The sound velocities of twenty-five sediment samples were measured, both on the research vessel and in the laboratory, and the data for sixteen typical sediment samples are presented here. The sound velocity measurement on the deck were carried out immediately after the sediment cores were collected from seafloor and the time elapsed between the collecting and the measurement was about 30-60 minutes. The sound speed were measured using the whole sediment sample core on the deck and in the laboratory, respectively. The sound velocities of sediment samples were measured at temperatures ranging from ~18.0–20.5°C on the deck and ~19.5–21.5°C in the laboratory. The measurements were conducted using the same equipment, at frequencies ranging from 25–250 kHz, both in the deck and laboratory ( Figure 2 ). The field bench-top acoustic system was composed of a portable bench and a digital sonic meter, with a measurement accuracy equal to ±0.1 mm of the sample length. Meanwhile, the laboratory bench-top acoustic system was composed of an automatic bench and a digital sonic meter, with the same measurement accuracy as in the field tests. The digital sonic meter was equipped with a handheld computer, Bluetooth and a high-speed analogue-to-digital data converter (A/D), with a low-power consumption and a small volume and weight (http://www.cqbtsk.com.cn/english/Sonic/WSD-3.htm). The portable field bench included a dismountable bottom beam, two fixed units and two support units, and the laboratory bench included a movable and locatable slide platform, a fixed holder, guide rail, sliding holder, V-shaped support and a dividing ruler. The purpose of the beams, the platform, and the support units are to fix and support the sediment cores. During the measurement, the sediment core was placed on the measuring platform at first. The acoustic measurements were conducted along with the length of the cores. The planar emitting transducer was pressed onto one side of the sample core section and the receiving transducer was pressed onto the other side of the sample core. A pulse signal was emitted at the central frequency of each emitting transducer. This method uses travel time differences between the acoustic signal through sediment samples and identical reference cores filled with water to measure compressional wave phase speed. The take-off point of a sound wave signal were point out using the conventional manual identification method. The measurement accuracy of the sediment core length of the bench was 0.1 mm, and the digital sonic instrument had a sampling rate of 10 MHz (that is, the accuracy of travel time is ± 0.1 μs) and a sampling length of 4,096 samples. So, the accuracy of sound speed of bench-top acoustic system was estimated to be better than ± 0.1% for a typical sample with a length of 30 cm and velocity of 1500 m/s.

Together with the measurement of the sound velocities of sediment samples, the particle composition, physical and mechanical properties and microstructure were analysed in the laboratory. Laboratory procedures to determine the geotechnical properties of the sediment samples followed the specifications for oceanographic survey GB/T 12763.8-2007 and the standard for geotechnical testing method GB/T 50123-1999, belonging to National Standard of the People’s Republic of China. A laser particle analyser (Malvern 2000, UK) was used to determine the composition, mean grain size, sand content, clay content and sorting coefficient; the sediments were then classified using Shepard’s (1954) ternary diagram based on the obtained sand and clay contents. The physical properties analysed included the wet bulk density, dry bulk density, water content, porosity, permeability coefficient and plasticity index. The wet bulk density is the weight of the mineral solids and porewater per unit volume and was measured using a cutting-ring method using a steel ring sampler (diameter 6 cm, height 2 cm) which was pushed into the sediment sample (Zhou et al., 2008). The water content was measured via oven-drying. The dry bulk density was calculated from the wet bulk density and water content (Zheng et al., 2011). Porosity was calculated by the specific gravity and dry bulk density (Zheng et al., 2011), and the permeability coefficient was measured using the water head difference from the permeability test (Clearman, 2007). The parameters mechanical properties measured in laboratory were the penetration and shear strengths, wherein the penetration strength was measured using a penetrometer (Lei and Xiao, 2002), and shear strength was measured using a miniature Vane shear apparatus (Xu et al., 2017). The liquid limit and plastic limit of sediments were measured using liquid-plastic limit combined tester (Meng et al., 2015), and the plasticity index is the difference of liquid limit and plastic limit. Microstructural observation were performed using a scanning electron microscope (SEM; QUANTA 200, Thermo Fisher Scientific, USA), at the magnifications of ×100, ×800 and ×1500.

