Zinc isotope systematics in snow and ice accretions in Central European mountains

Chemical Geology 388 (2014) 130–141

Contents lists available at ScienceDirect

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

Zinc isotope systematics in snow and ice accretions in Central European mountains Petra Voldrichova a,b, Vladislav Chrastny a, Adela Sipkova a, Juraj Farkas a, Martin Novak a,⁎, Marketa Stepanova a, Michael Krachler c, Frantisek Veselovsky a, Vladimir Blaha a, Eva Prechova a, Arnost Komarek d, Leona Bohdalkova a, Jan Curik a, Jitka Mikova a, Lucie Erbanova a, Petra Pacherova a a

Division of Geochemistry and Laboratories, Czech Geological Survey, Geologicka 6, 152 00 Prague 5, Czech Republic Department of Analytical Chemistry, Faculty of Science, Charles University, Albertov 6, 128 43 Prague 2, Czech Republic c European Commission Joint Research Centre, Institute for Transuranium Elements, 76125 Karlsruhe, Germany d Faculty of Mathematics and Physics, Charles University, Sokolovska 83, 186 75 Prague 8, Czech Republic b

a r t i c l e

i n f o

Article history: Received 18 March 2014 Received in revised form 2 September 2014 Accepted 5 September 2014 Available online 16 September 2014 Editor: Carla M. Koretsky Keywords: Zinc Isotopes Deposition Atmosphere Snow Ice accretion

a b s t r a c t Zinc (Zn) pollution negatively affects human and ecosystem health. We quantified atmospheric Zn inputs at six remote mountain-top locations in the Czech Republic (Central Europe), and used δ66Zn isotope ratios to identify Zn from different pollution sources. The study sites were located at an elevation of approximately 1000 m near the state borders with Germany and Poland. During two winter seasons (2009–2010), over 400 samples of vertical deposition (snow) and horizontal deposition (ice accretions) were collected. Zinc pollution levels were generally low. Zinc concentrations in snow and ice accretions were less than twice as high in the east, compared to the west. Across the sites, over 90% of Zn was present in a weak-acid soluble form. Zinc concentrations were 5 times higher in ice accretions, which formed from small droplets originating in the basal cloud layer, rich in pollutants, than in snow. In contrast, droplets resulting in snow formation were larger and scavenged less pollution due to their smaller surface area. δ66Zn of Pribram sphalerite (west) and smelter-derived fly ash (west) were low, −0.23 and −0.47‰, respectively. Olkusz sphalerite (east) had a higher δ66Zn of 0.02‰. δ66Zn of snow ranged from −0.60 to 0.68‰. Ice accretions had δ66Zn between −0.67 and 0.14‰. At the three eastern sites, δ66Zn of ice accretions was lower than δ66Zn of snow, suggesting the presence of volatilized smelter-derived or coalburning derived Zn. δ66Zn of ice accretions at two of the three western sites was higher than δ66Zn of snow. Different δ66Zn values of snow and ice accretions from the same site reflected different pollution sources, which may have been situated at different distances from the receptor site. δ66Zn of the soluble Zn fraction was higher than δ66Zn of the insoluble Zn fraction, possibly also indicating a different origin of these two Zn fractions. Zinc isotope heterogeneity in the atmosphere of remote areas indicates that δ66Zn can be a useful tool in pollution provenance studies. © 2014 Published by Elsevier B.V.

1. Introduction Zinc is the second most important transition metal in the human body, due to its unique role in enzymes and proteins (Viers et al., 2007). Excessive supply of Zn, however, has a detrimental effect on living organisms (Cloquet et al., 2008). Despite its only moderate overall toxicity, Zn pollution is of concern in many industrial and developing countries (Shiel et al., 2010). Metallurgy, coal burning, tire wear, zinccoated roofs, fertilizers and pesticides can be major sources of Zn to the environment (Borrok et al., 2008). Air pollution by a suite of trace metals, including Zn, has been quantified in many countries over the past 40 years (Freydier et al., 1998; Simonetti et al., 2000; Luck and Ben Othman, 2002). Zinc contamination of urban atmospheres is still ⁎ Corresponding author. E-mail address: [email protected] (M. Novak).

http://dx.doi.org/10.1016/j.chemgeo.2014.09.008 0009-2541/© 2014 Published by Elsevier B.V.

increasing, in contrast to many other trace metals (Thapalia et al., 2010). Anthropogenic Zn emissions in developing countries are also increasing (Dolgopolova et al., 2006). Over the past 15 years, Zn isotope ratios have been used to trace sources and transport pathways of pollution (Albarede, 2004). Zinc has five stable isotopes (64Zn, 66Zn, 67Zn, 68Zn, and 70Zn). Zinc isotope abundances are measured by thermal ionization mass spectrometry (TIMS), or multi-collector plasma-source mass spectrometry (MC ICP MS). The ratio of the two most abundant Zn isotopes (66Zn/64Zn; abundances of 28 and 48%, respectively) is used to report Zn isotope systematics. The commonly used δ66/64Zn (δ66Zn) ratio expresses the relative ‰ deviation of the Zn isotope composition of a sample from that of a standard. Zinc exists in nature in only one oxidation state (+2). The range of δ66Zn values of geological materials on Earth is relatively narrow, not exceeding 2‰. Several authors (Pokrovsky et al., 2005; Borrok et al.,

P. Voldrichova et al. / Chemical Geology 388 (2014) 130–141

2009; Cloquet et al., 2006, 2008) have compared the ranges of δ66Zn values among various rock types and environmental sources. Weiss et al. (2007) have summarized our current knowledge of Zn isotope fractionations in nature. Igneous processes in the mantle and crust do not fractionate Zn isotopes. Zinc is a relatively volatile element, and its condensation occurs under lower temperatures compared to other base metals, such as copper (Sivry et al., 2008). Due to a kinetic fractionation, Zn vapor emitted from a smelter is isotopically light (i.e., has low δ66Zn values; Juilliot et al., 2011). Whereas globally distributed sphalerites exhibit an average δ66Zn value of 0.20‰ (Marechal and Albarede, 2002), smelter emissions are characterized by values close to −0.30‰ (Sonke et al., 2008). δ66Zn values of rocks and unpolluted environmental samples cluster between 0.20 and 0.50‰ (Thapalia et al., 2010). Most anthropogenic Zn samples are isotopically lighter (0.10 to 0.30‰; John et al., 2007). Zinc isotopes have been successfully applied as tracers in many polluted and unpolluted ecosystems (Luck et al., 1999; Mattielli et al., 2006; Weiss et al., 2007; Viers et al., 2007; Bigalke et al., 2010). Some features of Zn isotope systematics, however, complicate the use of δ66Zn as a tracer. It has been shown, for example, that δ66Zn values in a single ore deposit may vary by as much as 0.70‰ (Mason et al., 2005). Similarly, the zinc isotope composition of an individual plant may vary by as much as 1.60‰ (Cloquet et al., 2008). With no clearcut isotope composition of mixing end-members, source apportionment at a receptor site may be difficult (Chen et al., 2008). The Czech Republic has been known for a sharp (5- to 10-fold) pollution gradient, with an industrialized, polluted northwest, and nearly unpolluted rural south (Novak et al., 2008). After the years of peak industrial pollution (the late 1980s), industrial emission rates of various pollutants have decreased substantially. Recently, a 300-km shift in the highest industrial pollution has been reported from the northwest to the northeast, from the North Bohemian soft-coal basin to the Lower Silesia stone-coal basin (Bohdalkova et al., 2012). The most visible legacy of the North Bohemian industrialization is spruce die-back, caused by acid rain (Novak et al., 2007; Oulehle et al., 2013a). The Lower Silesian conurbation is notorious for the highest air-borne dust levels in today's Europe, and low life expectancy (Erbanova et al., 2008). Atmospheric input of pollutants into ecosystems is comprised of wet and dry deposition. The main forms of wet deposition are vertical deposition (rain and snow) and horizontal deposition (fog and ice accretions; Moldan and Cerny, 1994). Dry deposition increases with surface roughness, and may bring considerable amounts of pollutants. Larger particles tend to be deposited closer to the pollution source than finer particles. In industrial regions, ice accretions are often characterized by higher concentrations of pollutants, compared to snow (Dousova et al., 2007). Snow and ice accretions have never been analyzed for Zn isotope composition, despite their potential for isotope fingerprinting in environmental health studies. Here we present δ66Zn values of snow and ice accretions collected at 6 mountain-top locations in the Czech Republic, Central Europe. Winter seasons were chosen because higher demand for electricity results in higher emissions from coalburning power plants, compared to other seasons. Furthermore, frequent temperature inversions slow the dispersion of pollutants from the industries, and the concentrations of pollutants in the air increase. Our objective was to compare the Zn isotope composition between vertical and horizontal deposition in a region characterized by a sharp spatial gradient in industrial pollution. We hypothesized that Zn isotope systematics may differ between vertical and horizontal deposition, possibly reflecting different provenance of these two pools of Zn. Our second objective was to quantify and compare input fluxes of soluble and insoluble Zn at remote upland locations of Central Europe. Previous literature often used the insoluble form of Zn for isotope analysis, using total HF digests (Dong et al., 2013). The soluble form was rarely analyzed for δ66Zn, yet it was unclear which form supplies higher fluxes of atmospheric Zn into ecosystems.

