Dynamics and diagenesis of trace metals in sediments of the Changjiang Estuary

Science of the Total Environment 675 (2019) 247–259

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Dynamics and diagenesis of trace metals in sediments of the Changjiang Estuary Liqin Duan a,b,c,d,⁎, Jinming Song a,b,c,d,⁎, Xianmeng Liang a,c, Meiling Yin a,c, Huamao Yuan a,b,c,d, Xuegang Li a,b,c,d, Chengzhe Ren a,c, Bu Zhou a, Xuming Kang e, Xuebo Yin b,c,d,f a

CAS Key Laboratory of Marine Ecology and Environmental Sciences, Institute of Oceanology, Chinese Academy of Sciences, Qingdao 266071, PR China Laboratory for Marine Ecology and Environmental Sciences, Qingdao National Laboratory for Marine Science and Technology, Qingdao 266237, PR China University of Chinese Academy of Sciences, Beijing 100049, PR China d Center for Ocean Mega-Science, Chinese Academy of Sciences, Qingdao 266071, PR China e Key Laboratory of Testing and Evaluation for Aquatic Product Safety and Quality, Ministry of Agriculture, Yellow Sea Fisheries Research Institute, Chinese Academy of Fishery Sciences, Qingdao 266071, PR China f CAS Key Laboratory of Marine Geology and Environment, Institute of Oceanology, Chinese Academy of Sciences, Qingdao 266071, PR China b c

H I G H L I G H T S

G R A P H I C A L

A B S T R A C T

• Seasonal dynamics of trace metals at SWI of the Changjiang Estuary are absent. • Mechanism that drove the diffusive release of trace metals at SWI was studied. • Seasonal cycling of trace metals was more obvious in the seasonal hypoxic zone. • Solid profiles of trace metals supported their dynamic variations in porewater. • Benthic diffusion was an important contributor to trace metals in water column.

a r t i c l e

i n f o

Article history: Received 31 December 2018 Received in revised form 10 April 2019 Accepted 12 April 2019 Available online 16 April 2019 Editor: Filip M.G. Tack Keywords: Trace metals Porewater Early diagenesis Seasonal cycling

a b s t r a c t The seasonal dynamics and diagenesis of trace metals at two contrasting coastal sites were studied to determine the mechanism that drove the diffusive release of trace metals from sediments in the Changjiang Estuary. Porewater trace metal concentrations were 53.4–4829 nM for Zn, 11.0–344 nM for Cu, 7.75–221 nM for Cr, 2.71–61.1 nM for Co, 0.822–42.7 nM for Pb and 0.037–4.22 nM for Cd. The concentrations and profiles of trace metals in the porewater and solid phase displayed obvious regional and seasonal variations. This variation was mainly reflected in the surface layer and the depth of the suboxic and anoxic layers. Regionally, surface porewater trace metal concentrations in the seasonal hypoxic region were higher than those in the aerobic region due to changes in the redox conditions being beneficial to the release of trace metals. Seasonally, surface porewater trace metal concentrations decreased in summer compared to spring due to their removal by forming metal sulfides in summer. Solid profiles of the trace metals supported their dynamic variations in the porewater. The partition coefficient suggested that the formation of Fe/Mn (hydr)oxides was effective for the removal of trace metal in oxidizing condition, while the formation of sulfides was conducive to the removal of trace metals in reducing

⁎ Corresponding authors at: CAS Key Laboratory of Marine Ecology and Environmental Sciences, Institute of Oceanology, Chinese Academy of Sciences, Qingdao 266071, PR China. E-mail addresses: [email protected] (L. Duan), [email protected] (J. Song).

https://doi.org/10.1016/j.scitotenv.2019.04.190 0048-9697/© 2019 Elsevier B.V. All rights reserved.

248

L. Duan et al. / Science of the Total Environment 675 (2019) 247–259

Benthic fluxes The Changjiang Estuary

condition. The combination of porewater with solid phase data suggested that the dynamics of Cu, Zn, Cr and Co were mainly controlled by Fe and Mn diagenesis, the dynamics of Cd were affected by S cycling, and the dynamics of Pb were disturbed by anthropogenic inputs and benthic organism activities. Estimation of benthic fluxes indicated that sediments were an important source of trace metals in the water column. The contributions of trace metals by sediments to the water column of the Changjiang Estuary were only one order of magnitude lower than those by riverine fluxes. © 2019 Elsevier B.V. All rights reserved.

1. Introduction Sediments of estuaries and coastal waters have been considered an important sink for trace metals discharged by riverine input, biological activities and human activities. However, when water and sediment conditions (e.g., redox and pH) change and/or are influenced by hydrodynamics and biological disturbance, trace metals may be remobilized and even diffused back to the water column, likely resulting in secondary pollution (Ho et al., 2012; Vizzini et al., 2013; Zhang et al., 2014). It has been reported that the benthic diffusive fluxes of trace metals from sediments to the water column were equivalent to or even exceeded riverine influxes in many coastal areas (Rivera-Duarte and Flegal, 1997; Santos-Echeandia et al., 2009; Warnken et al., 2001; Williams et al., 1998). Therefore, the effect of trace metal migration and diffusion on marine water ecosystems cannot be ignored. Although sediments have an important impact on seawater quality, the mechanisms for determining whether trace metals are permanently buried in sediments or remobilized to the water column have not been well established. Diffusion of trace metals from marine sediments to the water column is accomplished through the sediment-seawater interface (SWI). Particularly, as the link between sediments and the overlying waters, porewater diffusion is a more important and direct dynamic release of trace metals from sediments into the overlying waters (AleksanderKwaterczak and Zdechlik, 2016). Various processes, as well as redox, pH, and sulfide, influence the distribution of trace metals between solid and liquid phases and control whether trace metals are sequestered in sediments or remobilized to the overlying waters (SantosEcheandia et al., 2009; Wang and Wang, 2017). Thus, the variations in trace metal concentrations in porewater can be used as sensitive indicators of their transportation and transformation in sediments. The mobility, transformation and release processes, such as sorption, precipitation, dissolution, advection, diffusion, bioirrigation and bioturbation, mainly depend on early diagenesis and are influenced by the physicochemical conditions in both sediments and the overlying waters (Kalnejais et al., 2015; Santos-Echeandia et al., 2009). Microbial mineralization of organic matter below the SWI as a driving force of early diagenesis is known to strongly impact the transport mechanisms of trace metals, especially when Fe and Mn (hydr)oxides and sulfates are the terminal electron acceptors (Kalnejais et al., 2015). Generally, when organic matter is mineralized, reductive dissolution of Fe and Mn (hydr)oxides occurs and is accompanied by remobilization of associated trace metals, which supplies trace metals to porewater (Charriau et al., 2011; Dang et al., 2015; Rigaud et al., 2013). Dissolved trace metals move upwards and downwards and can further be readsorbed onto or coprecipitated with newly formed Fe and Mn mineral phases (e.g., (hydr)oxides, carbonates and sulfides) or form insoluble metal sulfide phases when sulfide is produced by sulfate reduction bacteria (Morse and Luther, 1999; Sundby, 2006). Conversely, the oxidation of sulfides can release dissolved trace metals into porewater (Wang and Wang, 2017). Consequently, the dynamics of trace metals in the SWI are controlled by oxidative degradation of organic matter and dissolution and formation of Fe and Mn minerals and insoluble metal sulfide phases. The relatively importance of these reactions at the SWI is crucial to understanding trace metal cycling.

Sediments are subject to strong seasonal and spatial variations in organic and terrigenous matter fluxes, temperatures, biological activities and redox conditions in estuaries and coastal areas; thus, there are considerable regional and seasonal variations in early diagenetic processes, redox zones and redox boundary layers near the SWI (Kalnejais et al., 2015; Santos-Echeandia et al., 2009; Wang and Wang, 2017). These processes can lead to obvious spatial and seasonal differences in the migration and exchange mechanisms of trace elements at the SWI. Thus, prediction of trace metal cycling is particularly challenging due to changes in spatial and seasonal conditions. Although material exchange at the SWI is rapid in coastal waters, the exchange fluxes of trace metals at the SWI under different environmental conditions are uncertain (Oehler et al., 2015; Torres et al., 2014). Therefore, understanding the effects of various environmental conditions on the exchange processes of trace metals at the SWI is important to quantify the trace metal budget. In recent years, due to progressive human activities, seasonal hypoxic zones in estuaries and coastal waters have been formed and are constantly expanding. The occurrence of hypoxia in bottom water inevitably leads to active trace metal exchange at the SWI, further aggravating the influence on the water column. There are typical seasonal hypoxic zones in the Changjiang Estuary, which has been considered one of the largest coastal hypoxic zones in the world (Chen et al., 2007; Zhao et al., 2015). However, the variability in processes that impact trace metal mobilization, seasonal changes in porewater concentrations and benthic fluxes have not been studied in the Changjiang Estuary. The main objectives of the study were to (1) determine the diagenetic reactions of Fe and Mn in the Changjiang Estuary; (2) discuss the porewater dynamics of trace metals and their regional and seasonal variations in the Changjiang Estuary; (3) evaluate the partitioning of trace metals between porewater and solid phases in the Changjiang Estuary; and (4) quantify the diffusive fluxes of trace metals in different seasons in the Changjiang Estuary. 2. Materials and methods 2.1. Study area The Changjiang Estuary is one of the world's most productive continental shelf areas. The Changjiang River, as the world's fourth largest river based on suspended load, has average freshwater influx of ~9.25 × 1011 m3 and sediment influx of ∼5 × 108 t annually (Duan et al., 2017). The Changjiang Estuary has been regarded as one of the largest coastal hypoxic areas in the world. The hypoxia occurrence in the Changjiang Estuary varies regionally and seasonally. The season for hypoxia events usually begins in early June and may persist as late as October (Zhou et al., 2017). In June, hypoxia appears in waters along the coast to the south of the Changjiang Estuary. In July, the hypoxic zone appears in the northwestern Changjiang Estuary. In August, two hypoxic zones develop in the northern and southern Changjiang Estuary, with the northern zone being more severe. The hypoxic zone in the southern Changjiang Estuary persists in September and October. This seasonal hypoxic variation is influenced by the pycnocline between the Changjiang River freshwater discharge and the upwelling water from the Taiwan Warm Current (TWC) and the abundant bottom detritus from surface phytoplankton blooms (Zhang et al., 2007;

