Rare earth element and yttrium geochemistry in sinking particles and sediments of the Jiaozhou Bay, North China: Potential proxy assessment for sediment resuspension

Marine Pollution Bulletin 144 (2019) 79–91

Contents lists available at ScienceDirect

Marine Pollution Bulletin journal homepage: www.elsevier.com/locate/marpolbul

Baseline

Rare earth element and yttrium geochemistry in sinking particles and sediments of the Jiaozhou Bay, North China: Potential proxy assessment for sediment resuspension

T

Jin Liua,b,d, Jinming Songa,b,c,d, , Huamao Yuana,b,c,d, , Xuegang Lia,b,c,d, Ning Lia,b,c,d , Liqin Duana,b,c,d ⁎

⁎

a

Key Laboratory of Marine Ecology and Environmental Sciences, Institute of Oceanology, Chinese Academy of Sciences, Qingdao 266071, China Laboratory for Marine Ecology and Environmental Science, Qingdao National Laboratory for Marine Science and Technology, Qingdao 266237, China University of Chinese Academy of Sciences, Beijing 100049, China d Center for Ocean Mega-Science, Chinese Academy of Sciences, Qingdao 266071, China b c

ARTICLE INFO

ABSTRACT

Keywords: Rare earth elements Jiaozhou Bay Trapped particles Sediments Resuspension proxy

To exploit the resolving ability of rare earth element and yttrium (REY) in resuspension binary mixing model, and discover potential new REY-related resuspension proxy, this preliminary research studied the geochemical signature of REY in different Jiaozhou Bay samples including surficial/core sediments and settling trap-collected particles. Close quantitative relation for bulk concentration in particles, sediments and fine-grained fraction of major river sediments around the Yellow Sea, approved the priority contribution of catchment detrital materials. Moreover, common characteristics occurred for compartment-specific partitioning REY signatures in six operated-defined fractions, and multiple REY normalization pattern indexes (i.e. Y/Ho divergence, and Ce/Eu anomalies). All constrain the application of REY in resuspension discrimination of marginal shallow seas. However, linearity with different slopes and intercepts were plotted for the MREE bulge index versus HREE/ LREE figure in reducible amorphous Fe-oxides fraction, which could provide new discrimination perceptions.

Rare earth elements and yttrium (REY, the latter being included because of similar primary mineral sourcing and chemical reactivity as the lanthanides) are of great interest because their unique chemical properties make them especially powerful tracers of myriad fundamental geochemical processes (Pattan et al., 2005; Prakash et al., 2012; Song and Choi, 2009; Wang et al., 2016). The systematic variation in atomic mass, ionic radius, and electron configuration (i.e., the lanthanide contraction effect; Byrne and Kim, 1990), as well as variation in oxidation state across the REY lead to series wide trends in incorporation into (or exclusion from) secondary minerals and/or complexes with organic/inorganic ligands (Akagi, 2013; Bau and Koschinsky, 2009; Qiu et al., 2015; Reynard et al., 1999; Tang and Johannesson, 2010). Consequently, the REYs have a universal application in the study of soil and sediment provenance identification (Leybourne and Johannesson, 2008; Mao et al., 2010; Prakash et al., 2012; Song and Choi, 2009); tracers for agricultural microelement fertilizer or feed additive (Mihajlovic et al., 2014; Zhang and Shan, 2001); various biogeochemical processes such as mineral dissolution, redox fluctuations, and

biological cycling (Nozaki and Alibo, 2003; Vazquez-Ortega et al., 2016; Vazquez-Ortega et al., 2015); paleo-climate and paleo-environment reconstruction (Chen et al., 2015; Zhao et al., 2013). Recently, REE was used in the binary mixing model to correct for the effects of active resuspension and constrain particulate organic carbon (Corg) flux in the East China Sea (Hung et al., 2016). The quantification of resuspension provides crucial information to elucidate the mechanisms regarding to particle cycling and sedimentation, also evaluate the processes influencing the uptake, storage and transformation of carbon, nutrients and trace metals in marine environment (Bloesch, 1994; Gardner et al., 1985; Hung et al., 2013; Ye et al., 2013). Hence, regarding their less lability in bio-uptake/diagenesis process and distinctive bulk concentration for resuspended and the first settling materials (Song and Choi, 2009), the more consistent and convincing resuspension ratios acquired by REE approach would be expected than those obtained by other substitutive proxies, such as biogenic components (Corg, organic nitrogen and phosphorus), terrestrial elements (aluminum, iron, manganese and silicon) and unique natural/

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

https://doi.org/10.1016/j.marpolbul.2019.04.044 Received 28 February 2019; Received in revised form 15 April 2019; Accepted 16 April 2019 Available online 10 May 2019 0025-326X/ © 2019 Elsevier Ltd. All rights reserved.

Marine Pollution Bulletin 144 (2019) 79–91

J. Liu, et al.

Fig. 1. (a) East China marginal seas and the location of Jiaozhou Bay (inside the red rectangle). (b) Sampling stations in Jiaozhou Bay. The black triangle indicated one sediment core and six surface sediment sampling stations, the hollow pentagon revealed trap stations. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

anthropogenic fallout radioisotopes (Avnimelech et al., 1999; Gasith, 1976; Hung et al., 2016; Hung et al., 2013; Matisoff and Carson, 2014; Matisoff et al., 2017). However, like trace metals, chemical reactions with soluble ligands, organic matter, clay minerals, carbonates, phosphates, sulphur, and Fe/Mn-(oxy)hydroxides, as well as variations in pH and redox status have all been reported to influence REY fractionation, transport, and fate in natural low-temperature surface environment (Caetano et al., 2009; Du Laing et al., 2009). The oceanic processes related to resuspension in marginal shallow seas would superimposed certain influences on the original and/or ‘preformed’ continental signal of terrigenous particles (Bayon et al., 2004; Caetano et al., 2009). A prime and most important example is the hydrogenous Fe/Mn-oxides precipitates dispersed within marine sediments in the form of coatings around detrital particles (Bayon et al., 2004; Prakash et al., 2012). The adsorption, complexation, reversible scavenging, partial reductive solution and oxidative precipitation processes must have shaped the REY record of this Fe/Mn-oxyhydroxide phase and regulated the abundance of dissolved REY in seawater (Hathorne et al., 2015; Leybourne and Johannesson, 2008; MarmolejoRodriguez et al., 2007). Lanthanide series fractionation patterns, Y/Ho divergence, and Ce/Eu anomalies can be used to quantify the relative contributions of these oceanic transformation activities. Normalizing to average shale values (e.g. North American shale composite, NASC), the typical seawater REE fractionation reveals preferential scavenging of light REEs (LREE, LaeNd) compared to heavy REEs (HREE, DyeLu, Y not included), which form more stable carbonate complexes and have stronger affinity for negatively charged sites of organic molecules (Hathorne et al., 2015). Middle REE (MREE, SmeTb) enrichment or MREE bulge in sub-oxic pore waters are largely attributed to the reductive dissolution of Fe-oxyhydroxides that may subsequently be taken up onto other minerals like authigenic carbonate fluorapatite (Haley et al., 2004; Zhang et al., 2016), while increased LREE retention with respect to HREE was performed in the transitional zones where oxyhydroxides generated (Caetano et al., 2009). Even similar ionic radius, the anomalously low affinity of Y for Fe-oxyhydroxides relative to Ho has been observed during the earliest stages of basalt weathering (Thompson et al., 2013) and made comparable reactive particle shuttle mechanism in the marine realm (Pack et al., 2007). Redox effects on REY fractionation can be effectively extrapolated from the redox sensitive Ce/Eu anomalies, which is defined as the difference ratio of the predicted Ce/Eu based on neighboring REYs to the measured Ce/Eu (Shields and Stille, 2001). Deviate from the relative conservative

dissolved 3+ oxidation state, Ce is exceptional for its ready oxidization to highly insoluble Ce(IV)O2 (Hathorne et al., 2015). A contrasting transformation of Eu3+ to Eu2+ also decouple Eu from its neighboring REE, but this might restricted in strongly reducing conditions and hydrothermal systems (Vazquez-Ortega et al., 2015). Other phases like carbonate or phosphorus have been proposed as possible important sink of REY in marine sediments (Hannigan and Sholkovitz, 2001; Rasmussen et al., 1998). Besides crystallographic effect of minerals (Prakash et al., 2012), organic matter is also responsible for different solid phase association of REYs and plays an important role in the earlydiagenetic reaction (Yang et al., 2017). To the extent that different oceanic biogeochemical transformations associated with sediment resuspension, including early diagenetic changes in sea-floor and highly selective resuspension of fine-particles (Gardner et al., 1985), might strengthen particle-seawater interaction and affect the signatures and/or concentrations of REY in sinking particles, we postulated that the resolving ability of REY concentration as potential resuspension proxy might shrink. Nevertheless, specific mobilization of REY may be strongly influenced by the above mechanisms, and provide new insights into the discrimination between re-suspended and settling particles. Upon the critical position of understanding material exchange through water column and settling particles (Akagi et al., 2011), we would explore the degree to which the REY characters in re-suspended particles is shaped by related physical, chemical and biological processes. We will probe how REY signatures change in marginal seas as moving from the complex of exchangeable cations, across a range of metal (oxy)hydroxide crystallinity, into oxidisable organic matter and residuum. Specifically, (1) to quantify the concentrations of REYs in settling particles received by sediment traps, surface sediments and core sediment samples; (2) the chemical speciation of REYs were compared and discussed under the resuspension framework; and (3) to relate aforementioned resuspension-related mechanisms and distributions of REYs within the different geochemical fractions. The Jiaozhou Bay (JZB) (35°58′–36°18′ N, 120°04′–120°23′ E), is a typical semi-enclosed bay situated on the west of the Yellow Sea and the southeast of the Shandong Peninsula (Fig. 1a). The water area of JZB is approximately 370 km2 with an average depth of 7 m, and the maximum water depth of 71 m is near the mouth of the bay (Song et al., 2016). The river system along JZB is more developed in the northern continental area. About 10 mountain-stream derived rain-source rivers, such as the Daguhe, Moshuihe, Baishahe, Licunhe, and Yanghe Rivers, 80

