DSpace at VNU: Influence of salinity intrusion on the speciation and partitioning of mercury in the Mekong River Delta

DSpace at VNU: Influence of salinity intrusion on the speciation and partitioning of mercury in the Mekong River Delta t...

Influence of salinity intrusion on the speciation and partitioning

of mercury in the Mekong River Delta

Nguyen Thi Van Hac, Suthipong Sthiannopkaod, Seunghee Hana,⇑

a School of Environmental Science and Engineering, Gwangju Institute of Science and Technology (GIST), Gwangju 500-712, Republic of Korea

b Faculty of Environment, Ho Chi Minh City University of Technology, District 10, Ho Chi Minh City, Viet Nam

c Faculty of Environment, Ho Chi Minh City University of Natural Resources and Environment, Tan Binh District, Ho Chi Minh City, Viet Nam

d Department of Environmental Engineering, Dong-A University, Busan 604-714, Republic of Korea Received 26 February 2012; accepted in revised form 11 December 2012; available online 29 December 2012

Abstract

The lower Mekong and Saigon River Basins are dominated by distinctive monsoon seasons, dry and rainy seasons Most

of the Mekong River is a freshwater region during the rainy season, whereas during the dry season, salt water intrudes approximately 70 km inland To understand the role of salinity intrusion controlling Hg behavior in the Mekong and Saigon River Basins, Hg and monomethylmercury (MMHg) in surface water and sediment of the Mekong River and in sediment of the Saigon River were investigated in the dry season Sediment Hg distribution, ranging from 0.12 to 0.76 nmol g1, was mainly controlled by organic carbon distribution in the Mekong River; however, the location of point sources was more important in the Saigon River (0.21–0.65 nmol g1) The MMHg concentrations in Mekong (0.16–6.1 pmol g1) and Saigon (0.70–8.7 pmol g1) sediment typically showed significant increases in the estuarine head, with sharp increases of acid volatile sulfide Unfiltered Hg (4.6–222 pM) and filtered Hg (1.2–14 pM) in the Mekong River increased in the estuarine zone due to enhanced particle loads Conversely, unfiltered MMHg (0.056–0.39 pM) and filtered MMHg (0.020–0.17 pM) was similar between freshwater and estuarine zones, which was associated with mixing dilution of particulate MMHg by organic- and MMHg-depleted resuspended sediment Partitioning of Hg between water and suspended particle showed tight correlation with the partitioning of organic carbon across study sites, while that of MMHg implied influences of chloride: enhanced chlo-ride in addition to organic matter depletion decreased particulate MMHg in the estuarine zone Primary production was an important determinant of inter-annual variation of particulate Hg and sediment MMHg The bloom year showed relatively low particulate Hg with low C/N ratio, indicating biodilution of Hg In contrast, the percentage of MMHg in sediment increased significantly in the bloom year, likely due to greater availability of metabolizable fresh organic matter The overall results emphasize that Hg behavior in the lower Mekong River Basin is strongly connected to the local monsoon climate, via alterations in particle loads, biological productivity, and availability of sulfate, chloride and organic matter

Ó 2012 Elsevier Ltd All rights reserved

1 INTRODUCTION Interactions between Hg and organic matter influence

the transport and bioavailability of Hg in riverine and

estuarine environments (Laurier et al., 2003; Choe et al., 2003; Conaway et al., 2003; Han et al., 2007; Lee et al.,

2011) Dissolved Hg distribution across watersheds is ex-plained by dissolved organic carbon (DOC) distribution

in a number of river systems (Peckenham et al., 2003; Schuster et al., 2008; Brigham et al., 2009) In estuarine systems, complexation of Hg with dissolved organic mat-ter, coupled with colloidal coagulation, is reported to influence estuarine mixing behavior (Stordal et al., 1996; 0016-7037/$ - see front matter Ó 2012 Elsevier Ltd All rights reserved.

http://dx.doi.org/10.1016/j.gca.2012.12.018

⇑ Corresponding author Tel.: +82 62 7152438; fax: +82 62

7152434.

E-mail address: shan@gist.ac.kr (S Han).

www.elsevier.com/locate/gca Geochimica et Cosmochimica Acta 106 (2013) 379–390

Choe et al., 2003; Lee et al., 2011) and bioavailability of

Hg (Pan and Wang, 2004; Zhong and Wang, 2009)

Monomethylmercury (MMHg) is a toxic form of Hg that

biomagnifies in aquatic food chains (Chen et al., 2008;

Gantner et al., 2009) Previous studies note that

biogeo-chemical factors (e.g., sulfate, sulfide, and organic matter

contents) are critical in the production of MMHg in

estu-arine sediments (Hammerschmidt et al., 2004;

Hammers-chmidt and Fitzgerald, 2006; Han et al., 2007) For

example, increasing Hg(II) methylation rate along with

increasing salinity is shown in estuarine sediments to be

associated with sulfate availability to sulfate-reducing

bac-teria (Hollweg et al., 2009)

