1 INTRODUCTION

Cold seeps, a phenomenon of natural fluid leakage, are formed by the upward migration of fluids mainly composed of water, hydrocarbon, hydrothion, and fine-grained sediments (Joseph, 2017) and generally related to the dissociation of gas hydrates (Chen and Cathles, 2005; Feng and Chen, 2015) or the upward migration of gas and oil beneath the seafloor along geologically weak zones, which are widely distributed around the world and play an indicative role in the deep exploration of oil and gas (Suess, 2014; Feng et al., 2018).

Seeps are closely connected to material circulations (such as carbon and sulfur etc.) and of great importance to trace the deep origin of life (Boetius and Wenzhöfer, 2013; Suess, 2014). Typically characterized by rich methane, cold seeps play a crucial role in reconstructing paleoclimate and detecting deep hydrocarbon resources (Feng et al., 2018). Each cold seep presents its unique nature as it is heterogeneous both in time and space, stressing the necessity of specific and regional research on cold seep.

After the first discovery at a depth of 3 200 m on the Florida cliff in the Gulf of Mexico (Paull et al., 1984), cold seeps and cold-seep carbonates were found in both active and passive continental margins (Suess, 2014; Feng et al., 2018). To date, cold seeps have been found from the tropics to the poles, and from shallow continental shelves to abyssal plains or trenches (Roberts and Aharon, 1994; Campbell, 2006; Roberts et al., 2010; Suess, 2014).

During the upward migration of cold-seep fluid, methane in the fluid and sulfate diffusing downward into the sediments cause anaerobic oxidation of methane (AOM) and sulfate reduction reactions (Eq.1) (Reeburgh, 2007).

Under the joint action of methane-oxidizing archaea and sulfate-reducing bacteria, the alkalinity of the surrounding fluid increases, causing methanederived carbonate to locally supersaturate and form authigenic carbonate in the sulfate methane transition zone (SMTZ) (Eq.2) (Raiswell, 1988; Chen et al., 2015; Peketi et al., 2015).

Cold-seep carbonate is an important product of cold seep in shallow seafloor burial, whose morphology, mineralogy, and other characteristics contain abundant information of cold seep (Chen et al., 2005; Suess, 2005; Han et al., 2008). Researches on cold-seep carbonates have revealed the distribution and activity of cold seeps in geological history and the interaction between the formation environment of cold-seep carbonates and surrounding environment, especially in the case of paleo-cold seeps wherein fluid activity ceased (Naehr et al., 2007; Feng and Chen, 2015; Yang et al., 2018). Further investigation is required to explore the characteristics of fluid evolution in response to the geochemical characteristics of different occurrences. In addition, the carbon and oxygen isotopes of cold-seep carbonates vary across different sea areas, reflecting the variations in fluid leakage rates and sedimentary environment of cold-seep fluids from different sources (Chen et al., 2005; Han et al., 2008, 2014; Tong et al., 2013; Feng and Chen, 2015; Liang et al., 2017; Feng et al., 2018). Trace elements, especially redoxsensitive elements (e.g. Mo, U, Cr, Co) and the rare earth elements (REE) of cold-seep carbonates, show great potential to help reconstruct the paleoenvironment (Ge et al., 2010; Smrzka et al., 2020).

Shenhu area where the first hydrate trial situated in the South China Sea (SCS) is a vital platform for hydrate and cold-seep research in China. In 2004, several cold-seep carbonates were detected by China's Haiyang-4 ("ocean" in Chinese) vessel in this sea area where gas hydrate samples were extracted thereafter; and the first trial production of gas hydrate was successfully completed with a total gas production of 309 000 m³ in July 2017, which strongly proved the significance of cold-seep activities in the development of gas hydrates (Suess, 2005; Zhang et al., 2007, 2019). Studies have shown that different sedimentary environments and fluid evolution characteristics tend to produce cold-seep carbonates with various morphological characteristics, mineral compositions, and geochemical characteristics (Feng et al., 2009; Han et al., 2013; Lu et al., 2015). Previous studies on cold-seep carbonates in this area focused on the geochemical characteristics of fluid, biomarkers, and fluid evolution of chimney-shaped cold-seep carbonates (Ge et al., 2010; Feng et al., 2018; Zhang et al., 2019).