Marine Pollution Bulletin 144 (2019) 79–91

J. Liu, et al.

converged radially into the bay, but there are no big controlling river that greatly affects the sediment source and hydrological condition. With narrow entrance, the poor water-exchange between the inside and outside of the JZB may seriously constrict the diffusion of pollutants and the self-purification of the bay (Liang et al., 2018). The JZB has been strongly influenced by anthropogenic activities, for not only bearing a large number of industrial and agricultural/living solid wastes, sewage from land-sourced rivers and sewage outlets, but also accepting the input of atmospheric dry/wet dust-fall and the pollutants emitted by aquiculture. Three sample types were comprised in the whole sampling strategy (Fig. 1b): (1) Settling particles were collected at two mooring McLane® Mark 78H Time Series deployed stations — station TS1 with a water depth of 15 m was conducted in July 2016, and station TS2 with a water depth of 10.5 m was conducted in December 2016. The placement of both sediment traps were five meters above the sea floor. Setting a 24-h sampling time in advance for each bottle, we retrieved eight and seven samples from stations TS1 and TS2, respectively. (2) Six surface sediment (SS, numbered from S3 to S8) samples scattered over the entire JZB were gained during mid-August 2016 campaign using a Van Veen grab sampler. (3) One sediment core (CS5) of 20 cm length was obtained at the same location with station S5 using a box sampler during the SS sampling campaign. In the laboratory, all traps were screened for zooplankton ‘swimmers’ > 200 μm prior to filtering. Particulate concentrations of major components (Al2O3, Fe2O3, Mn) and REY (La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Y, Ho, Er, Tm, Yb, Lu) were determined in samples filtered onto pre-combusted HCl-washed glass fiber (GF/F) filters. The sediment core was sectioned at 2-cm intervals and a total of 10 subsamples were obtained. All surface and core sediment subsamples were stored in plastic bags and immediately deep-frozen and lyophilized. Dried aliquots were ground using an agate mortar and pestle for homogenization and prepared for analysis. Data from this study are reported on a dry weight basis. Bulk extracted concentrations for major components and REY were processed using a complete acid digestion protocol with 0.6 mL ultrapure HNO3 and 2 mL ultrapure HF in closed Teflon bombs at 150 °C for approximately 100 mg homogenized samples (Duan et al., 2012). Then 0.5 mL HClO4 was added and evaporated to dryness. The residue was treated with 1 mL HNO3 and 1 mL H2O and heated in the closed Teflon digestion vessel at 120 °C for 12 h, followed by diluting (1:5 v:v) using ultrapure water and transferred (with filter remnants) to acid-washed 50 mL low-density polyethylene bottles. Organic phosphorus (OP) was calculated as the difference between total P (TP) and inorganic P (IP), where TP and IP were determined based on HCl-extractable P (1 mol L−1 HCl, 24 h) of combusted (550 °C, 2 h) and non-combusted sediment, respectively (Aspila et al., 1976). Total carbon (TC) and OC measurements were made on lyophilized samples ground to 200 mesh. The untreated subsamples were analyzed for TC with a Perkin-Elmer model 2400 elemental analyzer. The concentrations of OC were determined in a similar way after treating another set of subsamples with HCl (4 mol L−1) several times until ebullition ceased in order to remove inorganic carbon prior to analysis. Concentrations of inorganic carbon (IC) were equated to the difference between TC and OC. To elucidate the overall impact of numerous alterations (especially mineral effect of Fe redox cycling) induced by resuspension in changing the records of REY chemical-signature, the modified sequential selective extraction procedure was operated to separate six operationally defined fractions (Poulton and Canfield, 2005; Rauret et al., 1999): exchangeable and carbonates fraction (F1), poorly crystalline Fe oxide fraction (F2), crystalline Fe oxides fraction (F3), magnetite fraction (F4), organic matter fraction (F5) and residual fraction (F6). Briefly, about 0.10 g of homogenized sediment/particle samples were weighed and loaded into a 50 mL centrifuge tube, and F1 fraction was extracted with 10 mL 0.11 mol L−1 acetic acid, shaken for 16 h at 22 °C. The slurry was then centrifuged at 4000g for 30 min to pelletize the sediment, and the supernatant was filtered (< 0.45 μm nylon

syringe filter) into 30 mL metal-free tubes and acidified with 2–3 drops of concentrated ultrapure nitric acid. F2 (e.g., ferrihydrite, lepidocrocite) were extracted with 1 mol L−1 hydroxylamine-hydrochloride in 25% v:v acetic acid for 48 h. F3 (e.g., goethite, hematite) were extracted with freshly-prepared citrate-buffered sodium dithionite (50 g L−1 in 0.35 mol L−1 acetic acid + 0.2 mol L−1 sodium citrate buffer solution at pH 4.8) for 2 h. Extraction of F4, which may also extract ilmenite and metals complexed with humic acid-like polymers (Siregar et al., 2004), was performed using 0.2 mol L−1 ammonium oxalate + 0.17 mol L−1 oxalic acid solution at pH 3.2 for 6 h. Finally, the oxidizable F5, which consists primarily of organic-bound fraction (strongly complexed by organics or present in biomass) but could include minor quantities of sulfide minerals, was extracted for ~2 h with periodic agitation at 85 °C with 30% hydrogen peroxide (pH 2 with HNO3), followed by continuous agitation for 16 h with 3.2 mol L−1 ammonium acetate in 20% (v:v) nitric acid at room temperature (Tessier et al., 1979). The residual fraction contained primarily silicate minerals and insoluble organic matter, and equated the difference between total concentrations and the sum of these five extract fractions. Between each step, samples were rinsed by vortexing the pellet with 10 ml of 0.01 mol L−1 potassium chloride solution, centrifuging to collect supernatant, and combining the filtered rinse solution with the extract. Extract solutions were stored at 4 °C until analysis by ICP-MS equipped with a dynamic reaction cell to eliminate mass interferences (Thermo, iCAP-Q) for major components and REY after dilution in trace-metal grade 2% HNO3. Extract solutions that were not reacted with samples were analyzed as method blanks to assess potential contamination from extract chemicals. The quality assurance and quality control were controlled by method blanks, field duplicate samples and standard reference materials. The relative percent difference for major components and REYs identified in paired duplicate samples was all < 10%. Blank samples were implemented throughout all the experiments. To explicitly evaluate the analytical precision, all samples were determined in triplicate. Precision, expressed as relative standard deviation, was better than 10%. Accuracy of total analysis was assured using the certified standard reference materials (GSD-9 and GSS-15), the recoveries of Al2O3, Fe2O3 and Mn were 99.34%, 101.23%, 91.03% in GSD-9 and 95.15%, 99.22%, 88.30% in GSS-15, respectively. Their analogues for REY were in the range of 95–105% (GSD-9) and 90–108% (GSS-15). To enable comparison across the full suite of sample types, REY concentration data for sediment and trapped particle samples were normalized by the NASC concentrations. Magnitude of cerium anomaly (Ce/Ce∗), europium anomaly (Eu/Eu∗), Y/Ho ratio, MREE bulge index (MREE/MREE∗) and the HREE to LREE ratio (HREE/LREE) was calculated as

Ce/Ce =

CeNASC LaNASC × PrNASC

Eu/Eu =

EuNASC SmNASC × GdNASC

Y/Ho = YNASC ÷ HoNASC

MREE/MREE =

2 × average(MREENASC ) average(LREENASC) + average(HREENASC )

average(HREENASC ) HREE = LREE average(LREENASC ) where REYNASC corresponds to REY concentration normalized by the NASC value. Ratios significantly greater than (less than) 1.0 represent positive (negative) anomalies and/or depletion (enrichment) with respect to the referenced levels. The calculation of mean values, the standard error of a mean value (E), minimum/maximum values, Pearson correlation coefficients and multiple comparison tests between TS1, TS2, SS and core sediments CS5 were executed with MATLAB® R2014a. Unless otherwise indicated, for all statistical analyses the 81

Marine Pollution Bulletin 144 (2019) 79–91

J. Liu, et al.

Table 1 Concentrations of rare earth elements and other chemical parameters for trapped settling particles and surface/core sediments ( ± standard deviations of the mean values, letters represent significant differences at p < 0.05). Parameters

TS1

TS2

SS

CS5

La (μg g−1) Ce (μg g−1) Pr (μg g−1) Nd (μg g−1) Sm (μg g−1) Eu (μg g−1) Gd (μg g−1) Tb (μg g−1) Dy (μg g−1) Y (μg g−1) Ho (μg g−1) Er (μg g−1) Tm (μg g−1) Yb (μg g−1) Lu (μg g−1) ∑REE (μg g−1) Al2O3 (mg g−1) Fe2O3 (mg g−1) Mn (μg g−1) IP (μg g−1) OP (μg g−1) OC (mg g−1) IC (mg g−1)

48.8 ± 1.37 ab 92.09 ± 2.53 a 9.98 ± 0.26 a 34.53 ± 0.99 a 6.39 ± 0.18 a 1.24 ± 0.04 a 5.21 ± 0.15 a 0.74 ± 0.02 a 4.13 ± 0.1 a 23.18 ± 0.67 a 0.8 ± 0.02 a 2.29 ± 0.06 a 0.36 ± 0.01 a 2.14 ± 0.06 a 0.34 ± 0.01 a 232.22 ± 6.35 a 147.8 ± 3.6 a 60.1 ± 2.2 a 1620.1 ± 66.5 a 545.99 ± 12.93 a 289.19 ± 13.28 a 6.32 ± 0.14 a 4.61 ± 0.17 a

53.63 ± 2.51 a 88.82 ± 3.47 a 9.61 ± 0.35 a 33.54 ± 1.28 a 6.15 ± 0.22 a 1.22 ± 0.04 a 5.1 ± 0.18 a 0.73 ± 0.03 a 4.05 ± 0.15 a 22.79 ± 0.9 a 0.79 ± 0.03 a 2.25 ± 0.09 a 0.34 ± 0.01 a 2.09 ± 0.08 a 0.33 ± 0.01 a 231.44 ± 9.27 a 141.9 ± 3.7 a 58.4 ± 1.5 a 1300.5 ± 39.9 b 530.37 ± 10.13 a 240.29 ± 6.94 ab 6.07 ± 0.11 a 3.27 ± 0.19 b

