Satellite-derived sea surface remote sensing data enable large-scale subsurface sound speed profile (SSP) inversion; however, considerable inversion errors commonly arise due to the lack of in situ constraints. To address this limitation, this Letter proposes a bidirectional long short-term memory-based framework that integrates sparse observations to improve inversion accuracy. Experiments in the Kuroshio Extension and the Philippine Sea demonstrate that sparse observations yield improved inversion accuracy, with remarkable improvement without sea surface data. Depth-layer sensitivity analysis reveals that optimal sparse sampling depth intervals differ under scenarios with and without sea surface data, offering guidance for rational underwater observation deployment.

Efficient and high-precision underwater acoustic field prediction is crucial for underwater target detection, autonomous vehicle path planning, and other naval applications. Traditional numerical models suffer from high computational complexity. This study proposes a hybrid physics-based and data-driven approach for low-frequency underwater acoustic field prediction. First, the convolutional autoencoder is constructed to extract bathymetric features. These, along with source depth, are input into a convolutional neural network to predict range-dependent modal coefficients, which are integrated with normal-mode theory to compute the acoustic field. Finally, the residual network further refines the prediction. Using coupled-mode solutions as ground truth, the proposed neural network achieves error improvement margins of 1.0-3.0 dB in shallow-water environments, and 1.0-5.0 dB in deep-sea conditions, compared to adiabatic solutions. Relative to end-to-end neural network baselines, the present method delivers performance improvements of approximately 1.5 dB in shallow-water scenarios, and approximately 0.7 dB in deep-sea environments, with particularly enhanced performance at 25 Hz. For low-frequency acoustic field computation in deep-sea settings, the neural network demonstrates a 180-200 times computational speedup in modal coefficients calculation over numerical models, significantly enhancing the efficiency of acoustic field prediction.

This study introduces a convolutional neural network based method for rapid prediction of underwater acoustic propagation fields, addressing the high computational cost of traditional methods. By analyzing regional terrain features and constructing a training dataset, the model learns acoustic transmission loss patterns across various terrain conditions. Tests in the Western Pacific demonstrate a root mean square error of 3.48 dB for non-smoothed fields, with an average prediction time of 1.95 ms per batch (10 samples). This method highlights the potential for fast acoustic propagation predictions using simplified inputs, offering a promising direction for real-time applications.

Complex island-reef topography significantly affects underwater acoustic propagation by increasing transmission loss, expanding acoustic shadow zones, and reducing coherence, which collectively degrade underwater detection performance. Taking the Xisha (Paracel) Islands as a representative environment, this study employs twodimensional and three-dimensional BELLHOP ray models to analyze the influence of buoy deployment depth and position on the acoustic field distribution and detection coverage. On this basis, a distributed buoy depth-diversity optimization method combining acoustic field complementarity with a genetic algorithm (GA) is proposed. Simulation results show that the proposed method increases the effective detection ratio from 12.61% to 31.17%, effectively expanding the detection range. Furthermore, coherence loss induced by complex topography reduces the array gain by approximately 1.5 dB; neglecting this effect leads to an overestimation of the detection range by about 8%.

The marine environment during typhoons significantly disturbs sea surface boundary conditions, mixed layer structure, and ambient noise, thereby influencing the acoustic field structure and operational depth selection of underwater platforms. This study takes Typhoon Mawar (2023) as the research background and integrates environmental data, BELLHOP acoustic propagation simulations, and noise field modeling to characterize the variation of underwater acoustic conditions.Based on these acoustic characteristics, an optimization framework incorporating game-theoretic methods is developed to determine the depth strategy of the underwater platform. The resulting optimization problem is solved using linear programming. The results show that the optimal depth range for the underwater platform lies in the mid-depth sound-shadow layer before and after the typhoon's passage. During the typhoon, the combined effects of mixed layer deepening and wind-generated noise significantly reduce the passive detection probability in both the mid-water and shallow high-noise layers.Monte Carlo simulations further validate the effectiveness of the optimized strategy. Compared with the uniform depth distribution strategy, the optimization strategy based on the Nash equilibrium solution can significantly reduce the passive detection probability of the underwater platform.