131

2. Materials and methods 2.1. Study sites In 2009, we established a new hydrogeochemical monitoring network in high-elevation regions of the Czech Republic, less than 1 km from state borders with Germany, Poland and Slovakia. All study sites were situated on mountain peaks at elevations of about 1000 m, and were unforested. For this study, we collected data at six sites (Fig. 1, Table 1). Three sites (KAP, LOU and PRA) were situated in the western part of the Czech Republic. The other three sites (POM, VRH and ELK) were located in the eastern part of the country. The study sites were remote from local pollution sources, approximately 5–10 km away from the nearest settlement, road, or trail.

2.2. Sampling Snow and ice accretions were collected during two winter seasons between February 2009 and April 2010. The mean interval between snow samplings was 7 days, and the mean interval between ice accretion samplings was 8 days. Both sample types on each sampling date were collected in triplicate. The snow samples were taken from snowpack surfaces (30 by 10 by 3 cm) on the mountain summit. The distance between replicate snow samples was 20 to 50 m. Three wooden poles carrying ice accretion samplers (Figs. S1–S3; Electronic Annex) were erected at fixed positions at mutual distances of 20 to 50 m. Each sampling device consisted of a horizontal wooden bar attached to a wooden pole 1.5 m above the snow pack surface. The samplers were polyethylene (PE) rectangles (14 by 8 by 2 cm) with a high surface area. Each PE rectangle was covered by 784 “thorns” (14 by 2 by 0.6 mm) on both sides. Each stand carried two pairs of these rectangles, installed perpendicularly (N–S, E–W), and one additional pair of rectangles made of PE mesh (14 by 8 by 0.1 cm), also installed perpendicularly. As we found by in-situ trials, under some meteorological conditions, large-surface mesh was a more efficient rime scavenger than the thorns. Under other meteorological conditions, the opposite was true. The distance between adjacent pairs of rectangles on the bar was 25 cm. All six rectangles (four with thorns, two made of mesh) were used to collect one combined sample of ice accretions. Three ice accretion samples were collected simultaneously at each study site. In all, we collected 237 snow samples and 179 ice accretion samples. These samples were analyzed for Zn and Sc concentrations. Thirty five selected snow samples and 9 selected ice accretion samples were analyzed for Zn isotope ratios. Relatively Zn-rich samples were selected for the isotope analysis. At each site, snow samples selected for δ66Zn measurements included both winters, and were complemented by at least one ice accretion sample. Pre-cleaning of ice accretion samplers took 48 h. The PE rectangles were submerged in semiconductors-purity HNO3, diluted with deionized water (DW; Milipore, MiliQ 18.2 MΩ cm) in a 1:6 ratio, for 24 h. After rinsing, the rectangles were submerged in DW for another 24 h. Each snow and ice accretion sample was placed in a 1.5 L pre-cleaned polypropylene (PP) container and kept at −20 °C. Three samples of sphalerite (ZnS), one sample of hemimorphite [Zn4 Si 2O 7 (OH)2 ·H2 O], and one sample of tetrahedrite [(Cu,Fe,Ag, Zn)12Sb4S13] finely intergrown with chalcopyrite (CuFeS2) were selected from private collections, and analyzed for δ66Zn. Fly ash from the Pribram base-ores smelter and soft coal from Sokolov–Antonin were also analyzed for δ 66 Zn. Additionally, 4 samples of the b0.05 mm fraction of the 0 horizon of forest soils from two Zn mining regions, Olkusz (Poland) and Pribram (Czech Republic), were analyzed for δ66Zn. The soil sampling procedure was described in detail by Chrastny et al. (2012).

132

P. Voldrichova et al. / Chemical Geology 388 (2014) 130–141

Fig. 1. Study sites. Large point sources of pollution include coal-burning power plants and base-metal smelters. Diffuse pollution is related to traffic.

(67Zn/64Zn = 1.14), to highly ‘overspiked’ (67Zn/64Zn = 8.552) solutions. The raw 67Zn/64Zn ratios and δ66/64Zn values from these variably spiked solutions are presented in Fig. S4. Our double-spike setup was well calibrated and robust to possible variations in the sample-tospike ratio. This was documented in the δ66/64Zn values that scattered close to 0‰ from mixtures whose 67Zn/64Zn ratios ranged from 0.450 (underspiked) up to 2.084 (overspiked). Finally, we also observed that two highly ‘overspiked’ standards, with 67Zn/64Zn of 4.937 and 8.552 yielded slightly but systematically lower (about 0.1‰) δ66/64Zn values, and therefore we aimed not to overspike our natural samples to such high levels (i.e., 67Zn/64Zn N 4.9).

2.3. Preparation and calibration of a Zn double spike We adopted the double-spike approach in order to correct for isotope fractionation effects on measured δ66/64Zn values (Tanimizu et al., 2002). These effects could originate either from the sample purification procedure or from instrumental fractionation (i.e., mass bias) during the measurements by MC ICP MS. For this study, we used a 67 Zn–70Zn double-spike solution that was prepared by mixing two single-spike solutions. The 67Zn single-spike solution was prepared by dissolving about 10 mg of Zn metal oxide spike (ISOFLEX, 67Zn isotope enrichment = 89.60%) in a concentrated suprapure HNO3. Similarly, the 70Zn single spike was prepared by dissolving about 9 mg of a metal oxide spike (ISOFLEX, 70Zn isotope enrichment = 99.53%) in an adjusted volume of concentrated HNO3. These two single-spike solutions were then mixed proportionally to yield the following calculated isotope ratios: 66Zn/64Zn = 2.470516, 67Zn/64Zn = 56.914564, 68 Zn/64Zn = 3.213595, and 70Zn/64Zn = 22.625372. For the 67Zn/64Zn ratio, the optimal sample-to-spike ratio was then close to unity. The associated errors were negligible. The mixture of all four Zn isotopes was such that 90–95% of the isotopes 64Zn and 66Zn came from the sample, and 90–95% of the isotopes 67Zn and 70Zn came from the double spike. In order to test the sensitivity of measured δ66/64Zn values to variable sample-to-spike mixing ratios, we prepared a set of mixtures (NIST 682 and the double spike) whose isotope compositions ranged from highly ‘underspiked’ (67Zn/64Zn = 0.450), through ‘normal’

2.4. Analysis Sample processing took place in a laminar flow box (ISO class 7). The mass of each ice and snow sample was determined before melting. Precipitation samples were acidified (0.6 M HNO3; Romil) while melting and filtered (0.45 μm). The filtrate was analyzed for soluble Zn and Sc concentrations by sector-field ICP MS (Element 2, Thermo Fisher Scientific) at the University of Heidelberg (Germany). Ultra-clean techniques were used throughout the analysis, as previously adapted for the determination of trace elements in polar ice (Shotyk and Krachler, 2009). A self-aspirating PFA nebulizer, connected to a desolvating sample introduction system was used. The Zn detection limit was 0.001 μg L− 1. The river water reference material SLRS-5 (National Research Council Canada) with a certified Zn concentration

Table 1 Study site characteristics. Site KAP Kaprad LOU Kamenna loucka PRA Pramenac POM Pomezni hreben VRH Pricny vrch ELK Velky Polom

Bohemian Forest Mts. Bohemian Forest Mts. Ore Mts. Giant Mts. Jeseniky Mts. Beskydy Mts.