34.24 ± 2.93 c 66.17 ± 5.32 b 7.28 ± 0.58 b 25.46 ± 2.04 b 4.79 ± 0.36 b 0.96 ± 0.05 b 3.89 ± 0.29 b 0.56 ± 0.04 b 3.09 ± 0.27 b 17.18 ± 1.57 b 0.59 ± 0.06 b 1.71 ± 0.18 b 0.26 ± 0.03 b 1.59 ± 0.18 b 0.25 ± 0.03 b 168.03 ± 13.82 b 131.4 ± 4.7 a 39 ± 4.2 b 655.5 ± 27.3 c 550.22 ± 39.37 a 137.3 ± 24.59 c 4.19 ± 0.66 b 1.32 ± 0.22 c

45.24 ± 2.14 b 88.18 ± 4.21 a 9.59 ± 0.48 a 33.57 ± 1.62 a 6.13 ± 0.29 a 1.2 ± 0.06 a 5.03 ± 0.23 a 0.71 ± 0.03 a 4.03 ± 0.18 a 22.66 ± 1.12 a 0.78 ± 0.04 a 2.24 ± 0.1 a 0.34 ± 0.02 a 2.07 ± 0.1 a 0.34 ± 0.02 a 222.1 ± 10.57 a 144.6 ± 6.6 a 54.7 ± 2 a 718.48 ± 23.12 c 599.13 ± 15.72 ab 194.42 ± 6.07 b 6.4 ± 0.11 a 1.9 ± 0.15 c

significance level applied was α = 0.05. Table 1 lists the mean concentrations and standard errors for REY and other chemical parameters in settling particles and surface/core sediments. For the total REE concentrations (∑REE), different sample groups showed the following means and standard deviations of meanvalue: 232.22 ± 6.35 and 231.44 ± 9.27 μg g−1 for particles in the TS1 and TS2 station, 168.03 ± 13.82 and 222.1 ± 10.57 μg g−1 for the surface and core sediments, respectively. The ∑REE concentrations of sinking particles in this study (Table 1) are consistent with the ∑REE concentration range of ~200 to 268 μg g−1 in the fine-grained fraction (< 20 μm) of major river sediments around the Yellow Sea (Song and Choi, 2009), including the Yangtze and Yellow River in China, the Han, Geum and Yeongsan River in Korean. The extremely low ∑REE concentrations of bulk settling particles collected monthly using sediment trap set at the highly productive North Pacific Ocean were found to oscillate from 3.03 to 22.6 μg g−1, where biogenic silica, carbonates and organic carbon can be as high as 75–91% of the total mass flux (Akagi et al., 2011). It is clearly apparent from the above comparison and Table 1 that the continental input as river-transported fine-grained particles occupy a dominant position in the sediment trap collected samples. The ∑REE concentrations in sediments observed in JZB are in agreement with that in Bohai Bay surface sediment (averaging at 184.86 μg g−1 (Xu et al., 2012)), which have certain geographic sourcerock overlap and comparable contamination status. Moreover, the ∑REE concentrations of SS and CS5 were higher than those of the Yellow Sea sediment (123.01 μg g−1), the East China Sea sediment (109.03 μg g−1) and pelagic sediment (125.93 μg g−1) (Xu et al., 2010). For individual REY element, the similar uniform systematic variation pattern was discovered among those sampling subsets/groups: the mean REY concentrations of TS1, TS2 and CS5 generally exhibited no significant difference with the exception of La, and are significantly larger than that of SS (Table 1). Comparing with the major components, only Fe2O3 had the REY-related distribution. Although the concentrations of Al2O3 presented parallel numerical changes, they did not significantly different in statistical meaning among sediments and particles (Table 1). This common characteristic in resuspension framework is certainly in part due to the grain-size effect, as already observed for trace metal elements, in which higher REY adsorptive retention associated with Fe (oxy)hydroxides and layer aluminosilicate clay minerals than coarser minerals, such as quartz and feldspar, is predictable in

finer settling and resuspended particles (Liu et al., 2016; Liu et al., 2018). Therefore, the non-significant difference between Al normalization levels of most REYs reinforced this inference (Table S1). The order of magnitude of the REY concentrations (Table 1), Ce > La > Nd > Y > Pr > Sm > Gd > Dy > Er > Yb > Eu > Ho > Tb > Tm > Lu, generally followed the tendency of decreasing abundance with increasing atomic number and the Oddo–Harkins rule, which indicated more frequent for even atomic numbers than adjacently larger and smaller odd atomic numbers (Mihajlovic et al., 2014). This sequence is within the range for some classical REY references as described by Laveuf and Cornu (2009). To eliminate the abnormal distribution of natural abundance and acquire better tracing application, the NASC-normalized REY pattern was commonly used for distinguishing subtle fractionations and anomalies in the REY coherent group, where Y is included to the left of Ho, consistent with its ionic radius (Leybourne and Johannesson, 2008). The REY pattern in our samples was displayed in Fig. 2, which should reflect both the physical/mechanical dispersion of catchment host lithologies, the hydromorphic components transport from aqueous phase (soil solution, steam and sea-water), biological recycling and organic matter impact (Akagi et al., 2011; Laveuf and Cornu, 2009). A general declining trend with increasing atomic number, also the zigzag fluctuations around Dy was observed through the whole database (Fig. 2), but the latter feature was demonstrated as an artificial misinterpret because of the Upper Continental Crust (UCC) normalized REY pattern smoothing this part (data not shown). Comparing few previous REY investigations in settling particles, they reported a similar compositional pattern: an increase with increasing atomic number in the LREE region and a slight kink at Eu or Gd (Akagi et al., 2011; Lerche and Nozaki, 1998), which was quite controversy with our study. Nevertheless, it is common phenomenon that the HREE part experiences preferential leaching during weathering (Mihajlovic et al., 2014). More HREE depletion than LREE relative to NASC might be due to the easier soluble HREE-complexes, as the energy of desolvation is a function of the atomic number. Comparatively, LREE are more susceptible to occur as free species (Vazquez-Ortega et al., 2016). Therefore, the mixing weathering processes in supergene conditions would bring more HREE with solution migration, leaving the enrichment of LREEs in detritus (Vazquez-Ortega et al., 2015). Again, the unified normalization REY pattern in sinking particles and sediments, and its 82

Marine Pollution Bulletin 144 (2019) 79–91

J. Liu, et al.

Fig. 2. Total REY fractionation patterns for the settling particles and sediments.

character, confirmed the priority contribution of catchment detrital materials. The mean relative contribution of various REY species in sequential leachates is presented in Fig. 3, and the geochemical fractions of the REYs differ slightly inside and outside of each sample group. The majority of the REY, over 50%, have been found in residual fraction (F6). Approximately one-third is basically bound to reducible amorphous Fe oxides (F2), indicating the main host phase for labile REY pool. The water soluble, exchangeable and carbonates fraction (F1), organic matter and sulphides fraction (oxidisable fraction, F5), together with other two crystalline iron minerals fraction (F3 and F4) only take a minority percentage, which is 2.07–3.29%, 0.59–4.75%, 1.06–4.30% and 2.44–5.93%, respectively. Vazquez-Ortega et al. (2016) also reported more Short Range Order Fe-oxides (e.g. ferrihydrite) than Long Range Order Fe-oxides (e.g. goethite) in hillslope pedons, but they had

the larger abundant fraction for Organo-metal colloids. This result is not surprising given that their organic matter concentration is much higher. Despite the extraction steps differ to make the direct comparison problematic, particularly for methods excluding specific reagents for differentiation of Fe-bearing minerals, the consensus for most oceanic sediments and soils can be highlighted: the poorly crystallized Fe-oxides seems to play a crucial role for REY migration and transformation (Bayon et al., 2004; Caetano et al., 2009; Mihajlovic et al., 2014; Prakash et al., 2012; Xu et al., 2012; Yan et al., 1999). Regarding REY's compartment-specific partitioning signatures, several possible mechanisms should be taken into account: (1) Provenance information induced by weathering of primary minerals; (2) organic and inorganic complexation (e.g. sulfate, carbonate, and hydroxide ligands); and (3) adsorption and co-precipitation with secondary minerals (Laveuf and Cornu, 2009; Tang and Johannesson, 2010).

Fig. 3. The average percentages of REYs and Fe/Mn in the six defined fractions. (F1 exchangeable and carbonates fraction, F2 poorly crystalline Fe oxide fraction, F3 crystalline Fe oxides fraction, F4 magnetite fraction, F5 organic matter fraction, and F6 residual fraction). 83

Marine Pollution Bulletin 144 (2019) 79–91

J. Liu, et al.

Fig. 4. a. REY fractionation patterns for the F1, F2 and F3 fractions released during the sequential extraction in the settling particles and sediments, error bars (standard deviations of mean values) were calculated for each sample groups. b. REY fractionation patterns for the F4, F5 and F6 fractions released during the sequential extraction in the settling particles and sediments, error bars (standard deviations of mean values) were calculated for each sample group.

