On a coastal beach of very mild slope (close to or less than 1∶100), the multiple wave breaking may occur when waves propagate to the coast, companied by the wave recovery process between two breaking. In the current models for calculation of wave height, the wave breaking is usually simulated by the wave energy dissipation. But the establishment of energy dissipation describing wave breaking effect in these models did not automatically take the multiple wave breakings into account, and especially did not accurately simulate the wave recovery in this process. The present study presents a new model of the wave energy dissipation which can overcome this problem. The model is established by re-determining the stable wave energy, the saturated wave height-depth ratio and the dissipation coefficient and by introducing the judgment condition for the occurrence of wave recovery. The model has a very small value of the energy dissipation over wave recovery region and can properly describe the wave motion in wave recovery process. Comparison with measured wave heights shows that the model can give the wave heights on the plane and barred beaches of various bottom slopes (1∶10 to 1∶100), which agree well with the measurement, especially the multiple wave breaking and wave recovery process can be recognized automatically in the numerical simulations.

Based on the temperature and salinity grid data of SODA from1980 to 2015 and combined with Argo data, the formation mechanism and seasonal variability of the barrier layer thickness in the Bay of Bengal in Indian Ocean Dipole (IOD) events are revealed. The results show that the barrier layer changed significantly in the IOD event year. In the peak of pure IOD events (pIOD) and the pIOD+El Niño events (September−November), except for the 5 m thickening of the barrier layer in the southwest of the Bay of Bengal, the barrier layer become thinner range from 5 m to 15 m in the equatorial region and the region from Andaman Sea to the northern Bay of Bengal. This form disappears as the pIOD weakens. However, the form is maintained until March to May of the following year in the pIOD+El Niño events. The variation of barrier layer in the nIOD year is roughly opposite to that of the pIOD year. Further analysis shows that the change of isothermal layer caused by the remote forcing of the equatorial wind anomaly in the IOD event is the main cause of the variability of barrier layer. In the IOD+ENSO events, the equatorial wind anomaly is maintained for more than three months after the IOD disappear, making the anomaly of barrier layer last longer in the Bay of Bengal. In addition to the equatorial forcing, the Ekman pumping in the Bay of Bengal and the variability of the salt of mixed layer also have an effect on the interannual variation of the barrier layer thickness.

In this study, we investigated clay mineral assemblages together with other provenance indicators of Core ARC7-LIC retrieved from the southern Alpha Ridge, in order to reveal the changes in sediment deposition, ocean circulation in the western Arctic Ocean, and the development of surrounding ice sheet through the mid-late Pleistocene (～MIS 29). Changes in clay mineral composition in Core ARC7-LIC suggest a transition of Siberian sourced material during MIS 29−13 towards a North American sourced material during MIS 12−1. This transition reflects the change of the circulation in the western Arctic Ocean before and after the Mid-Brunhes Event, which features an amplification of glacial-interglacial cycles in the post Mid-Brunhes Event. An exceptionally high smectite peak characterizes MIS 12, which is inferred to be from North American source. Laurentide ice sheet discharges icebergs and fine material in the western Arctic Ocean since MIS 16. The amplitude of ice sheet growth and decay increases since MIS 12. During MIS 6, 4 and 3, the asynchronous variations of Ca/Al and kaolinite suggest heterogeneous development of Laurentide ice sheet on the Canadian Arctic Archipelago and Alaska-Mackenzie sides.

Hydrothermal venting systems mostly develop on the localized bathymetric highs of mid-ocean ridges with different spreading rates, and porous-elastic thermodynamic numerical models, which is filled with Darcian fluid can intuitively model the geometry, temperature structure and venting location of the hydrothermal systems in oceanic crust. Simulation results and the experimental-analytical model reveal that bathymetric highs with different sizes show a varied influence on the geometry of hydrothermal convection, the greater and wider bathymetric highs, the greater deflection of underlying plume to the bathymetric highs. Adopting the cross-axis bathymetry of the EPR 9°17′N and Lucky Strike hydrothermal fields into the model, our simulations produce vents fitting the realistic location of these two fields well. Finally, our schematic model for bathymetric relief deflecting underlying plumes indicates that the lower bathymetry and its underlying porous oceanic crust act as the main recharge zone, while the bathymetric high focuses the most of the discharge hot fluid due to the reduce of the overlying pressure of seawater.

In order to solve the problem of spatial distribution of fracture system at the end of methane leakage system, based on the high-resolution three-dimensional seismic data in the north of South China Sea, the spatial structure and distribution characteristics of fractures in bottom simulating reflector (BSR) distribution area are described by using visualization and coherent body technology. The geological genetic types of fractures are described. The relationship between fractures and other types of transport systems on methane gas accumulation is discussed. The fractures in the upper part of BSR interface are far less than those in the lower part of BSR interface, which makes the methane gas supply larger than that in the process of hydrate accumulation. It has a general indicating role in studying hydrate accumulation and detecting methane gas leakage. According to the development scale of fractures, the study area can roughly identify four types of fractures, they are each short fractures, long fractures, fracture bundles, and fracture groups, which enhance the leakage capacity of fluids in turn. These fractures often coexist in multiple types in the stratum, or form a leakage system together with other geological structures. These results and understandings are of great significance to improve the gas hydrate accumulation model and mechanism of methane leakage system in deep-water basin.