Location

Elevation (m)

Mean annual temperature (°C)

Mean annual precipitation (mm)

Prevailing wind direction

48°51′N, 13°46′E 49°22′N, 12°46′E 50°42′N, 13°43′E 50°43′N, 15°49′E 50°13′N, 17°22′E 49°30′N, 18°40′E

941 925 935 1183 954 1006

4.7 4.8 4.3 1.5 4.9 5.1

953 1016 857 1415 1018 1295

West Northwest Northwest Northwest Northwest Northwest

P. Voldrichova et al. / Chemical Geology 388 (2014) 130–141

of 0.845 ± 0.095 μg L−1 was analyzed at regular intervals between samples. Our experimental value of 0.820 ± 0.038 μg Zn L−1 (n = 40) was in excellent agreement with the certified value. The Sc detection limit was 0.006 ng L−1. Repeated analysis of SLRS-5 gave an average Sc concentration of 9.10 ± 0.70 ng L−1 (n = 40), which is in agreement with the published value of 8.7 ± 1.5 ng L−1 (Heimburger et al., 2013). Further analytical work was performed at the Czech Geological Survey, Prague (Czech Republic). Insoluble particles were mixed with 3.5 mL of concentrated HF and 1.5 mL of concentrated HNO3 (both UpA grade, Romil Ltd., Cambridge) in a PFA digestion vessel (Savillex, USA). The mixture was kept at 160 °C for 24 h. The residue was evaporated to dryness and 1.5 mL of concentrated HNO3 was added. The step was repeated three times. 1.8 mL of concentrated H2O2 was added, and the residue was evaporated to dryness and dissolved in 4 mL 8 M HCl (Romil). One milliliter of this solution was evaporated to dryness, dissolved in 5 mL of 0.6% HNO3 (Romil) and analyzed for Zn and Sc concentrations. Zinc for isotope analysis was separated from solution by elution chromatography, using a modified procedure by Chapman et al. (2006). 0.6 mL of AG–1MP 100–200 mesh (Bio-Rad) resin, suspended in DW, was pipetted to a PE column (total volume of 10 mL). The column was topped up with 0.1 M HCl and drained. This step was repeated three times. The resin was then pre-conditioned with 6 mL of 8 M HCl. One milliliter of the sample in 8 M HCl was added. Matrix removal was performed by adding 1.5 mL of 8 M HCl. To separate Cu, a total amount of 11.5 mL of 6 M HCl was used. Subsequently, to separate Fe, a total amount of 11 mL of 3.5 M HCl was added. To collect Zn, 0.1 M HCl was added in 3 steps: 1, 2, and 4 mL. The overall Cu–Fe–Zn separation efficiency is seen in Fig. S5. The Zn recovery was better than 93% for all samples. The eluted Zn was pooled for isotope analysis. The procedural blank was less than 1% of the overall intensity measured on mass 66. Zinc isotope analysis was carried out on a MC ICP MS Neptune (Thermo Fisher Scientific). The amount of Zn in each measured sample was approximately 500 ng. The introduction interface included Ni cones (X cones). Samples were introduced into the Ar plasma via desolvating nebulizer Aridus II (Cetac, USA) at a fluid flow rate of 50 μL min−1. Zinc isotopes of masses 64, 66, 67, 68, and 70 were measured simultaneously. The integration time was 8 s, the idle time was 3 s, and each measurement consisted of 40 cycles. Washout times between samples lasted 200 s. For a Zn concentration of ca. 100 μg L− 1, we obtained an intensity of 10 V for the 64Zn isotope. For each sample, blank intensities (b0.005 V) were subtracted from the measured intensities. Isobaric interference of 64Ni on 64Zn were monitored by means of 62Ni signals and found to be negligible for the entire data set. 136Ba2 + was monitored at 69 amu to allow correction for any doubly-charged barium interference on mass 68. We used the 67Zn/64Zn and 70Zn/64Zn raw ratios to correct the instrumental mass bias. The procedure was based on nested-iteration subtraction of the double-spike composition (Bullen, 2007). The 66Zn/64Zn ratios are reported in the δ66Zn notation relative to the SRM NIST 683 standard. Fig. S6 gives a time series of standard measurements. The external reproducibility of the δ66Zn measurements was ± 0.03 (2 SD; Fig. S6).

2.5. Statistical analysis Zinc concentration data were pre-processed by subtracting procedural blanks. Statistical analysis was conducted using the R software, version 3.0.2. (R Core Team 2013). For each sample type (ice accretions soluble and insoluble, snow soluble and insoluble), the six sites were compared by one-way analysis of variance (ANOVA). In each comparison, statistically significant differences between at least two sites were determined. Multiple comparisons were then performed using the Tukey–Kramer method. In figures and tables, multiple-comparison labels (a, ab, etc.) are given. Letters were chosen alphabetically,

133

corresponding to the highest, second highest, third highest, etc. values. Sites which do not share a common letter differ significantly at the 5% level. 2.6. Transport trajectories of air-borne Zn Trajectories of air masses bringing Zn to two study sites (PRA and VRH) were identified using the HYSPLIT model (http://ready.arl.noaa. gov/HYSPLIT.php). The meteorological database GDAS was chosen for the calculations. Input data included localization and elevation of receptor sites. Air-mass trajectories were computed at 6-hour intervals up to the time span (age) of each precipitation sample. Air-mass trajectories were plotted separately for year 1 and year 2. At each site, only those days were included for which δ66Zn values of snow were measured. Air-mass trajectories were not included if snow samples were older than 5 days. 3. Results 3.1. Zn concentrations in snow and ice accretions The concentrations of soluble Zn species were significantly higher than the concentrations of insoluble Zn species. This was true for both snow and ice accretions, and throughout the observation period (Table 2). Across the sites, snow contained on average 11 ± 2 μg of soluble Zn L−1, while ice accretions contained on average 60 ± 6 μg of soluble Zn L−1 (means ± standard errors). The variability in Zn concentrations is seen in Table S1. Across the sites, snow and ice accretions had a similarly low concentration of HF-extracted Zn (2 ± 0.4 μg L−1; 6 and 30 times lower compared to soluble Zn in snow and ice accretions, respectively). Overall, Zn concentrations in the more abundant soluble form were significantly higher (p b 0.05) in ice accretions than in snow (by five times on average). The year-to-year variability in soluble Zn concentrations was lower for snow (Fig. 2a) than for ice accretions (Fig. 2b). However, in only one case out of 10 the year-to-year variability was significant (p b 0.05; soluble Zn in ice accretions at ELK, with the second winter more polluted). In most cases, individual study sites had similar soluble Zn concentrations (Fig. 2a,b). For snow, KAP in the southwest had the lowest soluble Zn concentrations, while ELK in the northeast had the highest soluble Zn concentrations (Table 2). The difference was statistically significant at the p = 0.05 level. In ice accretions, PRA (northwest) had the lowest soluble Zn concentrations, while ELK (northeast) had the highest soluble Zn concentrations (Table 2), but the difference was statistically insignificant. The previously reported 5-to-10 fold pollution gradient for other pollutants, such as sulfur (S), and lead (Pb) between areas around PRA in the northwest and around KAP in the southwest (e.g., Novak et al., 2005, 2008) was non-existent. There were no statistically significant differences for any Zn form in Table 2 for these two sites.

Table 2 Mean Zn concentrations in soluble and insoluble Zn forms in snow and ice accretions. Data for the combined winter seasons 2009–2010. Different letters in superscript mark statistically different values (p b 0.05). Site

Mean Zn concentration (μg L−1) ± SE Snow

Ice accretions

Soluble KAP LOU PRA POM VRH ELK

4.65 14.0 8.34 6.13 16.0 18.8

± ± ± ± ± ±

Insoluble d

0.7 3.7abc 1.1bcd 1.4cd 2.4ab 2.3a

1.28 1.27 1.62 0.48 3.01 3.26

± ± ± ± ± ±

Soluble mn

0.8 0.3mn 0.4mn 0.2n 0.4m 1.0m

59.1 54.0 51.2 45.8 60.7 86.7

± ± ± ± ± ±

Insoluble rs

14.0 13.3rs 5.6rs 5.2s 10.1rs 9.8r

0.83 1.73 2.76 2.57 1.94 2.32

± ± ± ± ± ±

0.2z 0.6z 0.8z 0.6z 0.7z 0.7z

134

P. Voldrichova et al. / Chemical Geology 388 (2014) 130–141

insoluble Zn in both vertical and horizontal precipitation (snow and ice accretions) by water fluxes. As seen in Fig. 2c, the soluble Zn input into the ecosystems prevailed by far over the insoluble Zn input. The largest proportion of insoluble Zn input was observed at VRH in the east (18% of the total Zn input). Very small proportions of insoluble Zn input were seen at KAP and PRA in the west (1% of the total Zn input). Across the sites, 94% of the Zn deposition flux occurred in the soluble form. As also seen in Fig. 2c, the Zn deposition flux was greater in the east than in the west. In winter, ELK and VRH, the sites closest to the Lower Silesian industrial region, exhibited the highest Zn input flux into the ecosystems. Across the sites, the Zn input flux slightly increased with increasing total precipitation (R2 = 0.26), however, this increase was statistically insignificant (p N 0.05). 3.3. Sc concentrations and fluxes in precipitation Scandium concentrations (Table S1) averaged 12, 63, 29 and 50 ng L− 1 in soluble snow, insoluble snow, soluble ice accretions, and insoluble ice accretions, respectively. Scandium concentrations were roughly 1000 times lower than Zn concentrations. Scandium deposition fluxes were similar at five sites, LOU, PRA, POM, VRH and ELK (Fig. 3). KAP had a 10 times lower Sc deposition flux than those five sites. 3.4. δ66Zn of pollution sources

Fig. 2. Concentrations of soluble Zn in snow (a) and ice accretions (b) at mountain-top locations. Water-soluble Zn constitutes on average 94% of total Zn. Different letters denote significantly different values at the p b 0.05 level. Total deposition (c) represents the seasonal Zn input flux from the atmosphere to the ecosystems.

