Journal of Contaminant Hydrology 177–178 (2015) 122–135

Contents lists available at ScienceDirect

Journal of Contaminant Hydrology journal homepage: www.elsevier.com/locate/jconhyd

Tracking natural and anthropogenic origins of dissolved arsenic during surface and groundwater interaction in a post-closure mining context: Isotopic constraints Mahmoud Khaska a,b,⁎, Corinne Le Gal La Salle a,b, Patrick Verdoux a, René Boutin c a b c

Univ. Nimes, EA 7352 CHROME, rue du Dr Georges Salan, 30021 Nimes, France Aix-Marseille Université, CNRS-IRD-Collège de France, UM 34 CEREGE, Technopôle de l'Arbois, BP80, 13545 Aix-en-Provence, France LHyGeS Laboratory, UMR 7517, University of Strasbourg, Blessig street 1, Strasbourg cedex 67084, France

a r t i c l e

i n f o

Article history: Received 21 October 2014 Received in revised form 3 February 2015 Accepted 16 March 2015 Available online 7 April 2015 Keywords: Arsenic pollution Tailing dams Stream water 87 Sr/86Sr δ18O and δ2H Surface water–groundwater interaction

a b s t r a c t Arsenic contamination of stream waters and groundwater is a real issue in Au–As mine environments. At the Salsigne Au–As mine, southern France, arsenic contamination persists after closure and remediation of the site. In this study, natural and anthropogenic arsenic inputs in surface water and groundwater are identified based on 87Sr/86Sr, and δ18O and δ2H isotopic composition of water. In the wet season, downstream of the remediated zone, the arsenic contents in stream water and alluvial aquifer groundwater are high, with values in the order of 36 μg/L and 40 μg/L respectively, while upstream natural background average concentrations are around 4 μg/L. Locally downgradient of the reclaimed area, arsenic concentrations in stream water showed 2 peaks, one during an important rainy event (101 mm) in the wet season in May, and a longer one over the dry period, reaching 120 and 110 μg/L respectively. The temporal variations in arsenic content in stream water can be explained i) during the dry season, by release of arsenic stored in the alluvial sediments through increased contribution from base flow and decreased stream flow and ii) during major rainy events, by mobilization of arsenic associated with important surface runoff. The 87Sr/86Sr ratios associated with increasing arsenic content in stream waters downstream of the reclaimed area are significantly lower than that of the natural Sr inherited from Variscan formations. These low 87Sr/86Sr ratios are likely to be associated with the decontaminating water treatment processes, used in the past and still at present, where CaO, produced from marine limestone and therefore showing a low 87Sr/86Sr ratios, is used to precipitate Ca3(AsO4)2. The low Sr isotope signatures will then impact on the Sr isotope ratio of (1) the Ca-arsenate stored in tailing dams, (2) effluent currently produced by water treatment process and (3) groundwater draining from the overall site. Furthermore, Δ2H shows that the low 87Sr/86Sr ratio, arsenic rich water is characterized by an evaporated signature suggesting a potential influence of water impacted by evaporation during storage in decantation lagoons. This study shows the suitability of Sr and stable isotopes of water as tracers to differentiate natural and anthropogenic sources of arsenic release or other trace elements from mining context where CaO is used for water treatment. © 2015 Elsevier B.V. All rights reserved.

⁎ Corresponding author at: Univ. Nimes, EA 7352 CHROME, rue du Dr Georges Salan, 30021 Nimes, France. Tel.: +33 466 709 971; fax: +33 466 709 989. E-mail address: [email protected] (M. Khaska).

http://dx.doi.org/10.1016/j.jconhyd.2015.03.008 0169-7722/© 2015 Elsevier B.V. All rights reserved.

M. Khaska et al. / Journal of Contaminant Hydrology 177–178 (2015) 122–135

1. Introduction The environmental impacts of arsenic mining activities and their effects on ecosystems and human health are observed in stream waters and groundwater several decades after mine closure (Bates et al., 1992; Casiot et al., 2009; Daus et al., 2000; Mukherjee et al., 2008a; Wang and Mulligan, 2006; Williams, 2001; Wong et al., 1999). During mining activities, toxic elements such as arsenic are dispersed mainly by aerial transport of dusts, gas and fumes from the mine and extraction plants to the surrounding environment over distances of about 10 km or more. After mine closure, the waste rock dumps and tailing dams are impacted by mechanical and chemical weathering and generate acid mine drainages (AMD) due to oxidation of sulfur and As-bearing minerals by run-off waters and infiltrated rainwater. AMD commonly have an acidic pH and high contents of dissolved metalloids and anions such as SO4 and AsO4 (Sánchez-Rodas et al., 2005; Smedley et al., 1996). Significant amounts of As and metalloids are released by AMD into the environment (run-off and stream waters, groundwater, soils, sediments) and introduced into the food chain (Smedley and Kinniburgh, 2002; Smith et al., 2002; Vicente-Martorell et al., 2009). The As contents of AMD and stream waters impacted by AMD may reach 1 to 10 mg/L or more (Lottermoser, 2003; Smedley and Kinniburgh, 2002). Adsorption and desorption are two of the major processes controlling arsenic mobility in soils, sediments and aquifers (Dousova et al., 2012; Morin and Calas, 2006; Mukherjee et al., 2008b; Peters, 2008; Polizzotto et al., 2008). Arsenic preferentially adsorbs on iron oxides, carbonates, clays and organic materials (Smedley and Kinniburgh, 2002). The sorption processes are controlled by prevailing physico-chemical parameters mainly the redox conditions, pH, and temperature (Handley et al., 2013; Sharif et al., 2008; Welch and Lico, 1998; Zheng et al., 2004). In addition, increasing ionic strength decreases the amount of adsorbed arsenic as a result of electrostatic interactions (Bauer and Blodau, 2006; Mohan and Pittman, 2007). Reductive dissolution of arsenic bearing oxides (iron, manganese, aluminum), promotes the release of arsenic adsorbed to their surface (Bose and Sharma, 2002; Handley et al., 2013; Nickson et al., 1998; Pedersen et al., 2006). These processes have been proposed as mechanisms for arsenic mobilization in Bangladesh, Vietnam and Cambodia (Smedley and Kinniburgh, 2002). Similarly, acidification of the environment, leading to the dissolution of oxides, induces an increase in arsenic concentrations (Masscheleyn et al., 1991). The Salsigne mining district (Fig. 1a and b) located in the Orbiel River watershed (Aude, southern France) is typical of a contaminated site after mine decommissioning. The main mining activity at Salsigne began in the nineteenth century for iron, then in 1908, for gold and arsenic. The open pit and mine were closed in 1998 and 2004, respectively. Recurrent proof of arsenic contamination of stream waters, alluvial aquifers and vegetables in the Orbiel Valley, was obtained both before and after mine closure. Consequently, a large campaign of decontamination, confining and rehabilitation was carried out from 2001 to 2006, and focused on mill tailings disposal and processing plants. Nevertheless, since 2006, waters with high arsenic contents (commonly in the range of 1 to 80 mg/L) were discovered in the Orbiel Valley.