The Mekong River originates at an elevation of about

5100 m in the Tibetan Plateau of western China and flows

southward through China, Myanmar, Laos, Thailand,

Cambodia, and Vietnam, before discharging into the

South China Sea (Edwin, 2009) It has a total length of

4800 km and drains an area of 795,000 km2, with a mean

annual water discharge of 470 km3yr1, making it the

lon-gest river in Southeast Asia (Dai and Trenberth, 2002)

The river flows over rock for about 80% of its length

be-fore it enters the alluvial plain of Cambodia and Vietnam

(Xue et al., 2010, 2011) The Mekong River Delta in

southern Vietnam is composed of Holocene alluvial

sedi-ments that were rapidly deposited beginning 8000 yr BP

(144 106ton yr1; Xue et al., 2010, 2011) The delta is

an essential component of life for millions of South Asians

residing in the vicinity of the river and its tributaries,

pro-viding the main freshwater source for urban, industrial,

agricultural, and fishery uses (Osborne, 2000; Baran

et al., 2007) Despite this importance, surprisingly little

information is available regarding Hg levels in the

Me-kong River as compared to other major rivers (e.g., the

Amazon and Nile Rivers)

The Saigon River originates from Phum Daung in

southeastern Cambodia, flows south for approximately

225 km, and empties into the Nha Be River, which

dis-charges into the South China Sea, 20 km north of the

Me-kong Delta (Lambert et al., 2010) The Saigon River is

joined by the Ben Cat River just upstream of Ho Chi Minh

City, and it encircles Ho Chi Minh City, the largest city in

Vietnam Since the 1980s, Ho Chi Minh City has

experi-enced continually increasing population density and rapid

industrial growth, which have negatively affected the

envi-ronmental quality of the urban river (Thuy et al., 2007; Vo,

2007) Large volumes of untreated domestic and industrial

wastewater, along with substances from accidental spills,

are released directly into the river (Minh et al., 2007)

Monitoring of metals pollution in the Saigon River

indi-cates that surface river-water and sediments are severely

contaminated with Cd, Cr, Cu and Zn (Thuy et al.,

2007); nevertheless, no literature is available regarding

lev-els of Hg pollution in the water and sediment of the Saigon

River

The lower Mekong River and Saigon River Basins are

dominated by distinctive monsoon seasons The dry season

occurs from November to April, with an average discharge

rate of 2000 m3s1, and the rainy season occurs from May

to October, with an average discharge rate of 40,000 m3s1

for the Mekong River (Hoa et al., 2007) While most of the Mekong Delta is a freshwater region during the rainy sea-son, salt water intrudes approximately 70 km inland in the dry season, with vertical stratification in salinity (Wolanski et al., 1996, 1998; Cenci and Martin, 2004) In the wet season, salinity intrusion is observed near the river mouth, but this does not extend more than a few kilometers inland (Wolanski et al., 1996) In contrast to the Mekong River, the Saigon River has a constant flow rate (54 m3s1; People’s Committee of Ho Chi Minh City, 2006) throughout the year, due to flow regulation from the Dau Tieng Reservoir (George et al., 2004) Salinity intrusion is observed from the lower Saigon River, approx-imately 10 km inland from the Nha Be River (www.eng.hochiminhcity.gov.vn)

Salinity intrusion influences the speciation and bioavail-ability of Hg in low-discharge rivers (Chiffoleau et al., 1994; Laurier et al., 2003; Covelli et al., 2006) Like per-manent estuaries, interactions between Hg and organic matter have primary importance in the transport and bio-availability of Hg in these rivers (Laurier et al., 2003; Turner et al., 2004) Hg–chloro complexation could in-crease the mobility of Hg, even though the majority of

Hg may remain in organic fraction (Ramalhosa et al.,

2005) Salinity intrusion is often accompanied by forma-tion of a high turbidity zone as a result of intense hydro-dynamic energy of tidal currents (Laurier et al., 2003; Covelli et al., 2006) In a high turbidity zone, the quantity and quality of particulate organic matter influence redistri-bution of Hg between water and suspended particles, which may cause changes in Hg solubility and bioavail-ability (Laurier et al., 2003; Turner et al., 2004; Covelli

et al., 2006; Cana´rio et al., 2008) A high turbidity zone often provides an ideal site for diagenetic processes of trace metals, in association with increased microbial activ-ity and flocculation of colloidal particles (Roth et al., 2001; Laurier et al., 2003; Covelli et al., 2006) Increased MMHg concentration in the high turbidity zone at the mouth of the Isonzo River is considered to result from in-tense microbial methylation in the low-oxygen bottom-water (Covelli et al., 2006)

In the current study, we aim to understand the role of salinity intrusion on Hg speciation in riverine water and sediment We hypothesize that: (1) salinity intrusion in-creases sediment MMHg production due to the enhanced availability of sulfate to Hg(II)-methylating microbes; (2) salinity intrusion increases particle solubility of Hg and MMHg due to chloride complexation and/or particulate organic matter dilution by resuspended sediment; and (3) salinity intrusion increases water column MMHg concen-tration, as a result of intense microbial activity in the high turbidity zone To test these hypotheses, we examine Hg and MMHg distributions in surface river water, suspended particles, and sediments in relation to relevant biogeo-chemical variables (e.g., salinity, suspended particulate matter (SPM), sulfate, chlorophyll-a (Chl-a), DOC, partic-ulate organic carbon (POC), and particpartic-ulate nitrogen (PN)); and of sediment (e.g., total organic carbon (TOC), total nitrogen (TN) and acid volatile sulfide (AVS))