84

Marine Pollution Bulletin 144 (2019) 79–91

J. Liu, et al.

greater ∑REE/Fe weight ratios of 7.48–33.47 × 10−3 in F2 than 0.29–3.45 × 10−3 of F3 in this study. Moreover, the smaller PZC of Feoxides (e.g. goethite, lepidocrocite and hematite; around 5.3–6.5) than amorphous Fe minerals (e.g. ferrihydrite; around 5.3–8.8) can render F2 more susceptible to have negative residual charges in marine pH situations (Appel et al., 2003), and help to account for the higher REY capacity. To straightened out the differential fractionation between LREEs, MREEs and HREEs in F2-F3-F4, we need to partition the FeeMn oxyhydroxide leachates within our studied JZB particles and sediments. Their component may derive from (1) the direct delivery of detrital ‘preformed’ FeeMn oxides (i.e. riverine input and aeolian dust fall, the former one is controlling) and/or (2) authigenic precipitation from local seawater (Bayon et al., 2004). After reductively dissolve, Fe2+ and Mn2+ are released and re-crystallized under anaerobic conditions, be capable of recording a hydrogenous (seawater-sourced) REY signature. Thus the REY pattern of neoprecipitated FeeMn oxyhydroxides experienced early diagenesis is anticipated to depart from the original ‘preformed’ pattern (Caetano et al., 2009). However, the ‘authigenic’ end member represented by FeeMn coatings on recent Cape Basin forams, displays a light-REE depletion relative to MREE and HREE (Bayon et al., 2004), wiped off the possibility of consideration REY amount from neoprecipitated minerals. The increased retention of LREE in amorphous Fe-oxyhydroxides has been reported in previous researches (Caetano et al., 2009; Lecomte et al., 2017; Yan et al., 1999), and can be implied using correlations between acid-reductive extracted individual REY and Fe in this study (Table 2). Significantly strong correlations to insignificant week correlations were found from LREE (r = 0.601–0.912) to HREE (r = 0.3–0.507) through MREE (r = 0.455–527) in F2 fraction (Table 2). The more stable Fe-Mn-oxides in F3 fraction might intrinsically presented MREE enrichment under the influences of Mn-oxyhydroxides (Abbott et al., 2015; Prakash et al., 2012; Song and Choi, 2009). Because Fe in F3 were relative trivial and less active when comparing F2, the respective opposite trend of Mnoxyhydroxides and Fe-oxyhydroxides in REY adsorption might both play an important role and trigger MREE enriching, unlike the dominant expression of Fe-oxides in F2. The Mn/Fe molar ratios in F2 fraction (i.e. 288.7 ± 18.8–549.6 ± 22.8 × 10−4 mol:mol) were one order of magnitude greater than their analogues in F3 fractions (i.e. 34.83 ± 0.24–39.35 ± 0.67 × 10−4 mol:mol), might not consistent with this reference (Fig. 3). Thus we attribute the both performance of Mn and Fe oxyhydroxides in F3 fractions to the substantially less REY retaining ability for crystalline Fe oxides than Mn oxides. The significantly positive correlation coefficients between Mn and REYs in F3 (ranged from 0.822 to 0.902), together with the less analogues in F2 (ranged from 0.138 to 0.778) also witness Mn-oxyhydroxides' effect (Table 2). The last F4-magnetite fraction is quite complex for their partial biological formation mechanism (Cooper et al., 2000; Konhauser Kurt, 2006). The correlations among OP, OC with LREEs in F4 were greater than their analogues F2 and F3, for example, r = 0.800 for OP versus La and r = 0.883 for OC versus La in F4, r = 0.767 for OP versus La and r = 0.665 for OC versus La in F3, r = 0.738 for OP versus La and r = 0.610 for OC versus La in F2 (Table 2). The implication being that the increased affinity of LREE is the reason for more HREE depletion in F4 than F2. Another special attention should be paid is the phosphorus. The strong affinity of REYs to phosphate compounds has been evaluated (Hathorne et al., 2015; Mihajlovic et al., 2014; Vazquez-Ortega et al., 2016; Yang et al., 2017), and they typically contain thousands of μg g−1 of REYs. The significant correlation among F2-IP and REYs, particularly HREE and MREE (Table 2), consistent with phosphate mineral enrichment in MREE (Vazquez-Ortega et al., 2015). But the specific role of phosphate needs further verification to explore the connection between various P speciation and REYs.

1. F1 The NASC-normalized patterns showed slightly MREE enrichment in all SS, also MREE and LREE enrichment than HREE for CS5, but the convex shapes did not recurring for TS1 and TS2 (Fig. 4a). There are two possible explanations for this difference. One is that the vertical fluid REY profile has been modelled to influence adsorption processes. Typical seawater REY depth profile is featured by the increasing REY concentrations with water depth, accompanied by the HREE enrichment (Abbott et al., 2015; Hathorne et al., 2015). Thus, in like manner, irreversible scavenging mechanism for particle reactive LREE in which adsorption onto particles over balanced with desorption could be functioned in this study (Oka et al., 2009), causing the REY adsorbed to the sinking particles an opposite distribution with the oceanic HREE enhancement. Likewise, the apparent partition coefficients between particulate and dissolved REYs were found to be higher for the LREE than the HREE (Garcia-Solsona et al., 2014). For anoxic pore-fluids interacted with surficial sediments, REY patterns tend to be characterized by a MREE bulge pattern in the shallow ferruginous zone (Haley et al., 2004; Kim et al., 2012). The electrostatic attraction on exchange sites on surfaces and interfaces of negatively charged inorganic and organic complexes of stagnant sediments would undoubtedly inherited its environmental attributes (Yan et al., 1999). Another aspect is the REYs incorporated into the CaCO3 crystals. Because the ionic radius of lanthanide's trivalent ions, ranging between 0.861 Å for Lu3+ and 1.03 Å for La3+, is similar to that of Ca2+, REYs can easily replace Ca into carbonate mineral lattice, notably through isomorphic exchange (Polyakov and Nearing, 2004). The inorganic parent carbonates are considered to have a low ∑REE with LREE enrichment, but most of them are dissolved during pedogenesis, and the REE‑carbonate complexes precipitated minerals are expected to be richer in HREE than LREE for bicarbonates more inclined to complex HREEs than LREEs with increasing pH (Akagi, 2013; Compton et al., 2003). And the biogenic carbonate may display a slight HREE enrichment with respect to seawater (Akagi et al., 2004; Zhang et al., 2016). The average IC concentration was 4.61 ± 0.17 mg g−1 for TS1 and 3.27 ± 0.19 mg g−1 for TS2, significantly greater than that of 1.32 ± 0.22 mg g−1 for SS, and that of 1.9 ± 0.15 mg g−1 for CS5 (Table 1), which is possibly caused by the autochthonous calcareous shell corrosion at sea-floor (Rutgers van der Loeff and Boudreau, 1997). This can lead to more exposure of unprotected marine OC to microbial decomposition and zooplanktonic feeding, and a greater terrigenous/riverine fraction of IC can be inferred for SS/CS5, when comparing TS1/TS2 (Liu et al., 2019). Therefore, the higher HREE/LREE ratios in SS/CS5 (mean HREE/LREE ratio = 0.67 and 0.59 for SS and CS5) than TS1/TS2 (mean HREE/LREE ratio = 0.49 and 0.39 for TS1 and TS2) might result from these higher inorganic riverine carbonate fractions in sediments. 2. F2–F3–F4 Since Mn was usually registered with Fe for the similar pathways of biogeochemistry in aquatic sediments (Liu et al., 2018), we included this element in our discuss. Extracts targeting amorphous Fe hydroxides (F2) were depleted in HREE (HREE/LREE ranged from 0.36 to 0.46) (Fig. 4a). The small fraction of REY associated with F3 dominantly presented ‘MREE bulge’ or ‘hat-shaped’ distribution (MREE/MREE* ranged from 1.24 to 1.47), but the equally scarce F4 fraction showed more HREE depletion than F2 (HREE/LREE ranged from 0.26 to 0.38) (Fig. 4a and b). With one order of magnitude higher for F2 than F3 and F4, younger, amorphous Fe is suggested to responsible for retaining the REYs. As the ionic radii differences between REEs and Fe and Mn determined the incorporation of REYs into Fe-Mn-oxide structure constrained, adsorption became the main mechanism for REYs scavenging onto Fe-Mn-oxides surface (Braun et al., 1993). Therefore, the irregular structure in F2 might cause REYs to be more easily embedded compared to the regular structure of the older Fe-Mn-oxides, manifested by the 85

Marine Pollution Bulletin 144 (2019) 79–91

J. Liu, et al.

Table 2 Coefficients of determination (r) for bivariate linear regressions of REYs against Mn, Fe2O3, IP, OP and OC in different fractions. REY

F2

F3

Mn La Ce Pr Nd Sm Eu Gd Tb Dy Y Ho Er Tm Yb Lu ⁎ ⁎⁎

0.778 0.549⁎⁎ 0.512⁎⁎ 0.508⁎⁎ 0.431⁎ 0.409⁎ 0.432⁎ 0.355 0.307 0.428⁎ 0.267 0.259 0.138 0.177 0.168 ⁎⁎

Fe2O3

IP

OP

0.912 0.635⁎⁎ 0.607⁎⁎ 0.601⁎⁎ 0.527⁎⁎ 0.502⁎⁎ 0.522⁎⁎ 0.455⁎ 0.404⁎ 0.507⁎⁎ 0.365⁎ 0.360 0.262 0.305 0.300

−0.031 0.307 0.372⁎ 0.379⁎ 0.454⁎ 0.473⁎⁎ 0.437⁎ 0.495⁎⁎ 0.521⁎⁎ 0.461⁎ 0.563⁎⁎ 0.573⁎⁎ 0.601⁎⁎ 0.611⁎⁎ 0.599⁎⁎

0.738 0.835⁎⁎ 0.807⁎⁎ 0.807⁎⁎ 0.743⁎⁎ 0.731⁎⁎ 0.756⁎⁎ 0.702⁎⁎ 0.655⁎⁎ 0.756⁎⁎ 0.635⁎⁎ 0.630⁎⁎ 0.514⁎⁎ 0.532⁎⁎ 0.522⁎⁎

⁎⁎

OC ⁎⁎

F4

Mn

0.610 0.891⁎⁎ 0.894⁎⁎ 0.888⁎⁎ 0.885⁎⁎ 0.873⁎⁎ 0.893⁎⁎ 0.875⁎⁎ 0.863⁎⁎ 0.867⁎⁎ 0.854⁎⁎ 0.855⁎⁎ 0.824⁎⁎ 0.819⁎⁎ 0.821⁎⁎ ⁎⁎