123

In such environments it is essential to differentiate between remnant anthropogenic pollution and the contribution of naturally contaminant rich surface water and groundwater. The aim of this paper is to test a suite of geochemical tracers to assess the origin of high metalloids concentrations in the vicinity of reclaimed mining sites. Common treatment processes used on mining sites are the addition of lime, both in ore and effluents treatment, and storage in decantation basins. Hence we propose to use a combination of tracers including stable isotopes of water and 87Sr/86Sr isotopic composition. The 87Sr/86Sr ratio is chosen as a tracer of groundwater origin, mixing processes, water–rock interaction, and anthropogenic contamination in environments with contrasting imprints, as presented in Faure (1986) and illustrated in recent studies (Aquilina et al., 2002; Cary et al., 2014; Deng et al., 2009; Khaska et al., 2013; Négrel and Petelet-Giraud, 2005; Petelet-Giraud et al., 2003; Pierson-Wickmann et al., 2009; Vengosh et al., 2013). The stable isotopes of water (δ18O, δ2H) provide information on the recharge processes and the origin of groundwater (Craig, 1961). Deviations from the Global Meteoric Water Line (GMWL) indicate modification of the groundwater by evaporation (Dansgaard, 1964; Faure, 1986). In this paper the As concentrations of surface water and groundwater are assessed first, both upstream and downstream of the ore processing site. Then information derived from the isotope tracers, 87Sr/86Sr ratios and 18O and 2H is investigated. Supporting observations are provided based on local alluvial sediments analyses (Section 2.1 in Supporting information). Finally, the origin and release processes of As into the stream water are discussed and a conceptual model is derived showing rough estimates of fluxes and highlighting the isotopic tracers approach. This study tracks the transport and fate of arsenic between surface and groundwater based on Sr isotopes. 2. Site overview The Salsigne Au–As–Fe deposit (Aude, southern France) (Fig. 1a and b) occurs as stock works and lodes several meters wide (4–15 m) in the southern foothills of the Montagne Noire (S1, Supplemental Information). The ore is located in late Variscan NS normal faults that crosscut stacked and southwardrecumbent nappes of low grade metasedimentary rocks. These formations are Paleozoic in age and include quartzites, schists, calc-schists and limestones (Berger et al., 1993; Demange et al., 2006). The major ore minerals of the Salsigne mining district comprise pyrite, arsenopyrite, pyrrhotite, chalcopyrite, and bismuthinite and the minor ones galena, sphalerite, scheelite and wolframite (Marcoux and Lescuyer, 1994). In the northern part of the study area, the Orbiel River crosscuts the Variscan metamorphic basement. The Orbiel Valley is therefore deeply embanked from site S1 to S5 (Fig. 1a and b) by metamorphic formations. More to the South, Paleocene and Eocene formations overlay the Paleozoic basement. There, the Orbiel Valley widens and recent alluvial deposits are increasingly abundant from site S5 to site S9 (Fig. 1a and b). The remediated zone is located at the limit of the northern and southern parts of the Orbiel Valley. The site history is complex and lime has been used throughout the ore process, fumes and effluent treatments, and as a result of this usage several lagoons, tailing dams and storage areas existed or still exist on site. The current study does not aim

124

M. Khaska et al. / Journal of Contaminant Hydrology 177–178 (2015) 122–135

Fig. 1. (a) Simplified geological map of the Orbiel basin with the rock sampling sites with the location of the main Variscan mineralized deposits (B, C, F, M, N, S and V), mining processing area and rock sampling site (R1 to R6). (b) Sampling location map for groundwater, stream waters, mining drainages and soils in the Orbiel basin.

at reconstructing the complex functioning of the site but rather focuses on the method to assess its impact on the local stream water. Hence, the site as a whole is considered as a potential source of Ca-arsenates. Currently, a Water Treatment Plant (WTP) is located at the bottom of the remediated zone along the Orbiel River. It collects highly contaminated water ([As] = 20–80 mg/L) from the reclaimed sites (ADEME, 2001) and uses lime treatment to reduce arsenic contents. On the left bank of the river a series of decantation lagoons collect drainage water from a single tailing dam. 3. Material and methods The study was performed in two steps. The first step was carried out during two campaigns in June and December 2011 respectively. The dissolved As contents of the 2011 water

samples were determined on 9 surface waters (sites S1 to S9), 12 groundwaters (sites G1 to G12) of the alluvial aquifer from upstream to downstream of the ore processing site (Table 1) and two groundwater samples were collected from the Variscan basement. Then, two monitoring sites (sites S1 and S5) were selected along the Orbiel River for monthly survey during a full hydrological cycle in 2012 from January 2012 to January 2013. In addition Orbiel water samples and three Orbiel influents (IN 1, 2 and 3 (Fig. 1b)) were collected on a major rainy event (101 mm) that occurred on the 22/05/2012 (Fig. 1b; Table 2). No more than 2 rainy events of this magnitude were recorded over the last decade (www.hydro. eaufrance.fr). The location of these two sites was chosen to correlate variations of arsenic in a pristine area (input) with those observed downstream of the mining area (output). Additional water samples were collected during a one-day

M. Khaska et al. / Journal of Contaminant Hydrology 177–178 (2015) 122–135

125

Table 1 Physicochemical parameters (T, EC, pH, Eh) and dissolved As content of surface water and alluvial groundwater for samples collected from the Orbiel valley during 2011 campaign. Sampling location

Sampling date

Depth (m)

T° (°C)

EC (μS/cm)

pH

Eh (v)

As(tot) (μg/L)

Start of dry season S1 S2 S2 S2 S3 S4 S5 S6 S7 S8 S9 G1 G2 G3 G4 G5 G7 G8 G10 G12 Vb1 Vb2

03/06/2011 03/06/2011 03/06/2011 03/06/2011 03/06/2011 03/06/2011 03/06/2011 03/06/2011 03/06/2011 24/06/2011 24/06/2011 24/06/2011 24/06/2011 30/06/2011 30/06/2011 30/06/2011 30/06/2011 30/06/2011 30/06/2011 30/06/2011 30/06/2011 30/06/2011

– – – – – – – – – – – 3 3 3 15 15 4 3 4 7 8 4

14.1 14.2 14.2 14.2 14.6 14.2 14.6 14.6 18.0 13.4 14.0 17.0 16.3 14.9 15.0 17.6 15.9 16.8 15.6 15.0 13.0 14.0

88 114 220 138 168 175 213 260 663 278 352 132 518 1127 2180 1356 358 366 346 1935.0 1053 870

7.2 6.7 6.8 6.7 7.5 7.7 7.2 7.3 7.7 7.4 7.9 6.7 6.9 7.7 7.7 7.1 7.1 6.6 6.9 7.2 7.1 6.9

0.35 0.28 0.34 0.35 0.40 0.37 0.23 0.37 0.42 0.37 0.37 0.28 0.30 0.04 0.03 0.04 0.28 0.30 0.27 0.23 0.30 0.35

3.0 3.5 2.5 3.3 4.8 5.9 27.7 33.0 42.6 30.7 20.0 4.1 9.4 547.0 18.1 20.2 39.8 17.5 11.5 1.5 2.3 1.3

Rainy season S1 S2 S4 S5 S7 S8 G1 G2 G3 G4 G5 G6 G7 G8 G9 G10 G11

08/12/2011 08/12/2011 08/12/2011 08/12/2011 08/12/2011 08/12/2011 08/12/2011 08/12/2011 12/12/2011 09/12/2011 12/12/2011 12/12/2011 12/12/2011 15/12/2011 15/12/2011 15/12/2011 15/12/2011

– – – – – – 3 3 3 15 15 3 4 3 3 4 5

10.4 10.7 11.6 12.0 13.3 12.0 12.0 14.7 12.0 14.7 13.5 14.9 13.6 11.2 16.4 13.5 16.0

86 118 160 217 360 346 116 571 1300 1842 1415 638 419 327 410 475 470

7.4 7.7 7.9 7.7 8.3 8.5 7.6 7.3 7.4 7.2 7.8 6.9 7.4 7.5 6.9 7.2 7.0

0.38 0.36 0.34 0.19 0.35 0.27 0.43 0.39 0.13 0.04 0.37 0.44 0.38 0.43 0.93 0.46 0.33

2.7 3.1 4.8 32.4 40.5 36.0 3.9 9.1 136.0 11.1 76.1 6.5 39.8 37.6 24.8 11.7 10.3

rainy event on sites S1 and S5. The water samples were analyzed for total dissolved As and water samples collected during the monthly survey were analyzed for Sr contents and for 87Sr/86Sr ratios. 6 rock samples of Variscan schists R1 to R6 (Fig. 1b) were collected and analyzed for Sr isotope composition (Table 2). Finally, sediments were collected on the first 50 cm of the alluvial terrace at 8 depths below ground surface (Section 1.1, Supporting information). The mineralogical composition of the sediments was determined using a binocular microscope. Complementary analyses were obtained by Environmental Scanning Electron Microscopy (ESEM) on both sediments and stream pebble coating. For water samples collected in 2011, the physicochemical parameters (T, EC, pH, Eh) were measured in the field on raw water in a flow-through cell, in order to prevent equilibration with the atmosphere. The domestic wells were flushed for at least 30 min before groundwater sampling. All water samples

