Ke Qu

Time-series machine learning models represented by long short-term memory (LSTM) networks provide an effective way to obtain high-precision sound speed profiles (SSPs) quickly and at low cost, which can meet the practical application requirements of underwater sonar systems. However, in sea areas with frequent strong internal solitary waves, the large-amplitude sound speed anomalies caused by them will seriously interfere with model learning in the form of strong outlier features, resulting in a sharp drop in SSP prediction accuracy and significant degradation of the generalization stability and robustness of the model. To address this problem, this paper proposes a time-series SSP prediction method based on a bidirectional long short-term memory (Bi-LSTM) network. First, Empirical Orthogonal Function (EOF) decomposition is used to realize the low-dimensional feature representation of SSPs, and then the bidirectional time-series feature capture capability of Bi-LSTM is used to predict the SSP sequence with large disturbances caused by strong internal solitary waves. Multiple groups of comparative experiments based on the measured temperature chain data in the continental slope area of the South China Sea show that the Bi-LSTM model has a significant improvement in prediction accuracy and robustness compared with the classical LSTM model. Among them, the Bi-LSTM model with EOF decomposition achieves a correlation coefficient of 0.995 and an average Root Mean Square Error (RMSE) as low as 0.387 m/s. Under the condition of internal solitary wave disturbance, the classical LSTM is difficult to effectively capture the large abrupt change in sound speed, while the proposed Bi-LSTM model can still achieve accurate prediction of the SSP in the disturbance section, and has both the feature recognition and evolution prediction capabilities for the strongly nonlinear internal solitary wave process. This method provides effective technical support for the rapid and large-scale reconstruction of the sound speed field under the disturbance of strong internal solitary waves.

Dissolved oxygen (DO) is a primary environmental regulator of shrimp physiology and behavior in intensive aquaculture systems. Whether shrimp acoustic emissions quantitatively reflect oxygen-driven behavioral modulation under operational pond conditions, however, remains uncertain due to the difficulty of isolating biologically relevant signals from complex soundscapes. In this study, passive acoustic monitoring was conducted in commercial outdoor ponds culturing Litopenaeus vannamei. A periodic-coding non-negative matrix factorization approach was applied to separate putative shrimp-associated acoustic components from broadband background noise and to obtain stable time–frequency representations of acoustic activity. Temporal variations in the extracted acoustic intensity were compared with simultaneously measured DO concentrations. Rather than relying on global correlation, phase-specific analyses revealed that the putative shrimp-associated acoustic component exhibited consistent positive associations with DO dynamics during both rising and declining phases, whereas background noise showed only weak and non-coherent relationships with DO. These results indicate that the observed acoustic–oxygen relationship is non-stationary and context-dependent. Given the observational nature of the study and potential confounding influences (e.g., aeration and other environmental factors), these findings, which are based on observations from a single pond over a limited recording period (62.85 h) under specific operational conditions, should be interpreted with caution and regarded as a proof-of-concept rather than evidence of general applicability. Nevertheless, the results are consistent with the hypothesis that population-level acoustic activity may reflect environmentally modulated behavioral responses. This highlights the potential of soundscape-based approaches as non-invasive tools for supporting aquaculture monitoring, while emphasizing the need for further validation under controlled and multi-site conditions.

Due to the limited number of deep sound speed profile (SSP) samples, the existing wide-area SSP inversion methods cannot estimate the full-depth SSP. In this paper, the full-depth SSP inversion is achieved by adding physical mechanism constraints to the neural network inversion algorithm. A dimensionality reduction approach for SSP perturbation, based on the hydrodynamic mechanism of seawater, is proposed. Constrained by the characteristics of ocean stratification, a self-organizing map is employed to invert the depth of the sound channel axis and reconstruct the SSP from the sea surface to the sound channel axis. The SSP from the sound channel axis to the seabed is reconstructed by integrating the characteristics of the sound channel axis and the sound speed gradient characteristics of the deep sea isothermal layer. The efficacy of the method was validated by the Argo data from the South China Sea. The average root mean square error of the reconstructed full-depth SSP is 2.85 m/s. Additionally, the average error of transmission loss prediction within 50 km is 2.50 dB. The proposed method is capable of furnishing effective full-depth SSP information without the necessity of any in situ measurements, thereby meeting the requirements of certain underwater acoustic applications.

