Science of the Total Environment 527–528 (2015) 384–392

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Tracing metal sources in core sediments of the artificial lake An-Dong, Korea: Concentration and metal association Mansik Choi a,⁎, Jongkyu Park a, Dongjin Cho a, Dongjun Jang a, Miseon Kim a, Jongwoo Choi b a b

Department of Ocean Environmental Sciences, Chungnam National University, Daejeon 305-764, Republic of Korea Environmental Measurement & Analysis Center, National Institute of Environmental Research, Incheon 404-708, Republic of Korea

H I G H L I G H T S • • • • •

Highly enriched metals were found in environmental media near the Zn smelter. Metal concentrations in lake sediments were the highest at the core bottom (1980). Metal concentrations decreased to the background level from 1980 to 2000. Cd and Zn concentrations increased again from 2000 to the present. Ore mining and Zn smelter contributed sequentially to lake sediments.

a r t i c l e

i n f o

Article history: Received 10 December 2014 Received in revised form 29 April 2015 Accepted 4 May 2015 Available online 14 May 2015 Editor: F.M. Tack Keywords: Trace metals Zn smelter Lake sediments Soil Stream water Stream sediment

a b s t r a c t The concentration and source of trace metals in the artificial lake An-Dong, which has widespread abandoned mines and a Zn smelter upstream of the drainage basin, were investigated. Soils (18 ea), stream waters (15 ea) and sediments (15 ea) in the main channel and five tributaries downstream of the Zn smelter towards the lake (~50 km downstream) were collected. And two core sediments were also taken from the middle of the lake. All samples were analyzed for trace metals in bulk and in a 1 N HCl-leached fraction. Although the soil and stream sediments consisted mostly of sand-sized grains, concentrations of metals (Cu, Zn, Cd and Pb) were very high in all samples, including soils, stream waters and sediments at sites near the Zn smelter. However the metal concentrations decreased rapidly downstream, suggesting that the area of impact of the smelter lies within 5 km. Highly enriched metal concentrations were also found in dated core sediments from the lake; while the highest concentrations of Co, Ni, As, Cu, Zn, Cd and Pb were detected in the bottom of the sediment core (dated 1980) they decreased towards 2000, and only Cu, Zn and Cd concentrations increased again in present-day samples. Since the temporal variation in metal concentrations appeared consistent with historical variation in ore mining and Zn smelter production rates, a model combining the production rates of each was developed, which estimated 3%, 12% and 7% contributions from Zn smelter compared to ore mining production rate to levels of Cu, Cd and Zn, respectively, suggesting the different pathways by different sources. In addition, analysis of Cd/Zn and Cu/Zn ratios showed that contamination from ore mining decreased from 1980 to 2000, and smelting processes were most likely responsible for metal enrichment (Cu, Cd and Zn) from 2000 to the present. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Trace metals released into aquatic environments, whether originating from natural weathering or from human activities, may harm human health and ecosystems, through bio-accumulation and biomagnification (Oehme, 1978; Vernet, 1991). Individual countries have developed environmental conservation protocols to decrease metal concentrations in the environment, and remediation works are often ⁎ Corresponding author. E-mail address: [email protected] (M. Choi).

http://dx.doi.org/10.1016/j.scitotenv.2015.05.013 0048-9697/© 2015 Elsevier B.V. All rights reserved.

undertaken in metal-contaminated areas. Before the implementation of regulations and the onset of remediation works, investigation of the behavior and fate of metals in a target environment should be undertaken. Metals released into the environment through atmospheric and/or riverine pathways are eventually deposited in sediments within an aquatic reservoir, hence the sediment layer may preserve the history of metal contamination in the drainage basin (Monna et al., 2000; Choi et al., 2007; Cheng and Hu, 2010). Lake An-Dong is an artificial lake constructed in 1977 by damming to reserve agricultural and industrial water and to generate hydroelectric

M. Choi et al. / Science of the Total Environment 527–528 (2015) 384–392

power. It has a surface area of around 51.5 km2 and a drainage area of about 1584 km2. In the northern upstream area, there are widespread abandoned mines. These include Yonhwa mine (1961–1987 yr), Keumho mine (previously Janggun, 1976–present), and the Sukpo Zinc (Zn) smelter, producing Zn products from 1973 to the present, on the site of Yonhwa mine. Ore mining contributes mine wastes and tailings to the environment, while metal-smelting processes are inevitably accompanied by gaseous and particulate emissions, sewage waters and solid wastes (Dudka and Adriano, 1997). Soils around the mining tails and waste ore piles have been found to be contaminated with Zn and Pb (KRETC, 2005) and sediments in lake An-Dong showed higher concentrations of Cu, Zn, Cd, Pb and Hg than the effective range low (ERL) level of the National Oceanic and Atmospheric Administration (NOAA) (Park et al., 2012a). Furthermore, some fish in this lake, such as bluegill and bass, were found to harbor higher Hg concentrations than those in other rivers and lakes in Korea (Byeon et al., 2010). It is therefore necessary to identify sources of metal deposits in lake An-Dong, to prevent further contamination and develop remediation protocols. Although Pb isotope characterization in some lake sediments indicated that mining tails and ore wastes contributed to Pb contamination (Park et al., 2012b), more detailed investigation of the extent of smelter waste contamination in the lake is required. Sampling of various monitoring media such as soil, stream waters, sediments in the main stream and tributaries, and temporal variation analysis using core sediments is necessary to develop a comprehensive recovery plan. This study investigated metal concentrations and metal enrichment in soils downwind of the smelter, in stream waters and sediment downstream of the smelter but upstream of the lake, and in dated core sediments taken within the lake. In addition, the contamination contribution from the smelter as well as the mining was assessed by

385

comparing the contamination history in sediments with the production history of mining and smelting activities. 2. Samples and method 2.1. Samples A synopsis of metal concentrations throughout environmental media including soils, stream water and sediments sampled along the main stream and in five tributaries between the smelter and the lake in June 2012 is shown in Fig. 1. Two gravity core sediments were collected to investigate the temporal variation of metal contamination in July 2013 (Fig. 1). Firstly, 18 soil samples at 10 sites were collected at two depths (upper: 2–3 cm from the surface, lower: 10–15 cm). As the average wind direction in this area (An-Dong) over the past 10 years was predominantly northwest and southeast (KMA, 2003–2012), sample sites were chosen downwind of the smelter (up to 16.5 km from the smelter). The five sites closest to the smelter were located at an elevation around 50 m higher than the smelter, placing them above the top of the smelter chimney. All soil samples were largely coarsegrained sand with small amounts of fine-grained materials. The soil samples were collected in vinyl bags and delivered to a laboratory. A total of 15 stream water and sediment samples were collected from sites located approximately midway between the smelter and the lake, from both the main river channel and tributaries. Surface sediments submerged below the water level (2–3-cm depth), were collected from the channel bank by hand using a plastic scoop, and stream waters from the middle of the stream were collected by hand using a pole sampler. All stream sediments, including those from the main channel and tributaries (with the exception of sample SS 11),

R5,SS7 S1-5 Sukpo Zn smelter R1-3,SS1-4 S6 R4,SS5

Yonwha

N

SS6

S8

S7 Tri-3 S10

Tri-2 S9

R6, SS8 Tri-1

Tri-4 Keumho

R7, SS9

R8, SS10

Soil Stream water & sediments Core sediments

5km

Tri-5

CHINA

East/Japan Sea

R9, SS11 CHINA KOREA

Jujin Domok Yellow Sea

Lake An-Dong

Study area

JAPAN

Fig. 1. Sampling sites for soils (S1–S10), stream waters (R1–R9) and stream sediments (SS1–SS11) in the main channel, those in tributaries (Tri1–Tri4) and core sediments (Jujin and Domok) in lake An-Dong.

386

M. Choi et al. / Science of the Total Environment 527–528 (2015) 384–392

consisted of gravelly sand-sized materials. Sample SS 11 was composed of mostly fine-grained materials with a high organic matter content. Two core sediments of lengths 44 cm and 52 cm were collected at Jujin and Domok, respectively, located in the center of the lake, using a gravity corer. They were delivered to the laboratory, and subsamples were stored in vinyl bags after slicing with a plastic blade at 1-cm spacing. The core from Jujin was highly bioturbated throughout, and the upper 25 cm of the Domok core also showed high bioturbation and brownish-colored features, while the lower part consisted of dark-grayish colored and laminated layers. After freeze drying, gravel sized materials (such as plant roots, rock fragments) were removed from soil and stream sediments using a 2-mm grid sieve. After sieving, all solid samples were ground and homogenized in an agate mortar prior to chemical analysis. Stream waters were acidified with purified concentrated nitric acid (5 mL/L), then and filtered using a vacuum filtration system and a 0.45-μm pre-cleaned membrane filter, after preservation for 1 month at room temperature. 3. Methods Metal concentrations in both the bulk sample and a 1 N HCl-leached fraction were analyzed following Song and Choi (2009). For analysis, 0.2-g samples of powdered bulk material were digested using a mixed acid solution (HF + HNO3 + HClO4). Five milliliters of saturated boric acid and 1 mL of HClO4 were added, and the digested solution was evaporated. The evaporation residues were dissolved in 20-mL 1% HNO3. Metals in the 1 N HCl-leached fraction were extracted by shaking for 24 h in a horizontal shaker after mixing of 20-mL 1 M HCl with a 0.2-g powdered sample. After centrifugation, the supernatants were collected and diluted appropriately in 1% HNO3. Total dissolvable metals in stream waters were measured by direct injection to the instrument without further treatment. Aluminum (Al) and iron (Fe) were measured using inductively coupled plasma atomic emission spectrometry (ICP– AES, PerkinElmer Ltd.), and trace metals (Mn, Cr, Co, Ni, Cu, Zn, As, Sn, Cd, Pb) were measured using inductively coupled plasma mass spectrometry (ICP–MS, Thermo-Finnigan Ltd.) at the Korea Basic Science Institute (KBSI), Korea. The reliability of the analytical method was verified using certified sediment reference materials, MESS-3 from the National Research Council of Canada (NRCC); 83% (Ni) to 107% (Cd) recovery was obtained for the total digestion process and 15% (Cr) to 95% (Cd) recovery for the 1 N HCl extraction (Table 1). The deposition age in core samples was determined by estimating the sedimentation rate using 210Pbex measurement and the constant initial concentration model (Carroll and Lerche, 2003; San Miguel et al., 2004) assuming constant sedimentation rate and 210Pbex fluxes. Total 210Pb activity in sediment was measured by α-spectrometry, using an electro-deposited plate after total digestion at KBSI. Supported 210 Pb activity was obtained by measuring 226Ra concentrations using γTable 1 Measured and certified values of major and trace metals in the 1 N HCl leached and total digested sediment reference material, MESS-3 from NRCC (National Research Council Canada) (n = 3). Certified values

Al Fe Ca Mg Cr Ni Cu Zn As Cd Pb

% % % % mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg

Mean

Error

8.59 4.34 1.47 1.6 105 46.9 33.9 159 21.2 0.24 21.1

0.23 0.11 0.06 4 2.2 1.6 8 1.1 0.01 0.7

1 N HCl leached

Total digestion

Recovery

Mean

Stdev

Mean

Stdev

(%)

0.4 0.5 0.3 2 0.24 0.03 0.1

8.34 4.32 1.39 1.73 108 38.8 30.0 163 19.4 0.26 21.2

0.07 0.05 0.02 0.02 1 0.6 0.2 2.29 0.28 0.06 0.3

97.1 99.5 94.9 108.1 102.5 82.7 88.4 102.3 91.5 106.5 100.6

16.2 25.7 23.7 117 10.6 0.21 18.6

spectrometry at KBSI. The estimated deposition rates of core sediments were 1.53 ± 0.44 cm/yr (n = 8, r2 = 0.67) and 3.37 ± 2.70 cm/yr (n = 7, r2 = 0.24) for Domok and Jujin, respectively. Organic carbon was measured using a CHN analyzer after the dissolution of carbonates by 1 N HCl solution at the Center for Research Facilities at Chungnam National University. 4. Results and discussion 4.1. Synoptic features of metal enrichment Metal concentrations in soils, stream sediments and waters of the main channel and tributaries are summarized in Supplementary Tables 1–3. Since metal concentrations in bulk samples can vary depending on the physical and geochemical substrates, especially for sediments (Horowitz, 1991), enrichment factors relative to upper continental crust (UCC, Taylor and McLennan, 1995) were used to describe the synoptic features of metal enrichment by applying the following equation: E:F:ðMÞ ¼

ðM=AlÞsample ðM=AlÞUCC

:

Here, M and Al indicate the concentrations of the metal of interest and of aluminum (Al), respectively, and E.F.(M) is the enrichment factor of the metal relative to UCC. Since the enrichment of metals might be assessed only by comparing with the background concentration, the enrichment factor relative to UCC may under- or over-estimate the metal contamination, especially for the environmental samples within the drainage area with the metal-rich formations. Thus, of the metals investigated, the metals only with higher enrichment factor (N2.5) were presented. For soils and sediments, Cr, Cu, Zn, As, Cd and Pb were enriched; measured concentrations for these analytes in the 1 N HClleached fraction and enrichment factors for bulk samples are shown with distance from the smelter in Figs. 2 and 3, respectively. Extremely high concentrations of Cu (N200 mg/kg), Zn (N 2500 mg/kg), Cd (N20 mg/kg) and Pb (N350 mg/kg) in soils and Zn (N 2500 mg/kg), Cd (N20 mg/kg) and Pb (N79 mg/kg) in sediments were detected in the two samples closest to the smelter; these concentrations decreased rapidly with increasing distance from the smelter. Spatial variations in metal enrichment in soils and sediments coincided with concentrations measured in the 1 N HCl-leached fraction, which indicates that metals exist in environmentally labile forms and are affected by contamination sources (Jung, 2001; Lee et al., 2001). Although enrichment factors for Cr were higher than 2.5 in soils, there were few differences in the 1 N HCl leached concentrations between the upper and lower layers of the soil column. Furthermore, similar concentrations of As were detected in the 1 M leached fraction of the closest sample and in soils collected at a distance of 3–5 km from the smelter (Fig. 2). Compared to the sediments taken upstream of the smelter, concentrations and enrichment factors of Zn, Cd and Pb were the highest in the closest samples and lake sediments; in particular, extremely high enrichment factors for Zn (43fold), Cd (136-fold) and Pb (4-fold) were found in samples taken 1.6 km from the smelter (Fig. 3). The enrichment factors decreased rapidly to levels found in upstream sediments at about 2 km (Pb)–31.2 km (Zn, Cd) from the smelter. There was little difference in As enrichment between upstream and downstream sediments, while lake sediments showed enrichment factors 4–5-fold higher than those in upstream sediments. In tributaries, samples tri-1 and tri-3 showed higher enrichment factors for Pb, Cu and As, and Zn, As and Pb, respectively, than main channel sediment showing the possibility of As supply to the lake from mining waste piles in the tributaries. However, no tributaries exhibited higher enrichment factors for Cd than the main channel. