Based on the adaptive boosting algorithm (AdaBoost) combined with the extreme learning machine (ELM), the strong classifier of ELM-AdaBoost with strong robustness and high precision is thus constructed by iterating, adjusting, and optimizing the weights between each ELM classifier. ELM-AdaBoost method can enhance the stability of the existing ELM classifier. In this paper, the data collected by side scan sonar in the Zhujiang River Estuary was used to classify and identify three types of typical sediments as rock, sand, and mud. The average classification accuracy of new method exceeds 90%, which is better than the average classification accuracy of a single ELM classifier of 85.95%. It is also superior to other traditional classifiers (i.e. LVQ and BP) and it takes much less time to classify than traditional classifiers. The experimental result shows that the proposed ELM-AdaBoost method can be effectively applied to the classification and identification of seabed sediment and can meet the needs of real-time classification of seabed sediment.

The distribution and diffusion of suspended particulate matter (SPM) in coastal waters play an important role on ecological environment, coastal geomorphic evolution, aquaculture and coastal engineering. The northern Yellow River Delta is a strong coastal erosion area due to the diversion of Diaokou course. Revealing the variation characteristics and laws of SPM concentration in this area is the basis for maintaining the safety of protection projects. Using the well-validated model to retrieve the SPM concentration in the coastal waters, the cross validation results of Landsat-8 and Sentinel-2 satellite sensors show that the SPM concentrations retrieved by the two sensors have strong consistency, and the two kinds of satellite data can be used together. The seasonal variation of SPM concentration in the study area is obvious. The concentration of SPM is higher in spring and winter, and lower in summer. Autumn is the season when SPM concentration changes from low to high. In winter and spring, the wind and waves are large in this area. Under the combined action of waves and current, strong sediment resuspension occurs, which is the main source of SPM. The construction of spur dike group altered the spatial and temporal distribution of SPM to a certain extent.

Based on the outputs of eight earth system models involved in the Coupled Model Intercomparison Project Phase 6 (CMIP6), this study assessed the simulation skill of the Southeast Indian subantarctic mode water (SEISAMW) of these models by comparing with observations. Moreover, this study investigated the projected long-term trends in subduction rate, volume and properties of the SEISAMW under medium and high greenhouse gas emission scenarios (i.e., SSP245, SSP585). The results show that the CMIP6 models generally have produced artificially greater mixed layer depth and smaller upper layer potential density in comparison with those of the Argo observation. Consequently, the simulated SEISAMW in the CMIP6 models are generally with larger subduction rate and smaller potential density. Meanwhile, the subduction regions of the SEISAMWs show significant differences among the analyzed CMIP6 models, which are attribute to lateral induction in the mixed layer. Furthermore, in the historical, SSP245 and SSP585 outputs, the SEISAMWs show consistent decreasing trends in subduction rate and volume, increasing trend in temperature, and decreasing trends in salinity and potential density. The long-term trends of the SEISAMWs are largest under SSP585 scenario, followed by the SSP245 scenario and historical simulation. The projected trends of SEISAMW can be explained by the following mechanism: the temperature and freshwater flux in the southeastern Indian Ocean upper layer tend to increase under enhanced radioactive forcing, resulting in shoaling in mixed layer and flattening of the mixed layer gradient. As a result, the trends of SEISAMWs in subduction rate, volume and water properties show larger values in accordance with stronger radioactive forcing.

Using the direct observation data of the MSS90L turbulence profiler obtained in the summer cruise of the Changjiang River Estuary scientific investigation and experimental study (NORC2019-03-02) in July 2019, in this paper, the distribution of the turbulent kinetic energy dissipation rate ε and the vertical turbulent diffusion coefficient Kz are calculated and analyzed. The turbulent kinetic energy dissipation rate is 1.72×10⁻¹⁰ W/kg to 2.95×10⁻⁵ W/kg, and the vertical turbulent diffusion coefficient is 3.24×10⁻⁷ m²/s to 4.55×10⁻² m²/s. The distribution of the turbulent kinetic energy dissipation rate and the vertical turbulent diffusion coefficient are similar, both of which are the strongest in the upper layer, the next in the bottom layer and the weakest in the middle layer. In the upper layer, due to wind stress, the turbulent kinetic energy dissipation rate and vertical turbulent diffusion coefficient are larger; the stratification is stronger at the thermocline, which suppresses the dissipation of turbulent kinetic energy and vertical turbulent mixing. The secondary circulation of the salinity front will promote the separation of the low salt water mass, and the vertical circulation caused by the front will enhance the turbulent mixing of the ocean. The energy exchange between the low salt water mass and the outside world is less, and the turbulent kinetic energy dissipation rate is weaker. There are obvious upwelling and downwelling in the Changjiang River Estuary, which are produced by the secondary circulation of the front; the existence of upwelling and downwelling promotes the dissipation of turbulent kinetic energy and turbulent mixing.

