Influence of dissolved organic matter on dissolved vanadium speciation in the Churchill River estuary (Manitoba, Canada)

Chemosphere 154 (2016) 367e374

Contents lists available at ScienceDirect

Chemosphere journal homepage: www.elsevier.com/locate/chemosphere

Influence of dissolved organic matter on dissolved vanadium speciation in the Churchill River estuary (Manitoba, Canada)
line Gue
guen b, * Yong Xiang Shi a, Vaughn Mangal a, Ce a b

Environment and Life Sciences Graduate Program, Trent University, Peterborough, ON, Canada Department of Chemistry, Trent University, Peterborough, ON, Canada

h i g h l i g h t s
Temporal and spatial changes in dissolved V speciation using DGT.
Positive relationship found between protein-like DOM and DGT-labile V.
Surface sediment is a significant source of DGT-labile V during spring freshet.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 6 November 2015 Received in revised form 15 March 2016 Accepted 25 March 2016 Available online 9 April 2016

Diffusive gradients in thin films (DGT) devices were used to investigate the temporal and spatial changes in vanadium (V) speciation in the Churchill estuary system (Manitoba). Thirty-six DGT sets and 95 discrete water samples were collected at 8 river and 3 estuary sites during spring freshet and summer base flow. Dissolved V concentration in the Churchill River at summer base flow was approximately 5 times higher than those during the spring high flow (27.3 ± 18.9 nM vs 4.8 ± 3.5 nM). DGT-labile V showed an opposite trend with greater values found during the spring high flow (2.6 ± 1.8 nM vs 1.4 ± 0.3 nM). Parallel factor analysis (PARAFAC) conducted on 95 excitation-emission matrix spectra validated four humic-like (C1eC4) and one protein-like (C5) fluorescent components. Significant positive relationship was found between protein-like DOM and DGT-labile V (r ¼ 0.53, p < 0.05), indicating that protein-like DOM possibly affected the DGT-labile V concentration in Churchill River. Sediment leachates were enriched in DGT-labile V and protein-like DOM, which can be readily released when river sediment began to thaw during spring freshet. © 2016 Elsevier Ltd. All rights reserved.

Handling Editor: Martine Leermakers Keywords: DGT-Labile species DOM High resolution mass spectrometry PARAFAC Vanadium

1. Introduction Vanadium (V) is a transition metal widely distributed in natural environments (air, water, soil and biota) with potential mutagenic and carcinogenic properties (Stern et al., 1993). Natural sources of V in aquatic environments encompass rock weathering and sediment leaching (Hope, 1997; Rühling and Tyler, 2001; Shiller and Mao, 2000). Vanadium can also be introduced into aquatic environments through industrial activities especially from oil refineries and power plants that are burning V-rich fuel oil or coal (Evans and Landergran, 1975; Moskalyk and Alfanti, 2003; Kirk et al., 2014;
guen et al., 2006). Other sources of V to aquatic environments Gue are fertilizers, sewage sludge, and discharge of domestic

* Corresponding author.
guen). E-mail address: [email protected] (C. Gue http://dx.doi.org/10.1016/j.chemosphere.2016.03.124 0045-6535/© 2016 Elsevier Ltd. All rights reserved.

wastewater (Bhatnagar et al., 2008; McBride and Cherney, 2004). Vanadium is essential for normal cell growth at concentrations below 0.10 mM (Salice et al., 1999) but excessive amounts can cause adverse health effects on human and aquatic species. For example, the total V lethal concentration (LC50) was 37e118 mM for juvenile rainbow trout (Salmo gairdneri R.) (Frank et al., 1996), 39e59 mM for zebra fish (Brachydanio rerio) (Bishayee et al., 2010), and 65 mM for guppies (Poecilia reticulata) (Bishayee et al., 2010). In surface water, V with the oxidation state of þ5 is the dominant and most toxic form to cells (Sabbioni et al., 1991) and organisms (Ma and Fu, 2009). The toxicity of trace metals in aquatic environments is greatly influenced by its chemical speciation and particularly the concentration of its free form and small soluble complexes. On the other hand, metal bound to dissolved organic matter (DOM), iron and aluminum oxyhydroxides are not readily bioavailable (Koukal et al.,

368

Y.X. Shi et al. / Chemosphere 154 (2016) 367e374

2007). A change in DOM composition and concentration may affect bioavailability and mobility of metals in waters (Koukal et al., 2003). The diffusive gradients in thin-films (DGT) technique is a timeintegrated, passive sampler technique for the in situ determination of free and weakly bound metals in natural waters (Davison and Zhang, 1994). The DGT device is composed of a diffusive layer, conventionally a polyacrylamide hydrogel, and a binding layer, € typically ferrihydrite (Luo et al., 2010; Osterlund et al., 2010; Price et al., 2013) or zirconia (Guan et al., 2015) impregnated within the polyacrylamide hydrogel. DGT with ferrihydrite binding gel has been deployed in synthetic and natural freshwater to measure V in € his highest oxidation state þ5 (Luo et al., 2010; Osterlund et al., 2010; Panther et al., 2013; Price et al., 2013). Despite the fact that spring freshet is characterized by a dramatic increase in flux of dissolved and particulate species €lemann et al., 2005; Kuzyk et al., 2008; Mann et al., 2012), it is (Ho also the least sampled period. The discharge of Churchill River
ry et al., 2005) causes during freshet periods (up to 1320 m3/s; De significant changes in the estuary including increased stratification and turbidity, reduced marine and freshwater nutrient supply, and supplying the large freshwater pool into the interior ocean (Kuzyk et al., 2008). These conditions suppressed phytoplankton productivity while increasing organic matter supply by the river (Kuzyk et al., 2008). However, DOM concentration and composition as well as V concentration and speciation are still largely unknown during this period. Earlier studies on dissolved V speciation showed the DOM adsorption effect was significant in river and estuary (Shiller and Mao, 2000). Lu et al. (1998) found that aquatic humic substances strongly complexed vanadate and VOþ 2 . Excitation-emission matrix (EEM) fluorescence spectroscopy and parallel factor analysis (PARAFAC) have been used to assess the composition (e.g. proteinand humic-like based on peak position in the EEM; Coble, 1996) and
guen, 2013; origin of DOM in natural waters (e.g. Dainard and Gue Kothawala et al., 2013; Walker et al., 2013). Recent advances in ultra-high resolution mass spectroscopy allowed the identification of structural composition of DOM (e.g. D'Andrilli et al., 2013; Hertkorn et al., 2008; Mangal et al., 2016), a prerequisite for exploring the relationship between DOM and metal complexation. In this paper, the objectives were (1) to determine the concentration of dissolved and DGT-labile V in the Churchill estuary system during spring freshet and at base flow regime; (2) to assess the impact of DOM on V speciation. The influence of DOM concentration and fluorescence composition (Coble, 1996) on V speciation will be discussed; (3) to determine dissolved V speciation in sediment leachate in the Churchill River. Although Shiller and Mao (2000) showed that sediment leaching and rock weathering were important sources of fluvial dissolved V, it is not clear how it would affect its dissolved speciation (i.e. DGT-labile concentration). 2. Materials and methods 2.1. Site description and sampling The Churchill River, with an average discharge of 1200 m3/s
ry et al., 2005), drains a portion of the Canadian Shield and (De empties into western Hudson Bay. The drainage basin
ry et al., 2005) is dominated by flat muskeg plains (288880 km2; De and shallow lakes. Since 1977, about 90% of the river flow upstream of Southern Indian Lake (~320 km above the Bay) is diverted into the Nelson River to generate power. Ten study sites (R1-R7, E1-E3; Fig. 1) located in the enclosed estuary (approximately 13 km long and up to 3 km wide) in the lower end of the Churchill River were sampled in 2013e2015. Triplicate DGT units were deployed in the Churchill River

