In Situ, High-Resolution Profiles of Labile Metals in Sediments of Lake Taihu

Characterizing labile metal distribution and biogeochemical behavior in sediments is crucial for understanding their contamination characteristics in lakes, for which in situ, high-resolution data is scare. The diffusive gradient in thin films (DGT) technique was used in-situ at five sites across Lake Taihu in the Yangtze River delta in China to characterize the distribution and mobility of eight labile metals (Fe, Mn, Zn, Ni, Cu, Pb, Co and Cd) in sediments at a 3 mm spatial resolution. The results showed a great spatial heterogeneity in the distributions of redox-sensitive labile Fe, Mn and Co in sediments, while other metals had much less marked structure, except for downward decreases of labile Pb, Ni, Zn and Cu in the surface sediment layers. Similar distributions were found between labile Mn and Co and among labile Ni, Cu and Zn, reflecting a close link between their geochemical behaviors. The relative mobility, defined as the ratio of metals accumulated by DGT to the total contents in a volume of sediments with a thickness of 10 mm close to the surface of DGT probe, was the greatest for Mn and Cd, followed by Zn, Ni, Cu and Co, while Pb and Fe had the lowest mobility; this order generally agreed with that defined by the modified BCR approach. Further analyses showed that the downward increases of pH values in surface sediment layer may decrease the lability of Pb, Ni, Zn and Cu as detected by DGT, while the remobilization of redox-insensitive metals in deep sediment layer may relate to Mn cycling through sulphide coprecipitation, reflected by several corresponding minima between these metals and Mn. These in situ data provided the possibility for a deep insight into the mechanisms involved in the remobilization of metals in freshwater sediments.


Introduction
Heavy metals in aquatic ecosystems come mainly from Nature (background levels) and anthropogenic activities [1]. Over the past few decades, there has been an increase in the discharge of untreated or partially treated sewage from industry, agriculture, and urban life. The resulting deposition of specific heavy metals has contributed to significant heavy metals pollution [2][3][4].
the DGT profiles. The mechanisms responsible for the remobilization of these metals in sediments were further revealed by relating the distributions of DGT-labile fractions to potentially governing factors. Figure 1 shows the five Lake Taihu sites selected for this study; samples were collected in October 2014. These sites can be classified into three ecological types (Table 1). Sites 1 and 2 are located in algal-dominated regions (Meiliang and Zhushan Bays) with polluted characteristics. Site 5 is located in the macrophyte-dominated region (East Lake Taihu); this is an aquaculture region with fresh aquatic vegetation and small aquatic animals in the water column which feed crabs. Sites 3 and 4 are in the central and west part of the lake, respectively. There is no obvious macrophyte coverage in the two site regions, but they frequently suffer from algal blooms.

Description of Sampling Sites
Int. J. Environ. Res. Public Health 2016, 13,884 3 of 17 sediments were further revealed by relating the distributions of DGT-labile fractions to potentially governing factors. Figure 1 shows the five Lake Taihu sites selected for this study; samples were collected in October 2014. These sites can be classified into three ecological types (Table 1). Sites 1 and 2 are located in algal-dominated regions (Meiliang and Zhushan Bays) with polluted characteristics. Site 5 is located in the macrophyte-dominated region (East Lake Taihu); this is an aquaculture region with fresh aquatic vegetation and small aquatic animals in the water column which feed crabs. Sites 3 and 4 are in the central and west part of the lake, respectively. There is no obvious macrophyte coverage in the two site regions, but they frequently suffer from algal blooms.

DGT Preparation and Field Deployment
In a preliminary experiment, we used a typical DGT probe with a 0.90 mm-thickness diffusion layer to measure metals in Lake Taihu at a vertically spatial resolution of 3 mm, with a typical deployment time of 24 h. However, several trace metals could not be detected due to low masses accumulated by DGT. Consequently, a DGT probe with a thin diffusion layer was used in order to get higher accumulation masses of trace metals. Such a type of DGT has been increasingly used for high-resolution study [37]. Previous studies have shown that the uses of thin diffusion layers down to 0.01 mm thickness do not differ in terms of DGT response from that of a conventional diffusion layer (0.90 mm) [38]. For this study, the binding gel impregnated with Chelex-100 resin (Bio-Rad,