In this study, we quantify the mixed-layer salinity (MLS) budget in the Indian Ocean using the salinity budget equation along with Argo salinity, SODA ocean current, evaporation, and precipitation data. The results show that the distribution and seasonal variability for the terms of freshwater flux, advection and entrainment in the Indian Ocean, and their contributions to the variations of MLS. Regarding the climatological annual mean, the negative contribution (15.14%) of the horizontal advection term to MLS is much larger than its positive contribution (9.89%). This implies that the horizontal advection transports low-salinity water to the salty regions, resulting in a decrease in MLS there. The distribution and seasonal variability of the freshwater flux term are generally consistent with those of precipitation, with a positive contribution (13.70%) higher than negative contribution (7.81%). It suggests that the freshwater flux term increases MLS in the Indian Ocean because of the evaporation overwhelming the precipitation. Analysis on the regions with significant seasonal variations of MLS indicates that from June to November, the Southwest Monsoon Current transports low-salinity water from the western equatorial Indian Ocean to the western Arabian Sea, causing salinity in that region to decrease. The advection of high-salinity water from the southern and central to the northern of Bengal Bay is the main reason for the increase in salinity there.

Using in-situ hydrographic observations from January to December 2017, 3D thermohaline structure and seasonal variation in the Zhanjiang Bay is investigated in this study. The results show that: (1) In 2017, the annual mean temperature is 23−27℃, the mean salinity is 19−27, the mean potential density is between 11−17 kg/m³, and the mean buoyancy frequency (N²) is about 7×10⁻⁵−5×10⁻³ s⁻² in the Zhanjiang Bay. The vertical structure and horizontal distribution of N² are similar to that of the temperature, while the distributions of potential density and salinity are similar. (2) The temperature has significant seasonality with the highest value in summer, followed by autumn, and the lowest in winter. The maximum temperature difference between winter and summer reaches 15℃, while the seasonal mean of the salinity varies slightly. The ebb and flood have less influence on temperature and salinity, comparing with their seasonal variations. The thermocline is the strongest in summer with the maximum gradient reaches 0.7℃/m at 10 m, whereas it shallows to 5 m in spring and fall, and the water well mixed in winter. The halocline is prominent in summer and fall, with the maximum gradient 1.1 m⁻¹. The seasonal variation of the thermohaline in the upper layer and lower layer are consistent. (3) For the horizontal distribution, the temperature decreases and the salinity increases from the bay head, the mid bay, the bank region, the shoal region to the bay mouth. The average temperature difference between the bay head and the bay mouth is 2.3℃, and the salinity difference is 2.7. The temperature-salinity (θ-S) diagram shows banding distributions with one end as the low-temperature and high-salinity water in the bay mouth, one end as the high-temperature and low-salinity water in the bay head, and the other water masses are between them. Different bands are shown in different seasons.

Ocean waves can affect the roughness of the ocean surface, and the waves generated by tropical cyclones impact the momentum and energy fluxes across the air-sea interface. In this study, the impacts of ocean waves on the momentum and energy fluxes under tropical cyclones is examined by using the tropical cyclone observation dataset IBTrACS (International Best Track Archive for Climate Stewardship) and the simulations from wave model WW III (WAVEWATCH III). It is found that the intensity of tropical cyclones increased by about 1 m/s every decade in the past 30 years, but the change of translation speed is not obvious. For the stronger tropical cyclones, the difference for momentum flux and the energy flux between the air-sea interface can be significantly increased by the waves. Owing to the asymmetry of wind and surface wave fields under tropical cyclones, momentum difference and energy difference also demonstrate asymmetric distribution: the area with larger momentum difference is behind the moving direction of tropical cyclone, while energy difference is the largest in the right-rear quadrant, and both are relatively smaller in the left front quadrant. The inverse wave age is highly correlated with momentum difference and energy difference, and the correlation coefficient is about 0.95, indicating that the younger the surface wave, the more momentum and energy absorbed and stored by surface wave field. The inverse wave age increases with the increase of tropical cyclone translation speed, and the speed is positively correlated with momentum difference and energy difference according to the correlation coefficient above 0.8. Therefore, the surface waves affect the distribution and magnitude of momentum and energy input from the atmosphere to the ocean under tropical cyclones. In the future study of ocean boundary dynamics and thermodynamics, especially the study of tropical cyclones, considering the influence of ocean wave evolution is necessary.