There was a slight trend toward higher Zn concentrations in all four forms of atmospheric deposition (i.e., soluble and insoluble snow and ice) when data for the three eastern sites (POM, VRH, ELK) were combined, in comparison with the combined data from the three western sites (KAP, LOU, PRA). Zinc concentrations across the eastern sites were 1.6, 1.6, 1.2, and 1.3 times higher than Zn concentrations across the western sites for soluble Zn in snow, insoluble Zn in snow, soluble Zn in ice accretions, and insoluble Zn in ice accretions, respectively. Zinc concentrations in any form in the east were never more than twice those in the west.

Sphalerite from Olkusz (southern Poland) contained isotopically heavier Zn (δ66Zn of 0.02‰, Table 3) than Pribram sphalerite (central Czech Republic; δ66Zn of −0.23 to −0.07‰). Zinc in tetrahedrite from Slovakia was also isotopically light (−0.25‰). In contrast, hemimorphite, a secondary low-temperature mineral, contained isotopically heavy Zn (0.62‰). Fly ash sampled near a base-metal smelter in Pribram released isotopically light Zn (−0.47‰) into the atmosphere. Soft coal from Sokolov (northwest) also contained isotopically light Zn (−0.17‰). The overall range of δ66Zn values of potential pollution sources was 1.10‰ (Table 3). 3.5. Zn isotopes in snow The range of δ66Zn values of the soluble snow fraction was 1.28‰ (from −0.60 to 0.68‰; solid bars in Fig. 4). Across the sites, the mean δ66Zn of the soluble snow fraction was −0.01 ± 0.05‰. δ66Zn values in snow on average greater than 0‰ were observed at two sites (0.16‰ at VRH, and 0.11‰ at KAP). The lowest mean δ66Zn values were observed at LOU and PRA (− 0.13 and − 0.12, respectively). Mean δ66Zn at POM and ELK were close to zero (−0.02, and −0.01‰, respectively). At POM, we observed one of the most negative δ66Zn

3.2. Zn fluxes via atmospheric input Zinc deposition rates (Fig. 2c) were calculated from the available data after some simplifications had been made. We took advantage of similar elevations of all our study sites (close to 1000 m), and assumed that ice accretions at such elevations contributed 10% of snowpack water (Hindman et al., 1983). Total precipitation at each site was known (Table 1). We further assumed that the winter-time precipitation represented 45% of the annual total precipitation. This estimate was based on long-term hydrological monitoring of small upland catchments in the Czech Republic (data for 1994–2012; Oulehle et al., 2013b). The Zn deposition rates were calculated by multiplying concentrations of soluble and

Fig. 3. Deposition flux of Sc, a conservative geogenic trace element.

P. Voldrichova et al. / Chemical Geology 388 (2014) 130–141 Table 3 δ66Zn values for selected pollutants (ore minerals, fly ash) in Central Europe. Site

Sample type

δ66Zn (‰)

Olkusz, Poland Pribram, Czech Republic

Sphalerite Dark sphalerite Red sphalerite Hemimorphite Tetrahedrite + chalcopyrite Fly ash Soft coal

0.02 −0.23 −0.07 0.62 −0.25 −0.47 −0.17

Pribram, Czech Republic Maria Bana, Slovakia Pribram, Czech Republic Sokolov, Czech Republic

values in the data set (−0.45‰). All δ66Zn data for atmospheric deposition are given in Table S2. In three randomly selected snow samples, we also determined the δ66Zn values of the insoluble fraction using HF digests (Table 4). In all three cases, the insoluble Zn fraction had lower δ66Zn values than the soluble Zn fraction, on average by 0.22‰.

3.6. Zn isotopes in ice accretions δ66Zn values of the soluble fraction of ice accretions were between −0.67 and 0.14‰, averaging −0.28 ± 0.09‰ (open bars in Fig. 4). In the eastern Czech Republic (sites POM, VRH and ELK), δ66Zn of ice accretions was always lower than the mean δ66Zn of snow. In contrast, at LOU and PRA in the western Czech Republic, the mean δ66Zn of ice accretions was higher than the mean δ66Zn of snow. On two occasions, a snow and ice accretion sample was taken simultaneously: At VRH (east), δ66Zn of ice accretion was lower by 0.39‰ than that of snow. At PRA (west), δ66Zn of snow was lower by 0.30‰ than that of ice accretions. Plots of δ66Zn values versus 1/Zn concentration can be useful for identification of mixing two isotopically distinct pollution sources, or Zn fractionation effects. Fig. 5 gives such a plot for the soluble Zn fraction in ice accretions. Data from all six sites plotted close to a straight line (R2 = 0.47; p b 0.05), with one PRA sample as an outlier. Polluted sites with higher Zn concentrations exhibited relatively low δ66Zn values of ice accretions.

3.7. Inter-annual comparisons for snow For the soluble snow fraction, we plotted δ66Zn values versus 1/Zn concentration separately for each year (Fig. 6). Across the sites, no simple relationship was found between the two variables. None of the individual sites exhibited positive or negative trends between δ66Zn values and 1/Zn concentrations. Instead, some sites in Fig. 6 were characterized by distinct clusters of data points. One example was ELK in 2010. In some cases, nearly constant Zn concentrations were accompanied by changing δ66Zn values (KAP in 2009). In other cases, nearly constant δ66Zn values were accompanied by changing Zn concentrations (VRH in 2009). At most sites, Zn concentrations and δ66Zn values in snow were different in 2009 and 2010. In 2009, clusters of data in Fig. 6 seemed to be clearly separated for some sites, e.g., PRA and KAP. This distinction was non-existent in the following year. In 2010, Zn concentration and isotope characteristics at PRA and KAP overlapped.

3.8. Effect of distance from pollution sources on δ66Zn Soil dust particles were analyzed for Zn isotope ratios at a distance of 1 km from a smelter in Pribram, and 1–5 km from a smelter in Olkusz (Fig. 7). At both sites, δ66Zn of primary ore minerals was close to 0‰. At both sites, soil Zn was isotopically lighter. With an increasing distance from the smelter, δ66Zn of soil dust decreased. At Olkusz, this decrease was relatively large, close to −1.50‰.

135

3.9. Contrasting air-mass trajectories in year 1 and 2 The prevailing air-mass back-trajectories at PRA and VRH differed in year 1 and year 2 (Fig. 8). At PRA, pollutants were brought from the north and northwest in year 1, and from the west and south in year 2. In year 1, soluble Zn concentrations averaged 8 μg L − 1 , whereas in year 2, these Zn concentrations were only 3 μg L− 1. At VRH, pollutants were brought from the northwest in year 1, and from the southeast in year 2. In year 1, soluble Zn concentrations averaged 7 μg L − 1 , while in year 2, the average Zn concentration was much higher (22 μg L− 1).

4. Discussion 4.1. Comparison between Zn, Sc and water fluxes Typically, industrial sources do not release Sc into the environment. Most air-borne Sc is of geogenic origin. By comparing Figs. 2c and 3, it appears that elevated Zn deposition fluxes at the eastern sites VRH and ELK (compared to LOU, PRA and POM) were anthropogenic: they cannot be explained by elevated dust inputs, which would simultaneously bring geogenic Sc and geogenic Zn. Since our study sites are characterized by low pollution levels, we do not use Sc, or any other geogenic element, to calculate Zn enrichment factors (EF). According to Reimann and de Caritat (2005), using EFs to detect human influence on element cycles in remote areas with low pollution levels should be avoided. Local influences, such as chemical heterogeneity of geogenic dust and bigeochemical cycling, make the non-anthropogenic background too variable. As shown by the lack of a positive correlation between Zn deposition and precipitation totals, water fluxes (and elevation) were not the main control of the Zn input into the ecosystems in our study.

