This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright Author's personal copy

Chemical Geology 276 (2010) 104–118

Contents lists available at ScienceDirect

Chemical Geology

journal homepage: www.elsevier.com/locate/chemgeo

Sulfate and strontium water source identification by O, S and Sr isotopes and their temporal changes (1997–2008) in the region of , central-eastern

M. Tichomirowa a,⁎, C. Heidel a, M. Junghans b, F. Haubrich c, J. Matschullat a a TU Bergakademie Freiberg, Institute of Mineralogy, Brennhausgasse 14, D-09599 Freiberg, Germany b TU Bergakademie Freiberg, International Centre, Lessingstr. 45, D-09599 Freiberg, Germany c TU Dresden, Institute of Soil Science and Site Ecology, Pienner Str. 7, D-01737 Tharandt, Germany article info abstract

Article history: Various waters (bulk atmospheric precipitation, groundwater, river water, flowing mine water, mine drainage Received 16 February 2010 galleries and flooding mine water) were studied in the region of Freiberg to achieve an improved understanding Received in revised form 7 June 2010 of sources and mixing processes in the hydrological cycle and to specify the main sources that determine the Accepted 10 June 2010 δ18 δ34 87 86 isotope composition of dissolved sulfate and strontium ( OSO4, SSO4 and Sr/ Sr) in river water. The Editor: J.D. Blum combined use of isotope and chemical data led to identification of three major sulfate sources: i) sulfate from groundwater, ii) sulfate from the oxidation of local ore sulfides that remained in unflooded und flooded parts after Keywords: mine decommissioning, and iii) “industrial” sulfate released from abundant waste, slag and tailing deposits left Sulfate sources behind after a long mining and industrial history of the region. The same three components (groundwater, pore Precipitation water from oxidation of residual ores and industry source) were identified by analyzing Sr isotope ratios. River However, the Sr isotopes pointed out a fourth, yet unrecognized source that contributed to water of the flooded Stable isotope tracers mine section. The data show that the isotope composition of water of the river Freiberger was mainly Acid mine drainage determined by uncontaminated groundwater diluted with atmospheric precipitation, whereas the other two Strontium end-members played only a minor role for this river water. However, for the river Triebisch the isotope compositions of dissolved sulfate and strontium record a strong impact of end-members II and III after recharging the flooding water into the river. – δ18 A twelve year monitoring (1997 2008) revealed a strong decrease of OSO4 values of water from bulk – ‰ ‰ δ34 atmospheric precipitation as of 2006 2008 (7 )comparedto1997(14 ), while SSO4 values remained constant (5‰). This isotope shift was probably caused by a lower contribution of “primary” sulfates to atmospheric precipitation in recent years. Groundwater and river water (Freiberger Mulde) also showed δ18 – ‰ δ18 decreasing OSO4 values albeit to a lower degree (by 4 5 ). A decrease of OSO4 values was again found in flowing mine water (by 5–6‰) mainly affected by end-member I (uncontaminated groundwater). The flood event in August 2002 likely washed out large parts of “older” groundwater and thus enabled the relatively large δ18 oxygen isotope decrease ( OSO4) of groundwater and river water within a few years. The previously high input from the “industrial” source (waste and slag deposits) decreased from 1997 to 2002 in some drainage galleries and flooding water that discharge their water into local rivers. This led to lower δ18Oand δ34S values of dissolved sulfates in these waters accompanied by a concentration decrease of several elements (Cl, Na and Mg). However, no further decrease of the impact of this source could be detected after the storm and flood event in August 2002. © 2010 Elsevier B.V. All rights reserved.

1. Introduction Until 1990, this region received the highest S-deposition in Central 3 Europe with local annual mean SO2-concentrations of about 100 μg/m Central-eastern Germany, together with south-western Poland and (e.g., Zimmermann et al., 2003). Many factories and power plants in East the northern , belongs to the former “Black Triangle”. Germany were shut down or equipped with desulfurization and dust removal systems after the German re-unification. As a result, a major

decrease of industrial SO2 emissions and a corresponding decrease of S- ⁎ Corresponding author. TU Bergakademie Freiberg, Institute of Mineralogy, Brenn- deposition were observed, especially in the vicinity of coal burning hausgasse 14, D-09599 Freiberg, Germany. Tel.: +49 3731 393528; fax: +49 3731 394060. power plants (Matschullat et al., 2000; Zimmermann et al., 2003, 2006). E-mail addresses: [email protected] (M. Tichomirowa), A similar tendency of decreasing SO emissions and concentrations [email protected] (C. Heidel), 2 [email protected] (M. Junghans), [email protected] occurred in Northern Bohemia, Czech Republic, after modernization (F. Haubrich), [email protected] (J. Matschullat). of coal burning power plants in 1996–1997 (Novák et al., 2001;

0009-2541/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.chemgeo.2010.06.004 Author's personal copy

M. Tichomirowa et al. / Chemical Geology 276 (2010) 104–118 105

Bridges et al., 2002). In addition to pollution from coal burning power region the methods have been applied to identify sulfate sources in plants, the Freiberg region has an 800-year of mining and related mine waters (Haubrich and Tichomirowa, 2002; Tichomirowa et al., industries, leaving behind various legacies of these former activities. 2003; Junghans and Tichomirowa, 2009). These studies identified two Related acid mine drainage has an impact not only on mine water but main sulfate sources in mine water: i) sulfate from groundwater and also on river water, even on a larger scale. The Freiberg ore deposit ii) sulfate from sulfide oxidation. contributes 37% of the Zn and 5% of the Cd contamination of the Sr isotopes likewise characterize solute Sr sources and may River (Martin et al., 1994). The region thus deserves special attention contribute to better understand water mixing processes (e.g., Negrel to trace changes in the environment which corresponds to changing et al., 1997, 2001; Siegel et al., 2000; Petelet-Giraud et al., 2003; Singh et industrial activities. al., 2006). Heidel et al. (2007) indicated three different Sr sources in Stable isotope analyses of sulfur and oxygen have been used for mine water of the Freiberg region: i) Sr from anthropogenic sources via decades to identify and quantify sources of sulfates and related mixing groundwater transport into the mine, ii) Sr released from gangue processes (e.g., Holt and Kumar, 1991; Krouse and Grinenko, 1991; carbonates, and iii) Sr released from weathering of host rocks of gangue Krouse and Mayer, 2000; Nordstrom et al., 2007). In the Freiberg ore.

Fig. 1. A. Location of and the Freiberg region in Germany. B. The Freiberg area with the location of the historical city, the river Freiberger Mulde and the tributaries Münzbach (Mb) and Halsbach (Hb), the main drainage system RSS, selected ore veins and mine adits northwest of the city of Freiberg, and the location of important waste and slag deposits. The location of atmospheric precipitation collectors (prIM and prRZ), and sample locations of groundwater (gw 1, gw 68, GW 3004, 3001, 2338, 4133 and 4080), river water (FM 1–10, Mb, Hb, T1 and T2), flowing mine water (SH 1–2), drainage galleries (GS and HU), flooding water (RZUL) and drainage water at the RSS portal are shown. The location of the RSS portal is not to scale, it is about 10 km north from sample location FM 3. Author's personal copy

106 M. Tichomirowa et al. / Chemical Geology 276 (2010) 104–118

This study provides new evidence for source identification by several smaller adits (e.g., Hauptstolln Umbruch — HU, Glücksilberstern– combining isotope studies of water (O-isotopes), dissolved sulfate (S- Stolln — GS) into the Mulde River without contact with the flooded and O-isotopes), and dissolved strontium (87Sr/86Sr) with chemical workings (Figs. 1 and 2). data. Compared with our previous investigations, a larger variety of The long-lasting mining history together with other industrial water types (atmospheric precipitation, groundwater, river water and activities left behind several industrial waste deposits, mainly located mine water) was investigated to achieve an improved understanding along the river Freiberger Mulde (Fig. 1). While local sulfide ore was of the role of different sources in the hydrological cycle in the Freiberg dominantly mined and processed, ore from other regions of the world region. The new data should help identifying the major sulfate and Sr was also used in the 20th century. Particularly a location in the southeast sources that determine the isotope composition of dissolved sulfate of the old historical city centre of Freiberg (Fig. 1,nearsamplepointsFM and strontium in river water. Finally, the new data are compared with 6 and 7) has a long history of changing industrial activities (e.g., silver, our previous results to reveal temporal changes of sulfate sources and lead and arsenic smelting, coining and lead recycling). Various mixing processes over a 12 years period (1997–2008). chemicals (acids, bases and metals) were used and later discarded in tailings over the centuries. We refer to these contaminated locations, 2. Study area together with former slag and waste deposits, as “industrial waste deposits”. Today, the chemical composition of the groundwater of the Freiberg is located in Saxony, central-eastern Germany (Fig. 1A). The contaminated locations is continuously monitored by government Freiberg polymetallic sulfide ore deposit was nearly continuously mined authorities. over 800 years to a maximum depth of 740 m. The ore contains galena, The annual mean precipitation is about 600 mm for the Freiberg pyrite, sphalerite, arsenopyrite, and chalcopyrite as well as quartz, region. In August 2002 there was an important storm and flood event minor carbonate gangue with some barite and fluorite (Baumann et al., in the region, which led to the flooding of local rivers and finally the 2000). From 1168 to 1915 the deposit was mined for Ag (about 30 tons Elbe river (the Elbe water level rose 9 m). It is important to note that per year were obtained in the 16–18th centuries) but also for Cu, Pb, and the Freiberg region received almost 200 mm precipitation within four

As. Later, mainly Zn and FeS2 were extracted until 1968, when the mine days (10–13 August 2002; Haubrich and Tichomirowa, 2004) during was decommissioned due to uneconomic conditions (e.g., Baumann et this flood event. al., 2000). Between 1968 and 1971, 2.6 million m3 of mine workings were 3. Water sampling and preservation flooded to the level of the deepest adit, the Rothschönberger Stolln (RSS), at about 225 m below surface (Figs. 1 and 2). Mine water of the flooded Bulk atmospheric deposition was sampled at two locations (prIM, part (RZUL) emerges in the Reiche Zeche shaft at the level of RSS. prRZ) over a period of two to three months each from November 2006 Afterwards, it mixes with other drainage water and discharges via the to December 2008 in funnel shape collectors in the open field. Triebisch River into the Elbe River (Figs. 1 and 2). In addition, 3.6 m3/min Groundwater samples were pumped from monitoring wells. Uncon- of mine water drains from the upper unflooded part of the mine via taminated groundwater (gw1 — one sample in 2002, gw 68 — one