L. Duan et al. / Science of the Total Environment 675 (2019) 247–259

Zhu et al., 2016). Based on the unique seasonal hypoxia in bottom waters of the Changjiang Estuary, it is important to determine the mechanisms that drive the benthic metal cycling and fluxes. 2.2. Sample collection Five cores (~30 cm) of undisturbed sediments and approximately 100 mL of the overlying waters were collected with a Soutar-type box corer at four sites in May and September 2017 in the Changjiang Estuary onboard the R/V “Kexue 3” cruise. The station locations were chosen to be representative of the different environments. Sites C1 and C2 are located in the seasonal hypoxic region, where the bottom water is hypoxic in summer and oxic in the other seasons. Sites C3 and C4 are located in the aerobic region, where the bottom water is oxic during the whole year. Among these sites, three cores (C1, C2 and C3) were collected in May and another 2 cores (C2 and C4) were collected in September. The sampling sites are shown in Fig. 1. The overlying water samples were collected from 2 cm above the SWI by syringe. The samples were immediately filtered through 0.45 μm acid-washed pore size acetate cellulose filters and then placed into acid-washed plastic bottles and preserved by adding trace metal grade nitric acid to pH b 2. The sediment cores were sampled to extract the porewater. Two acid-washed PVC tubes (10 cm diameter and 30 cm length) were placed inside the box corer to obtain undisturbed sediment cores at each site. One PVC tube was employed to collect a sediment core for porosity determination. Another PVC tube with predrilled holes was used to collect a sediment core for porewater and solid samples. For the PVC tubes, holes of 2.5 mm diameter were drilled and staggered at a 0.5 cm interval between 0 and 5 cm, 1 cm interval between 5 and 10 cm and 2 cm interval between 10 cm and the bottom. The holes were sealed with tape before coring. After obtaining the sediment cores, the PVC tubes were immediately sealed at the top and bottom to maintain anoxic conditions and placed into a glove box filled with N2. Porewater was collected using Rhizon core solution samplers, which are porous polymer tubes with glass fiber strengthener and a membrane pore size of 0.12–0.18 μm. Briefly, Rhizon samplers (Rhizosphere Research Products, Wageningen, Netherlands) of 10 cm length and 2.5 mm diameter were

249

inserted horizontally through holes drilled in the core at different depths in the profiles at each site to ensure that the porous membrane was centered in the core. Porewater was obtained by connecting the Rhizon sampler to a plastic syringe with a polyethylene plunger and creating a vacuum. The sample volume ranged from 5 to 10 mL, and the sampling duration was 20–60 min. The extracted porewater was divided into three aliquots. One extracted subsample was acidified with trace metal grade nitric acid and stored at −20 °C until laboratory analysis for trace metals. One extracted subsample was transferred into a vial containing a solution of zinc acetate in 2 M sodium hydroxide for sulfide analysis. One extracted subsample was frozen until analysis for sulfate. After complete porewater removal, the solid cores were sliced into 1 cm thick slices for the upper 10 cm and 2 cm thick slices below 10 cm inside a N2-filled glove box. These sliced sediments were transferred to acid-washed plastic bags previously purged with N2 and then frozen for solid analysis. The sediment cores for porosity analysis were also sectioned via the same method. All sampling equipment in contact with the samples for trace metal analysis was acid washed and rinsed in 18.2 MΩ Milli-Q water. 2.3. Chemical analysis 2.3.1. Parameter analysis The dissolved oxygen (DO) and pH values of the overlying waters were immediately determined after sampling. The pH was determined with a pH meter. The DO concentration was determined using a SevenExcellence Multiparameter (Mettler Toledo, US) equipped with a DO probe. The DO values measured by probe were calibrated by the Winkler titration method (Grasshoff et al., 1999). Sulfate concentrations in the overlying waters and porewaters were determined using a Thermo Dionex ICS-5000+ ion chromatograph. Dissolved sulfide in the overlying waters and porewaters collected as ZnS onboard were measured colorimetrically by the Cline method (Cline, 1969) using a UV-VIS spectrophotometer. Porosity was determined gravimetrically by drying sediments to a constant weight in a 60 °C oven and assuming a grain density of 2.6 g/mL (Berner, 1980). Acid volatile sulfide (AVS) in the solid

Fig. 1. Locations of the sampling stations. The cores at sites C1 and C3 were sampled in May 2017, the core at site C4 was sampled in September 2017, and the cores at site C2 were sampled in both May and September 2017.

250

L. Duan et al. / Science of the Total Environment 675 (2019) 247–259

sediments was determined using the cool diffusion method (Burton et al., 2008; Zhu et al., 2012). Briefly, 6 M HCl together with 0.1 M ascorbic acid was loaded into reactors containing preweighed wet sediments (3–5 g) and then immediately sealed. Evolving H2S gas was absorbed and trapped as ZnS precipitant in a separate vial containing alkaline zinc acetate solution. The trapped sulfide was measured colorimetrically by the Cline method (Cline, 1969) using a UV-VIS spectrophotometer.

2.5. Statistical analysis

2.3.2. Fe, Mn and trace metal determination Concentrations of Fe, Mn and trace metals in the overlying waters were determined using modified methods based on Sawatari et al. (1995). The concentrations were determined by inductively coupled plasma mass spectrometry (ICP-MS) after preconcentration using a trace-metal-grade Ga standard solution and NH4OH at pH 9.0 and then dissolution in 5% HNO3. The same method was also used to measure trace metals in the porewaters. The precision of the method was better than 10% for triplicate preconcentration. Fe and Mn in the porewaters were determined by inductively coupled plasma optical emission spectrometry (ICP-OES) after the porewater samples were diluted 10–15 times with 5% HNO3. The precision of the method was better than 10% for triplicate dilution. Both the preconcentration and dilution were performed to avoid the impact of “salt” concentrations. Certified reference materials NASS-6 and GBW80040 were used to assess the accuracy of the analytical procedure for the overlying water and porewater, with recoveries of 85–115%. Trace metal contents in the solid samples were determined after digestion. Briefly, freeze-dried and ground solid samples were digested with HF-HNO3-H2O2 using microwave high-pressure digestion (MARS 6 CLASSIC, USA). An aliquot of digestion was determined by ICP-MS for trace metals (Cu, Zn, Pb, Cr, Cd and Co). Another aliquot of digestion was determined by ICP-OES for Fe and Mn. The precision of the analysis was b10% for all metals in triplicate. Certified reference materials GSS15 and GSD-9 were analyzed for evaluating the accuracy of the sediment analytical procedure. The determined values for all metals were within the certified values. Procedural blanks for the overlying water, porewater and sediment and quality control sample were processed and examined after every 10 samples. The results were blank corrected.

3.1. Porewater Fe, Mn and sulfide

2.4. Benthic flux calculation The diffusive fluxes of trace metals from sediment to the overlying water were calculated by the concentration gradients using Fick's first law of diffusion (Berner, 1984): F ¼ Ø  Ds  ð∂C=∂zÞ where F is the diffusive flux of metals (μg/m2/d), ø is the porosity of sediment (dimensionless), Ds is the molecular diffusion coefficient (cm2/s), C is the porewater concentration (μg/L), Z is the depth below the SWI (cm), and ∂C/∂z is the concentration gradient found at the SWI (μg/L/cm). Ds is calculated from the following empirical equations (Ullman and Sandstrom, 1987): Ds ¼ D0 ðØ≤0:7Þ Ds ¼ Ø2 D0 ðØN0:7Þ where D0 is the diffusion coefficient in water. The values of D0 for each metal were taken from Li and Gregory (1974). The concentration gradient was estimated from porewater profiles, with the assumption that the gradient across the SWI was one-dimensional and could be approximated as the linear fitting slope of the metal concentration at the SWI.