0.822 0.879⁎⁎ 0.877⁎⁎ 0.901⁎⁎ 0.897⁎⁎ 0.878⁎⁎ 0.891⁎⁎ 0.880⁎⁎ 0.901⁎⁎ 0.900⁎⁎ 0.899⁎⁎ 0.902⁎⁎ 0.880⁎⁎ 0.902⁎⁎ 0.870⁎⁎ ⁎⁎

Fe2O3

OP

0.727 0.824⁎⁎ 0.825⁎⁎ 0.858⁎⁎ 0.842⁎⁎ 0.817⁎⁎ 0.817⁎⁎ 0.844⁎⁎ 0.836⁎⁎ 0.852⁎⁎ 0.863⁎⁎ 0.834⁎⁎ 0.782⁎⁎ 0.840⁎⁎ 0.802⁎⁎

0.767 0.822⁎⁎ 0.825⁎⁎ 0.843⁎⁎ 0.836⁎⁎ 0.831⁎⁎ 0.853⁎⁎ 0.815⁎⁎ 0.832⁎⁎ 0.822⁎⁎ 0.827⁎⁎ 0.846⁎⁎ 0.880⁎⁎ 0.841⁎⁎ 0.821⁎⁎

⁎⁎

OC ⁎⁎

F5

Mn

0.665 0.729⁎⁎ 0.724⁎⁎ 0.771⁎⁎ 0.739⁎⁎ 0.727⁎⁎ 0.687⁎⁎ 0.777⁎⁎ 0.703⁎⁎ 0.721⁎⁎ 0.776⁎⁎ 0.702⁎⁎ 0.638⁎⁎ 0.726⁎⁎ 0.690⁎⁎ ⁎⁎

0.896 0.882⁎⁎ 0.892⁎⁎ 0.898⁎⁎ 0.886⁎⁎ 0.863⁎⁎ 0.875⁎⁎ 0.824⁎⁎ 0.799⁎⁎ 0.854⁎⁎ 0.802⁎⁎ 0.833⁎⁎ 0.848⁎⁎ 0.813⁎⁎ 0.768⁎⁎ ⁎⁎

Fe2O3

OP

0.805 0.846⁎⁎ 0.803⁎⁎ 0.817⁎⁎ 0.858⁎⁎ 0.866⁎⁎ 0.858⁎⁎ 0.835⁎⁎ 0.827⁎⁎ 0.825⁎⁎ 0.837⁎⁎ 0.838⁎⁎ 0.846⁎⁎ 0.857⁎⁎ 0.823⁎⁎

0.800 0.811⁎⁎ 0.814⁎⁎ 0.816⁎⁎ 0.813⁎⁎ 0.791⁎⁎ 0.799⁎⁎ 0.765⁎⁎ 0.747⁎⁎ 0.801⁎⁎ 0.757⁎⁎ 0.772⁎⁎ 0.783⁎⁎ 0.697⁎⁎ 0.685⁎⁎

⁎⁎

OC ⁎⁎

OC

0.883 0.839⁎⁎ 0.879⁎⁎ 0.867⁎⁎ 0.819⁎⁎ 0.789⁎⁎ 0.786⁎⁎ 0.750⁎⁎ 0.716⁎⁎ 0.792⁎⁎ 0.718⁎⁎ 0.727⁎⁎ 0.760⁎⁎ 0.741⁎⁎ 0.684⁎⁎ ⁎⁎

0.197 0.016 0.177 0.194 0.221 0.248 0.302 0.354 0.365 0.483⁎⁎ 0.382⁎ 0.422⁎ 0.498⁎⁎ 0.447⁎ 0.509⁎⁎

Correlation is significant at the 0.01 level (two-tailed). Correlation is significant at the 0.05 level (two-tailed).

Ce/Ce* (> 1) was discovered to occur in F1, F2, and F3, while significantly fewer negative Ce/Ce* (< 1) was happened for F4 and F5 (Fig. 5). As cerium fractionation is strongly influenced by oxidation-reduction reactions (Laveuf and Cornu, 2009), the greater Ce mass fraction relative to ambient La and Pr in reducible oxides than oxidizable organic/sulphidic matter is feasible. The Fe-Mn-oxides have been previously investigated to possess positive Ce anomalies from the oxidation of Ce3+ under oxic conditions through abiotic processes (Thibault de Chanvalon et al., 2016). It occurred at Eh values of around 0.3 V for circumneutral pH (Laveuf and Cornu, 2009), induced cerianite formation and promoted its solid phase incorporation (Bau and Koschinsky, 2009; Bau et al., 1996). More specifically, the distribution coefficients (Kd) for Ce reported by Ohta and Kawabe (2001) point out higher Ce affinity of Mn-oxides than that for Fe-(oxy)hydroxides. However, the Ce/Ce* displayed a decreasing sequence of CS5 (1.201 ± 0.003) > SS (1.167 ± 0.013) > TS1 (1.155 ± 0.004) > TS2 (1.019 ± 0.006) in F2 fraction, distinct with Mn/Fe molar ratios in the same partial extracts, which was TS1 (549.6 ± 22.8 × 10−4 mol:mol) > TS2 (484.8 ± 48.4 × 10−4 mol:mol) > CS5 (328.7 ± 6.2 × 10−4 mol:mol) > SS (288.7 ± 18.8 × 10−4 mol:mol). Once more, the dissimilar sequence of CS5 (1.169 ± 0.003) > SS (1.122 ± 0.023) > TS1 (1.192 ± 0.01) > TS2 (1.061 ± 0.007) for Ce/Ce* and TS2 (39.35 ± 0.67 × 10−4 mol:mol) > TS1 (38.42 ± 0.19 × 10−4 mol:mol) > SS (37.77 ± 2.51 × 10−4 mol:mol) > CS5 (34.83 ± 0.24 b × 10−4 mol:mol) for Mn/Fe molar ratios were presented in F3 fraction. Hence, the much higher abundance of Fe in the non-refractory/labile fractions could be responsible for more effective Ce sequestration into Fe-(oxy) hydroxides simply because of their greater abundance of reactive interfacial sites (Vazquez-Ortega et al., 2016). Similarly, the positive Ce anomalies in Fe-oxides were also observed in marine ferromanganese crusts (Bau and Koschinsky, 2009). The suggestion being that oxidative scavenging of Ce is not restricted to Mn(IV)-oxides, but also includes Fe (III)-oxides in marine systems (Hathorne et al., 2015). Besides Fe-Mnoxides, other phase (e.g., organic matter) is expected to be a stronger scavenger to capture the composition of circumambient aqueous waters and presumably oxidized Ce (Leybourne and Johannesson, 2008). But this conclusion is inferred by the strong bound between Ce3+ and Ce4+ and organic ligands in DOC-rich waters, there is no direct evidence for the remanent sedimentary organic matter. The significantly negative F5Ce/Ce* is not suitable for the above situation, and might result from the Ce(IV)O2 transformed into soluble Ce3+ in reducing condition. Therefore, future investigations should involve closer examination of the REYs in the dissolved, colloidal, and sedimentary organic matter fractions in these samples.

3. F5 The HREE enriched patterns were observed in organic F5 fraction (Fig. 4b), consistent with preferential complexation of HREEs after REE release from organic matter degradation (Haley et al., 2004). The higher HREE/LREE ratios in CS5 and SS (averaging at 3.02 and 2.41, respectively) than TS1 and TS2 (averaging at 1.55 and 1.56, respectively) suggested greater HREE enhancement for old, remnant organic matter in sediments, also implying higher freshly-formed biogenic particle content in sinking materials. Additionally, the significant stronger correlations between HREE and OC than their analogues among LREE/MREE and OC further confirmed the above viewpoints (Table 2). The underneath mechanization was that the capacity to complex, adsorb or chelate positively charged trivalent REYs, originated from negative charged carboxyl and phenolic hydroxyl groups in organic matter, usually favor REE of intermediate and heavy mass (i.e. MREE and HREE), because of their smaller ionic radii (Abbott et al., 2015; Vazquez-Ortega et al., 2016). 4. F6 The normalization level of the residual F6 fraction decreased from La to Lu in both sediments and settling particles (Fig. 4b). The indication being that HREEs can easily form soluble complexes e.g. with carbonate, dissolved organic matter, and tend to be more mobile than LREEs (Xu et al., 2012). Thus the HREE's higher potential availability than LREE lead them to be more preferentially leached. On the other hand, Taylor and McLennan (1985) mentioned that as LREEs have a larger effective ion radius than HREEs they can tend to be less incorporated in silicate mineral structures. Furthermore, the sigmoidal appearance was displayed in F6, which can be divided into four contiguous curved sections with more obvious shapes for LaeNd and YeLu interval (Fig. 4b). Prakash et al. (2012) and Ohta and Kawabe (2000) have gave the compelling evidence of tetrad effect associated with REE‑carbonate complexation constants also existed in the distribution coefficients between ferromanganese-oxide and seawater, so we can reasonably accommodate this mechanism when considering REY's liberation. The tetrad effect that represented quarter, half, three-quarter and completely filled 4f orbitals of elements over the lanthanide series had produced the remnant residual fraction an analogue appearance. Notable positive/negative Ce and Eu anomalies, which were not shown in the total composition, occurred in different fractions regardless of sample group (Fig. 5). Comparing with the analogical Ce/Ce* in total concentration and residual F6 fraction, the moderately higher positive 86

Marine Pollution Bulletin 144 (2019) 79–91

J. Liu, et al.

Fig. 5. The Ce and Eu anomalies in the total and six defined fractions of settling particles and sediments, error bars (standard deviations of mean values) were calculated for total concentration and each fraction within sample groups. Alphabetical letters before the hyphen represent significant differences (p < 0.05) between different sample groups in the fixed fraction, while latter ones denote significant differences (p < 0.05) between different fractions in the fixed group.

