Accurately predicting Arctic sea ice concentration (SIC), especially during the melting season at sub-seasonal scales, is essential for advancing our knowledge of global climate change. Currently, statistical models for SIC prediction face three major challenges: limited spatial resolution and forecast timeliness; inadequate representation of sea ice dynamic processes; and difficulties in predicting SIC during the melting season. To address these challenges, this study introduces super-resolution SIC Transformer (SR-SICFormer), a deep learning-based, data-driven model designed for fine-scale, long-term Arctic SIC prediction. We also propose a novel physical constraints loss function, ConIce loss, which integrates thermodynamic and dynamic sea ice processes into the training procedure, aiming to improve the model’s predictive accuracy. Satellite remote sensing data is used for training and validating the model. Experimental results demonstrate that SR-SICFormer outperforms traditional statistical and numerical models in extended-range SIC prediction tasks, achieving a 5× spatial super-resolution factor and a 15-day forecast period. The model achieves a root mean square error (RMSE) of 0.0504, a correlation coefficient (r) of 0.9420, a peak signal-to-noise ratio (PSNR) of 30.32 dB, and a structural similarity index measure (SSIM) of 0.9297. For the 2× super-resolution and 60-day forecast during the melting season, SR-SICFormer maintains strong performance, keeping SIC residuals within [-0.3, 0.3]. Additionally, the ConIce loss function effectively preserves both sea ice extent (SIE) and internal concentration distributions, ensuring that extended-range forecast results closely match ground truth in both SIE and concentration.

High-resolution (HR) sea ice concentration (SIC) is essential for polar ocean monitoring and Arctic research. Super-resolution (SR) techniques improve the spatial resolution of SIC data, but existing methods often fail to capture fine-scale structures, perform poorly at ice edges, and lack accuracy near sea-land boundaries. To address these challenges, we propose a Topography-Enhanced Multi-Level Residual Attention U-Net (TE-MRAU-Net) for SIC downscaling. To address these challenges, this paper proposes a topography-enhanced multi-level residual self-attention U-Net downscaling model (TE-MRAU-Net). TE-MRAU-Net integrates three innovative modules: the HR Topography Feature (HRTF) module, which introduces static topographic constraints to effectively improve reconstruction accuracy along sea-land boundaries; the Multi-Level Residual (MLRes) module, which enhances the model's ability to extract fine-scale sea ice features in super-resolution predictions; and the Spatial Attention Connector (SAC) module, which strengthens spatial modeling and structural consistency, particularly improving reconstruction performance in marginal sea ice edges and lower-latitude Arctic regions. Experimental results indicate that, in the pan-Arctic region, TE-MRAU-Net outperforms other models in various evaluation metrics and spatial distribution reconstruction. For the $5\times$ SR prediction task, the model achieves a mean absolute error (MAE) of 0.0049, a root mean square error (RMSE) of 0.0081, a correlation coefficient ($r$) of 0.9993, a peak signal-to-noise ratio (PSNR) of 40.21dB, a structural similarity index measure (SSIM) of 0.9764, a balanced accuracy (BACC) of 0.9428, a mean opinion score (MOS) of 4.63, and a subjective image quality assessment (SIQA) of 9.75, greatly enhancing the model's ability to capture detailed SIC features.

