The ocean conditions beneath the ice cover play a key role in understanding the sea ice mass balance in the polar regions. An integrated high-frequency ice-ocean observation system, including Acoustic Doppler Velocimeter, Conductivity-Temperature-Depth Sensor, and Sea Ice Mass Balance Array (SIMBA), was deployed in the landfast ice region close to the Chinese Zhongshan Station in Antarctica. A sudden ocean warming of 0.14℃ (p < 0.01) was observed beneath early-frozen landfast ice, from (−1.60 ± 0.03)℃ during April 16–19 to (−1.46 ± 0.07)℃ during April 20–23, 2021, which is the only significant warming event in the nearly 8-month records. The sudden ocean warming brought a double rise in oceanic heat flux, from (21.7 ± 11.1) W/m² during April 16–19 to (44.8 ± 21.3) W/m² during April 20–23, 2021, which shifted the original growth phase at the ice bottom, leading to a 2 cm melting, as shown from SIMBA and borehole observations. Simultaneously, the slowdown of ice bottom freezing decreased salt rejection, and the daily trend of observed ocean salinity changed from +0.02 d⁻¹ during April 16–19, 2021 to +0.003 d⁻¹ during April 20–23, 2021. The potential reasons are increased air temperature due to the transit cyclones and the weakened vertical ocean mixing due to the tide phase transformation from semi-diurnal to diurnal. The high-frequency observations within the ice-ocean boundary layer enhance the comprehensive investigation of the ocean’s influence on ice evolution at a daily scale.

Sea ice surface roughness (SIR) affects the energy transfer between the atmosphere and the ocean, and it is also an important indicator for sea ice characteristics. To obtain a small-scale SIR with high spatial resolution, a novel method is proposed to retrieve SIR from Sentinel-1 synthetic aperture radar (SAR) images, utilizing an ensemble learning method. Firstly, the two-dimensional continuous wavelet transform is applied to obtain the spatial information of sea ice, including the scale and direction of ice patterns. Secondly, a model is developed using the Adaboost Regression model to establish a relationship among SIR, radar backscatter and the spatial information of sea ice. The proposed method is validated by using the SIR retrieved from SAR images and comparing it to the measurements obtained by the Airborne Topographic Mapper (ATM) in the summer Beaufort Sea. The determination of coefficient, mean absolute error, root-mean-square error and mean absolute percentage error of the testing data are 0.91, 1.71 cm, 2.82 cm, and 36.37%, respectively, which are reasonable. Moreover, K-fold cross-validation and learning curves are analyzed, which also demonstrate the method’s applicability in retrieving SIR from SAR images.

Accurate simulation of the evolution of freak waves by the wave phase focusing method requires accurate linear and nonlinear properties, especially in deep-water conditions. In this paper, we analyze the ability to simulate deep-water focused waves of a two-layer Boussinesq-type (BT) model, which has been shown to have excellent linear and nonlinear performance. To further improve the numerical accuracy and stability, the internal wave-generated method is introduced into the two-layer Boussinesq-type model. Firstly, the sensitivity of the numerical results to the grid resolution is analyzed to verify the convergence of the model; secondly, the focused wave propagating in two opposite directions is simulated to prove the symmetry of the numerical results and the feasibility of the internal wave-generated method; thirdly, the limiting focused wave condition is simulated to compare and analyze the wave surface and the horizontal velocity of the profile at the focusing position, which is in good agreement with the measured values. Meanwhile the simulation of focused waves in very deep waters agrees well with the measured values, which further demonstrates the capability of the two-layer BT model in simulating focused waves in deep waters.

Quality control (QC) is an essential procedure in scatterometer wind retrieval, which is used to distinguish good-quality data from poor-quality wind vector cells (WVCs) for the sake of wind applications. The current wind processor of the China-France Oceanography Satellite (CFOSAT) scatterometer (CSCAT) adopts a maximum likelihood estimator (MLE)-based QC method to filter WVCs affected by geophysical noise, such as rainfall and wind variability. As the first Ku-band rotating fan-beam scatterometer, CSCAT can acquire up to 16 observations over a single WVC, giving abundant information with diverse incidence/azimuth angles, as such its MLE statistical characteristics may be different from the previous scatterometers. In this study, several QC indicators, including MLE, its spatially averaged value (MLE_m), and the singularity exponents (SE), are analyzed using the collocated Global Precipitation Mission rainfall data as well as buoy data, to compare their sensitivity to rainfall and wind quality. The results show that wind error characteristics of CSCAT under different QC methods are similar to those of other Ku-band scatterometers, i.e., SE is more suitable than other parameters for the wind QC at outer-swath and nadir regions, while MLE_m is the best QC indicator for the sweet region WVCs. Specifically, SE is much more favorable than others at high wind speeds. By combining different indicators, an improved QC method is developed for CSCAT. The validation with the collocated buoy data shows that it accepts more WVCs, and in turn, improves the quality control of CSCAT wind data.