A large, active submarine cold seep, named Haima cold seep, was discovered by the remotely operated vessel (ROV) Haima ("hippocampus" in Chinese) in 2015 and named after the ROV name. The Haima cold seep is abundant in authigenic carbonates, exposed gas hydrates, and widely developed coldseep biota (Liang et al., 2017; Feng et al., 2018).

Here, we firstly observed and described the morphology of the cold-seep carbonates from Shenhu area and then carried out a detailed study on the mineralogical and geochemical characteristics (element compositions and isotopes of carbon and oxygen) of the carbonates from Shenhu. Focusing on Shenhu area, we further compare Shenhu cold-seep carbonates with those from Haima cold seep, in order to understand the regional specificity of Shenhu coldseep activities, meanwhile, to demonstrate the necessity of more detailed cold-seep research in various regions, especially in the SCS (known as its complex tectonic settings).

2 GEOLOGICAL SETTING

Shenhu area is located in the Baiyun Sag of the Zhuer Depression of the Zhujiang (Pearl) River Mouth Basin (Fig. 1), which is in the transition zone between the northern continental slope and the Central Sea Basin of the SCS. The seabed has a highly undulating topography with various geomorphic types, such as sea valleys, domes, and erosion grooves, with water depth in the range of 1 000–3 000 m, water temperature of 3.3–3.7 ℃, and geothermal gradient of 45–67.7 ℃/km (Su et al., 1989; Yu, 1990; Zhang et al., 2018). Rich oil and gas resources have been discovered in this area (He et al., 2009; Li et al., 2016). The neotectonics led to the formation of a large active diapir belt, while fluid discharge and overflow occurred on the seabed, resulting in the abnormal development of mud diapirs and gas chimneys and the formation of arch anticlines at different depths, which provided fluid passages for the migration of gas hydrates (Ge et al., 2010; Wu et al., 2010; Liang et al., 2014; Su et al., 2018).

The Haima cold seep is located in the uplift zone of the Qiongdongnan Basin in the northwestern region of the SCS (Gong et al., 1997; Xie et al., 2006). The combination of thick sedimentary sequence and high geothermal gradient, coupled with faults or diapirs, significantly promoted the generation and migration of natural gas in the basin (Sun et al., 2017). Bottomsimulating reflectors (BSR) and gas chimneys are widely distributed at the bottom of the basin, which further indicates the presence of gas hydrates in the area (Hui et al., 2016; Liang et al., 2017).

3 MATERIAL AND METHOD

A total of 17 samples were investigated in this study. 15 samples were from Shenhu area (SH2017-1–SH2017-15) and were obtained during the 2017 survey trawl (China's Haiyang-4 vessel, the Guangzhou Marine Geological Survey). The Shenhu samples were collected at water depths of 500–700 m. The other 2 samples (ROV2 and ROV12) from the Haima cold seep were collected from the seabed by the ROV of China's Haiyang-6 in 2015. In the study area of Haima cold seep, the temperature of bottom water was 3.0 ℃, and the water depth was approximately 1 370–1 390 m.

All samples were washed with seawater, dried naturally, and stored at room temperature (25 ℃) in the Key Laboratory of Marginal Sea and Oil and Gas Exploration, School of Marine Science, Sun Yat-sen University. The morphological characteristics of each sample were observed, photographed, and recorded in detail.

The X-ray diffraction (XRD) analysis of the carbonates was carried out at the School of Chemistry, Sun Yat-sen University. Firstly, small pieces were cut from the fresh surface of the samples and were ground to powder less than 200 meshes using an agate mortar. Then, a Bruker D8 Advance X-ray diffractometer was used for the analysis. The parameters were set as Cu target Kα ray, power of 40 kV/30 mA, 2θ scanning range of 3°–85°, incidence slit of 1 mm, and scanning speed of 4°/min. The MDI Jade9 software was used to analyze the d values and relative strength of the obtained XRD spectra to determine the mineral composition. Meanwhile, according to the ratios of the maximum diffraction peaks of the minerals obtained in the XRD spectra to the maximum diffraction peaks of the standard minerals, the percentage contents of the main minerals were semiquantitatively determined, and the analysis accuracy was ±5 weight (wt.)%.