At different frequencies, the differences in the sound velocities of the silty sand samples collected from the northern Huanghai Sea and Bohai Sea measured on-site and in the laboratory were generally ~85–130 m/s, and those measured in the laboratory were significantly higher than those measured on the deck. The discrepancies in the sound velocities were slightly different at different frequencies. Meanwhile, it was also found that the dispersion of the sound velocity at different measurement frequencies, which was obtained in the two measurement methods, was consistent ( Figures 3A–C ).

There were four types of sediments collected in the study area of the Yellow Sea and Bohai Sea – silty sand, sandy silt, silt and clayey silt. These different types of sediments exhibited different degrees of granularity in terms of sand content, clay content, mean particle size and the sorting coefficient ( Figure 4 ). For the sand content, samples of silty sand had significantly higher concentrations, ranging from 48.5–63.2%, while samples of silt and clayey silt had much lower concentrations, ranging from 0.2–21.7%, and samples of silt exhibited a moderate concentration in the range of 21.4–37.0%. With respect to clay content, samples of clayey silt had substantially higher concentrations, ranging from 26.6–36.9%. Meanwhile, samples of silty sand had much lower concentrations, ranging from 7.1–15.9%, and samples of sandy silt and silt had middling concentrations, ranging from 12.7–24.7% and 13.8–24.8%, respectively.

The mean particle sizes in samples of silty sand were generally higher, ranging from 6.8–8.4 μm, while those of other types of sediments were substantially lower. This was the most obvious for the samples of sandy silt and clayey silt, which ranged in size from 3.4–6.6 μm and 3.9–7.6 μm, respectively. For the samples of silt, the average sizes of individual grains were generally concentrated in the range of 7.0–7.5 μm. With respect to the sorting coefficient, the different types of sediments in the study area were >1.0 and the silty sand had the highest average sorting coefficient, ranging from 2.2–3.1. The silty sand was poorly sorted with good gradation. The sorting coefficient of the clayey silt range from a similar 2.1–3.2. Compared with the other types of sediments in the study area, the sorting coefficients of the sandy silt and silty sediments were generally lower, while the degree of sorting was generally higher.

This study is focused on the sediments collected in the northern Huanghai and Bohai seas, which mainly comprise silty sand, sandy silt, silt and clayey silt. The sediment samples that Meng et al. (2012) collected in the southern Huanghai Sea involved 10 types of sediment, including super-fine sand, silty sand, clayey sand, sandy silt, coarse silt, medium silt, clayey silt, silty clay, clay and sand–silt –clay. Therefore, compared with the characteristics of the sediments in the southern Huanghai Sea, where the sand content ranged from ~0.3–77.3%, clay content ranged from ~8.9–89.5% and the mean particle size ranged from ~0.002–0.129 mm, the sediments in the northern Huanghai and Bohai seas represent a narrower scope and are generally concentrated. However, in terms of the silty sand, sandy silt, silt and clayey silt, the scope of the granular characteristics of the sediment samples collected in this study is broader than in the study of the southern Huanghai Sea by Meng et al. (2012). These differences are related to the quantitative restrictions imposed by the sediment sampling in this work. Nevertheless, on it can be preliminarily estimated that the distributions of the sediments in the northern Yellow Sea and Bohai Sea are typical (Wang et al., 1996).

Different types of sediments in the northern Huanghai Sea and Bohai Sea have physical properties that are obviously different based on the measured wet bulk density, dry bulk density, water content, porosity, plasticity index and permeability coefficient ( Figure 5 ). For wet bulk density and dry bulk density, the samples of different sediment types were arranged in the order of: silty sand > sandy silt > silt > clayey silt, and respectively in the range of 1.58–2.06 g/cm³ and 1.07–1.53 g/cm³. The water contents of different types of sediments in the study area were high, and the moisture content of the clayey silt was the highest. The difference in the water contents among the other three types of sediments was not obvious, as they ranged from 35–45%. With respect to the porosity, clayey silt sediments exhibited the highest porosity and a low degree of consolidation, while the silty sand has the lowest porosity and a high degree of consolidation. Preliminary results showed that it these traits were closely related to the water lost from sediment samples, wherein the samples of sandy silt were more discrete and changed within a wider range. The difference between silt and sandy silt samples was insignificant, and they changed within a standard range.