Aiming at the problem of accuracy degradation caused by the mixing of spatiotemporal disturbance patterns in sound speed profile (SSP) inversion using the traditional geographic grid division method, this study proposes an acoustic stability pre-clustering framework that integrates principal component analysis and machine learning clustering. Disturbance mode principal component analysis is first used to extract characteristic parameters, and then a machine learning clustering algorithm is adopted to pre-classify SSP samples according to acoustic stability. The SSP inversion experimental results show that: (1) the SSP samples of the South China Sea can be divided into three clusters of disturbance modes with statistically significant differences. (2) The regression inversion method based on cluster attribution reduces the average error of SSP inversion for data from 2018 to 1.24 m/s, which is more than 50% lower than what can be achieved with the traditional method without pre-clustering. (3) Transmission loss prediction verification shows that the proposed method can produce highly accurate sound field calculations in environmental assessment tasks. The acoustic stability pre-clustering technology proposed in this study provides an innovative solution for the statistical modeling of marine environment parameters by effectively decoupling the mixed effect of SSP spatiotemporal disturbance patterns. Its error control level (<1.5 m/s) is 37% higher than that of the single empirical orthogonal function regression method, showing important potential in underwater acoustic applications in complex marine dynamic environments.

Sound speed profile (SSP) inversion based on remote sensing parameters allows for the acquisition of global quasi-real-time SSPs without the need for on-site measurements, thereby fulfilling the requirements of many acoustic applications. This study makes two enhancements to the single empirical orthogonal function regression (SEOF-R) method. First, the k-means clustering algorithm is utilized to cluster SSP samples, ensuring the consistency of perturbation modes in the physical regression. Second, baroclinic modes are employed to derive a novel SSP basis function, named the ocean mode basis, which accurately characterizes the inversion relationship. Validation experiments using data from the South China Sea yield promising results. Compared with the SEOF-R method, the reconstruction error of the improved approach is reduced by 27%, with an average reconstruction error of 1.73 m/s. The average prediction transmission loss error decreases by 70%, reaching 1.29 dB within 50 km. The grid-free processing and low sample dependence of the proposed method further enhance the applicability and accuracy of remote sensing-based SSP inversion.

The remote sensing-based inversion of sound speed profile (SSP) enables the acquisition of high-spatial-resolution SSP without in situ measurements. The spatial division of the inversion grid is crucial for the accuracy of results, determining both the number of samples and the consistency of inversion relationships. The result of our research is the introduction of a physics-inspired self-organizing map (PISOM) that facilitates SSP inversion by clustering samples according to the physical perturbation law. The linear physical relationship between sea surface parameters and the SSP drives dimensionality reduction for the SOM, resulting in the clustering of samples exhibiting similar disturbance laws. Subsequently, samples within each cluster are generalized to construct the topology of the solution space for SSP reconstruction. The PISOM method significantly improves accuracy compared with the SOM method without clustering. The PISOM has an SSP reconstruction error of less than 2 m/s in 25% of cases, while the SOM method has none. The transmission loss calculation also shows promising results, with an error of only 0.5 dB at 30 km, 5.5 dB smaller than that of the SOM method. A physical interpretation of the neural network processing confirms that physics-inspired clustering can bring better precision gains than the previous spatial grid.