Fig. 2. Schematic profile through the Freiberg mining district with accessible mine levels of the ore vein “Schwarzer Hirsch Stehender”, sample locations of flowing mine water (SH 1 and SH 2), and upper drainage galleries HU in the unflooded part. Details are given for the sampling point of flooding water at the outfall (RZUL), three sampling points for the main adit RSS, and two sampling points of the river Triebisch — a tributary of the Elbe river. Author's personal copy

M. Tichomirowa et al. / Chemical Geology 276 (2010) 104–118 107 sample in 2007) and contaminated (industrial waste deposits) measurement on a Finnigan MAT Delta plus mass spectrometer with groundwater were collected at different locations (GW 2338, GW dual inlet system. Reproducibility for this method was ≤0.1‰. The 3001, GW 3004, GW 4080 and GW 4133 — sampled once in December δ18O values of total dissolved sulfates were analyzed, after precipi-

2006; Fig. 1). River water was sampled in 1998 and 2006–2007 along tation as BaSO4 salts with a BaCl2 solution, using pyrolysis (e.g., the Freiberger Mulde river at different locations: FM 1, 2 and 5 reflect Kornexl et al., 1999)bycontinuousflow isotope ratio mass river water which is uncontaminated by industrial activities, and FM spectrometry (CF-IRMS). The sulfur isotope composition of precipi-

3, 4, 6–10 represent river water before and after main adit discharges tated BaSO4 was analyzed by CF-IRMS using an elemental analyzer (Fig. 1). The small tributaries Münzbach (Mb) and Halsbach (Hb) coupled to the mass spectrometer (e.g., Giesemann et al., 1994). The were sampled once in 1998 while the Triebisch river (T1 and T2) was values were calibrated using one internal BaSO4 standard for drift sampled three times in 2006–2007 (Fig. 1). correction and two international reference standards. We used the Flowing mine water, which drains along the mined ore vein standards IAEA-SO-5 (12.1‰) and IAEA-SO-6 (−11.3‰) for the “Schwarzer Hirsch Stehender” in the unflooded parts (SH 1 and SH 2), calibration of oxygen isotopes resulting in a value of 8.7‰ for NBS 127 was sampled at two different mine levels in 1997 (Haubrich and (Brand et al., 2009). The standards IAEA-SO-2 (22.7‰) and IAEA-SO-3 Tichomirowa, 2002) and again in 2000–2002 (Junghans and Tichomir- (−32.3‰) were used for the calibration of sulfur isotopes resulting in owa, 2009; Fig. 2). We provide new data for these two locations for the a value of 21.6‰ for NBS 127 (Ding et al., 2001). All samples were time period 2005–2007 (8 sampling campaigns). New samples from analyzed at least in triplicate and mean values are given in the Tables drainage galleries (GS — sampled five times from 2004 to 2007, HU — and Figures. The long-term reproducibility of oxygen isotopes of sampled seven times from 2004 to 2007) in the upper unflooded part sulfates is ≤0.5‰, that of sulfur isotopes ≤0.3‰, although the internal were taken at their portals before discharging into the Freiberger Mulde error of three successive measurements often was smaller. river (Figs. 1 and 2). To characterize mine water from the flooded part in Water samples were evaporated at 85 °C for Sr isotope analysis. the Reiche Zeche Shaft (about 4 m3/min), we sampled directly at the The evaporated volume varied between 2 and 40 mL, depending on Sr outfall (Fig. 2: RZUL — 60 samples from 1997 to 2007). The deepest mine concentrations. The residues were dissolved in 2.5 N HCl (1 mL). adit, the Rothschönberger Stolln (RSS) with a length of about 24 km, Strontium was extracted by cation-exchange chromatography with was sampled at different locations (Fig. 2): (i) before mixing with mine about 4 mL Dowex 50WX8 (200–400 mesh) and 2.5 N HCl as eluent. water from the Reiche Zeche shaft (RSS 1 — sampled three times from After evaporation and dissolution of the residue in distilled water 2006 to 2007), (ii) after this mixing (RSS 2 — sampled three times from (10 μL), the sample was loaded onto a single tantalum filament with

2006 to 2007) and (iii) before its discharge into the river Triebisch (RSS 1N H3PO4 and measured with the thermal ionization mass spec- 3–11 samples from 2003 to 2007). trometer Finnigan MAT 262. Relative errors were less than 0.005% for All samples from groundwater, river water, and mine water were 87Sr/86Sr. Replicate analyses of the NBS 987 standard (n=26) gave an immediately analyzed for pH, temperature, electrical conductivity, average 87Sr/86Sr ratio of 0.71029±0.00005 (2σ), slightly higher than redox potential, and dissolved oxygen. Samples were filtered through a the certified value of 0.71025 (Faure, 2001). The total procedure blank 0.45 μmporesizefilter and split into subsamples (e.g., for isotope was negligible (b0.5 ng). analyses of dissolved sulfate or strontium and for element analyses), which were stored in polyethylene or polypropylene bottles. Samples 5. Results for dissolved metal analysis were acidified with HNO3. Water samples 18 for δ OH O analysis were not filtered and stored in brown glass bottles. 2 5.1. Bulk precipitation All samples were stored at a temperature of 8 °C prior to analysis. The flow rate of flowing mine water (SH 1 and SH 2) was measured Mean sulfate concentrations in atmospheric deposition from 2006 with a beaker or bucket by determining the volume five times and to 2008 (Table 1: 2.48–2.94 mg/L) decreased only slightly compared calculating the mean. The flow rate of the Freiberger Mulde river was to 1997 (Haubrich and Tichomirowa, 2004: 3.2 mg/L). Sulfur isotope taken from the database of the Saxonian State Agency for Environ- ratios of sulfate did not show large variations and the mean sulfur ment and Geology (http://www.umwelt.sachsen.de/lfug/hwz/MP/ isotope composition (Table 1: 4.8–5.0‰) was very similar to values 566010/index.html). from previous years for the Freiberg region (Tichomirowa et al., 2004: 4.6–4.8‰) and for the German state of Saxony in general (Tichomir- 4. Methods – ‰ δ18 owa et al., 2007:45 ). The mean OSO4 value for atmospheric sulfate collected at both Freiberg locations (Table 1: 6.1 and 7.6‰) Alkalinity was determined by acidic titration. Cl, F, NO , and SO 3 4 was about 7‰ lower than that determined in 1997 (Haubrich and concentrations were measured by ion chromatography; Na, K, Ca, and Tichomirowa, 2004: 14.2‰). Mg by flame atomic absorption spectrometry (F-AAS), and Li by flame atomic emission spectrometry (F-AES). The water samples were analysed for Al, As, Cd, Cu, Fe, Mn, Pb, and Zn by ICP-AES, ICP-MS or 5.2. Groundwater

ETA-AAS, depending on their concentration. Only SO4 concentrations were analyzed in atmospheric deposition. All analyses were done Physico-chemical field parameters as well as chemical data show according to quality control protocols in the geochemical analytical that several samples of groundwater were affected either by contam- laboratories of the Institute of Mineralogy, TU Bergakademie Freiberg. ination with ore bodies or by pollution from abundant industrial waste The isotope compositions were determined at the isotope deposits. The main indicators for such contamination were low pH laboratory of the same institute. Oxygen and sulfur isotope results values, increased electrical conductivities, and increased concentrations 18 34 are reported as δ O and δ S values in per mil (‰ vs. VSMOW and of ore-related elements (e.g., Zn, SO4,CdandCu,compareTable 2). VCDT, respectively), where: Accordingly, only the samples gw 68 and gw 1 indicated unaffected groundwater. In contrast, groundwater samples labelled with an hi asterisk in Table 2 was affected by contamination to various degrees. δðÞ‰ = R = R −1 × 1000 sample standard δ18 The OSO4 value of the unaffected groundwater gw 1, sampled in 2007, ‰ δ18 was about 4 lower than the OSO4 value of sample gw 68, sampled in δ34 δ18 34 32 18 16 δ34 for Sor O, R= S/ Sor O/ O, respectively. 2002, whereas their SSO4 values were very similar (Table 2). Its Sr The δ18O water values were determined using the equilibration isotope value (Table 2) was higher (0.71772) compared to samples from method with CO2 after Epstein and Mayeda (1953), and subsequent mineralized groundwater (0.71347 to 0.71649). Author's personal copy

108 M. Tichomirowa et al. / Chemical Geology 276 (2010) 104–118

Table 1 5.4. Mine water Sulfate concentrations and isotope data of atmospheric precipitation.

