Mercury speciation,bioavailability and risk assessment on soil-rice systems from a watershed impacted by abandoned Hg mine-waste tailings

Jian Pang·Jialiang Han·Xuelu Fan·Chan Li·Xian Dong·Longchao Liang·Zhuo Chen

Abstract Mercury(Hg)is a global pollutant and can be accumulated in the food chain,posing exposure risks to humans.In this study,rice plants and corresponding rhizosphere soil samples were collected from a watershed of the Wawu River Basin that is heavily impacted by historic Hg mining and retorting activities.Total mercury(THg)and methylmercury(MeHg)in rice grains,as well as other tissues and soil samples,were measured.Five soil Hg fractions,as well as soil parameters,were also determined.The results show that the average concentrations of THg and MeHg in rice grains were 14±7.0μg kg-1 and 7.2±4.0μg kg-1.Soil organic-bound(Hg-o)and strong complex-bound (Hg-s)were the main Hg fractions,accounting for 44%of the total.To estimate the Hg–ligand interaction in the soils,soil-N/Hg(R=0.451,p＜0.05),-S/Hg(R=0.372,p＜0.1),and-OM/Hg ratio(R=0.320,p＜0.5)with MeHgsoil were observed with signif icant positive correlations,indicating that the formation of Hg–OM,Hg–N–OM or Hg–S–OM complexes could prevent Hg(II)from methylation in soils.The signif icant positive correlations of the-N/Hg ratio,-S/Hg ratio and-OM/Hg ratio with MeHg in rice tissues suggested that Hg methylation and MeHg demethylation occurred throughout the rice paddy ecosystem.The estimated MeHg daily intake(EDI)was 0.075±0.041μg kg-1 bw d-1 and was lower than the RfD level of 0.1μg kg-1 bw d-1 recommended by the US EPA.However,approximately 29%of the hazardous index(HI)of MeHg in grain exceeded 1,posing a potential threat to local populations,particularly pregnant women and children.

Keywords Mercury and methylmercury·Rice·Mercury speciation and bioavailability·Paddy soil·Risk assessment

1 Introduction

Mercury(Hg)is a global pollutant(Lindqvist et al.1991),and its mobility,bioavailability and toxicological effects are highly dependent on Hg chemical forms(Ullrich et al.2001).Methylmercury(MeHg),an organic Hg form,has been considered one of the most toxic Hg forms(Mergler et al.2007).Under certain environmental conditions,inorganic mercury(IHg)can be converted into MeHg by anaerobic bacteria(Ullrich et al.2001).Owing to the fatsolubility and long half-life period,MeHg can be bioaccumulated and biomagnif ied in the food chain(Qiu et al.2008).Asa consequence,up to 95%of total Hg(THg)can be bioaccumulated in organisms like f ish,posing health risks to biota(WHO 1990;Clarkson 2002).MeHg concentrations in f ish can be more than 106times greater than in ambient water(Stein et al.1996).Fish consumption is currently considered the dominant pathway of MeHg exposed to humans(Mergler et al.2007).

China has one of the world’sgreatest Hg depositsin the form of cinnabar ores,most of which are allocated in the Guizhou Province(Feng et al.2008).Guizhou islocated in the center of the circum-Pacif ic mercuriferous belt and the total reserves of metal Hg have reached 80,000 tons(Qu 2004;Qiu et al.2005a,b).An approximately 600-year history of intensive mining and retorting activities have generated huge quantities of Hg-containing waste residue,wastewater and exhausts in the local ecosystem,causing heavily Hg-contaminated soil,water,air and biota(Tan et al.1997;Feng et al.2008).In Hg mining areas,rice paddies have been heavily contaminated with Hg and have caused accumulation of MeHg in rice(Qiu et al.2008;Meng et al.2010;Zhang et al.2010a,b,c).Elevated MeHg concentrations,as high as 174μg kg-1in edible parts of rice,were f irst reported in the Wanshan Hg mining area(Qiu et al.2008).Recently,studies suggest that rice consumption can be the main pathway of MeHg exposure to populations dwelling in abandoned Hg mining areas(Feng et al.2008;Qiu et al.2008;Zhang et al.2010a,b).

