The site SH3 and site SH1 samples used in this study were obtained by the Guangzhou Marine Geological Survey in the Shenhu area of the SCS in 2007 (Su et al, 2015). Site SH3 is of particular interest because of its high hydrate saturation and thicker hydrate-bearing sediments. The sediments are mainly muddy but include silt with clay or sand as a secondary component. In this study, 31 samples were obtained from site SH3 from 0–25 m below seafloor, and the main trace elements were measured separately. And the site SH1 is very closed to the site SH3. At site SH1, the sediments are same as site SH3. However, there is no gas hydrate in sediments. In this paper, 11 samples collected from the site SH1. The δ³⁴S values, main and trace elements were measured separately.

Core 973-4 has four distinctive lithological intervals, as summarised from Wang (2013). The sediments in the intervals of 12–450 cm and 450–530 cm are composed of celadon silty clay and grey clayey silt, respectively. The 530–603 cm interval consists of celadon clayey silt with abundant foraminiferal shells. The 603–1 375 cm interval consists of grey silty clay with patchy hydrogen-sulphide textures (Zhang et al, 2018b). In this study, we selected 64 samples from site 973-4 and conducted analytical tests.

Major elements in the sediments were analysed at the Analytical and Testing Center of the Guangzhou Institute of Energy Conversion, Chinese Academy of Sciences. X-Ray fluoresccene spectrometry (XRF; Thermo ARL ADVANTta IntelliPowerTM 2000, Waltham, MA, USA) was used to determine the primary element contents. The measured XRF spectral data were converted to elemental and oxide contents using UniQuant semi-quantitative analysis software. The whole rock trace element contents were analysed using an inductively coupled plasma mass spectrometer (ICP-MS; Agilent 7700e, Santa Clara, CA, USA) at the analysis and testing centre of the Wuhan Shangpu Analysis Technology Co., Ltd.

Total organic carbon (TOC) was measured using a Heraeus (Ulm, Germany) CHN-O Rapid Elemental Analyzer. Before testing, the appropriate sediments were selected, excess calcium carbonate was removed by adding 10% HCl, and the sample was diluted several times with distilled water before being placed in an oven at 50°C. The TOC test instrument accuracy was greater than 1%. The aforementioned experimental process was conducted at the Guangzhou Institute of Geochemistry, Chinese Academy of Sciences. Total sulphur and total carbon tests were performed using an element analyser (Vario EL cube; Elementar, Alb-Donau-Kreis, Germany) with an accuracy of 0.1%. To burn and decompose it, the rock powder sample was oxidised catalytically in oxygen at a high temperature, producing a mixture containing C and S gases. These mixed gases were then placed into contact with tungsten oxide and copper to convert them into CO₂ and SO₂ gases, at which point the sample was separated by a column. A thermal conductivity detector was used to calculate the C and S component contents by comparing the test sample with a standard sample. The aforementioned pretreatment and experimental procedures were completed at the Analytical Testing Center of the Guangzhou Institute of Energy Research, Chinese Academy of Sciences.

Reduced inorganic sulphur (RIS) in the samples was extracted using CrCl₂ and HCl. This reduction method can release sulphur contained in all sulphide lattices, so the obtained sulphur content was almost equal to the total content of chromium reducible sulphur (CRS). The reduced sulphur was converted to H₂S gas, which was blown out by the carrier gas (nitrogen), and then precipitated by passing it through an AgNO₃–NH₃H₂O solution to obtain silver sulphide. The silver sulphide was filtered, dried, and weighed. Then, the weight percentage of CRS in the deposit was calculated according to the weight of precipitated silver sulphide. The detailed experimental steps can be found in Panieri et al (2017).

During the experiment, quantitative addition of pyrite as a standard sample resulted in a recovery rate of 88%–92%. The silver sulphide was then sent to the State Key Laboratory of Biogeology and Environmental Geology of China University of Geosciences (Wuhan, China) for analysis of δ³⁴S using an elemental analysis–isotope ratio mass spectrometer (Delta V Plus). All results were reported in standard delta notation as per mil deviations from the Vienna-defined Canyon Diablo Troilite (VCDT). The standard deviation of the measurements was less than 0.2‰ (VCDT). Measurement errors of about 0.2‰ (1σ) were calculated from replicate analyses of the IAEA international standards IAEA S1 (−0.3‰), IAEA S2 (+22.7‰), and IAEA S3 (−32.3‰).