δ18 δ18 δ34 Different mine waters were distinguished: flowing mine water (SH Sampling date SO4 OH2O OSO4 SSO4 Comments 1 and SH 2), water from drainage galleries in the unflooded part of the VSMOW VSMOW VCDT mine (GS and HU), flooding water (RZUL), and drainage water at ‰‰‰ mg/L flooding level (RSS1, 2 and 3; see Fig. 2). prRZ Flowing mine waters. The observed decrease of pH, increase of – − Nov Dec/2006 4.38 8.9 6.4 6.2 electrical conductivity and increasing concentrations of ore-related Jan–Feb/2007 –– 7.5 5.7 Mar–May/2007 – −4.7 3.6 4.0 elements (Zn, Pb, Cd, Cu and SO4; mean values in Table 4, single values Jun–Jul/2007 – −6.1 2.3 3.6 in Appendix Table 2) along the flow path from SH 1 to SH 2 (Fig. 2) Aug–Sep/2007 2.63 −8.0 8.4 4.0 was caused by the addition of highly mineralized water from the Oct–Nov/2007 – −9.1 9.4 5.7 oxidation of remaining sulfide ores in “clayey deposits” (Haubrich and – − Dec Jan/2008 3.60 8.7 12.1 5.9 Tichomirowa, 2002; Tichomirowa et al., 2003; Junghans and Ticho- Feb–Mar/2008 2.95 −8.6 10.1 4.4 Apr–May/2008 2.46 −8.5 3.9 4.2 mirowa, 2009). The low pH values within these clayey deposits Jun–Jul/2008 2.59 −4.1 0.8 4.1 released K, Al, and Mg from silicate ore-hosting wall rocks, where K – − Aug Sep/2008 2.59 6.0 3.7 4.9 was precipitated in newly formed mineral illite [KAl2(AlSi3O10)(OH)2] Oct–Dec/2008 2.34 −11.7 5.0 5.1 and partly in jarosite [(Pb,K)Fe3(SO4)2(OH)6](Tichomirowa et al., Mean 2.94 −7.7 6.1 4.8 2003). These waters had remarkably very low concentrations of Cl, F, prIM Na, K, and Ca. Nov–Dec/2006 3.03 −7.9 5.0 5.7 The addition of highly mineralized sulfate-bearing water from Jan–Feb/2007 –– 5.5 5.5 34 18 sulfide oxidation resulted in a decrease of δ SSO and δ OSO values – – − 4 4 Mar May/2007 7.6 7.6 4.9 from SH 1 to SH 2 (Table 4), as shown in previous sampling campaigns Jun–Jul/2007 – −7.5 7.0 4.8 Aug–Sep/2007 2.38 −6.8 10.9 3.4 at these two locations (e.g., Haubrich and Tichomirowa, 2002; Oct–Nov/2007 – −10.1 10.3 5.5 With snow Junghans and Tichomirowa, 2009). However, compared to earlier Dec–Jan/2008 3.56 −9.2 9.9 6.6 With snow sampling campaigns (1997–1998: Haubrich and Tichomirowa, 2002 18 (δ OSO of snow=4.6‰) 18 4 and 2000–2002: Junghans and Tichomirowa, 2009), δ OSO values Feb–Mar/2008 1.77 −10.3 8.8 5.0 With snow 4 were much lower while δ34S values remained nearly constant Apr–May/2008 2.95 −0.6 9.9 4.5 SO4 Jun–Jul/2008 2.81 −5.8 7.2 4.5 (Fig. 3). In 1997, flowing mine water (SH 1, SH 2) formed a mixing line Aug–Sep/2008 1.98 −5.5 5.5 4.4 with groundwater and pore water from sulfide oxidation as being the Oct–Dec/2008 1.36 −11.8 3.8 4.8 end-members (Fig. 3A). While samples from 2000 to 2002 plot close − Mean 2.48 7.6 7.6 5.0 to this mixing line (albeit with larger scatter, Fig. 3B), this tendency – not determined. was hardly recognizable for samples from 2005 to 2007 (Fig. 3C). Nonetheless, both isotope values of sulfates (S and O) at location SH 2 were dependent on the sulfate concentration for the time period 2005–2006 (Fig. 4). They must therefore represent the addition of 5.3. River water sulfate released from sulfide oxidation processes as shown for the sampling periods 1997–1998 and 2000–2002. However, the two River water had pH values between 7.0 and 8.1 (mean values in samples from 2007 did not follow this trend, neither in Fig. 3C nor in Table 3; single values in Appendix Table 1). Most element concentra- Fig. 4, and probably recorded a further time-related decrease of their fl δ18 tions (except for Al) depended on water ow rates and were highest OSO4 values. There was no clear difference in the Sr isotope for the sampling campaign on 04/05/2007 with the lowest flow rate composition between sample locations SH 1 and SH 2 (Table 4: about (Appendix Table 1). The locations FM 5 and T1 yielded the lowest 0.7186). element concentrations and were least affected by contamination. The drainage gallery GS is located beneath industrial waste Many elements already had distinctly higher concentrations down- deposits resulting from industrial mining and smelting as well as stream at location FM 6 (Table 3, Appendix Table 1). Zn released by from lead recycling activity (Fig. 1). Samples contained very high sphalerite oxidation (Tichomirowa et al., 2003) showed the highest concentrations of ore-related elements (e.g., Zn, Pb, Cu, and Cd). increase from FM 5 to FM 6 (about 15-fold) and further increased after Different from the flowing mine waters SH 1 and SH 2 described the HU adit discharged into the river (FM 10). At sampling point FM above, this water also yielded high concentrations of Cl, F, Na, and Ca; 10, a mean Zn concentration of about 436 μg/L was found compared to probably added by industrial activities (Table 5). Sulfate from this an initial value of about 10 μg/L at sample location FM 5 (Table 3, location showed an isotope composition (δ34S and δ18O) distinct from δ34 δ18 Appendix Table 1). Cd from sphalerite (Tichomirowa et al., 2003) other mine waters. Compared to 1997, both SSO4 and OSO4 values occurred in dissolved form and showed a similar behaviour. The sulfur (Junghans and Tichomirowa, 2009: 4.8‰ and 8.4‰, respectively) isotope composition of dissolved sulfate in FM river water from 2006 decreased with time (Table 5: 2.7‰ and 4.4‰, respectively). This to 2007 was nearly the same as in 1998, while the oxygen isotope drainage gallery had a low Sr isotope ratio (Table 5: 0.71443) similar composition of sulfate decreased by about 5‰ (Table 3). The Sr to the nearby mineralized groundwater GW 4133 (Table 2: 0.71429). isotope composition of all samples from the river Freiberger Mulde The drainage gallery HU drains mine water from the upper was very similar (Table 3: 0.71521–0.71632). unflooded part of the Freiberg mine (Fig. 2). It contained increased The RSS adit delivered high loads of various elements into the concentrations of ore-related elements released during sulfide Triebisch river as shown by the strong concentration increase from oxidation (e.g., Zn, Pb, Cu, and Cd) with values ranging between location T1 to T2 (Table 3: e.g., 186-fold increase for Zn and 20-fold those of SH 1 and SH 2 and concentrations above those at location SH δ18 increase for Cd). The ore-related elements Zn and Cd were released 2 (e.g., Cl, F, Na, Ca, and Al; Table 5). Only a slight decrease of OSO4 ‰ δ34 from the oxidation of remaining ores in the mine workings. Cu, Ni, Mg, values was observed with time (by 3 ), while SSO4 values and K showed a similar increase as Zn and Cd, albeit to a much lower remained constant (Fig. 3). The Sr isotope composition varied degree. The addition of RSS water led to a strong increase of the Sr between 0.71561 and 0.71753 (Table 5). isotope composition in the river Triebisch (Table 3: 0.71168 for T1 and Water from the flooded part (almost 500 m depth) of the Freiberg 0.71748 for T2). mine rose in the Reiche Zeche shaft (RZUL) where it emerged with a Author's personal copy

M. Tichomirowa et al. / Chemical Geology 276 (2010) 104–118 109

Table 2 Physico-chemical field parameters, chemical and isotope data of groundwater samples.

Assigned group Unit gw 68 gw 1 GW 2338a GW 3001a GW 3004a GW 4080a GW 4133a GWAa GWAa GWAa GWAa GWBa

Sample depth m b.g. 8 15 14 14 14 6 9 Sampling date dd/mm/yyyy 25/07/2002 08/06/2007 15/12/2006 13/12/2006 15/12/2006 14/12/2006 14/12/2006 T °C 11.7 14.6 14.2 11.5 11.4 10.8 12.2 EC μS/cm 50.3 458 707 4920 1959 1765 1394 pH 5.4 5.8 5.8 3.2 5.8 3.4 3.6 Eh mV – 426 519 679 538 703 677

O2 mg/L 6.2 7.3 –––––

HCO3 mg/L 8.0 23.2 –––––

NO3 mg/L 24.2 16.5 –––––

SO4 mg/L 58 111 267 4320 621 965 696 Na mg/L 7.4 17.1 ––––– K mg/L 1.4 3.98 ––––– Ca mg/L 20.1 44.9 ––––– Mg mg/L 6.2 12.3 ––––– F mg/L ––––––– Cl mg/L 9.4 55 ––––– Al μg/L 29 9 – 130 140 –– Cd μg/L 3 2 170 3300 24 780 590 Cu μg/L 4 6 8200 18000 44 70000 6000 Fe μg/L 75 159 ––––– Mn μg/L 20 9 ––––– Ni μg/L – 6 ––––– Pb μg/L 4 12 50 13 5 240 820 Zn μg/L 43 102 – 410000 16000 5800 360 Sr μg/L – 171 ––––– δ18 ‰––− − − − − OH2O (VSMOW) 9.9 7.2 8.0 9.9 9.2 10.2 δ18 ‰ − − − − OSO4 (VSMOW) 5.3 1.0 3.9 5.4 3.1 4.9 1.6 δ34 ‰ SSO4 (VCDT) 5.3 5.1 1.1 0.2 2.7 1.6 1.0 87Sr/86Sr – 0.71772 0.71521 0.71649 0.71644 0.71347 0.71429

– not determined. a Groundwater influenced by industrial tailings from former mining and other industrial activities.

flow rate of about 4 m3/min (Fig. 2; single values are given in oxygen and a higher temperature (Table 5). It contained increased Appendix Table 3, and mean values in Table 5). Water from the concentrations of ore-related elements (e.g., Zn, Pb, Cu, and Cd) as fl δ18 out ow differed from other mine water by its low content of dissolved well as elements like Cl, F, Na, Ca, and Al (Table 5). The OSO4 values

Table 3 Mean physico-chemical field parameters, chemical and isotope data of river water samples.

Unit Mean FM 1– 4 Mean Mb+Hb Mean FM 5 Mean FM 6 Mean FM 7 Mean FM 8 Mean FM 9 Mean FM 10 Mean T1 Mean T2

Sampling period 1998 1998 2006–2007 2006–2007 2006–2007 2006–2007 2006–2007 2006–2007 2006–2007 2006–2007

Q L/min –– 215 T °C –– 9.4 8.8 9.7 9.0 9.1 8.8 9.0 11.8 EC μS/cm –– 220 322 362 355 369 369 460 701 pH –– 7.7 7.5 7.7 7.6 7.5 7.4 7.7 7.5 Eh mV –– 435 385 390 435 420 367 440 486

O2 mg/L –– 10.1 12.0 11.8 11.9 10.9 11.5 10.2 10.0

HCO3 mg/L –– 33.0 33.9 41.9 40.7 40.3 39.1 92.7 106

NO3 mg/L –– 22.4 20.9 22.8 22.7 23.0 23.0 24.0 11.8

SO4 mg/L –– 42 56 54 54 58 59 90 221 Na mg/L –– 10.0 18.5 21.5 21.0 21.4 22.0 19.9 26.8 K mg/L –– 2.83 3.21 3.18 3.41 3.46 3.52 4.40 5.00 Ca mg/L –– 20.0 31.7 35.9 34.3 34.7 34.5 49.1 83.9 Mg mg/L –– 4.93 5.34 5.42 5.75 5.91 5.95 11.7 20.3 F mg/L –– 0.25 0.56 0.68 0.59 0.64 0.66 0.20 1.20 Cl mg/L –– 13.0 42.0 48.2 45.4 46.7 46.4 40.8 48.0 Al μg/L –– 136 287 274 208 282 237 144 200 Cd μg/L –– b.d.l. 2.3 2.0 2.9 4.5 5.0 0.4 12.0 Cu μg/L –– 3.7 5.6 6.5 7.8 11.5 11.0 3.0 11.7 Fe μg/L –– 158 196 270 202 238 198 161 440 Mn μg/L –– 16 55 38 41 90 76 32 569 Ni μg/L –– 2.9 4.2 3.9 4.7 4.8 5.3 3.7 22 Pb μg/L –– 6.7 13.8 14.2 9.9 12.1 13.0 5.7 12.7 Zn μg/L –– 10 147 115 198 384 436 12.7 2 357 Sr μg/L –– 93 110 102 106 110 97 178 315 δ18 ‰– – − − − − − − − − OH2O (VSMOW) 9.7 9.7 9.6 9.6 9.5 9.4 9.2 9.4 δ18 ‰ − OSO4 (VSMOW) 5.6 6.0 0.7 0.4 0.5 0.8 0.5 0.1 1.5 1.3 δ34 ‰ SSO4 (VCDT) 5.5 5.1 5.4 5.1 5.4 5.2 5.0 4.8 5.0 1.8 87Sr/86Sr –– 0.71594 0.71525 0.71527 0.71521 0.71523 0.71632 0.71168 0.71748