For the plasticity index, different types of sediments showed more obvious regularity, which increased in the order of: silty sand< sandy silt< silt< clayey silt. The plasticity index of silty sand was in the range of ~6.5–8.5, while for sandy silt it was ~9.0–11.0, for silt, the range was ~12.5–14.0, and clayey silt ranged from ~14.5–19.5. Finally, the permeability coefficient showed different degrees of regularity for different types of sediments, which increased in the order of clayey silt<silt<sandy silt<silty sand. The permeability of silty sand was in the range of 6.0×10^-7–7.5×10^-7 cm/s, while for sandy silt, it ranged from 4.5×10^-7–6.8×10^-7 cm/s; the permeability range for silt was 3.2×10^-7–5.0×10^-7 cm/s and that of clayey silt was 3.5×10^-7–5.5×10^-8 cm/s.

Compared with the physical properties of the sediments collected in the northern Huanghai and Bohai seas, the sediment types in the southern Huanghai are complex and their properties vary widely. Meng et al. (2012) analysed 10 sediment samples collected from this region in a laboratory, and found a wet bulk density of 1.50–2.07 g/cm³, a water content of 25.0–107.4%, a porosity of 0.40–0.74 and a permeability coefficient ranging from 9.0×10^-8–1.5×10^-7 cm/s. The consistency of the results between our two studies may be seen in the fact that the physical properties of seafloor sediments correspond well to their types, and coarser sediments have higher values for bulk density and permeability, and lower ones for water content and porosity.

The penetration and shear strengths of sediments are important parameters to characterize their macroscopic mechanical properties, which are vital for marine geotechnical engineers. The penetration strength indicates the vertical bearing capacity of the sediment, which is determined by the degree of compaction. The shear strength indicates the maximum shear force that the sediment can bear, which is mainly affected by the adhesion and friction angle of the sediment. Based on laboratory measurements of penetration and shear strength for different types of sediment samples from the northern Huanghai and Bohai seas ( Figure 6 ), it was found that those of silty sand were the highest. The sandy silt and silt exhibited middle values, with insignificant differences between them, whereas clayey silt sediments had very low penetration strengths and high shear strengths. The distributions of the mechanical properties of these sediments could also be obtained in the same manner for the laboratory measurements of the sediments from the southern Huanghai Sea (Meng et al., 2012). However, for the sediments in the north, the differences were particularly notable.

Combining the analyses shown in Figures 4 – 6 and research of Meng et al. (2012) in the southern Huanghai Sea, we identified four types of sediments – silty sand, sandy silt, silt and clayey silt, which could be divided into three categories for measurement in the laboratory. Among them, silty sand constituted a category, sandy silt and silt could be grouped into a single category and clayey silt was the final category. Comparatively, sandy silt and silt have small differences in their particle sizes (i.e., sand content, clay content, average particle size and sorting coefficient), physical properties (i.e., wet bulk density, dry bulk density, water content, porosity, plasticity index and permeability) and mechanical properties (i.e., penetration and shear strength). In this study, the sediment samples were classified according to Shepard’s ternary diagram used for classification according to the Marine Investigation Regulations (GB/T12763.8-2007). In marine geotechnical engineering, sediments are divided into sand, silt and clay according to their plasticity indices (Jia et al., 2011). According to the plasticity index method used in marine geotechnical engineering, there are four types of sediments in the study area, as classified according to the Marine Investigation Regulations, which can be broadly divided into the aforementioned three categories. Among them, silty sand and sandy silt can be roughly classified as sand, while silt and clay are classified as such.