The analysis of major and trace elements (including REE) was done at the University of Queensland, Australia. The procedure is as follows. First, about 200-mg samples were weighed and placed in a Teflon centrifuge tube, then washed twice with 2-mL MilliQ water. Two milliliter acetic acid was then added in and sonicated for 1 h, followed by heated at 80 ℃ for 12 h for complete dissolution. After centrifugation, the supernatant was collected into a pre-cleaned Teflon beaker. After the supernatant was dried up at 120 ℃, 2-mL 7-mol/L HNO₃ was added. Upon fully dissolution followed by dry-down again, 2-mL 1-mol/L HNO₃ was added for dissolution for 12 h. Then, 50-μL solutions were extracted and added with 0.5-mL internal standard and 4.45-mL 2% HNO₃, which were all precisely weighed. After centrifugating for 10 min, the major and trace elements of sample solution were analyzed on Agilent 7900 inductively coupled plasmamass spectrometry (ICP-MS). BHVO-2, W-2, and JCP-1 were used as calibration standards and BIR-1 as cross check. The enrichment factor (EF) of trace elements in carbonates was calculated by X_EF=[(X/Ti)_sample/(X/Ti)_PAAS], where X represents the index element and PAAS represents the standardization of the Post Archean Australian Shale (PAAS; McLennan, 1989). Al in pelagic seawater will be scavenged by organic matter, so it may be unsuitable to use Al to represent terrestrial sediments here (Murray and Leinen, 1996). Therefore, Ti was used in this study to replace Al as the standard of terrestrial sediments. Enrichment factors can be used effectively to evaluate the degree of autogenous enrichment of trace elements (Tribovillard et al., 2006). Generally, an enrichment factor greater than 1 indicates enrichment relative to the standard, an enrichment factor greater than 3 indicates detectable enrichment, and an enrichment factor greater than 10 indicates moderate to strong enrichment (Algeo and Tribovillard, 2009).

In this paper, Ce/Ce^*=2Ce_N/(La_N+Nd_N). Log(Ce/Ce^*) stands for Ce anomaly. Eu/Eu^*=2Eu_N/(Sm_N+Gd_N), Pr/Pr^*=2Pr_N/(Ce_N+Nd_N), where the subscripted N represents the relative value of the standardization of PAAS (McLennan, 1989).

The measurements of carbon and oxygen isotopic compositions were also done at the University of Queensland, Australia. The powder samples reacted with 105% orthophosphoric acid at 90 ℃ to release CO₂ in the Dual Inlet system and Multiprep on-line sampler system. The product of CO₂ was imported into IsoPrime 100 mass spectrometer. The results were presented relative to VPDB standard. Analytical precision of C isotope and O isotope was better than 0.08‰ and 0.10‰, respectively.

4 RESULT 4.1 Hand specimen observation

The 15 cold-seep carbonates collected from Shenhu area (Fig. 2) are numbered as SH2017-1–SH2017-15 (represented by 1–15 in this paper). Most of the samples are bluish-black with irregular crust form or in the form of chimneys and blocks. Yellowish-gray surface of some samples show obvious oxidized features, holes, and biological debris. Crust cement is visible in several samples. Samples 1, 4, 7, 8, and 10 appear as a chimney. The diameters of the central fluid channel and the outer ring of sample 1 with an evident arc-shaped cross section, are approximately 3 cm and 7 cm, respectively, which shows the traces of biological activities. Sample 4 has a central fluid channel with a diameter of 2 cm and contains bioclasts. Sample 10 has the most complete chimney shape with a diameter of about 3 cm. There are clear bluish-gray fluid migration channels without obvious cement features shown on samples 7 and 8.

Sample 2 is a compact block with crust cement on the surface.

Other samples are described as crust with the length of 6.5–10.5 cm. Millimeter-sized holes are observed on the surface of samples 3, 9, and 15, while sample 6 is darker than other samples exhibits large pores. Sample 9 presents cement and also channels for fluid migration shown on its cross section. Sample 11 also has crust cement and traces of tubular biological activity on its cross section. Samples 12 and 13 contain several bioclasts. Sample 14 that is denser than other samples show no obvious cementation characteristics. Sample 15 that is in a long strip-like shape has several fluid channels in the cross section.