High-precision identification of acoustic convergence zones (CZs) and acoustic shadow zones (SZs) is a core prerequisite for deep-sea sonar performance prediction and long-range underwater target detection. However, in data-scarce marine environments, traditional acoustic identification methods suffer from high environmental sensitivity and significant computational costs, while pure data-driven deep learning methods face dilemmas such as a lack of physical consistency and poor generalization on small samples. To address these issues, a three-level cascaded recognition framework based on physics-prior-augmented deep learning is proposed in this paper, enabling accurate segmentation of CZs and intelligent classification of sound field types under data-scarce scenarios. In this framework, physical acoustic principles are incorporated exclusively as priors through a training dataset generated by a Gaussian beam acoustic propagation code (Bellhop) and through hand-crafted geometric features derived post hoc from the initial segmentation outputs. Taking a typical deep-sea area in the Northwest Pacific Ocean as the research object, a hybrid dataset comprising 5000 simulated transmission loss images and 500 simulated images from a geographically distinct sea area is constructed. The sound field is categorized into four types: strong convergence, usable convergence, weak convergence, and shadow zone. In the first stage, the ResNet-34 backbone is improved by integrating deformable convolution and a global statistical feature module, which, combined with a joint loss function, achieves high-precision pixel-level segmentation of CZs and SZs, with the regional gray contrast reaching 86.9%. In the second stage, a customized dual-channel VGG16 architecture is designed to fuse the extracted geometric priors and visual features, achieving a sound field classification accuracy of 89.91%. In the third stage, a hybrid data augmentation technique combining Mixup and convolutional autoencoder is adopted alongside a transfer learning strategy to mitigate the data scarcity under cross-domain conditions, boosting the small-sample classification accuracy to 84.45%. The experimental results demonstrate that the models in each stage of the proposed framework significantly outperform traditional methods and baseline networks. This study provides a novel methodology and technical support for intelligent sound field identification in data-scarce marine environments. Finally, the core contributions and current limitations are summarized, and future research directions, such as constructing a dynamic hydrological parameter feedback mechanism and identifying three-dimensional complex sound fields, are prospected.

Fine-scale detailed estimation of sea ice concentration (SIC) is pivotal for maritime safety, scientific exploration, and environmental surveillance. However, current datasets frequently present challenges due to their limited resolution, thereby hindering fine-scale analysis of sea ice conditions. This paper introduces a novel Deep Self-Attention Residual U-Net (Deep SARU-Net) architecture to address the limitations inherent in existing super-resolution estimation techniques. By harnessing distinctive multi-stage self-attention mechanisms, orthogonal rectangular convolutional kernels, and residual modules, this architecture significantly augments both the precision and generalizability of SIC super-resolution estimation tasks. Experimental results demonstrate that in the vicinity of the Chukchi Sea, the Deep SARU-Net method exhibits superior performance in terms of both RMSE and SSIM values compared to other models, showcasing its efficacy. Furthermore, generalization analyses across diverse sea regions confirm the model’s universality.

Accurate reconstruction of ocean subsurface temperature (ST) is fundamental for understanding ocean dynamics and quantifying climate change. However, existing methods still face three major limitations: inadequate modelling of temporal multiscale variability, incomplete representation of spatial multiscale structure, and inefficient exploitation of coupled multivariate information. Here we propose a deep learning framework, STMSNet, that integrates multi-source remote sensing observations including sea surface temperature (SST), sea level anomaly (SLA), and sea surface wind (SSW) to reconstruct three-dimensional ST. To more effectively capture temporal multiscale variability, we first design a Temporal Scale Mixer (TS Mixer) based on temporal convolution, which enhances the representation of time-dependent features across distinct timescales. Second, to improve spatial multiscale feature learning and fusion, we develop a Spatial Feature Fusion Module (SFFM) built upon a Hierarchical Atrous Spatial Pyramid Pooling (HASPP) module. Third, to better exploit the coupled information among oceanic variables, we introduce an adaptive multi-task learning (MTL) architecture that uses salinity reconstruction as an auxiliary task, providing complementary information for the primary temperature reconstruction. This design enhances the model's accuracy and reliability. Experiments demonstrate that STMSNet reliably reproduces the spatial distribution and vertical structure of ST. The combination of temporal–spatial multiscale modules with the adaptive MTL framework markedly strengthens the representation of complex ocean variability. Relative to representative machine learning and deep learning baselines, STMSNet achieves consistently higher accuracy, particularly in dynamically active regions including internal wave regions, coastal upwelling zones, and mesoscale eddy fields. Overall, the proposed spatiotemporal multiscale reconstruction framework offers a promising route towards high-precision estimation of ocean ST.