DGT Preparation and Field Deployment
In a preliminary experiment, we used a typical DGT probe with a 0.90 mm-thickness diffusion layer to measure metals in Lake Taihu at a vertically spatial resolution of 3 mm, with a typical deployment time of 24 h. However, several trace metals could not be detected due to low masses accumulated by DGT. Consequently, a DGT probe with a thin diffusion layer was used in order to get higher accumulation masses of trace metals. Such a type of DGT has been increasingly used for high-resolution study [37]. Previous studies have shown that the uses of thin diffusion layers down to 0.01 mm thickness do not differ in terms of DGT response from that of a conventional diffusion layer (0.90 mm) [38]. For this study, the binding gel impregnated with Chelex-100 resin (Bio-Rad, Hercules, CA, USA) was prepared based on Zhang and Davison [39]. When assembling the DGT probe, the Chelex gel was covered directly by a Durapore ® PVDF membrane (HVLP00010, Millipore, Shanghai, China) in a Perspex holder. The PVDF membrane has a pore size of 0.45 µm and a thickness of 0.10 mm, which was used as the thin diffusion layer. A sponge sheet was attached on the back of the Perspex holder, to identify the SWI location when retrieving the probe from the sediment after deployment [40]. These changes effectively facilitate field sampling with the DGT probe.
To prepare for the sampling event, the DGT probes were deoxygenated with nitrogen overnight and stored in a container filled with deoxygenated NaCl solution (0.1 M). DGT probes were inserted into the sediments using a customized releasing device [41] and deployed for 24 h. After retrieval, each probe was rinsed to remove visible sediment particles, and the SWI position was marked. The probes were placed in a sealed plastic bag at air temperature to prevent moisture loss and transported to the laboratory for analysis. The temperature of the overlying water was also recorded.

Sediment Sampling
The sediment cores at each site were collected using a gravity core sampler within a few minutes after insertion of the DGT device. Each core was sliced into 0.5 cm sections to a depth of 5 cm, and then into 1.0 cm sections from a depth of 5 to 10 cm. Each section was protected by high purity nitrogen gas. The sediment sections were lyophilized at −56 • C after cryopreservation, and then ground up to pass through a 100-mesh (0.15 mm) sieve. The samples were stored at 4 • C before analysis.

Sample Analysis
The Chelex binding gels were cut off at the marked SWI position; the gels were then removed from the DGT probes and vertically sliced into 3 mm sections. Each slice was transported into a centrifuge tube, and a 1.0 M HNO 3 solution was added to elute the metals in the binding gel [35]. The Mn, Zn, Ni, Cu, Pb, Co and Cd concentrations were analyzed using an ICP-MS (NexION 300, Perkin-Elmer, Waltham, MA, USA), and the Fe concentration was analyzed using the phenanthroline colorimetric method with a Epoch Microplate Spectrophotometer (BioTek, Winooski, VT, USA).
Basic chemical properties of the sediment samples were analyzed using standard methods [42]. The total heavy metal concentrations in sediment samples were determined using microwave digestion with a mixture of HF-HNO 3, followed by ICP-MS or spectrophotometer analysis after approximate dilution. The pH value was measured in a deionized water suspension, at a solid/liquid ratio of 1:5 (5 g:25 mL, after 2 h of shaking and equilibration) using a pH electrode. The organic matter content in the sediments was measured in terms of total organic carbon (TOC) using a TOC analyzer (TOC-V CPN, Shimadzu, Kyoto, Japan).
Fractionation of metals was performed using a modified BCR method [43]. Briefly, a weight of 1.0 g dried sediment was added into 100 mL centrifuge tube and was sequentially extracted using 40 mL of 0.11 M acetic acid for 16 h, 40 mL of 0.5 M hydroxyammonium chloride for 16 h, 10 mL of 8.8 M hydrogen peroxide for 1 h (pH 2, repeat one time) and then 50 mL 1 M ammonium acetate (pH 2) for 16 h. The extracted metals were water and acid soluble (F1), reducible (F2), and oxidisable forms (F3), respectively. The residue after the four extractions was digested using aqua regia and the measured metals were residual form (F4).

Data Processing and Statistical Analysis
The DGT measurements were interpreted as time-averaged flux (F DGT ) for each metal (pg·cm −2 ·s −1 ): where M is the mass (pg) of labile metals taken up by DGT. A is the exposure area of the gel (cm 2 ), and t is the deployment time (s).
Relative mobility was used to assess the mobility of different metals, and is defined as the ratio of metals accumulated by DGT to the total metals in sediment: where M is the mean accumulation mass by DGT in vertical profiles (ng·cm −2 ). A is the exposure area of DGT probe in the sediment (cm 2 ). The variable m is the mean total metal content in a volume of sediment (ng) with a thickness of 10 mm close to the surface of DGT probe. Modeling of DGT-induced diffusion showed that the metals accumulated by DGT mostly come from this thickness of sediment layer [44]. The m is calculated using total metal concentration, water content and sediment density.

Sediment pH and TOC
Sediment pH and TOC can influence metal lability, by changing the partition of labile forms [24,45,46]. Figure 2 presents the distribution profiles of pH and TOC in sediments of five sites. The pH value increased downward in surface sediments from the SWI to depths of 40, 50, 80, 20 and 30 mm at Sites 1 to 5, respectively, followed by a steady stage or a slight decrease until to the bottom of the profiles. The pH values at Site 4, ranging from 7.12 to 7.64, were evidently higher than those of other sites; the pH values at Site 3, ranging from 6.53 to 7.26, were followed after Site 4. Algal-and macrophyte-dominated regions showed similar pH variations across the profiles, with the pH values ranging from 6.22 to 6.99. Relative mobility was used to assess the mobility of different metals, and is defined as the ratio of metals accumulated by DGT to the total metals in sediment: where M is the mean accumulation mass by DGT in vertical profiles (ng·cm −2 ). A is the exposure area of DGT probe in the sediment (cm 2 ). The variable is the mean total metal content in a volume of sediment (ng) with a thickness of 10 mm close to the surface of DGT probe. Modeling of DGT-induced diffusion showed that the metals accumulated by DGT mostly come from this thickness of sediment layer [44]. The is calculated using total metal concentration, water content and sediment density.