Rice plantations are the most prevalent use of land in South and East Asia,where rice is the dominant foodstuff,and approximately 70%of the daily dietary energy of more than 2 billion people comes from rice and its byproducts(FAO 2002,2006).China is the largest producer of rice,generating 29%of the global rice supply in 2009(IRRI 2012).Rice paddies,as typical wetland environments,have usually been kept asf looded reduction environmentsduring the rice growing period.This has resulted in favorable conditions for Hg methylation for microorganisms,which isconsidered an important sourceof MeHg in theterrestrial ecosystem(Meng et al.2010,2011).The mobility and methylation of Hg in temporarily f looded paddies are determined by a range of factors,such as pH,dissolved organic matter(OM),sulfur,iron,redox potential,and dissolved Hg concentration(Liu et al.2014;Zhao et al.2016).Previous studies have indicated that Hg methylation islargely facilitated by a subset of sulfate-reducing bacteria(SRB)or iron-reducing bacteria(IRB)in anoxic conditions(Gilmour et al.1992;Fleming et al.2006).Particularly,the OM,which containsvarious O-,N-,and S-bearing ligands,not only servesasan important electron donor for SRB,but also as a direct binder to Hg(II)(Skyllberg and Drott 2010;Graham et al.2012).

Although paddy soil has been conf irmed to be the primary MeHg source to rice,no signif icant positive correlation between soil THg and MeHg and rice THg has been reported(Gnamusˇet al.2000;Meng et al.2010,2011;Rothenberg et al.2011;Zhang et al.2012).Water soluble and organo-chelated Hg in soil is readily available to biota(Schuster 1991;Wallschla¨ger et al.1998;Boszke et al.2006;Covelli et al.2009),while Hg in paddy soil,particularly in Hg mining regions,can be bound to different matrices,resulting in various capabilities of mobility,bioavailability and potential toxicity(Issaro et al.2009).Hence,thepaddy soil Hg speciesand their interconversions play important roles in the Hg accumulation of rice.However,the transfer of different Hg species from soil to rice still remains far from being well understood.Few studies have focused on the relationships between different Hg speciesin soilsand Hg in rice in watershed impacted by abandoned Hg mine-waste tailings.

In this study,rice plants and corresponding rhizosphere soils were collected from a watershed that has been strongly impacted by historic Hg retorting activities.The main objectives are(1)to reveal the distribution and accumulation of THg and MeHg in rice plants and rhizosphere soils,(2)to characterize the inf luence of soil Hg speciation on Hg bioaccumulation in rice,(3)to elucidate key factorsof soil parametersthat determine the production and bioaccumulation of MeHg in rice and,(4)to assess the health exposure risk of both IHg and MeHg via rice consumption.

2 Materials and methods

2.1 Study area

The study region of the Wawu River watershed(E:109°23′–109°43′,N:27°53′–27°70′)covers an area of approximately 343 km2in Wanshan,Guizhou Province,Southwest China.The Wawu River f lows into the Jinjiang River at the conf luence of Yangtou Town.The main tributaries of the Wengman,Aozhai and Xiaxi originate from the Yanwuping(YWP),Meizixi(MZX)and Dashuixi(DSX),which are located at 8.3 km northeast,2.5 km southwest and 2.5 km east of Wanshan Town,respectively.The region is characterized by a subtropical monsoon humid climate with an average annual rainfall of 1400 mm and a mean temperature of 13.4°C.Large quantities of mine-waste calcines have been deposited adjacent to the sites of the MZX,DSX and YWP.The calcines nearby the MZX and DSX sites were approximately 2.83 million m3and 44.5 million m3,respectively,and there were approximately 3337 tons of mine tailings near the YWP.The historic intensive Hg smelting activities have resulted in serious Hg contaminations to the ambient air,water,soil and biota(Qiu et al.2005a,b;Li et al.2008a,b,2009).

2.2 Sampling and preparation

Whole rice plants and corresponding rhizosphere paddy soil samples were collected within the Wawu River watershed during the harvest season of September 2015(Fig.1).Brief ly,at each sampling site,approximately 2–3 clusters of whole rice plants within a plot area of 50 m×50 m were randomly collected and pooled as one f inal sample.After the separation of grains,the whole plant was thoroughly washed in situ,with water from the adjacent rivers.At each site,approximately 1.5 kg of rhizosphere composited soils were collected simultaneously.Soil from the roots of 2–3 clusters was mixed and represented a f inal sample.All samples were kept in Ziplock bags to prevent cross-contamination.Then,all samples were stored in coolers(+4°C)and transported to the laboratory for further processing prior to speciation analysis.