Cl− (mg/L) 6.7 7.0 7.3 7.5 8.3 8.8 7.4

7.8 12.5 55.7 56.0 10.4 10.6 11.5

7.8 8.0 9.0 13.1

11.0 22.1 27.6 61.8 14.6 11.4 9.8 11.1 11.8 11.7

were filtered in the field through a 0.45 μm MF-Millipore membrane fitted on a Sartorius poly-carbonate filter holder, then stored below 4 °C in 60 mL HDPE bottles pre-cleaned with hot 10% HNO3 and deionized water. The filtered water samples for As and Sr analyses were acidified with ultra-pure 14 N HNO3. Dissolved As concentrations were measured by MC-ICP mass spectrometry after mineralization at the LERES laboratory (EHESP of Rennes) certified COFRAC. The uncertainty for the dissolved As concentrations was less than 22%. Sr contents were determined at the LyGeS laboratory (University of Strasbourg) using ICP-mass spectrometry with a precision of ±2% based on an indium internal standard of 20 ppb. The chemical separation of Sr, based on the method of Pin et al. (2003), was performed at the GIS laboratory (University of Nîmes) with a specific ion exchange resin (5RS-BS-S) under Class 100 laminar flow hoods in a clean room. 87Sr/86Sr ratios were determined at the GIS laboratory, using a TRITON Ti thermal ionization mass spectrometer with

126

M. Khaska et al. / Journal of Contaminant Hydrology 177–178 (2015) 122–135

Table 2 Dissolved As and Sr contents, 87Sr/86Sr ratios and δ18O–δ2H compositions of Orbiel stream water monthly sampled during 2012 at S1 and S5. 8 rock samples (7 sandy schists and 1 black limestone) were analyzed for 87Sr/86Sr. Sites

Sampling date

Site 1

07/01/2012 07/02/2012 07/03/2012 07/04/2012 07/05/2012 22/05/2012 07/06/2012 10/07/2012 07/08/2012 07/09/2012 07/10/2012 09/11/2012 07/12/2012 07/01/2013 07/01/2012 07/02/2012 07/03/2012 07/04/2012 07/05/2012 22/05/2012 07/06/2012 10/07/2012 07/08/2012 07/09/2012 07/10/2012 09/11/2012 07/12/2012 07/01/2013 22/05/2012 22/05/2012 24/06/2012 22/05/2012 30/06/2011 01/07/2011 30/06/2011 12/12/2011 01/07/2011 11/12/2011 30/06/2011 12/12/2011

As

Sr

87

Sr/86Sr

33 34 33 35 40 27 34 46 52 53 48 44 37 30 43 50 50 56 53 77 47 70 104 110 95 70 59 52 – 754 88 29 663

0.71445 0.71496 0.71491 0.71455 0.71471 0.71475 0.71485 0.71483 0.71489 0.71498 0.71489 0.71510 0.71506 0.71489 0.71400 0.71361 0.71345 0.71311 0.71367 0.71228 0.71366 0.71328 0.71285 0.71288 0.71299 0.71310 0.71368 0.71366 0.71542

(μg/L)

Site 5

IN1 IN2 IN3 IN4 Vb1 Vb1 G3 G3 G4 G4 G5 G5 CaO Ca-arsenate Ca-arsenate Ca-arsenate Ca-arsenate Schist (Orbiel pebble) Schist (Orbiel pebble) R1 (lower Cambrian schist) R2 (lower Cambrian schist) R3 (lower Cambrian schist) R4 (lower Cambrian schist) R5 (middle Cambrian schist) R6 (lower Cambrian limestone)

2.4 2.5 2.4 2.6 2.9 6.4 3.1 3.9 5.2 3.5 3.5 3.9 2.8 2.8 13.9 20.9 22.8 24.0 18.1 120.5 17.0 48.0 101.3 110.0 89.9 48.1 19.2 20.5 126 414 25 1.3 2.3 547 136 18 11 20 76

an internal normalization for isotope fractionation to an 86 Sr/88Sr ratio of 0.1194. For the entire chemical procedure, the Sr total blank was less than 1 ng. A typical internal precision of ± 0.000005 (2σ) was obtained for the 87Sr/86Sr ratios and was always better than 0.000009. Rock samples were powdered and 10 mg was dissolved for 87Sr/86Sr ratio analysis. Over the course of the study, repeated analyses of the NBS987 standard were conducted to test the reproducibility of the 87 Sr/86Sr ratio measurements, and gave a mean 87Sr/86Sr value

330 308 5066 3630 2600 168,714 154,101

2σ (×10−6)

0.71558 0.71527 0.71526 0.71548 0.71310 0.71227 0.71299 0.71123

6 8 7 8 6 6 6 6 4 5 6 7 9 6 4 6 4 4 4 4 5 4 5 6 4 5 3 6 7 6 6 8 5 7 5 4 4 8

0.71354 0.70759 0.71010 0.71024 0.71076 0.71092 0.71427 0.71601 0.71884 0.71856 0.71674 0.71638 0.72260 0.71489

7 11 6 8 8 9 6 7 6 5 4 3 4 5

δ18O

δ2H

Δ2H

(‰ vs. SMOW) −7.1 −7.4 −7.3 −7.3 −7.2 −7.1 −7.2 −7.1 −7.1 −6.9 −7.0 −6.8 −6.9 −7.1 −7.0 −7.2 −7.3 −7.2 −7.2 −6.6 −7.2 −7.1 −6.8 −6.9 −6.9 −6.7 −6.8 −7.0

−44.4 −46.7 −46.4 −46.4 −46.0 −45.3 −46.4 −45.3 −46.0 −44.2 −44.6 −42.6 −42.6 −44.9 −44.0 −46.2 −46.1 −45.6 −46.0 −43.0 −46.0 −44.8 −43.8 −44.2 −44.0 −42.2 −42.5 −44.2

−6.4 −6.1 −6.1 −6.6 −6.5 −5.6

−41.7 −40.0 −40.8 −41.3 −41.0 −37.4

−0.67 −0.77 −0.52 −0.17 0.01 0.26 0.10 0.26 0.51 0.16 −0.02 −0.71 −1.12 −0.66 −0.48 −0.09 −0.29 −0.48 −0.14 1.17 0.07 −0.22 0.81 0.54 0.41 0.22 −0.34 −0.45

of 0.710245 ± 6 × 10−6 (2σ, n = 65). 87Sr/86Sr ratios are reported with 5 significant digits. A Picarro L2130-I liquid-vapor isotope analyzer was used for δ18O and δ2H measurements at the GIS laboratory. Internal water standards calibrated with respect to V-SMOW and SLAPP were used as working standards. The overall uncertainties are between ±0.1 and ±0.4‰ for δ18O and δ2H respectively. Replicates of deionized water (n = 90; 2σ) gave internal uncertainties for δ18O and δ2H of the same order of magnitude.

M. Khaska et al. / Journal of Contaminant Hydrology 177–178 (2015) 122–135

4. Results 4.1. As contents in surface and groundwater Large variations in the As contents were observed both in stream waters and in alluvial groundwater (Fig. 2a and b). The observed trends are similar over both campaigns. The As contents of stream water samples from the northern Orbiel

50

Stream water (June-July 2011 ) Stream water (December 2011) Drinking Water Guidelines Russec stream water

Dissolved As (µg/L)

40

a S7

30

S5 20

S9 DWG

10

S1 0 0

2

4

6

8

10

12

14

16

Distance (km) 50

Alluvial groundwater (June-July 2011 )

b

Alluvial groundwater (December 2011)

G7

Dissolved As (µg/L)

40

Groundwater of unmineralized Paleozoic basement

G8

30

G9 20

DWG

10

Vb1

a 0 0

2

G12

Vb2

4

6

8

10

12

Distance (km) 1000

c

Dissolved As (µg/L)