4.2 XRD test result

The XRD analysis shows that the carbonate contents of the samples from Shenhu area are in the range of 20%–65% with an average of 43%. The carbonate content of sample 2 is the highest and that of sample 10 is the lowest (only 20%). The carbonate minerals in the samples are mainly aragonite with different proportions of calcite and dolomite. The carbonate minerals of samples 1–13 are aragonite and calcite, and those of samples 14 and 15 are calcite and dolomite. The clastic quartz content is more than 50% in some samples and that of sample 10 is even as high as 70%. The types of clay minerals are mainly montmorillonite, illite, and chlorite.

The carbonate contents of the two samples from Haima cold seep are significantly different, with that of ROV2 being 65% and ROV12 being 10%. The carbonate minerals are mainly aragonite and calcite, while the quartz content of ROV12 is 85%. The types of clay minerals are mainly kaolinite, chlorite, and illite (Fig. 3).

4.3 Main and trace elements

The concentrations of some elements in the samples are listed in Table 1. MgO contents of the 15 samples from Shenhu area vary significantly in the range of 0.95%–14.12%, while higher values are shown on samples 14 and 15, i.e., 12.99% and 14.12%, respectively. Fe contents of all the Shenhu samples are low in the range of 0.14%–0.35%. The highest MnO content in Shenhu samples is 0.36% while the lowest is 0.05%. TiO₂ contents are in the range of 3.08–10.72 μg/g with an average of 5.65 μg/g.

MgO contents of the two samples from Haima cold seep are 5.86% for ROV12 and only 0.96% for ROV2. TiO₂ contents of the samples from Haima cold seep are in the range of 12.50–16.60 μg/g and higher than those of the samples from Shenhu area.

Sr contents of the Shenhu samples vary from 705.5 to 19 645.4 μg/g. In addition, the contents of Ni, Zn, and Ba are very high. Mo and U contents are 0.02–0.67 and 2.21–25.79 μg/g, respectively. U contents of the two samples from Haima cold seep (ROV12 and ROV2), are 26.26 and 11.93 μg/g, while Mo contents are only 0.04 and 0.24 μg/g.

The REE results for the cold-seep carbonate samples are depicted in Table 2. To define the REE sources of our samples, we standardized the REE values based on Post-Archean Australian Shale (PAAS, McLennan, 1989) and created the REE distribution curves (Fig. 4).

The REE contents of the Shenhu carbonates vary dramatically in the range of 15.46–76.17 μg/g with an average of 43.98 μg/g. The La_N/Sm_N and Yb_N/Sm_N values of the samples are in the range of 0.53–0.89 (mean value of 0.65) and 0.52–0.75 (mean value of 0.60), respectively. These values are lower than the respective ratios for modern seawater (La_N/Sm_N=0.57–1.37, Yb_N/Sm_N=3.31–6.47) (Alibert, 2016). The Y/Ho ratios of the Shenhu samples are in the range of 26.64–30.98 with an average of 29.32.

The total REE contents of the Haima samples are lower than those of the Shenhu samples. The La_N/Sm_N and Yb_N/Sm_N ratios of the Haima samples are 0.51–0.76 (mean value of 0.63) and 0.62–0.71 (mean value of 0.66), respectively, which are similar to those of the Shenhu samples. The Y/Ho ratios of the Haima samples are in the range of 28.33–34.42 with an average of 31.38.

4.4 Carbon and oxygen isotopic composition

The carbon and oxygen isotopic compositions of all the samples are shown in Table 3. The Shenhu samples have relatively negative carbon isotopic composition. The values of δ¹³C vary from -22.34‰ to -59.30‰ VPDB with an average of -41.59‰ VPDB. The difference between the δ¹³C values of the two Haima samples is insignificant (-37.91‰ VPDB for the ROV2 and -37.91‰ VPDB for ROV12).

The oxygen isotopic compositions of the Shenhu samples in the range of 2.05‰–3.90‰ and with an average value of 2.75‰, are obviously heavier than that of seawater. The δ¹⁸O values of the Haima samples show obvious enrichment, and the δ¹⁸O value of ROV2 is as high as 3.76‰.