At site 973-4, all the data changed dramatically at 400–600 cm below seafloor. Among them, MgO and SrO showed opposite trends. The indicators of primary productivity of Cu/Al and Ni/Al show the same trend of change and both decrease rapidly at 400–600 cm below seafloor. The changing trends of U and Mo show opposite trends. However, at site SH3, there are two obviouly change of main and trace element at 7–15 m below seafloor and 17.5–20 m below seafloor, all the data have increased invariably. But the study found that all mineral indicators have a greater increase at 8–15 m below seafloor than at 17.5–22 m below seafloor. The main changes at site SH1 in the 25–36 m below seafloor position. In the 25–36 m below seafloor, all the data declined.

At site SH3, the δ³⁴S values increase from −49‰ to −35‰ at 10–15 m below seafloor, and later, the δ³⁴S values increase from −35‰ to −11‰ at 17.5–20 m below seafloor. At site SH1, the δ³⁴S values increase at 25–36 m below seafloor. At site 973-4, the δ³⁴S increase from −40‰ to 10‰ at 650–900 cm below seafloor.

Generally, in oxic and suboxic sediments, reduced sulphur and organic carbon contents show a correlation with an average S/C ratio of 0.36 (Berner, 1981; Li et al, 2016a, 2016b) (Fig. 2). In euxinic sediments, similar to those of the Black Sea, the S/C ratio is always higher than 2.05 (Leventhal, 1983). Moreover, in methane-rich sediments, where AOM is the predominant process, AOM-induced pyrite formation may increase the S/C ratio of sediments (Lim et al, 2011; Sato et al, 2012). At site SH3, the S/C ratio of the 0–8 m below seafloor interval was very close to 0.36 (Fig. 2). However, in the same layer, the U/Al and Mo/Al ratios were not enriched (Figs 3e and k), indicating an oxidising environment. This result is consistent with the S/C ratio in the oxic sediments of the Black Sea. The S/C ratios in the 8–15 m and 18–22 m below seafloor intervals were higher than 0.36 (Fig. 2). This result may indicate that there are two stages of AOM at SH3. However, at site 973-4, the S/C ratio of the 700–869 cm below seafloor interval was higher than 0.36 (Fig. 2), and the ratio was close to 0.36 at other layers of site 973-4, indicating that AOM occurs only in the 600–800 cm below seafloor interval of site 973-4 (Fig. 2).

Total sulphur (TS) and TOC showed a positive correlation in the 0–6 m below seafloor interval of site SH3, where the average S/C ratio was 0.36. The changes with depth of TiO₂, Al₂O_3, and SiO₂ were also examined. These increased significantly in the 0–6 m below seafloor depth, indicating a high land-derived input of organic matter to the sediments (Figs 4a–e). The low reactivity of land-derived organic matter and the normal oxic marine environment may account for the lower S/C ratio. At site SH3, high contents of CRS were observed in the 8–15 m and 18–22 m below seafloor intervals (Fig. 5). In these two layers, the concentration of TS increased significantly (Fig. 2); however, the TOC content remained constant, with an average of 0.35 (Fig. 2), which excluded the mechanism of additional CRS formation in a euxinic environment. The main reason for this is AOM. At the same time, high values of δ³⁴S (Fig. 5), greater than the value of average seawater sulphate δ³⁴S of 21‰, occurred in the 8–15 m and 18–22 m below seafloor intervals (Fig. 5a). Many studies have shown that a significant positive bias of δ³⁴S indicates the occurrence of AOM (Wu et al, 2011; Egger et al, 2016; Lin et al, 2017; Zhang et al, 2018b).