estuary system (Fig. 1A) from August 23 to August 27, 2013 (summer base flow; Fig. 1B) and from May 8 to May 16, 2014 (spring prefreshet; Fig. 1B) and from May 12 to May 21, 2015 (spring freshet). Mean water level in the Churchill River was 22.51 ± 0.02 m in late August 2013, 24.07 ± 0.31 m pre-freshet and 23.90 ± 0.19 m freshet (Government of Canada (2015)). Due to site accessibility in spring, all sites could not be monitored at all seasons. The estuarine DGT units (E1-E3) were deployed in summer 2013 for 4e5 h during high tide over a 6-day period for an overall accumulation period of 24e30 h in order to assess the influence of marine-derived DOM. The prefreshet (R1 and R5) and freshet DGT units (R2-R3 and R6) were deployed for 2e4 d in moving waters (300e350 m3/s) after drilling a hole through the river ice. The diffusive boundary layer was assumed to be negligible (Denney et al., 1999; Gimpel et al., 2001). R2 was the only site in the vicinity of a shallow wetland. Water samples were daily collected at each site during the time of DGT deployment and immediately filtered through a precombusted, 0.7 mm glass fiber filter (Whatman) (Dainard and
guen, 2013) and a 0.22 mm nitrocellulose filter (Millipore) for Gue DOM analysis. Dissolved V samples were obtained after filtration through 1 mm Nuclepore membrane filter (Whatman) and stored at 4  C until analysis. Temperature of river was measured daily using a digital thermometer (Accumet AP85; ThermoScientific). 2.2. DGT preparation Each DGT unit was composed of a 0.5 mm precipitated ferrihydrite binding gel layer (Luo et al., 2010), a 0.8 mm diffusive acrylamide-based gel layer and a 0.45 mm cellulose nitrate filter (Zhang and Davison, 1999). The binding and diffusive gels were casted using 0.25 and 0.5 mm thick, acid-bathed, polystyrene spacers, respectively. The binding gel was composed of a 0.5 mm diffusive gel immersed in 1 M iron nitrate solution for 12 h and in 2N-morpholino ethanesulfonic acid (MES buffer) solution between 10 and 40 min (Luo et al., 2010). No significant difference in accumulated V was found in the 10 min soaked gel (p > 0.05). The binding gel was then rinsed with Milli-Q water (MQW) for 24 h and stored in MQW. The gel thicknesses (i.e. 0.5 and 0.8 mm for diffusive and binding gels, respectively) were verified using a digital caliper to an accuracy of 0.02 mm. The DGT preparation was conducted in metal-free 10,000 class clean room to minimize metal contamination. Blank concentrations were assessed by measuring the mass of metal present in binding gels on unit brought to the field but not deployed. The field blank of binding gels (n ¼ 3) was 0.37 nM, which was much lower than the lowest DGT-labile V concentrations measured in the Churchill River (i.e. 4.3 nM). Binding gels was eluted by 1 M nitric acid (HNO3; metal-free) for 24 h before ICP-MS analysis (XSeries II, Thermo). Indium and rhodium were used as internal standards. The accuracy of the ICPMS measurements was assessed using SLEW-3 and SLRS-5 reference water (National Research Council, Canada). The measured V concentrations were within 5% of the certified values. The time-integrated concentration of DGT-labile V in the bulk solution, CDGT was calculated based on Fick's First Law of Diffusion (Zhang and Davison, 1995): CDGT ¼ (M* Dg)/(t*A*D) where M is mass of V accumulated in the binding gel, Dg is thickness of diffusive gel and filter membrane, t is deployment time, A is sampling area exposed to bulk solution and D is diffusion coefficient of vanadium in diffusive gel and filter membrane (D ¼ 6.26  106 cm2/s at 25  C; Luo et al., 2010). The D values were corrected to in situ average temperatures using the Stokes-Einstein relation (Li and Gregory, 1974).

Y.X. Shi et al. / Chemosphere 154 (2016) 367e374

A

369

B 25.0

E3 E2

24.5

R7 R6 R5

R4 R3 R2 R1

Water level [m]

E1 24.0 23.5 23.0 22.5 22.0 Mar/13

Jul/13

Nov/13

Mar/14

Jul/14

Nov/14

Mar/15

Jul/15

Fig. 1. (A) Sampling sites in the Churchill River estuary in prefreshet (May 2014), freshet (May 2015) and summer base flow (August 2013) and (B) the corresponding water level in the Churchill River (Government of Canada, 2015).