5 DISCUSSION 5.1 Possible factors affecting morphology

Cold-seep carbonates in the SCS occur in the form of blocks, chimneys, branches, columns, and biochemical reefs (Han et al., 2008; Tong et al., 2013; Feng et al., 2018). Researches have shown that both internal and external factors influence the formation of cold-seep carbonates (Mazzini et al., 2006; Magalhães et al., 2012). Internal factors include the composition and flow rate of fluid in cold-seep seepage and biological activities (Mazzini et al., 2006; Suess, 2014), while external factors include regional tectonics, chemical properties of bottom water, and local erosion (Schlüter et al., 1998; Borowski et al., 1999; Stakes et al., 1999; Díaz-del-Río et al., 2003; Mazzini et al., 2006; Suess, 2014). In this study, the carbonates from Shenhu area were procured from the same site, and there was little difference among the samples in terms of bottom water, regional geology or other external factors. Therefore, it is speculated that mainly internal factors affected the occurrence of the samples. Generally, the block form and dense texture indicate a strong compaction and deep formation depth; chimney shape may inherit from fluid channels, which reflects significant pipeline development and is usually associated with hydrocarbon-rich fluid migration through vertical fractures or hydrocarbon fluid saturation in the pores of submarine sediments (Díaz-del-Río et al., 2003; Ge et al., 2010; Nyman et al., 2010; Feng et al., 2018); crust form may appears in the shallow depth close to the seafloor, possibly suffering the erosion of undercurrent (Magalhães et al., 2012). In samples 1 and 4, the obvious biological features indicate the medium-high flux of seepage fluid. Dense texture of sample 2 is not conducive tofl uid migration, reflecting a greater compaction, and sample 2 could have precipitated at deeper depth during a lower flux. Evident channels shown in our samples present a relatively focused fluid migration (Fig. 2, samples 7 and 8). The quality and quantity of holes observed in the samples may be the indicator of either difference in the strength of fluid leakage or biological activity, and higher porosity generally provides a better fluid migration environment with less compaction (Fig. 2).

5.2 Sedimentary environment of cold-seep carbonates

The mineralogy of cold-seep carbonates can well reflect the sedimentary environment (Greinert et al., 2001; Naehr et al., 2007). The presence of SO₄^2- in pore water shows stronger inhibition effect on the precipitation of calcite and dolomite than on that of aragonite (Aloisi et al., 2000). Therefore, aragonite usually precipitate in the shallower SMTZ, while calcite and dolomite tend to form more deeply (Paull and Ussler Ⅲ, 2008; Gong et al., 2018). The Sr content in our samples from both Shenhu area and Haima cold seep is significantly higher than that of common marine carbonates (610 μg/g), of which low magnesium calcite is usually dominated carbonate facies (Turekian and Wedepohl, 1961). The Sr content of cold-seep carbonates is related to their mineralogy, and Ca²⁺ in aragonite is preferentially replaced by Sr²⁺. Therefore, the content of Sr in aragonite is significantly higher than that in calcite (Peckmann et al., 2001), which is consistent with the results of XRD analysis. Furthermore, the content of aragonite in cold-seep carbonates with Sr content greater than 104 μg/g is relatively high, which results from a shallower formation depth influenced by the active cold-seep activity. The distinct proportion of calcite in samples 2, 7, and 8 indicates the relatively high deposition depth and low cold-seep fluid flow (Fig. 2). Specially, a certain proportion of dolomite appears in samples 14 and 15. Though the precipitation mechanism of dolomite remains mysterious, it generally tends to precipitate under the conditions of high alkalinity, low sulfate concentration and strong biological activities (Warren, 2000; Tong et al., 2012; Lu et al., 2018). Zhang et al. (2012) suggested that it was not the consumption of sulfate but presence of dissolved sulfide that promotes the dehydration of Mg²⁺, which is more conducive to the formation of dolomite under the condition of strongly reduced sulfide. Magalhães et al. (2012) believed the formation of dolomite is related to the erosion of undercurrent. Both samples 14 and 15 show a relatively high MgO content, which may induce the formation of dolomite. However, different occurrences appear in samples 14 and 15: high porosity and distinct fluid migration traces in sample 15 while low porosity and dense texture in sample 14. More diffuse seepage features in sample 15 may be strongly associated with the methane and sulphate gradients and may form during cementation of subsurface permeable layers. It indicates the strong microbial activity in sample 15 as well. Sample 14 may precipitate in the deeper depth and be exposed to the seafloor due to the heavy erosion of undercurrent.

Additionally, the low carbonate contents of samples 5, 10, 11, 14, and 15 with the relatively high contents of quartz emphasize the great effect of terrestrial sediments and deeper formation depth in lower flux seepage. Similarly, clay minerals, which are around 10% in content, may also influence the formation of carbonates.