As mentioned previously, the S/C ratio at site 973-4 was higher than 0.36 in the 700–869 cm below seafloor interval and close to 0.36 in the other layers (Fig. 2). This indicates a potential AOM effect in the 700–869 cm below seafloor interval. The concentrations of TiO₂ and Al₂O₃ (Figs 4a–e) decreased at 700–890 cm below seafloor, indicating that the input of TDM was low. However, Shao et al (2007) considered that there are contour current sediments at the top of the SMTZ at site 973-4, there is no significant increase of authigenic carbonate in the modern SMTZ (600–860 cm below seafloor). At site 973-4, the study found a marked decrease in carbonate at the top of the SMTZ. The value of δ³⁴S increased obviously at 700–860 cm below seafloor. Combined these data confirm the occurrence of AOM at 700–860 cm below seafloor, the same increase trend of CRS was found at 27–37 m below seafloor at site SH1 (Fig. 6).

As discussed above, there are two SMTZ, at 8–15 m below seafloor and 19–25 m below seafloor, respectively, at site SH3. Moreover, high Ba concentrations (exceeding 420×10⁻⁶) occur in the 8–15 m and 19–23 m below seafloor intervals. The peak Ba concentrations within these zones exceed 480×10⁻⁶. The Ba fronts (defined here as intervals with Ba>430×10⁻⁶) at site SH3 are diffuse, extending over several meters with multiple peaks. For example, at site SH3, a Ba front extends from 8 m to 15 m below seafloor and contains at least two peaks with Ba>430×10⁻⁶. Additionally, a second Ba front was also found at the top of the current SMTZ (19–25 m below seafloor) of site SH3 (Dickens, 2001). In general, two primary components contribute to high Ba concentrations. For example, high concentrations of aluminosilicate phases and barite may increase the Ba content. The Ba from aluminosilicates is generally immobile, but the Ba hosted in barite is influenced by ${\rm {SO}}_4^{2-} $. During AOM, most of the ${\rm {SO}}_4^{2-} $ is consumed by methane, so barite decomposes at the bottom of the SMTZ. The Ba²⁺ migrates upward until it reaches an environment rich in ${\rm {SO}}_4^{2-} $, where it re-stabilises gradually. Therefore, the Ba front can help identify the position of the SMTZ. However, it is necessary to eliminate the major elements influencing the Ba fronts, and an alternative means of displaying downcore changes in sedimentary Ba content is to normalise Ba to Al (Dickens and Owen, 1996). The downcore profiles of Ba/Al at sites SH3 and 973-4 are very similar in shape to the profiles of bulk sediment Ba (Fig. 7). The similarity of the two profiles strongly suggests that the observed Ba fronts originate from digenetic accumulations of labile Ba. A double Ba peak distribution was found at the top of the SMTZ and the paleo-SMTZ, indicating instability of the paleo-SMTZ and the SMTZ through geological history and that there is a trend of vertical movement. In contrast to the Ba front at site SH3, there is a decrease in concentration of Ba at the top of the SMTZ at site 973-4. At the same time, most of the major and trace elements are lower. Shao et al (2007) considered that contour current sediments occur at the top of the SMTZ at site 973-4, and Xie et al (2019) thought that the contour current sediments caused the SMTZ to migrate downward quickly. These conclusions are consistent with the results of this study. This is the reason why there is no Ba front at the top of SMTZ. Previous studies have shown that the formation of barium sulphate in the SMTZ requires two conditions: a stable ion source and a relatively stable SMTZ position. Because the presence of an SMTZ makes the supply of Ba²⁺ stable and sufficient, the Ba peak in the study area is more affected by the stability of the SMTZ. Li et al (2017) considered that the movement of the SMTZ is influenced by the methane flux from the bottom sediments. When the methane flux is higher, the SMTZ is shallower; and when the methane flux is lower, the SMTZ is deeper. Lin et al (2018) found that the sedimentation rate also influences movement of the SMTZ. Therefore, to understand the movement mechanism of the SMTZ, it is necessary to compare the change in methane flux at the bottom and the change in deposition rate at the top of the SMTZ fully.