The mesoscale eddy (ME) has a significant influence on the convergence effect in deep-sea acoustic propagation. This paper use statistical approaches to express quantitative relationships between the ME conditions and convergence zone (CZ) characteristics. Based on the Gaussian vortex model, we construct various sound propagation scenarios under different eddy conditions, and carry out sound propagation experiments to obtain simulation samples. With a large number of samples, we first adopt the unified regression to set up analytic relationships between eddy conditions and CZ parameters. The sensitivity of eddy indicators to the CZ is quantitatively analyzed. Then, we adopt the machine learning (ML) algorithms to establish prediction models of CZ parameters by exploring the nonlinear relationships between multiple ME indicators and CZ parameters. Through the research, we can express the influence of ME on the CZ quantitatively, and achieve the rapid prediction of CZ parameters in ocean eddies. The prediction accuracy (R) of the CZ distance (mean R: 0.9815) is obviously better than that of the CZ width (mean R: 0.8728). Among the three ML algorithms, Gradient Boosting Decision Tree has the best prediction ability (root mean square error (RMSE): 0.136), followed by Random Forest (RMSE: 0.441) and Extreme Learning Machine (RMSE: 0.518).

The Stokes production coefficient (E₆) constitutes a critical parameter within the Mellor-Yamada type (MY-type) Langmuir turbulence (LT) parameterization schemes, significantly affecting the simulation of turbulent kinetic energy, turbulent length scale, and vertical diffusivity coefficient for turbulent kinetic energy in the upper ocean. However, the accurate determination of its value remains a pressing scientific challenge. This study adopted an innovative approach by leveraging deep learning technology to address this challenge of inferring the E₆. Through the integration of the information of the turbulent length scale equation into a physical-informed neural network (PINN), we achieved an accurate and physically meaningful inference of E₆. Multiple cases were examined to assess the feasibility of PINN in this task, revealing that under optimal settings, the average mean squared error of the E₆ inference was only 0.01, attesting to the effectiveness of PINN. The optimal hyperparameter combination was identified using the Tanh activation function, along with a spatiotemporal sampling interval of 1 s and 0.1 m. This resulted in a substantial reduction in the average bias of the E₆ inference, ranging from O(10¹) to O(10²) times compared with other combinations. This study underscores the potential application of PINN in intricate marine environments, offering a novel and efficient method for optimizing MY-type LT parameterization schemes.

Synthetic aperture radar (SAR) and wave spectrometers, crucial in microwave remote sensing, play an essential role in monitoring sea surface wind and wave conditions. However, they face inherent limitations in observing sea surface phenomena. SAR systems, for instance, are hindered by an azimuth cut-off phenomenon in sea surface wind field observation. Wave spectrometers, while unaffected by the azimuth cutoff phenomenon, struggle with low azimuth resolution, impacting the capture of detailed wave and wind field data. This study utilizes SAR and surface wave investigation and monitoring (SWIM) data to initially extract key feature parameters, which are then prioritized using the extreme gradient boosting (XGBoost) algorithm. The research further addresses feature collinearity through a combined analysis of feature importance and correlation, leading to the development of an inversion model for wave and wind parameters based on XGBoost. A comparative analysis of this model with ERA5 reanalysis and buoy data for of significant wave height, mean wave period, wind direction, and wind speed reveals root mean square errors of 0.212 m, 0.525 s, 27.446°, and 1.092 m/s, compared to 0.314 m, 0.888 s, 27.698°, and 1.315 m/s from buoy data, respectively. These results demonstrate the model’s effective retrieval of wave and wind parameters. Finally, the model, incorporating altimeter and scatterometer data, is evaluated against SAR/SWIM single and dual payload inversion methods across different wind speeds. This comparison highlights the model’s superior inversion accuracy over other methods.