4.2. Gradients in atmospheric Zn pollution In Fig. 2c, the two eastern-most sites, ELK and VRH, show the highest Zn depositions in 2009–2010. The studied area was affected by the proximity of industrial centres in three countries (Fig. 1). Prior to 1989, these countries (Czechoslovakia, East Germany and Poland) had centrally planned economies and belonged to the worst polluters in Europe (Novak et al., 2005). Annual emission rates of many pollutants peaked between 1985 and 1995. Atmospheric Zn loads in those years are not known. We can, however, compare our data in Table 2 with Zn concentrations in small forested catchments of the GEOMON network (Fottova, 2011). Open-area (bulk) precipitation was collected monthly at LIZ near LOU in the south, at JEZ near PRA in the northwest, at MOD near POM in the northeast, and at CER near ELK in the east. The oldest Zn concentration data for vertical deposition in these catchments were collected in 1995. In 1995, Zn concentrations in rain and snow averaged 22, 18, 44 and 24 μg L−1 at LIZ, JEZ, MOD and CER, respectively. From 1995 to 2010, Zn concentrations in vertical deposition across the sites decreased on average by 50%. The largest decrease was observed in the area of POM (86%), and the smallest decrease in the area of ELK (21%). These data confirm the previously mentioned shift in maximum pollution in the Czech Republic to the east. In our study, soluble Zn concentrations in snow (11 μg Zn L− 1) were lower compared to rainfall data from the Mediterranean (28 μg Zn L− 1; Luck and Ben Othman, 2002), and similar to rainfall data from tropical Africa (11 μg Zn L − 1 ; Freydier et al., 1998). Soluble Zn concentrations in snow in the Czech mountains were six times lower than those in snow in southern Quebec (Simonetti et al., 2000). In northern Quebec, soluble Zn concentrations in snow were three times lower than those in the Czech mountains (Simonetti et al., 2000).

136

P. Voldrichova et al. / Chemical Geology 388 (2014) 130–141

Fig. 4. Isotope composition of water-soluble Zn in snow (solid bars) and ice accretions (open bars). Histograms of δ66Zn values for the six study sites. Down arrows mark δ66Zn value of major ore deposits in Central Europe, x signs indicate mean δ66Zn values at individual sites. Solid dots above bars indicate the same sampling dates for snow and ice accretions.

4.3. Zn enrichment in ice accretions relative to snow It is unclear whether in pre-industrial times horizontal deposition exhibited higher concentrations of trace elements than vertical deposition. Hence, it is difficult to ascribe elevated Zn concentrations in ice

accretions, compared to snow, solely to anthropogenic air pollution. In Central Europe, previous studies reported faster decreases in pollutant concentrations in ice accretions than in snow. For example, at a mountain-top site located between POM and VRH, Dousova et al. (2007) observed a decrease twice as fast in arsenic (As) concentrations

P. Voldrichova et al. / Chemical Geology 388 (2014) 130–141

137

Table 4 Comparison of δ66Zn between the soluble and insoluble portions of vertical atmospheric deposition in winter. δ66Zn of the insoluble portion was measured following HF digestion of particles N 0.45 μm. Site

Sampling date

Age of snow (days)

δ66Zn (‰) soluble

δ66Zn (‰) insoluble

VRH VRH VRH

March 1, 2009 March 17, 2009 April 3, 2009

4 4 5

0.06 0.07 0.12

−0.13 0.04 −0.22

in ice accretions compared to snow over a period of 20 years. Arsenic concentrations were still three times higher in ice accretions than in snow at the later date (Dousova et al., 2007). Our data in Table 2 showed on average five times higher Zn concentrations in ice accretions than in snow 20 years after the period of peak industrial pollution. The Zn enrichment in ice accretions likely resulted from different mechanisms of formation of horizontal and vertical deposition. Ice accretions form from supercooled droplets originating in the basal cloud layer, rich in pollutants. Ice accretions are captured by high surface area objects on the Earth's surface. In contrast, snow forms when water vapor in a cloud freezes on condensation nuclei. The diameter of droplets forming horizontal deposition (fog, ice accretions) is smaller than the diameter of droplets forming vertical deposition (rain, snow). The total surface area per unit water mass is higher for ice accretions than for snow. Consequently, ice accretions are a more efficient scavenger of airborne contaminants than snow (Moldan and Cerny, 1994, and references therein). Figs. S7 and S8 show SEM images of insoluble particles in snow and ice accretions in the northeast of the Czech Republic, close to POM. Angular-shaped particles, up to 0.1 mm in diameter, predominate in snow, whereas smaller isometric-to-spheroidal particles (b0.01 mm in diameter) abound in ice accretions. These images illustrate the different characters of particles in vertical and horizontal precipitation in Central Europe. 4.4. Predominance of soluble Zn in atmospheric input

Fig. 6. Contrasting Zn isotope and concentration characteristics of each study site in 2009 and 2010.

The proportion of soluble and insoluble Zn species is related to sample pH. We did not measure pH values for individual samples of snow and ice accretions in this study. The reason was that we adopted a methodology used to melt ice cores from polar regions (Krachler et al., 2005; Zheng et al., 2007). To prevent irreversible precipitation of trace metals on the surface of the containers, we acidified each sample immediately after the melting started (see Section 2.4). However, we do have information on the pH of present-day atmospheric deposition from other

studies. Across monitored Czech catchments, snow in 2010 had a pH of 5.08 ± 0.1 (mean ± SE; data from LIZ, JEZ, MOD and CER; Fottova, 2011). The mean pH of horizontal deposition was between 4.0 and 5.0 (Fisak et al., 2009), just slightly lower, compared to snow (for further references, see Bohdalkova et al., 2012). Natural waters contain Zn in a predominantly dissolved form at pH b 5.8 (Majer et al., 2012). Samples analyzed for Zn concentrations and isotope ratios in this study still fell in this high Zn-solubility field. It is not surprising that on average 94% of all Zn in atmospheric deposition was present in the soluble form. Wherever pollutant Zn was released to the atmosphere in a solid phase, the size of the dust particles was small enough for almost complete dissolution

Fig. 5. Relationship between δ66Zn and 1/Zn concentration for ice accretions.

Fig. 7. Decreasing δ66Zn values of soil zinc with an increasing distance from a smelter.

138

P. Voldrichova et al. / Chemical Geology 388 (2014) 130–141

Fig. 8. Back-trajectories of air masses, preceding snow samplings at PRA and VRH. The two years of observations differed in the source regions of pollution. More Zn was brought to PRA from the north (Leipzig, Saxony) in year 1, compared to year 2 (Bavaria, West Germany). More Zn was brought to VRH in year 2 from the southeast (the industrial conurbation of Ostrava), compared to year 1 (Lubin, northwest).

of Zn in water droplets prior to freezing. Hence, it is important to include dissolved Zn species in isotopic studies. 4.5. Isotopic tracing of sources vs. processes In contrast to some riverine systems where Zn behaves conservatively (Chen et al., 2008), Zn isotopes in the atmosphere do fractionate (Juilliot et al., 2011, and references therein). At a receptor site, source apportionment may be hampered by isotope effects that occurred in the atmosphere. Isotopic tracing of Zn sources in the atmosphere of Central Europe can only be successful if the isotope difference between individual pollution sources is larger than the isotope effects during Zn release by anthropogenic sources and its transport to receptor sites (see Sections 4.7 and 4.8). 4.6. The role of geogenic Zn A synthesis of existing environmental δ66Zn data has indicated that, on average, anthropogenic Zn sources tend to contain isotopically lighter Zn than geogenic Zn sources (Pokrovsky et al., 2005; Borrok et al., 2009; Cloquet et al., 2006, 2008). Across our study sites, the average δ66Zn of the prevailing soluble Zn species in ice accretions was lower than that in snow (− 0.28 vs. − 0.01‰). It follows that, in general, ice accretions contained more anthropogenic Zn, whereas snow may have contained a larger proportion of geogenic Zn. Thus far, the relative distance of individual Zn sources for each receptor site cannot be determined. Whole-rock δ66Zn values for the main geological units in Central Europe have not been determined. 4.7. Comparison between δ66Zn of pollution sources and deposition In all ore districts in Fig. 1, primary minerals are processed. Therefore, in our interpretations, we will leave out the secondary mineral hemimorphite, typical of the oxidation zone. We will also disregard the Lubin base metal district in Poland, where the production of Cu and Ag predominates, with Zn present only as a minor constituent. In Pribram, the volume of dark sphalerite (Table 3) exceeds that of red-colored sphalerite by more than 10 times. Therefore, we will consider the low δ66Zn of the dark sphalerite (− 0.20‰) to be representative of bulk Pribram Zn isotope composition. This value is lower compared to most other Zn ore districts (0.16 ± 0.2‰; Sonke et al., 2008). As seen in Fig. 4, δ66Zn of snow at the three western sites (KAP, LOU, and PRA) was scattered around the Pribram sphalerite δ66Zn. Pribram Zn may have influenced δ66Zn snow at these three sites. PRA, for example,