– not determined. b.d.l. — below detection limit. Author's personal copy

110 M. Tichomirowa et al. / Chemical Geology 276 (2010) 104–118

Table 4 2002). The second component usually led to higher concentrations of Mean physico-chemical field parameters, chemical and isotope data of flowing mine ore-related elements (Zn, Pb, Cu, and Cd). water SH 1 and SH 2. Plotting isotope data of sulfate versus the reciprocal sulfate Unit SH 1 SH 2 concentration is another way to identify sulfate sources (Fig. 6). The fi na 10 16 mixing between end-members I and II should de ne a line in this diagram type similar to Fig. 3. Mean values of the three time periods Sampling period 2005–2006 2005–2007 were used for most water types. However, only a single analysis was Q L/min 61.3 23.2 available for groundwater uncontaminated by pollution from waste T °C 8.8 9.9 fi EC μS/cm 312 425 deposits and/or sul de oxidation processes for each time period. It δ34 pH 5.7 3.7 remains unclear whether these values are representative. The SSO4 Eh mV 570 691 value of uncontaminated groundwater is similar to those of O2 mg/L 11.6 11.1 precipitation and river water (e.g., Tables 1 and 3; n=44). Therefore, HCO3 mg/L 3.40 b.d.l. 34 we consider the δ SSO value of end-member I as representative, NO mg/L 53.7 42.9 4 3 because δ34S values of atmospheric precipitation remain constant SO4 mg/L 102 137 SO4 Na mg/L 8.20 8.72 during the transit through soil horizons (Mayer et al., 1995). Neither K mg/L 3.00 2.92 the sulfur isotope composition of sulfates from groundwater nor that Ca mg/L 32.7 35.3 of pore water has changed with time (Fig. 3). Therefore, the sulfur Mg mg/L 9.10 10.3 isotope composition is regarded as a suitable basis to define the F mg/L 0.10 0.08 Cl mg/L 6.80 6.98 mixing line. Accordingly, a line was drawn in Fig. 6A, B, C starting from Al μg/L 140 796 pore water (end-member II) through sampling points SH 2 and SH 1 Cd μg/L 14.7 67.2 (since it was shown that their isotope composition was controlled by μ Cu g/L 17.6 171 mixing processes from end-members I and II — see Section 5.4) Fe μg/L 51.9 882 Mn μg/L 149 383 towards groundwater. The sulfate concentration of groundwater was Ni μg/L b.d.l. 18.3 adjusted to this line because of its large uncertainty caused by the Pb μg/L 56.6 959 limited data set (adjusted groundwater is shown as a cross). Then, this Zn μg/L 1754 6551 sulfate concentration was used to define end-member I in diagrams Sr μg/L 124 122 δ18 ‰ − − for oxygen isotope composition too. O H2O (SMOW) 9.9 9.9 δ18 ‰ − − In addition to flowing mine waters SH 1 and SH 2 water from the O SO4 (SMOW) 1.5 1.9 δ34 ‰ S SO4 (CDT) 4.3 3.0 drainage gallery HU also plots along this mixing line, indicating sulfate 87Sr/86Sr 0.71866 0.71861 from both above mentioned sources. However, water from the – not determined. drainage gallery GS as well as the flooding water RZUL did not plot b.d.l. — below detection limit. along this line, thus indicating a third sulfate source. This source (end- a — n number of samples. member III) was chosen for Fig. 6A and D to span a triangle that includes all sampling points. Its isotope composition and sulfate decreased only slightly with time (by 3–4‰, Fig. 3). The Sr isotope concentration in 1997–1998 was similar to the drainage gallery GS, composition showed a narrow range (Table 5: about 0.7170). with high concentrations of ore-related elements but also of other The deepest and largest mine adit RSS represents drainage water at elements, attributed to industrial activities (e.g., Cl, F, Na and Ca). the flooding level and contributed about 36 m3/min into the river Accordingly, the sulfate of end-member III was attributed to industrial Triebisch, a tributary of the river Elbe. After mixing with water from waste deposits (i.e., sulfate produced in industrial processes and later the flooded part (RSS 2), the water in the RSS adit travelled about deposited in tailings). Pyrite oxidation experiments reported only a δ18 δ18 10 km, where both precipitation of minerals and further mixing small difference between OSO4 and OH2O values between 0 and occurred, before it discharged into the river Triebisch. It carried high 4.1‰ (e.g., Balci et al., 2007; Tichomirowa and Junghans, 2009; Heidel loads of Zn and other elements at its portal (RSS 3), despite an almost et al., 2009) in agreement with values for end-member II (Junghans δ18 δ18 neutral pH value (Table 5). Similar to RZUL, the OSO4 values and Tichomirowa, 2009). Hence, the large difference between OSO4 – ‰ δ18 ‰ decreased only slightly with time (by 2 3 , Table 5). The Sr isotope and OH2O values for end-member III (e.g., about 15 for GS) ratio of this adit was relatively high (Table 5: 0.71883 for RSS 3). excludes the possibility that this sulfate was formed by sulfide New data were plotted with those from our previous investiga- oxidation as end-member II. tions by calculating mean annual values where possible (Fig. 5). Some of the contaminated groundwater samples showed an Various water types in the region of Freiberg recorded a decrease in obvious contribution of this industrial sulfate source. Two different δ18 their OSO4 values with time. However, the degree of this decrease types of contaminated groundwater were distinguished based on δ34 δ18 clearly differed between water types. In contrast, the SSO4 values their OSO4 values (GWA*: samples GW 2338*, 3001*, 3004*, 4080*; remained nearly constant for most water types but decreased for GWB*: GW 4133*), with the second being much more affected by the some mine waters with time (Fig. 5). sulfate of end-member III (Fig. 6). The sample of GWB* was taken directly beneath an industrial waste deposit (Fig. 1). It had a much 6. Discussion higher Pb but a lower Zn concentration compared to other contaminated groundwater samples (Table 2 — no measurements of 6.1. Sulfate sources Cl, Na and Ca were conducted). This sample was more affected by the industrial source (end-member III). In conclusion, the identification of δ18 –δ34 fi The OSO4 SSO4 diagram (Fig. 3) can be used to identify different this third sulfate end-member, which was not de ned in previous δ18 sulfate sources. Mixing of water with two sulfate sources would result in investigations, results from a combination of isotope data ( OSO4, fl δ18 δ34 a straight line in this diagram. This was shown for owing mine water OH2O, SSO4) and chemical data (e.g., high Cl, Ca, and Pb). SH 1 and SH 2 for the time period 1997–1998 (Fig. 3A) resulting in the The sulfur and oxygen isotope composition of dissolved sulfate in identification of two sulfate sources: (i) groundwater uncontaminated river water from the Freiberger Mulde is quite similar to that of by pollution from waste deposits and/or sulfide oxidation processes groundwater (Fig. 3), whereas its sulfate concentration is much lower fi δ18 δ34 (end-member I), and (ii) pore water formed from remaining sul des (Fig. 6). The impact of end-members II and III on OSO4 and SSO4 due to their oxidation (end-member II; Haubrich and Tichomirowa, values of Freiberger Mulde river water is therefore negligible (Fig. 6). Author's personal copy

M. Tichomirowa et al. / Chemical Geology 276 (2010) 104–118 111

Fig. 3. Sulfur and oxygen isotope composition of sulfate of various water types for three sampling campaigns: A) 1997–1998: data for mine water from Haubrich and Tichomirowa (2002), data for river water and RZUL (this study), B) 2000–2002, data for mine water from Junghans and Tichomirowa (2009), data for atmospheric precipitation from Tichomirowa et al. (2004 — only S isotope values), data for RZUL (this study), and C) 2005–2007, all data from this study. Mixing lines for SH 1 and SH 2 were defined by Haubrich and Tichomirowa (2002) in panel A and are shown for comparison as dotted lines in panels B and C.

However, the isotope composition of dissolved sulfate of the Triebisch due to the addition of flooding water from RZUL. The two end- river after mixing with flooding water (T2) shows clear contributions members for this mixing process were RSS 1 and RZUL. Similarly, line from end-members II and III. “b” in Fig. 7A represents the addition of water from RSS 3 into the river Triebisch. The arrow shows the resulting shift in Sr concentration and 6.2. Strontium sources Sr isotope composition from T1 to T2. Fig. 7B shows mean values for Sr isotope composition and Sr Sr isotope compositions can be used to identify Sr sources. Fig. 7A concentrations given in Tables 3–5. Only a single value was available indicates two mixing processes. Line “a” shows that Sr concentrations for groundwater gw 1. The end-member II represented the same increased and Sr isotope ratios slightly decreased from RSS 1 to RSS 2, source as determined by using S and O isotopes of sulfates, namely Author's personal copy

112 M. Tichomirowa et al. / Chemical Geology 276 (2010) 104–118

Fig. 4. Sulfur and oxygen isotope composition of dissolved sulfate in flowing mine water samples SH 1 and SH 2 versus their reciprocal sulfate concentration for the sampling period 2005–2007. A good correlation emerges for sampling point SH 2, if values from 2007 are not included. pore water from sulfide oxidation. The mixing line between end- suggested. This assumption is supported by the groundwater sample members I and II was drawn in the same way as in Fig. 6: starting from GW 4133* taken directly beneath an industrial waste deposit with a very pore water and drawing through samples SH 1 and SH 2. Thus, in similar 87Sr/86Sr isotope ratio (this sample is not shown in Fig. 7B accordance with O and S isotopes (Fig. 6), end-member I should because no Sr concentration was measured). It becomes obvious from represent uncontaminated groundwater. This was supported by Fig. 7B, that the end-members I and III had similar Sr isotope similar Sr isotope compositions of the river water from the Freiberger compositions, but the latter had much higher Sr concentrations. The Mulde (representing a mixture of groundwater and precipitation). increase of Sr concentrations along the river from sampling point FM 5 Hence, the single value of groundwater sample gw 1 is not a good towards FM 6 (grey arrow in Fig. 7B) can be explained either by addition proxy for the Sr concentration of end-member I. of uncontaminated groundwater (to rain water) or by addition from The drainage gallery GS did not plot along the mixing line between industrial waste deposits. The latter interpretation is more likely due to end-members I and II (Fig. 7B). Hence, another Sr source contributed to the parallel increase of concentrations of other elements (Table 3: Cl, Na, its high Sr concentrations. The mixing lines to end-member III were andCa).However,thisinfluence has only a minor impact on the Sr drawn towards GS to include this sampling point in the triangle. In isotope ratio because of the similarity of 87Sr/86Sr values for both end- δ18 δ18 “ ” 87 86 accordance with OSO4 and OH2O values of GS, an industrial source members (I and III). The Sr/ Sr values of the Freiberger Mulde river with high Sr concentrations and relatively low 87Sr/86Sr isotope ratios is water were probably mainly determined by the 87Sr/86Sr value of

Table 5 Mean physico-chemical field parameters, chemical and isotope data of drainage galleries GS and HU in the unflooded mine part, flooding water RZUL, and the main drainage gallery RSS in the flooded part.