di‑carbonate species Ln(CO3)2− dominated the inorganic carbonate complexes for HREE (including Ho), while positively charged monocarbonate (YCO3+) prevailed for LREE (and Y), even these two element having nearly identical ionic radii. Thus the higher particle reactivity of Ho on positively charged Fe-oxyhydroxide surfaces, compared to yttrium, can be more scavenged, leading to greater negative Y/Ho anomalies (Luo and Byrne, 2004; Prakash et al., 2012). Regarding the specific sediment column, free Fe2+, released from dissimilated reduction reactions under the deep anaerobic condition, is upward transported to the shallow more oxidation environment to be recombined and form low crystalline or amorphous iron (hydrogen) oxides, and the REYs are co-precipitated or adsorbed from the pore water. The significantly lower F2 than F3 in SS, together with the sudden decreased F2-Y/Ho in shallow depth profile of CS5 and the overall declining trend of F2-Y/Ho in deep section (Fig. 7) would reflect this mechanism. Comparatively, the fluctuated F3-Y/Ho around its analogues in sinking particles which proved irrelevant or little relationship between diagenetic reactions and Y/Ho fractionation in old crystalline Fe minerals, further enhanced our explanation (Fig. 7). Thompson et al. (2013) also proposed that organic matter could exert a suppressing effect on Y/Ho fractionation. The dramatic monotonous reduction from 1.37 to nearly unit was targeted in F5-Y/Ho depth profile (Fig. 7), suggesting that organic matter mineralization would enhance the suppressing effect. The divergences of REY pattern indices, including cerium anomaly against the yttrium to holmium ratio, middle REE bulge index versus the HREE to LREE ratio, between settling particles and sediments were drawn in Fig. 8. For total composition, the common overlapping were occurred for both pair-to-pair contrast graphs (Fig. 8), consistent with the above discuss for the principal role of allochthonous particles incorporated into both surface sinking autochthonous and bottom sedimentary end-members in coastal environments. This conclusion was also expressed on the depth profile of Ce/Ce* and Eu/Eu* for CS5, whose vacillation generally surrounded the mean value of TS1 and TS2 (Fig. S1 and S2), also the indistinguishable distribution for the REY pattern indices in F3, F4, F5 and F6 fractions (Fig. S3 and S4). Such situation constraint the resolving ability of REY in resuspended and the first settling materials. Nevertheless, hope came from the most

Because the reduction of Eu requires strong reducing conditions rarely encountered in supergene conditions (Laveuf and Cornu, 2009), the relative quantitative changes of Eu/Eu* is not as much as Ce/Ce* (Fig. 5). The most pronounced positive Eu anomaly appeared in F6 residual phase (Fig. 5), reflecting the relative abundance of plagioclase feldspar for its more stable complexes with Eu2+ rather than trivalent lanthanides (Möller and Muecke, 1984). Furthermore, associating the Eu anomaly of the other partial extracts to that of the corresponding residual F6 fraction analysis in our study provides a means for evaluating the relative mobility of Eu in the local environment (Leybourne and Johannesson, 2008). The less Eu/Eu* from F1 to F5 than F6 (Fig. 5) reveals that Eu was either less mobile than Sm and Gd during hydromorphic transport, or that Eu has been preferentially sequestered into a newly formed minerals and/or phase than Sm and Gd that could not be removed by the sequential leaching solution. The divergent Y/Ho fractionation behavior due to preferential loss of yttrium with the lowest Misono softness in basalt weathering profiles has been recently shown. This process is caused by an empty 4f electron structure in Y3+, decreasing its polarizability and diminishing its tendency to form covalent metal-ligand bonds with organic acids and secondary minerals (Thompson et al., 2013). So the exclusion of Y (relative to Ho) during precipitation of secondary Fe-(oxy)hydroxides in soils has been previously reported (Vazquez-Ortega et al., 2016). Fig. 6 revealed the Y/Ho ratios in the total and six defined fractions of settling particles and sediments for near-shore environment, but the absolute value of > 1 indicated Y enrichment relative to Ho. We account this phenomenon to the controlling behavior of refractory detritus, similar to the mechanism performed in REY pattern for various Fe minerals. Nevertheless, the inner-changes of Y/Ho can still provide some information about neoformed hydrodynamic activities. The decreasing Y/ Ho sequence of F2 > F3 > F4 in TS1 and TS2 was observed (Fig. 6), as solid-state transformation and ripening occurs in the Fe-oxides, REY are expelled from the structure and subjected to increasing depletion, and such an effect therefore appears escalated for Y (Laveuf and Cornu, 2009; Thibault de Chanvalon et al., 2016). On the contrary, this situation did not applied for SS and CS5 which have lower and parallel F2 level than F3 (Fig. 6). The implication is that early diagenetic changes in the sediment might trigger such alteration. In seawater (pH = 7.9), 87

Marine Pollution Bulletin 144 (2019) 79–91

J. Liu, et al.

Fig. 6. The Y/Ho ratios in the total and six defined fractions of settling particles and sediments, error bars (standard deviations of mean values) were calculated for total concentration and each fraction within sample groups. Alphabetical letters before the hyphen represent significant differences (p < 0.05) between different sample groups in the fixed fraction, while latter ones denote significant differences (p < 0.05) between different fractions in the fixed group.

vulnerable F1 and F2 fraction (Fig. 8), demonstrating certain separate in pair-to-pair contrast graphs. For the F1 fraction, pore-water fluxes from reducing sediments (Haley et al., 2004), and sedimentary organic matter (Hathorne et al., 2015), can potentially produce a MREE enrichment in liquid phase. The highest HREE/LREE in F1 fraction was accounted for by inorganic riverine carbonates content (i.e. the continental source). Consequently, the discrepancy adsorption between the bottom deposited degrading/ continental end-member and the upper autochthonous neoformed endmember could be anticipated in the plot of MREE/MREE* against HREE/LREE. Yet, a linear relationship was obtained for F1 whole dataset (Fig. 8): MREE/MREE* = 1.03 × HREE/LREE + 0.65, r = 0.93, p = 0.00, demonstrated that only quantitative, not distinct tendency/

slope, have taken place. Another issue is introduced for F2, the linear relationship was MREE/MREE* = 0.86 × HREE/LREE + 0.70 (r = 0.77, p = 0.00) for F2 whole dataset, but equation of MREE/ MREE* = 2.28 × HREE/LREE + 0.15 (r = 0.94, p = 0.00) with greater correlation coefficient was calculated for F2 particle dataset (TS1 and TS2), and equation of MREE/MREE* = 0.81 × HREE/LREE + 0.71 (r = 0.82, p = 0.00) was calculated for F2 sediment dataset (SS and CS5, except the outlier of S3). The slope and intercept inconsistency in F2 revealed different provenance and/or allocation mechanism for reducible amorphous FeeMn oxides. Redox cycles at sediment-seawater interface have refreshed the original/’performed’ vulnerable poorly crystalline Fe oxide pool (e.g., ferrihydrite, lepidocrocite) to certain extent, made them more prone to record the mixed seawater-porewater

Fig. 7. Depth profile of Y/Ho ratios in the total and six defined fractions for core sediments CS5 (shadows indicate the mean value ± standard error of their analogues in settling particles). 88

Marine Pollution Bulletin 144 (2019) 79–91

J. Liu, et al.

Fig. 8. Rare earth element pattern indices (all NASC normalized); (above) cerium anomaly against the yttrium to holmium ratio, (bottom) middle REE bulge index versus the HREE to LREE ratio.

Fe-oxyhydroxides, while the MREE bulge patterns in F3 fraction reflected the respective opposite trend of Mn-oxyhydroxides and Feoxyhydroxides in REY adsorption. 3. Even not shown in the total composition, moderately positive Ce anomalies were found in reducible oxides fractions and pronounced negative Ce anomalies occurred for oxidizable organic/sulphidic fractions, largely because the cerium fractionation is strongly influenced by oxidation-reduction reactions. Minor inner-variation happened to Eu anomalies, manifested its less mobility than Sm and Gd during hydromorphic transport. Moreover, the Y/Ho fractionation produced by the anomalously low affinity of yttrium for iron (Fe) (oxyhydr)oxides relative to lanthanides with similar ionic radius (e.g., Ho) provided important information about solid-state transformation and ripening, also early diagenetic reactions. 4. Most prominent aspect for this study was the separated distribution of REY pattern indices, including cerium anomaly against the yttrium to holmium ratio, middle REE bulge index versus the HREE to LREE ratio, between settling particles and sediments. Although total concentration and other bounding phase (from F3 to F6) of particles and sediments overlapped to each other in pair-to-pair contrast graphs, distinctive slopes and intercepts in middle REE bulge index versus the HREE to LREE ratio comparison figure for F2 fraction revealed different provenance and/or allocation mechanism for reducible amorphous FeeMn oxides in sediments and sinking particles. Combined with the magnitude differences of F1 adsorption processes, the patch allocation of Y/Ho ratios against the Ce anomalies for F1 and F2 fractions, it is possible to create a precedent for proxies differentiation for the binary resuspension mixing model.

signature of lower ratios of MREE/MREE* versus HREE/LREE than the fresh river water. Considering the hydrogenetic FeeMn-oxides is responsible for Y/Ho and Ce fractionation, the Y/Ho ratios plotted against the Ce anomalies were fairly separated for F1 and F2 fraction (Fig. 8), but not showing linearity, suggested patch characteristics for new and old FeeMn-oxides. Despite the disparity existed for REY pattern indices, these data is not enough to support the binary mixing and compute the resuspension ratio in settling particles, because the ‘authigenic/biogenic’ end-member was absent in present framework. Redox interface with possible presence among the water column, not the traditional sediment-seawater interface, can greatly influence the formation of hydrogenetic FeeMn-oxides for resuspended particles. Therefore, until the redox interface was assured nearby the sediment-seawater interface and these proxies remained relative stable during the resuspension transportation, the identified proxies in Fig. 8 cannot be applied for the discrimination of resuspended and freshly-deposited particles. On the basis of bulk composition of rare earth elements and yttrium (REY), compartment-specific partitioning REY signatures in six operated-defined fractions, and divergences of REY pattern indices, we reached the following conclusions: 1. The total and individual REY concentrations of settling particles (TS1 and TS2) and core sediments (CS5) generally exhibited no significant difference, and are significantly larger than that of surface sediments (SS), partially induced by grain-size effect. The close relationship of REY concentration and normalization pattern between sinking particles, sediments and fine-grained fraction of major river sediments around the Yellow Sea, confirmed the priority contribution of catchment detrital materials. 2. Reducible amorphous Fe oxides (F2) were validated to be the main host phase for labile REY pool, much more than exchangeable and carbonates fraction (F1), crystallized Fe-oxides fraction (F3), magnetite fraction (F4), and organic matter fraction (F5). The majority of the REY, over 50%, was still found in residual fraction (F6). A convex shapes for sediment F1 dataset did not duplicate for sinking particle F1 dataset, primarily because different vertical fluid REY profile can strongly influence adsorption processes, also various inorganic riverine transported and ocean biogenic carbonates distribution. The HREE depletion patterns were discovered for F2 fraction due to the preferential scavenging of LREEs in amorphous