380 S Noh et al / Geochimica et Cosmochimica Acta 106 (2013) 379–390

2 MATERIALS AND METHODS

2.1 Sample collection and pre-treatment

Water and sediment samples were collected from 14 sites

in March 2010 and seven sites in April 2011 along the Tien

(Fig 1), and from eight sites along the Saigon River in

March 2010 (Fig 2) Ultra-clean sample handling protocols

were used to prevent sample contamination Unfiltered

sur-face water samples were collected from a depth of

approx-imately 0.5 m in acid-cleaned Teflon bottles, using a

peristaltic pump system equipped with Teflon tubing

Fil-tered surface water samples were collected using the same

method with polyethersulfone filter capsules (MilliporeÒ,

0.45 lm) connected to the tubing outlet At each site, three

independent samples were collected in Teflon, polyethylene,

and borosilicate glass bottles The Teflon bottles of

unfil-tered and filunfil-tered water for determination of Hg and

MMHg were preserved with high-purity HCl (0.2% v/v),

and stored at 4°C within 12 h Polyethylene bottles of

unfil-tered water were kept at 4°C for measurements of SPM,

POC, PN, alumina, and chlorophyll-a concentrations

Borosilicate glass vials of filtered water were collected for

determination of DOC and sulfate concentration Filtered

water samples were collected only from the Tien River

The ancillary parameters, e.g., pH, temperature, salinity,

and oxidation–reduction potential (ORP), were recorded

in situ using a multi-parameter probe (Thermo Scientific)

The unfiltered water samples for SPM, POC, PN,

alu-mina, and chlorophyll-a were filtered in the laboratory

within 24 h of sampling Glass fiber filters (WhatmanÒ,

GF/F) were used for particle collection for chlorophyll-a,

alumina, POC, and PN measurements, and 0.4 lm

polycar-bonate membranes were used for SPM measurements The

filters containing particles were dried immediately for mea-surement of SPM and alumina, and filters for

chlorophyll-a, POC, and PN were frozen at20 °C until analysis Sed-iment samples were collected from the top-20 cm layer

Fig 1 Map of sampling sites in Tien River, Vietnam.

Fig 2 Map of sampling sites in Saigon River, Vietnam.

using a stainless steel grab sampler to determine Hg,

MMHg, and TOC; and TN and AVS concentrations The

collected sediments were stored in sealed polyethylene bags

and frozen at20 °C until analysis

2.2 Hg and MMHg analysis

The Hg in water samples was analyzed followingEPA

method 1631 (2002) Aqueous samples were oxidized with

bromine monochloride (BrCl) for a minimum of 12 h After

oxidation, excess oxidant was destroyed with

hydroxyl-amine hydrochloride solution prior to analysis Divalent

Hg in these samples was reduced to elemental Hg by SnCl2

solution and the elemental Hg was trapped on gold traps

The Hg0 released from gold traps by thermal desorption

was fed into a cold vapor atomic fluorescence spectrometer

(CVAFS; Model III, Brooks Rand) An acceptable

calibra-tion curve was achieved daily with an r2of at least 0.99 The

relative differences of duplicate analyses averaged 11 ± 8%

(n = 144) and recovery of the matrix spike averaged

98 ± 25% (n = 7) Recovery of certified reference material

(European Commission, BCRÒ-579, coastal seawater,

1.9 ± 0.5 ng kg1) averaged 105 ± 6% (n = 10)

Details of the analytical protocols for MMHg analysis

in an aqueous phase were as given by EPA method 1630

(2001) Acidified water samples were distilled with

ammo-nium pyrrolidine dithiocarbamate (APDC) solution and

the distillates were then converted to gaseous MMHg by

aqueous-phase ethylation using tetraethylborate solution

The volatile MMHg was then purged and trapped onto

Te-naxÒtraps, which were flash-heated in a nitrogen stream

The released Hg species were thermally separated on a

GC column, then detected by CVAFS (Model III, Brooks

Rand) An acceptable calibration curve was achieved daily

with an r2of at least 0.99 The relative difference of

dupli-cate analyses was typically 18 ± 5% (n = 52), whereas the

matrix spike and certified reference material

(ERM-CE464, IRMM, 5.5 ± 0.2 mg kg1) recovery averaged

99 ± 6% (n = 12) and 99 ± 6% (n = 12), respectively

The concentrations of particulate Hg (PHg) and

partic-ulate MMHg (PMMHg) were determined from the

differ-ences between the unfiltered Hg (or MMHg) and filtered

Hg (or MMHg), normalized to the SPM concentration

The Hg found in sediment was analyzed following the

appendix toEPA method 1631 (2001) The Hg in sediment

(0.5–1.5 g) was digested overnight with 10 ml of aqua regia

Sediment digests were then diluted with 0.07 N BrCl

solu-tion and used for Hg determinasolu-tion, using the same method

as for aqueous Hg samples The relative difference of

duplicate analyses was typically 16 ± 24% (n = 48), whereas

certified reference material (ERM-CC580, IRMM, 132 ±

3 mg kg1) recovery ranges were 92 ± 6% (n = 12)