The subsurface temperature profile plays a vital role in sound speed profiling and acoustic propagation. To obtain accurate subsurface temperature profiles without relying on sea surface data, this paper proposes the application of the Transformer model, a deep learning architecture, for the reconstruction of subsurface temperature profile. The approach leverages in-situ Argo observations, utilizes the self-attention mechanism of the Transformer to capture the complex nonlinear relationships within the subsurface temperature profile, thereby reconstructing the subsurface temperature profile. To address the relatively large reconstruction errors in complex oceanic environments, the paper introduces a strategy that integrates sparse in-situ observations into the model training process. By leveraging the self-attention mechanism, the model is guided to focus on temperature variations in the vicinity of known observation points, thereby improving reconstruction accuracy. The experimental results in both the Kuroshio Extension and the Philippine Sea demonstrate that the proposed method effectively enhances the reconstruction accuracy of the subsurface temperature profile. Furthermore, this paper also analyzes the impact of fusing sparse observations from specific depth layers on the performance of the reconstructed results.

Underwater acoustic propagation is influenced by water column properties, seabed topography, and source frequency, with existing numerical models exhibiting varied performance across different conditions. This study evaluates the frequency adaptability of three acoustic models—BELLHOP (geometric ray-based), RAM (parabolic equation), and KRAKEN (coupled mode)—under diverse seabed topographies, including deep-sea-flat (25–2000 Hz), shallow-sea-flat (25–10000 Hz), and gentle/steep-slope seabed (25–800 Hz). Flat seabed scenarios use the Scooter model as a benchmark, while sloping seabed scenarios are compared against analytical solutions. Results indicate that in a 200 m deep flat ocean environment, BELLHOP achieves high accuracy for frequencies above 200 Hz, KRAKEN performs comparably to RAM below 50 Hz, and RAM excels below 200 Hz. In a 4000 m deep flat ocean, RAM outperforms at frequencies below 100 Hz, while BELLHOP performs well above 100 Hz. For sloping seabed environments with slopes less than 6.5°, RAM demonstrates stability below 100 Hz, while BELLHOP performs better above 100 Hz; for slopes greater than 6.5°, RAM remains stable below 50 Hz, with BELLHOP outperforming above 50 Hz. KRAKEN is found unsuitable for sloping seabed simulations. These findings provide quantitative guidance for selecting acoustic models based on frequency and seabed topography.

In the sound propagation experiment conducted in the South China Sea, a comparison between measured and simulated data revealed differences in the sensitivity of various acoustic models to terrain characteristics under identical conditions. To systematically analyze this sensitivity, the models Bellhop, RAM, and KRAKEN were compared against the Couple model under different terrain conditions at receiving depths of 100m and 300m. The results showed that KRAKEN significantly reduces multiple reflections from the seabed and sea surface in sharply undulating terrain. Error analysis indicated that for errors in terrain location, height, and angle, the average errors of Bellhop and RAM at a receiving depth of 10 0m were 0.67dB, 1.5dB, and 1.33dB lower, respectively, than those at 300m. Furthermore, analysis within the frequency range of 0-500Hz demonstrated that RAM exhibits better stability than Bellhop at receiving depths of 0-200m. This study provides theoretical support for sound propagation simulation under complex terrain and model selection in realistic marine environments..

Ocean Wireless Sensor Networks (referred to as Ocean Sensor Networks, OSNs) are an essential component and technical means of three-dimensional ocean sensing and monitoring systems. They utilize the capabilities of information perception and interaction among nodes to acquire the required data within maritime areas, serving as a critical tool for marine resource exploitation and management. In OSNs, the collected data are meaningful only when they are geo-located. In this case, the importance of localization in the context of OSNs is well-established. However, the limitations of the acoustic communication underwater and complex ocean environment take adverse impacts on accurate localization. Therefore, an upsurge of interest has been made for localization in OSNs. To conclude the recent achievements, this paper takes a brief survey on localization in OSNs.

The transmission loss (TL) matrix is vital in underwater acoustics, ocean exploration,[1] and sonar design, yet traditional numerical methods like ray tracing, finite element methods, and parabolic equations struggle with the high computational complexity and reliance on prior environmental parameters in complex marine settings, especially during large-scale parameter sweeps or Monte Carlo simulations where costs escalate exponentially, limiting their practicality. Diffusion models, an emerging generative approach excelling in image and speech synthesis via progressive denoising, adeptly capture dynamic variations in high-dimensional data.[2-3] This study introduces a diffusion-based method to predict TL matrices in large-depth regions, using Bellhop-generated TL data to train the model on implicit distributions tied to source depth, frequency, and seabed topography, enhanced by condition embeddings and attention mechanisms. Experimental results show it efficiently generates accurate TL matrices across multiple source depths with significantly reduced computational cost, providing an innovative solution for rapid acoustic field computation, design, and optimization.