Fig.1 Map of sampling sites(WK Wukeng,DSX Dashuixi,XX Xiaxi,MZX Meizixi,AZ Aozhai,YWP Yanwuping,LLS Liulongshan)

In the laboratory,the whole rice plants were thoroughly washed with tap water,then with Milli-Qwater threetimes.After rinsing,therice plantswere separated into leaf,stem,and root with ceramic scissors.All samplesweredried with a Freeze Dryer(FDU-1100,Japan).For grain,polished rice was obtained by shelling and removing the hull and bran.Then,all sampleswereground with a Micro Plant Grinding Machine(IKA-A11,Germany).The soil samples were ground with a mortar and passed through a 200μm mesh size sieve and packed in Ziplock bags.During the sample preparation processes,the tools(such as scissors)and machines(such as FDU-1100 and IKA-A11)were rinsed three times with ethanol,to avoid cross-contamination.

2.3 Analytical methods

2.3.1 THg

For rice,approximately 0.05–0.2 g sample was added to 5 ml Milli-Q water and a 5 ml fresh mixture of HNO3/H2SO4(v/v 4:1)and was digested in a water bath at 95°C for 3 h(Zhang et al.2010a,b,c).THg was determined by the dual-stage gold amalgamation method and cold vapor atomic f luorescence spectrometry(CVAFS,Tekran 2500,Tekran Inc,Canada),preceded by BrCl oxidation,SnCl2reduction,preconcentration,and thermoreduction to Hg0(US EPA 2002).

For soil,approximately 0.2 g sample was weighed into plastic tubes,then 5 ml Milli-Q water and 5 ml fresh aqua regia were added for digestion in a water bath at 95°C for 3 h.THg was determined by using a F732 V Atomic Absorption Spectrometry Mercury Analyzer(AAS,F732-V;Shanghai Huaguang,China).

2.3.2 MeHg

For rice,approximately 0.1–0.2 g sample was diluted with 5 ml 25%KOH–methanol solution and heated in 75°C water for 3 h(Liang et al.1996).MeHg in the sample was extracted with CH2Cl2,then back-extracted from the solvent phase into an aqueous ethyl phase.The sample extraction was followed by distillation,addition of 2 M acetate buffer,ethylation with 1%sodium tetraethylborate and purging and trapping of MeHg onto Tenax(US EPA 2001).MeHg was determined using gas chromatographycold vapor atomic f luorescence spectrometry(GC-CVAFS,Brooks Rand Model III,Brooks Rand Laboratories,Seattle,WA),following US EPA Method 1630.

For soil,approximately 0.05–0.1 g sample was prepared using 1.5 ml CuSO4-methanol solvent and 7.5 ml 25%HNO3,heated in 75°C water for 3 h(Liang et al.1994).The following steps were similar to the processing of rice,and the MeHg analysis was also according to the USEPA Method 1630.

2.3.3 Soil Hg fractions

Sequential extraction of the Hg fractions in soil referred to Bloom et al.(2003),with a modif ication.Five operational Hg fractions:water-soluble-bound(Hg-w),simulated gastric acid-bound(Hg-g),fulvic acid-bound(Hg-f),humic acid-bound(Hg-h)and strong complex-bound(Hg-s)fractions were extracted.All Hg species were analyzed by CVAFS.Detailed information on the extraction steps is listed in Table 1.

2.3.4 Soil parameters

To determine pH,approximately 10 g of a soil sample was added to 25 ml Milli-Q water,shaken for 5 min at room temperature and allowed to stand for 1–3 h.The clear solution was used to measure the pH value by PHS-3C pH meter(NY-T,1377-2007).For organic matter(OM),approximately 0.25 g of a soil sample was mixed with 0.05 g HgSO4,2.5 ml K2Cr2O7,and 3.75 ml H2SO4.The mixed solution was digested for 30 min at 135°C and diluted with Milli-Q water to 50 ml.The absorbance was measured at 585 nm by using a 72 UV–Vis spectroscope(HJ,615-2011).To measure total nitrogen(TN)and total sulfur(TS),approximately 0.1 g of a soil sample was weighed in aluminum tin cups and determined by Vario MACRO Cube Elemental analyzer(Elementar analysensysteme,Hanau;Germany).