Pearson correlation analysis performed by SPSS 16.0 was used to determine the relationships among the trace metals and Fe and Mn in the porewater and solid phases. A value of p b 0.05 (2-tailed) was considered a significant correlation. Origin Pro 9.0 software was used to plot the data. 3. Results

Because the cycling and benthic fluxes of trace metals are strongly influenced by the redox conditions and diagenesis in sediments, porewater Fe, Mn, sulfide and sulfate were determined. The porewater profiles of Fe, Mn, sulfide and sulfate are shown in Fig. 2. Generally, the porewater profiles of Fe and Mn displayed an increasing trend downwards until peak appearance and then decreased and kept stable low values at the bottom layer. However, their profile variations and peak depths were different among the regions and seasons. In the seasonal hypoxic region, dissolved Fe and Mn concentrations in the porewaters were 0.893–33.3 μM for Fe and 0.903–70.1 μM for Mn in spring and 35.6–114 μM for Fe and 24.2–50.6 μM for Mn in summer. Dissolved Mn displayed lower values within the upper 3 cm depth in spring, with values of 3.59 μM for core C1 and 43.1 μM for core C2. Dissolved Mn also had lower values within the upper 1.5 cm depth in summer, with a value of 39.0 μM for core C2. Below the upper layer, dissolved Mn was marked by peaks in both seasons. Specifically, dissolved Mn showed higher values at 3–10 cm depth, with the maximum (70.1 μM) at 5 cm depth in spring, while it displayed higher values at 2–4 cm depth, with the maximum (50.6 μM) at 2 cm depth in summer. Below the middle layer, dissolved Mn decreased and kept stable low values. Similarly, dissolved Fe showed lower values in the surface or subsurface porewater, with values of 3.91 μM in spring and 66.4 μM in summer. The higher values of dissolved Fe occurred at 12–23 cm depth, with the maximum (33.3 μM) at 17 cm depth in spring, while it displayed higher values at 2–4 cm depth, with relatively high values at 3–4 cm depth in summer. Below the middle layer, dissolved Fe decreased and kept steady. Dissolved sulfide generally displayed an increasing trend downwards and typically reached the higher concentrations between 15 and 25 cm depth in spring, whereas it presented a maximum at 5 cm depth in summer. Sulfate displayed a decreasing trend with depth. The seasonal variation in dissolved Fe, Mn and sulfide profiles was evident in the seasonal hypoxic region, and their peaks moved upwards from spring to summer. In the aerobic region, dissolved Fe and Mn concentrations were 0.893–50.4 μM for Fe and 9.14–29.3 μM for Mn in spring and 10.5–107 μM for Fe and 15.0–58.0 μM for Mn in summer. Dissolved Mn had higher values at 5–12 cm depth, with the maximum (29.3 μM) at 5 cm depth in spring, while it had higher values at 2–10 cm depth, with the maximum (58.0 μM) at 6 cm depth in summer. Dissolved Fe generally displayed a sharply decreasing trend downwards in both seasons, with higher values at 1–4 cm depth in spring and 3.5–7 cm depth in summer. Fe removed from the porewater was shallower than Mn in both seasons, but both of them were never completely removed from the porewater (Fig. 2). Dissolved sulfide displayed an increasing trend downwards and reached higher concentrations at 14–20 cm depth in spring, whereas it presented a stable trend except for a maximum (2.71 μM) at 0.5 cm depth in summer. Sulfate displayed a decreasing trend with depth in spring, whereas it presented a stable trend except for a value at the surface in summer. 3.2. Trace metals in the overlying water and porewater Trace metal concentrations in the overlying waters were in the ranges of 4.77–22.1 nM for Cu, 46.1–91.0 nM for Zn, 0.170–1.76 nM for Pb, 8.56–12.3 nM for Cr, 0.200–0.315 nM for Cd and 0.401–3.94 nM for Co

L. Duan et al. / Science of the Total Environment 675 (2019) 247–259

251

Fe (μM) Mn(μM) Sulfide (μM) Sulfate (mM) Sulfide (μM) Sulfate (mM) Mn(μM) Spring 0 30 60 90 1200 3 620 25 30 Spring 0 40 80 0 25 30 40 800 3 620 5 5 C1 C2 0 0 Depth (cm)

Depth (cm)

-5 -10 -15

-5 -10 -15

-20

-20

-25

-25

-30

-30

Depth (cm)

Fe (μM) Mn(μM) Sulfide (μM) Sulfate (mM) Spring 0 30 60 90 1200 40 800 3 620 25 30 C3 5 0 -5 -10 -15 -20 -25 -30

0

0

-5

-5

Depth (cm)

Depth (cm)

Fe (μM) Mn(μM) Sulfide (μM) Sulfate (mM) Fe (μM) Mn(μM) Sulfide (μM) Sulfate (mM) Summer 0 30 60 90 1200 3 620 25 30 Summer 0 40 800 25 30 60 1200 40 800 3 6 20 C2 5 C4 5

-10 -15

-10 -15

-20

-20

-25

-25

-30

-30 Fig. 2. Porewater profiles of Fe, Mn, sulfide and sulfate in cores C1–C4 of the Changjiang Estuary in spring and summer.

in spring and 17.9–24.5 nM for Cu, 55.4–225 nM for Zn, 3.27–6.98 nM for Pb, 27.4–28.7 nM for Cr, 0.632–0.934 nM for Cd and 3.19–4.26 nM for Co in summer. Concentrations of trace metals in the overlying waters showed a significant increase from spring to summer. Dissolved trace metal concentrations in the porewaters were 1–2 magnitudes higher than those in the overlying waters. In the seasonal hypoxic region, porewater trace metal concentrations were 117–1499 nM for Zn, 22.1–263 nM for Cu, 11.0–221 nM for Cr, 3.90–61.1 nM for Co, 1.32–42.7 nM for Pb and 0.187–4.22 nM for Cd in spring and 512–4829 nM for Zn, 29.1–344 nM for Cu, 27.4–63.4 nM for Cr, 4.26–19.0 nM for Co, 1.99–10.3 nM for Pb and 0.374–3.10 nM for Cd in summer. The average porewater concentrations in spring were higher than those in summer for Cu, Cr, Co and Pb, with the opposite result for Zn and Cd. In the aerobic region, porewater trace metal concentrations were 53.4–236 nM for Zn, 11.0–103 nM for Cu, 7.75–69.4 nM for Cr, 2.71–27.9 nM for Co, 0.822–21.6 nM for Pb and 0.037–0.788 nM for Cd in spring and 200–1342 nM for Zn, 28.6–103 nM for Cu, 30.8–77.2 nM for Cr, 4.38–19.4 nM for Co, 2.12–10.3 nM for Pb and 0.587–3.02 nM for Cd in summer. The average porewater concentrations in spring were higher than those in summer for Cu and Co, with the opposite result for Zn, Cr, Pb and Cd. In addition, the average porewater concentrations for all trace metals in the seasonal hypoxic region were generally higher than those in the aerobic region in spring. However, the concentration variation for porewater trace metals in summer was complex. The average concentrations for Zn and Cu were higher in the seasonal hypoxic region, whereas those for the other metals were higher in the aerobic region.

The porewater concentrations of trace metals are plotted as a function of depth in Fig. 3. According to their redox behaviors, trace metals were grouped in three categories. First, porewater Cu, Zn and Cr had similar profiles. In spring, all three trace metals were characterized by elevated concentrations below the interface in both regions. These trace metals decreased downwards between 0 and 5 cm depth and then maintained a stable trend. In summer, the Cu, Zn and Cr profiles were complicated. Specifically, these trace metals had peaks at 2–3 cm depth in the seasonal hypoxic region, whereas their concentrations decreased downwards between 0 and 5 cm depth in the aerobic region. The peak locations of Cu and Zn were shallower in spring than in summer. Second, porewater Cd and Pb presented sporadic peaks at various depths. In spring, peaks of Cd and Pb appeared at surface or subsurface layers, and then they displayed a general decrease downwards until 10 cm or 15 cm depth. Below 10 cm or 15 cm depth, peaks also appeared at depths of ~25 cm and ~15 cm. In summer, the fluctuation was more intense but without obvious variation trend except for peaks at surface and middle layers (depths of ~15 cm and ~5 cm). Third, porewater Co presented vertical and seasonal variations similar to those of Mn. The Co peaks mainly appeared in the middle layer (2–12 cm depth). 3.3. Solid data 3.3.1. Fe, Mn and AVS Fe, Mn and AVS in solid phase displayed an obvious regional variation (Fig. 4). Solid Fe and Mn contents in the seasonal hypoxic

252

L. Duan et al. / Science of the Total Environment 675 (2019) 247–259

Depth (cm)

Spring Cu (nM) Cr (nM) Zn (nM) Cd (nM) C1 5 0 175 350 0 2500 50000 120 240 0.0 2.3 4.6

Pb (nM) 25

50

0

50

0

Co (nM) 31

62

0 -5 -10 -15 -20 -25 -30

Spring Cu (nM) Cr (nM) Zn (nM) Cd (nM) C2 5 0 175 350 0 2500 50000 120 240 0.0 2.3 4.6

Depth (cm)

0

0

Pb (nM) 25

Co (nM) 31

62

0 -5 -10 -15 -20 -25 -30

Depth (cm)

Spring Cu (nM) Cr (nM) Co (nM) Zn (nM) Pb (nM) Cd (nM) C3 5 0 175 350 0 2500 50000 120 240 0.0 2.3 4.6 0 25 50 0 31 62 0 -5 -10 -15 -20 -25 -30

Depth (cm)

Cr (nM) Summer Cu (nM) Zn (nM) Cd (nM) C2 5 0 175 350 0 2500 50000 120 240 0.0 2.3 4.6

Pb (nM) 25

50

0

Co (nM) 31

62

0 -5 -10 -15 -20 -25 -30

Cr (nM) Summer Cu (nM) Zn (nM) Cd (nM) C4 5 0 175 350 0 2500 50000 120 240 0.0 2.3 4.6

Depth (cm)

0

0

Pb (nM) 25

50

0

Co (nM) 31

62

0 -5 -10 -15 -20 -25 -30 Fig. 3. Porewater profiles of trace metals in cores C1–C4 of the Changjiang Estuary in spring and summer.

region were equal to those in the aerobic region in spring; however, their contents in the seasonal hypoxic region were lower than those in the aerobic region in summer. Contrary to solid Fe and Mn, AVS content in the seasonal hypoxic region was higher than that in the aerobic region. Fe, Mn and AVS profiles in solid phase displayed an obvious seasonal variation. In spring, Fe and Mn in solid phase showed profiles

contrary to those of porewater. Generally, solid Fe and Mn showed higher values at the surface. Below the surface layer, they decreased downwards until the minimum at depths of approximately 7 cm and 9 cm and then kept low values. However, the profile of AVS was similar to that of dissolved sulfide. The AVS profile generally presented very low values at the surface layer and then increased with depth, especially showing an obvious increase beginning at ~5 cm depth.