G3

G3

100

G5

G7, G8, G9

Valley were low, ranging from 2.5 to 5.9 μg/L, remaining below the World Health Organization (WHO) Drinking Water Guidelines (DWG) of 10 μg/L for As (WHO, 1993, 2011). From site 5 onward, the As concentrations of stream waters in the southern Orbiel Valley increased by a factor of 10 with values ranging from 28 to 42 μg/L (Fig. 2a). These values are in good agreement with the average As contents of 8 μg/L and 45 μg/L reported for Orbiel stream water samples collected upstream and downstream from the remediated zone during the 2009–2010 survey (Girard, 2011). Further downstream, the As contents in the Orbiel stream water decreased slightly down to 20 μg/L at site S9 (Fig. 2a). Surface water was also sampled from the Russec (S7), as the main tributary of the Orbiel stream (Fig. 1b). There, arsenic concentrations were of 40 μg/L, falling in the same range as in the Orbiel stream. A similar trend in the As contents was observed for groundwater with low arsenic content (4 to 10 μg/L) upstream from site 5 and higher As concentrations downstream reaching values of 40 μg/L. Marked high anomalies of 547 and 136 μg/L are noticeable at site G3 (Fig. 2c), which is located between the Orbiel River and the remediated zone (Fig. 1b). Then in the most southerly sites of the transect, groundwater As concentrations decreased down to 1.5 μg/L similar to upstream. In contrast, the As concentrations of groundwater collected in the Variscan basement (sites Vb1 and Vb2), east and upgradient of the ore processing site (Fig. 1b), are low ranging from 1.3 to 2.3 μg/L. Upstream from the ore processing site, groundwater Eh range from 0.28 to 0.40 V vs. normal hydrogen electrode (NHE), while locally, at site 5, groundwater Eh decreases down to 0.03 V (Fig. 4). Eh in the Orbiel water also decreases slightly at site S5 dropping from 0.37 down to 0.19 V vs. NHE. Down gradient from site 5, Eh rises back to incoming values. Under prevalent Eh and pH conditions the predominant As species in solution is As (V) (Fig S1, Supporting information). Concomitantly, in the vicinity of site S5, in the bed of the Orbiel River, the pebbles are covered with a red to ochraceous coating of As bearing iron-oxyhydroxide. These red pebbles are not observed upstream, at site S4, nor downstream from site S6 onwards. Finally, Arsenic is omnipresent in the alluvial environment of site S5. The leachate from the sediment sample collected there shows high arsenic concentrations (Table S2, Supporting information) with low 87Sr/86Sr ratios. This suggests that locally the sediments of the alluvial aquifer are highly impacted by the imported Sr and high As concentrations (see Section 2.1, Supporting information). 4.2. Temporal variations of As concentrations in surface water

10

G1 G12 1 0

127

2

4

6

8

10

12

Distance (km) Fig. 2. (a) Dissolved As contents in the Orbiel stream water from sites S1 to S9. (b) Dissolved As contents in the Orbiel groundwater from sites G1 to G12, (c) Reduction of (b) shows the high As contents observed in the environment of the site G3 highlighted in gray. The dissolved As contents of groundwater from the Russec and Orbiel influent (sites G7, G8, G9) and of groundwater from the unmineralized Variscan basement (sites Vb1 and Vb2) are indicated for comparison.

At site S1, the dissolved As contents of the Orbiel stream waters recorded over the one year remained low, varying only slightly from 2.4 to 6.4 μg/L (Fig. 3, Table 2) and representing the background concentrations. In contrast, the dissolved As contents at site S5, were higher and more variable than those sampled at site S1, ranging from 13.9 to 120.5 μg/L (Fig. 3, Table 2). At this site, arsenic concentrations showed base line values around 20 μg/L and revealed two peaks: i) the first As peak was sudden, very brief, and clearly related to the rainy event of 22/05/2012, marked by a high As content (120.5 μg/L); ii) the second As peak rose slowly and lasted over the five

128

M. Khaska et al. / Journal of Contaminant Hydrology 177–178 (2015) 122–135

a

Upstream (S1) Downstream (S5) Daily precipitation record

22/05/2012

120

120

100

Dissolved As (µg/L)

100 80

As peak 1 80

As peak 2

60

60 40 40

Daily precipitation (mm/day)

140

20

20

0 01/12

0 02/12

03/12

04/12

05/12

06/12

07/12

08/12

09/12

10/12

11/12

12/12

01/13

Fig. 3. Monthly variations in dissolved As contents of the Orbiel stream water (from January 2012 to January 2013) at sites S1 and S5.

months of the dry season. Both As peaks were also noticeable at site S1, but in much lower concentrations (Fig. 3). 4.3. Strontium isotopes as a proxy for arsenic source A search was made for a suitable isotope tracer likely to differentiate natural from anthropogenic As inputs. Sr isotopes could provide appropriate tracers due to the use of CaO in the past and at present on the site. CaO is a Sr-rich compound, produced by calcination of marine limestones. Such a compound and its derived whitewash should provide lower Sr isotopic compositions than those of the Variscan schists from the upper Orbiel Valley (Capo et al., 1998; Faure, 1986). To remove arsenic from the contaminated waters whitewash has been used to co-precipitate arsenic as Ca3(AsO4)2 at pH 12.5 (Dutré and Vandecasteele, 1995; Gao et al., 2006; Martínez-Villegas et al., 2013; Román-Ross et al., 2006; Vandecasteele et al., 2002). Ca3(AsO4)2 has been stored on site and is still currently produced by the WTP. Dissolution of

0.35

Upstream

Downstream

0.30

Eh (V)

0.25

S5

0.20 0.15 0.10

G3

0.05

G4

G5

0.00 0

10

20

30

40

50

60

Distance/Upstream (km) Fig. 4. Variations of ORP in alluvial groundwater showing the occurrence of a reducing zone in the alluvial aquifer at the mining site.

calcium arsenate from mining residues has been shown to provide high As levels (Martínez-Villegas et al., 2013). Currently, the decontaminated water released in the Orbiel environment by the WTP as well as groundwater seeping from this site should provide an intermediate 87Sr/86Sr ratios between those of CaO and the Variscan basement rocks. Hence the Sr isotopic composition should therefore serve as a suitable tracer to differentiate between natural and anthropogenic arsenic sources in the Salsigne mining district. The 87Sr/86Sr ratios and Sr contents of the stream water samples collected in 2012 are compared to those of the Variscan schists (Table 2; Fig. 6b). In addition, in order to simulate the 87Sr/86Sr ratios of the Ca3(AsO4)2 produced on site, a laboratory test was conducted. CaO lime, similar to that used on site, was added to arsenic rich water collected from mining waste dumps when the liquid-tosolid (L/S) ratio is fixed at 1.7. The 87Sr/86Sr ratios of the precipitated Ca3 (AsO4)2 (further referred to as laboratory derived Ca-arsenate) range from 0.71010 to 0.71092 (Table 2). Both WTP effluent and Ca-arsenate precipitate are expected to show the same 87Sr/86Sr ratio as Sr isotopes do not fractionate in the environment (Faure, 1986). At site S1, the 87Sr/86Sr ratios of stream water samples were high and ranged from 0.71444 to 0.71510, (Fig. 5). In these schists, Rb-rich minerals such as biotite and muscovite have accumulated 87Sr since Variscan times by radioactive decay of 87 Rb. At present, the weathering of these minerals and of kaolinite, newly formed in situ from micas and feldspars, provides a radiogenic Sr signature to run-off waters that feed the Orbiel watershed (Blum et al., 1993). Indeed, bulk rock analyses of the Variscan schists of the upper Orbiel Valley provided similar radiogenic 87Sr/86Sr ratios ranging from 0.71427 to 0.72260 (Table 2; Fig. 6b). Similarly, Orbiel influent stream water measured downstream from the mining area showed high 87Sr/86Sr ratios, ranging from 0.71346 to 0.71558 (Table 2 Fig. 6b). In contrast, the 87Sr/86Sr ratios of Orbiel stream waters sampled at site S5 were clearly less radiogenic and more

M. Khaska et al. / Journal of Contaminant Hydrology 177–178 (2015) 122–135

200 175

129

0.716 Dissolved As (S5)

Dissolved As (S1)

Sr isotopic composition (S5)

Sr isotopic composition (S1)

As peak1 125

As peak 2

87Sr/ 86Sr

Dissolved As (µg/L)

0.715 150

0.714

100 0.713

75 50

0.712 25 0.711

0

Fig. 5. Dissolved As content and 87Sr/86Sr ratio of Orbiel stream water collected monthly at sites S1 and S5. Note the inverse relationship of As content and 87Sr/86Sr ratios at site S5.

variable over time (0.71227 ≤ 87Sr/86Sr ≤ 0.71400) than those measured at site S1. The lowest 87Sr/86Sr ratio of the Orbiel stream was observed at site 5 during the 22/05/2012 rainy event concomitant to the As peak 1 (Fig. 5). Two weeks after this rainy event, both the As and Sr concentrations went back to base line values, i.e., similar to those recorded before this event,

during the five months of the dry season corresponding to the second As peak period, a new inverse relationship was recorded. During that period, the evolution of the Sr isotopic composition was gradual, as was that of the As contents, but the 87 Sr/86Sr ratios did not reach the lowest radiogenic values observed during the first As peak (Fig. 5).