Unit GS HU RZUL RZUL RZUL RZUL RZUL RSS 1 RSS 2 RSS 3 RSS 3

na 4 7 5 24 6 11 14 3 3 6 4

Sampling period 2004–2007 2004–2007 1997 2000–2002 2003 2004–2005 2006–2007 2006–2007 2006–2007 2003–2005 2006–2007

T °C 10.4 8.4 18.5 18.8 18.0 18.8 18.8 13.4 14.6 12.5 13.7 EC μS/cm 2351 802 2676 1885 1469 1597 1661 701 930 838 851 pH 4.8 7.1 6.1 6.1 6.1 6.1 6.1 7.0 6.6 7.3 7.2 Eh mV 598 435 359 467 505 464 450 343 391 461 482

O2 mg/L 8.0 11.1 1.1 1.1 2.8 0.6 1.2 9.7 8.2 7.7 7.8

HCO3 mg/L 2.3 39.8 96.0 77.0 72.0 78.3 86.1 106 97.5 92.4 105

NO3 mg/L 12.6 8.00 3.48 2.00 2.04 0.86 1.50 7.31 5.74 7.60 7.52

SO4 mg/L 919 279 1150 918 704 811 812 246 372 300 289 Na mg/L 180 27.0 245 89.1 55.0 59.0 54.8 21.7 30.1 33.5 30.9 K mg/L 17.8 4.94 22.3 11.7 9.40 9.62 9.49 4.40 5.60 4.87 5.41 Ca mg/L 238 81.0 299 289 208 236 240 83.0 164 104 106 Mg mg/L 41.5 28.1 55.8 53.4 40.7 45.3 47.7 24.4 37.6 24.0 25.6 F mg/L 2.17 0.70 1.78 1.60 1.87 2.25 2.47 0.41 1.15 1.71 1.75 Cl mg/L 353 73.5 212 128 90.0 103 104 43.6 58.2 54.8 54.2 Al μg/L 6671 559 686 435 480 409 633 66.8 257 280 284 Cd μg/L 1138 48.8 72.0 82.0 111 68.6 74.9 17.9 28.3 22.0 35.0 Cu μg/L 2188 54.6 38.6 27.3 27.6 17.5 30.4 8.73 16.0 6.50 33.3 Fe μg/L 251 1847 2612 905 867 350 761 852 1086 165 658 Mn μg/L 16025 4450 11120 11291 4380 5853 6204 483 1744 1048 1106 Ni μg/L 643 153 –– –– 80.8 13.6 23.1 39.5 224 Pb μg/L 2486 117 42.2 34.7 41.0 51.0 60.4 15.9 18.2 – 145 Zn μg/L 47238 4801 14660 15605 11660 12318 12391 3452 5445 3512 3922 Sr μg/L 452 211 –– –– 682 235 336 425 406 δ18 ‰ − − − − − − − − − − − O H2O 9.0 9.5 9.6 9.4 9.5 9.4 9.5 9.7 9.7 9.7 9.7 (VSMOW) δ18 ‰ − − − − O SO4 4.4 1.4 3.9 1.0 1.0 0.4 0.2 1.9 1.6 0.8 0.9 (VSMOW) δ34 ‰ S SO4 (VCDT) 2.7 1.1 2.1 1.2 1.0 1.0 1.1 1.1 1.1 0.9 1.0 87Sr/86Sr 0.71443 0.71607 –– – 0.71693 0.71716 0.71861 0.7179 – 0.71883

– not determined. a n — number of samples. Author's personal copy

M. Tichomirowa et al. / Chemical Geology 276 (2010) 104–118 113

Fig. 5. Sulfur (left column) and oxygen (right column) isotope composition of dissolved sulfate in various water types. If possible, annual mean values are given, shown by thick lines. Single values are shown by different signs (■●○△). pr — atmospheric precipitation, gw — groundwater, rw — river water Freiberger Mulde. The annual mean value of river water was calculated from FM 1 to 4 for 1998 and from FM 5 to 10 for 2006 and 2007 including several sampling campaigns throughout the year. Dotted lines connect annual mean and single values if no data were available. Author's personal copy

114 M. Tichomirowa et al. / Chemical Geology 276 (2010) 104–118

Fig. 6. Sulfur (left column) and oxygen (right column) isotope composition of dissolved sulfate in various water types versus their reciprocal sulfate concentration for the three sampling campaigns 1997–1998, 2000–2002, and 2005–2007. If possible, mean values were used for these time intervals. Roman numbers show the sulfate end-members I, II, and III. The grey triangle includes all mixing compositions. gw — groundwater, rw — river water, pw — pore water. See text for discussion. groundwater (Fig. 7B). Contrary, the Triebisch river water (T2) shows a 6.3. Temporal changes in the isotope composition (S and O) of sulfates clear impact of end-member III caused by the addition of flooding water (Fig. 7B). The mixing line between end-members I and II obviously changed According to the triangle in Fig. 7B, the flooding water (RZUL) its position with time (Fig. 5). It rotated clockwise in Fig. 6A, B, C from should be influenced by a fourth Sr source. The higher Sr concentra- 1997–1998 to 2005–2007 due to decreasing sulfate concentrations in tion in the flooding water (RZUL) was interpreted as dissolution of groundwater, whereas the sulfur isotope compositions of groundwa- carbonates from gangue material (Heidel et al., 2007) since the ter and flowing mine water (SH 1 and SH 2) remained constant. The flooding water had a higher pH value (Table 5: 6.1) compared to influence of the industrial sulfate source (end-member III) likely flowing mine waters SH 1 and SH 2 (Table 4: 5.7 and 3.7, respectively). decreased from 1997 to 2007 for GS, RZUL and RSS 1–3asreflected by However, the measured 87Sr/86Sr isotope ratios of three carbonate δ34 δ18 their lower SSO4 and OSO4 values, provided that neither the samples (data are from Heidel et al., 2007: 0.71010–0.71563) were isotope composition nor the sulfate concentration of end-member III rather low and maybe not representative for all carbonate veins in the had changed with time. Flowing mine water (SH 1 and SH 2) did not flooded part. The high Sr concentrations of the flooding water may at plot directly on the mixing line between end-members I and II for least partially result from mixing with end-member III, as indicated by δ18 OSO4 values (Fig. 6D, E, F). This was probably caused by an inexact S and O isotopes of sulfates (Fig. 6) and elevated concentrations of Cl, value of the oxygen isotope composition of sulfate from uncontam- Na, and Ca. However, a fourth yet unrecognized Sr source with 87Sr/ inated groundwater (caused by the limited data set), which might be 86Sr isotope ratios higher than end-member III (0.713–0.714) should slightly higher for Fig. 6D and E. There was a slight clockwise rotation have contributed to flooding water (RZUL: 0.71693–0.71716). of the mixing line between end-members I and II from 1997–1998 to Author's personal copy