Introduction: The trend of sound speed profile (SSP) inversion is towards wide-area sound speed estimation. However, the traditional inversion method of dividing the latitude and longitude grids has limitations in terms of significantly lower accuracy when samples are lacking. k-means clustering algorithm (K-means) can divide the training class to achieve high accuracy estimation.Method: This paper proposes a grid-free pre-classification inversion scheme based on empirical orthogonal function (EOF) vectors. The scheme is based on the K-means to classify the samples according to the perturbation mode of the SSP. After classification, the SSP inversion is carried out using the self-organizing map algorithm (SOM). The experimental sea area is selected from the South China Sea, and the inversion results are evaluated using root mean square error (RMSE) as the criterion.Result: The inversion results show that the inversion error is 2.1 m/s for the pre-classification solution and 2.7 m/s for the solution without pre-classification, a steady improvement of more than 20% in the inversion error. Accuracy is also improved by 2.14 m/s in the depth range where the sound speed perturbance is greatest.Discussion: This pre-classification scheme has smaller inversion errors and the classification results are reasonable in terms of distribution in time and space. It provides a feasible solution for SSP inversion in sea areas where samples are lacking.

Introduction: In underwater acoustic applications, the three-dimensional sound speed distribution has a significant impact on signal propagation. However, the traditional sound speed profile (SSP) measurement method requires a lot of manpower and time, and it is difficult to popularize. Satellite remote sensing can collect information on a large ocean surface area, from which the underwater information can be derived.Method: In this paper, we propose a method for reconstructing the SSP based on an extensible end-to-end tree boosting (XGBoost) model. Combined with satellite remote sensing data and Argo profile data, it extracts the characteristic matrix of the SSP and analyzes the contribution rate of each order matrix to reduce the introduction of noise. The model inverts the SSP above 1000 m in the South China Sea by using the root mean square error (RMSE) as the precision evaluation index.Result: The results showed that the XGBoost model could better reconstruct the SSP above 1000 m, with a RMSE of 1.75 m/s. Compared with the single empirical orthogonal function regression (sEOF-r) model of the linear regression method, the RMSE of the XGBoost model was reduced by 0.59 m/s.Discussion: For this model, the RMSE of the inversion results was smaller, the robustness was better, and the regression performance was superior to that of the sEOF-r model at different depths. This study provided an efficient tree boosting model for SSP reconstruction, which could reliably and instantaneously monitor the 3D sound speed distribution.

Given that spatiotemporal measurement of the subsurface profile over a wide range are difficult to obtain, surface observations from satellites are often used to estimate the sound speed profile (SSP). This paper proposes a multisource method based on the self-organizing map (SOM) to improve the estimation of the SSP by merging surface observations with satellite data. Surface observations from the Kuroshio Extension Observatory (KEO) were used to supplement satellite observations (anomalies in the measured sea level and sea surface temperature) to this end. Different combinations of the surface parameters were assessed, their errors were analyzed, and differences between the results before and after the multisource parameters were used are discussed. The proposed method significantly increased the accuracy of estimating the SSP when the parameters obtained from in situ measurements were used, with a root mean square error of 2.18 m/s, less than a third of the error obtained when only satellite observations were used. The proposed method provides a new approach to determining an accurate three-dimensional structure of the sound speed when various surface observations are available.

Sound speed profile (SSP) inversion is usually performed by linear statistical regression, such as the single empirical orthogonal function regression (sEOF-r) model. However, due to the complex dynamic activities of the ocean, the relationship between parameters is not strictly linear, often resulting in an unsatisfactory inversion result. In this study, an algorithm based on the random forest (RF) integrated learning model, for SSP inversion, was proposed. Using the sea surface temperature anomaly (SSTA) and sea surface height anomaly (SSHA) data, the sound speed profile of the upper 1000 m layer in the South China Sea was reconstructed, and its accuracy was evaluated through the root mean square error (RMSE). The accuracy of the evaluation demonstrated that the RF model proposed here could reconstruct the SSP in the upper 1000 m layer better than the sEOF-r can. Compared with the latter, the average reconstruction accuracy of the RF model was improved by 0.56 m/s. The linear regression of the sEOF-r model fell short of expectations in the regression between surface and subsurface parameters. By removing the constraints of linear inversion, the nonlinear regression of the RF model showed a smaller RMSE and better robustness in the reconstruction process and was superior to the sEOF-r model at all depths. As a result, it provided an effective integrated learning model for SSP reconstruction.