The carbonates of the Haima cold seep contain significantly higher content of aragonite, suggesting the shallower precipitation depth and stronger coldseep activity than that of Shenhu area. Significant influence of terrestrial deposition also imprinted in Haima samples as the low carbonate content of ROV12 and the certain contain of clay minerals are shown.

In recent years, the enrichment factors of trace elements, especially Mo and U, have been commonly used to trace the redox condition during the formation of cold-seep carbonate (Algeo and Tribovillard, 2009; Palomares et al., 2012; Sato et al., 2012; Tribovillard et al., 2012; Hu et al., 2015). Although there are small deposits of both U and Mo in the ocean, U is dissolved in water in the high state as UO₂(CO₃)₃^4- with a residence time of up to 450 ka in oxidized seawater, while Mo exists in seawater as MoO₄^2- with a residence time of up to 780 ka (Algeo and Tribovillard, 2009; Hu et al., 2015). Under the anoxic condition without free H₂S in the water column, U (Ⅵ) is reduced to UO₂ during precipitation at the redox boundary of Fe (Ⅱ–Ⅲ), while Mo needs to be deposited at sedimentary depths where H₂S is present (Zheng et al., 2002; Piper and Calvert, 2009; Chen et al., 2016; Smrzka et al., 2020). Hence, the deposition of U takes priority over that of Mo, resulting in U_EF>Mo_EF. Mo and U are significantly enriched in our samples from both regions, which indicates the anoxic formation environments and an overlap or at least a close proximity of the iron reduction zone and the SMTZ in the active cold seep (Chen et al., 2016; Lin et al., 2021) (Fig. 5).

The REE composition is often used to determine the redox conditions and trace the fluid properties during carbonate formation (Haley et al., 2004; Chen et al., 2005; Himmler et al., 2010; Bayon et al., 2011; Rongemaille et al., 2011; Han et al., 2014). The REE composition of carbonates is usually strongly influenced by the surrounding seawater and pore water. The similar patterns of REE distribution standardized by PAAS among all the samples are interpreted as a consistent sedimentary fluid environment (Fig. 4). REE distribution patterns of the samples have typical features of that of anoxic pore water with an middle rare earth element (MREE) enrichment pattern, illustrating that the carbonates may have precipitated from the pore water. Previous studies have shown that the REE distribution patterns of pore water are related to depth and that the MREE enrichment pattern generally appears in the reduction zone of iron (Haley et al., 2004; Kim et al., 2012), revealing the anoxic precipitation environment of the carbonates.

The Y/Ho ratio is useful for distinguishing among different types of water and their respective sediments (Nothdurft et al., 2004). Jakubowicz et al. (2015) suggested that owing to the interaction with bedrocks and sediments, typical seepage fluids (without any influence of seawater) have a low Y/Ho value, which may be close to those of the upper continental crust (27.5) and chondrite (25–28) (Taylor and McLennan, 1985; Kamber et al., 2005). Seawater usually shows a high Y/Ho value of 60–150 (Nozaki et al., 1997; Lawrence and Kamber, 2006). The Y/Ho values of the Shenhu area and Haima samples are similar, ranging from 26.64 to 34.42. These values are slightly higher than the Y/Ho value of typical seepage fluid and significantly lower than that of seawater, suggesting that the samples were less affected by seawater.

Ce anomaly is usually used to clarify the sedimentary environments of cold-seep carbonates. A negative Ce anomaly after PAAS standardization usually indicates an oxidation environment, while an absent or positive Ce anomaly indicates a reduction environment. Previous studies have shown that late diagenesis may change the main Ce/Ce^* values of carbonates (Feng et al., 2009; Hu et al., 2014), leading to a positive correlation between Ce/Ce^* and REE, and a negative correlation between Ce/Ce^* and Dy_N/Sm_N (Shields and Stille, 2001). We can see from Fig. 6 that Ce/Ce^* has no correlation with REE or Dy_N/Sm_N, so the results of Ce/Ce^* reflect the original Ce anomaly and can indicate the original sedimentary environment of the samples. Traditionally, cold-seep carbonates are supposed to precipitate in a reductive environment, with no or positive Ce anomaly. However, in recent years, studies on cold-seep carbonates in Mexico, Black Sea, Congo, and other places have shown negative Ce anomalies, indicating aerobic environments (Feng et al., 2009; Ge et al., 2010; Tong et al., 2012). It is speculated that the cold seep may have had a temporary aerobic environment, which may be related to the change in fluid leakage rate (Feng et al., 2009, 2010; Birgel et al., 2011; Tong et al., 2012). The lg(Ce/Ce^*) values of the Shenhu samples range from -0.01 to 0.06, most suggesting no or positive Ce anomaly. Samples 11 and 14 show a slightly negative Ce anomaly. The formation environments of samples 9 and 15 may have been slightly more reductive than those of the other samples because they show true positive Ce anomaly. Sample ROV12 shows no Ce anomaly while Sample ROV2 shows an obviously negative Ce anomaly, indicating that the latter might have experienced a temporary oxidation environment or may have been mixed with seawater with negative Ce negative anomaly (Hu et al., 2014) (Fig. 7).