0.7155

a

L1

0.7150 0.7145 0.7140 As peak 2

87Sr/86Sr

0.7135

Downstream (S5) Upstream water (S1)

0.7130

D5B

Basement groundwater - Vb2 Variscan schists and limestone

0.7125

Orbiel alluvial groundwater

As peak1

Ca- arsenate

0.7120

Variscan schists

Rainwater

0.7115

Mine drainage Variscan schists water Site1

D5A

0.7110 0.7105

Site 5 Alluvial aquifer groundwater Ca- arsenate

Ca- arsenate C 0.7080

0.7120

0.7160 87Sr/86Sr

2000

2500

0.7100 0

500

1000

1500

b 0.7200

3000

0.7240

3500

1/Sr (mmol/L) Fig. 6. (a) Sr mixing diagram showing the influence of Ca-arsenate (D5A), alluvial groundwater and soil samples (D5B). The water rock interaction between rainwater and Variscan basement rocks is represented by the line L1. (b) Variations of 87Sr/86Sr in water, rock and other compartments of the Orbiel Valley.

130

M. Khaska et al. / Journal of Contaminant Hydrology 177–178 (2015) 122–135

On the 87Sr/86Sr vs. 1/[Sr] mixing diagram (Fig. 6), all stream water samples from site S1 fall along the L1 regression line with a y-intercept of 0.71515, which suggests binary mixing. The least radiogenic end-member of this mixing may be the local rainwater with 87Sr/86Sr ratios ranging from 0.70919 and 0.71314 (Négrel and Roy, 1998). The most radiogenic endmember is most likely the Sr released from weathered Variscan metasedimentary formations of the upper Orbiel Valley, in agreement with the 87Sr/86Sr ratios of the Variscan schists and the groundwater hosted by these schists (Table 2). At site 5 the surface water signature, collected during the rainy season (from January to June 2012) i.e., not affected by the winter or summer peaks, plots in a tight cluster slightly less radiogenic than that of site 1 (Fig. 6). Those and the samples collected during the 22/05/2012 rainy event define a mixing domain D5A between two end-members, the most radiogenic being the S5 base line stream water and the least radiogenic being the Ca-arsenate. Stream waters sampled during the dry season define a second mixing domain D5B following a mixing trend between background water at site S5 and the local alluvial groundwater as defined by sites G3, G4 and G5. The 87 Sr/86Sr ratios of the local alluvial aquifer as observed on G3, G4 and G5 are intermediate between those of CaO and those of the Variscan basement rock. 4.4. δ18O and δ2H of arsenic contaminated water Orbiel stream waters collected in 2012 at sites S1 and S5 and groundwater, located nearby the site S5 and sampled in 2011, were analyzed for δ18O and δ2H compositions (Table 2). The data are in agreement with the average isotopic composition of rainwater estimated for the studied area by Millot et al. (2010). The observed isotopic composition at the site S1 and S5 apparently falls within the same range with only slight variations over time (from −7.4 to −6.7‰ vs. V-SMOW and from −46.7 to −42.2‰ vs. V-SMOW for δ18O and δ2H respectively). Set apart the positive anomaly linked to the 22/05/2012 rainy event, δ18O and δ2H variations over time could reflect the influence of seasonal variations of the O and H isotopic compositions of meteoric waters with some enrichment in heavy isotopes in summer. In addition, the δ18O and δ2H measured at site S5 show slightly, but systematically, higher contents than those measured at site S1 (~ +0.15 and +0.5 in δ18O and δ2H respectively). This suggests the influence of a more enriched end-member. While those variations remain close to the analytical uncertainty, having been measured on the same run the systematic shift can be considered as representative. Water samples corresponding to the As peak related to the 22/05/2012 rainy event (Fig. 7a and b) are also marked by a significant enrichment (+0.6‰ and +5 in δ18O and δ2H respectively). On a δ2H vs. δ18O diagram (Fig. 7c and d) most of the Orbiel stream water samples plot near the Local Meteoric Water Line (LMWL) (δ2H = 7.4 (±0.1) × δ18O + 7.3 (±0.7)) (Payne, 1991). While, groundwater from sites G3, G4 and G5 plots below the LMWL on a regression line with a slop of 5.0 (RL, Fig. 7c), this implies the influence of water affected by an evaporation process. Located downwards of the tailing dams and ponds, this groundwater signature suggests that the decantation basins may hold a pool of water affected by evaporation.

5. Discussion 5.1. Evidence of contamination based on dissolved As variations

87

Sr/86Sr ratios with

At site S1, radiogenic Sr ratios (0.71444 ≤ 87Sr/86Sr ≤ 0.71510) and relatively low dissolved As contents (2.4 to 6.4 μg/L) in the Orbiel stream water represent the regional geochemical background. At site S5 the less radiogenic and more variable 87Sr/86Sr ratios (0.71227 ≤ 87Sr/86Sr ≤ 0.71400) and higher and more variable dissolved As contents (13.9 ≤ [As] ≤ 120.5 μg/L) attest of the overall site impact. At site S5, the As peak associated with low 87Sr/86Sr ratio during the 22/05/12 rainy event (reaching values of 120.5 μg/L and 0.71228, respectively (Fig. 5)) suggests a notable release of dissolved As associated with non radiogenic Sr into the Orbiel River at site S5. In the Sr mixing diagram (Fig. 6), the endmembers of the mixing (D5A) are the stream water not affected by the winter or summer peaks and an end-member reflecting the influence of the Ca-arsenates used on site in the water treatment process. All these results suggest that the As peak 1 was mainly of anthropogenic origin and arose directly from the nearby remediated ore processing area. Sources for high As and low 87Sr/86Sr ratio may be leaching by surface runoff of contaminated alluvial sediments (Berg et al., 2008; Polizzotto et al., 2008). This process may be enhanced by dissolution of As rich carbonates such as calcium arsenates material in the superficial horizon of the alluvial deposits (cf Section 2.1, Supporting information). Furthermore, the leachate of the alluvial sediments shows similar Sr composition to that of the alluvial groundwater (Fig. S3, Supporting information). The dissolution of carbonate may be enhanced by initially slightly acidic rainfall and surface runoff (Bossy et al., 2012). Subsequently, high pH (N8.5) of sediment leachate (Section 2.1, Supporting information) is expected to favor desorption of As (Peters, 2008; Smedley and Kinniburgh, 2002). As seen before in the Sr mixing diagram, surface water samples collected during the dry season and corresponding to the As peak 2 fall within the second mixing domain (D5B) (Fig. 6). It was observed that the D5B tend toward an endmember slightly more radiogenic than that of the D5A mixing domain. That increase in the 87Sr/86Sr ratio was explained by a more pronounced influence of the Sr signature of detrital materials issued from the erosion of the Variscan rocks. This implies a slightly different pathway for anthropogenic As where water circulate through the local alluvial sediment terraces and where the Sr isotope signature is slightly increased by contact with the detrital sediments issued from weathering and erosion of Variscan rocks. High arsenic concentration of the 50 cm of sediment leachate and intermediate 87Sr/86Sr ratio (Fig S3 and Table S2, Supporting information) suggest that sediments act either as a pathway or as a secondary source of anthropogenic As. The occurrence of As bearing iron-oxyhydroxide coating the river bed attests that Fe(II)-rich groundwater in reductive conditions discharges in the oxygenated stream water which subsequently generates co-precipitation of As with Fe oxyhydroxides. It is further underlined that no high As concentration is observed in groundwater of the Variscan schists (sites Vb1 and Vb2; Fig. 1b; Table 1) outcropping north of the reclaimed area,

M. Khaska et al. / Journal of Contaminant Hydrology 177–178 (2015) 122–135 -41.0 22/05/2013

(‰ vs SMOW)

-6.8 -7.0 -7.2

2H

Downstream (S5)

-43.0

-44.0

-45.0

Downstream (S5)

-46.0

Upstream (S1)

Upstream (S1)

-36

c

-42

09/12

08/12

07/12

06/12

05/12

04/12

03/12

d

-40 -42 -44 Upstream (S1) Downstream (S5) Alluvial groundwater Local meteoric water line (LMWL) Groundwater regression line (RL)

-46

-7.1

-6.6 18O

-6.1

-5.6

(‰ vs SMOW)

As peak1 (S5), 22/05/2012

-44

-45 S1, 22/05/2012

2H

(‰ vs SMOW)

-43

2H

(‰ vs SMOW)