The variability in seafloor topography is a primary factor influencing sound propagation. Complex and varied seabed topographies generate differing degrees of sound propagation effects, thereby altering the sound propagation loss compared to a flat seabed. This paper analyzes and statistically processes the slope data of the selected area's terrain. It utilizes acoustic models to simulate and compare the sound propagation loss in convergence zones and near the deep-sea sound channel axis under different slope conditions. The findings indicate that an upslope terrain reduces propagation loss in convergence zones, and the propagation loss is minimized near the deep-sea sound channel axis when the downslope gradient is 7 degrees. The paper also examines and discusses the impact of different bottom sediment parameters on sound propagation loss and the backscattering of acoustic energy in upslope conditions. Finally, the study analyzes the horizontal refraction effect in the three-dimensional sound field under complex terrain conditions from ray trajectories perspective.

In the context of acoustics, underwater sound propagation is intricately influenced by conditions of the marine environment. These conditions dynamically alter the sound speed profile (SSP), introducing a high level of variability and uncertainty into SSP predictions. Such uncertainties are incorporated into the acoustical model, affecting the accuracy of forecasting sound propagation characteristics, including ray trajectory and propagation loss. This study introduces and applies a stochastic ray model designed to estimate the ray trajectory and propagation loss within the oceanic environment. We specifically examine the effects of SSP changes on sound propagation. Our approach integrates ray theory with an adaptive dynamic orthogonal differential equation to model stochastic evolution fields. Initially, this research utilizes SSPs gathered from the Philippine Sea and Kuroshio Extension region. We investigates the impact of SSP perturbation on the ray trajectory and propagation loss. Following this preliminary analysis, we present a stochastic ray model tailored for rapid estimation of sound field characteristics, premised on effects of the SSP disturbance. The proposed model achieves high accuracy while substantially diminishing the computational complexity compared with the Bellhop model.

Low-frequency vibrational modes in infrared (IR) and Raman spectra, often termed molecular fingerprints, are sensitive probes of subtle structural changes and chemical interactions. However, their inherent weakness and susceptibility to environmental interference make them challenging to detect and analyze. To tackle this issue, we developed a deep learning denoising protocol based on an attention-enhanced U-net architecture. This model leverages the inherent correlations between high- and low-frequency vibrational modes within a molecule, effectively reconstructing low-frequency spectral features from their high-frequency counterparts. We demonstrate the effectiveness of this method by recovering low-frequency signals of trans-1,2-bis(4-pyridyl)ethylene (BPE) adsorbed on an Ag surface, a representative system for surface enhancement Raman spectroscopy (SERS). Notably, the trained model exhibits promising transferability to SERS spectra acquired under different surface and external field conditions. Furthermore, we applied this method to experimental IR and Raman spectra of BPE, achieving high-quality, low-frequency spectral recovery.

This article introduces an iterative technique aimed at jointly estimating target location and sound propagation speed in a 3-D underwater environment characterized by an isogradient sound speed profile (SSP). The developed algorithm is based on the modified polar representation (MPR) framework and offers a unified solution for underwater near and far-field target localization. This solution enables precise identification of the source position in the near field and estimation of the direction of arrival (DOA) in the far field. Prior MPR-based solutions typically assume sound propagation along a straight path at a known, constant speed within the medium, an assumption which fails in complex underwater environments, resulting in excessive distance estimation bias and thus impacting positioning accuracy. Our article resolves this concern by introducing compensation for the stratification effect of the SSP. The proposed algorithm first employs a constrained weighted least squares (WLSs) approach, incorporating fixed weight matrices and constraint terms to generate an initial solution. Subsequently, through stratification compensation and iterative refinement, the accuracy of source localization is enhanced. Extensive simulations, conducted using BELLHOP, demonstrate that the positioning algorithm can attain Cramer-Rao lower bound (CRLB) accuracy under moderate noise levels. Its superior performance in challenging underwater environments is further validated through comparative analysis with several existing algorithms.

Underwater temperature profiles hold significant importance in contemporary ocean research and development. There exists a temporal and spatial connection between Sea Surface Temperature (SST) and underwater temperature profiles. Utilizing this spatio-temporal relationship, we propose a predictive reconstruction based on satellite remote sensing data, along with its visualization software. Our model, based on the Convolutional Neural Network (CNN), uses historical SST and Sea Level Anomaly (SLA) data from satellite remote sensing to predict the corresponding SST and SLA data for the next ten days. Furthermore, a model combining Empirical Orthogonal Functions (EOF) with a Deep Evidence Regression Network (DERN) utilizes historical underwater temperature profiles and SST information to train a reconstruction model. This establishes a relationship between surface and underwater temperatures. The predicted SST and SLA data for the next ten days, obtained using the prediction model, are used to reconstruct underwater temperature information. The predictive reconstruction module is visualized and integrated into the Earth visualization software, facilitating a more intuitive observation of the global predictive reconstruction results.