2.3. Sediment leachate and DGT deployment A bulk sample of surface sediment (0e20 cm) was collected from river sites R2 and R3 in May 20, 2015 and stored at 20  C until analysis. Sediment samples were then oven-dried for 48 h and sieved to <2 mm for use in laboratory evaluations. For DGT deployment experiments, 30 g of oven-dried sediment sample was weighed into acid-washed glass containers. Vanadium concentration in the acid-washed containers was 0.04 nM, which was well below concentrations measured in sediment. The samples were hydrating with 300 ml of Milli-Q water (MQW; Millipore) and gently mixed using a variable speed nutating mixer (Fisher Scientific) for 42 h at 20 ± 1  C (Hooda et al., 1999). Leachate samples were filtered through 1 mm nitrocellulose filter (Millipore) for dissolved V and DOM analysis. Triplicate DGT units were deployed in filtered leachate and continuously stirred for 8 h. DGT binding gel was eluted and analyzed as previously for units deployed in the Churchill estuary system. 2.4. DOM analysis Dissolved organic carbon (DOC) concentration was determined by a Shimadzu total organic carbon analyzer (TOC-VCPH). Three to five replicate injections (150 mL each) were performed for each sample, resulting in a typical variation coefficients < 2%. The absorbance spectrum was obtained using a Shimadzu UV-2550 spectrometer. MQW blanks were run after each sample to prevent cross-contamination. The EEM spectra were obtained by scanning excitation wavelength from 250 to 500 nm and emission wavelength from 300 to 600 nm, both in increments of 5 nm in ratio (S/R) mode using a Horiba Jobin-Yvon Fluoromax 4. EEMs were corrected using the manufacturer-generated emission and excitation correction factors and calibrated to Raman units using the area under the Raman scatter peak of MQW (Lawaetz and Stedmon, 2009). Background scattering was removed by subtracting an EEM of ultrapure MQW measured daily. Absorbance-based correction was used to correct for inner filter effect with accuracy better than 5% (Gu and Kenny, 2009). 2.5. High resolution mass spectrometry analysis River water and sediment leachate DOM were diluted in 60:40

(v/v) trace metal grade methanol (Sigma Aldrich): MQW and continuously injected into a Thermo electrospray ion source of a Thermo Orbitrap (resolving power ¼ 140 000 at m/z 400) housed in the Trent Water Quality Centre. Samples were injected at a rate of 50 mL/min and analyzed in negative ionization mode. Mass spectra were externally calibrated following Thermo protocol and internally calibrated using CH2 homologous series naturally present in the sample. The accuracy and precision in the relevant mass range (m/z 200e1000) was always <1 ppm. Molecular formula assignment, data filtering and processing were performed as per Mangal
guen (2016). and Gue 2.6. Statistical analysis Parallel factor analysis (PARAFAC) was conducted on 95 EEMs using Matlab R2010a, following the guidelines outlined in the DOMFluor toolbox (Stedmon and Bro, 2008). Emission wavelengths of 500e600 nm were cut as a means of eliminating fluorescent scattering that would otherwise be detrimental to modeling. This was acceptable because excitation and emission maxima of PARAFAC modeled fluorescent components were not expected to fall within these ranges based on previous findings (Coble, 1996, 2007). The model was constrained to non-negative values and run for three to seven components. The appropriate number of components was determined by split-half analyses (Stedmon and Bro, 2008). Composition of fluorescent components were determined by dividing each component by total fluorescence (e.g. C1% ¼ C1/ (C1 þ C2 þ C3 þ C4þC5)*100). Centered log ratio transformation (CLR) was run on the proportions of each fluorescent component
guen, for dimension reduction of compositional data (Cuss and Gue 2015). Principal component analysis (PCA) was conducted using the CLR-transformed proportions of PARAFAC components, DGT-labile V and dissolved V concentration, and DOC using XLSTAT 2014. 3. Results and discussion 3.1. Dissolved organic carbon concentrations Concentrations of DOC in Churchill River (R1 - R7) and estuary (E1 - 3) ranged from 0.34 to 1.90 mM and from 0.25 to 0.75 mM, respectively (Fig. 2). The average river DOC concentration was

370

Y.X. Shi et al. / Chemosphere 154 (2016) 367e374 2.0

DOC (mM)

1.5

1.0

.5

0.0

E1

E2

E3

R1

R2

R3

R4

R5

R6

R7

Fig. 2. Average values of DOC at three estuary (E1-E3) and six river sites (R1-5 R7) in pre-freshet (white) and freshet (dark gray) and summer (gray).

1.06 ± 0.41, 0.65 ± 0.23 and 0.98 ± 0.24 mM in, pre-freshet, freshet and summer base flow respectively, which is consistent with other Arctic rivers (Stedmon et al., 2011). Spring freshet DOC values were significantly lower than at pre-freshet and summer flow (p < 0.05). DOC in R6 and R3 were significantly greater in summer than during freshet (p < 0.05; Fig. 2) probably due to dilution with DOC-poor melting waters. No significant differences were observed between pre-freshet and summer values (p > 0.05). The average estuary DOC concentration was 0.26 ± 0.10 mM in summer and 0.65 ± 0.25 mM in spring, which was consistent with other studies (Castelle et al., 2009). Estuary DOC was significantly greater during pre-freshet (p < 0.05) than summer base flow probably due to higher freshwater pulses of DOC (higher discharge and water level) from Churchill River to the estuary.

3.2. Fluorescent DOM (FDOM) A five-component model was split-half validated on 95 EEMs (Fig. S1). The components were compared with those in the OpenFluor database (Murphy et al., 2014). Component 1 (C1) had two excitation maximum (Ex) at <250 and 310 nm and one emission peak at 450 nm. This component was previously labeled as terrestrial humic-like in rivers (Mostofa et al., 2010; Hong et al., 2012; Dong et al., 2014) and coastal environments (Dainard and
guen, 2013; Murphy et al., 2008; Søndergaard et al., 2003) Gue and sediment (Yuan et al., 2014). Component C2 (Ex/Em ¼ 275/ 475 nm) was attributed to UV humic-like DOM (Coble, 1996; Mostofa et al., 2010; Hong et al., 2012). Component C3 displayed two excitation maxima (<250 and 295 nm) and one emission maximum at 400 nm. This peak was previously attributed to UV and microbial humic-like DOM (Coble, 1996; Murphy et al., 2008; Hong et al., 2012; Dong et al., 2014; Kowalczuk et al., 2009; Osburn et al., 2011; Yamashita et al., 2010b; Murphy et al., 2013; Kothawala et al., 2012, 2013; Cawley et al., 2012; Stedmon et al., 2003, 2007; Osburn and Stedmon, 2011; Walker et al., 2009; Shutova et al., 2014). C4 (Ex/Em ¼ <250/440 nm) was attributed to UV humic-like DOM (Coble, 1996; Mostofa et al., 2010; Hong et al., 2012; Murphy et al., 2011; Yamashita et al., 2010b; Shutova et al., 2014). Component C5 (Ex/Em ¼ 275/320 nm) was the only protein-like (tyrosine-like; Coble, 1996) component (Yamashita et al., 2010a, 2010b; Dong et al., 2014; Murphy et al., 2006; Kothawala et al., 2013; Yuan et al., 2014) found in this study. Humic-like C1eC3 dominated the FDOM pool at all seasons