The microstructural characteristics of sediments are important factors that determine their physical mechanics and engineering properties, and also have important impacts on their acoustic properties. In this work, three microscopic structures of different magnifications (×100, ×800 and ×1500) were examined by electron microscopy in the sediments from the northern Huanghai Sea and Bohai Sea. The typical microstructure electron microscope images of four different sediment samples are shown in Figure 7 . This microscope images represent the dried particles alignment of different sediments.

Silty sand had a sedimentary structure based on sand particles, and silt particles supplemented to become the skeleton, with a small amount of clay filling in. It can be seen from the upper panel of Figure 7 that the samples of silty sand had large skeletal particles and their tendency toward a preferred orientation is obvious. The fine particles are filled and attached to larger particles with large pores, which can be approximated as a single grain structure (Hamilton and Bachman, 1982). The content of the sand in sandy silt was less than that of the silty sand, and the content of fine grain sizes increased. It can be seen from the second panel (row) of Figure 7 that the particle sizes of the sedimentary skeleton were much smaller than in the silty sand, with no clear trend of particle orientation. Additionally, the sediments shown in this panel exhibit moderate roundness, and the voids are filled with more fine-grained particles, while also being attached to the skeleton, with more and smaller pores. The sand content in the silt was even lower, and the main component was silt; the clay content increased relative to the sandy silt. It can be seen from the second panel from the bottom in Figure 7 that the abundance of large particles that make up the skeleton is greatly reduced, and there is an obvious grain orientation. The silt grains exhibit poor roundness, and a large number of fine particles fill in and attach to the skeleton, with more and smaller pores than in the sandy silt. Finally, the sand content in clayey silt was very small, as the main components were fine-grained particles and clay. As can be seen from the bottom panel of Figure 7 , the clayey silt forms and approximate bundled sheet structure (Hamilton and Bachman, 1982; Meng et al., 2012), wherein large particles are completely surrounded by fine particles, which become aggregates owing to complex physicochemical interactions.

The observation of the microstructural characteristics of sediments involves relatively matured research methods; however, the physical models of sediment geotechnical and acoustic behaviours and properties based on such microstructures are currently lagging behind. Techniques for measuring the porosity sediments from microstructural images have now become possible. Changes in the clay microstructure during consolidation have also been studied and linked to changes in the shape, alignment and spatial arrangement or distribution of clay crystals and aggregates (Zheng et al., 2014). It is also now possible to predict sediment permeability based on the size of the pores pathways observed and the amount and degree of interconnection among pores (Vaughan et al., 2002). Nevertheless, models for predicting compressional or shear wave velocities based on sediment structure and the physicochemical interactions between sediment grains and pore fluids have remained unsuccessful, and much research is still required (B.H. Liu et al., 2013).

Based on the comparative analysis of the sound velocities measured on the deck and the laboratory for four types of sediments collected from the northern Huanghai and Bohai seas (i.e., silty sand, sandy silt, silt and clayey silt), we found the thresholds of V p − f / V p − l , which ranged from 0.92–0.94, 0.90–0.92, 0.98–0.99 and from 0.99–1.01, respectively ( Figure 8 ). According to our comparative analyses, the V p − f / V p − l values of the four types of sediment in the southern Huanghai Sea are generally within the thresholds of the northern Huanghai Sea and Bohai Sea, and their deviations are less than 5%, which demonstrates that the calculated threshold range of V p − f / V p − l in this study is statistically significant.

Temperature is one of the most important factors affecting the sound field in the ocean. A clear theoretical system of its influence has now been formed. However, the influence of temperature on the acoustic properties of seafloor sediments is still in the stage of semi-qualitative analysis and a clear understanding of the mechanism of influence has not yet been formed.