-38

-48 -7.6

02/12

-47.0

01/12

01/13

12/12

11/12

10/12

09/12

08/12

07/12

06/12

05/12

04/12

03/12

02/12

01/12

-7.6

01/13

-7.4

22/05/2013

11/12

(‰ vs SMOW) 18O

b

-42.0

-6.6

12/12

a

10/12

-6.4

131

-46 Average of Orbiel stream waters at S5 Average of Orbiel stream waters at S1

-47 -7.5

-7.3

-7.1 18O

-6.9

-6.7

-6.5

(‰ vs SMOW)

Fig. 7. Monthly variations in δ18O (a) and δ2H (b) of the Orbiel stream water at S1 and S5. δ2H vs. δ18O diagram (c) shows the influence of local groundwater (G3, G4, G5) on surface water at S5. (d) Enlargement of (c) showing the end member of the mixing lines. ML1 and ML2 represent mixing lines between groundwater and average δ18O at site S1 and δ18O during the 22/05/2012 rainy event respectively.

on the left bank of the Orbiel River. This enables to disregard significant direct arsenic contamination from the Variscan basement groundwater in this area. 5.2. Impact of evaporated waters on groundwater traced by δ2H and δ18O The small enrichment in δ18O and δ2H observed at site S5 with respect to site S1 (Fig. 7) can be explained considering the δ18O and δ2H compositions of local alluvial groundwater (Craig, 1961). On the δ2H vs. δ18O diagram (Fig. 7 c and d) a mixing line (ML1, Fig. 7d) can be defined between groundwater contaminated by arsenic from wells located nearby the tailings dams (sites G3, G4 and G5) and the average of the water samples from site S1. Average surface water at site S5 plots along that mixing line confirming the influence of groundwater at that site. Such mixing can also be observed on a single event of the 22/05/2012 represented by a second mixing line (ML2, Fig. 7d). This mixing trend suggests the influence of highly enriched waters having undergone an evaporation process and probably originated from the decantation basins or ponds, collecting the drainage water from

tailing dams. This finding suggests that (1) in the close environment of site 5, groundwater in alluvial sediments on the left bank of the Orbiel River may be impacted by drainage waters issued from decantation ponds, in agreement with the high As contents measured in the groundwater, and that, (2) the difference between the δ2H and δ18O averages of stream water samples at sites S1 and S5 may be due to a small contribution of groundwater affected by drainage waters from the lagoons (Table 2; Fig. 7 d). 5.3. Arsenic coupled with 87Sr/86Sr ratio and Δ2H In summary, in the environment of the decommissioned ore processing site, it was shown that the sharp increase in As concentration in both surface and groundwater is associated with i) low Sr isotope ratio, enabling us to identify the influence of Ca-arsenates, and with ii) δ18O and δ2H enriched water enabling us to trace the influence of water marked by evaporation. The plots of 87Sr/86Sr vs. As, 87Sr/86Sr vs. Δ2H and As vs. Δ2H enable to confirm the observed trends (Fig. 8 a, b and c), where Δ2H is the deuterium excess (Δ2H = δ2H − (7.4 × δ18O) (after

132

M. Khaska et al. / Journal of Contaminant Hydrology 177–178 (2015) 122–135

0.7155

a

0.7150

Upstream S1 Downstream S5

87Sr/86Sr

0.7145 0.7140

F½xDO ¼ F½xUP þ F½xWTP þ F½xGW

0.7135

Peak 2

Q DO ¼ Q UP þ Q WTP þ Q GW : 0

50

100

150

As (µg/L) 9

b

8 vs SMOW)

ð3Þ

Peak 1

0.7120

2H (‰

ð1Þ

Q DO  ½xDO ¼ Q UP  ½xUP þ Q WTP  ½xWTP þ Q GW  ½xGW ð2Þ

0.7130 0.7125

Where [x] and Q are the concentration and the flow respectively. The abbreviations DO, UP, WTP, GW stand for downstream, upstream, water treatment plant and groundwater respectively. To evaluate groundwater flow, a simple Cl mass balance is established for the stream section between site S1 and site S5, upstream from the mining site and just downstream of the ore processing area respectively. Combining Eq. (3) in Eq. (2), one obtains:

7

Q GW ¼



Q UP  ½xDO −½xUP þ Q WTP  ½xDO −½xWTP
: ½xGW −½xDO

ð4Þ

6 Upstream S1

5 0

50

100

150

As (µg/L)

0.7155

c

Upstream S1

0.7150

Downstream S5

0.7145 87Sr/86Sr

Outside of rainy events, at least three sources contribute to both the flow and the As flux at site 5: incoming stream flow, groundwater flow and water treatment plant outlet. Hence flow and tracer fluxes can be written as follows:

0.7140 0.7135 Peak 2

0.7130

Peak 1

0.7125 0.7120 5.0

6.0

7.0 2H (‰

8.0

9.0

vs SMOW)

Fig. 8. Increasing As concentrations (a) and decreasing 87Sr/86Sr ratios (b) related to δ2H enrichment, resulting probably from the influence of evaporation of basin waters. Evolution of 87Sr/86Sr ratios with Δ2H enrichment (c). The dashed and dash-dot curves represent mixing curves.

Dansgaard, 1964) used to show the influence of mixing with an end-member having undergone evaporation process. In the 3 graphs, incoming surface water at site S1 plot in a cluster characterized by relatively low As concentrations, high 87 Sr/86Sr, and low deuterium excess while surface water at site S5 clearly tends toward high As concentration, low 87 Sr/86Sr ratio and high deuterium excess characteristic of the observed local alluvial groundwater. 5.4. Conceptual model and mass balance of arsenic flux A simple conceptual model of arsenic fluxes is attempted based on the current available knowledge (Fig. 9).

To take into account the observed seasonal variations, 2 mass balances are established between site S1 and site S5, one for the low and one for the high flow period as described below. Average flow rates monitored at the gauging station (Fig. 1b) range from 4.1 to 6.1 m3/s for low and high-flow conditions respectively (data from ADES database during the period between 2001 and 2010; http://www.hydro.eaufrance. fr/). For high flow conditions, chloride concentrations in stream water are taken from the measurement of the first campaign in June and December 2011, at site S1 and S5 (Table 1). For low flow conditions, chloride concentration in stream water was estimated taking into account the evapotranspiration coefficient calculated from Meteo France weather station which leads to a concentration factor of 1.4 between the wet and dry season. Those chloride concentration estimates are similar to those monitored by Marchal (2007). Groundwater Cl concentrations are derived from the alluvial aquifer data (G3, G4 and G5) both for low and high flow conditions (Table 1; Table S3 in Supporting information). Based on this data, the obtained groundwater flow is in the order of a few hundreds of liters per second varying between 0.6 and 0.2 m3/s for low and high flow conditions respectively. Those estimates are coherent with the observed stream flow increase across the site as evidenced by direct flow measurement during 1999 and 2000 (ADEME, 2001). Arsenic concentrations, in stream water, considered for the calculations of arsenic fluxes are derived from the monthly monitoring at sites 1 and 5 for low and high flow period (Table 2) and range between 2.7 and 3.8 at site S1, and between 19.5 and 69.5 at site 5 (Table S3, Supporting information). Average arsenic concentrations in the alluvial groundwater are derived from G3, G4 and G5 piezometers data (Table 1), completed by five years monthly monitoring data from 4 wells located on the ore processing site (Table S3, Supporting information) (Fig. 1a) (Nguyen et al., 2007 and data ADES database http://ww.ades.eufrance.fr/). The discharge of the WTP is ~25 m3/h in average with an average arsenic concentration of 0.94 mg/L (ADEME, 2001; Girard, 2011).

M. Khaska et al. / Journal of Contaminant Hydrology 177–178 (2015) 122–135

133

Fig. 9. Conceptual hydro-geochemical model for As pathways in the river-aquifer system at the study site. The model provides estimate of the As flux (a), and the geochemical processes for As increase in surface and groundwater (b).