As mentioned above, all samples present more than one type carbonate mineral phases and different redox condition features (e.g. lg(Ce/Ce^*) and Fig. 7), possibly indicating multi-stage formation. Feng et al. (2018) concluded that several regional factors including temporal and spatial variation of cold-seep fluid flow would generate the heterogeneity in the facies of seep carbonates. Here, we have ventured to speculate that the aragonite-dominated carbonates are formed close to the seafloor and the lower seepage flux would induce the precipitation of dense carbonate with a higher proportion of calcite (samples 2, 7, and 8). Low carbonate content along with high debris content and main carbonate phases (calcite and dolomite) in samples 14 and 15 indicate a deeper formation depth in a low fluid seepage and a relatively complex formation mechanism possibly induced by high Mg content and strong microbial activity.

Although the information of the Haima samples is limited, we can infer from their geochemical characteristics that their precipitation is influenced by both seawater and seepage fluids and may have experienced a short-term oxidized environment.

5.3 δ¹³C and δ¹⁸O analysis for source of cold-seep fluid

Cold-seep carbonates generally inherit the carbon isotopic composition of the cold-seep fluid via AOM. Hence, this characteristic of the sample can be used to determine the source of carbon, and its variation reflects the mixing degree of different carbon sources (Suess, 2014; Feng et al., 2018). Several studies on the carbon oxygen isotopes of cold-seep carbonates have been conducted in Shenhu area and Haima cold seep (Feng et al., 2018). Our Shenhu samples have a relatively negative carbon isotopic composition with an average value of -41.59‰, consistent with the result of previous studies (Feng et al., 2018; Fig. 8), which is lighter than those of normal marine carbonates (δ¹³C ranging from -3‰ to 3‰ VPDB; Anderson and Arthur, 1983), indicating a degree of AOM function and mixed methane source between biogenic and thermogenic methane (Ge et al., 2011; Zhang et al., 2017). Additionally, δ¹³C of our samples varies widely, and most of crust carbonates except for samples 3, 11, and 14 presents more depleted ¹³C than that of the other two shape types. As discussed above, morphology is correlated firmly with the formation environment, and the deeper where carbonate formed may spark a stronger effect of thermogenic methane with higher δ¹³C. Other factors such as the permeable condition in the sediment, heterogeneity of methane properties and seawater will also influence δ¹³C of carbonates.

The δ¹³C values of the two Haima cold-seep carbonates, i.e., ROV2 and ROV12, are similar to those of previous studies (Fig. 8), identifying a mostly biogenic origin of methane (Liang et al., 2017; Guan et al., 2018). Considering that the only 10% carbonate content of ROV12 implies significant influence of background deposition, the addition of seawater may lead to the bias of ¹³C during its precipitation.

The oxygen isotopic compositions of the cold-seep carbonates are generally close to that of seawater and mainly affected by the formation temperature, the mineral composition, and δ¹⁸O of pore water (Aloisi et al., 2000). The oxygen isotopes of the samples from both Shenhu area and Haima cold seep are obviously heavier than that of seawater. The reason for ¹⁸O enrichment may be that ¹⁸O-rich fluids participate in the precipitation of cold-seep carbonates. The main sources of ¹⁸O-rich fluids are the dissociation of gas hydrates and dehydration of clay minerals (Feng and Chen, 2015; Liang et al., 2017; Feng et al., 2018). Gas hydrate dissociation has been detected in the Shenhu area (Suess, 2005; Zhang et al., 2007, 2019). Studies on the genesis of gas hydrates in the Shenhu area show that the source of gas hydrates is a mixture of thermogenic and microbial gases, and the value of δ¹³C of gas from the dissociation of hydrates is between -40‰ and -60‰ (Zhang et al., 2017), which does not conflict with our results of C isotope and has sufficient potential as a source of methane in the study area. Hence, water from hydrate dissociation may participate in the precipitation, increasing the δ¹⁸O value of the cold-seep carbonates (Feng et al., 2018).