2.3.5 QA/QC

Method blanks,the standards curve,certif ied reference material(CRM),and relative standard deviation(RSD)were employed for quality assurance and quality control for THg and MeHg analyses in rice and soil(Table 2).

For soil,the THg in the estimated value of estuarine sediments(ERM CC580)was 130±0.005 mg kg-1,which is comparable with the certif ied value of 132±0.003 mg kg-1.The estimated value of MeHg was 0.075±0.006 mg kg-1,which is comparable with the certif ied value of 0.076±0.004 mg kg-1.The method detection limits were 0.1μg kg-1for THg and 0.003μg kg-1for MeHg.The method blank wasbelow the limit for THg and MeHg.

For rice,themethod detection limitsfor THg and MeHg were the same as in the soil samples.The average THg concentration of citrus leaves (GBW 10020) was 0.140±0.005 mg kg-1,whereas the certif ied value was 0.152±0.020 mg kg-1.An average MeHg concentration of Lobster hepatopancreas (TORT-2) was 0.140±0.006 mg kg-1, with a certif ied value of 0.152±0.013 mg kg-1.The method blank was below the limit for THg and MeHg.

2.4 Statistical analysis

Statistical analysis of the data was performed using the SPSSStatistics 22.0 software(IBM,International Business Machines Corporation,America,2013)and Origin Pro 8.5 software(Origin Lab,America,2011).The correlationcoeff icients(R)and signif icance probabilities(p)were adopted among the data.

Table 1 The specif ic extraction steps of Hg fractions in soil

Table 2 Information on QA/QC of measurement data

3 Results and discussion

3.1 Soil

3.1.1 THg and MeHg

As shown in Fig.2,the THg(THgsoil)concentrations of soils,obtained from different areas,ranged from 0.43 to 110 mg kg-1,with an average value of 23±28 mg kg-1.Most of the paddy soil samples exhibited rather high THg values,exceeding the Chinese national standard limit of 1.5 mg kg-1for paddy soils(GB15618,1995).The soil sample collected from XX showed the highest THg value,which is 36±39 mg kg-1on average,followed by AZ,with 28±24 mg kg-1;and WM,with 23±29 mg kg-1.The lowest value of 6.9±2.8 mg kg-1,on average,was observed in WW(Fig.2a).The high THg concentrations observed in XX and AZ suggested that the historically intensive Hg mining and retorting activities might be responsible for this high THg value.

The average MeHg concentration in the soil samples was 1.3±1.0μg kg-1,with values ranged from 0.3 to 4.2μg kg-1.The proportion of MeHg to THg in the soil was 0.018%,with a range of 0.00084%–0.072%,suggesting that soil Hg was the main source in IHg forms(Qiu et al.2005a,b).The highest MeHg concentration of 2.0±1.2μg kg-1on average was recorded in samples from AZ,followed by XX,with 1.4±1.3μg kg-1on average;WM,with 0.87±0.39μg kg-1and WW,with 0.87±0.39μg kg-1(Fig.2b).The average MeHg concentrations were at a similar level between the Wengman River and Wawu River,which were 2.3 times lower than those observed in the Aozhai River.A large amount of Hg0that was released into the atmosphere during Hg retorting activities may eventually be deposited into its surroundings,causing in situ MeHg production(Zhao et al.2016).

To estimate the Hg–ligand interaction in soils,the soil-N/Hg ratio,-S/Hg ratio and-OM/Hg ratios were calculated(Yin et al.2018).Signif icant positive correlations of soil-N/Hg ratio(R=0.451,p＜0.05),-S/Hg ratio(R=0.372,p＜0.1),and-OM/Hg ratio(R=0.320)with MeHgsoilwere observed(Fig.3a–c),indicating that the elevated N/Hg ratio,S/Hg ratio and OM/Hg ratio in the soils inhibited the production of MeHg,possibly due to the formation of Hg–OM, Hg–N–OM or Hg–S–OM complexes,which prevented Hg(II)from methylation in the soils(Skyllberg et al.2006).