The radionuclides ⁷Be, ²¹⁰Po and ²¹⁰Pb transported by the atmosphere can be used as important tracers for studying the material deposition flux of the Arctic Ocean’s atmosphere, modern ocean sedimentation and the transport of materials into the sea ice. They have been widely used in the study of air mass movement, soil erosion, and particle circulation processes in water systems. This paper reports the activities of ⁷Be, ²¹⁰Po and ²¹⁰Pb in the surface snow of the high-latitude ice floes of the Arctic Ocean in 2018. The activity concentrations of ⁷Be, ²¹⁰Po and ²¹⁰Pb are 33.6−632.68 mBq/L, 36.2−87.5 mBq/L, and 30.9−194.49 mBq/L, respectively. The activity concentrations of ⁷Be and ²¹⁰Pb in the surface snow of the Arctic Ocean are lower than those in the mid-latitude continental areas. The results show that the activity concentrations of ⁷Be in snow increased with the increase of latitude. The activity ratio of ²¹⁰Po/²¹⁰Pb ranged from 0.70 to 1.48 (with an average of 0.93), ²¹⁰Po is almost in equilibrium with ²¹⁰Pb. It indicates that the age of the surface snow is “older”.

The geochemistry, occurrence phase and enrichment mechanism of REE and PGE were studied in cobalt-rich crusts from the Seamount Xufu in the Pacific by ICP-OES, ICP-MS and phase analysis methods. The results showed that, REE and PGE were enriched in cobalt-rich crusts, REE contents were 1 842−2854 μg/g, and PGE contents were 144−1180 ng/g. REE and PGE contents in the old layers were higher than that in the new layers, and the phosphatization in the old layers might play an active role in the REE and PGE enrichment. Moreover, LREE contents were more than HREE contents, and REE diagrams showed the positive Ce anomalies and no Eu anomalies, so Ce was enriched in cobalt-rich crusts. Meanwhile, there was an obvious contrast between PGE, and PPGE contents were more than IPGE contents. PGE diagrams showed the positive Pt anomalies and negative Pd anomalies, so Pt was enriched and Pd was poor in cobalt-rich crusts. In addition, REE in the new layers were mainly enriched in the iron oxide phase, and the enrichment ratios were 65.40%−70.05%. While REE in the old layers were mainly enriched in the residual phase, and the enrichment ratios were 62.27%−65.77%, because it might be related to the phosphate. Meanwhile, PGE in the new or old layers were mainly enriched in the iron oxide phase, and the enrichment ratios were 63.66%−69.51%, and also the residual phase ranged from 29.20% to 34.68% had an impact on PGE enrichment. Therefore, REE and PGE enrichment were presumed to the iron oxide colloid particles adsorption. In the marine oxidative environment, the soluble Ce³⁺ and Pt²⁺ were oxidized to insoluble Ce⁴⁺ and Pt⁴⁺ respectively, and the precipitates were adsorbed into the iron oxide phase, so REE and PGE were enriched in cobalt-rich crusts.

The seabed soil layer will gradually accumulate pore pressure under the action of wave cyclic load, reduce the stability of the soil layer, and threaten offshore engineering. In order to study the accumulation mechanism of pore water pressure, the discrete element pore density flow method was proposed, and the discrete element analysis software MatDEM was improved to realize the simulation of the accumulation process of seabed sediment pore pressure. Based on the field test device and the mechanical parameters of the soil body, a discrete element model was established. By comparing field test and numerical simulation results, it was found that after applying the wave load to the seabed sediment, a higher pore pressure was generated in the surface soil body and gradually transferred to the deep layer. Under the action of cyclic wave load, the cumulative range of pore pressure between soil particles gradually increased. When the pore pressure accumulation time was long enough, the pore pressure in the soil layer converged to the average of the maximum and minimum loads applied. As the number of load cycles increased, the pore pressure at all depths increased cumulatively until the soil body liquefied and the internal soil particles became resuspended sediments. Under the action of periodic wave loads, the soil particles moved after being liquefied and suspended, and the shallow particles had a large displacement, and the overall performance of the soil was circular arc movement.