The coherence of this model was verified by calculating the theoretical concentrations of arsenic in stream water at site 5 based on the calculated flows. The values obtained for arsenic concentrations are similar to those observed at site 5 (Table S3, Supporting information). This model suggests that the groundwater arsenic flux is in the range of some hundreds of grams per hour (Table S3, Supporting information), about 5 to 17 times higher than that of the incoming surface water. This attests that the major component of anthropogenic arsenic in stream water at site 5 is derived from the groundwater flux in low and high flow

periods. In contrast the flux of the WTP ([As] b 0.025 kg/h) appears to be negligible with respect to the others. The model may be further refined with continuous monitoring to improve the arsenic fluxes estimate. Based on current observation a simple geochemical system governing As release to surface water can be proposed. The prevalence of reducing conditions observed both in groundwater and in minor proportion in stream water at the ore processing site, as well as iron coating on the stream bed, attests of groundwater discharge. In reduced condition iron occurs in groundwater as Fe (II) in solution and both Fe (II) and

134

M. Khaska et al. / Journal of Contaminant Hydrology 177–178 (2015) 122–135

As are discharged to stream water explaining the observed As concentration increase at site 5. Enhanced proportion of base flow with decreasing stream flow during the dry season induces more pronounced As concentrations in summer (data from ADES database during the period between 2001 and 2010; http://www.hydro.eaufrance.fr/). Once in the stream and in a more oxygenated environment, less reductive conditions induce oxidation of Fe(II) to Fe (III) and precipitate of Fe-oxyhydroxide, as attested by iron coating on stream pebbles. Finally, arsenic may co-precipitate with the reduced iron forming As bearing Fe-oxyhydoxide precipitate as observed on the SEM spectrum of the pebble coating. Further investigations would be required to refine the redox processes involved. During the period of storm event, arsenic release processes may include shear leaching of As rich surfaces and sediments. 6. Conclusions This study showed the potential for using the 87Sr/86Sr, δ18O, δ2H to track the origin of remaining arsenic released after closure of a Au–As mine, where lime water treatment is extensively used. Low 87Sr/86Sr ratio shows the influence of the lime used in the water treatment process, producing Ca-arsenates contrasting with the radiogenic 87Sr/86Sr ratio provided by crustal Variscan rocks. Increasing As contents associated with decreasing 87Sr/86Sr ratio confirms the anthropogenic origin of As issued from Ca-arsenates stored on site. Sr isotope ratio can therefore constitute a suitable tracer to differentiate the natural from the anthropogenic origin of arsenic in stream waters or groundwater when the solid and/or liquid effluents from mining activities contain part of the arsenic as Ca-arsenates. δ2H and δ18O provide useful additional tracers of water marked by evaporation process and possibly issued from decantation ponds collecting drainage waters from the site. Those tracers represent powerful tools to further investigate the origin of As at this site and other mining sites in similar context. Hence, we demonstrate that high arsenic concentrations in surface and groundwater are mostly associated with leaching of Ca-arsenate issued from water treatment process. The As peaks are due to a higher proportion of base flow in the dry season and to As mobilization by surface runoff during heavy rainfall events. Acknowledgments M. Khaska was awarded a doctoral scholarship funded by the University of Tishreen. This research is a part of the ongoing research program «Transdisciplinary approach to environmental risks» funded by the University of Nîmes. The Orbiel Valley inhabitants and Salsigne ancient miners are thanked for their warm welcome, informative discussions and assistance in the field to sample groundwater. Jean-Marie Taulemesse of the Alès School of Mines is thanked for technical support in the ESEM analyses. The mayor of Conques-sur-Orbiel is thanked for his authorization to collect water samples from the municipal well. We thank Jemma Burton and Diana Warwick for English editing.

Appendix A. Supporting information Material and methods and additional details on the historical mining and remedial campaign, sampling and analyses, leaching experiments and arsenic speciation can be found online at http://dx.doi.org/10.1016/j.jconhyd.2015.03.008. References ADEME, 2001. Etude hydrolique de la zone de SEPS pour la matrice des eaux pluviales, réhabilitation du site de la Combe du saut, R-EEG- 01 -GP −009. Aquilina, L., Ladouche, B., Doerfliger, N., Seidel, J.L., Bakalowicz, M., 2002. Origin, evolution and residence time of saline thermal fluids (Balaruc springs, southern France): implications for fluid transfer across the continental shelf. Chem. Geol. 192, 1–21. Bates, M.N., Smith, A.H., Hopenhayn-Rich, C., 1992. Arsenic ingestion and internal cancers: a review. Am. J. Epidemiol. 135, 462–476. Bauer, M., Blodau, C., 2006. Mobilization of arsenic by dissolved organic matter from iron oxides, soils and sediments. Sci. Total Environ. 354, 179–190. Berg, M., Trang, P.T.K., Stengel, C., Buschmann, J., Viet, P.H., Van Dan, N., Giger, W., Stüben, D., 2008. Hydrological and sedimentary controls leading to arsenic contamination of groundwater in the Hanoi area, Vietnam: the impact of iron–arsenic ratios, peat, river bank deposits, and excessive groundwater abstraction. Chem. Geol. 249, 91–112. Berger, G., Boyer, F., Debat, P., Demange, M., Freytet, P., Marchal, J., Mazeas, H., Vautrelle, C., 1993. Notice explicative, Carte géol. France (1/50 000), feuille Carcassonne (1037). BRGM Ed. Orléans; 1993. Blum, J.D., Erel, Y., Brown, K., 1993. 87Sr/86Sr ratios of sierra Nevada stream waters: implications for relative mineral weathering rates. Geochim. Cosmochim. Acta 57, 5019–5025. Bose, P., Sharma, A., 2002. Role of iron in controlling speciation and mobilization of arsenic in subsurface environment. Water Res. 36, 4916–4926. Bossy, A., Grosbois, C., Hendershot, W., Beauchemin, S., Crouzet, C., Bril, H., 2012. Contributions of natural arsenic sources to surface waters on a high grade arsenic-geochemical anomaly (French Massif Central). Sci. Total Environ. 432, 257–268. Capo, R.C., Stewart, B.W., Chadwick, O.A., 1998. Strontium isotopes as tracers of ecosystem processes: theory and methods. Geoderma 82, 197–225. Cary, L., Benabderraziq, H., Elkhattabi, J., Gourcy, L., Parmentier, M., Picot, J., Khaska, M., Laurent, A., Négrel, P., 2014. Tracking selenium in the Chalk aquifer of northern France: Sr isotope constraints. Appl. Geochem. 48, 70–82. Casiot, C., Egal, M., Elbaz-Poulichet, F., Bruneel, O., Bancon-Montigny, C., Cordier, M.-A., Gomez, E., Aliaume, C., 2009. Hydrological and geochemical control of metals and arsenic in a Mediterranean river contaminated by acid mine drainage (the Amous River, France); preliminary assessment of impacts on fish (Leuciscus cephalus). Appl. Geochem. 24, 787–799. Craig, H., 1961. Isotopic variations in meteoric waters. Science 133, 1702–1703. Dansgaard, W., 1964. Stable isotopes in precipitation. Tellus 16, 436–468. Daus, B., Mattusch, J., Paschke, A., Wennrich, R., Weiss, H., 2000. Kinetics of the arsenite oxidation in seepage water from a tin mill tailings pond. Talanta 51, 1087–1095. Demange, M., Pascal, M.-L., Raimbault, L., Armand, J., Forette, M.C., Serment, R., Touil, A., 2006. The Salsigne Au–As–Bi–Ag–Cu deposit, France. Econ. Geol. 101, 199–234. Deng, Y., Wang, Y., Ma, T., 2009. Isotope and minor element geochemistry of high arsenic groundwater from Hangjinhouqi, the Hetao Plain, Inner Mongolia. Appl. Geochem. 24, 587–599. Dousova, B., Buzek, F., Rothwell, J., Krejcova, S., Lhotka, M., 2012. Adsorption behavior of arsenic relating to different natural solids: soils, stream sediments and peats. Sci. Total Environ. 433, 456–461. Dutré, V., Vandecasteele, C., 1995. Solidification/stabilisation of arseniccontaining waste: leach tests and behaviour of arsenic in the leachate. Waste Manag. 15, 55–62. Faure, G., 1986. Principles of Isotope Geology. Second ed. . Gao, S., Goldberg, S., Herbel, M.J., Chalmers, a.T., Fujii, R., Tanji, K.K., 2006. Sorption processes affecting arsenic solubility in oxidized surface sediments from Tulare Lake Bed, California. Chem. Geol. 228, 33–43. Girard, A., 2011. Site de Salsigne (Aude) Gestion et surveillance 2007–2010 Site de Salsigne (Aude) Gestion et surveillance 2007–2010, BRGM/RP-59620. Handley, K.M., McBeth, J.M., Charnock, J.M., Vaughan, D.J., Wincott, P.L., Polya, D.A., Lloyd, J.R., 2013. Effect of iron redox transformations on arsenic solidphase associations in an arsenic-rich, ferruginous hydrothermal sediment. Geochim. Cosmochim. Acta 102, 124–142. Khaska, M., Le Gal La Salle, C., Lancelot, J., ASTER team, Mohamad, A., Verdoux, P., Noret, A., Simler, R., 2013. Origin of groundwater salinity (current seawater vs. saline deep water) in a coastal karst aquifer based on Sr and Cl