M. Tichomirowa et al. / Chemical Geology 276 (2010) 104–118 115

“Primary” sulfates have very high δ18O values (40–45‰; Holt et al., 1982; Holt and Kumar, 1984). This was confirmed by measurements of sulfates in flue gases (Holt and Kumar, 1984:26–40‰) and in natural precipitation near a power plant (Jamieson and Wadleigh, 1999:29– 42‰). In contrast, “secondary” sulfate, being the ultimate oxidation product and sink of sulfur gases of biological, volcanic, and anthropo- genic origins, has much lower δ18O values, depending on the oxidation pathway and oxidants (Holt et al., 1981, 1982, 1983; Alexander et al., 2002; Savarino et al., 2003). Therefore, the decrease of δ18O values of atmospheric sulfate in the Freiberg region could indicate a lower portion of “primary” sulfate compared to “secondary” sulfate. However, the decrease of “primary” sulfate in atmospheric precipitation should be observed not only on a local scale, but on a broader regional scale. Table 6 gives a comparison of the oxygen isotope composition of sulfate from atmospheric precipitation of the Freiberg region and other regions for the investigated time period. The usual range of δ18O values of sulfate from atmospheric precipitation varies from 5 to 16‰ δ18 (Krouse and Mayer, 2000). On a regional scale, the OSO4 values from Freiberg in 1997–1998 (14.1–14.2‰) agreed with values from the 87 86 Fig. 7. Mean Sr/ Sr ratios versus mean reciprocal Sr concentrations in various water nearby Czech location (Novák et al., 2007: location Jezery with 13.1‰ types. Panel A shows two mixing lines, demonstrating the addition of water from RZUL into – RSS 1 and the resulting mixed water RSS 2 (line a) as well as the addition of water from RSS in 2000 2001) and the location in SW Poland (Jedrysek, 2000, 3 into T1 and the resulting mixed water T2 (line b). Fig. 7B shows mean values from Wroclaw: 14.3‰ in 1993–1995). For the same Polish location Tables 3–5 and single values for gw 1 (Table 2). Data for pore water as well as for Wroclaw, Gorka et al. (2008) reported mean values of 14.5‰ and carbonates from Heidel et al. (2007). The triangle shows the same three end-members as 12.5‰ for 2004 and 2005, respectively. Hence, atmospheric sulfate in Fig. 6. The mixing line between end-members I and II was drawn similar to Fig. 6 by had still high δ18O values in Wroclaw, whereas δ18O values in the starting from pore water and extending through SH 1 and SH 2. The mixing lines toward SO4 SO4 end-member III were drawn towards GS to include this sampling point in the triangle. The Freiberg region, which lies about 300 km west of Wroclaw, already grey arrow shows the shift from sampling point FM 5 towards FM 6. The dashed line decreased. Unfortunately, it is unclear if the values from Wroclaw for represents a possible mixing line between pore water and flooding water (RZUL). See text δ18 2005 may report a decrease of OSO4 values. Interestingly, the lowest for discussion. δ18 OSO4 value was observed in Wroclaw after a continuous fortnight δ18 ‰ rainfall episode, when the OSO4 value decreased from 14.0 to 9.0 . 2000–2002 (Fig. 6D, E) due to a decrease of the sulfate concentration This was interpreted as a washout effect of “primary” sulfate from δ18 fl in groundwater, while the OSO4 values of groundwater and owing long range transport (Gorka et al., 2008). mine water (SH 1 and SH 2) remained constant. In contrast, the However, no data were found at a regional scale with mean δ18O further clockwise rotation of this mixing line in Fig. 6F (2005–2007) values below 10‰ (Table 6). Hence, the decreasing influence of fi δ18 “ ” wascausedbyasignicant decrease of the OSO4 values of primary sulfate as the major reason for the observed decrease of ‰ δ18 groundwater (by 4 ) in addition to its lowered sulfate concentration. OSO4 values in the Freiberg region remains at least doubtful. The isotope composition of the uncontaminated groundwater in Alternatively, changes in oxidation pathways and/or oxidants (e.g., Fig. 6F should likely be even lower, as indicated by the position of SH 1 greater influence of water vapour with lower δ18O values) and/or and SH 2 below the proposed mixing line between end-member I and II. Table 6 Comparison of δ18O values (VSMOW, ‰) of sulfate in atmospheric precipitation from the Freiberg region and other regions from literature data from 1997 until 2008. δ18 6.4. Possible reasons for decreasing OSO4 values in atmospheric Year Freiberg region Literature values Source precipitation 1997 14.2 This work 1998 14.1 This work It has been shown, that sea salt influence in atmospheric 2000 – 13.1; 14.2 a precipitation of the region can be neglected for atmospheric precipita- 2001 – 13.1; 14.2 a 10.0±2.1 b tion since non-marine sulfate contributes between 95 and 99% of total 11.9 c sulfate in the Erzgebirge region (Zimmermann et al., 2006), consistent 2002 – 14.1 c with the limited inland penetration of sea salt particles (e.g., Gustafsson 2003 – 12.4 c δ18 2004 – 10.4 c and Franzén, 2000). For non-marine atmospheric sulfate, OSO4 values depend on sources and on oxidation pathways of atmospheric sulfur From 0 to 20 d 14.5 e components (mainly SO2). Atmospheric sources are usually divided into 13.6 f “primary” and “secondary” sulfate. “Primary” sulfate was formed before 2005 – 12.5 e its release into the atmosphere, e.g., during combustion at high 15.4 f temperatures in industrial stacks of power plants and smelters (e.g., 11.0 c 2006 5.7 This work Holt et al., 1982; Holt and Kumar, 1984). “Secondary” sulfate was 2007 7.5 This work formed after SO2 emission into the atmosphere at lower temperatures 2008 6.6 This work (e.g., Holt et al., 1981, 1982, 1983). This low temperature oxidation may a — Novák et al. (2007): mean values from two locations (LIZ and JEZ) in Czech either follow a homogenous or a heterogeneous pathway. Aqueous- Republic. phase reactions in cloud and fog droplets are ascribed to a heteroge- b — Norman et al. (2006): mean value Fraser Valley, Canada. neous oxidation pathway. Gas-phase photochemically initiated reac- c — Einsiedl et al. (2007): mean values in Franconian Alb, S Germany. tions occur during homogeneous oxidation (Holt et al., 1981, 1982, d — Wasiuta et al. (2005): snowpacks from the Prince of Wales Icefield. e — Gorka et al. (2008): mean values in Wroclaw, SW Poland. 1983). Various oxidants having different oxygen isotope compositions — − f Jenkins and Bao (2006): mean values in Baton Rouge, LA, USA. (e.g., OH ,O3 and H2O2) can oxidize SO2 to sulfate (e.g., Jamieson and Note that δ18O values of water are similar only for central Europe but can be quite Wadleigh, 1999). different for other regions of the world (e.g., Canada and USA). Author's personal copy

116 M. Tichomirowa et al. / Chemical Geology 276 (2010) 104–118

“ ” δ34 sources of secondary sulfate (biological, anthropogenic, volcanic) or investigators also found a decrease of SSO4 values during transit δ18 their relationship may have led to the observed decrease of OSO4 through the soil (e.g., Novák et al., 2003). The difference between δ18 values in atmospheric precipitation at Freiberg. Interestingly, arctic mean OSO4 values of sulfate from bulk precipitation and ground- δ18 ‰ snow showed large variations of its OSO4 values (Table 6: from 0 to water as well as river water from 1997 to 1998 was about 8 and 20‰), although it was assumed that only 5% was made up of decreased to about 6‰ in 2007. “primary” sulfate from long range transport (Wasiuta et al., 2005). The flowing mine water at sampling point SH 1 was only slightly δ18 High OSO4 values were attributed to secondary oxidation by ozone affected by mixing with sulfate from oxidation processes (e.g., (having very high δ18O values) instead of OH−. Savarino et al. (2003) Haubrich and Tichomirowa, 2002; Junghans and Tichomirowa, δ18 − ‰ reported OSO4 values between 5.5 and 23.5 for background 2009; compare Figs. 3, 5 and 6). Its mean sulfate concentration sulfate in Antarctic ice cores for a time about 750 years ago (about decreased from about 140 mg/L in 1997 (Haubrich and Tichomirowa, 18 1260 AD) and a δ OSO value of 3.4‰ for the time about 20 years ago 2002) to about 130 mg/L in 2000–2002 (Junghans and Tichomirowa, 4 − (1990 AD). They state that OH is usually the main oxidant for SO2 2009) and to about 100 mg/L in 2005–2006 (Table 4), and may be a oxidation in the atmosphere. better estimate of sulfate concentrations of groundwater instead of − Oxidation of SO2 by hydroxyl OH groups was shown to cause the single values in Table 2. Therefore, sulfate concentration in δ18 δ18 seasonal variations of OSO4 values positively correlated with OH2O groundwater in the Freiberg area decreased with time although they δ18 values (e.g., Holt et al., 1979, 1981). This correlated change in OSO4 are still on a high level compared to other regions in Europe. The δ18 and OH2O values was observed in the 1960s and 1970s at various sulfate concentration decrease in groundwater is attributed to a time- regions (Cortecci and Longinelli, 1970; Longinelli and Bartelloni, 1978; delayed response to the major decrease of industrial SO2 emissions, Holt et al., 1979, 1981) and was explained as heterogeneously oxidized starting from 1990, and a corresponding decrease in SO2 concentra- sulfate. Heterogeneous oxidation by local rain water was likely the most tions in the atmosphere of a region with previously very high important pathway at times of high atmospheric SO2 concentrations pollution levels. Low SO2 concentrations were achieved in the throughout the industrial countries. Freiberg region in 2000 AD (mean monthly concentrations reached In contrast, Jamieson and Wadleigh (1999) reported an inverse trend b10 μg/m3) and decreased only slightly thereafter (Tichomirowa et δ18 δ18 between OH2O and OSO4 values in rainwater, which was supposed al., 2004). Relatively low mean sulfate concentrations in atmospheric to have high contribution of primary sulfate from a nearby power plant. precipitation were already reached in 1997 (about 3.2 mg/L; variation δ18 Similarly, there was a weak negative correlation between OH2O and between 0.3 and 51 mg/L), with mean monthly SO2 concentrations δ18 2 μ 3 OSO4 values for sample location prRZ (R =0.62) but no correlation between 8 and 25 g/m (Haubrich and Tichomirowa, 2004). δ18 δ18 for location prIM. Obviously, both locations often had distinct OSO4 Unfortunately, OSO4 values of bulk precipitation were not δ18 values (Table 1), despite a distance below 3 km (Fig. 1). The OH2O measured in the region between 1998 and 2006. Therefore, the values for sample location prRZ agreed with the usual pattern: lower oxygen isotope shift in precipitation sulfate was discovered at the end δ18 δ18 fi OH2O values in winter and higher OH2O values in summer. This may of 2006 for the rst time. However, this shift should have occurred δ18 fl point to an undisturbed situation for the location prRZ compared to earlier as indicated by the shift of OSO4 values of owing mine sample location prIM (Fig. 1). waters SH 1 and SH 2 in 2005 (Fig. 5), because this sulfate already δ18 In general, low OSO4 values of atmospheric sulfate can be passed the soil horizon. explained by low temperature oxidation of local SO2. For example, in Mass balance calculations of the impact of rainwater (about 2–3mg/ 3+ − presence or absence of a Fe -catalyst or neutralizing NH3, HSO3 L sulfate, mean annual precipitation about 600 mm) cannot explain the δ18 δ18 (aqueous) oxidation would result in a OSO4 value between 0 and shift in OSO4 of groundwater (about 100 mg/L). However, during the ‰ δ18 − ‰ fl 8 when oxidized with water having a OH2O value of 8 (Holt et ood event in August 2002 the Freiberg region received almost 200 mm δ18 – al., 1981). This OH2O value is quite similar to that of the Freiberg precipitation within four days (10 13 August 2002; Haubrich and δ18 “ ” region (Table 1). The described negative correlation between OH2O Tichomirowa, 2004). We propose that older groundwater with higher δ18 δ18 fl and OSO4 values for location prRZ could not be caused by lower OSO4 values was at least partly ushed during this event and exported temperatures in winter (leading to higher fractionation between both by river water and mine drainage galleries. This would explain the δ18 δ18 components), because of the large change in OSO4 values. Therefore, relatively large shift of OSO4 values in groundwater, river water and either the sulfate source or the oxidant is believed to have changed in flowing mine water within only four years. winter (both with higher δ18O values). 6.6. Reasons for decreasing δ18O values of dissolved sulfate in drainage 6.5. Reasons for decreasing δ18O values of dissolved sulfate in galleries groundwater, river water, and flowing mine water The most complete dataset exists for sampling point RZUL from δ18 δ18 There is convincing evidence that the OSO4 values in surface and 2000 to 2007 (Appendix Table 3). The decrease of OSO4 values of subsurface waters (groundwater and river water) of the region of dissolved sulfate from 1997 until 2002 was accompanied by a δ18 Freiberg decreased in the last years. The shift in OSO4 values of bulk decrease of SO4, Na, and Cl concentrations in water (Fig. 8, line A). δ18 fl precipitation is the driving force that led to decreasing OSO4 values in During the ood event in August 2002, between 20 and 30% of the groundwater and river water, the latter being largely influenced by base unflooded part of the mine workings (several hundred thousands m3) flow and consequently by groundwater. However, the sulfate concen- were flushed and temporarily flooded (Kolitsch et al., 2005). Large tration in river water of the Freiberger Mulde was much lower (about amounts of highly mineralized solutions were mixed with ground- 50 mg/L) compared to local groundwater (about 100 mg/L), probably water and exported. A significant increase of ore-related element due to dilution with atmospheric precipitation (Tables 1 and 3). concentrations (e.g., Zn, Cd and Cu) occurred. These decreased to The sulfate from precipitation enters the soil layer with a local previous levels only one year after the flood (Kolitsch et al., 2005). thickness of about 1 m. Complex processes of S-retention occur in The sulfate concentration showed a different pattern along with its soils with organic S-bonding, sulfate adsorption, and precipitation of isotope composition as well as with concentrations of some other aluminum hydroxyl sulfate minerals. Thus, mean transit times of S in elements (e.g., Cl with R2 =0.88, Na with R2 =0.83, Mg with 2 2 δ18 2 soils range from several years to many decades (Mayer et al., 2001; R =0.62, Ca with R =0.51, and OSO4 with R =0.54). Concentra- δ34 Novák et al., 2004). While the SSO4 values of precipitation sulfate tions of these elements decreased due dilution with a large volume of δ18 fl remain constant during the transit through soil horizons, their OSO4 surface water (precipitation and river water) ushed into the mine values decreased by several per mil (Mayer et al., 1995). Other workings during the flood event. The input of large water volumes led Author's personal copy