(Fig. 3). The fluorescence of the humic-like C1eC3 increased from pre-freshet to summer base flow at R1 and R5 (p < 0.05) but remained stable from freshet to summer at R3 and R6. All R sites except R1, showed an increase in protein-like intensity (i.e. C5) from summer to pre-freshet/freshet conditions. C5 has been associated with the autochthonous production of DOM (Determann et al., 1996) which was relatively strong during spring freshet when irradiance and nutrient inputs from the river increased (Parsons et al., 1988). Some differences and similarities in the relative composition of PARAFAC components between flow regimes were apparent (Fig. 4). Humic-like (C1%eC2%) was more abundant during summer base flow than at pre-freshet (R1 and R5) and freshet (R6 and R3) (p < 0.05) (Fig. 4). On the other hand, protein-like C5% was significantly greater during freshet than summer base flow at R3 and R6 and significantly greater during pre-freshet than summer base flow at R5 (p < 0.05) (Fig. 4E). The temporal shifts in fluorescence composition were consistent with a shift from poorly degraded plant litter during spring freshet to more degraded, humified DOM during base flow conditions. Similar temporal patterns were reported in larger Arctic Rivers (Stedmon et al., 2011; Walker et al., 2013) C5% in surface sediment averaged 28.6e34.1% of the total fluorescence at R2-3, which was at least 2 times greater than in surface water (14.7e16.3%; Fig. 4E). The high protein contents in sediment leachates were also confirmed using high resolution mass spectrometry (Fig. S3). While a variety of compounds such as lignin, lipid and tannins can be determined, compounds with protein characteristics can be found with and O/C ratio ranging from 0.15 to 0.5 and H/C ratios ranging from 1.5 to 2.25 (Ohno et al., 2013; D'Andrilli et al., 2013). Total dissolved compounds associated with protein structures in sediment leachate ranged from 359 at R2 to 380 at R3, contributing to 39.4% and 26.1% of total assigned peaks respectively. Total dissolved compounds associated with protein structures in surface water ranged from 541 at R2 to 750 at R3, contributing to 26.3% and 15.7% of total assigned peaks respectively. Percentage of dissolved compounds associated with protein structures of the total dissolved compound in sediment was about 50% higher than in surface water. Together these results confirm that sediment leaching represented a significant source of proteinlike DOM in water. 3.3. V speciation in Churchill River sediment Dissolved V concentration in sediment leachates ranged from 90.0 ± 4.4 nM at R3 to 580.2 ± 9.7 nM at R2, with an average of 335.1 nM which was up to 50 times more than the average (6.8 nM) in the Churchill River surface waters (Fig. 5). The high V concentrations in sediment leachates confirm that surface sediment constitutes a significant source of dissolved V in rivers (Emerson and Huested, 1991; Shiller and Mao, 1999, 2000). On the other hand, DGT-labile V in sediment leachates ranged from 50.4 ± 11.4 nM at R7 to 565.5 ± 28.3 nM at R3. Like surface waters (Fig. 5A and C), more DGT-labile V was found in sediment leachates at R2 (97 ± 2% vs 56 ± 13%). Together these results confirm that sediment leaching can be a significant source of dissolved V and DGT-labile V in rivers. 3.4. V speciation in surface waters Concentrations of DGT-labile and dissolved V at R1-R7 ranged from 1 to 3 nM (Fig. 5A) and from 3 to 56 nM (Fig. 5B), respectively. The average dissolved V concentration (10.3 ± 9.4 nM) was consistent with previous studies (Shiller and Boyle, 1987; Ramani et al., 2014; Taylor et al., 2012). Average DGT-labile and dissolved V concentrations at estuary

Y.X. Shi et al. / Chemosphere 154 (2016) 367e374 .30

.30

.30

A

B

C

.25

.25

.20

.20

.20

.15

C3/DOC

.25

C2/DOC

C1/DOC

371

.15

.15

.10

.10

.10

.05

.05

.05

0.00 R1

R2

R3

R4

R5

R6

0.00

R7

0.00 R1

R2

R3

R4

R5

R6

R7

R1

R2

D

R4

R5

R6

R7

E

.25

.25

.20

.20

C5/DOC

C4/DOC

R3

.30

.30

.15

.15

.10

.10

.05

.05

0.00

0.00 R1

R2

R3

R4

R5

R6

R7

R1

R2

R3

R4

R5

R6

R7

Fig. 3. Average DOC-normalized PARAFAC loadings for (A) C1, (B) C2, (C) C3, (D) C4 and (E) C5 at R1-R7 in pre-freshet (white), freshet (dark gray) and summer base flow (gray).

50

50

A

50

B

C 40

30

30

30

C3%

C2%

40

C1%

40

20

20

20

10

10

10

0

0

R1

R2

R3

R4

R5

R6

R7

0

R1

R2

R3

50

R4

R5

R6

R7

R1

R2

R3

R4

R5

R6

R7

50

D

E 40

30

30

C4%

C5%

40

20

20

10

10

0

0

R1

R2

R3

R4

R5

R6

R7

R1

R2

R3

R4

R5

R6

R7

Fig. 4. Average PARAFAC abundance (%) for (A) C1, (B) C2, (C) C3, (D) C4 and (E) C5 at R1-R7) in pre-freshet (white), freshet (dark gray) and summer base flow (gray).

sites (E1-3) ranged from 6 to 17 nM (Fig. 5A) and from 200 to 660 nM (Fig. 5B), respectively. The estuarine dissolved concentrations were higher than those reported in the open ocean (10e40 nM; Jeandel et al., 1987). However the summer E sites were located in the surf zone in shallow waters where the interaction between oxic waters and V-rich sediments (up to 1400 nM; Emerson and Huested, 1991) increased the solubility of V may explain the high dissolved V concentration found in this study in aerobic water (Niencheski et al., 2014). Dissolved riverine V concentrations were 5.6 fold greater in summer than during freshet whereas the abundance of DGT-labile