The MMHg in sediment was analyzed following the

pro-cedure described by Choe and Gill (2003) The sediment

MMHg (0.5–1.0 g) was digested with acidic potassium

bro-mide solution and extracted into methylene chloride An

aliquot of methylene chloride was then back-extracted to

Milli-Q water by purging out methylene chloride with

nitro-gen gas The extracted MMHg was measured using the

same method described for aqueous samples The relative

difference of duplicate analyses was typically 15 ± 14% (n = 73), whereas certified reference material (ERM-CC580, IRMM, 75 ± 4 lg kg1) recovery averaged

96 ± 8% (n = 8)

2.3 Particle and sediment compositions The SPM was measured by filtering 0.2 L of water through a pre-weighed polycarbonate membrane (What-manÒ, 0.4 lm) shortly after sample collection Particles were dried at 60°C for 12 h, and the SPM loads were then weighed For determination of organic carbon and nitrogen

in SPM and sediment, the GF/F filters bearing the SPM and approximately 10 mg of sediment were freeze-dried Freeze-dried samples were acidified with HCl solution to re-move inorganic carbon, then measured with an elemental analyzer interfaced to a continuous flow isotope ratio mass spectrometer (EA-IRMS, ThermoQuest) Alumina on fil-ters was analyzed with a wavelength dispersive X-ray fluo-rescence spectrometer (Axios Mineral, PANalytical) The samples for chlorophyll-a were obtained by filtering 0.1– 0.4 L of water through GF/F filters Phytoplankton pig-ments retained on the GF/F filters were extracted in 90% acetone (Liu et al., 2007) The chlorophyll-a in the extracts was measured at 750, 665, 645, 630 and 480 nm wave-lengths with an ultraviolet–visible spectrophotometer (Opt-izen, Mecasys) DOC and sulfate concentrations in filtered water samples were measured by TOC analyzer (Multi N/C

3100, Analytik Jena) and ion chromatography (DX-120, Dionex), respectively For determination of AVS, approxi-mately 10 g of sediment sample was acidified with 20 ml of 6.0 M deoxygenated HCl and held under a nitrogen gas flow for 2 h Volatilized sulfide was collected in traps filled with sulfide antioxidant buffer (SAOB; EPA, 1996), then measured with a sulfide electrode (Kim et al., 2006) Sigma PlotÒ, version 11.2 (Systat software, Inc.) was used for all statistical analyses Linear regression analysis yielded the coefficient of determination (r2) and significant probability (p) Differences between two independent groups were determined using one-way ANOVA All statis-tical results were reported as significant at a level of

p < 0.05 Linear correlation analyses yielded the Pearson’s correlation coefficient (r) for parameters passing the nor-mality test (Shapiro–Wilks nornor-mality test) The results were considered statistically significant if p-values were less than 0.05

3 RESULTS 3.1 Geochemical settings

In the Tien River, the mean water temperature across sampling sites was 31 ± 0.9°C in 2010 and 30 ± 0.8 °C in

2011, and mean pH was 7.8 ± 0.1 in 2010 and 7.4 ± 0.3

in 2011 Saline intrusion was observed from 0 to 50 km from the coast in 2010 and from 0 to 20 km in 2011 ( Sup-porting Information, Table S1) In the brackish zone of the Tien River, the residence time of saline bottom-water can be quite long, as indicated by lower ORP High concen-trations of SPM (>80 mg L1) were found from the

brack-382 S Noh et al / Geochimica et Cosmochimica Acta 106 (2013) 379–390

ish zone, and were associated with hydrodynamic energy of

tidal currents: the Mekong River Delta has 3.5-m

seminal tides from the South China Sea and irregular 1-m

diur-nal tides from the Gulf of Thailand (Hoa et al., 2007) In

the Saigon River, the mean water temperature and pH for

2010 were 30 ± 0.9°C and 7.7 ± 0.3, respectively, and

sal-ine intrusion was observed up to 10 km from the Nha Be

River (Supporting Information, Table S2) Like the Tien

River, the brackish zone showed lower ORP than the

fresh-water zone, and a high turbidity zone was found at sites 7

and 8

Concentration and compositional characteristics of the

SPM (i.e., alumina, POC, PN, and atomic C/N ratio) were

compared between freshwater and brackish zones of the

Tien River (Table 1) Alumina content, used as a proxy

for fine particles, significantly (p < 0.05) increased in the

saline zone, supporting the periodic occurrence of sediment

resuspension Levels of POC varied from 1.6% to 9.3%

(mean 4.0 ± 2.4%) in 2010 and from 1.7% to 4.8% (mean

3.5 ± 1.2%) in 2011 POC decreased in the brackish zone

in both years, implying dilution of fluvial particles by

or-ganic-depleted resuspended sediment The PN showed a

similar trend to the POC The strong correlation between

POC and PN (r2= 0.95, p < 0.05, linear regression)

sug-gests that N is predominantly bound to particulate organic

matter

3.2 Sediment characteristics

In the Tien River, TOC in sediment ranged from 2.2% to

3.9% (mean 2.7 ± 0.58%) in 2010, and from 1.2% to 2.4%

(mean 1.7 ± 0.5%) in 2011 In the Saigon River, TOC in

sediment ranged from 2.6% to 3.7% (mean 3.2 ± 0.32%)