Fig.2 The distribution of THg(a)and MeHg(b)concentrationsin rhizospheresoil and ricetissuesfor samplesfrom four tributaries(XX Xiaxi River,AZ Aozhai River,WM Wengman River,WW Wawu River)(units of THggrain:μg kg-1)

Fig.3 The correlationsof MeHg in soil and ricewith soil-N/Hg ratio,-S/Hg ratio and-OM/Hg ratio[MeHgsoil(a–c),MeHggrain(d–f),MeHgroot(g–i),MeHgstalk(j–l),and MeHgleaf(m–o)]

3.1.2 Soil Hg fractions

Soil Hg fraction concentrations indicated that organicbound Hg(Hg-o:Hg-f and Hg-h)and strong complexbound (Hg-s)were the main forms,accounting for approximately 44%of the total amount.The bioavailable bound Hg fraction(Hg-b:Hg-w and Hg-g)in the soil samples,which was considered to be dominated by active Hg2+and contributed to Hg methylation or was directly absorbed by plants(Boszke et al.2006;Covelli et al.2009),generally accounted for 0.029%of the total.Soil samples from XX,AZ,and WM exhibited the highest concentrations of Hg-s,followed by Hg-o,and the lowest of Hg-b,while soils from WW showed the highest values of Hg-o(Fig.4a,b).The reason for the high Hg-o in WW could be the large biochemical oxygen demand(BOD)and chemical oxygen demand(COD)in the water(Li et al.2012),which can readily absorb Hg released from upstream mine-waste tailings.

Both Hg-b(R=0.687,p＜0.01)and Hg-s(R=0.917,p＜0.01)were signif icantly correlated to THgsoil,as well as to each other(R=0.428,p＜0.05)(Table 3).The positive collection between Hg-b and Hg-s might indicate that Hg-s can be transformed to Hg-b.A recent study proposed that the nonbioavailable fraction in soil could be transformed to more toxic or potentially bioavailable fractions,hence promoting MeHg production during rice cultivation(Wu et al.2018).The transformation between Hg-s and Hg-b has been a great concern to scientists.The increased Hg-b in paddy soil might be a hint of readily accumulated Hg in biota that could eventually enter and biomagnify in the food chain.

Fig.4 Thedistribution of Hg fractionsconcentration in rhizospheresoil(a),thepercentageof Hg fractionsin soil THg(b)from four tributaries(XX Xiaxi River,AZ Aozhai River,WM Wengman River,WW Wawu River)(Hg-w:water-soluble;Hg-g:simulated gastric acid bound;Hg-f:fulvic acid bound;Hg-h:humic acid bound and Hg-s:strong complex bound)

3.2 Rice

3.2.1 THg and MeHg in grain

The THgconcentration(THggrain)was14±7.0μg kg-1on average,ranging from 4.1 to 34μg kg-1.The XX exhibited the highest THg concentration of 17±9.1μg kg-1on average, followed by AZ at a concentration of 15±3.9μg kg-1,WM at 13±4.3μg kg-1,and the lowest of 11±3.9μg kg-1,observed in WW(Fig.2a).Samples collected from sites2 and 7 from XX,site 11 from AZ,and site 14 from WM exhibited high values,exceeding themaximum Hg level of 20μg kg-1recommended by the Ministry of Health Standardization Administration of China(2017)(Fig.5).Among thosefour sites,both site7 and site 11areadjacenttoabandoned artisanal Hgretortingfacilities and were strongly impacted by atmospheric Hg deposition.This newly deposited Hg is considered to be highly bioavailableand isreadily transformed into MeHg,resulting in high THggrainlevels(Meng et al.2011).

The MeHg concentration (MeHggrain) was 7.2±4.0μg kg-1on average,ranging from 2.0 to 17μg kg-1.Similar to THggrain,the highest average MeHg concentration of 8.1±5.3μg kg-1was recorded in XX,followed by AZatconcentrationsof 7.7±2.2μg kg-1,WM at 6.5±5.2μg kg-1,and WW at 6.5±3.2μg kg-1(Fig.2b).High MeHggrainlevels were found near the abandoned Hg mining sites and artisanal Hg smelters,which was consistent with the distribution of THggrain.Moreover,MeHggraincontributed to a large portion of Hg in rice(FMeHggrain),reaching an average of 51%,with a range of 25%–79%.The high MeHg ratios in grain certainly bring a greatrisktothehealthof localresidents,sinceMeHgisreadily bioavailable(Liet al.2008a,b).

