize a model designed for one sensor to others. Furthermore, discrepancies in data distribution between training and test sets result in a notable decline in model performance, impeding model sharing across satellite hyperspectral sensors and hindering the interpretation of wetland scenes. Domain adaptation methods leveraging Generative Adversarial Networks (GANs) have been extensively researched and applied in the realm of cross-sensor land feature classification. Nevertheless, these data-level cross-domain classification strategies typically focus on band selection or alignment using relatively similar data to address image differences, without addressing spectral variability or incorporating pseudo-labels to enhance classification accuracy. Noise changes aggravate the distribution characteristics and model differences of vegetation in classification tasks. This has a negative impact on subsequent classification accuracy. To alleviate these problems, we have designed a linear unbiased stochastic network classification framework based on adversarial learning. The framework employs a style randomization algorithm to simulate spectral drift. It generates simulated images to enhance the model’s generalization ability. Supervised contrastive learning is utilized to prevent redundant learning of the same training images. Domain discrimination and domain-invariant characteristics are considered. We optimize the generator and discriminator using inter-class and intra-class contrast loss functions. The dual regularization training method is adopted, and non-redundant expansion is realized. It achieves similarity and addresses offsets. This method minimizes computational cost. Cross-sensor classification experiments were conducted, with comparative tests performed on a self-made wetland dataset. This method demonstrates significant advantages in wetland vegetation classification. According to the visualization results, our classification strategy can be used for cross-domain vegetation classification in coastal wetlands. It can also be applied to other small-satellite hyperspectral images and cross-satellite multispectral data, reducing on-site sampling costs and proving cost-effective. Full article
(This article belongs to the Special Issue Ocean Climate: Deep Learning, Statistical Methods and Dynamical Modeling)
►▼ Show Figures

Figure 1

21 pages, 4737 KiB  
Article
Study on the Impact of Input Parameters on Seawater Dissolved Oxygen Prediction Models
by Wenqing Li, Jing Lv, Yuhang Wang and Xiangfeng Kong
J. Mar. Sci. Eng. 2025, 13(3), 536; https://doi.org/10.3390/jmse13030536 - 11 Mar 2025
Viewed by 381
Abstract
The concentration of dissolved oxygen (DO) in seawater is a core ecological indicator in aquaculture, and its accurate prediction is of great value for the management of marine ranching. In response to the lack of exploration on the optimization mechanism of input parameters [...] Read more.
The concentration of dissolved oxygen (DO) in seawater is a core ecological indicator in aquaculture, and its accurate prediction is of great value for the management of marine ranching. In response to the lack of exploration on the optimization mechanism of input parameters in existing DO prediction studies, this study, based on observational data from the Goji Island marine ranching, constructed a technical framework of “parameter screening—model optimization—ecological analysis”. By integrating correlation analysis, principal component analysis (PCA), and multi-model comparison (SVM, MLP, and RF) methods, this study systematically revealed the input parameter optimization strategies and its ecological correlation mechanism. The research findings are as follows: (1) Parameter optimization can significantly improve model accuracy, and the model performance is optimal after eliminating the low-correlation parameter (Tur) (RMSE = 0.039, MAE = 0.030, R2 = 0.884). (2) The absence of key parameters (such as Sal) will lead to a significant decrease in prediction accuracy (the R2 reduction rate reaches 71.6%). (3) The parameter importance ranking is Tem > pH > Sal > Chl-a > Tur, among which Tem explains 42.3% of the variation in DO. The intelligent parameter optimization framework proposed in this study provides theoretical support for the development of a marine ranching DO monitoring system, and its technical path can be extended to the prediction of other water environment indicators. Future research will develop a parameter adaptive selection algorithm, conduct the dynamic monitoring of multi-scale environmental factors, and achieve the intelligent optimization and verification of model parameters. Full article
(This article belongs to the Special Issue Ocean Climate: Deep Learning, Statistical Methods and Dynamical Modeling)
►▼ Show Figures