Sediment pH and TOC
Sediment pH and TOC can influence metal lability, by changing the partition of labile forms [24,45,46]. Figure 2 presents the distribution profiles of pH and TOC in sediments of five sites. The pH value increased downward in surface sediments from the SWI to depths of 40, 50, 80, 20 and 30 mm at Sites 1 to 5, respectively, followed by a steady stage or a slight decrease until to the bottom of the profiles. The pH values at Site 4, ranging from 7.12 to 7.64, were evidently higher than those of other sites; the pH values at Site 3, ranging from 6.53 to 7.26, were followed after Site 4. Algal-and macrophyte-dominated regions showed similar pH variations across the profiles, with the pH values ranging from 6.22 to 6.99. Macrophyte-dominated region at Site 5 had the greatest contents of TOC, followed by algaldominated regions at Sites 1 and 2 and then transitional regions at Sites 3 and 4. The contents of TOC were on average 1.01%, 1.68%, 0.61%, 0.74% and 2.34% for Sites 1 to 5, respectively. Site 5 was also in an aquaculture region, where fresh aquatic vegetation and small aquatic animals are annually added in the water column to feed the crabs. Both the aquatic macrophyte decomposition and aquacultural activities contributed to the high content of TOC. The TOC had a downward decrease from 2.88% at Macrophyte-dominated region at Site 5 had the greatest contents of TOC, followed by algal-dominated regions at Sites 1 and 2 and then transitional regions at Sites 3 and 4. The contents of TOC were on average 1.01%, 1.68%, 0.61%, 0.74% and 2.34% for Sites 1 to 5, respectively. Site 5 was also in an aquaculture region, where fresh aquatic vegetation and small aquatic animals are annually added in the water column to feed the crabs. Both the aquatic macrophyte decomposition and aquacultural activities contributed to the high content of TOC. The TOC had a downward decrease from 2.88% at the SWI to 2.37% at a depth of 25 mm at Site 5. Slightly downward decreases of TOC were also observed in surface sediments of Sites 3 and 4, while no evident trend appeared at other two sites.

Total Concentrations of Metals
The vertical distributions of the eight metals in total concentrations are shown in Figure 3. Generally, the concentration of each metal kept stable with the depth of sediment. The relative standard error (RSD, n = 10) for each metal concentrations from the five sites are 18%, 16%, 12%, 18%, 14%, 17%, 12% and 24% for Fe, Mn, Zn, Ni, Cu, Pb, Co and Cd, respectively. All the metals showed a downward decreasing trend in Site 2 below the depth of 75 mm. Besides, Cd and Pb exhibited a downward decreasing trend at Site 3 below the depth of 35 mm. the SWI to 2.37% at a depth of 25 mm at Site 5. Slightly downward decreases of TOC were also observed in surface sediments of Sites 3 and 4, while no evident trend appeared at other two sites.

Total Concentrations of Metals
The vertical distributions of the eight metals in total concentrations are shown in Figure 3. Generally, the concentration of each metal kept stable with the depth of sediment. The relative standard error (RSD, n = 10) for each metal concentrations from the five sites are 18%, 16%, 12%, 18%, 14%, 17%, 12% and 24% for Fe, Mn, Zn, Ni, Cu, Pb, Co and Cd, respectively. All the metals showed a downward decreasing trend in Site 2 below the depth of 75 mm. Besides, Cd and Pb exhibited a downward decreasing trend at Site 3 below the depth of 35 mm. . Site 2 had the greatest total concentrations for all the eight metals. This site is in Zhushan Bay which has been seriously polluted by sewage emissions and from agricultural non-point sources [14,47]. The higher concentrations of all the metals above the 75 mm depth may reflect a 35-year pollution history based on an average sedimentation rate of 2.1 mm·yr −1 in Taihu [10]. In contrast, Site 3 showed the least accumulations of Mn, Zn, Ni and Cd in sediments, reflecting that this site was much less influenced by anthropogenic activity. In addition, Sites 4 and 5 showed the least accumulations of Cu and Pb and Co in sediments, respectively. . Site 2 had the greatest total concentrations for all the eight metals. This site is in Zhushan Bay which has been seriously polluted by sewage emissions and from agricultural non-point sources [14,47]. The higher concentrations of all the metals above the 75 mm depth may reflect a 35-year pollution history based on an average sedimentation rate of 2.1 mm·yr −1 in Taihu [10]. In contrast, Site 3 showed the least accumulations of Mn, Zn, Ni and Cd in sediments, reflecting that this site was much less influenced by anthropogenic activity. In addition, Sites 4 and 5 showed the least accumulations of Cu and Pb and Co in sediments, respectively.