3.2.2 THg and MeHg in roots,stalks and leaves

THg concentrationsin tissuesexhibited the highest valuesin roots,followed by leaves,with the lowest in stalks,ranging from 0.06 to 7.7,0.10 to 1.2,and 0.02 to 0.31 mg kg-1,respectively.The XX exhibited the highest value of 2.4±2.6 mg kg-1in roots,indicating that the heavily Hgcontaminated soilsfor root Hg levelswasdirectly associated withsoil Hglevels(Fayand Gustin2007).Boththeleavesand stalks from AZ exhibited the highest values of 0.74±0.44 mg kg-1and 0.12±0.10 mg kg-1,respectively.Sinceplantscanabsorb Hg0fromtheair intotheleaves and stalks(Meng et al.2012),thehigh valuesof THgstalkand THgleaffound in AZlikely suggestahighlevelof atmospheric Hg in ambient air.

Theaverage MeHg concentration in roots(MeHgroot)was 2.9±1.5μg kg-1,with a range of 0.78–7.3μg kg-1.The AZ exhibited the highest MeHgrootof 3.1±0.86μg kg-1,followed by XX at 3.0±2.1μg kg-1, WM at 2.9±2.1μg kg-1,and WW at 2.5±1.0μg kg-1.In contrast,the lowest MeHg concentration (MeHgstalk)of 0.72±0.50μg kg-1onaveragewasobserved instalks,with a range of 0.18–1.8μg kg-1.The XX exhibited the highest MeHgstalkof 1.0±0.66μg kg-1,and the lowest of 0.46±0.25μg kg-1wasobservedat WW(Fig.2b).Among the different tissues,the leaf tissue exhibited the lowest average MeHg concentration (MeHgleaf) of 0.41±0.25μg kg-1with a range of 0.13–1.2μg kg-1.Similarly,the AZ exhibited the highest MeHgleafof 0.56±0.39μg kg-1, followed by XX at 0.42±0.25μg kg-1,WW at 0.37±0.094μg kg-1,and WM at 0.26±0.12μg kg-1(Fig.2b).

The proportion of MeHg to THg in stalks(FMeHgstalk)is 1.25%,with a range of 0.28%–4.5%.Then,the proportion of MeHg to THg in roots(FMeHgroot)is0.47%,with a range of 0.038%–12%.MeHgleafcontributed the smallest proportion of Hg in rice(FMeHgleaf),only reaching 0.14%,on average,with a range of 0.025%–0.33%.

MeHgSoil 1 Table3Pearson’scorrelationmatrix,thelinearcorrelationcoefficientsamongtheHglevels(THg,MeHgandsoilHgfractions)inrice-soilsystem MeHgGrain MeHgLeaf MeHgStalk MeHgRoot Hg-s Hg-o Hg-b THgSoil THgGrain THgLeaf THgStalk THgRoot 1 0.296 1 0.318 0.525**1 0.711**0.612**0.536**1 0.482*0.361 0.525**0.454*1 1 0.124 1 0.095 0.428*1 0.687**0.151 0.917**1 0.265 1 0.518**0.561**1 0.714**0.536**0.413*1 0.317 0.470*0.27 0.860**THgRoot THgStalk THgLeaf THgGrain THgSoil Hg-b Hg-o Hg-s MeHgRoot MeHgStalk MeHgLeaf MeHgGrain MeHgSoil**Correlationissignificantatthe0.01level;*correlationissignificantatthe0.05level

Fig.5 THg and MeHg concentrationsin grain of ricegrown in the Wawu River Basin(WK Wukeng,DSX Dashuixi,XX Xiaxi,MZX Meizixi,AZ Aozhai,YWP Yanwuping,LLS Liulongshan)

3.3 Factors affecting Hg levels in rice

3.3.1 Soil THg and MeHg

The THgsoilsignif icantly correlated to THgroot(R=0.860,p＜0.01),followed by THgleaf(R=0.561,p＜0.01),THgstalk(R=0.413,p＜0.05),and THggrain(R=0.265)(Table 3).Among rice tissues,THggrainwas signif icantly positively correlated with THgstalk(R=0.536,p＜0.01)and THgleaf(R=0.518,p＜0.01),indicating that atmospheric Hg may be one of the Hg sources in the aerial parts of rice(Meng et al.2012).