L. Duan et al. / Science of the Total Environment 675 (2019) 247–259

Fe (%) 3 4

5 0.03

Mn (%) 0.06 0.09 0

AVS (μg/g) 180 360 Spring 2 0 C2

-5

-5

-10

-10

Depth (cm)

Depth (cm)

Spring 2 0 C1

-15 -20

Mn (%) AVS (μg/g) 5 0.03 0.06 0.09 0 180 360

Fe (%) 3 4

AVS (μg/g) Mn (%) 5 0.03 0.06 0.09 0 180 360

-15 -20

-25

-25

-30

-30

Spring 2 0 C3

Fe (%) 3 4

253

Fe (%) 3 4

Mn (%) AVS (μg/g) 5 0.03 0.06 0.09 0 180 360

Fe (%) 3 4

AVS (μg/g) Mn (%) 5 0.03 0.06 0.09 0 180 360 Summer 2 0 C4 -5

Depth (cm)

-5 -10 -15 -20 -25 -30

Summer 2 C2 0

-10

Depth (cm)

Depth (cm)

-5

-15 -20

-10 -15 -20

-25

-25

-30

-30 Fig. 4. Solid profiles of Fe, Mn and acid volatile sulfide (AVS) in cores C1–C4 of the Changjiang Estuary in spring and summer.

In summer, the solid Fe and Mn profiles in the seasonal hypoxic region were different from aerobic region. In the seasonal hypoxic region, solid Fe and Mn displayed minimum values at the upper 0–2 cm. Below 2 cm depth, solid Fe and Mn presented variations similar to those of porewater Fe and Mn. The solid Fe and Mn values increased downwards until 5 cm depth, with maximum values of 4.19% for Fe and 0.077% for Mn, and then decreased until 12 cm depth and subsequently kept low values. The profile of AVS was similar to that of dissolved sulfide. The AVS profile generally presented very low values (~0) at the surface 0–4 cm depth and then increased downwards until reaching a maximum at 5 cm depth; then, AVS decreased until 8 cm depth and subsequently kept low values. In the aerobic region, the variation in the solid Fe and Mn profiles was small compared to that in the porewater Fe and Mn profiles. Similar to porewater Fe and Mn, solid Fe and Mn showed minimum values at the surface layer. Solid Fe and Mn presented higher values at the subsurface layer, with 4.90% for Fe and 0.085% for Mn. Below the subsurface layer, solid Fe kept a constant value; however, solid Mn generally displayed an increase downwards. The AVS profile generally displayed an increase downwards from 8.67 μg/g at the surface to 162 μg/g at the bottom. The AVS content in summer was higher than that in spring, with the highest AVS value occurring in the seasonal hypoxic region in summer.

3.3.2. Trace metals Solid trace metals also presented regional and seasonal variations (Fig. 5). Solid trace metal contents in the surface layer were higher in spring than in summer and were also higher in the aerobic region than in the seasonal hypoxic region. Surface enrichment was only found in the aerobic region in spring. The variations in the solid profiles of most trace metals were generally similar to those in the solid Fe and Mn profiles, with some differences among sites. In core C1 of the seasonal hypoxic region, trace metals Cu, Zn and Cr showed a decreasing trend downwards, whereas Cd and Pb presented the opposite trend in spring. Co displayed the same profile as Mn, except for an abnormal high value at 4 cm depth. In core C2, the solid profiles of trace metals presented obvious seasonal variations. In spring, all trace metals displayed coincident profile distributions, showing slight increases downwards until reaching a maximum at 7 cm depth and then decreased and kept stable. In summer, the trace metals, except for Cd, presented profiles similar to those of Fe and Mn. These trace metals showed an increase from the surface to 3 cm depth and kept higher values between 3 and 6 cm depth with peaks at different depths; then, the values decreased below 6 cm and kept stable values, which were still higher than the values at the upper 3 cm depth. Although Cd has a profile similar to those of the other trace metals, Cd had an abnormally high value at 12 cm depth.

254

L. Duan et al. / Science of the Total Environment 675 (2019) 247–259

Cr (μg/g) Spring Cu (μg/g) Zn (μg/g) C1 0 15 30 45 60 120 180 90 135

Pb (μg/g)

Cd (μg/g) 0

1

2

20

40

Co (μg/g) 20

40

Depth (cm)

-5 -10 -15 -20 -25 -30

Cr (μg/g) Spring Cu (μg/g) Zn (μg/g) C2 0 15 30 45 60 120 180 90 135

Pb (μg/g)

Cd (μg/g) 0

1

2

20

40

Co (μg/g) 20

40

Depth (cm)

-5 -10 -15 -20 -25 -30

Cr (μg/g) Spring Cu (μg/g) Zn (μg/g) C3 0 15 30 45 60 120 180 90 135

Pb (μg/g)

Cd (μg/g) 0

1

2

20

40

Co (μg/g) 20

40

Depth (cm)

-5 -10 -15 -20 -25 -30

Summer Cu (μg/g) Cr (μg/g) Zn (μg/g) C2 0 15 30 45 60 120 180 90 135

Pb (μg/g)

Cd (μg/g) 0

1

2

20

40

Co (μg/g) 20

40

Depth (cm)

-5 -10 -15 -20 -25 -30

Summer Cu (μg/g) Cr (μg/g) Zn (μg/g) C4 0 15 30 45 60 120 180 90 135

Cd (μg/g) 0

1

2

Pb (μg/g) 20

40

Co (μg/g) 20

40

Depth (cm)

-5 -10 -15 -20 -25 -30 Fig. 5. Solid profiles of trace metals in cores C1–C4 of the Changjiang Estuary in spring and summer.

In the aerobic region, the solid profiles of trace metals were less fluctuant, indicating a stable sedimentary environment. Generally, the trace metals, except for Cd and Pb, presented profiles similar to those of Fe and Mn in both spring and summer. However, Zn and Cu had abnormally high values at 10 cm depth in spring, consist with the maximum of AVS.

The profile of solid Cd was the same as that of AVS. The profile of Pb was similar to that of Fe and Mn in core C2 in both seasons. However, in core C3, Pb displayed a vertical variation similar to that of Fe between the surface and 15 cm depth and similar to that of AVS below 15 cm depth. There was no similarity between Pb and Fe and Mn or AVS in cores C1 and C4.

L. Duan et al. / Science of the Total Environment 675 (2019) 247–259

4. Discussion 4.1. Early diagenesis of Fe and Mn 4.1.1. Porewater Fe and Mn cycling 4.1.1.1. Surface porewater. The diagenetic process in porewater is crucial for understanding the migration mechanism, cycling and diffusive flux of trace metals at the SWI (Dang et al., 2015; Kalnejais et al., 2015; Santos-Echeandia et al., 2009; Ramírez-Pérez et al., 2015). Despite the differences in absolute porewater concentrations between regions and seasons, the profiles of dissolved Fe, Mn and sulfide in all cores were similar and showed distinct variations with depth. Generally, Fe and Mn (hydr)oxides were reduced to Fe2+ and Mn2+ and then released into porewater at suboxic and/or anoxic zones where organic matter was degraded due to bacterial activities. Subsequently, Fe2+ and Mn2 + spread upwards to the surface layer and may be removed by forming minerals. Due to the different temperatures, DO and organic matter in the overlying waters, the depths and thicknesses of the oxic, suboxic and anoxic zones varied based on regions and seasons. The most obvious variation and reactions occurred at the surface layer. In the seasonal hypoxic region, there were very low dissolved Fe and Mn concentrations at depths of 0.5–1.5 cm and 0–1 cm in spring, suggesting that the removal of Fe and Mn and penetrative depth of O2 was lower than 1.5 cm. Due to the oxic condition, the removal of Fe and Mn was mainly by forming Fe and Mn (hydr)oxides in spring. The depth of low Mn concentrations was shallower than that of low Fe due to the relatively slow oxidation process of Mn2+ compared with that of Fe2+ (Jørgensen and Kasten, 2006). In summer, upward profiles and elevated release of dissolved Fe and Mn indicated that O2 osmosis to sediments was low and thus enhanced the reduction of Fe and Mn (hydr)oxides. However, solid Fe and Mn contents in summer were similar to those in spring, indicating that the reductive release of Fe and Mn (hydr)oxides was removed again. The removal mechanisms for Fe and Mn were different. Due to hypoxic conditions in summer, Fe diagenesis was closely related to organic matter and S. Fe removal mainly depended on bonding with S2− to form FeS and/or pyrite (FeS2) (Canfield, 1989; Rickard and Luther, 2007). Dissolved sulfide and AVS were determined, and their concentrations were higher in summer than in spring, suggesting that the surface removal of Fe was due to the formation of FeSx. However, the Mn2+ variation in porewater was mainly controlled by autogenetic Mn oxides, which were insoluble within a certain pH range (7.0–8.5) in the oxygenating environment by forming a Mn carbonate phase other than MnS, unless the Mn concentration in the porewater reached abnormally high levels (hundreds of μM) (Aller et al., 2004; Elderfield et al., 1981; Kalnejais et al., 2015). The pH values in the sediments of the Changjiang Estuary were 7.04–7.62. Dissolved inorganic carbon concentrations were high in summer in this area (Qu et al., 2017). The porewater Mn concentrations in the seasonal hypoxic region were b80 μM. All of these evidences suggested that the surface removal of Mn was due to the formation of MnCO3. Moreover, the incomplete surface removal of Mn also indicated its control by carbonates (Gao et al., 2009; Wang and Wang, 2017). The seasonal cycle in porewater chemistry at this area was driven by DO, temperature and organic matter flux. In the overlying water of the seasonal hypoxic region, the water temperature increased from 12.0 to 16.0 °C to 25.0–28.0 °C, DO concentrations decreased from 7.21 mg/L to 3.44 mg/L, and the supply of fresh organic matter increased from spring to summer. These changes resulted in the O2 penetrative depth becoming shallower and the peaks of dissolved Fe and Mn becoming closer to the SWI in summer. In the aerobic region, the removal of Fe occurred in both seasons with dissolved Fe concentration close to zero at depths of 0–0.5 cm in spring and 0.5–1 cm in summer. The aerobic region presented an oxic condition during the whole year; thus, the seasonal variation in surface Fe and Mn was related to the water temperature and organic matter flux