Fractionation of Metals
The metals were fractioned into four forms using the revised BCR procedure [43]. The speciation of each metal in the five sampling sites is present in Figure 4. The distribution of these forms could be divided into four types among the eight metals. The first type was Fe, which was dominated by residual form (F4). This form accounted for on average 81% to 86% of total Fe in sediments for the five sites, respectively. The acid-soluble form then contributed to on average 12% to 17% of total Fe. The dominance of the residual form reflected that Fe was much inert in sediment.
The second type was Mn and Cd, both of which were dominated by water and acid-soluble form (F1) followed by reducible (F2). In heavily-polluted Site 2, the extracted Cd was mostly composed by the water and acid-soluble form (accounting for 58%-85% of the total extracted Cd). In Site 3 containing the least total concentration of Cd, the water and acid-soluble Cd was dominated above the depth of about 30 mm, followed by a sharp decrease to the bottom. The contrast changes between the two sites reflected that the acid-soluble Cd was much labile and could reflect the contamination of Cd in sediment. This was consistent with the previous recognition by others [20,48].
The third type was Pb, which was dominated by reducible form (F2) in sediments. The reducible Pb accounts for on average 77%, 83%, 76%, 77% and 80% of the total extracted Pb in Sites 1 to 5, respectively. The proportion of residual form had a notable increase in Site 5 which contained the greatest TOC. Plach and Warren have revealed that the mobilization of Pb was controlled by natural organic matter (NOM) in NOM-rich freshwater sediments [49]. Since the reducible form of metals in BCR fractionation method was considered less labile than acid-soluble form [43], Pb in sediments of Lake Taihu should be much less labile in comparison to Cd.
Other metals (Co, Ni, Cu and Zn) showed an intermediate situation for the distribution of the four forms. The F1 and F4 were the major form for Ni and Co, while F1 and F2 were the major forms for Cu and Zn. In heavily-polluted Site 2, the contribution of F1 had an evident increase in relative to the residual form, reflecting an increase in lability of the four metals; In Site 5 containing the greatest content of TOC, the contribution of F3 and/or F4 for Co, Ni and Cu increased notably in relative to other sites, reflecting an increase in their inertness in sediments.    The distributions of DGT-labile Co were similar to those of Mn at the five sites. Sites 1, 3 and 5 showed an increase and then decrease after a low-level stable stage. Sites 2 and 4 showed the sharp increases of labile Co from middle depths followed by relatively steady stages to the bottom of the profiles. The distributions of FDGT among Cu, Zn and Ni were similar, except for those at Site 2. The FDGT had large decreases with depth in the surface 30 mm for the three elements, followed by slight decreases to the bottom of the profiles. Similar decreases in the surface sediments appeared for Cu at Site 1 and Ni at Site 4. Sites 1, 3 to 5 then remained relative steady with small fluctuations or had decreasing trends to the bottom with depth. Site 2 had different distributions of FDGT among the three elements. The FDGT of Cu and Ni showed a downward increase and decrease in the surface 20 mm and 10 mm, respectively, followed by a decrease and stable stage to a depth 57 mm and 78 mm. After that, they both had downward increases to the bottom of the profiles. The FDGT of Zn had an increasing trend from the SWI to a depth of 36 mm, followed by a consistent decrease and then a stable stage to the bottom. The F DGT of Mn at Sites 2 and 4 had similar profiles: they remained steady with very low values in the surface sediment layers, followed by sharp increases to middle depths and then remained steady to the bottom of the sediments. Sites 1, 3 and 5 also had a sharp increasing phase of F DGT after a low-level stage in the surface sediment layer, followed by a downward decrease to the bottom of the sediment. The depths showing the onset of the sharp increasing phase of DGT-labile Mn were 10, 30, 18, 40 and 40 mm at Sites 1 to 5, respectively.