(Table 2) Little lateral variation of TOC, TN, and C/N

ra-tios were noted in the Tien and Saigon Rivers (Supporting

Information, Table S3) Temporally, atomic C/N ratios in

the Tien River were significantly (p < 0.05) lower in 2011

(range 3.4–6.5) than in 2010 (range 9.6–15)

AVS in sediment of the Tien River ranged from 0.053 to

6.5 lmol g1in 2010 and from 0.20 to 20 lmol g1in 2011

(Table 2) Mean AVS concentration was significantly

high-er (p < 0.05) in the brackish zone compared to the

freshwa-ter zone in 2010 The same distribution patfreshwa-tern was found

in the Saigon River, where AVS ranged from 0.14 to

0.45 lmol g1 in the freshwater zone, and from 5.7 to

17 lmol g1 in the brackish zone (p < 0.05) The primary

factors that control AVS in river and estuarine sediments

are the availabilities of reactive Fe(II), dissolved sulfate,

and metabolizable organic carbon (Morse et al., 2007) As

the concentrations of TOC and TN, and C/N ratio were

consistent across sites, increased AVS at the saline zone

could be attributable to enhanced sulfate availability

3.3 Total Hg and MMHg in sediment

The Hg concentrations found in Tien River sediment

ranged from 0.32 to 0.76 nmol g1 dry weight (mean

0.50 ± 0.14 nmol g1) in 2010 and from 0.12 to

0.24 nmol g1dry weight (mean 0.18 ± 0.040 nmol g1) in

2011 (Table 2); in the Saigon River, it ranged from 0.21 Ta

ORP (mV)

SPM (mg

1 ) Alumina (mmol

1 )

1 ) POC (%)

PN (%)

C/N (mo

1 )

C (lgL

1 )

to 0.65 nmol g1 (mean 0.44 ± 0.16 nmol g1) These ranges are comparable to those reported for urban rivers and estuaries moderately contaminated with Hg: the Patux-ent River (0.29–0.79 nmol g1;Benoit et al., 1998) and Bay

of Fundy (0.050–0.70 nmol g1; Sunderland et al., 2006) There was no statistical (p > 0.05) difference between the freshwater and saline zones in the Tien River, with weak in-creases at sites 9 and 10, located close to My Tho City ( Sup-porting Information, Table S3) The Saigon River also showed relatively constant sediment Hg, with weak peaks

at sites 1 and 7, located near Dau Tieng Reservoir and

Ho Chi Minh City, respectively

In the Tien River, sediment MMHg ranged from 0.16 to 6.1 pmol g1dry weight (mean 1.2 ± 1.6 pmol g1) in 2010, and from 0.57 to 5.2 pmol g1 dry weight (mean 1.8 ± 1.6 pmol g1) in 2011 (Table 2) In the Saigon River, sediment MMHg ranged from 0.70 to 8.7 pmol g1(mean 3.7 ± 2.5 pmol g1) in 2010, comparable to data reported for the Patuxent River (0.49–4.0 pmol g1; Benoit et al.,

1998), the Bay of Fundy (0.25–7.38 pmol g1;Sunderland

et al., 2006), and San Francisco Bay (0.5–5.0 pmol g1;

Conaway et al., 2003) In the Tien River, strong MMHg peaks were found near the estuarine head (sites 9, 10, and 18;Supporting Information, Table S3) In the Saigon

Riv-er, MMHg concentrations were significantly (p < 0.05) higher in the brackish zone than in the freshwater zone, with a strong peak near the estuarine head, like the Tien (site 4)

3.4 Total Hg and MMHg in surface water

In the Tien River, Hg levels in unfiltered river water (UHg) ranged from 11 to 222 pM (mean 61 ± 65 pM) in

2010 and from 4.6 to 55 pM (mean 22 ± 23 pM) in 2011 (Fig 3;Supporting Information, Table S4) These concen-tration ranges are similar or lower than those reported for urban rivers in China and urban estuaries in North Amer-ica moderately contaminated with Hg: the Yalujiang River (154–344 pM; Zhang and Wong, 2007), East River (55–

244 pM; Liu et al., 2012), New York/New Jersey Harbor (30–550 pM;Balcom et al., 2008), and San Francisco Bay (0.73–440 pM;Conaway et al., 2003) Increased UHg levels were found from the saline zone in both years In the fresh-water zone, a maximum peak of UHg was detected at site 6, which might be a local runoff effect from Cao Lanh City Temporally, UHg was about four times higher in 2010 than