V was significantly reduced from pre-freshet/freshet to summer base flow at all sites (Fig. 5C). For example, DGT-labile V at R1 accounted for 81% in pre-freshet conditions but only 15% in summer base flow. Sediment leaching is known to be a major source of dissolved V in aquatic systems (Emerson and Huested, 1991; Shiller and Mao, 2000). During pre-freshet period, bottom river sediment was still frozen, limiting the fresh leaching of dissolved V. As spring freshet progressed, river water temperatures increased above the freezing point and the bottom river sediments began to melt releasing large amount of dissolved V which were mostly DGTlabile as revealed by our sediment leachate experiments. This

372

Y.X. Shi et al. / Chemosphere 154 (2016) 367e374 1000

A

Dissolved V (nM)

DGTnM) Labla ilb eilVe (Vn(M )

100

10

1

0.1

100

10

1

E1

levl(e%d) V (%) LabileDGTV / Dlaisbsio

B

E2

E3

R1

R2

R3

R4

R5

R6

R7

E1

E2

E3

R1

R2

R3

R4

R5

R6

R7

100

C D

10

1

E1

E2

E3

R1

R2

R3

R4

R5

R6

R7

Fig. 5. Average concentrations of (A) DGT-labile V and (B) dissolved V and (C) abundance of DGT-labile V in pre-freshet (white), freshet (dark gray) and summer base flow (gray).

process explains in part the increase in DGT-labile V associated with spring freshet. The DGT-labile V contribution was significantly higher at the R sites than that at the E sites (42.4% vs 3.1%), likely due to the interactions with V-rich river sediment. Dissolved V concentrations were strongly positively correlated to DOC (r ¼ 0.48), humic-like C1eC4 (r ¼ 0.39e0.66) and, C2% and C4% (r ¼ 0.36 and 0.66), suggesting that dissolved V was strongly affected by DOC and humic-like intensity and abundance. Similar results were found by Shiller and Mao (2000). Lu et al. (1998) also reported that aquatic humic substances strongly complex V in aqueous solution. Dissolved and DGT-labile V concentrations were also weakly correlated with protein-like DOM. Stronger

relationships were found at some sites. For example, DGT-labile V concentrations were positively correlated with C5% at R1 at all seasons (r ¼ 0.92, p < 0.05), R6 during freshet (r ¼ 0.62, p < 0.05) and R2 during freshet (r ¼ 0.89, p < 0.05). The influence of proteinlike DOM on DGT-labile V concentrations agreed well with the variation in DOM quality in sediment where protein-rich sediment leachates showed greater DGT-labile V concentrations. 3.5. Principal component analysis and influence of hydrography To comprehensively assess spatial and temporal differences in dissolved V speciation, principal component analysis (PCA) was

A

B C5

1

5 Protein-like

0.75

R5 pre-freshet

4

0.5

Humic-like

summer

3

pre-freshet

F2 (20.58 %)

0.25 C1 C4

2

C2

1

freshet

0

-0.25

C3

0

-0.5 -0.75

-1

-1

-2

R1 pre-freshet -1

-0.75

-0.5

-0.25 0 0.25 F1 (73.40 %)

0.5

0.75

1

-3

-2

-1

0

1

2

3

4

5

6

Fig. 6. Results of principal component analysis (PCA) of V speciation and DOM characterization. (A) Loading plots for PC1 and PC2. (B) Sample scores for the first two principal components.

Y.X. Shi et al. / Chemosphere 154 (2016) 367e374

performed using the DOC-normalized loadings, compositional loadings of the five PARAFAC components (Fig. 6). The first two principal components accounted for 73.4 and 20.6% of the total variance, respectively. The humic-like loadings (C1-4) displayed high positive F1 values whereas protein-like loading and abundance had the lowest positive F1 values (Fig. 6A). The humic-like and protein-like components had dissimilar distributions along F1, suggesting that DOM quality (humic-like and protein-like) controls F1 (Fig. 6A). DGT-labile V was significantly positively correlated with F2 scores (r ¼ 0.68, p < 0.05). Therefore, F2 can represent DGT-labile V. The loadings of protein-like C5 showed the most positive F2 loadings, suggesting that DGT-labile V was affected by the abundance of protein-like C5. The scores plot (Fig. 6B) showed that pre-freshet/freshet and summer samples were displayed in opposite quadrants, confirming the seasonal impact on DOM and metal speciation. Freshet samples had higher F2 scores than summer samples suggesting that the abundance of C5 was influenced by spring melt conditions. This is congruent with higher concentration and composition of C5 during freshet than in summer (Figs. 3E and 4E). The pre-freshet samples were site-dependent with R5 associated with the same quadrant as the freshet samples, indicating that R5 during pre-freshet had high abundance of C5. This is probably because of proximity of a shallow wetland, a V-rich and C5-rich environment. 4. Conclusions This study assessed the concentrations of dissolved and DGTlabile V and DOC in the Churchill River estuary system during spring pre-freshet, freshet and in summer base flow. The DGT-labile V in Churchill River increased significantly in response to the increase of protein-like DOM. Sediment seems to be the major source for dissolved and DGT-labile V and protein-like DOM. The results of this study demonstrate the importance of studying the effects of DOM complexation on V speciation in river environments, which affects vanadium bioavailability and toxicity. The combination of DGT, PARAFAC and PCA is a powerful tool for interpreting relationships between optical properties both between distinct DOM sources, and V speciation. Acknowledgements This study would not have been possible without the assistance of Antoine Perroud and his sample analysis, expertise and knowledge of instrumentation. We particularly thank Yu Zhu, Jordan Balch, Antoine Perroud, Dave Allcorn and Jackie Verstege for lab and field assistance. We also thank Naomi Stock for her invaluable assistance with the high resolution mass spectrometry analysis. We are also grateful for the helpful comments of two anonymous reviewers. Financial support for this study was awarded from NSERC, Churchill Northern Study Center (CNSC) and Canada Research Chair program. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.chemosphere.2016.03.124. References Bhatnagar, A., Minocha, A.K., Pudasainee, D., Chung, H., Kim, S., Lee, G., Min, B., Jeon, B., 2008. Vanadium removal from water by waste metal sludge and cement immobilization. Chem. Eng. J. 144, 197e204. Bishayee, A., Waghray, A., Patel, M.A., Chatterjee, M., 2010. Vanadium in the detection, prevention and treatment of cancer: the in vivo evidence. Cancer