A high ¹⁸O may also reflect a local low-temperature area, while a relatively low ¹⁸O may reflect a local high-heat flow (Feng and Chen, 2015; Liang et al., 2017; Yang et al., 2018). The dissociation of hydrates and the possible dehydration of clay minerals in this area affect the enrichment of ¹⁸O in carbonates by oreforming fluids, thus leading to the illusion of lower temperature. The age of the Shenhu samples in this study is unknown and hence, the specific temperature of the formation of the cold-seep carbonates cannot be determined. Therefore, the contribution of formation temperature and methane-rich fluid to the ¹⁸O values of the cold-seep carbonates in this area requires further study.

The samples from the Haima cold-seep area were collected at a temperature of approximately 3 ℃ from the bottom of the sea, and the ¹⁴C dating results are in the range of 5.1–6.1 ka and 2.9–3.9 ka (Liang et al., 2017). Many studies have shown that the sea level of the SCS had remained relatively stable over the past 6 000 years, with no significant change in the sea surface temperature (Pelejero et al., 1999; Zhao et al., 2006; Steinke et al., 2011). Therefore, we use the temperature of underlying water as the reference value of diagenetic temperature. It was assumed that the samples were formed from the current bottom water at a temperature of 3.0 ℃ and δ¹⁸O value of 0‰ VSMOW. The oxygen isotopic composition was calculated using Eq.3 when aragonite is in equilibrium with seawater. The oxygen isotopic composition at equilibrium between calcite and seawater was calculated using Eq.4. The conversion Eq.5 between δ¹⁸O VPDB and δ¹⁸OVSMOW was carried out using the equation proposed by Coplen et al. (1983). The definition of the oxygen isotopic fractionation factor, α, is given by Eq.6.

Here, T is the absolute temperature of seawater (K), δ¹⁸O_{aragonite/calcite} in Eq.6 and δ ¹⁸O_water used here are given in the δ-notation (‰) relative to the VSMOW standard.

The oxygen isotopic compositions of aragonite and calcite at equilibrium with seawater were 3.2‰ VPDB and 2.5‰ VPDB, respectively, which were lower than the measured δ¹⁸O values of the Haima samples, which further confirms that ¹⁸O-rich fluid was involved in the precipitation. The source of biogenic methane indicates that they originated from the shallow strata, and the temperature and pressure conditions in the source area could not facilitate the transformation from montmorillonite to illite. In addition, gas hydrate has been recovered by the gravity and piston core from the cold seep (Liang et al., 2017). Therefore, we speculated that the source of ¹⁸O-rich fluid here was mainly the dissociation of hydrates in the shallow strata.

6 CONCLUSION

Morphology and mineralogy of the studied coldseep carbonates from Shenhu area reflect the different environment of formation and fluid migration. Enrichment of U and Mo and REE features (including MREE enrichment, absent or slightly positive Ce anomaly) in the samples indicate the anoxic precipitation environment. And some samples may experience a periodical oxidation or influence by seawater, showing the negative Ce anomalies. Compared with the samples from the two areas, we conclude that the formation depths of the Haima samples were shallower and represent a stronger coldseep activity.

Cold-seep carbonates from both Shenhu area and Haima cold seep are methane-derived, while the former mainly originate from a mixture of thermogenic and microbial gases, and the latter from biogenic gas. The ¹⁸O-rich fluid was mainly from the dissociation water of gas hydrates.

Our study makes the comparison between Shenhu area and Haima cold seep to get better understanding of the cold seep system in SCS and is conductive to the exploration of gas hydrate. However, lack of analyses such as microphase observation by scanning electron microscope due to loose samples, and limited data restrict our further analysis. Hence, more samples and comprehensive studies will be needed in the future.

7 DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding authors upon reasonable request.

8 ACKNOWLEDGMENT

We are grateful to the two anonymous reviewers for the valuable suggestions.