2011 in the lower-river and estuarine zone DHg averaged 20% of UHg (4.7–44%), and spatial and temporal varia-tions of DHg were similar to those of UHg PHg levels were decreased in the lower-river and estuarine zone compared

to the upper river (>100 km from coast), a major contrast

to the UHg and DHg distributions Temporally, PHg was higher in 2010 than in 2011

In the Tien River, MMHg concentrations in unfiltered water (UMMHg) ranged from 0.056 to 0.39 pM (mean 0.20 ± 0.085 pM) in 2010 and 0.14 to 0.28 pM (mean 0.20 ± 0.047 pM) in 2011 (Fig 3;Supporting Information, Table S4) These ranges are within the range previously re-ported for Hg-contaminated urban rivers and estuaries: Patuxent River (0.05–1.2 pM;Benoit et al., 1998) and San

TOC (%)

TN (%)

C/N (mo

AVS (lmol

1 )

Hg (nmol

1 )

MMHg (pmol

1 )

384 S Noh et al / Geochimica et Cosmochimica Acta 106 (2013) 379–390

Francisco Bay (0.050–2.3 pM;Conaway et al., 2003)

Ex-cept for the steep decrease in the upper river, there was

no significant spatial variation for UMMHg across study

sites and there were no significant (p > 0.05) differences

be-tween 2010 and 2011 Dissolved MMHg (DMMHg) in the

Tien River ranged from 0.020 to 0.17 pM (mean 0.090 ± 0.052 pM) in 2010 and from 0.066 to 0.094 pM (mean 0.079 ± 0.010 pM) in 2011 In general, DMMHg was 19% to 78% of UMMHg (mean 45 ± 18%) and spatial and temporal variations in DMMHg followed those of UMMHg There was a distinct spatial variation for partic-ulate MMHg (PMMHg), which was significantly lower (p < 0.05) in the saline zone compared to the upper- and lower-river zones

4 DISCUSSION 4.1 Variation of sediment Hg

In the Tien River, TOC was a major factor governing lateral and temporal distribution of sediment Hg (Table 3,

Fig 4) A similar correlation was found in other rivers and estuaries (Hammerschmidt and Fitzgerald, 2004; Heyes

et al., 2006; Hollweg et al., 2009); this may be attributed

to strong affinity between Hg and thiolic binding sites on TOC (Ravichandran, 2004; Skyllberg et al., 2007) No lin-ear correlation was found between TOC and sediment Hg for the Saigon River, with a relatively narrow range of TOC Point source location might be more important for the Saigon River, since relatively high sediment Hg levels were found from the runoff sites of Dau Tieng Reservoir and Ho Chi Minh City Increased metal concentrations (e.g., Cd, Cr, Cu, and Zn) in urban parts of the Saigon

Riv-er have been reported previously, and untreated urban and industrial wastewaters were identified as major sources of metal pollution (Thuy et al., 2007)

4.2 Variation of sediment MMHg The sediment MMHg in the Tien and Saigon Rivers showed significant linear correlation with AVS (Table 3), suggesting that biogeochemical processes that produce min-eral FeS are associated with MMHg production In 2010, peak MMHg percentages were recorded at the estuarine head of the Tien (site 10) and Saigon Rivers (site 4;

Fig 5) Enhanced sulfate, associated with salinity intrusion, appears to increase Hg(II) methylation rate and AVS ( Hol-lweg et al., 2009) Relatively low MMHg percentage in the saline zone, despite large AVS, could be related to the inhi-bition effect of dissolved sulfide (Benoit et al., 2001; Drott

et al., 2007) A number of studies report that dissolved sul-fide shows strong inhibition of Hg(II) methylation rate in coastal sediments (Conaway et al., 2003; Heyes et al., 2006; Sunderland et al., 2006) Interestingly, in the present study, there is a significant positive linear correlation be-tween TN and MMHg in both rivers (Table 3) Assuming that TN is more inorganic (e.g., ammonium) than organic, based on the lack of correlation between TOC and TN, a sediment redox condition might explain high MMHg It

is commonly shown that reducing sediment provides favor-able conditions for microbial Hg(II) methylation ( Sunder-land et al., 2006) Applying a multiple linear regression for MMHg, it was found that AVS explained 41% of vari-ability in MMHg (p < 0.05, n = 17) and that TN explained

an additional 15% of variability in MMHg ([MMHg

Fig 3 Unfiltered Hg (UHg), dissolved Hg (DHg), and particulate

Hg (PHg) with distance to the coast in 2010 (a) and 2011 (b), and

unfiltered MMHg (UMMHg), dissolved MMHg (DMMHg), and

particulate MMHg (PMMHg) with distance to the coast in 2010 (a)

and 2011 (b) in Tien River.

(pmol g1)] = 0.19[AVS (lmol g1)] + 7.45[TN (%)] 1.7,

p < 0.05, n = 17) Besides AVS and sediment redox

condi-tion, TOC often plays a main role in controlling MMHg

(or %MMHg) in marine sediment (Conaway et al., 2003;

Hammerschmidt and Fitzgerald, 2004; Lambertsson and

Nilssons, 2006; Sunderland et al., 2006; Hollweg et al.,

2009) However, sediment TOC did not exhibit significant

correlation with MMHg in the Tien and Saigon Rivers (

Ta-ble 3), implying that organic content was not a limiting fac-tor for Hg(II) methylation rate across our study sites