373

Lett. 294, 1e12. €fer, J., Blanc, G., Dabrin, A., Lanceleur, L., Masson, M., 2009. Gaseous Castelle, S., Scha mercury at the airewater interface of a highly turbid estuary (Gironde Estuary, France). Mar. Chem. 117, 42e51. Cawley, K.M., Butler, K.D., Aiken, G.R., Larsen, L.G., Huntington, T.G., McKnight, D.M., 2012. Identifying fluorescent pulp mill effluent in the Gulf of Maine and its watershed. Mar. Pollut. Bull. 64, 1678e1687. Coble, P.G., 1996. Characterization of marine and terrestrial DOM in seawater using excitation-emission matrix spectroscopy. Mar. Chem. 51, 325e346. Coble, P.G., 2007. Marine optical biogeochemistry: the chemistry of ocean color. Chem. Rev. 107, 402e418.
guen, C., 2015. Relationships between molecular weight and fluoresCuss, C., Gue cence properties for size-fractionated dissolved organic matter from fresh and aged sources. Water Res. 68, 487e497.
guen, C., 2013. Distribution of PARAFAC modeled CDOM compoDainard, P.G., Gue nents in the North Pacific ocean, Bering, Chukchi and Beaufort seas. Mar. Chem. 157, 216e223. Davison, W., Zhang, H., 1994. In situ speciation measurements of trace components in natural waters using thin-film gels. Nature 367, 546e548. D'Andrilli, J., Foreman, C.M., Marshall, A.G., McKnight, D.M., 2013. Characterization of IHSS Pony Lake fulvic acid dissolved organic matter by electrospray ionization Fourier transform ion cyclotron resonance mass spectrometry and fluorescence spectroscopy. Org. Geochem. 65, 19e28. Denney, S., Sherwood, J., Leyden, J., 1999. In situ determinations of labile Cu, Cd and Mn in river waters using DGT. Sci. Total Environ. 239, 71e80.
ry, S.J., Stieglitz, M., McKenna, E.C., Wood, E.F., 2005. Characteristics and trends of De river discharge into Hudson, James, and Ungava bays, 1964-2000. J. Clim. 18, 2540e2557. Determann, S., Reuter, R., Willkomm, R., 1996. Fluorescent matter in the eastern Atlantic Ocean. 2. Vertical profiles and relation to water masses. Deep Sea Res. Part I 43, 345e360. Dong, Q., Li, P., Huang, Q., Abdelhafez, A.A., Chen, L., 2014. Occurrence, polarity and bioavailability of dissolved organic matter in the Huangpu River, China. J. Environ. Sci. 26, 1843e1850. Emerson, S.R., Huested, S.S., 1991. Ocean anoxia and the concentrations of molybdenum and vanadium in seawater. Mar. Chem. 34, 177e196. Evans Jr., H.T., Landergran, S., 1975. Vanadium. In: Wedepohl, K.H. (Ed.), Handbook of Geochemistry. Springer-Verlag, Berlin. Chap. 23. Frank, A., Madej, A., Galgan, V., Petersson, L.R., 1996. Vanadium poisoning of cattle with basic slag. Concentrations in tissues from poisoned animals and from a reference, slaughter-house material. Sci. Total Environ. 181, 73e92. Gimpel, J., Zhang, H., Hutchinson, W., Davison, W., 2001. Effect of solution composition, flow and deployment time on the measurement of trace metals by the diffusive gradient in thin film technique. Anal. Chim. Acta 448, 93e103. Government of Canada, 2015. Real-time Hydrometric Data Graph for Churchill River above Red Head Rapids. Retrieved on Nov 12, 2014 from. http://wateroffice.ec. gc.ca/. Gu, Q., Kenny, J.E., 2009. Improvement of inner filter effect correction based on determination of effective geometric parameters using a conventional fluorimeter. Anal. Chem. 81, 420e426. Guan, D.X., Williams, P.N., Luo, J., Zheng, J.L., Xu, H.C., Cai, C., Ma, L.Q., 2015. Novel precipitated zirconia-based DGT technique for high resolution imaging of oxyanions in waters and sediments. Environ. Sci. Technol. 49, 3653e3661.
guen, C., Guo, L., Wang, D., Tanaka, N., Hung, C.C., 2006. Chemical characteristics Gue and origin of dissolved organic matter in the Yukon River. Biogeochem 77, 139e155. Hertkorn, N., Frommberger, M., Witt, M., Koch, B.P., Schmitt-Kopplin, Ph., Perdue, E.M., 2008. Natural organic matter and the event horizon of mass spectrometry. Anal. Chem. 80, 8908e8919. €lemann, J.A., Schirmacher, M., Prange, A., 2005. Seasonal variability of trace Ho metals in the Lena River and the southeastern Laptev Sea: impact of the spring freshet. Glob. Planet. Change 48, 112e115. Hong, H., Yang, L., Guo, W., Wang, F., Yu, X., 2012. Characterization of dissolved organic matter under contrasting hydrologic regimes in a subtropical watershed using PARAFAC model. Biogeochem 109, 163e174. Hooda, P.S., Zhang, H., Davison, W., Edward, A.C., 1999. Measuring bioavailable trace metals by diffusive gradients in thin film (DGT): soil moisture effects on its performance in soils. Eur. J. Soil Sci. 50, 285e294. Hope, B.K., 1997. An assessment of the global impact of anthropogenic vanadium. Biogeochem 37, 1e13. Jeandel, C., Caisso, M., Minster, J.F., 1987. Vanadium behaviour in the global ocean and in the Mediterranean Sea. Mar. Chem. 21, 51e74. Kirk, J.L., Muir, D.C.G., Gleason, A., Wang, X., Lawson, G., Frank, R.A., Lehnherr, I., Wrona, F., 2014. Atmosphericdeposition of mercury and methylmercury to landscapes and waterbodies of the Athabasca oil sands region. Environ. Sci. Technol. 48, 7374e7383. Kothawala, D., von Wachenfeldt, E., Koehler, B., Tranvik, L., 2012. Selective loss and preservation of lake water dissolved organic matter fluorescence during longterm dark incubations. Sci. Total Environ. 433, 238e246. Kothawala, D.N., Stedmon, C.A., Muller, R.A., Weyhenmeyer, G.A., Kohler, S.J., Tranvik, L.J., 2013. Controls of dissolved organic matter quality: evidence from a large-scale boreal lake survey. Glob. Chang. Biol. 20, 1101e1114.
guen, C., Pardos, M., Dominik, J., 2003. Influence of humic substances Koukal, B., Gue on the toxic effects of cadmium and zinc to the green alga Pseudokirchneriella subcapitata. Chemosphere 53, 953e961.