Figure 1

24 pages, 8896 KiB  
Article
A Prediction of Estuary Wetland Vegetation with Satellite Images
by Min Yang, Bin Guo, Ning Gao, Yang Yu, Xiaoli Song and Yanfeng Gu
J. Mar. Sci. Eng. 2025, 13(2), 287; https://doi.org/10.3390/jmse13020287 - 4 Feb 2025
Viewed by 644
Abstract
Estuarine wetlands are the transition zone between marine, freshwater, and terrestrial ecosystems and are more ecologically fragile. In recent years, the spread of exotic vegetation, specifically Spartina alterniflora, in the Yellow River estuary wetlands has significantly encroached upon the habitats of native [...] Read more.
Estuarine wetlands are the transition zone between marine, freshwater, and terrestrial ecosystems and are more ecologically fragile. In recent years, the spread of exotic vegetation, specifically Spartina alterniflora, in the Yellow River estuary wetlands has significantly encroached upon the habitats of native species such as Phragmites australis, Suaeda glauca Bunge, and Tamarix chinensis Lour. With advances in land prediction modeling, predicting wetland vegetation distribution can aid management and decision-making for ecological restoration. We selected the core area as the study object and coupled the hydrological model MIKE 21 with the PLUS model to predict the potential future distribution of invasive and dominant species in the region. (1) Based on the fine classification results from satellite images of GF1/G2/G5, we gained an understanding of the changes in wetland vegetation types in the core area of the reserve in 2018 and 2020. (2) Using public data such as ERA5 and GEO as input for basic environmental data, using MIKE 21 to provide high-spatial-resolution hydrodynamic parameters for the PLUS model as an environmental driver, we modeled the spatial distribution of various wetland vegetation in the Yellow River estuary wetland in Dongying under different artificial restoration measures. (3) We predicted the 2022 distribution of typical vegetation in the region, used the classification results of GF6 as the actual distribution, compared the spatial distribution with the actual distribution, and obtained a kappa coefficient of 0.78; the predicted values of the model are highly consistent with the true values. This study combines the fine classification results of vegetation based on hyperspectral remote sensing, the construction of a coupled model, and the prediction effect of typical species, providing a reference for constructing and optimizing the vegetation prediction model of estuarine wetlands. It also allows scientific and effective decision-making for the management of ecological restoration of delta wetlands. Full article
(This article belongs to the Special Issue Ocean Climate: Deep Learning, Statistical Methods and Dynamical Modeling)
►▼ Show Figures

Figure 1

20 pages, 10179 KiB  
Article
Fusion of In-Situ and Modelled Marine Data for Enhanced Coastal Dynamics Prediction Along the Western Black Sea Coast
by Maria Emanuela Mihailov, Alecsandru Vladimir Chirosca and Gianina Chirosca
J. Mar. Sci. Eng. 2025, 13(2), 199; https://doi.org/10.3390/jmse13020199 - 22 Jan 2025
Viewed by 939
Abstract
This study explores the use of Temporal Fusion Transformers (TFTs), an AI/ML technique, to enhance the prediction of coastal dynamics along the Western Black Sea coast. We integrate in-situ observations from five meteo-oceanographic stations with modelled geospatial marine data from the Copernicus Marine [...] Read more.
This study explores the use of Temporal Fusion Transformers (TFTs), an AI/ML technique, to enhance the prediction of coastal dynamics along the Western Black Sea coast. We integrate in-situ observations from five meteo-oceanographic stations with modelled geospatial marine data from the Copernicus Marine Service. TFTs are employed to refine predictions of shallow water dynamics by considering atmospheric influences, with a particular focus on wave-wind correlations in coastal regions. Atmospheric pressure and temperature are treated as latitude-dependent constants, with specific investigations into extreme events like freezing and solar radiation-induced turbulence. Explainable AI (XAI) is exploited to ensure transparent model interpretations and identify key influential input variables. Data attribution strategies address missing data concerns, while ensemble modelling enhances overall prediction robustness. The models demonstrate a significant improvement in prediction accuracy compared to traditional methods. This research provides a deeper understanding of atmosphere-marine interactions and demonstrates the efficacy of Artificial intelligence (AI)/Machine Learning (ML) in bridging observational and modelled data gaps for informed coastal zone management decisions, essential for maritime safety and coastal management along the Western Black Sea coast. Full article
(This article belongs to the Special Issue Ocean Climate: Deep Learning, Statistical Methods and Dynamical Modeling)
►▼ Show Figures

Figure 1

19 pages, 2084 KiB  
Article
An End-to-End Ocean Environmental Noise Anomaly Detection Framework Combining Time–Frequency Information and Expert Gating
by Libin Du, Mingyang Liu, Zhichao Lv, Chuanhe Tan, Junkai He and Fei Yu
J. Mar. Sci. Eng. 2025, 13(1), 141; https://doi.org/10.3390/jmse13010141 - 15 Jan 2025
Viewed by 719
Abstract
The detection and optimization of ocean environmental noise anomalies play a crucial role in enhancing the safety of marine engineering applications and ecological protection. Current anomaly detection methods for ocean environmental noise often suffer from issues of accuracy and robustness. To address these [...] Read more.
The detection and optimization of ocean environmental noise anomalies play a crucial role in enhancing the safety of marine engineering applications and ecological protection. Current anomaly detection methods for ocean environmental noise often suffer from issues of accuracy and robustness. To address these challenges, this paper first proposes an end-to-end framework that combines time–frequency information and expert gating, significantly improving the precision of noise sequence generation. Secondly, a Gamma distribution-based residual analysis method for anomaly detection is designed, enhancing the robustness of anomaly detection. Finally, an anomaly optimization module is developed to improve data quality, enabling efficient noise anomaly detection and optimization. Our experimental results demonstrate that the proposed model significantly outperforms traditional models in multi-frequency noise prediction, with strong robustness in anomaly detection and high generalization performance. The proposed framework offers a novel approach for analyzing the causes of noise anomalies and optimizing models. Additionally, the research outcomes provide efficient technical support for deep-sea exploration, equipment optimization, and environmental protection. Full article
(This article belongs to the Special Issue Ocean Climate: Deep Learning, Statistical Methods and Dynamical Modeling)
►▼ Show Figures

Figure 1

Show export options expand_more Show export options expand_less
Select all
Export citation of selected articles as:
 
Displaying articles 1-5
Back to TopTop