The sound speed profile (SSP) perturbations by oceanic dynamical processes are transferred to the sound field via acoustical models, introducing computational inaccuracies. Hence, an accurate estimation of the SSP holds significant importance. This letter introduces a method for SSP estimation based on the multipath delay structure of large depth vertical arrays. This approach capitalizes on the sensitivity of array's multipath delay structure to SSP perturbations and estimates SSP using a single explosive charge signal, thus minimizing the reliance on extensive acoustic data. The method primarily comprises three components: the establishment of a sound speed perturbation model, construction of a synthetic dataset, and parameter estimation based on the genetic algorithm. This letter employs the proposed model to generate foundational samples and maximizes the coverage of authentic SSPs, thereby enhancing the precision of SSP estimation. The acoustic observational SSP is obtained from deep-sea regions of the Western Pacific to compare with estimated SSP, the result of traditional method and publicly available online datasets. The results indicate that the method proposed in this letter achieves commendable accuracy, and the root-mean-square error (RMSE) between the estimated SSP and the observed SSP is only 1.3 m/s.

In this paper, a modified high-efficiency Convolutional Neural Network (CNN) with a novel Supervised Contrastive Learning (SCL) approach is introduced to estimate direction-of-arrival (DOA) of multiple targets in low signal-to-noise ratio (SNR) regimes with uniform linear arrays (ULA). The model is trained using an on-grid setting, and thus the problem is modeled as a multi-label classification task. Simulation results demonstrate the robustness of the proposed approach in scenarios with low SNR and a small number of snapshots. Notably, the method exhibits strong capability in detecting the number of sources while estimating their DOAs. Furthermore, compared to traditional CNN methods, our refined efficient CNN significantly reduces the number of parameters by a factor of sixteen while still achieving comparable results. The effectiveness of the proposed method is analyzed through the visualization of latent space and through the advanced theory of feature learning.

In this paper, the combination of the deep evidential regression network (DERN) and empirical orthogonal function (EOF) methods, namely, the DERN-EOF method, was proposed to reconstruct the subsurface temper-ature profiles. Firstly, the DERN method was compared with the previous machine learning methods used in the subsurface temperature field reconstruction, for example the Self-organizing map method. The new network was shown to have certain advantages of higher accuracy and the capability to provide uncertainty estimation, which enables the reconstruction of the subsurface field with higher belief. Analysis of the method performance sug-gests that the DERN could improve reconstruction accuracy by almost 20% and that the uncertainty statistics agrees well with the data distribution. Then the DERN method combined with the EOF method was used to reconstruct subsurface temperature profiles at a global scale. Results suggest that the reconstruction error in highly dynamic regions, like the Kuroshio Extension and Gulf Stream regions, is significantly higher than that in other regions, and the uncertainties increase with the reconstruction error, suggesting that the proposed method can provide a relatively believable uncertainty estimation.

Underwater acoustic localization (UWAL) is extremely challenging due to the multipath nature of extreme underwater environments, the sensor position uncertainty caused by unpredictable ocean currents, and the lack of underwater observation data due to sparse array, which all affect localization performance. Addressing these issues, this paper proposes a simple and effective underwater acoustic localization method using the time difference of arrival (TDOA) measurements based on the multipath channel effect of the underwater environment. By introducing the calibration source, localization performance was improved, and the sensor position error was corrected. The Cramér–Rao lower bound (CRLB) was derived, and the proposed method was able to achieve the CRLB with small deviation. Numerical simulations confirm the improved performance of the proposed method, including (1) a 20 dB and 30 dB reduction in the CRLB for far and near source scenarios, respectively, indicating improved accuracy and reliability when estimating unknown sources; (2) better Mean Squared Error (MSE) performance compared to existing methods and an efficiency of over 90% in low noise and above 80% in moderate noise in several scenarios, with a delayed threshold effect; and (3) achieving CRLB performance with only three sensors in a 3D space, even under moderate noise, while existing methods require at least five sensors for comparable performance. Our results demonstrate the efficacy of the proposed method in enhancing the accuracy and efficiency of source localization.