374

Y.X. Shi et al. / Chemosphere 154 (2016) 367e374


, P., Reinhardt, A., Ferrari, B., Wilkinson, K.J., Loizeau, J.L., Koukal, B., Rosse Dominik, J., 2007. Effect of Pseudokirchneriella subcapitata (Chlorophyceae) exudates on metal toxicity and colloid aggregation. Water Res. 41, 63e70. Kowalczuk, P., Durako, M.J., Young, H., Kahn, A.E., Cooper, W.J., Gonsior, M., 2009. Characterization of dissolved organic matter fluorescence in the South Atlantic Bight with use of PARAFAC model: interannual variability. Mar. Chem. 113, 182e196. Kuzyk, Z.A., Macdonald, R.W., Granskog, M.A., Scharien, R.K., Galley, R.J., Michel, C., Barber, D., Stern, G., 2008. Sea ice, hydrological, and biological processes in the Churchill River estuary region, Hudson Bay. Estuar. Coast Shelf Sci. 77, 369e384. Lawaetz, A.J., Stedmon, C.A., 2009. Fluorescence intensity calibration using the Raman scatter peak of water. Appl. Spec. 63, 936e940. Li, Y.-H., Gregory, S., 1974. Diffusion of ions in sea water and in deep-sea sediments. Geochim. Cosmochim. Acta 38, 703e714. Lu, X., Johnson, W.D., Hook, J., 1998. Reaction of vanadate with aquatic humic substances: an ESR and 51V NMR study. Environ. Sci. Technol. 32, 2257e2263. Luo, J., Zhang, H., Santner, J., Davison, W., 2010. Performance characteristics of diffusive gradients in thin films equipped with a binding gel layer containing precipitated ferrihydrite for measuring arsenic(V), selenium(VI), vanadium(V), and antimony(V). Anal. Chem. 82, 8903e8909. Ma, Z., Fu, Q., 2009. Comparison of hypoglycemic activity and toxicity of vanadium (IV) and vanadium (V) absorbed in fermented mushroom of coprinus comatus. Biol. Trace Elem. Res. 132, 278e284.
guen, C., 2015. Examining concentrations and molecular weights of Mangal, V., Gue thiols in microorganism cultures and in Churchill River (Manitoba) using a fluorescent-labeling method coupled to asymmetrical flow field-flow fractionation. Anal. Bioanal. Chem. 407, 4305e4313.
guen, C., 2016. Molecular characterization of phytoMangal, V., Stock, N., Gue plankton dissolved organic matter and sulfur components using high resolution Orbitrap mass spectrometry. Anal. Bioanal. Chem. http://dx.doi.org/10.1007/ s00216-015-9295-9. Mann, P.J., Davydova, A., Zimov, N., Spencer, R.G.M., Davydov, S., Bulygina, E., Zimov, S., Holmes, R.M., 2012. Controls on the composition and lability of dissolved organica matter in Siberia's Kolyma River basin. J. Geophys. Res. http:// dx.doi.org/10.1029/2011JG001798. McBride, M.B., Cherney, J., 2004. Molybdenum, sulfur, and other trace elements in farm soils and forages after sewage sludge application. Commun. Soil Sci. Plant Anal. 35, 517e535. Moskalyk, R.R., Alfanti, A.M., 2003. Processing of vanadium: a review. Min. Eng. 16, 793e805. Mostofa, K.M.G., Wu, F., Liu, C., Fang, W.L., Yuan, J., Ying, W.L., Wen, L., Yi, M., 2010. Characterization of Nanming River (southwestern China) sewerage-impacted pollution using an excitation-emission matrix and PARAFAC. Limnology 11, 217e231. Murphy, K.R., Ruiz, G.M., Dunsmuir, W.T.M., Waite, T.D., 2006. Optimized parameters for fluorescence-based verification of ballast water exchange by ships. Environ. Sci. Technol. 40, 2357e2362. Murphy, K.R., Stedmon, C.A., Waite, T.D., Ruiz, G.M., 2008. Distinguishing between terrestrial and autochthonous organic matter sources in marine environments using fluorescence spectroscopy. Mar. Chem. 108, 40e58. Murphy, K.R., Hambly, A., Singh, S., Henderson, R.K., Baker, A., Stuetz, R., Khan, S.J., 2011. Organic matter fluorescence in municipal water recycling schemes: toward a unified PARAFAC model. Environ. Sci. Technol. 45, 2909e2916. Murphy, K.R., Stedmon, C.A., Graeber, D., Bro, R., 2013. Fluorescence spectroscopy and multi-way techniques. PARAFAC. Anal. Methods 5, 6557e6566. Murphy, K.R., Stedmon, C.A., Wenig, P., Bro, R., 2014. OpenFluor e an online spectral library of auto-fluorescence by organic compounds in the environment. Anal. Methods 6, 658e661. Niencheski, L.F., Windom, H.L., Moore, W.S., 2014. Controls on water column chemistry of the southern Brazilian continental shelf. Cont. Shelf Res. 88, 126e139. Ohno, T, Ohno, P.E., 2013. Influence of heteroatom pre-selection on the molecular formula assignment of soil organic matter components determined by ultrahigh resolution mass spectrometry. Anal. Bioanal. Chem. 405, 3299e3306. Osburn, C.L., Stedmon, C.A., 2011. Linking the chemical and optical properties of dissolved organic matter in the Baltic-North Sea transition zone to differentiate three allochthonous inputs. Mar. Chem. 126, 281e294. Osburn, C.L., Wigdahl, C.R., Fritz, S.C., Saros, J.E., 2011. Dissolved organic matter composition and photoreactivity in prairie lakes of the U.S. Great Plains. Limnol. Oceanogr. 56, 2371e2390. € Osterlund, H., Chlot, S., Faarinen, M., Widerlund, A., Rodushkin, I., Ingri, J., Baxter, D.C., 2010. Simultaneous measurements of As, Mo, Sb, V and Wusing a ferrihydrite diffusive gradients in thin films (DGT) device. Anal. Chim. Acta 682, 59e65. Panther, J.G., Stewart, R.R., Peter, R.T., Bennett, W.W., Welsh, D.T., Zhao, H., 2013.