Hamilton (1971, 1972) concluded that the acoustic properties of pore water in sediments are the most sensitive to temperature changes and are also the main factors that control the acoustic properties of sediments with temperature. Zou et al. (2008) found three trends in sound velocities with temperature based on temperature-controlled physical model experiments in a laboratory; these were: (1) the positive growth of sound velocity (STPIK), (2) the negative growth of sound velocity (STNIK) and (3) the fluctuation of sound velocity (STWK). Because the test samples for these three trends were all silt, with little differences in porosity, density or water content, it is difficult to clearly interpret these results in terms of the actual effect of such trends in sediment sound velocities with temperature. In this study, it was speculated that seafloor sediments have complex structures due to their different sedimentary histories (Hamilton, 1976; Lu et al., 2003), and the testes samples were all unconsolidated sediments with a loose structure. The difference between loose and unconsolidated structures composed of pore particles of different shapes and geometries leads to the complexity of the relative motions of seawater through pores and the solid-phase particles under thermal motion, possibly resulting in the formation of the three trends observed by Zou et al. (2008). In other words, the microstructure of loose sediments is a critical factor affecting the trends in sound velocity with temperature.

Based on the laboratory measurements of sound velocities with temperature for nine different types of samples, Zeng et al. (2009) showed that the sound velocity decreased with temperature in the order of sand > clay > silt, presumably because as temperature rises, changes in the seepage state cause the porosity of the sediment to increase, such that the seawater volume increases, resulting in a low sound velocity. Thus, sound velocity increases with temperature in the clayey sand and clayey sand–silt. It is speculated that in the clayey silt, moisture does not easily flow due to the high clay content. Furthermore, the composition of clay minerals and sands play an important role in the response of sediment sound velocities with temperature. In other words, the difference between the grain sizes of sedimentary components is the important factor influencing the variation of the sound velocity with temperature for different types of sediments. Hong et al. (2011) analysed the variation of seafloor sediment sound velocities with temperature in the South China Sea. The sound velocities of the sediments in this area varied slowly and exhibited obvious fluctuations, but the overall trend was almost linearly decreasing. We believe that the mechanism underling the effect of temperature on the sound velocity of seafloor sediments was equivalent to the comprehensive effect on the equivalent elastic modulus and density of the sediments.

In the temperature-controlled physical model experiments in laboratories, such as the studies of Zou et al. (2008) and Zeng et al. (2009), the range of temperatures measured has been 0–30°C, and the range of sound velocities has been 1300–1600 m/s. As for the temperature-controlled sound velocity measurement of samples from the South China Sea (Hong et al., 2011), the range of variation in the sound velocity was 1700–1880 m/s within a similar temperature range as for the aforementioned studies, and the variations in the sound velocities of different types of sediments differed. The measured sound velocities of different sediment samples varies within the range of 1400-1600 m/s over the measured temperature range (Meng et al., 2012). In the shipboard field and laboratory sound velocity measurements of sediment samples from the Huanghai and Bohai seas, the measurement temperature ranged from ~18.0–20.5°C, and the temperature difference in the same group was less than 3°C. Based on the present laboratory measurements of sediment sound velocity, it is assumed that temperature had little effect on the sound velocities measured (about 8 m/s), as such a small difference in temperature (≤2.5°C) can be neglected.

The particle size composition of seafloor sediments is one of the important parameters in seafloor sedimentology and soil mechanics, an important parameter can reflect the sedimentary origin, environment and processes of the sediments. Characterizing the size composition of sediments is important for laboratory analyses of seafloor sediment samples (Lu and Liu, 2008). Research of the relationship between the particle size composition and acoustic properties of seafloor sediments began in 1960s and 1970s. Akal (1972) and Hamilton (1976) extensively studied the effects of particle size on the sound velocities of seafloor sediments in the world oceans, and established the empirical relationships of grain size parameters, such as sand content, clay content and mean particle size, with their acoustic properties. In subsequent works, Liang and Lu (1983), Prasad (2002), X.L. Liu et al. (2013) and Kan et al. (2014) have analysed the effect of sediment particle size on the acoustic properties of shallow sediments in different marine areas, and established an empirical ground acoustic model that describes the relationship between sediment particle size and geoacoustic properties, which are suitable for different areas. These studies have revealed that the influence of sediment particle size on the acoustic properties of marine sediments has a clear mathematical relationship that may be quantified during macroscopic characterization. The deviations of on-site sound velocity measurements from laboratory measurements of different types of sediments have obvious differences with respect to their classification and particle size compositions, but what are the effects? What is the mechanism underlying such impacts? The answers to these two questions need to be understood from the point of view of marine soil mechanics to understand the mechanism that causes the deviation between laboratory and on-site sound velocity measurements.