M. Tichomirowa et al. / Chemical Geology 276 (2010) 104–118 117

sulfate from oxidation processes of local (sometimes low grade) ore sulfides that remained after mine decommissioning in the unflooded und flooded part. In addition to these two sulfate sources, identified in previous investigations (Haubrich and Tichomirowa, 2002; Junghans and Tichomirowa, 2009), a third sulfate source had a significant impact on contaminated groundwater and several types of mine water. End- member III was sulfate from abundant waste, including former tailings, and slag deposits, left behind after a long lasting mining and industrial history of this region. This end-member showed not only a different δ18 δ34 isotope composition of dissolved sulfate ( OSO4 and SSO4) at similar δ18 OH2O, but also obvious differences in concentrations of some dissolved elements (Cl, Pb, Na, and Ca). The same three end-members (uncontaminated groundwater, pore water from oxidation of residual ores and their host rocks, industrial water) were identified by applying Sr isotope ratios. However, Sr isotopes even showed a fourth yet unrecognized source that contributed to water of the flooded part of the mine. The data indicate that the isotope composition of river water from the Freiberger Mulde was mainly influenced by uncontaminated ground- water diluted with atmospheric precipitation, whereas the other two end-members played only a minor role. Contrary, the isotope composition of dissolved sulfate and strontium in the river Triebisch is largely influenced by end-member III after mixing with flooding water. ‰ δ18 A large decrease (by 7 )of OSO4 values of sulfate in bulk atmo- spheric precipitation was observed as of 2006 (compared to 1997), δ34 while SSO4 values remained constant. This isotope shift was probably caused by a lower contribution of “primary” sulfates to atmospheric precipitation. Alternatively, “secondary” sulfate sources (mainly of biological, and anthropogenic origin) could have changed their contributions and/or the main oxidant has changed. Groundwater and δ18 river water (Freiberger Mulde) also showed a decrease of their OSO4 valuesbuttoalowerdegree(by4–5‰). We propose that the oxygen 18 Fig. 8. Oxygen isotope composition of dissolved SO4 from flooding mine water (RZUL) δ isotope shift in atmospheric precipitation caused the OSO4 shift in versus reciprocal concentrations of sulfate, Na and Cl. Lines A and B record time trends groundwater and river water. The flood event in August 2002 very likely as discussed in the text. washed out large parts of “older” groundwater and thus enabled the relatively large isotope shift of groundwater and river water within a to a replacement of the flooded part as indicated by a sharp increase of few years. In addition, a decrease of sulfate concentrations in dissolved oxygen concentrations (from 0.5 to about 4 mg/L) and to a groundwater resulted (although it was still at a high level of about δ18 fl temporary decrease of water temperature (from 19.0 °C to about 100 mg/L in 2006). OSO4 values also decreased in owing mine water, 16.6 °C), lasting to December 2003 (Kolitsch et al., 2005 and Appendix which was mainly affected by end-member I (uncontaminated

Table 3). From 2004 on, both SO4 and Cl concentrations slightly groundwater). However, this isotope shift was not recorded in drainage increased until 2006–2007, approaching values close to those of 2002 galleries and flooding water, which have a much greater water volume. before the flood event (Fig. 8, line B). On the other hand, the Na The contribution of “industrial” sulfate (end-member III) has concentrations remained nearly constant as of 2003 (Fig. 8), most decreased with time in water from the drainage adits GS and RSS, as likely because large parts of the Na-source were flushed out during well as in the flooding water RZUL, especially for the time from 1997 the flood event. to 2002. This led to lower δ18O and δ34S values of dissolved sulfates in δ18 fl The decrease of OSO4 values from 1997 to 2002 of the ooded part these water types and was accompanied by a decrease of concentra- (RZUL, line A in Fig. 8) is interpreted as a result of a lower input from tions of Cl, Na, and Mg. No further decrease could be detected after the “industrial” sulfate (end-member III), as recorded by a similar tendency storm and flood event in August 2002, however. δ18 for Cl concentrations. This trend may be valid for drainage water from GS Our data showed that the OSO4 value of atmospheric sulfate can δ18 fi — and RSS 3 for that time period (Figs. 3 and 6). In 2003, the OSO4 values signi cantly change locally within a few years. The isotope shift was in GS and HU waters increased (Fig. 5), possibly caused by a higher at least partly — transferred to dissolved sulfate in groundwater and contribution from sulfate end-member III. This is difficult to validate river water in a short time period of less than ten years, indicating that based on single analyses only. Supporting evidence comes from RZUL the transit time of S in soil was less than a decade. However, this with increasing sulfate and Cl concentrations as of 2004, clearly “transit time” was probably accelerated due to the flushing effect of indicating an increased input of end-member III (line B in Fig. 8). the regional storm and flood event in 2002. δ18 However, this is not accompanied by a rise in OSO4 values, probably δ18 subdued by the concurrent decrease of OSO4 values of groundwater. Acknowledgements

7. Conclusions The authors thank Angelika Braun, Klaus Bombach, Rositta Liebscher, Heidrun Meinhardt, and Sabine Mühlberg for field The combined use of isotope and chemical data allowed the identi- assistance, and Rositta Liebscher and Heidrun Meinhardt for labora- fication of three major sulfate sources in various water types of the tory assistance. We gratefully acknowledge the financial support for C. Freiberg region. End-member I was sulfate from groundwater, ulti- Heidel by the Deutsche Forschungsgemeinschaft and for M. Junghans mately derived from atmospheric deposition. End-member II was by the Heinrich Böll Foundation. We thank two anonymous reviewers Author's personal copy