Acknowledgements This research was funded by the National Basic Research Program of China (973 Program) (No. 2015CB452902); the Aoshan Technology Innovation and Aoshan Talents Program Supported by Qingdao National Laboratory for Marine Science and Technology (No. 2016ASKJ02-4); the National Key Research and Development Plan Sino-Australian Centre for Healthy Coasts (No. 2016YFE0101500), the Youth Innovation Promotion Association of the Chinese Academy of Science (No. 2016191) and the Joint Fund between Natural Science Foundation of China and Shandong Province (No. U1606404). 89

Marine Pollution Bulletin 144 (2019) 79–91

J. Liu, et al.

Appendix A. Supplementary data

Kim, J.H., Torres, M.E., Haley, B.A., Kastner, M., Pohlman, J.W., Riedel, M., Lee, Y.J., 2012. The effect of diagenesis and fluid migration on rare earth element distribution in pore fluids of the northern Cascadia accretionary margin. Chem. Geol. 291, 152–165. Konhauser Kurt, O., 2006. Bacterial iron biomineralisation in nature. FEMS Microbiol. Rev. 20, 315–326. Laveuf, C., Cornu, S., 2009. A review on the potentiality of Rare Earth Elements to trace pedogenetic processes. Geoderma 154, 1–12. Lecomte, K.L., Sarmiento, A.M., Borrego, J., Nieto, J.M., 2017. Rare earth elements mobility processes in an AMD-affected estuary: Huelva Estuary (SW Spain). Mar. Pollut. Bull. 121, 282–291. Lerche, D., Nozaki, Y., 1998. Rare earth elements of sinking particulate matter in the Japan trench. Earth Planet. Sci. Lett. 159, 71–86. Leybourne, M.I., Johannesson, K.H., 2008. Rare earth elements (REE) and yttrium in stream waters, stream sediments, and Fe-Mn oxyhydroxides: fractionation, speciation, and controls over REE plus Y patterns in the surface environment. Geochim. Cosmochim. Acta 72, 5962–5983. Liang, X., Song, J., Duan, L., Yuan, H., Li, X., Li, N., Qu, B., Wang, Q., Xing, J., 2018. Source identification and risk assessment based on fractionation of heavy metals in surface sediments of Jiaozhou Bay, China. Mar. Pollut. Bull. 128, 548–556. Liu, J., Ye, S., Allen Laws, E., Xue, C., Yuan, H., Ding, X., Zhao, G., Yang, S., He, L., Wang, J., Pei, S., Wang, Y., Lu, Q., 2016. Sedimentary environment evolution and biogenic silica records over 33,000 years in the Liaohe delta, China. Limnol. Oceanogr. 62, 474–489. Liu, J., Ye, S., Yuan, H., Ding, X., Zhao, G., Yang, S., He, L., Wang, J., Pei, S., Huang, X., 2018. Metal pollution across the upper delta plain wetlands and its adjacent shallow sea wetland, northeast of China: implications for the filtration functions of wetlands. Environ. Sci. Pollut. Res. 25, 5934–5949. Liu, J., Song, J., Yuan, H., Li, X., Li, N., Duan, L., 2019. Biogenic matter characteristics, deposition flux correction, and internal phosphorus transformation in Jiaozhou Bay, North China. J. Mar. Syst. 196, 1–13. Luo, Y.R., Byrne, R.H., 2004. Carbonate complexation of yttrium and the rare earth elements in natural waters. Geochim. Cosmochim. Acta 68, 691–699. Mao, G., Hua, R., Gao, J., Zhao, K., Long, G., Lu, H., Yao, J., 2010. Rare earth element and trace element features of gold-bearing pyrite in the Jinshan gold deposit, Jiangxi Province. Acta Geol. Sin. 84, 614–623. Marmolejo-Rodriguez, A.J., Prego, R., Meyer-Willerer, A., Shumilin, E., Sapozhnikov, D., 2007. Rare earth elements in iron oxy-hydroxide rich sediments from the Marabasco River-Estuary System (pacific coast of Mexico). REE affinity with iron and aluminium. J. Geochem. Explor. 94, 43–51. Matisoff, G., Carson, M.L., 2014. Sediment resuspension in the Lake Erie nearshore. J. Great Lakes Res. 40, 532–540. Matisoff, G., Watson, S.B., Guo, J., Duewiger, A., Steely, R., 2017. Sediment and nutrient distribution and resuspension in Lake Winnipeg. Sci. Total Environ. 575, 173–186. Mihajlovic, J., Staerk, H.-J., Rinklebe, J., 2014. Geochemical fractions of rare earth elements in two floodplain soil profiles at the Wupper River, Germany. Geoderma 228, 160–172. Möller, P., Muecke, G.K., 1984. Significance of Europium anomalies in silicate melts and crystal-melt equilibria: a re-evaluation. Contrib. Mineral. Petrol. 87, 242–250. Nozaki, Y., Alibo, D.S., 2003. Importance of vertical geochemical processes in controlling the oceanic profiles of dissolved rare earth elements in the northeastern Indian Ocean. Earth Planet. Sci. Lett. 205, 155–172. Ohta, A., Kawabe, I., 2000. Rare earth element partitioning between Fe oxyhydroxide precipitates and aqueous NaCl solutions doped with NaHCO3: determinations of rare earth element complexation constants with carbonate ions. Geochem. J. 34, 439–454. Ohta, A., Kawabe, I., 2001. REE(III) adsorption onto Mn dioxide (delta-MnO2) and Fe oxyhydroxide: Ce(III) oxidation by delta-MnO2. Geochim. Cosmochim. Acta 65, 695–703. Oka, A., Hasumi, H., Obata, H., Gamo, T., Yamanaka, Y., 2009. Study on vertical profiles of rare earth elements by using an ocean general circulation model. Glob. Biogeochem. Cycles 23. Pack, A., Russell, S.S., Shelley, J.M.G., van Zuilen, M., 2007. Geo- and cosmochemistry of the twin elements yttrium and holmium. Geochim. Cosmochim. Acta 71, 4592–4608. Pattan, J.N., Pearce, N.J.G., Mislankar, P.G., 2005. Constraints in using cerium-anomaly of bulk sediments as an indicator of paleo bottom water redox environment: a case study from the Central Indian Ocean Basin. Chem. Geol. 221, 260–278. Polyakov, V.O., Nearing, M.A., 2004. Rare earth element oxides for tracing sediment movement. CATENA 55, 255–276. Poulton, S.W., Canfield, D.E., 2005. Development of a sequential extraction procedure for iron: implications for iron partitioning in continentally derived particulates. Chem. Geol. 214, 209–221. Prakash, L.S., Ray, D., Paropkari, A.L., Mudholkar, A.V., Satyanarayanan, M., Sreenivas, B., Chandrasekharam, D., Kota, D., Raju, K.A.K., Kaisary, S., Balaram, V., Gurav, T., 2012. Distribution of REEs and yttrium among major geochemical phases of marine Fe-Mn-oxides: comparative study between hydrogenous and hydrothermal deposits. Chem. Geol. 312, 127–137. Qiu, X., Liu, C., Wang, F., Deng, Y., Mao, G., 2015. Trace and rare earth element geochemistry of the Upper Triassic mudstones in the southern Ordos Basin, Central China. Geol. J. 50, 399–413. Rasmussen, B., Buick, R., Taylor, W.R., 1998. Removal of oceanic REE by authigenic precipitation of phosphatic minerals. Earth Planet. Sci. Lett. 164, 135–149. Rauret, G., Lopez-Sanchez, J., F., Sahuquillo, A., Rubio, R., Davidson, C., Ure, A., Quevauviller, P., 1999. Improvement of the BCR three step sequential extraction procedure prior to the certification of new sediment and soil reference materials. J. Environ. Monit. 1, 57–61. Reynard, B., Lecuyer, C., Grandjean, P., 1999. Crystal-chemical controls on rare-earth