From Figure 3 and Figure 8 , it can be seen that the silty sand, sandy silt, clayey silt and silt samples from the Huanghai and Bohai seas have significantly different V p − f / V p − l ranges, while for the same type of sediments, the V p − f / V p − l is concentrated within a narrow range and the deviation is small. The results show that the V p − f / V p − l value of silty sand is similar to that of sandy silt, both of which are notably less than 1.0. The V p − f / V p − l _l of the silt is also less than but closer to 1.0, while for clayey silt, values both greater than and less than 1.0 exist. From the V p − f / V p − l _l value classifications, silty sand and sandy silt can be grouped into one class, while the classes for silt and clayey silt remain discrete. Based on the characteristics of the particle size components of sediment samples in this study, and according to the Shepard’s classification method of Marine Investigation Regulations (GB/T12763.8-2007) and the plasticity index method of marine geotechnical engineering, silty sand and sandy silt can be roughly classified as sand, and silt and clayey silt can be classified as such.

For superficial sediments, the external environmental factors that cause errors in the shipboard and laboratory acoustics tests are the changes in the temperature of sediment samples and the vibration of sediment samples during transport. From the analyses described in section 5.2, it can be assumed that the change in temperature had little effect on the errors of either on-site or laboratory acoustics tests of sediment samples in this study. Therefore, the impact of vibration during the transport of sediment samples can be taken as the main factor causing the deviation of the laboratory-measured velocities from the on-site ones. As shown in Figure 3 and Figure 8 , the influence of vibration during the transport of sediment samples obviously differed between silty sand and sandy silt, and between silt and clayey silt. Based on analogous analyses in marine sediment surveys and marine geotechnical classification methods, the difference in the above influencing mechanisms could be analysed from the perspective of marine soil mechanics.

In the study of marine soil mechanics, sandy soil has the characteristic of micro-vibrational liquefaction, and clay exhibits obvious thixotropic characteristics under the action of dynamic forces. As the transitional soil between sand and clay, silt has both the liquefaction characteristics of sandy soil and the thixotropic characteristics of clay (Jia et al., 2011). The liquefaction of sand refers to the phenomenon that the saturated sand entered a liquid state under the action of vibration; due to the increase of pore water pressure, the sand changes from a solid state to a liquid state from the decrease in the effective stress. The mechanism is the saturated fine sand has a tendency to move and become denser under vibrational forces, wherein the stress migrates from the sand framework to the water, and due to the poor penetration of fine sand, the pore water pressure will increase sharply. When the pore water pressure reaches the total stress value, the effective stress declines to 0. The particles will then be suspended in the water, and the sand will liquefy (Liu et al., 2017). After a period of time following liquefaction, the drainage and consolidation processes will occur, resulting in more intense consolidation (Li et al., 2008). Meanwhile, the thixotropic property of clay is such that when the structure of clay is disturbed, the electric double layer is destroyed, resulting in reduced strength. However, with the increase of standing time, a new equilibrium is formed among the soil particles, ions and water molecules, and the soil strength recovers gradually. Nevertheless, the recovered characteristics generally do not differ greatly from the initial sediment strength; this property is called soil thixotropy.

Based on the analyses presented here, it can be inferred that silty sand and sandy silt both have the dynamic liquefaction characteristics of sandy soils, such that the liquefaction and re-consolidation processes that occur during sample transport are important mechanisms that lead to significantly higher laboratory sound velocities than those recorded on the deck. Clayey silt has the dynamic thixotropy characteristic of clay, and the thixotropic resumption process during sample transport is also an important mechanism that leads to significantly higher laboratory sound velocities than those from the on-deck. With the dual characteristics of liquefaction and thixotropy, the processes of liquefaction, re-consolidation and thixotropic resumption of silt–sand occurring in the process of sample transport are the important factors that lead to significantly higher laboratory sound velocities than those measured on the deck for silts. The dual characteristics of silt–sand, with dynamic liquefaction–thixotropy, determine that its V p − f / V p − l ranges from those of sandy silt and silty sand to that of clayey silt.