M. Khaska et al. / Journal of Contaminant Hydrology 177–178 (2015) 122–135 isotopes. Case study of the La Clape massif (southern France). Appl. Geochem. 37, 212–227. Lottermoser, B.G., 2003. Mine Wastes: Characterization, Treatment and Environmental Impacts (277 pp.). Marchal, J., 2007. Suivi de la sortie d'eau de l'ancienne mine de La Caunette, commune de Lastours (Aude). BRGM-55926-FR. BRGM Ed., Orléans. Marcoux, E., Lescuyer, J.L., 1994. Les minerais sulfo-arséniés aurifères de Salsigne, Aude, France. Can. Mineral. 159–177. Martínez-Villegas, N., Briones-Gallardo, R., Ramos-Leal, J.A., Avalos-Borja, M., Castañón-Sandoval, A.D., Razo-Flores, E., Villalobos, M., 2013. Arsenic mobility controlled by solid calcium arsenates: a case study in Mexico showcasing a potentially widespread environmental problem. Environ. Pollut. 176, 114–122. Masscheleyn, P.H., Delaune, R.D., Patrick, W.H., 1991. Effect of redox potential and pH on arsenic speciation and solubility in a contaminated soil. Environ. Sci. Technol. 25, 1414–1419. Millot, R., Petelet-giraud, E., Guerrot, C., Négrel, P., 2010. Multi-isotopic composition (d 7 Li–d 11 B–d D–d 18 O) of rainwaters in France: origin and spatio-temporal characterization. Appl. Geochem. 25, 1510–1524. Mohan, D., Pittman, C.U., 2007. Arsenic removal from water/wastewater using adsorbents — a critical review. 142, 1–53. Morin, G., Calas, G., 2006. Arsenic in soils, mine tailings, and former industrial sites. Elements 2, 97–101. Mukherjee, A., Bhattacharya, P., Savage, K., Foster, A., Bundschuh, J., 2008a. Distribution of geogenic arsenic in hydrologic systems: controls and challenges. J. Contam. Hydrol. 99, 1–7. Mukherjee, A., von Brömssen, M., Scanlon, B.R., Bhattacharya, P., Fryar, A.E., Hasan, M.A., Ahmed, K.M., Chatterjee, D., Jacks, G., Sracek, O., 2008b. Hydrogeochemical comparison and effects of overlapping redox zones on groundwater arsenic near the Western (Bhagirathi sub-basin, India) and Eastern (Meghna sub-basin, Bangladesh) margins of the Bengal Basin. J. Contam. Hydrol. 99, 31–48. Négrel, P., Petelet-Giraud, E., 2005. Strontium isotopes as tracers of groundwater-induced floods: the Somme case study (France). J. Hydrol. 305, 99–119. Négrel, P., Roy, S., 1998. Chemistry of rainwater in the Massif Central (France): a strontium isotope and major element study. Appl. Geochem. 13, 941–952. Nguyen, D., Blaise, M., Marchal, J.-P., 2007. Suivi de la qualité des eaux souterraines en aval des installations classées 100 sites en languedocroussillon. Rapport final. Nickson, R., McArthur, J., Burgess, W., Ahmed, K.M., Ravenscroft, P., Rahman, M., 1998. Arsenic poisoning of Bangladesh groundwater. Nature 395, 338. Payne, B., 1991. On the statistical treatment of environmental isotope data in hydrology. Isotope Technique in Water Resources Development. IAEA-SM 319/36, pp. 273–290. Pedersen, H.D., Postma, D., Jakobsen, R., 2006. Release of arsenic associated with the reduction and transformation of iron oxides. Geochim. Cosmochim. Acta 70, 4116–4129. Petelet-Giraud, E., Negrel, P., Casanova, J., Négrel, P., 2003. Variability of Sr-87/ Sr-86 in water draining granite revealed after a double correction for atmospheric and anthropogenic inputs. Hydrol. Sci. J. 48, 729–742. Peters, S.C., 2008. Arsenic in groundwaters in the Northern Appalachian Mountain belt: a review of patterns and processes. J. Contam. Hydrol. 99, 8–21.

135

Pierson-Wickmann, A., Aquilina, L., Weyer, C., Molénat, J., Lischeid, G., 2009. Acidification processes and soil leaching influenced by agricultural practices revealed by strontium isotopic ratios. Geochim. Cosmochim. Acta 73, 4688–4704. Pin, C., Joannon, S., Bosq, C., Le Fèvre, B., Gauthier, P.-J., 2003. Precise determination of Rb, Sr, Ba, and Pb in geological materials by isotope dilution and ICP-quadrupole mass spectrometry following selective separation of the analytes. J. Anal. At. Spectrom. 18, 135–141. Polizzotto, M.L., Kocar, B.D., Benner, S.G., Sampson, M., Fendorf, S., 2008. Nearsurface wetland sediments as a source of arsenic release to ground water in Asia. Nature 454, 505–508. Román-Ross, G., Cuello, G.J., Turrillas, X., Fernández-Martínez, A., Charlet, L., 2006. Arsenite sorption and co-precipitation with calcite. Chem. Geol. 233, 328–336. Sánchez-Rodas, D., Luis Gómez-Ariza, J., Giráldez, I., Velasco, A., Morales, E., 2005. Arsenic speciation in river and estuarine waters from southwest Spain. Sci. Total Environ. 345, 207–217. Sharif, M.U., Davis, R.K., Steele, K.F., Kim, B., Hays, P.D., Kresse, T.M., Fazio, J.A., 2008. Distribution and variability of redox zones controlling spatial variability of arsenic in the Mississippi River Valley alluvial aquifer, southeastern Arkansas. J. Contam. Hydrol. 99, 49–67. Smedley, P., Kinniburgh, D., 2002. A review of the source, behaviour and distribution of arsenic in natural waters. Appl. Geochem. 17, 517–568. Smedley, P.L., Edmunds, W.M., Pelig-Ba, K.B., 1996. Mobility of arsenic in groundwater in the Obuasi gold-mining area of Ghana: some implications for human health. Geol. Soc. Lond. Spec. Publ. 113, 163–181. Smith, A.H., Lopipero, P.A., Bates, M.N., Steinmaus, C.M., 2002. Public health. Arsenic epidemiology and drinking water standards. Science 296, 2145–2146. Vandecasteele, C., Dutré, V., Geysen, D., Wauters, G., 2002. Solidification/ stabilisation of arsenic bearing fly ash from the metallurgical industry. Immobilisation mechanism of arsenic. Waste Manag. 22, 143–146. Vengosh, A., TY Lindberg, T., Merola, B.R., Ruhl, L., Warner, N.R., White, A., Dwyer, G.S., Di Giulio, R.T., 2013. Isotopic imprints of mountaintop mining contaminants. Environ. Sci. Technol. 47, 10041–10048. Vicente-Martorell, J.J., Galindo-Riaño, M.D., García-Vargas, M., Granado-Castro, M.D., 2009. Bioavailability of heavy metals monitoring water, sediments and fish species from a polluted estuary. J. Hazard. Mater. 162, 823–836. Wang, S., Mulligan, C.N., 2006. Occurrence of arsenic contamination in Canada: sources, behavior and distribution. Sci. Total Environ. 366, 701–721. Welch, A.H., Lico, M.S., 1998. Factors controlling As and U in shallow ground water, southern Carson Desert, Nevada. Appl. Geochem. 13, 521–539. WHO, 1993. 2nd edition. Guidelines for Drinking-Water Quality 1. World Heal. Organ, p. 94. WHO, 2011. Arsenic in Drinking-Water 24. World Heal. Organ. Williams, M., 2001. Arsenic in mine waters: an international study. Environ. Geol. 40, 267–278. Wong, H.K., Gauthier, A., Nriagu, J., 1999. Dispersion and toxicity of metals from abandoned gold mine tailings at Goldenville, Nova Scotia, Canada. Sci. Total Environ. 228, 35–47. Zheng, Y., Stute, M., van Geen, A., Gavrieli, I., Dhar, R., Simpson, H.J., Schlosser, P., Ahmed, K.M., 2004. Redox control of arsenic mobilization in Bangladesh groundwater. Appl. Geochem. 19, 201–214.