Vertical F DGT Profiles of Metals
The distributions of DGT-labile Co were similar to those of Mn at the five sites. Sites 1, 3 and 5 showed an increase and then decrease after a low-level stable stage. Sites 2 and 4 showed the sharp increases of labile Co from middle depths followed by relatively steady stages to the bottom of the profiles.
The F DGT value of Cd remained quite steady with the depth at Sites 1, 2, 4 and 5, except that a downward decrease appeared in the surface 20 mm layer at Site 2. Only Site 3 showed a considerable variation of F DGT across the profile, reflected by an increase from the depth of 5 mm to 25 mm and then a consistent decrease to the bottom of the sediment.
The distribution of F DGT for Pb showed considerable fluctuations with sediment depth. All the sites had a downward decrease in the surface 10-20 mm layer. Site 5 then remained steady to the bottom; Sites 1 and 3 had downward increasing trends to the bottom; Sites 2 and 4 had large variation with the depth, but both had visible minima with smaller F DGT at the depth of 78 mm.
The distributions of F DGT among Cu, Zn and Ni were similar, except for those at Site 2. The F DGT had large decreases with depth in the surface 30 mm for the three elements, followed by slight decreases to the bottom of the profiles. Similar decreases in the surface sediments appeared for Cu at Site 1 and Ni at Site 4. Sites 1, 3 to 5 then remained relative steady with small fluctuations or had decreasing trends to the bottom with depth. Site 2 had different distributions of F DGT among the three elements. The F DGT of Cu and Ni showed a downward increase and decrease in the surface 20 mm and 10 mm, respectively, followed by a decrease and stable stage to a depth 57 mm and 78 mm. After that, they both had downward increases to the bottom of the profiles. The F DGT of Zn had an increasing trend from the SWI to a depth of 36 mm, followed by a consistent decrease and then a stable stage to the bottom.
Visible minima with smaller values of F DGT relative to their adjacent data appeared at the depth of around 60 mm at Site 1 for Pb, Cu, Zn and Ni (marked as B in Figure 5), at the depth of around 78 mm at Site 2 for Pb and Ni (marked as A in Figure 5), at the depth of around 54 mm at Site 4 for Pb, Zn and Ni (marked as C in Figure 5). Similar minima of F DGT have appeared in the Mn profiles.

Overall Lability and Spatial Heterogeneity of Metals in Sediments
The measurement with DGT can reflect the lability of target analytes in sediments, since the measured species are two labile fractions of the analytes in sediments with high mobility, including dissolved fractions in pore water and easily-exchanged fractions released from sediment solids to resupply the pore water concentration due to DGT perturbance [31]. Table 3 presents the mean F DGT and the relative standard deviations (RSD, n = 33) of the eight metals in vertical profiles across the five sites. A total of five metals (Zn, Ni, Cu, Pb and Co) had the greatest mean F DGT at Site 2, which was consistent with their total concentrations in sediments ( Table 2). Since the DGT-labile metals have been confirmed as bioavailable fractions with excellent exposure concentration response to toxic effects on aquatic organisms [50,51], the organisms at Site 2 located in Zhushan Bay may face elevated risks from exposure of the five metals with great mass accumulation and high lability in sediments [20]. Table 3. Mean F DGT (pg·cm −2 ·s −1 ) and RSD (%) of eight metals in vertical profiles at five sites in Lake Taihu a . The mean F DGT at Site 1 to 4 was highest for Mn, followed by Fe, Zn, Ni, Cd, Cu, Pb, and then Co. This order was different from that of Fe, Mn, Zn, Ni, Cu, Pb, Co and Cd according to their total concentrations (Table 2), reflected by greater F DGT values for Mn and Cd. It demonstrates that the sediments with lower total concentrations of Mn and Cd have stronger ability in resupply to pore water from the solid phase during the uptake of DGT. The Cd was generally considered as a labile element, and the sediment solids had been found to show a strong ability in resupply of porewater Cd in Esthwait Water, UK [32].

Site
The variability (relative standard error, RSD) of the DGT-measured metal fluxes along the vertical direction was further used to assess the spatial heterogeneity in lability for each metal in sediments. The RSDs were in the ranges of 65.5%-154.1%, 50.3%-90.0%, 11.1%-38.5%, 11.4%-26.4%, 5.2%-64.6%, 14.5%-29.1%, 36.3%-59.3% and 5.5%-22.8% for Fe, Mn, Zn, Ni, Cu, Pb, Co and Cd, respectively. Fe, Mn and Co had the highest RSD values among the eight metals. As Fe and Mn were both redox-sensitive elements in sediments [52], the redox reactions, particularly those with Fe-and Mn-involved or controlled, should be vital factors influencing the variation of labile metals in sediments. The Cd had the lowest RSD values, reflecting that the lability of Cd remained relatively stable and should be less affected by redox reactions in comparison with other metals.

Relative Mobility of Metals in Sediments
The DGT accumulates labile fractions of metals in situ with high mobility in sediment [32][33][34]. The DGT-accumulated mass was used to demonstrate the relative mobility of metals in comparison with the total amounts of metals in a certain volume of sediment. Table 4 presents the relative mobility of eight metals at five sites in Lake Taihu. There was a large variation in the mobility among the eight metals. The relative mobility of Cd and Mn were the greatest, followed by Zn, Ni, Cu and Co, while Pb and Fe had the lowest mobility. The order was just consistent with the order arranged according to their distributions in chemical forms, especially the proportion of F1. Guevara-Riba et al. [48] studied the sediment mobility using the modified BCR approach, and defined mobility as the ratio of metals extracted in 0.11 M acetic acid (F1) to the sum of fractions; they also found that Cd had the greatest mobility in comparison with Zn, Pb, Cu, Ni and Cr. As the labile fractions of metals from chemical extraction are operationally defined with debate in reflection of their mobility [19,21,29], the DGT-derived indicator should be more accurate. Seven of the metals (all except Cu) had the highest relative mobility in the samples from Site 5, which was different from the order assessed by F1. Site 5 is located in East Lake Taihu, and is famous for its aquaculture. The biological effects of crabs on the metals may have made the metals more labile, even though the total concentration was not the highest. As with the metals from Site 5, Cu at Site 4 and Pb at Sites 2 and 4 also deserve more attention because of their higher mobility, making these metals more labile and more easily accumulated in organisms.