may be influenced by winds from the south, i.e., from the region around Pribram (Fig. 8 left). Because the scatter at KAP, LOU and PRA in Fig. 4 was relatively broad, further unknown sources and/or processes may have affected the local δ66Zn values. Olkusz sphalerite (δ66Zn of 0‰) may have influenced mainly the two eastern-most study sites (VRH and ELK). Indeed, as seen in Fig. 4, δ66Zn values of snow at VRH and ELK were scattered around δ66Zn of the Olkusz sphalerite. Contribution of Zn-rich dust from the mining areas, resulting from low-temperature processes, such as ore crushing and transportation, and emissions of sphalerite particles from smelter stacks, to the Zn atmospheric input are likely. Again, the relatively broad scatter of δ66Zn at these receptor sites may suggest the contribution of additional Zn sources, unrelated to metal smelting. Literature data (Cloquet et al., 2006; Sivry et al., 2008; Sonke et al., 2008; Bigalke et al., 2010), and our analysis of Pribram fly ash, agree that Zn that was volatilized in smelters at a high temperature and then condensed in the atmosphere is isotopically light. δ66Zn of Pribram fly ash was − 0.47‰ (Table 3). Particles containing condensed hightemperature Zn are generally believed to be smaller than those originating from mechanical processes related to ore crushing. Horizontal deposition usually scavenges smaller particles, whereas larger particles tend to be vertically deposited by gravity (see Figs. S7 and S8). Hence, Zn in ice accretions influenced by ore smelting should be isotopically light. Zn isotope data for ice accretions (open bars, Fig. 4) were indeed low, but only for the three eastern sites (POM, VRH and ELK). At these sites, Zn in ice accretions is best explained as a product of ore smelting. δ66Zn of ice accretions at the two western sites LOU and PRA were higher than δ66Zn of snow. We conclude that anthropogenic Zn processed at high temperatures did not dominate horizontal deposition at LOU and PRA. Soft coal in the northern Czech Republic (the Sokolov basin) contained isotopically light Zn (− 0.17‰). Dust resulting from openpit coal mining could influence the nearby site PRA. Snow at PRA contained Zn with an average isotope composition of −0.15‰, a value similar to that of local coal dust. Previous studies reported isotopically heavy Zn in dust related to coal processing (0.70‰; Sivry et al., 2008). If such particles dominated at the receptor sites, we would expect high δ66Zn values in atmospheric deposition. As seen in Fig. 4, high values (0.80‰) in snow were observed only at VRH in the east. Slightly less positive δ66Zn values of snow (0.50‰) were observed at KAP, which, however, is located in the south, remote from the coal mining regions. Unfortunately, we were not able to obtain δ66Zn values of fly ash derived from coal incineration. In general, we would expect that Zn that was subject to high-temperature volatilization in coal-burning power

P. Voldrichova et al. / Chemical Geology 388 (2014) 130–141

plants would result in isotopically light Zn condensate, just like in the case of ore smelting. Such data, however, have not been reported. We note that some precipitation samples in Fig. 4 had δ66Zn values below or above those of the potential pollution sources. This may result either from Zn isotope fractionations, or from our incomplete knowledge of the variability of the isotope signatures of individual Zn sources. Pokorna et al. (2013) demonstrated that, in the Czech Republic, changing wind direction typically brings pollutants from several different directions in less than 24 h. 4.8. The role of Zn sources other than metallurgy and coal burning Overall, δ66Zn values of snow and ice accretions in Fig. 4 could have been influenced by many different Zn sources, most of which remain isotopically poorly defined. For example, we do not know the average Zn isotope signature of pollution resulting from recycling of Zn previously used in industries and households. Various agricultural Zn sources, such as fertilizers and pesticides, are an unlikely source of Zn at our study sites. A recent study using Zn isotopes has shown that agriculture-related Zn with its relatively high δ66Zn is almost quantitatively immobilized in soils (Chen et al., 2008). In contrast, urban Zn is more mobile and may enter the hydrological cycle (Chen et al., 2008). δ66Zn signature of tire wear is close to 0‰ (Thapalia et al., 2010). Flue gas from municipal waste combustion is characterized by δ66Zn of 0 to 0.13‰ (Cloquet et al., 2006). Galvanized steel produces a δ66Zn of 0.12 to 0.58‰ (John et al., 2007). Urban rainwater is characterized by δ66Zn of 0.18 to 0.20‰ (Chen et al., 2008). All these pollution sources are isotopically similar to some data in Fig. 4, and may have contributed to Zn found in atmospheric deposition. Higher uncertainty is related to negative δ66Zn values of snow (b− 0.40‰), occasionally found at KAP, PRA and POM (Fig. 4). Roof leaching may produce a δ66Zn of − 0.70 to − 0.20‰ (Chen et al., 2008), however this value is probably site-specific. Hydrologicallycontrolled transport of this urban pollution type to remote mountain summits is unlikely. 4.9. Relationship between δ66Zn values and Zn concentrations We observed a positive relationship between δ66Zn and 1/Zn concentration for ice accretion samples (Fig. 5), but not for snow. As we have seen in Table 2, ice accretions contained five times more Zn than snow. Our data are consistent with a previous observation by Weiss et al. (2007). These authors reported a positive correlation between δ66Zn and 1/Zn concentration of environmental samples in polluted regions only. In unpolluted regions, no such relationship had been observed. 4.10. δ66Zn values in the soil Sivry et al. (2008) and Mattielli et al. (2009) showed that δ66Zn in topsoil decreases with an increasing distance from a base-metal smelter. This decrease was interpreted to be a result of a changing character of deposited Zn. Close to the pollution source, mainly high-δ66Zn particles that were released to the atmosphere as solids (sphalerite, zincite, gunningite) are deposited. At greater distances (several km from the smelter), the deposited particles contain isotopically light Zn that underwent evaporation during smelting (Juilliot et al., 2011). Our data in Fig. 7 are consistent with this interpretation. The observed decrease in soil δ66Zn from the smelter to a distance of 5 km was remarkably large (1.50‰). It is comparable to the entire range of δ66Zn values in snow and ice accretions in this study combined (1.40‰). One implication is that at mountain-top study sites, which are 20–30 times more distant from the smelters than soils in Fig. 7, smelting-derived Zn should mainly be present in the low-δ66Zn form. This is in agreement with our previous conclusion that smelting-derived Zn was responsible for the low δ66Zn of ice accretions in the east (POM, VRH, ELK). The relatively

139

high δ66Zn values of ice accretions in the west (LOU, PRA) were probably unrelated to smelting. 4.11. Distinct δ66Zn values of insoluble Zn We are not aware of any previous comparisons between δ66Zn of soluble and insoluble Zn fractions in atmospheric deposition. Data in Table 4 showed that insoluble Zn was isotopically lighter than soluble Zn. Quantitative dissolution of particulate Zn does not fractionate Zn isotopes (Albarede, 2004). Hence, the source of both forms of Zn in the same sample may be different. If the insoluble particles were mainly geogenic, while the soluble Zn was mainly anthropogenic, we would have expected isotopically heavier Zn in the particles compared to the solutes (Pokrovsky et al., 2005; Borrok et al., 2009; Cloquet et al., 2006, 2008). Such was not the case, but our data set was too small for any conclusions. 4.12. Inter-annual differences in δ66Zn at the receptor sites Dissolved Zn originating from one large point source of pollution is often deposited continuously over long distances (tens to hundreds of kilometers). It is not clear whether this Zn, when deposited, has a constant isotope composition, regardless of the distance from the pollution source. In Fig. 6, we saw different δ66Zn values for the first and the second year for most sites. At the same time, clusters of values for individual sites only partly overlapped. The simplest interpretation of these patterns would be that the differences between year 1 and year 2 reflected a change in Zn source areas over time, depending on the prevailing wind direction, and that different sites tended to receive pollution from different sources (cf., Fig. 8). We also note that, in many cases, fresh snow contained more pollution than “old” snow (Bohdalkova et al., 2012). Even on a day-to-day basis, the source region and history of air masses played a more important role than possible dilution/concentration of the pollutant, due to changing daily precipitation rates. Detailed isotope fingerprinting of various regional sources of Zn pollution would be needed for more specific source apportionment. The important implication, however, is that Zn in the Central European atmosphere is not a well-mixed reservoir. 5. Conclusions For the first time, we used δ66Zn values to study vertical and horizontal atmospheric deposition of Zn in winter. We selected six remote mountain-top locations in the Czech Republic with no local sources of pollution, and compared Zn abundance and isotope composition in snow and ice accretions. The studied region was characterized by the presence of large anthropogenic pollution sources, including metallurgy and coal industry. We found that, in 2009–2010, Zn concentration in snow and ice accretions in different parts of the Czech Republic varied by only a factor of two. This was surprising, given the sharp (10-fold) northwest–south pollution gradient, described repeatedly for many pollutants, especially in the years of maximum pollution (1985–1995). Today, the northeastern part of the Czech Republic is slightly more polluted than the northwestern and southern parts. The soluble Zn pool constituted on average 94% of total Zn. Zinc insoluble in weak acids made up a mere 6% of total Zn. Zinc concentrations in ice accretions were on average five times higher compared to those in snow, reflecting larger surface area per unit mass. For a Zn isotope study, we chose soluble Zn, which was more abundant than insoluble Zn. At the three eastern sites, δ66Zn values of ice accretions were lower than those of snow. At two western sites, the opposite was true. We ascribe Zn in ice accretions in the east to smelting, which produces isotopically light Zn vapor. We suggest that vertical and horizontal deposition may bring Zn that originated from different sources at different distances. Overall, in a δ66Zn vs. 1/Zn-concentration plot, snow from individual sites formed clusters of data with partial overlap. Better isotope