Titanium dioxide-based DGT for measuring dissolved As(V), V(V), Sb(V), Mo(VI) and W(VI) in water. Talanta 105, 80e86. Parsons, T.R., Webb, D.G., Dovey, H., Haigh, R., Lawrence, M., Hopky, G.E., 1988. Production studies in the Mackenzie river Beaufort sea estuary. Polar Biol. 8, 235e239. Price, H.L., Teasdale, P.R., Jolley, D.F., 2013. An evaluation of ferrihydrite- and Metsorb-DGT techniques for measuring oxyanion species (As, Se, V, P): effective capacity, competition and diffusion coefficients. Anal. Chim. Acta 803, 56e65. Ramani, S., Dragun, Z., Kapetanovic, D., Kostov, V., Jordanova, M., Erk, M., HajrulaiMusliu, Z., 2014. Surface water characterization of three rivers in the lead/zinc mining region of Northeastern Macedonia. Arch. Environ. Contam. Toxicol. 66, 514e528. Rühling, A., Tyler, G., 2001. Changes in atmospheric deposition rates of heavy metals in Sweden. A summary of Nationwide Swedish Surveys in 1968/70-1995. Water Air Soil Poll. 1, 311e323. Sabbioni, E., Pozzi, G., Pintar, A., Casella, L., Garattini, S., 1991. Cellular retention, cytotoxicity and morphological transformation by vanadium(IV) and vanadium(V) in BALB/3T3 cell lines. Carcinogenesis 12, 47e52. Salice, V.C., Cortizo, A.M., Gomez Dumm, C.L., Etcheverry, S.B., 1999. Tyrosine phosphorylation and morphological transformation induced by four vanadium compounds on MC3T3E1 cells. Mol. Cell Biochem. 198, 119e128. Shiller, A.M., Boyle, E.A., 1987. Dissolved vanadium in rivers and estuaries. Earth Planet. Sci. Lett. 86, 214e224. Shiller, A.M., Mao, L., 1999. Dissolved vanadium on the Louisiana Shelf: effect of oxygen depletion. Cont. Shelf Res. 19, 1007e1020. Shiller, A.M., Mao, L., 2000. Dissolved vanadium in rivers: effects of silicate weathering. Chem. Geol. 165, 13e22. Shutova, Y., Baker, A., Bridgeman, J., Henderson, R.K., 2014. Spectroscopic characterisation of dissolved organic matter changes in drinking water treatment: from PARAFAC analysis to online monitoring wavelengths. Water Res. 54, 159e169. Søndergaard, M., Borch, N.H., Stedmon, C.A., 2003. Fate of terrigenous dissolved organic matter (DOM) in estuaries: aggregation and bioavailability. Ophelia 57, 161e176. Stedmon, C.A., Bro, R., 2008. Characterizing dissolved organic matter fluorescence with parallel factor analysis: a tutorial. Limnol. Oceanogr. Methods 6, 572e579. Stedmon, C.A., Markager, S., Bro, R., 2003. Tracing dissolved organic matter in aquatic environments using a new approach to fluorescence spectroscopy. Mar. Chem. 82, 239e254. Stedmon, C.A., Markager, S., Tranvik, L., Kronberg, L., Slatis, T., Martinsen, W., 2007. Photochemical production of ammonium and transformation of dissolved organic matter in the Baltic Sea. Mar. Chem. 104, 227e240. Stedmon, C.A., Amon, R.M.W., Rinehart, A.J., Walker, S.A., 2011. The supply and characteristics of colored dissolved organic matter (CDOM) in the Arctic Ocean: Pan Arctic trends and differences. Mar. Chem. 124, 108e118. Stern, A., Yin, X.F., Tsang, S.S., Davison, A., Moon, J., 1993. Vanadium as a modulator of cellular regulatory cascades and oncogene expression. Biochem. Cell Biol. 71, 103. Taylor, H.E., Antweiler, R.C., Roth, D.A., Alpers, C.N., Dileanis, P., 2012. Selected trace elements in the Sacramento River, California: occurrence and distribution. Arch. Environ. Contam. Toxicol. 62, 557e569. Walker, S.A., Amon, R.M.W., Stedmon, C., Duan, S., Louchouarn, P., 2009. The use of PARAFAC modeling to trace terrestrial DOM and fingerprint water masses in coastal Canadian Arctic surface waters. J. Geophys. Res. Biogeosci. 114, G00F06. http://dx.doi.org/10.1029/2009JG000990. Walker, S.A., Amon, R.M.W., Stedmon, C.A., 2013. Variations in high-latitude riverine fluorescent dissolved organic matter: a comparison of large Arctic rivers. J. Geophys. Res. Biogeosci. 118, 1689e1702.
, 2010a. FluoresYamashita, Y., Cory, R.M., Nishioka, H., Kuma, K., Tanoue, E., Jaffe cence characteristics of dissolved organic matter in the deep waters of the Okhotsk Sea and the northwestern North Pacific Ocean. Deep sea Res. II 57, 1478e1485. Yamashita, Y., Scinto, L.J., Maie, N., Jaffe, R., 2010b. Dissolved organic matter characteristics across a subtropical wetland's landscape: application of optical properties in the assessment of environmental dynamics. Ecosystems 13, 1006e1019. Yuan, D., Guo, N., Guo, X., Zhu, N., Chen, L., He, L., 2014. The spectral characteristics of dissolved organic matter from sediments in Lake Baiyangdian. North China. J. Gt. Lakes. Res. 40, 684e691. Zhang, H., Davison, W., 1995. Performance characteristics of the technique of diffusion gradients in thin-films (DGT) for the measurement of trace metals in aqueous solution. Anal. Chem. 67, 3391e3400. Zhang, H., Davison, W., 1999. Diffusional characteristics of hydrogels used in DGT and DET techniques. Anal. Chim. Acta 398, 329e340.