Supplementary data to this article can be found online at https:// doi.org/10.1016/j.marpolbul.2019.04.044. References Abbott, A.N., Haley, B.A., McManus, J., Reimers, C.E., 2015. The sedimentary flux of dissolved rare earth elements to the ocean. Geochim. Cosmochim. Acta 154, 186–200. Akagi, T., 2013. Rare earth element (REE)-silicic acid complexes in seawater to explain the incorporation of REEs in opal and the “leftover” REEs in surface water: new interpretation of dissolved REE distribution profiles. Geochim. Cosmochim. Acta 113, 174–192. Akagi, T., Hashimoto, Y., Fu, F.-F., Tsuno, H., Tao, H., Nakano, Y., 2004. Variation of the distribution coefficients of rare earth elements in modern coral-lattices: species and site dependencies. Geochim. Cosmochim. Acta 68, 2265–2273. Akagi, T., Fu, F.F., Hongo, Y., Takahashi, K., 2011. Composition of rare earth elements in settling particles collected in the highly productive North Pacific Ocean and Bering Sea: implications for siliceous-matter dissolution kinetics and formation of two REEenriched phases. Geochim. Cosmochim. Acta 75, 4857–4876. Appel, C., Ma, L.Q., Dean Rhue, R., Kennelley, E., 2003. Point of zero charge determination in soils and minerals via traditional methods and detection of electroacoustic mobility. Geoderma 113, 77–93. Aspila, K.I., Agemian, H., Chau, A.S.Y., 1976. A semi-automated method for the determination of inorganic, organic and total phosphate in sediments. Analyst 101, 187–197. Avnimelech, Y., Kochva, M., Hargreaves, J.A., 1999. Sedimentation and resuspension in earthen fish ponds. J. World Aquacult. Soc. 30, 401–409. Bau, M., Koschinsky, A., 2009. Oxidative scavenging of cerium on hydrous Fe oxide: evidence from the distribution of rare earth elements and yttrium between Fe oxides and Mn oxides in hydrogenetic ferromanganese crusts. Geochem. J. 43, 37–47. Bau, M., Koschinsky, A., Dulski, P., Hein, J.R., 1996. Comparison of the partitioning behaviours of yttrium, rare earth elements, and titanium between hydrogenetic marine ferromanganese crusts and seawater. Geochim. Cosmochim. Acta 60, 1709–1725. Bayon, G., German, C.R., Burton, K.W., Nesbitt, R.W., Rogers, N., 2004. Sedimentary FeMn oxyhydroxides as paleoceanographic archives and the role of aeolian flux in regulating oceanic dissolved REE. Earth Planet. Sci. Lett. 224, 477–492. Bloesch, J., 1994. A review of methods used to measure sediment resuspension. Hydrobiologia 284, 13–18. Braun, J.-J., Pagel, M., Herbilln, A., Rosin, C., 1993. Mobilization and redistribution of REEs and thorium in a syenitic lateritic profile: a mass balance study. Geochim. Cosmochim. Acta 57, 4419–4434. Byrne, R.H., Kim, K.H., 1990. Rare earth element scavenging in seawater. Geochim. Cosmochim. Acta 54, 2645–2656. Caetano, M., Prego, R., Vale, C., de Pablo, H., Marmolejo-Rodriguez, J., 2009. Record of diagenesis of rare earth elements and other metals in a transitional sedimentary environment. Mar. Chem. 116, 36–46. Chen, J., Algeo, T.J., Zhao, L., Chen, Z.-Q., Cao, L., Zhang, L., Li, Y., 2015. Diagenetic uptake of rare earth elements by bioapatite, with an example from Lower Triassic conodonts of South China. Earth Sci. Rev. 149, 181–202. Compton, J.S., White, R.A., Smith, M., 2003. Rare earth element behavior in soils and salt pan sediments of a semi-arid granitic terrain in the Western Cape, South Africa. Chem. Geol. 201, 239–255. Cooper, D.C., Picardal, F., Rivera, J., Talbot, C., 2000. Zinc immobilization and magnetite formation via ferric oxide reduction by Shewanella putrefaciens 200. Environ. Sci. Technol. 34, 100–106. Du Laing, G., Rinklebe, J., Vandecasteele, B., Meers, E., Tack, F.M.G., 2009. Trace metal behaviour in estuarine and riverine floodplain soils and sediments: a review. Sci. Total Environ. 407, 3972–3985. Duan, L., Song, J., Li, X., Yuan, H., Li, N., Xu, Y., 2012. Thallium concentrations and sources in the surface sediments of Bohai Bay. Mar. Environ. Res. 73, 25–31. Garcia-Solsona, E., Jeandel, C., Labatut, M., Lacan, F., Vance, D., Chavagnac, V., Pradoux, C., 2014. Rare earth elements and Nd isotopes tracing water mass mixing and particle-seawater interactions in the SE Atlantic. Geochim. Cosmochim. Acta 125, 351–372. Gardner, W.D., Southard, J.B., Hollister, C.D., 1985. Sedimentation, resuspension and chemistry of particles in the northwest Atlantic. Mar. Geol. 65, 199–242. Gasith, A., 1976. Seston dynamics and tripton sedimentation in the pelagic zone of a shallow eutrophic lake. Hydrobiologia 51, 225–231. Haley, B.A., Klinkhammer, G.P., McManus, J., 2004. Rare earth elements in pore waters of marine sediments. Geochim. Cosmochim. Acta 68, 1265–1279. Hannigan, R.E., Sholkovitz, E.R., 2001. The development of middle rare earth element enrichments in freshwaters: weathering of phosphate minerals. Chem. Geol. 175, 495–508. Hathorne, E.C., Stichel, T., Brueck, B., Frank, M., 2015. Rare earth element distribution in the Atlantic sector of the Southern Ocean: the balance between particle scavenging and vertical supply. Mar. Chem. 177, 157–171. Hung, C., Tseng, C., Gong, G., Chen, K., Chen, M., Hsu, S., 2013. Fluxes of particulate organic carbon in the East China Sea in summer. Biogeosciences 10, 6469–6484. Hung, C., Chen, Y., Hsu, S., Wang, K., Chen, J., Burdige, D.J., 2016. Using rare earth elements to constrain particulate organic carbon flux in the East China Sea. Sci. Rep. 6, 33880.

90

Marine Pollution Bulletin 144 (2019) 79–91

J. Liu, et al. element concentrations in fossil biogenic apatites and implications for paleoenvironmental reconstructions. Chem. Geol. 155, 233–241. Rutgers van der Loeff, M.M., Boudreau, B.P., 1997. The effect of resuspension on chemical exchanges at the sediment-water interface in the deep sea — a modelling and natural radiotracer approach. J. Mar. Syst. 11, 305–342. Shields, G., Stille, P., 2001. Diagenetic constraints on the use of cerium anomalies as palaeoseawater redox proxies: an isotopic and REE study of Cambrian phosphorites. Chem. Geol. 175, 29–48. Siregar, A., Kleber, M., Mikutta, R., Jahn, R., 2004. Sodium hypochlorite oxidation reduces soil organic matter concentrations without affecting inorganic soil constituents. Eur. J. Soil Sci. 56, 481–490. Song, Y.H., Choi, M.S., 2009. REE geochemistry of fine-grained sediments from major rivers around the Yellow Sea. Chem. Geol. 266, 328–342. Song, J., Duan, L., Yuan, H., 2016. The Evolution of the Chemical Environment in the Jiaozhou Bay. Science Press, Beijing. Tang, J.W., Johannesson, K.H., 2010. Ligand extraction of rare earth elements from aquifer sediments: implications for rare earth element complexation with organic matter in natural waters. Geochim. Cosmochim. Acta 74, 6690–6705. Taylor, S.R., McLennan, S.M., 1985. The Continental Crust: Its Composition and Evolution. Blackwell, Oxford. Tessier, A., Campbell, P.G.C., Bisson, M., 1979. Sequential extraction procedure for the speciation of particulate trace metals. Anal. Chem. 51, 844–851. Thibault de Chanvalon, A., Metzger, E., Mouret, A., Knoery, J., Chiffoleau, J.F., BrachPapa, C., 2016. Particles transformation in estuaries: Fe, Mn and REE signatures through the Loire Estuary. J. Sea Res. 118, 103–112. Thompson, A., Amistadi, M.K., Chadwick, O.A., Chorover, J., 2013. Fractionation of yttrium and holmium during basaltic soil weathering. Geochim. Cosmochim. Acta 119, 18–30. Vazquez-Ortega, A., Perdrial, J., Harpold, A., Zapata-Rios, X., Rasmussen, C., McIntosh, J., Schaap, M., Pelletier, J.D., Brooks, P.D., Amistadi, M.K., Chorover, J., 2015. Rare earth elements as reactive tracers of biogeochemical weathering in forested rhyolitic terrain. Chem. Geol. 391, 19–32.

Vazquez-Ortega, A., Huckle, D., Perdrial, J., Amistadi, M.K., Durcik, M., Rasmussen, C., McIntosh, J., Chorover, J., 2016. Solid-phase redistribution of rare earth elements in hillslope pedons subjected to different hydrologic fluxes. Chem. Geol. 426, 1–18. Wang, G., Hao, F., Li, P., Zou, H., 2016. Use of rare earth element geochemistry to constrain the source of dolomitizing fluid for dolomitization of the Lower Triassic Feixianguan Formation, Jiannan area, China. J. Pet. Sci. Eng. 138, 282–291. Xu, Y., Song, J., Duan, L., Li, X., Zhang, Y., Sun, P., 2010. Environmental geochemistry reflected by rare earth elements in Bohai Bay (North China) core sediments. J. Environ. Monit. 12, 1547–1555. Xu, Y., Song, J., Duan, L., Li, X., Yuan, H., Li, N., Zhang, P., Zhang, Y., Xu, S., Zhang, M., Wu, X., Yin, X., 2012. Fraction characteristics of rare earth elements in the surface sediment of Bohai Bay, North China. Environ. Monit. Assess. 184, 7275–7292. Yan, X., Kerrich, R., Hendry, M.J., 1999. Sequential leachates of multiple grain size fractions from a clay-rich till, Saskatchewan, Canada: implications for controls on the rare earth element geochemistry of porewaters in an aquitard. Chem. Geol. 158, 53–79. Yang, J., Torres, M., McManus, J., Algeo, T.J., Hakala, J.A., Verba, C., 2017. Controls on rare earth element distributions in ancient organic-rich sedimentary sequences: role of post-depositional diagenesis of phosphorus phases. Chem. Geol. 466, 533–544. Ye, S., Laws, E.A., Gambrell, R., 2013. Trace element remobilization following the resuspension of sediments under controlled redox conditions: City Park Lake, Baton Rouge, LA. Appl. Geochem. 28, 91–99. Zhang, S., Shan, X., 2001. Speciation of rare earth elements in soil and accumulation by wheat with rare earth fertilizer application. Environ. Pollut. 112, 395–405. Zhang, L., Algeo, T.J., Cao, L., Zhao, L., Chen, Z., Li, Z., 2016. Diagenetic uptake of rare earth elements by conodont apatite. Palaeogeogr. Palaeoclimatol. Palaeoecol. 458, 176–197. Zhao, L., Chen, Z., Algeo, T.J., Chen, J., Chen, Y., Tong, J., Gao, S., Zhou, L., Hu, Z., Liu, Y., 2013. Rare-earth element patterns in conodont albid crowns: evidence for massive inputs of volcanic ash during the latest Permian biocrisis? Glob. Planet. Chang. 105, 135–151.

91