140

P. Voldrichova et al. / Chemical Geology 388 (2014) 130–141

characterization of pollution sources is needed, but the observed isotope inhomogeneities in the Central European atmosphere hold promise for a successful application of δ66Zn to source apportionment in remote ecosystems. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.chemgeo.2014.09.008. Acknowledgments This work was supported by the European Commission (SOIL TrEC; 244118). We thank Dominik Weiss and Shuofei Dong of the University of London, Imperial College, for advice. References Albarede, F., 2004. The stable isotope geochemistry of copper and zinc. In: Johnson, C.M., Beard, B.L., Albarède, F. (Eds.), Reviews in Mineralogy and Geochemistry, Geochemistry of Non-traditional Stable Isotopes vol. 55, pp. 409–427 (ISBN 1529-6466). Bigalke, M., Weyer, S., Kobza, J., Wilcke, W., 2010. Stable Cu and Zn isotope ratios as tracers of sources and transport of Cu and Zn in contaminated soil. Geochim. Cosmochim. Acta 74, 6801–6813. http://dx.doi.org/10.1016/j.gca.2010.08.044. Bohdalkova, L., Novak, M., Voldrichova, P., Prechova, E., Veselovsky, F., Erbanova, L., Krachler, M., Komarek, A., Mikova, J., 2012. Atmospheric deposition of beryllium in Central Europe: comparison of soluble and insoluble fractions in rime and snow across a pollution gradient. Sci. Total Environ. 439, 26–34. http://dx.doi.org/10. 1016/j.scitotenv.2012.08.089. Borrok, D.M., Nimick, D.A., Wanty, R.B., Ridley, W.I., 2008. Isotopic variations of dissolved copper and zinc in stream waters affected by historical mining. Geochim. Cosmochim. Acta 72, 329–344. http://dx.doi.org/10.1016/j.gca.2007.11.014. Borrok, D.M., Wanty, R.B., Ridley, W.I., Lamothe, P.J., Kimball, B.A., Verplanck, P.L., Runkel, R.L., 2009. Application of zinc isotopes to track the sources and mechanisms of metal loading in a mountain watershed. Appl. Geochem. 24, 1270–1277. http://dx.doi.org/ 10.1016/j.apgeochem.2009.03.010. Bullen, T., 2007. Chromium stable isotopes as a new tool for forensic hydrology at sites contaminated by anthropogenic chromium. WRI-12. Taylor and Francis CRC Press. Chapman, J.B., Mason, T.F.D., Weiss, D.J., Coles, B.J., Wilkinson, J.J., 2006. Chemical separation and isotopic variations of Cu and Zn from five geological reference materials. Geostand. Geoanal. Res. 30, 5–16. http://dx.doi.org/10.1111/j.1751-908X.2006. tb00907.x. Chen, J., Gaillardet, J., Louvat, P., 2008. Zinc isotopes in the Seine river waters, France: a probe of anthropogenic contamination. Environ. Sci. Technol. 42, 6494–6501. http://dx.doi.org/10.1021/es800725z. Chrastny, V., Vanek, A., Teper, L., Cabala, J., Prochazka, J., Pechar, L., Drahota, P., Penizek, V., Komarek, M., Novak, M., 2012. Geochemical position of Pb, Zn and Cd in soils near the Olkusz mine/smelter, South Poland: effects of land use, type of contamination and distance from pollution sources. Environ. Monit. Assess. 184, 2517–2536. http://dx. doi.org/10.1007/s10661-011-2135-2. Cloquet, Ch., Carignan, J., Libourel, G., 2006. Isotopic composition of Zn and Pb atmospheric depositions in an urban/periurban area of northeastern France. Environ. Sci. Technol. 40, 6594–6600. http://dx.doi.org/10.1012/es0609654. Cloquet, Ch., Carignan, J., Lehman, M.F., Vanhaecke, F., 2008. Variation in the isotopic composition of zinc in the natural environment and the use of zinc isotopes in biogeosciences: a review. Anal. Bioanal. Chem. 390, 451–463. http://dx.doi.org/10.1007/ s00216-007-1635-y. Dolgopolova, A., Weiss, D.J., Seltmann, R., Kober, B., Mason, T.F.D., Coles, B., Stanley, C.J., 2006. Use of isotope ratios to assess sources of Pb and Zn dispersed in the environment during mining and ore processing within the Orlovka–Spokoinoe mining site (Russia). Appl. Geochem. 21, 563–579. http://dx.doi.org/10.1016/j.apgeochem.2005. 12.014. Dong, S.F., Weiss, D.J., Strekopytov, S., Kreissig, K., Sun, Y.B., Baker, A.R., Formenti, P., 2013. Stable isotope ratio measurements of Cu and Zn in mineral dust (bulk and size fractions) from the Taklimakan Desert and the Sahel and in aerosols from the eastern tropical North Atlantic Ocean. Talanta 114, 103–109. http://dx.doi.org/10.1016/j. talanta.2013.03.062. Dousova, B., Erbanova, L., Novak, M., 2007. Arsenic in atmospheric deposition at the Czech–Polish border: two sampling campaigns 20 years apart. Sci. Total Environ. 387, 185–193. http://dx.doi.org/10.1016/j.scitotenv.2007.06.028. Erbanova, L., Novak, M., Fottova, D., Dousova, B., 2008. Export of arsenic from forested catchments under easing atmospheric pollution. Environ. Sci. Technol. 42 (19), 7187–7192. http://dx.doi.org/10.1021/es800467j. Fisak, J., Stoyanova, V., Chaloupecky, P., Rezacova, D., Tsacheva, T., Kupenova, T., Marinov, M., 2009. Soluble and insoluble pollutants in fog and rime water samples. Soil Water Res. 4, S123–S130. Fottova, D. (Ed.), 2011. Hydrochemical Budgets in the Monitoring Network GEOMON. Project Report SP/1a6/151/07. Czech Geological Survey, Prague. Freydier, R., Dupre, B., Lacaux, J.P., 1998. Precipitation chemistry in intertropical Africa. Atmos. Environ. 32, 749–765. Heimburger, A., Tharaud, M., Monna, F., Losno, R., Desboeufs, K., Nguyen, E.B., 2013. SLRS5 elemental concentrations of thirty-three uncertified elements deduced from SLRS5/SLRS-4 ratios. Geostand. Geoanal. Res. 37, 77–85. http://dx.doi.org/10.1/j.1751908X.2012.00185.x.