255

rather than DO. In the overlying water of the aerobic region, the water temperature increased from 14.0 °C to 25.5 °C, and the supply of organic matter also increased with Chl a concentrations from 1.81 μg/L to 5.94 μg/L from spring to summer. Thus, the increase in organic matter degradation resulted in some Fe and Mn (hydr)oxides being reduced and released into the porewater. Additionally, bacterial activities in the suboxic layer were enhanced in summer, leading to additional release of Fe and Mn. The high concentration gradient favored upward diffusion, further contributing to the higher surface porewater Fe and Mn concentrations in summer than in spring. Compared to the seasonal hypoxic region, the aerobic region had relatively low surface porewater Fe and Mn concentrations. It suggested that although there was seasonal variation in the surface porewater Fe and Mn in the aerobic region, the seasonal concentration difference was small due to the main process of formation of Fe and Mn (hydr)oxides throughout the year. 4.1.1.2. Suboxic/anoxic porewater. Compared to the surface layer, the diagenesis process of Fe and Mn in the suboxic and anoxic layers was relatively coincident between regions and seasons. Generally, Fe and Mn were involved in the degradation of organic matter as electron acceptors in the suboxic layer. Fe and Mn (hydr)oxides were reduced by bacterial activities (Lourino-Cabana et al., 2014), and dissolved Fe2+ and Mn2+ were released into the porewater due to the low DO levels and then migrated upwards and downwards (Olson et al., 2017). Peaks of dissolved Mn appeared above Fe since Mn reduction was favored thermodynamically (Froelich et al., 1979; Santos-Echeandia et al., 2009). In addition, there appeared to be simultaneous release of sulfide with Fe, e.g., in the anoxic layer in spring and in the suboxic layer in summer, probably due to the coexistence of sulfate and ironreducing bacteria. In addition, peaks of dissolved Fe, Mn and sulfide moved upwards from spring to summer in both regions, leading to a shallower depth of the suboxic layer in summer. This finding was consistent with summer hypoxia. The decrease in porewater Fe and Mn concentrations in the anoxic layer was due to the formation and precipitation of a sequence of FeSx and MnCO3 compounds (Aller et al., 2004; Holmkvist et al., 2011; Ramírez-Pérez et al., 2015). However, the increase in sulfide in the anoxic layer was probably due to sulfate reduction (Canfield et al., 1993; Kalnejais et al., 2015). 4.1.2. Solid Fe and Mn variations The variations in the Fe and Mn profiles in the solid phase generally showed changes corresponding with those in porewater, supporting their diagenetic reaction. The solid Fe and Mn profiles and content variations presented obvious regional and seasonal differences. In the seasonal hypoxic region, due to the obvious seasonal change in the redox condition of the bottom water, the solid profiles of Fe and Mn showed significant differences. In spring, the overlying water was oxic; thus, surface enrichment in Fe and Mn in solid phases occurred due to the formation of Fe and Mn (hyr)oxides, corresponding to the removal of Fe and Mn from the surface porewater. On the other hand, dissolved Fe in the suboxic layer moved downwards to the anoxic layer and then combined with sulfide to form FeSx. Thus, high solid Fe contents also appeared in the anoxic layer. In summer, the peaks of Fe and Mn in both porewater and solid phases moved upwards, indicating that suboxic and anoxic layers moved upwards. In addition, the peaks of Fe and Mn in the solid phase were deeper than those in the porewater. Particularly, the corresponding peak depth of solid Fe and AVS suggested the upper reduction, release and downward migration of Fe to form FeSx. However, the peak values of Mn were lower than the Mn content in shale (850 ppm), suggesting that this accumulation of solid Mn could be mainly as a Mn carbonate, with less incorporation into pyrite due to low Mn and sulfide concentrations. In the aerobic region, due to the perennial oxidation conditions, the variation in the solid Fe and Mn profile was relatively stable in both spring and summer. These profiles had surface or subsurface high values that corresponded to low values in the porewater profile. This finding

256

L. Duan et al. / Science of the Total Environment 675 (2019) 247–259

was attributed to the formation of Fe and Mn (hydr)oxides in the surface layer after the upward migration of dissolved Fe and Mn from the suboxic layer, resulting in the surface removal of Fe and Mn from the porewater. The difference was that Fe and Mn in the solid phase displayed a downward increase, whereas Fe and Mn in the porewater phase presented stable low values below 10 cm depth in summer. This finding was ascribed to the downward migration of dissolved Fe and Mn from the suboxic layer to the anoxic layer to form FeSx and MnCO3, corresponding to high AVS. The obvious regional difference was presented in summer, when hypoxia formed in the seasonal hypoxic region, accompanied by the reduction of Fe and Mn oxides and the release of dissolved Fe and Mn into the porewater; thus, solid Fe and Mn contents were lower than those in the aerobic region. In addition, the hypoxia formation favored sulfate reduction in the suboxic and/or anoxic layers, with correspondingly higher AVS. 4.2. Porewater dynamics of trace metals 4.2.1. Surface porewater Trace metal behaviors were controlled by the early diagenetic reactions within sediments (Kalnejais et al., 2015; Olson et al., 2017). The peaks of dissolved trace metals generally occurred at approximately 1 cm below SWI in the surface zone. A number of processes contributed to the observed high values, including the release of trace metals associated with the reductive dissolution of Fe and Mn (hydr)oxides (Canavan et al., 2007), degradation of organic matter (Tapia and Audry, 2013), and oxidation of AVS phases (Kalnejais et al., 2015). In spring, there was at least a 0.5 cm deepening of the dissolved Fe front compared to summer, which suggested that the O2 penetration depth deepened over spring. Thus, the higher surface porewater concentrations of trace metals likely resulted from the oxidation and release of reduced trace metal species, such as those associated with organic matter and AVS phases. In particular, Cu and Zn were bioactive trace metals, so their release from organic matter was likely a source of Cu and Zn in surface porewater. In summer, the surface release of trace metals mainly depended on the reduction of Fe and Mn oxides. However, the decreasing DO concentration and the increasing organism activities introduced more reduced species, such as sulfide and organic matter, in summer, which recombined with released trace metals to form metal sulfides or organic matter bound fraction. Consequently, this process led to the relatively low dissolved trace metal concentrations at the surface layer in summer. 4.2.2. Ferruginous zone The ferruginous chemical zone was mainly distributed at 2–10 cm depth, corresponding to the suboxic layer, where the porewater Fe concentrations were high. This zone was generated due to the biotic and abiotic reductions of Fe and Mn (hydro)oxides. Fe and Mn oxides were known to adsorb a large range of trace metals (Kalnejais et al., 2015), which could be released into porewaters during reductive dissolution of Fe and Mn (hydr)oxides. The significant relationships between porewater trace metals and Fe were only observed in summer in the seasonal hypoxic area and in spring in the aerobic region. The correlation coefficients were 0.818 (p b 0.01) for Co, 0.479 (p b 0.05) for Cu, 0.458 (p b 0.05) for Zn and 0.579 (p b 0.01) for Cd in the ferruginous zone of core C2 in summer and 0.596 (p b 0.01) for Co, 0.480 (p b 0.05) for Zn, 0.635 (p b 0.01) for Cd and 0.557 (p b 0.01) for Pb in the ferruginous zone of core C3 in spring. These accumulations of porewater trace metals in the ferruginous zone were due to the occurrence of strong Fe (hydr)oxide reduction (Naylor et al., 2004; Wang and Wang, 2017). However, there was no correlation between trace metals and Fe in spring in the seasonal hypoxic region and in summer in the aerobic region, indicating that trace metals were associated with other phases (e.g., sulfides). Precipitation as sulfides and/or coprecipitation with FeSx solid phases above the sulfidic zone was likely