118 M. Tichomirowa et al. / Chemical Geology 276 (2010) 104–118 for their helpful comments substantial improving the focus of the Kolitsch, S., Junghans, M., Klemm, W., Degner, T., Baacke, D., 2005. Hydrochemical monitoring (1970–2003), depth profile and flow measurements in partly flooded paper. underground workings of the central polymetallic vein ore deposit of Freiberg/ Germany. Z. Geol. Wiss. 33, 51–80. 18 Appendix A. Supplementary data Kornexl, B.E., Gehre, M., Höfling, R., Werner, R.A., 1999. On-line d O measurement of organic and inorganic substances. Rapid. Commun. Mass Spectrom. 13, 1685–1693. Krouse, H.R., Grinenko, V.A., 1991. Stable isotopes: natural and anthropogenic sulphur Supplementary data associated with this article can be found, in in the environment. SCOPE 43. John Wiley and Sons, Chichester. the online version, at doi:10.1016/j.chemgeo.2010.06.004. Krouse, H.R., Mayer, B., 2000. Sulphur and oxygen isotopes in sulphate. In: Cook, P.G., Herczeg, A.L. (Eds.), Environmental Tracers in Subsurface Hydrology. Kluwer, pp. 195–231. References Longinelli, A., Bartelloni, M., 1978. Atmospheric pollution in Venice, Italy, as indicated by isotopic analyses. Water Air Soil Pollut. 10, 335–341. Alexander, B., Savarino, J., Barkov, N.I., Delmas, R.J., Thiemens, M.H., 2002. Climate Martin, M., Beuge, P., Kluge, A., Hoppe, T., 1994. Grubenwässer des Erzgebirges–Quelle driven changes in the oxidation pathways of atmospheric sulfur. Geophys. Res. Lett. von Schwermetallen in der Elbe. Spektrum der Wissenschaften 5, 102–107. 29, 30-1–30-4. doi:10.1029/2002GL014879. Matschullat, J., Maenhaut, W., Zimmermann, F., Fiebig, J., 2000. Aerosol and bulk Balci, N., Shanks III, W.C., Mayer, B., Mandernack, K.W., 2007. Oxygen and sulfur isotope deposition trends in the 1990's, Eastern Erzgebirge, Central Europe. Atmos. systematics of sulfate produced by bacterial and abiotic oxidation of pyrite. Environ. 34, 3213–3221. Geochim. Cosmochim. Acta 71, 3796–3811. Mayer, B., Fritz, P., Prietzel, J., Krouse, H.R., 1995. The use of stable sulfur and oxygen isotope Baumann, L., Kuschka, E., Seifert, Th., 2000. Lagerstätten des Erzgebirges. Enke, ratios for interpreting the mobility of sulfate in aerobic forest soils. Appl. Geochem. 10, Stuttgart. 300 pp. 161–173. Brand, W., Coplen, T.B., Aerts-Bijma, A., Böhlke, J.K., Gehre, M., Geilmann, H., Gröning, Mayer, B., Prietzel, J., Krouse, H.R., 2001. The influence of sulfur deposition rates on M., Jansen, H.G., Meijer, H.A.J., Mroczkowski, S.M., Qi, H., Soergel, K., Stuart- sulfate retention patterns and mechanisms in aerated forest soils. Appl. Geochem. Williams, H., Weise, S.M., Werner, R.A., 2009. Comprehensive inter-laboratory 16, 1003–1019. calibration of reference materials for δ18O versus VSMOW using various on-line Negrel, P., Fouillac, C., Brach, M., 1997. A strontium isotopic study of mineral and surface high-temperature conversion techniques. Rapid Commun. Mass Spectrom. 23, waters from the Cezallier (Massif Central, France): implications for mixing 999–1019. processes in areas of disseminated emergences of mineral waters. Chem. Geol. Bridges, K.S., Jickells, T.D., Davies, T.D., Zeman, Z., Hunova, I., 2002. Aerosol, precipitation 135, 89–101. and cloud water chemistry observations on the Czech Krusne Hory plateau Negrel, P., Casanova, J., Aranyossy, J.-F., 2001. Strontium isotopic systematics used to adjacent to a heavily industrialised valley. Atmos. Environ. 36, 353–360. decipher the origin of groundwaters sampled from granitoids: the Vienne Case Cortecci, G., Longinelli, A., 1970. Isotopic composition of sulfate in rain water; Pisa, Italy. France. Chem. Geol. 177, 287–308. Earth Planet. Sci. Lett. 8, 36–40. Nordstrom, D.K., Wright, W.G., Mast, M.A., Bove, D.J., Rye, R.O., 2007. Aqueous-sulfate Ding, T., Valkiers, S., Kipphardt, H., DeBiévre, P., Taylor, P.D.P., Gonfiantini, R., Krouse, R., stable isotopes — a study of mining-affected and undisturbed acidic drainage. In: 2001. Calibrated sulfur isotope abundance ratios of three IAEA sulfur isotope Church, S.E., von Guerard, P., Finger, S.E. (Eds.), Integrated Investigations of reference materials and V-CDT with reassessment of the atomic weight of sulfur. Environmental Effects of Historical Mining in the Animas River Watershed, San Geochim. Cosmochim. Acta 65, 2433–2437. Juan County, Colorado, USGS Prof. Paper 1651, pp. 391–416. chapter E8. Einsiedl, F., Schäfer, T., Northrup, P., 2007. Combined sulfur K-edge XANES spectroscopy Norman, A.-L., Anlauf, K., Hayden, K., Thompson, B., Brook, J.R., Li, S.-M., Bottenheim, J., and stable isotope analyses of fulvic acids and groundwater sulfate identify sulfur 2006. Aerosol sulphate and its oxidation on the Pacific NW coast: S and O isotopes cycling in a karstic catchment area. Chem. Geol. 238, 268–276. in PM2.5. Atmos. Environ. 40, 2676–2689. Epstein, S., Mayeda, T., 1953. Variations of O18 content of waters from natural sources. Novák, M., Jacková, I., Prechová, E., 2001. Temporal trends in the isotope signature of Geochim. Cosmochim. Acta 4, 213–224. air-borne sulfur in Central Europe. Environ. Sci. Technol. 35, 255–260. Faure, G., 2001. Origin of Igneous Rocks — The Isotopic Evidence. Springer, Berlin. Novák, M., Buzek, F., Harrison, A.F., Prechová, E., Jacková, I., Fottová, D., 2003. Similarity Giesemann, A., Jäger, H.-J., Norman, A.L., Krouse, H.R., Brand, W.A., 1994. On-line sulfur between C, N, and S stable isotope profiles in European spruce forest soils: isotope determination using an elemental analyzer coupled to a mass spectrom- implications for the use of δ34S as a tracer. Appl. Geochem. 18, 765–779. eter. Anal. Chem. 66, 2816–2819. Novák, M., Michel, R.L., Prechová, E., Stepanová, M., 2004. The missing flux in a 35Sbudget Gorka, M., Jedrysek, M.O., Strapoc, D., 2008. Isotopic composition of sulphates from for the soils of a small polluted catchment. Water, Air, Soil Poll. Focus 4, 517–529. meteoric precipitation as an indicator of pollutant origin in Wroclaw (SW Poland). Novák, M., Mitchell, M.J., Jacková, I., Buzek, F., Schweigstillová, J., Erbanová, L., Prikryl, R., Isot. Environ. Health Stud. 44, 177–188. Fottová, D., 2007. Processes affecting oxygen isotope ratios of atmospheric and Gustafsson, M.E.R., Franzén, L.G., 2000. Inland transport of marine aerosols in southern ecosystem sulfate in two catchments in Central Europe. Environ. Sci. Technol. 41, Sweden. Atmos. Environ. 34, 313–325. 703–709. Haubrich, F., Tichomirowa, M., 2002. Sulfur and oxygen isotope geochemistry of acid Petelet-Giraud, E., Luck, J.-M., Ben Othman, D., Negrel, P., 2003. Dynamic scheme of mine drainage — the polymetallic sulphide deposit “Himmelfahrt Fundgrube” in water circulation in karstic aquifers as constrained by Sr and Pb isotopes. Freiberg (Germany). Isot. Environ. Health Stud. 38, 121–138. Application to the Herault watershed, Southern France. Hydrogeol. J. 11, 560–576. Haubrich, F., Tichomirowa, M., 2004. Natürliche 34S- und 18O- Variationen von 1997 bis Savarino, J., Bekki, S., Cole-Dai, J., Thiemens, M.H., 2003. Evidence from sulfate mass 1998 in der Atmosphäre im Raum Freiberg/Sachsen und ihre Ursachen. Wiss. Mitt. independent oxygen isotopic compositions of dramatic changes in atmospheric Inst. Geol. Freiberg 27, 51–72. oxidation following massive volcanic eruptions. J. Geophys. Res. 108 (D21). Heidel, C., Tichomirowa, M., Matschullat, J., 2007. Lead and strontium isotopes as doi:10.1029/2003JD003737 ACH 7-1–7-6. indicators for mixing processes of waters in the former mine “Himmelfahrt Siegel, D.I., Bickford, M.E., Orrell, S.E., 2000. The use of strontium and lead isotopes to Fundgrube”, Freiberg (Germany). Isot. Environ. Health Stud. 43, 339–354. identify sources of water beneath the Fresh Kill landfill, Staten Island, New York, Heidel, C., Tichomirowa, M., Junghans, M., 2009. The influence of pyrite grain size on the USA. Appl. Geochem. 15, 493–500. final oxygen isotope difference between sulphate and water in aerobic pyrite Singh, S.K., Kumar, A., France-Lanord, C., 2006. Sr and 87Sr/86Sr in waters and sediments – oxidation experiments. Isot. Environ. Health Stud. 45, 321 342. of the Brahmaputra river system: silicate weathering, CO2 consumption and Sr flux. Holt, B.D., Kumar, R., 1984. Oxygen-18 study of high-temperature air oxidation of SO2. Chem. Geol. 234, 308–320. Atmos. Environ. 18, 2089–2094. Tichomirowa, M., Junghans, M., 2009. Oxygen isotope evidence for sorption of Holt, B.D., Kumar, R., 1991. Oxygen isotope fractionation for understanding the sulphur molecular oxygen to pyrite surface sites and incorporation into sulfate in oxidation cycle. In: Krouse, H.R., Grinenko, V.A. (Eds.), Stable Isotopes: Natural and experiments. Appl. Geochem. 24, 2072–2092. Anthropogenic Sulphur in the Environment. : Scope, 43. John Wiley and Sons, Tichomirowa, M., Pelkner, S., Junghans, M., Haubrich, F., 2003. Sulfide oxidation at the Chichester, pp. 27–41. polymetallic sulphide deposit Freiberg (Germany) and consequences for heavy Holt, B.D., Cunningham, P.T., Kumar, R., 1979. Seasonal variations of oxygen-18 in metal mobilisation. In: Schulz, H.D., Hadeler, A.D. (Eds.), Geochemical Processes in atmospheric sulfates. Int. J. Environ. Anal. Chem. 6, 43–53. Soil and Groundwater. Wiley-VCH Weinheim, pp. 356–379. Holt, B.D., Kumar, R., Cunningham, P.T., 1981. Oxygen-18 study of the aqueous-phase Tichomirowa, M., Bombach, K., Liebscher, R., 2004. Schwefel- und Sauerstoffisotopen- oxidation of sulfur dioxide. Atmos. Environ. 15, 557–566. werte der Atmosphäre in Sachsen: Messungen 2000–2004 und zusammenfassende Holt, B.D., Kumar, R., Cunningham, P.T., 1982. Primary sulfates in atmospheric sulfates: Interpretation. Wiss. Mitt. Inst. Geol. Freiberg 27, 73–82. estimation by oxygen isotope ratio measurements. Science 217, 51–53. Tichomirowa, M., Haubrich, F., Klemm, W., Matschullat, J., 2007. Regional and temporal Holt, B.D., Cunningham, P.T., Engelkemeir, A.G., Graczyk, D.G., Kumar, R., 1983. Oxygen- (1992–2004) evolution of air-borne sulphur isotope composition in Saxony, 18 study of non-aqueous phase oxidation of sulfur dioxide. Atmos. Environ. 17, southeastern Germany, central Europe. Isot. Environ. Health Stud. 43, 295–305. – 625 632. Wasiuta, V.L., Norman, A.-L., Marshall, S., 2005. Moisture sources and SO2 oxidation Jamieson, R.E., Wadleigh, M.A., 1999. A study of the oxygen isotopic composition of pathways: evidence from the snow pack sulphate from the Prince of Wales Icefield, precipitation sulphate in Eastern Newfoundland. Water Air Soil Pollut. 110, 405–420. Ellesmere Island. www.arcticnet.ulaval.ca/pdf/posters_2005/wasiuta_and_marshall. 2− Jedrysek, M.O., 2000. Oxygen and sulphur isotope dynamics in the SO4 of an urban pdf. precipitation. Water Air Soil Pollut. 117, 15–25. Zimmermann, F., Lux, H., Maenhaut, W., Matschullat, J., Plessow, K., Reuter, F., Jenkins, K.A., Bao, H., 2006. Multiple oxygen and sulfur isotope compositions of Wienhaus, O., 2003. A review of air pollution and atmospheric deposition dynamics atmospheric sulfate in Baton Rouge, LA, USA. Atmos. Environ. 40, 4528–4537. in southern Saxony, Germany, Central Europe. Atmos. Environ. 37, 671–691. Junghans, M., Tichomirowa, M., 2009. Using sulfur and oxygen isotope data for sulfide Zimmermann, F., Matschullat, J., Brüggemann, E., Plessow, K., Wienhaus, O., 2006. oxidation assessment in the Freiberg polymetallic sulfide mine. Appl. Geochem. 24, Temporal and elevation-related variability in precipitation chemistry from 1993 to 2034–2050. 2002, Eastern Erzgebirge, Germany. Water Air Soil Pollut. 170, 123–141.