Particle size is the main factor that affects the dynamic liquefaction and thixotropic properties of sands, silts and clays (Zheng et al., 2011), and indirectly affects the V p − f / V p − l of silty sand, sandy silt, silt and clayey silt. Based on a comparative analysis of the on-deck- and laboratory-measured sound velocities in the sediment samples from the northern Huanghai and Bohai seas, it can be concluded that the V p − f / V p − l of the silt samples was the least affected by the particle size composition, while those of sandy silt and silty sand were the most affected ( Figure 9 ). For the V p − f / V p − l of clayey sand, different granularity parameters have different influences. In the sample tests performed in this study, the clay content of clayey silt was more concentrated, so it was difficult to judge the influence of clay content on its V p − f / V p − l . However, we can clearly point out that the sorting coefficient and sand content had significant impacts ( Figures 9B–D ).

Based on the microstructural observations, it was found that both the silty sand and sandy silt had large skeletal particles, tendencies for preferred grain orientations and the fine particles were filled and attached to larger particles ( Figure 7 upper row and upper middle row). The large and small particles and their combined arrangements can greatly affect the infiltration and drainage of sediments, thus affecting their liquefaction and re-consolidation, and the resultant change in V p − f / V p − l . The clayey silt had an approximately bundled sheet structure (Hamilton and Bachman, 1982), with large particles that were completely surrounded by fine particles, some of which became aggregates owing to complex physicochemical interactions ( Figure 7 lower row). For clayey silts with relatively concentrated clay contents, the effects of sand content and particle size on the electric double layer and the microstructure of the aggregates were relatively prominent. Therefore, the variations in the sand content and sorting coefficients had significant effects on the thixotropic properties of the clayey silt, which in turn affected its V p − f / V p − l . As a feature of dynamic liquefaction, re-consolidation and thixotropic resumption, particle size parameters exhibit a coupled effect on the dynamic characteristics of silty sand, making it difficult to determine which particle size factor had a significant impact. This is also an important reason that the V p − f / V p − l of the silt was obviously higher than those of the sandy silt and silty sand.

Based on the discussion and analyses in sections 5.2 and 5.3, it has now been shown that the dynamic liquefaction and re-consolidation and the dynamic thixotropy resumption processes of sediment samples are the important external factors that lead to the differences between the on-site and laboratory sound velocity measurements of sediments in the Huanghai and Bohai seas. External environmental conditions often affect the acoustic properties of sediments by changing the sedimentary characteristics of their properties, which acts as a trigger mechanism, as the physical properties of sediments are important internal factors that determine their acoustic properties (B.H. Liu et al., 2013). Numerous studies have shown that there are relatively clear mathematical relationships among bulk density, water content, porosity and other physical properties of sediments and their sound velocities (Hamilton and Bachman, 1982; Bachman, 1985; Orsi and Dunn, 1990; Kim et al., 2001; B.H. Liu et al., 2013). At present, the consistent research results obtained have shown that the larger the bulk density, the lower the water content, the smaller the porosity and the higher the sound velocity of the sediment.

In this study, the physical properties of sediment samples from on-site were not measured. However, comparative analyses of the on-deck- and laboratory-measured physical properties of a large number of sediments collected by Li et al. (2013) in the southern Huanghai Sea showed that the wet bulk density of sediments measured in the laboratory is generally higher than that of sediments measured on the deck, and the laboratory measurements of water content are generally lower than those obtained on site. The results presented here are closely related to the dynamic liquefaction and re-consolidation of the samples of silty sand, sandy silt and silt during transport. After the sediment sample was liquefied for a period of time, the pore water was drained and re-consolidation (i.e., compaction) occurred, which led to the aforementioned changes in physical properties. From Figure 10 , it can be seen that there are still some measurement data that show that the laboratory measurements were not much different from the on-site index. These data are the result of sediment samples exhibiting dynamic thixotropy, such as clayey silt. For sediment samples with thixotropy resumption, little water loss and no significant changes in wet bulk density were observed during the dynamic–static change. Therefore, a comparative analysis of the measured data from the southern Huanghai Sea confirms that liquefaction, re-consolidation and the resumption of thixotropy in sediment samples affect their V p − f / V p − l values based on changes in sediment physical properties.