an important removal mechanism. Due to the low solubility of Cu, Pb, Zn, Cd and Cr sulfides, even if only trace levels of sulfide were present, these metals would precipitate as sulfide phases (Kalnejais et al., 2015; Olson et al., 2017; Rosenthal et al., 1995). In addition, there was no correlation between trace metals and Mn in both seasons, except for Co. Similar porewater profiles were observed between Co and Mn in both regions and seasons with correlation coefficients of 0.728–0.790 (p b 0.01), suggesting that Co was more easily adsorbed onto Mn oxides (Santos-Echeandia et al., 2009) and that sedimentary mobility of Co was strongly controlled by Mn. 4.2.3. Sulfidic zone The sulfidic chemical zone was located at below 15 cm depth, corresponding to the anoxic layer, where the porewater sulfide concentrations were high. The lower dissolved trace metal concentrations in the sulfidic zone than in the ferruginous zone (except for core C1) were attributed to the formation of metal sulfides and/or coprecipitation with FeSx. It was suggested that the removal of Cd and Pb from the sulfidic zone was mainly by forming metal sulfides, and the removal of Cu and Zn was by forming metal sulfides and being associated with FeSx (Kalnejais et al., 2015; Olson et al., 2017; Santos-Echeandia et al., 2009). Dissolved trace metal concentrations in the sulfidic zone of the aerobic region were relatively constant and lower than those in the surface layer probably due to the increasing availability of sulfide for forming insoluble metal sulfides. However, the peaks of trace metals (Cu, Pb, Zn, Cr and Cd) also appeared in the deep sediments (below 15 cm depth) of the seasonal hypoxic region, which covaried with sulfide peaks but were opposite to AVS peaks. These trace metals underwent significant increases in the periods corresponding to high sulfate reduction and high sulfide production. This covariation was unconformable to the high insolubility of metal sulfides. The observed porewater trace metal increase in deep layer can be explained their bond with dissolved organic matter (DOM), thus enhancing their (especially Cr) solubility (Dang et al., 2015; Lourino-Cabana et al., 2014). It was suggested that elevated porewater Cr concentration under reducing conditions may occur due to preferential complexation with DOM (Beck et al., 2008; O'Connor et al., 2015). The formation of dissolved metal-polysulfide complexes and metal-sulfide-DOM ternary interactions were other possible reasons for increasing porewater trace metal concentrations in the deep sulfidic zone despite the sulfide presence (Dang et al., 2015; Hoffmann et al., 2012; Wang and Tessier, 2009). 4.2.4. Seasonal migration of trace metals The variation range of trace metals between seasons was generally higher than the site variation, validating that the sedimentation dynamics of trace metals were significantly influenced by temporal variability. Due to the seasonal migrations in redox boundaries, the surface porewater trace metal concentrations varied seasonally. In summer, the redox horizons shoaled, and the ferruginous zone moved upwards. Dissolved Fe and Mn and adsorbed trace metals were released into the porewaters; thus, the trace metal profiles and the Fe and Mn profiles were coupled. The upper Cu and Zn maximums in summer had roughly equal gradients upwards and downwards. This finding suggested that Cu and Zn diffused both upwards and downwards and then were rescavenged by Fe and Mn (hydr)oxides at the surface layer and precipitated as sulfides at the deep layer, respectively. Due to the higher sulfide concentration and more reducing conditions, a larger fraction of dissolved trace metals precipitated as sulfides than readsorbed to Fe and Mn (hydr)oxides. Thus, it seems that there was a net transfer of trace metals from the oxide phase to the sulfide phase in summer. In spring, O2 penetrated downwards, and the reduced phases of trace metals close to the surface were oxidized, resulting in high values of porewater trace metals at the surface. The high values of dissolved trace metals at the surface were prone to diffuse across the SWI, where trace metals could be scavenged by freshly precipitated Fe and Mn(hydr)oxides (Kalnejais et al., 2015).

L. Duan et al. / Science of the Total Environment 675 (2019) 247–259

4.3. Trace metals in solid phase and partitioning coefficients Coupling porewater data with solid phase data was difficult because the porewater data were impacted on much shorter timescales than the solid phase data (Jahnke et al., 1982; Kalnejais et al., 2015). However, the porewater data supported that the surface enrichment in solid Fe and Mn in spring was from the oxidation of upward diffusing dissolved Fe and Mn. Correspondingly, a surface enrichment in solid trace metals in spring was also observed. There were two possible mechanisms to explain the surface enrichment in solid trace metals, which were reassociation with Fe and Mn (hydr)oxides or formation of sulfides. If trace metals were precipitated as sulfides, they would remain immobilized in summer when sedimentary conditions became more reducing, resulting in more surface enrichment in solid trace metals in summer. However, this surface enrichment was not observed in summer, further suggesting that formation of sulfides was not the main mechanism for surface enrichment in solid trace metals and that the scavenging of trace metals by Fe and Mn (hydr)oxides was the main mechanism in spring in the Changjiang Estuary. Although the significant relationship between trace metals and Fe and Mn concentrations in porewater was not observed in spring in the seasonal hypoxic region and in summer in the aerobic region, the coupling relationships between trace metals (except for Cd and Pb) and Fe (r = 0.487–0.940, p b 0.05) and Mn (r = 0.454–0.922, p b 0.05) in solid phases suggested that trace metals were associated with Fe and Mn (hydr)oxides. There was a mechanism maintaining surface enrichment of solid Fe and Mn and trace metals. When Fe and Mn (hydr)oxides were reduced, adsorbed trace metals were released to the porewater. When the sedimentary environment became oxic, dissolved trace metals were readsorbed to Fe and Mn (hydr)oxides, thereby causing the surface enrichment in solid trace metals. This process was more likely to occur in the aerobic region than in the seasonal hypoxic region due to the lower probability of encountering sulfides. Additionally, trace metals in solid phase had significant correlations with both Fe and Mn and AVS in summer in the seasonal hypoxic region, implying the joint control of Fe and Mn and S diagenesis on trace metal cycling. This result was consistent with the conclusion of porewater that simultaneous release of sulfide with Fe in the suboxic layer in summer was probably due to the coexistence of sulfate- and ironreducing bacteria. Cd in solid phase had significant correlations with AVS (r = 0.620–0.630, p b 0.01), suggesting that Cd cycling was controlled by S. Pb variation in solid profiles was complicated. In core C2, Pb had positive relationships with Fe (r = 0.648–0.791, p b 0.01) and Mn (0.687–0.695, p b 0.01) in both seasons, suggesting that Pb was mainly controlled by Fe and Mn diagenesis. However, in core C3, the Pb variation had a positive correlation with Fe above 15 cm depth (r = 0.876, p b 0.01) and with AVS below 15 cm depth (r = 0.521, p b 0.01), indicating that Pb was controlled by Fe and Mn in the surface and ferruginous zones but by S in the sulfidic zone. The relationship between Pb and Fe and Mn and AVS was not observed in cores C1 and C4, likely indicating that the Pb dynamics were disturbed by anthropogenic inputs and frequent benthic organism activities. Consequently, various factors, such as redox, Fe, Mn and sulfide, controlled the distribution of trace metals between solid and liquid phases in the sediments. The solid-solution interactions could be described using the partitioning coefficient Kd (L/kg), calculated as the ratio of the metal concentration in the solid phase (mg/kg) and in the dissolved phase (mg/L) (Wang and Wang, 2017). The logKd may be seen as a measure of the reactivity, transport and fate of chemicals in porewaters (Emili et al., 2016). The logKd values of trace metals are illustrated in Table 1. There was no significant difference for the logKd values with depth, which was indicative of a steady equilibrium between dissolved and solid phases. However, the differences were apparent in terms of order of magnitude among metals since the logKd values varied from 2.85 for Mn to 7.87 for Fe. The higher logKd values for Fe, Cr, Co and Pb were probably due to their trapping in the large reducible or

257

Table 1 Average LogKd values for trace metals in cores C1–C4 of the Changjiang Estuary in spring and summer. LogKd Spring

Summer

C1 C2 C3 C2 C4

Mn

Fe

Cu

Zn

Cr

Cd

Pb

Co

2.85 2.34 2.55 2.54 2.67

7.87 – 4.66 4.03 4.38

3.93 3.86 4.01 3.91 4.11

3.79 3.8 4.12 3.28 3.43

4.78 4.64 4.99 4.61 4.59

3.98 4.12 3.65 3.28 3.42

4.58 4.13 4.70 4.42 4.49

4.23 4.14 4.49 4.45 4.53

oxidizable fractions, but the lower logKd values for Cd, Cu and Zn were likely due to their maintaining in a large exchangeable fraction (Charriau et al., 2011; Wang and Wang, 2017). The logKd values of trace metals in the aerobic region were generally higher than those in the seasonal hypoxic region. This result was attribute to the formation of Fe/Mn (hydr)oxides in the aerobic region, which effectively removed porewater trace metals. 4.4. Benthic fluxes Trace metal cycling in sediments after deposition is crucial regarding benthic fluxes (Pakhomova et al., 2007; Santos-Echeandia et al., 2009; Tang et al., 2016). The trace metal concentrations in the surface porewaters were higher than those in the overlying waters in both seasons, indicating that sediments showed a distinct diffusion of trace metals from the surface porewater to the water column (Rigaud et al., 2013). This finding could be supported by the benthic exchange fluxes. Benthic fluxes of trace metals in the Changjiang Estuary were calculated (Table 2). Benthic fluxes of Mn and Fe at the SWI were higher than those of trace metals and were 0.122 × 103–1.03 × 103 nmol/cm2/yr for Mn and 0.313 × 103–0.476 × 103 nmol/cm2/yr for Fe in spring and 1.07 × 103–1.13 × 103 nmol/cm2/yr for Mn and 2.06 × 103–4.58 × 103 nmol/cm2/yr for Fe in summer. Benthic fluxes of Mn and Fe in spring were lower than those in summer. Contrarily, benthic fluxes of trace metals in spring were higher than those in summer. Among the trace metals, Zn had the maximum benthic fluxes, with 18.6–117 nmol/cm2/yr in spring and 31.6–32.6 nmol/cm2/yr in summer. It was followed by Cu and Cr, with benthic fluxes of 1.65–24.1 nmol/cm2/yr for Cu and 2.22–16.6 nmol/cm2/yr for Cr in spring and 1.40–1.72 nmol/cm2/yr for Cu and 0.698–1.06 nmol/cm2/yr for Cr in summer. Benthic fluxes for the rest of trace metals were relatively low, with 1.23–2.10 nmol/cm2/yr for Co, 0.678–4.15 nmol/cm2/yr for Pb and 0.042–0.208 nmol/cm2/yr for Cd in spring and 0.247–0.338 nmol/cm2/yr for Co, 0.032– 0.092 nmol/cm2/yr for Pb and 0.020–0.088 nmol/cm2/yr for Cd in summer. It could be seen that all benthic fluxes of trace metals displayed positive values, indicating upward fluxes and that sediments served as the source for trace metals in both spring and summer. Although benthic fluxes of trace metals were positive in all cores of the Changjiang Estuary, their values displayed seasonal and regional variations. Seasonally, the higher benthic flux of Mn and Fe in summer than in spring suggested that the reductive release of Mn and Fe oxides and diffusion from deeper sediments in summer was greater than the oxidative release in spring. This result further indicated that Mn mainly

Table 2 Benthic fluxes across the sediment-seawater interface for cores C1-C4 of the Changjiang Estuary in spring and summer. Benthic flux (nmol/cm2/yr) Spring

Summer

C1 C2 C3 C2 C4

Mn (×103)