In the Tien River, sediment %MMHg levels for 2011 were four to five times those for 2010, despite similar levels

of AVS (Table 2) The range of C/N was significantly lower

in 2011 (3.4–6.5) than in 2010 (9.6–15), reflecting larger contributions of fresh and labile organic matter in 2011 In-creased %MMHg (or Hg(II) methylation rate) influenced

by enrichment of fresh organic matter has been reported from Venice Lagoon (Kim et al., 2011) and Hg-contami-nated sites in Sweden (Drott et al., 2007) Enrichment of fresh organic matter in 2011 sediment may be related to en-hanced primary production (Table 2) Nutrients are the main factor controlling phytoplankton growth in this re-gion The 2011 sampling campaign was indeed preceded

by a rainfall event lasting several days, which may have caused eutrophication and excessive algal bloom

4.3 Variation of Hg and MMHg in surface water PHg and PMMHg content were not laterally uniform: Hg- and MMHg-enriched particles transported by river runoff appear to be diluted with Hg- and MMHg-depleted, resuspended particles in the high turbidity zone (log [SPM] > 1.9, Fig 6) Indeed, the ranges of PHg and PMMHg concentrations in the brackish zone were quite similar to sediment Hg and MMHg concentrations, respec-tively (Table 2) Dilution of PHg and PMMHg in the estu-arine high turbidity zone was opposite the findings reported for the Seine Estuary, where matured particles (high C/N ratio) in the high turbidity zone actually entrapped more

Hg and MMHg from the aqueous phase, resulting in in-creases in PHg and PMMHg in the estuarine high turbidity zone (Laurier et al., 2003)

FromFig 7a and b, it is evident that low POC content

in brackish particles explains PHg and PMMHg decreases

in the estuarine high turbidity zone Hg data for 2011 showed relatively steady and low PHg levels, with no posi-tive linear correlation with POC (Fig 7a) The ratio of

Chl-a to POC wChl-as used to differentiChl-ate suspended pChl-articles

dom-Table 3

Partial correlation matrix for Hg, MMHg, %MMHg against

salinity, TOC, TN, AVS, Hg, and MMHg in sediment samples of

the Tien River (a) and Saigon River (b).

Hg (nmol g 1 )

MMHg (pmol g 1 )

%MMHg (a) Tien River

<0.0001 0.086 0.47

AVS (lmol g1) 0.22 0.65 0.88

0.37 0.002 <0.0001

0.0001 21 (b) Saigon River

0.00088 8 Listed from top to bottom: Pearson’s correlation coefficient,

p-value and sample number Significant correlations are indicated in

bold type.

Fig 4 Correlation between Hg and total organic carbon (TOC) in sediments of Tien and Saigon Rivers (linear regression includes Tien data only).

386 S Noh et al / Geochimica et Cosmochimica Acta 106 (2013) 379–390

inated by labile organic matter (Chl-a/POC > 5 lg mg1)

from those dominated by refractory organic matter (

Cifu-entes et al., 1988; Liu et al., 2007;Supporting Information,

Fig S1) Labile organic particles cover a narrow range of

C/N, from 4.6 to 6.4, similar to the Redfield ratio For

2010, it is notable that freshwater SPM is dominated by

la-bile organics, while brackish SPM is dominated by

refrac-tory organics For 2011, however, most particles in

freshwater and brackish zones were labile; this might be

associated with an excessive algal bloom in 2011 (Table 1)

This suggests that biodilution may be one reason for low and steady PHg in 2011 A similar dilution effect for PHg, associated with phytoplanktonic bloom, was reported from the Seine Estuary (Laurier et al., 2003) For 2011, the biodilution effect appears to be weak for MMHg, since lin-ear regression was not distinct between 2010 and 2011 (Fig 7b) This could be related to higher MMHg produc-tion in 2011, shown as sediment %MMHg inTable 2 Luen-gen and Flegal (2009) report that as the bloom decays, MMHg in estuarine water significantly increases, likely due to MMHg production from sediment and/or remineral-ization from decaying phytoplankton

4.4 Partitioning of Hg and MMHg between solution and particles

The log-transformed particle–water partition coefficient,

Kd= [particulate Hg] (mol kg1)/[dissolved Hg] (mol L1),

of Hg averaged 5.1 ± 0.44 in 2010 and 4.9 ± 0.21 in 2011

in the Tien River (Fig 8) These ranges were similar to those found from other rivers (e.g., 4.8–5.7 in the Patuxent River,Benoit et al., 1998; 4.5–6.5 in the St Lawrence River,

Que´merais et al., 1998; and 2.8–6.6 for streams in Oregon and Wisconsin,Brigham et al., 2009) The log Kdof MMHg averaged 4.5 ± 0.77 in 2010 and 4.5 ± 0.50 in 2011 in the

Fig 5 Distribution of avid-volatile sulfide (AVS) and %MMHg at

each sampling site for 2010 Tien (a) and 2010 Saigon (b) Rivers.

AVS data are missing in sites 11 and 12 for (a).

Fig 6 Correlation between particulate Hg (PHg) or particulate

MMHg (PMMHg) and suspended particulate matter (SPM) in

surface waters of Tien River.

Fig 7 Correlation between particulate Hg (PHg) and particulate organic carbon (POC) (a) and between particulate MMHg (PMMHg) and POC (b) in surface waters of Tien River Linear regression model for (a) includes 2010 data only.