Mechanisms for Fe, Mn and Co Remobilization
The results have shown a sharp increasing phase for DGT-labile Fe with depth in sediments ( Figure 5). This increase phase has been observed with DGT measurements in both freshwater and seawater sediments [36,53]. It was attributed to the reductive dissolution of Fe (oxyhy)droxides under anoxic condition in deeper sediment layers and the resulted release and remobilization of Fe(II) [54,55]. The reducible Fe minerals in sediments are mainly ferrihydrite, lepidicrocite, goethite, hematite and akaganéite, while ferrihydrite and lepidicrocite may play a primary role in increasing the lability of Fe because they are more easily reducible in comparison with other Fe minerals [56,57].
Similar to the redox-sensitive Fe, the initial increasing phase of DGT-labile Mn was also attributed to reductive dissolution of Mn(IV) oxides. Both Mn(IV) oxides and Fe (oxyhy)droxides are major electron acceptors during the early diagenetic processes dominated by organic matter decomposition [52]. The depths showing the onset of the sharp increase of labile Mn were mostly shallower than those of Fe at four sites. It demonstrated that Mn(IV) oxides are more easily reduced than Fe (oxyhy)droxides, which is consistent with the terminal electron-accepting processes (TEAPs) in sediment where Mn(IV) reduction dominates followed by Fe(III) reduction [35,52]. On the other hand, oxidation of Mn(II) is slower than that of Fe(II), enable it to extend closer to the SWI if they contacts with oxygen. At Site 5, the depth showing the onset of the sharp increasing phase of DGT-labile Mn nearly equaled to that of labile Fe. A similar phenomenon has been observed in contaminated marine sediment. It was suggested that in strong reducing sediments, most released Mn(II) is adsorbed to freshly formed Fe oxides, rather than being oxidized [58]; The adsorbed Mn(II) can release again with Fe(III) reduction [45,58], causing simultaneous increases of DGT-labile Fe and Mn as observed.
The distributions of labile Co were very similar to those of labile Mn at the five sites ( Figure 5). This was further verified by positively significant correlations at very significant levels (p < 0.01) between them ( Figure 6). Their precise correspondence in profiles has been reported in both freshwater [54,59] and marine systems [33,35,[60][61][62]. Three mechanisms have been developed to explain their similarity in remobilization. (1) The reductive remobilization of Co and Mn required very similar redox conditions and they tended to occur simultaneously but independently; (2) Co may released from organic matter decomposition which supply electrons to cause the reduction of Mn(IV); (3) Co may be incorporated in Mn oxyhydroxides, and released with the reductive remobilization of Mn [35,61,63]. The last one should be the reason responsible for the similarity between the distributions of labile Co and Mn. It has been shown that Mn oxides had much greater preference than Fe oxides in trapping Co; Enrichment of Co within Mn oxides was 9 times that of Fe (oxyhydr)oxide from an investigation in marine sediment [62]. The presence of Co in individual Mn-oxide particles has also been verified by elemental analysis with electron microscopy [64]. And their association can be through sorption of Co(II) by Mn oxides [65]. Co(II) adsorbed to the oxide can be oxidized to Co(III) and subsequently incorporated into Mn oxide structure [66]. Co is then released from reductive dissolution of particulate Mn oxides, as verified in the water column of two freshwater lakes [63,65].
Int. J. Environ. Res. Public Health 2016, 13, 884 12 of 17 The distributions of labile Co were very similar to those of labile Mn at the five sites ( Figure 5). This was further verified by positively significant correlations at very significant levels (p < 0.01) between them ( Figure 6). Their precise correspondence in profiles has been reported in both freshwater [54,59] and marine systems [33,35,[60][61][62]. Three mechanisms have been developed to explain their similarity in remobilization. (1) The reductive remobilization of Co and Mn required very similar redox conditions and they tended to occur simultaneously but independently; (2) Co may released from organic matter decomposition which supply electrons to cause the reduction of Mn(IV); (3) Co may be incorporated in Mn oxyhydroxides, and released with the reductive remobilization of Mn [35,61,63]. The last one should be the reason responsible for the similarity between the distributions of labile Co and Mn. It has been shown that Mn oxides had much greater preference than Fe oxides in trapping Co; Enrichment of Co within Mn oxides was 9 times that of Fe (oxyhydr)oxide from an investigation in marine sediment [62]. The presence of Co in individual Mnoxide particles has also been verified by elemental analysis with electron microscopy [64]. And their association can be through sorption of Co(II) by Mn oxides [65]. Co(II) adsorbed to the oxide can be oxidized to Co(III) and subsequently incorporated into Mn oxide structure [66]. Co is then released from reductive dissolution of particulate Mn oxides, as verified in the water column of two freshwater lakes [63,65]. It should be noted that the dynamics of Mn(IV) reduction on controlling the remobilization of Co had a large variation among the 5 sites. The molar ratios of Co to Mn were on average 8.2 × 10 −4 , 3.7 × 10 −3 , 1.8 × 10 −3 , 6.1 × 10 −3 and 2.3 × 10 −3 for the Sites of 1 to 5, respectively. The values were similar to the Co/Mn ratio range of 1.1 × 10 −4 to 2.6 × 10 −3 reported by Lienamann et al. [64] from several studies of anoxic lakes. All the sites exhibited high values of Co/Mn ratio in the upper sediment layers followed by sharp decreases of the ratio ( Figure 6). The depths at the end of these decreases generally corresponded to the onset of the sharp increases of labile Mn (Figure 5), reflecting that the reductive dissolution of Mn(IV) has caused the relatively greater increase of DGT-labile Mn in comparison to Co and the decreases of Co/Mn ratio. The molar Co/Mn ratios had a small variation after their sharp decreases, ranging from 1.4 × 10 −4 at Site 5 to 7.6 × 10 −4 at Site 2. The reason for this variation in deeper It should be noted that the dynamics of Mn(IV) reduction on controlling the remobilization of Co had a large variation among the 5 sites. The molar ratios of Co to Mn were on average 8.2 × 10 −4 , 3.7 × 10 −3 , 1.8 × 10 −3 , 6.1 × 10 −3 and 2.3 × 10 −3 for the Sites of 1 to 5, respectively. The values were similar to the Co/Mn ratio range of 1.1 × 10 −4 to 2.6 × 10 −3 reported by Lienamann et al. [64] from several studies of anoxic lakes. All the sites exhibited high values of Co/Mn ratio in the upper sediment layers followed by sharp decreases of the ratio ( Figure 6). The depths at the end of these decreases generally corresponded to the onset of the sharp increases of labile Mn ( Figure 5), reflecting that the reductive dissolution of Mn(IV) has caused the relatively greater increase of DGT-labile Mn in comparison to Co and the decreases of Co/Mn ratio. The molar Co/Mn ratios had a small variation after their sharp decreases, ranging from 1.4 × 10 −4 at Site 5 to 7.6 × 10 −4 at Site 2. The reason for this variation in deeper sediment layers is not clear. Stockdale et al. [62] performed an incubation experiment with intact and aged/homogenised marine sediments, and found that the ageing and homogenization of the sediments has caused the decrease of the Co/Mn ratio from 1.1 × 10 −3 to 1.9 × 10 −4 . Co can be immobilized via formation of or incorporation into sulphide phases to a greater degree than Mn because that Co-sulphide phase is much lower soluble than Mn-sulphide phase. The variation of the Co/Mn may attribute to the immobilization of Co to a different degree via the formation of, or imcorporation into sulphide phases [62]. Further studies should be carried out to investigate the factors influencing the coupling between Mn and Co.