Fe (×103)

0.122 1.03 0.510 1.13 1.07

0.476 – 0.313 4.58 2.06

Co

Cu

Cr

Cd

Zn

Pb

1.79 1.23 2.10 0.338 0.247

24.1 1.65 8.66 1.40 1.72

16.6 3.02 2.22 1.06 0.698

0.208 0.042 0.042 0.088 0.020

117 70.2 18.6 32.6 31.6

4.15 1.06 0.678 0.092 0.032

258

L. Duan et al. / Science of the Total Environment 675 (2019) 247–259

existed in carbonates rather than sulfides. However, benthic fluxes for most of the trace metals in spring were higher than those in summer, which was attributed to the lower trace metal concentrations in the overlying waters and the oxidation of trace metal sulfides or those associated with organic matter in spring. As the main terrestrial input, the freshwater discharge of the Changjiang River was low in spring, resulting in the low dissolved trace metal inputs to the coastal waters. Additionally, due to the saturated DO in the overlying water in spring, trace metals associated with sulfides or organic matter were oxidized and released into the porewater, enhancing the trace metal concentration gradients between the porewater and the overlying water. Regionally, benthic fluxes for most of the trace metals in the seasonal hypoxic region were higher than those in the aerobic region in both seasons. This finding was attributed to the lower DO in the overlying waters and more drastic redox action at the SWI in the seasonal hypoxic region than in the aerobic region. In addition, there was high productivity in the seasonal hypoxic region. The algal decomposition could establish a relatively high redox condition appropriate redissolution and reduction of Fe oxides. Consequently, associated trace metals were released across the SWI, leading to more labile trace metals diffused into the overlying waters. Sediments have been shown to supply dissolved trace metals to the water column in quantities equivalent to or even greater than riverine fluxes in many regions (Santos-Echeandia et al., 2009; Warnken et al., 2001). In this study, the contribution of benthic fluxes of trace metals to the water column was comparable to their riverine fluxes. The riverine fluxes of trace metals were calculated by dissolved trace metal concentrations in the Changjiang River and freshwater discharge (9.25 × 1011 m3/yr). The average concentrations of trace metals in the Changjiang River were 2.57 μg/L for Cu, 7.06 μg/L for Zn, 1.12 μg/L for Pb, 3.93 μg/L for Cr, 0.150 μg/L for Cd and 0.410 μg/L for Co (Yang et al., 2014). Correspondingly, their riverine fluxes were 2.38 × 109 g/yr for Cu, 6.52 × 109 g/yr for Zn, 1.04 × 109 g/yr for Pb, 3.63 × 109 g/yr for Cr, 1.34 × 108 g/yr for Cd and 3.79 × 108 g/yr for Co. In terms of the area of 3.70 × 104 km2 for the study area, benthic fluxes were 1.78 × 108 g/yr for Cu, 1.30 × 109 g/yr for Zn, 9.21 × 107 g/yr for Pb, 9.09 × 107 g/yr for Cr, 3.32 × 106 g/yr for Cd and 2.49 × 107 g/yr for Co. It could be seen that although benthic fluxes for most of the trace metals were lower than their riverine fluxes, except the same order of magnitude for Zn, sediment was a significant source of trace metals to the water column in the Changjiang Estuary. 5. Conclusions The seasonal cycling of porewater trace metals and their implications for benthic fluxes in the Changjiang Estuary were studied. Porewater trace metal behaviors and benthic fluxes presented obvious regional and seasonal differences, which were controlled by diagenetic processes and redox conditions. Particularly, the obvious seasonal and regional variations in the profiles were mainly presented in the surface layer. Regionally, surface porewater trace metals in the seasonal hypoxic region were higher than those in the aerobic region. This finding was attributed to the oxidative release of metal sulfides in spring and the reductive release of Fe and Mn (hydr)oxides in summer in the seasonal hypoxic region. However, perennial oxidation conditions in the aerobic region were conducive to the continued existence of Fe and Mn (hydr)oxides and associated trace metals. Seasonally, trace metal concentrations in the surface porewater decreased from spring to summer, especially in the seasonal hypoxic region. This result was due to the occurrence of hypoxia and the increase in sulfide in summer, which favored the formation of metal sulfides to remove the trace metals released by reductive dissolution of Fe and Mn (hydr)oxides. Consequently, the formation of Fe/Mn (hydr)oxides was effective for trace metal removal in oxidizing conditions, whereas the formation of metal sulfides was the main way for trace metal removal in reducing conditions. Both porewater and solid profiles suggested that the dynamics of Cu, Zn,

Cr and Co were mainly controlled by Fe and Mn cycling, the dynamics of Cd were driven by S cycling and the dynamics of Pb were disturbed by \anthropogenic inputs and benthic organism activities in the Changjiang Estuary. Although benthic fluxes for most of the trace metals were lower than their riverine fluxes, sediment was a significant source of trace metals to the water column in the Changjiang Estuary. Acknowledgements This paper was supported by the National Natural Science Foundation of China (No. 41676068), Aoshan Technology Innovation Program of the Qingdao National Laboratory for Marine Science and Technology, China (No. 2016ASKJ02-4), National Natural Science Foundation of China-Shandong Joint Fund (No. U1606404), Youth Innovation Promotion Association of the Chinese Academy of Sciences (No. 2016191) and Foundation for Huiquan Young Scholar of Institute of Oceanology, Chinese Academy of Sciences, China. References Aleksander-Kwaterczak, U., Zdechlik, R., 2016. Hydrogeochemical characteristics of interstitial water and overlying water in the lacustrine environment. Environ. Earth Sci. 75, 1352. Aller, R.C., Hannides, A., Heilbrun, C., Panzeca, C., 2004. Coupling of early diagenetic processes and sedimentary dynamics in tropical shelf environments: the Gulf of Papua deltaic complex. Cont. Shelf Res. 24, 2455–2486. Beck, M., Dellwig, O., Schnetger, B., Brumsack, H.J., 2008. Cycling of trace metals (Mn, Fe, Mo, U, V, Cr) in deep pore waters of intertidal flat sediments. Geochim. Cosmochim. Acta 72, 2822–2840. Berner, R., 1980. Early Diagenesis: A Theoretical Approach. Princeton University Press, Princeton, p. 241. Berner, R., 1984. Sedimentary pyrite formation: an update. Geochim. Cosmochim. Acta 48, 605–615. Burton, E.D., Sullivan, L.A., Bush, R.T., Johnston, S.G., Keene, A.F., 2008. A simple and inexpensive chromium-reducible sulfur method for acid-sulfate soils. Appl. Geochem. 23, 2759–2766. Canavan, R., Van Cappellen, P., Zwolsman, J., Van den Berg, G., Slomp, C., 2007. Geochemistry of trace metals in a fresh water sediment: field results and diagenetic modeling. Sci. Total Environ. 381, 263–279. Canfield, D.E., 1989. Reactive iron in marine sediments. Geochim. Cosmochim. Acta 53, 619–632. Canfield, D., Thamdrup, B., Hansen, J., 1993. The anaerobic degradation of organic matter in Danish coastal sediments: iron reduction, manganese reduction and sulfate reduction. Geochim. Cosmochim. Acta 57, 3867–3883. Charriau, A., Lesven, L., Gao, Y., Leermakers, M., Baeyens, W., Ouddane, B., Billon, G., 2011. Trace metal behaviour in riverine sediments: role of organic matter and sulfides. Appl. Geochem. 26, 80–90. Chen, C.C., Gong, G.C., Shiah, F.K., 2007. Hypoxia in the East China Sea: one of the largest coastal low-oxygen areas in the world. Mar. Environ. Res. 64, 399–408. Cline, J., 1969. Spectrophotometric determination of hydrogen sulfide in natural waters. Limnol. Oceanogr. 14, 454–458. Dang, D.H., Lenoble, V., Durrieu, G., Omanović, D., Mullot, J.U., Mounier, S., Garnier, C., 2015. Seasonal variations of coastal sedimentary trace metals cycling: insight on the effect of manganese and iron (oxy) hydroxides, sulphide and organic matter. Mar. Pollut. Bull. 92, 113–124. Duan, L.Q., Song, J.M., Yuan, H.M., Li, X.G., Li, N., 2017. The use of sterols combined with isotope analyses as a tool to identify the origin of organic matter in the East China Sea. Ecol. Indic. 83, 144–157. Elderfield, H., McCaffrey, R., Luedtke, N., Bender, M., Truesdale, V., 1981. Chemical diagenesis in Narragansett Bay sediments. Am. J. Sci. 281, 1021–1055. Emili, A., Acquavita, A., Covelli, S., Spada, L., Di Leo, A., Giandomenico, S., Cardellicchio, N., 2016. Mobility of heavy metals from polluted sediments of a semi-enclosed basin: in situ benthic chamber experiments in Taranto's Mar Piccolo (Ionian Sea, Southern Italy). Environ. Sci. Pollut. Res. 23, 12582–12595. Froelich, P.N., Klinkhammer, G.P., Bender, M.L., Luedtke, N.A., Heath, G.R., Cullen, D., Dauphin, P., Hammond, D., Hartman, B., Maynard, V., 1979. Early oxidation of organic-matter in pelagic sediments of the Eastern Equatorial Atlantic-suboxic diagenesis. Geochim. Cosmochim. Acta 43, 1075–1090. Gao, Y., Lesven, L., Gillan, D., Sabbe, K., Billon, G., De Galan, S., Elskens, M., Baeyens, W., Leermakers, M., 2009. Geochemical behavior of trace elements in sub-tidal marine sediments of the Belgian coast. Mar. Chem. 117, 88–96. Grasshoff, K., Kremling, K., Manfred, E., 1999. Methods of Seawater Analysis. vol. 9. Wileyvch, New York, p. 600. Ho, H.H., Swennen, R., Cappuyns, V., Vassilieva, E., Gerven, T.V., Tran, T.V., 2012. Potential release of selected trace elements (As, Cd, Cu, Mn, Pb and Zn) from sediments in Cam River-mouth (Vietnam) under influence of pH and oxidation. Sci. Total Environ. 435–436, 487–498. Hoffmann, M., Mikutta, C., Kretzschmar, R., 2012. Bisulfide reaction with natural organic matter enhances arsenite sorption: insights from X-ray absorption spectroscopy. Environ. Sci. Technol. 46, 11788–11797.