The simulation of wave time series or wave elevation histories is of great significance for the accurate study of real wave-action on marine structures. In this paper, a method of simulating multi-directional irregular wave elevations is proposed based on a linear single-summation model. To evaluate the effectiveness of the proposed method, theoretical simulated regular waves, unidirectional and multidirectional irregular waves are deterministically simulated. The consistence between the simulated wave time series and the original ones is verified. The simulated accuracy range is further evaluated based on the quantitative error analysis along the spatial dimension represented by the ratio of rr/L_s (in which, rr denotes the separations of the two wave gauges, and L_s dotes the significant wave length). It is suggested that the optimum relative separations of the wave gauges should be less than 0.12L_s when the proposed method is used to reconstruct multidirectional irregular waves.

Based on extensive experiments on the mooring tension of a rectangular cylinder, the present investigation compares the mooring tension of a floating rectangular cylinder under freak and random waves and quantifies the difference through experimental measurements. The effects of the relative wave height, relative period and freak wave parameter α₁ on the mooring tension are investigated. The results show that, the freak wave parameter α₁ has significant effect on the mooring tension of the floater. With α₁=2.0～2.83, the maximum mooring tension under freak wave were 1.9 times larger than that under random wave. With H_s/d=0.032～0.097, the maximum mooring tension under freak wave are significantly larger than those under irregular waves, but the 1/3 large value and average value under freak and irregular waves are almost the same. For the effect of the relative period, the critical variation of mooring tension occur at period within T_p<T_0p. The time-frequency spectra of the mooring tension under freak and random waves are calculated by wavelet analysis to investigate the time-frequency structure characteristics and variation. The time-frequency characteristics of mooring tension are significantly different from those of random wave.

This paper constructs a convolutional neural network based on TensorFlow. According to the idea of migration learning, the classical handwritten digit recognition is introduced as an introduction. The influence of different cost functions and activation function combinations on the classification results of convolutional neural network models is evaluated. Taking HJ-1A/B sea ice images as experimental data source, we analysis the influence of different function combinations on remote sensing sea ice image classification. It turns out that the cross-entropy cost function and the ReLU activation function are optimally combined. The feasibility of CNN in remote sensing sea ice classification is proved, and the classification results of the sea ice images in the Bohai Sea are verified. The calibration accuracy of the labeled samples is 98.4%. The model is then used to identify the unlabeled test samples. The influence of the window size on the sea ice classification results is discussed, and the optimal window size is 2×2 in the 400×400 small-scale classification experiment. Finally, the identification and verification of the entire Bohai Sea area is carried out, and the effect is good.

Based on the 634 MODIS images acquired in 2017, a total of 1 793 internal waves information are extracted. The paper analyzes the spatial distribution of internal waves in three typical regions of the Andaman Sea, quantitatively counts the characteristic parameters such as the length of the leading waves and the area of the wave packets. The potential sources of internal waves are discussed by ray tracing method and estimates the generation period of internal waves. The results show that the internal waves in the northern Andaman Sea have a small spatial scale. The average length of the leading waves is about 107 km, the average wave packet area is about 1 860 km², and the propagation direction is mainly eastward and southwestward. The average length of the leading waves in the central Andaman Sea is about 133 km, and the average wave packet area is about 3503 km². More than 70% of the internal waves propagate in the northeast direction. The average length of the leading waves in the northern part of the Sumatra Island is about 131 km, the average wave packet area is about 2 997 km², and the propagation direction of internal waves is mainly eastward, northeastward and southeastward. There are 7 potential internal wave sources in the Andaman Sea. The internal waves generation period is between 11.5 hours and 13 hours, which has obvious characteristics of half-day period.

In this paper, three methods for continuous visualization of mesoscale eddies are proposed, which are based on the technique of 2D streamline visualization and technique of mesoscale eddies identification: the method of eddy visualization based on OW parameters, the method of eddy visualization based on grid template and the method of eddy visualization method based on vector template. These three methods are respectively based on Okubo-Weiss algorithm, Faghmous algorithm and Liu's algorithm for eddy recognition, and the visualization results of the flow field are filled into the eddy to obtain better visualization effect. In the process of visualization, we introduce the transfer function to conduct real-time interaction between the color and transparency of the streamline in the eddy, which can control the display effect of setting the velocity, vorticity, OW parameters and other information by setting the color and position of the Key point on the control interface. In addition, we also compared the advantages and disadvantages of the three methods in terms of performance and display effect. In terms of performance, the performance is from high to low: the method of eddy visualization based on OW parameters, the method of eddy visualization based on grid template and the method of eddy visualization method based on vector template. In terms of display effect, the method of eddy visualization based on OW parameters is the worst among the three, with more chaotic short lines and smaller eddy boundary, which is limited to the core region of the eddy. The method of eddy visualization based on grid template has better display effect than the first method, with fewer messy short lines and relatively complete eddy. However, due to the lack of high resolution of data, the eddy boundary appears jagged after being put up for more than one time. The method of eddy visualization method based on vector template has the best display effect. The eddy is complete and full. At the same time, since the eddy boundary is reconstructed and vectorized, the eddy boundary is smoother. Compared with the traditional method of continuous visualization of eddies with long time series, these three methods provide a beautiful, dynamic and more informative visualization method. At the same time, they can become a practical tool for researchers to study eddies due to the addition of transfer function.