The distribution of oil and gas resources is intricately connected to the underlying structure of the lithosphere. Therefore, investigating the characteristics of lithospheric thickness and its correlation with oil and gas basins is highly important. This research utilizes recently enhanced geological–geophysical data, including topographic, geoid, rock layer thickness, variable rock layer density, and interface depth data. Employing the principles of lithospheric isostasy and heat conduction, we compute the laterally varying lithospheric thickness in the China seas and adjacent areas. From these results, two pivotal parameters for different types of oil and gas basins were statistically analyzed: the minimum lithospheric thickness and the relative fluctuation in lithospheric thickness. A semiquantitative analysis was used to explore the connection between these parameters and the hydrocarbon abundance within the oil and gas basins. This study unveils distinct variations in lithospheric thickness among basins, with oil and gas rich basins exhibiting a thicker lithosphere in the superimposed basins of central China and a thinner lithosphere in the rift basins of eastern China. Notably, the relative fluctuations in lithospheric thickness in basins demonstrate significant disparities: basins rich in oil and gas often exhibit greater thickness fluctuations. Additionally, in the offshore basins of China, a conspicuous negative linear correlation is observed between the minimum lithospheric thickness and the relative fluctuation in lithospheric thickness. This study posits that deep-seated thermal upwelling results in lithospheric undulations and extensional thinning in oil and gas basins. Concurrently, sustained deep-seated heat influences sedimentary materials in basins, creating favorable conditions for oil and gas generation. The insights derived from this study contribute to a quantitative understanding of the intricate relationships between deep lithospheric structures and oil and gas basins. These findings provide valuable guidance for future oil and gas exploration in the studied areas.

Multiple borehole samples are collected from the Baiyun Depression in deep-water area of the northern South China Sea (SCS) in an effort to reconstruct transgression processes during the Paleogene based on palynalgal analysis. This study indicates that the Baiyun Depression generated a large group of palynopore assemblages and fluvial/lacustrine-related algae during the early and middle Eocene when the Wenchang Formation was deposited. The entire depression was dominated by fluvial and lacustrine facies before transgression. Its eastern and southeastern sags transitioned to shallow marine environment by generating a large abundance of marine dinoflagellates during the Enping deposition of the late Eocene. Meanwhile, the southern uplift zone simply yielded fluvial/lacustrine-related palynopores and algae, and was dominated by the fluvial and lacustrine environment during the early stage of the Enping Formation, prior to shifting into transitional setting in the later period. Northwestern sags remained extensive fluvial and delta facies without existence of marine dinoflagellates. It was until the depositional stage of the Zhuhai Formation (Oligocene) that the overall depression was strongly impacted from transgression process. Both eastern and southeastern sags were mainly under deep marine setting on a continental slope while northwestern and southern areas developed transitional facies. Although distribution and accumulation patterns varied greatly among sub-sags, the overall Baiyun Depression was characterized by widespread development of marine dinoflagellates. It should be noted that the northwestern sag also partly generated large-scale river delta deposits. Due to the eustatic rise and change of SCS spreading axis, the overall Baiyun Depression was mostly influenced by the deep marine environment on a continental slope during the early Miocene. Both northwestern sag and southern uplift zone were found plentiful marine dinoflagellates. In summary, transgression initiated from the eastern and southeastern Baiyun Depression before subsequently progressing into the farther west. Evolution of transgression process is also greatly consistent with the gradual westward expansion of the SCS.

Based on high-resolution 3D seismic data acquired in the Pearl (Zhujiang) River Mouth Basin of the northern South China Sea, this study investigated the geometry, spatial extension, and throw distribution of the post-rift normal fault through detailed seismic interpretation and fault modeling. A total of 289 post-rift normal faults were identified in the study area and can be classified into four types: (1) isolated normal faults above the carbonate platform; (2) isolated normal faults cutting through the carbonate platform; (3) conjugate normal faults, and (4) connecting normal faults. Throw distribution analysis on the fault planes show that the vertical throw profiles of most normal fault exhibit flat-topped profiles. Isolated normal faults above the carbonate platform exhibit roughly concentric ellipses with maximum throw zones in the central section whereas the normal faults cutting through the carbonate platform miss the lowermost section due to the chaotic seismic reflections in the interior of the carbonate platform. The vertical throws of conjugate normal faults anomalously decrease toward their intersection region on the fault plane whereas the connecting normal faults present two maximum throw zones in the central section of the fault plane. According to the symmetric elliptical distribution model of fault throw, an estimation was made indicating that normal faults cutting through the carbonate platform extended downward between −1 308 s and −1 780 s (two-way travel time) in depth and may not penetrate the entire Liuhua carbonate platform. Moreover, it is observed that the distribution of karst caves on the top of the carbonate platform disaccord with those of hydrocarbon reservoirs and the post-rift normal faults cutting through the carbonate platform in the study area. We propose that these karst caves formed most probably by corrosive fluids derived from magmatic activities during the Dongsha event, rather than pore waters or hydrocarbons.