L. Duan et al. / Science of the Total Environment 675 (2019) 247–259 Holmkvist, L., Ferdelman, T.G., Jørgensen, B.B., 2011. A cryptic sulfur cycle driven by iron in the methane zone of marine sediment (Aarhus Bay, Denmark). Geochim. Cosmochim. Acta 75, 3581–3599. Jahnke, R., Heggie, D., Emerson, S., Grundmanis, V., 1982. Pore waters of the central Pacific Ocean: nutrient results. Earth Planet. Sci. Lett. 61, 233–256. Jørgensen, B.B., Kasten, S., 2006. Sulfur cycling and methane oxidation. Marine Geochemistry. Springer, Berlin, pp. 271–309. Kalnejais, L.H., Martin, W.R., Bothner, M.H., 2015. Porewater dynamics of silver, lead and copper in coastal sediments and implications for benthic metal fluxes. Sci. Total Environ. 517, 178–194. Li, Y.H., Gregory, S., 1974. Diffusion of ions in sea water and in deep-sea sediments. Geochim. Cosmochim. Acta 38, 703–714. Lourino-Cabana, B., Billon, G., Lesven, L., Sabbe, K., Gillan, D.C., Gao, Y., Leermakers, M., Baeyens, W., 2014. Monthly variation of trace metals in North Sea sediments. From experimental data to modeling calculations. Mar. Pollut. Bull. 87, 237–246. Morse, J.W., Luther, G.W., 1999. Chemical influences on trace metal-sulfide interactions in anoxic sediments. Geochim. Cosmochim. Acta 63, 3373–3378. Naylor, C., Davison, W., Motelica-Heino, M., Van Den Berg, G.A., Van Der Heijdt, L.M., 2004. Simultaneous release of sulfide with Fe, Mn, Ni and Zn in marine harbour sediment measured using a combined metal/sulfide DGT probe. Sci. Total Environ. 328, 275–286. O'Connor, A.E., Luek, J.L., McIntosh, H., Beck, A.J., 2015. Geochemistry of redox-sensitive trace elements in a shallow subterranean estuary. Mar. Chem. 172, 70–81. Oehler, T., Martinez, R., Schückel, U., Winter, C., Kröncke, I., Schlüter, M., 2015. Seasonal and spatial variations of benthic oxygen and nitrogen fluxes in the Helgoland Mud Area (southern North Sea). Cont. Shelf Res. 106, 118–129. Olson, L., Quinn, K.A., Siebecker, M.G., Luther, G.W., Hastings, D., Morford, J.L., 2017. Trace metal diagenesis in sulfidic sediments: insights from Chesapeake Bay. Chem. Geol. 452, 47–59. Pakhomova, S.V., Hall, P.O.J., Kononets, M.Y.U., Rozanov, A.G., Tengberg, A., Vershinin, A.V., 2007. Fluxes of iron and manganese across the sediment-water interface under various redox conditions. Mar. Chem. 107, 319–331. Qu, B.X., Song, J.M., Yuan, H.M., Li, X.G., Li, N., Duan, L.Q., 2017. Comparison of carbonate parameters and air-sea CO2 flux in the southern Yellow Sea and East China Sea during spring and summer of 2011. J. Oceanogr. 73, 365–382. Ramírez-Pérez, A.M., de Blas, E., García-Gil, S., 2015. Redox processes in pore water of anoxic sediments with shallow gas. Sci. Total Environ. 538, 317–326. Rickard, D., Luther, G.W., 2007. Chemistry of iron sulfides. Chem. Rev. 107, 514–562. Rigaud, S., Radakovitch, O., Couture, R.M., Deflandre, B., Cossa, D., Garnier, C., Garnier, J.M., 2013. Mobility and fluxes of trace elements and nutrients at the sediment-water interface of a lagoon under contrasting water column oxygenation conditions. Appl. Geochem. 31, 35–51. Rivera-Duarte, I., Flegal, A.R., 1997. Porewater gradients and diffusive benthic fluxes of Co, Ni, Cu, Zn, and Cd in San Francisco Bay. Croat. Chem. Acta 70, 389–417. Rosenthal, Y., Lam, P., Boyle, E., Thomson, J., 1995. Authigenic cadmium enrichments in suboxic sediments: precipitation and postdepositional mobility. Earth Planet. Sci. Lett. 132, 99–111. Santos-Echeandia, J., Prego, R., Cobelo-García, A., Millward, G.E., 2009. Porewater geochemistry in a Galician Ria (NW Iberian Peninsula): implications for benthic fluxes of dissolved trace elements (Co, Cu, Ni, Pb, V, Zn). Mar. Chem. 117, 77–87.

259

Sawatari, H., Fujimori, E., Haraguchi, H., 1995. Multi-element determination of trace elements in seawater by gallium coprecipitation and inductively coupled plasma mass spectrometry. Anal. Sci. 11, 369–374. Sundby, B., 2006. Transient state diagenesis in continental margin muds. Mar. Chem. 102, 2–12. Tang, W.Z., Duan, S.H., Shan, B.Q., Zhang, H., Zhang, W.Q., Zhao, Y., Zhang, C., 2016. Concentrations, diffusive fluxes and toxicity of heavy metals in pore water of the Fuyang River, Haihe Basin. Ecotoxicol. Environ. Saf. 127, 80–86. Tapia, J., Audry, S., 2013. Control of early diagenesis processes on trace metal (Cu, Zn, Cd, Pb and U) and metalloid (As, Sb) behaviors in mining- and smelting-impacted lacustrine environments of the Bolivian Altiplano. Appl. Geochem. 31, 60–78. Torres, E., Ayora, C., Jiménez-Arias, J.L., García-Robledo, E., Papaspyrou, S., Corzo, A., 2014. Benthic metal fluxes and sediment diagenesis in a water reservoir affected by acid mine drainage: a laboratory experiment and reactive transport modeling. Geochim. Cosmochim. Acta 139, 344–361. Ullman, W.J., Sandstrom, M.W., 1987. Dissolved nutrient fluxes from the nearshore sediments of Bowling Green Bay, central Great Barrier Reef lagoon (Australia). Estuar. Coast. Shelf Sci. 24, 289–303. Vizzini, S., Di Leonardo, R., Costa, V., Tramati, C.D., Luzzu, F., Mazzola, A., 2013. Trace element bias in the use of CO2 vents as analogues for low pH environments: implications for contamination levels in acidified oceans. Estuar. Coast. Shelf Sci. 134, 19–30. Wang, F., Tessier, A., 2009. Zero-valent sulfur and metal speciation in sediment porewaters of freshwater lakes. Environ. Sci. Technol. 43, 7252–7257. Wang, W.H., Wang, W.X., 2017. Trace metal behavior in sediments of Jiulong River Estuary and implication for benthic exchange fluxes. Environ. Pollut. 225, 598–609. Warnken, K.W., Gill, G.A., Griffin, L.L., Santschi, P.H., 2001. Sediment-water exchange of Mn, Fe, Ni and Zn in Galveston Bay, Texas. Mar. Chem. 73, 215–231. Williams, M.R., Millward, G.E., Nimmo, M., Fones, G., 1998. Fluxes of Cu, Pb and Mn to the Northeastern Irish Sea: the importance of sedimental and atmospheric inputs. Mar. Pollut. Bull. 36, 366–375. Yang, Z.F., Xia, X.Q., Wang, Y.P., Ji, J.F., Wang, D.C., Hou, Q.Y., Yu, T., 2014. Dissolved and particulate partitioning of trace elements and their spatial–temporal distribution in the Changjiang River. J. Geochem. Explor. 145, 114–123. Zhang, J., Liu, S.M., Ren, J.L., Wu, Y., Zhang, G.L., 2007. Nutrient gradients from the eutrophic Changjiang (Yangtze River) estuary to the oligotrophic Kuroshio waters and reevaluation of budgets for the East China Sea shelf. Prog. Oceanogr. 74, 449–478. Zhang, C., Yu, Z.G., Zeng, G.M., 2014. Effects of sediment geochemical properties on heavy metal bioavailability. Environ. Int. 73, 270–281. Zhao, J., Feng, X.W., Shi, X.L., Bai, Y.C., Yu, X.G., Shi, X.F., Zhang, W.Y., Zhang, R.P., 2015. Sedimentary organic and inorganic records of eutrophication and hypoxia in and off the Changjiang Estuary over the last century. Mar. Pollut. Bull. 99, 76–84. Zhou, F., Chai, F., Huang, D.J., Xue, H.J., Chen, J.F., Xiu, P., Xuan, J.L., Li, J., Zeng, D.Y., Ni, X.B., Wang, K., 2017. Investigation of hypoxia off the Changjiang estuary using a coupled model of ROMS-CoSiNE. Prog. Oceanogr. 159, 237–254. Zhu, M.X., Liu, J., Yang, G.P., Li, T., Yang, R.J., 2012. Reactive iron and its buffering capacity towards dissolved sulfide in sediments of Jiaozhou Bay, China. Mar. Environ. Res. 80, 46–55. Zhu, J., Zhu, Z., Lin, J., Wu, H., Zhang, J., 2016. Distribution of hypoxia and pycnocline off the Changjiang estuary, China. J. Mar. Syst. 154, 28–40.