Tien River We found a negative linear relationship

be-tween log Kdof Hg (MMHg and OC) and log [SPM]; this

is known as a particle concentration effect (Benoit, 1995;

Fig 8) The POC content in SPM may play a critical role

in this correlation, as we found less POC under high SPM

conditions However, there should be additional factors

to explain the close negative correlation, since POC

con-tents were relatively constant when log [SPM] > 1.9 (

Sup-porting Information, Table S1) Increasing colloidal

organic matter along with increasing SPM may be

associ-ated with decreases in Kdof Hg, MMHg, and OC as a

func-tion of SPM (Honeyman and Santschi, 1991; Benoit and

Rozan, 1999; Lee et al., 2011)

The slope (0.52) of the linear regression between log Kd

of Hg and log [SPM] was similar to that (0.38) between

log Kdof OC and log [SPM], indicating the critical role of

organic matter in governing phase-partitioning of Hg (

Be-noit et al., 1998; Choe and Gill, 2003; Conaway et al.,

2003; Laurier et al., 2003; Sunderland et al., 2006)

Interac-tions between organic matter and Hg, attributable to strong

Hg binding with reduced sulfur-containing functional

groups (e.g., thiol), appear to control particle–water

parti-tion of Hg On the other hand, the slope (0.97) of the

lin-ear regression between log Kdof MMHg and log [SPM] was

substantially lower than those for Hg and OC We

specu-late that chloro-complexation of MMHg provides

addi-tional decreases in log Kdof MMHg in the estuarine zone

(log [SPM] > 1.9) It is known that interactions between

MMHg and organic ligand (L) are less significant

com-pared to those between Hg and L (Zhong and Wang,

2009) According to the Hg–Cl–L ligand complexation

model, HgL complex dominates dissolved Hg pool under

the Tien’s estuarine conditions ([Cl] = 0.02–0.3 M,

[DOC] = 180–360 lM), assuming L/DOC = (5–50) 106

and 22 < log K0< 25, Fitzgerald et al., 2007) On the

con-trary, the CH3Hg–Cl–L ligand complexation model

pre-dicts that CH3HgL competes with CH3HgCl under our

estuarine conditions, assuming L/DOC = (5–50) 106

and log K0= 13.6, Fitzgerald et al., 2007) There appears

to be more than enough L with sufficient affinity with Hg

to compete out Hg–chloro complexes, but this is not the case for MMHg in the estuarine zone of the Tien River

5 SUMMARY The Mekong River Delta, located in Vietnam, is a flat, low-lying area of highly complex rivers, channels, and flood plains Although there are more than 1000 man-made ca-nals in the Mekong Delta for inhibition of saltwater, trans-port, and land reclamation, saline intrusion still occurs and causes severe human suffering during dry seasons (Hoa

et al., 2007) In surface sediment, enhanced %MMHg was found at the estuarine head, along with increased AVS, emphasizing the importance of sulfate availability In sur-face water, UHg concentrations were greater in the estua-rine high turbidity zone compared to the upper and lower rivers, due to enhanced particle load Fractions of particu-late Hg appear to be remobilized in the estuarine high tur-bidity zone, since DHg also increased in the high turtur-bidity zone On the contrary, UMMHg and DMMHg did not in-crease in the estuarine high turbidity zone compared to the upper and lower rivers, likely due to large decreases in PMMHg Although the high turbidity zone appears to have elevated microbial activity, which manifests as low ORP, sulfide may play a critical role in constraining active Hg(II) methylation Regarding particle–water distribution of Hg and MMHg, we found increased solubility of Hg and MMHg in the estuarine zone, with a negative linear rela-tionship between log Kdof Hg (MMHg) and log [SPM] Be-tween 2010 and 2011, sediment and water conditions were highly variable in terms of Hg and organic quality Hg in suspended particles was depleted in the bloom year (2011), highlighting the importance of the biodilution effect

In contrast, sediment %MMHg increased in the bloom year, perhaps due to enrichment of labile fresh organic mat-ter in sediment, and subsequent enhancement of microbial activity Taken as a whole, Hg speciation and partitioning

in the lower Mekong River Basin were strongly influenced

by salinity intrusion, and major biogeochemical factors affecting Hg behaviors were particle loads, biological pro-ductivity, and concentrations of sulfate, chloride and or-ganic matter

ACKNOWLEDGEMENTS

We are grateful for the support of the Hanyang Research and Industry Cluster at Hanyang University This study was supported

by the Ministry of Science and Technology, Korea, through the Institute of Science and Technology for Sustainability (UNU & GIST Joint Program); and by the Ministry of Land, Transport, and Maritime Affairs through “Impacts of ocean acidification on the bioaccumulation and release of mercury by microbes”.

APPENDIX A SUPPLEMENTARY DATA Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/ j.gca.2012.12.018

Fig 8 K d (particle–water partition coefficient) of Hg, MMHg and

organic carbon (OC) as a function of suspended particulate matter

(SPM) concentration in Tien River.

388 S Noh et al / Geochimica et Cosmochimica Acta 106 (2013) 379–390