In order to further analyse the factors that may influence the differences in the sound velocities between the those measured in the laboratory and the on-deck for the sediments from the Huanghai and Bohai seas, we statistically analysed the relationship between the physical parameters measured in the laboratory and V p − f / V p − l . Due to the limitations of on-site testing of the physical properties of sediment samples, the description of the relationship between the sound velocities and the physical properties of sediment samples in this sedimentary acoustic study, a laboratory measurement index was adopted (Hamilton, 1972; Lu and Liu, 2008; Meng et al., 2012; B.H. Liu et al., 2013), Therefore, in this study, the physical properties of sediment samples were not tested on-site, and only the effects of physical parameters measured in the laboratory on V p − f / V p − l were analysed. ( Figure 11 ).

Because different types of sediments differ substantially in their physical properties, and different types of sediments were represented by fewer samples, the impacts of the physical properties of different types of sediment on V p − f / V p − l were not separately analysed. From a broad perspective, the impacts of the lab-measured bulk density (wet and dry bulk density), water content and plasticity index on V p − f / V p − l was more pronounced than those of laboratory-measured porosity and permeability ( Figure 11 ). The V p − f / V p − l decreased with the increase of bulk density when the wet bulk density was lower than 18.0 kN/m ³ and the dry bulk density was lower than 13.0 kN/m³. When the wet bulk density was higher than 18.0 kN/m³ and the dry bulk density was higher than 13.0 kN/m³, V p − f / V p − l increased with increasing bulk density ( Figures 11A, B ). Additionally, when the water content was above 40.0%, V p − f / V p − l showed a significant positive correlation with water content ( Figure 11C ). When the plasticity index was higher than 9.0, V p − f / V p − l also showed a significant positive correlation with the plasticity index ( Figure 11D ).

Porosity is one of the most important physical parameters, and also has exhibited the most significant relationship with sound velocity statistics in previous empirical acoustic models used for studying the sound velocities and physical properties of sediments (Pan et al., 2006; Meng et al., 2012; Kan et al., 2014). However, the effect of porosity on V p − f / V p − l is not significant, which shows that the absolute value of the sediment pore size is an important parameter that affects the sound velocity. However, the effect of the absolute value of the sediment pore size on the difference between the velocities measured on-site and those measured in the laboratory is not obvious. In our analyses, it could be seen that compaction and decreases in the porosity of sediments during dynamic liquefaction and re-consolidation were the important factors influencing the V p − f / V p − l of silty sand, sandy silt and silt. It can be inferred that the variation of porosity under dynamic actions is also likely to be an important factor affecting V p − f / V p − l , which needs further study.

Based on a comparative analysis of the physical properties of sediments from the southern Huanghai Sea measured on site (i.e., shipboard) and in the laboratory, it was found that the large bulk density and high water content measured in the laboratory indicated that the values measured on the deck were also generally high. In the process of dynamic liquefaction, re-consolidation and thixotropy, changes in water content and bulk density are also large (Meng et al., 2012). Therefore, the effect of laboratory-measured bulk density and water content on V p − f / V p − l can represent the effect of bulk density and water content on V p − f / V p − l . The plasticity index of sediments is a physical indicator determined by the sediment particle size, and is independent of sediment disturbances. The effect of the plasticity index on V p − f / V p − l can represent sediment type. Comparative analyses showed that the effect of the particle size ( Figure 9 ) and plasticity index ( Figure 11 ) on V p − f / V p − l is consistent, which therefore demonstrates that the two classification methods, Shepard’s ternary diagram of Marine Investigation Regulations (GB/T12763.8-2007) and the plasticity index of sediments in marine geotechnical engineering, are different in approach but equally satisfactory in their results in terms of the influence on V p − f / V p − l in the Huanghai and Bohai seas.