MeHggrainexhibited signif icant positive correlations with MeHgroot(R=0.525, p＜0.01), MeHgstalk(R=0.612,p＜0.01)and MeHgsoil(R=0.296).MeHgrootwas also positively correlated with MeHgsoil(R=0.454,p＜0.05).These results indicated that MeHg in soil is a potential source of MeHg in rice roots and rice grains(Meng et al.2014).

3.3.2 Soil Hg speciation

There was a signif icant negative correlation between FMeHggrainand Hg-s (%) (R=-0.408,p＜0.05)(Fig.6).A signif icant positive correlation between FMeHggrainand Hg-o(%)(R=0.336)was observed(Fig.6).No signif icant relationship of Hg-b(%)(R=0.094)with FMeHggraincan beobtained(Fig.6).This could be mostly attributed to the fact that the Hg fraction wasnot adirect factor in MeHggraindistribution(Zhao et al.2016).

Hg-b(%)was signif icantly positively correlated with FMeHgleaf(R=0.595, p＜0.01) and FMeHgroot(R=0.399,p＜0.1)(Fig.6).This is likely because the bioavailable bound Hg might be a major substrate for the methylation process of IHg(Covelli et al.2009).Hg-o(%)was signif icantly correlated with FMeHgroot(R=0.169),FMeHgstalk(R=0.186)and FMeHgleaf(R=0.340),suggesting that organic-bound Hg in rice can promote the production of the more toxic MeHg(Zhou et al.2015).Negative correlations of FMeHgroot(R=-0.149),FMeHgstalk(R=-0.152)and FMeHgleaf(R=-0.256,p＜0.1)with strong complex-bound Hg were observed(Fig.6).However,the MeHg accumulation in rice may be inf luenced by many factors.More studies are needed to further investigate the process of the methylation of the Hg fraction.

3.3.3 Other soil parameters(OM,TN and TS)

Soil-N/Hg ratio (R=0.373,p＜0.1),-S/Hg ratio(R=0.391,p＜0.1),and -OM/Hg ratio (R=0.320,p＜0.5)with MeHggrain(Fig.3d–f)showed signif icant positive relationships,indicating the distribution of MeHggrainwas largely controlled by the OM(N,S)-Hg interactions in the soils(Yin et al.2018).

Signif icant positive relationships of MeHgroot(R=0.527,p＜0.01),MeHgstalk(R=0.530,p＜0.01)and MeHgleaf(R=0.521,p＜0.01)with-N/Hg ratio(Fig.3g,j,m)were observed.In addition,-S/Hg ratio was signif icantly correlated with MeHgroot(R=0.467,p＜0.05),MeHgstalk(R=0.521,p＜0.01)and MeHgleaf(R=0.437,p＜0.05)(Fig.3h,k,n).There were signif icant correlations between-OM/Hg ratio and MeHgroot(R=0.488,p＜0.05),MeHgstalk(R=0.472,p＜0.05)and MeHgleaf(R=0.499,p＜0.05)(Fig.3i,l,o).These results suggest that Hg methylation and MeHg demethylation occurred throughout the rice paddy ecosystem(Zhao et al.2016).

3.4 Risk assessment

Estimated daily intake(EDI)and hazard index(HI)via rice consumption of residents were assessed based on data for both IHg and MeHg in the grain obtained in the present study.According to a report by Lin et al.(2008),IHg was def ined as the difference between THg and MeHg.The RfD referred to the Hg daily safe intake recommended by JECFA(2010)and the US EPA(1998),and the corresponding values of IHg and MeHg were 0.57,0.23μg kg-1bw d-1and 0.30,0.10μg kg-1bw d-1,respectively.An HIvalue lower than 1 means that the level of Hg will not cause noncarcinogenic Hg exposure risk to humans(Cao et al.2010).In contrast,when the HIvalue is greater than 1,potential health risks arising from carcinogenic Hg exposure should be of concern(Qiu et al.2005a,b).Assuming MeHg absorption in the body was 100%,the calculations were as follows:

where EDI is the estimated daily intake,μg kg-1bw d-1;CM is the concentration of Hg(IHg and MeHg)in rice grain,μg kg-1;IR is the daily intake rate,0.625 kg d-1(Feng et al.2013);BW is the average body weight,60 kg;HI is the hazard index.