Mechanisms for Other Metal Remobilization
Previous studies have suggested three mechanisms responsible for the remobilization of trace metals at the SWI: (1) decomposition of organic material; (2) reductive dissolution of Fe/Mn (oxyhydro)xides; (3) desorption or adsorption at a pH gradient [32]. In this study, downward decreases in surface 10-20 mm sediment layer appeared for Pb at 5 sites, Ni at 2 sites and Cu and Pb at 1 site. This phenomenon was seldom reported in previous studies. These decreases could not be attributed to the decomposition of organic matter, which has been found to cause downward increases of metals and pronounced surface maximum of DGT-labile metals nearby the SWI [1,24,35,45]. In addition, the decomposition of organic matter was not notable at Sites 1 to 4, reflected by the relatively stable distributions of TOC ( Figure 2). The effect of reductive dissolution of Fe/Mn (oxyhydro)xides was also excluded. This process should induce the remobilization of trace metals and the increase of DGT-measured flux/concentration in surface sediments as observed previously in the Rupel River, Belgium [54] and in a sulphidic freshwater [67]. The decreases of labile metals may attribute to the downward increases of pH values in the surface layer at the five sites. The increase of pH could promote the precipitation of metals and decrease the dissolution of metals, resulting in the decreases of DGT fluxes as observed. The vertical decreases in totally four metal fluxes at Site 5 were accompanied by the largest decrease of the pH value, while other sites only showed the decreases of 1-2 metals. This difference could partly support the above hypothesis.
There were several corresponding minima of F DGT between Mn and metals at and below the depths of 60 mm ( Figure 5). The correspondences were more evident for Pb and Ni, for which three sites appeared this phenomenon. It was followed by Zn and Cu, for which two and one sites appeared this phenomenon. It reflects that the remobilizations of Pb, Ni, Zn and Cu in the deep sediments may to a certain degree relate to Mn cycling. As sulphate is the primary electron acceptor in deeper sediment layer when the reactive Fe and Mn oxides are consumed, the elevated sulphide production may be a major mechanism in removal of Mn and the four metals via sulphide precipitation. Yin et al. [68] reported that the concentration of acid volatile sulfide (AVS) in the sediments of Lake Taihu were 5.09 µmol·g −1 and 2.94 µmol·g −1 in the sediments of two sites in Meilang Bay, Lake Taihu. The ASV showed sharp downward increases below the depth of 6 cm and had maxima at the depths of around 8 cm. The depths with the maxima generally corresponded to the depths appearing the localized minima of F DGT for Mn and the four trace metals.
The vertical distributions of metal fluxes at the middle depths had no detailed spatial structure. The release of metals may from various sources, and two or more mechanisms should work separately or together to determine the remobilization of metals. Simultaneous measurements of metals and potentially governing factors at a fine scale using DGT and other high-resolution methods (e.g., planar optode) may greatly facilitate the discovery of those mechanisms.
Positive correlations were obtained for labile Ni with labile Cu and with labile Zn at totally 4 sites (Figure 7). Similar correspondence of Ni with Cu and Zn had been observed from a local remobilization associated with macrofauna in a deeper sediment layer of the North-East Atlantic [33]. It demonstrates that the three elements had a close link between the geochemical behaviors, including similar source materials and recycling processes.