Determining the size of the methane flux is an important process. Under normal circumstances, the SMTZ is relatively shallow when the methane flux is relatively large and is relatively deep when the methane flux is relatively small. Thus, the SMTZ can partially indicate the amount of methane flux at the bottom. Dickens (2001) also suggested that an obvious Ba front requires at least 10⁴ to 10⁶ years to form. However, if the methane flux is highly variable or decreases over time, a small Ba front should exist at multiple depths. At site SH3, double Ba fronts at the paleo-SMTZ and the current SMTZ are not obvious. Therefore, the methane flux from the bottom must have been variable over geological time (Dickens, 2001). In fact, sedimentary organic matter can affect the diffusion of CH₄; thus, the CH₄ diffusion flux in sediments of the study area can be calculated according to Fick’s laws (Eqs (2) and (3)):

where J is the diffusion rate (mmol/(m²·a)), $ \varphi $ is the sediment porosity, Ds is the sediment diffusion coefficient (m²/a), c is the CH₄ concentration (mmol/dm³), and x is the sediment depth (m). Ds can be defined as follows:

where n=3 (lithology factor of silty clay) and Do=1.4×10⁻⁵ cm²/s (initial diffusion coefficient of methane at 20°C). In the Shenhu area, the average sediment porosity and grain density for sediment at site SH3 are 40% and 2.6 g/cm³, respectively (Jin et al, 2019).

According to Table 1, the methane flux at the bottom of the southwestern Taiwan Basin is significantly greater than the methane flux in the Shenhu area. However, it is strange that site 973-4 in the southwestern Taiwan Basin did not show a significant upward movement of the SMTZ due to the increase in the bottom methane flux. Instead, through a comprehensive analysis of the main trace elements and isotopes, it is believed that the SMTZ has migrated downward quickly. Analysis of sediment Mg/Ca and Sr/Ca ratios at site 973-4 (Fig. 8), site SH1, and site SH3 revealed that the sediments at site 973-4 received greater inputs of sediments than site SH1 and site SH3. Therefore, although there is certainly a high flux of methane at the bottom of site 973-4, the deposition rate at the top is significantly higher than the other two sites and the movement at the site is controlled mainly by changes in deposition rate. Affected by the rapid contour current sedimentation, the SMTZ at site 973-4 dropped quickly by approximately 1 m.

Site SH1 is the station least affected by debris input (Fig. 8) because Mg/Ca and Sr/Ca ratios (Fig. 8) of its sediment indicate that it has greater deposition of biological debris. Because site SH1 is very close to site SH3, the deposition rates at the top are basically the same. However, the difference is that the position of the SMTZ at site SH1 is about 10 m deeper than the current SMTZ position at site SH3. As the top deposition rate is basically the same, the main reason for the difference is the difference in bottom methane flux. During the GMGS1 voyage in 2007, hydrates were found in the 190–205 m below seafloor horizon of SH3. Hydrates are stable materials formed in a high-pressure and low-temperature environment and sometimes are susceptible to disasters such as seabed landslides. The impact of such events may cause decomposition, resulting in seafloor leakage of methane. Therefore, the methane flux at the bottom of site SH1 may be affected by the decomposition of natural gas hydrate at the bottom of site SH3. Therefore, the SMTZ distribution is obviously deeper than that at site SH3. This conclusion is consistent with the fact that the SMTZ is relatively shallow when the methane flux is relatively large.