Hindman, E.E., Borys, R.D., DeMott, P.J., 1983. Hydrometerological significance of rime ice deposits in the Colorado Rockies. Bull. Am. Water Res. Assoc. 19, 619–624. http://dx. doi.org/10.1111/j.1752-1688.1983.tb02779.x. John, S.G., Park, J.G., Zhang, Z., Boyle, E.A., 2007. The isotopic composition of some common forms of anthropogenic zinc. Chem. Geol. 245, 61–69. http://dx.doi.org/10. 1016/j.chemgeo.2007.07.024. Juilliot, F., Marechal, Ch., Morin, G., Jouvin, D., Cacaly, S., Telouk, P., Benedetti, M.F., Ildefonse, P., Sutton, S., Guyot, F., Brown Jr., G.E., 2011. Contrasting isotopic signatures between anthropogenic and geogenic Zn and evidence for post-depositional fractionation processes in smelter-impacted soils from Northern France. Geochim. Cosmochim. Acta 75, 2295–2308. http://dx.doi.org/10.1016/j.gca.2011.02.004. Krachler, M., Zheng, J., Fisher, D., Shotyk, W., 2005. Analytical procedures for improved trace element detection limits in polar ice from Arctic Canada using ICP-SMS. Anal. Chim. Acta. 530, 291–298. http://dx.doi.org/10.1016/j.aca.2004.09.024. Luck, J.M., Ben Othman, D., 2002. Trace element and Pb isotope variability during rainy events in the NW Mediterranean: constraints on anthropogenic and natural sources. Chem. Geol. 182, 443–460. Luck, J.M., Ben Othman, D., Albarède, F. and Télouk, P. 1999. Pb, Zn and Cu isotopic variation and trace elements in rain. In Geochemistry of Earth's Surface, ed. Armannsson, Balkema, pp. 103, ISBN 978-80-7075-780-2. Majer, V., Hruska, J., Zoulkova, V., Holeckova, P., Myska, O., 2012. Atlas of the Chemistry of Surface Waters, Czech Republic. Czech Geological Survey, (190 pp.). Marechal, C., Albarede, F., 2002. Ion-exchange fractionation of copper and zinc isotopes. Geochim. Cosmochim. Acta 66, 1499–1509. http://dx.doi.org/10.1016/S00167037(1)00815-8. Mason, T.F.D., Weiss, D.J., Chapman, J.B., Wilkinson, J.J., Tessalina, S.G., Spiro, B., Horstwood, M.S.A., Parrish, R.R., Spratt, J., Coles, B.J., 2005. Zn and Cu isotopic variability in the Alexandrinka volcanic-hosted massive sulphide (VHMS) ore deposit, Urals, Russia. Chem. Geol. 221 (3–4), 170–187. http://dx.doi.org/10.1016/j.chemgeo.2005. 04.011. Mattielli, N., Rimetz, J., Petit, J., Perdrix, E., Deboudt, K., Flament, P., Weiss, D., 2006. Zn–Cu isotopic study and speciation of airborne metal particles within a 5-km zone of a lead zinc smelter. Geochim. Cosmochim. Acta 70, A401. http://dx.doi.org/10.1016/j.gca. 2006.06.808. Mattielli, N., Petit, J.C.J., Deboudt, K., Flament, P., Perdrix, E., Taillez, A., Rimetz-Planchon, J., Weis, D., 2009. Zn isotope study of atmospheric emissions and dry depositions within a 5 km radius of a Pb–Zn refinery. Atmos. Environ. 43, 1265–1272. http://dx.doi.org/ 10.1016/j.atmosenv.2008.11.030. Moldan, B., Cerny, J. (Eds.), 1994. Biogeochemistry of small catchments, a tool for environmental research. SCOPE Report 51. John Wiley & Sons, Chichester. ISBN: 0-471-93723-1. Novak, M., Kirchner, J.W., Fottova, D., Prechova, E., Jackova, I., Kram, P., Hruska, J., 2005. Isotopic evidence for processes of sulfur retention/release in 13 forested catchments spanning a strong pollution gradient (Czech Republic, central Europe). Global Biogeochem. Cycles 19 (4), GB4012. http://dx.doi.org/10.1029/2004GB002396. Novak, M., Mitchell, M.J., Jackova, I., Buzek, F., Schweigstillova, J., Erbanova, L., Prikryl, R., Fottova, D., 2007. Processes affecting oxygen isotope ratios of atmospheric and ecosystem sulfate in two contrasting forest catchments in Central Europe. Environ. Sci. Technol. 41, 703–709. http://dx.doi.org/10.1021/es0610028. Novak, M., Erel, Y., Zemanova, L., Bottrell, S.H., Adamova, M., 2008. A comparison of lead pollution record in Sphagnum peat with known historical Pb emission rates in the British isles and the Czech Republic. Atmos. Environ. 42 (40), 8997–9006. http://dx. doi.org/10.1016/j.atmosenv.2008.09.031. Oulehle, F., Chuman, T., Majer, V., Hruska, J., 2013a. Chemical Recovery of acidified Bohemian lakes between 1984 and 2012: the role of acid deposition and bark beetle induces forest disturbance. Biogeochemistry 116, 83–101. http://dx.doi.org/10.1007/ s10533-013-9865-x. Oulehle, F., Fottova, D., Kram, P., Hruska, J., 2013b. GEOMON — Hydrogeochemical Database. Czech Geological Survey, Prague. Pokorna, P., Hovorka, J., Krouzek, J., Hopke, P.K., 2013. Particulate matter source apportionment in a village situated in industrial region of Central Europe. J. Air Waste Manag. Assoc. 63, 1412–1421. http://dx.doi.org/10.1080/10962247.2013.825215. Pokrovsky, O.S., Viers, J., Freydier, R., 2005. Zinc stable isotope fractionation during its adsorption on oxides and hydroxides. J. Colloid Interface Sci. 291, 192–200. http://dx. doi.org/10.1016/j.scis.2005.04.079. Reimann, C., de Caritat, P., 2005. Distinguishing between natural and anthropogenic sources for elements in the environment: regional geochemical surveys versus enrichment factors. Sci. Total Environ. 337, 91–107. http://dx.doi.org/10.1016/j. scitotenv.2004.06.011. Shiel, A.E., Weis, D., Orians, K.J., 2010. Evaluation of zinc, cadmium and lead isotope fractionation during smelting and refining. Sci. Total Environ. 408, 2357–2368. http://dx. doi.org/10.1016/j.scitotenv.2010.02.016. Shotyk, W., Krachler, M., 2009. Determination of trace element concentrations in natural freshwaters: how low is “low”, and how low we do need to go? J. Environ. Monit. 11, 1747–1753. http://dx.doi.org/10.1039/b917090c. Simonetti, A., Gariepy, C., Carignan, J., 2000. Pb and Sr isotopic compositions of snowpack from Quebec, Canada: inferences on the sources and deposition budgets of atmospheric heavy metals. Geochim. Cosmochim. Acta 64 (1), 5–20. http://dx.doi.org/10. 1014/S0016-7037/99)00207-0. Sivry, Y., Riotte, J., Sonke, J.E., Audry, S., Schäfer, J., Viers, J., Blanc, G., Freydier, R., Dupre, B., 2008. Zn isotopes as tracers of anthropogenic pollution from Zn-ore smelters The Riou Mort–Lot River system. Chem. Geol. 255, 295–304. http://dx.doi.org/10.1016/j. chemgeo.2008.06.038. Sonke, J.E., Sivry, Y., Viers, J., Freydier, R., Dejonghe, L., Andre, L., Aggarwal, J.K., Fontan, F., Dupre, B., 2008. Historical variations in the isotopic composition of atmospheric zinc deposition from a zinc smelter. Chem. Geol. 252, 145–157. http://dx.doi.org/10.1016/ j.chemgeo.2008.02.006.

P. Voldrichova et al. / Chemical Geology 388 (2014) 130–141 Tanimizu, M., Asada, Y., Hirata, T., 2002. Absolute isotopic composition and atomic weight of commercial zinc using inductively coupled plasma mass spectrometry. Anal. Chem. 74, 5814–5819. http://dx.doi.org/10.1012/ac0259257. Thapalia, A., Borrok, D.M., van Metre, P.C., Musgrove, M., Landa, E.R., 2010. Zn and Cu isotopes as tracers of anthropogenic contamination in a sediment core from an urban lake. Environ. Sci. Technol. 44, 1544–1550. http://dx.doi.org/10.1012/es902933y. Viers, J., Oliva, P., Nonell, A., Gelabert, A., Sonke, J.E., Freydier, R., Gainville, R., Dupre, B., 2007. Evidence of Zn isotopic fractionation in a soil–plant system of a pristine tropical watershed (Nsimi, Cameroon). Chem. Geol. 239, 124–137. http://dx.doi.org/10.1016/ j.chemgeo.2007.01.005.

141

Weiss, D.J., Rausch, N., Mason, T.F.D., Coles, B.J., Wilkinson, J.J., Ukonmaanaho, L., Arnold, T., Nieminen, T.M., 2007. Atmospheric deposition and isotope biogeochemistry of zinc in ombrotrophic peat. Geochim. Cosmochim. Acta 71, 3498–3517. http://dx.doi.org/10. 1016/j.gca.2007.04.026. Zheng, J.C., Shotyk, W., Krachler, M., Fisher, D.A., 2007. A 15.800-year record of atmospheric lead deposition on the Devon Island Ice Cap, Nunavut, Canada: natural and anthropogenic enrichments, isotopic composition, and predominant sources. Global Biogeochem. Cycles 21 (2), GB2027. http://dx.doi.org/10.1029/2006GB002897.