Conclusions
The mean FDGT at four sites was the highest for Mn, followed by Fe, Zn, Ni, Cd, Cu, Pb, and then Co. The variability (RSD) of the FDGT along the vertical direction were much greater for Fe, Mn and Co than for the other metals, reflecting that the redox reactions were vital factors influencing the variation of labile metals in sediments. The relative mobility, which was calculated based on the use of FDGT, followed the order of Cd, Mn, Zn, Ni, Cu and Co, Pb and Fe. Cadmium in sediments deserved more attention because of its high mobility. Both DGT-labile Fe and Mn had sharp increasing phases with depth in sediments. The depths showing the sharp increase of labile Mn were mostly shallower than those of Fe at four sites, demonstrating that Mn(IV) oxides are more easily reduced than Fe (oxyhy)droxides. Both the depths were almost the same at Site 5, likely reflecting that in strong reducing sediments most released Mn(II) is adsorbed to freshly formed Fe oxides, rather than being oxidized. The distributions of labile Co were very similar to those of labile Mn at the five sites, with the Co/Mn ratios of 8.2 × 10 −4 , 3.7 × 10 −3 , 1.8 × 10 −3 , 6.1 × 10 −3 and 2.3 × 10 −3 , respectively. It demonstrated that Co should be incorporated in Mn oxyhydroxides, and released with the reductive remobilization of Mn, while the variation of the Co/Mn was attributed to the immobilization of Co to a different degree via the formation of, or incorporation into sulphide phases. Downward decreases in surface 10-20 mm sediment layer appeared for labile Pb, Ni and Cu at several sites, which were attributed to the downward increases of pH values in the same depth. The remobilizations of labile metals in deep sediment layer, together with several corresponding minima with Mn, may attribute to coprecipitation by elevated sulphide. In addition, Ni, Cu and Zn had a close link in their geochemical behaviors, reflected by the positive correlations of labile Ni with labile Cu and Zn.

Conclusions
The mean F DGT at four sites was the highest for Mn, followed by Fe, Zn, Ni, Cd, Cu, Pb, and then Co. The variability (RSD) of the F DGT along the vertical direction were much greater for Fe, Mn and Co than for the other metals, reflecting that the redox reactions were vital factors influencing the variation of labile metals in sediments. The relative mobility, which was calculated based on the use of F DGT , followed the order of Cd, Mn, Zn, Ni, Cu and Co, Pb and Fe. Cadmium in sediments deserved more attention because of its high mobility. Both DGT-labile Fe and Mn had sharp increasing phases with depth in sediments. The depths showing the sharp increase of labile Mn were mostly shallower than those of Fe at four sites, demonstrating that Mn(IV) oxides are more easily reduced than Fe (oxyhy)droxides. Both the depths were almost the same at Site 5, likely reflecting that in strong reducing sediments most released Mn(II) is adsorbed to freshly formed Fe oxides, rather than being oxidized. The distributions of labile Co were very similar to those of labile Mn at the five sites, with the Co/Mn ratios of 8.2 × 10 −4 , 3.7 × 10 −3 , 1.8 × 10 −3 , 6.1 × 10 −3 and 2.3 × 10 −3 , respectively. It demonstrated that Co should be incorporated in Mn oxyhydroxides, and released with the reductive remobilization of Mn, while the variation of the Co/Mn was attributed to the immobilization of Co to a different degree via the formation of, or incorporation into sulphide phases. Downward decreases in surface 10-20 mm sediment layer appeared for labile Pb, Ni and Cu at several sites, which were attributed to the downward increases of pH values in the same depth. The remobilizations of labile metals in deep sediment layer, together with several corresponding minima with Mn, may attribute to coprecipitation by elevated sulphide. In addition, Ni, Cu and Zn had a close link in their geochemical behaviors, reflected by the positive correlations of labile Ni with labile Cu and Zn.