Site SH3 is the most complex of the three sites, and it is difficult to determine whether the deposition rate or the methane flux dominates the movement of the SMTZ. However, the detailed study of the Ba peak is conducive to the understanding of its basic situation. In general, diagenetic front composed of labile Ba exists in shallow sediments above the paleo-SMTZ and the current SMTZ of site SH3. Simple calculations demonstrate that these Ba fronts were formed in the past. The minimum time required to precipitate a Ba front can be estimated from the quantity of Ba in the front and the supply of Ba to the front. At site SH3, the total area concentration of the double Ba front above the paleo-SMTZ is 16.867×10⁻⁶; however, the value of the double Ba front at the current SMTZ is 32.98×10⁻⁶.The downward supply of solid Ba determines the minimum time required to generate the Ba in a front. The average concentrations of solid Ba above and below the front at the paleo-SMTZ at site SH3 were 380×10⁻⁶ and 400×10⁻⁶, respectively (Fig. 7), indicating a net loss of 20×10⁻⁶ Ba as solid particles are buried through the ${\rm {SO}}_4^{2-} $/CH₄ interface. The average concentrations of solid Ba above and below the front at the current SMTZ of site SH3 are 450×10⁻⁶ and 523×10⁻⁶, respectively (Fig. 7), indicating a net loss of 73×10⁻⁶ Ba as solid particles are buried through the ${\rm {SO}}_4^{2-} $/CH₄ interface. The average sediment porosity and grain density of sediment at site SH3 are 40% and 2.6 g/cm³, respectively (Jin et al, 2019). With a sedimentation rate of 13 cm/ka, the time required to form the Ba front at the top of the paleo-SMTZ was about 2 times required to form the Ba front at the bottom of the current SMTZ. Therefore, the formation time of the SMTZ was longer than that of the paleo-SMTZ. In fact, double Ba peaks appeared at the top of the paleo-SMTZ and current SMTZ of site SH3, indicating that the SMTZ of these two periods was not stable but rather was in a process of moving slowly.

At site 973-4, the SMTZ also showed a tendency to move vertically. However, the mechanism of the movement at site 973-4 was significantly different from that of the modern SMTZ of site SH3. At site 973-4, the TOC/TN ratio did not change from the shallow portion to the deep portion, indicating that the input of the detrital material was stable; however, the result showed a significant decrease at the top of the SMTZ. The carbonate and Ba contents also decreased significantly. Xie et al (2019) believed that the geological time at the top of the SMTZ of site 973-4 was during the last glacial maximum, when there was an obvious contour current deposition rate of 91 cm/ka. However, Fe₂O₃ and Al₂O₃ did not increase. Shao et al (2007) analysed many seismic sections in the northern part of the SCS and discovered the existence of contour current flow at the bottom of site 973-4. The contour current deposition at the bottom of the seabed was the main reason for the high rate of deposition of the sediment. The high sedimentation rate deposits caused the SMTZ to move downward quickly, made the biogenetic barite unstable and diluted the carbonate in the sediment. The high sedimentation rate sediments are also the main reason for the reducing environment of site 973-4. The clear enrichment of U–Mo at the top of the SMTZ indicates a closed reducing depositional environment. At the same time, the primary productivity at the top of the SMTZ was also significantly reduced because both Cu/Al and Ni/Al were significantly negative (Figs 3c and d). Thus, the contour current deposition had an important influence on the formation of the submarine sedimentary environment, the movement of the SMTZ, and the enrichment of elements.

At site SH1, the indices of detrital materials of Ti, Al, and Fe all decreased at the SMTZ (Fig. 6). In the paleo-SMTZ layer at site SH3, the sedimentary environment was very similar to that of site 973-4, a closed reducing sedimentary environment, and the indicators of terrestrial debris in the paleo-SMTZ layer were obviously increased. However, the difference of site 973-4 is that the input of TDM at the paleo-SMTZ of site SH3 was not very high, only 20 cm/ka (Hsu et al, 2014), so the migration of the SMTZ was relatively slow and enriched carbonate and pyrite formed. The Ba in this layer exhibits a double peak, which also confirms the slow movement of its SMTZ from the side. It is foreseeable that it may hardly find any signs of an SMTZ in a depositional environment with a high deposition rate but low methane flux because the rapidly moving SMTZ is not conducive to the formation of the various minerals and the occurrence of the related chemical reactions (Lu et al, 2017).

The SMTZ at site SH3 in the Shenhu area was the most stable. Hence, the AOM signal in its sediment was the strongest. This reduced input of terrigenous detrital materials was conducive to improvement of primary productivity. At the modern SMTZ, the Cu/Al and Ni/Al ratios at site SH3 were also increased (Figs 3c and j), indicating higher primary productivity. The U–Mo coupling system indicated that the depositional environment of the modern SMTZ was suboxic, and such a depositional environment is very beneficial to microbial activity. Therefore, the input of terrigenous debris is an important influencing factor.