The results for EDI and HI of both IHg and MeHg are shown in Table 4.The IHg average value of EDI in rice was 0.072±0.040μg kg-1bw d-1,ranging from 0.021 to 0.170μg kg-1bw d-1.The MeHg averagevalueof EDI was 0.075±0.041μg kg-1bw d-1,ranging from 0.021 to 0.180μg kg-1bw d-1.According to the evaluation method recommended by JECFA,the average HI of IHg and MeHg were 0.13,with a range of 0.04–0.30 and 0.33,with a range of 0.09–0.79.The average HIs of IHg and MeHg were 0.24,with arangeof 0.07–0.58,and 0.75,with a range of 0.21–1.80,as recommended by the USEPA.

Approximately 29%EDIof MeHg in whitericesamples exceeded the maximum dose of 0.1μg kg-1bw d-1MeHg developed by the United States Environmental Protection Agency(US EPA 1997).Among all samples,the EDIand HIof IHg in Wawu River Basin were slightly lower than the values shown in the previous study(Zhang et al.2010a,b,c).The EDI of MeHg in the present study was consistent with the results reported by Zhang et al.(2010a,b,c),but much lower than the results reported by Qiu et al.(2008).

According to the daily safe intake evaluation method recommended by JECFA,the HI of IHg and MeHg were less than 1.According to the USEPA,the HI of IHg does not exceed 1,but approximately 29%HI of MeHg exceeded 1,which is a potential threat to the health of local residentswho consumetherice on regular basis.It isurgent to adopt corresponding ecological restoration measures to debasesoil Hg and atmospheric Hg in Hg mining areasand reduce Hg exposure risks for local residents,especially sensitive groups,such as pregnant women and children.

4 Conclusions

In summary,we have measured the THg and MeHg values of rice plants,and the corresponding rhizosphere soil samples were collected from a watershed of the Wawu River Basin.Soil Hg fractions of Hg-w,Hg-g,Hg-f,Hg-h,and Hg-s and other important parameters(e.g.,pH,OM,TN and TS)were evaluated systematically.Our results revealed that rice plants and the corresponding rhizosphere soil samples collected from the Wawu River Basin contain signif icant Hg-contaminations.Soil Hg fractions of Hg-o and Hg-s are the main Hg forms and reached 44%in total.Signif icant correlations of MeHg in rice and soil with soil parameters(N,Sand OM)were observed,mainly caused by demethylation activities in soil throughout the rice paddy ecosystem.The distribution of MeHggrainwas largely controlled by the OM(N,S)-Hg interactions in the soil.The EDI value of MeHg was measured to be 0.075±0.041μg kg-1bw d-1,which was lower than the RfD level 0.1μg kg-1bw d-1recommended by the US EPA.However,it isworth noting that 29%of HIsof MeHg exceeded 1,indicating the potential MeHg exposure risks to local residents,particularly to sensitive populations,e.g.,pregnant women and children.Ongoing research isfocused on understanding the possible processes of Hg methylation in rice paddies and developing ecological restoration measurements to reduce Hg concentrations in the paddy soil and Hg mining areas.

Fig.6 The correlations of Hg(%)in Hg fraction with FMeHggrain(a),FMeHgroot(b),FMeHgstalk(c)and FMeHgleaf(d)(FMeHggrain,FMeHgroot,FMeHgstalk and FMeHgleaf were the proportion of MeHg to THg in grain,root,stalk and leaf)

Table 4 The EDI and HI ofIHg and MeHg for humanexposures in Wawu River Basin

AcknowledgementsFinancial support for this work was provided by the National Natural Science Foundation of China (NSFC:21767007),and the Science and Technological Program of Guizhou(2018-1111).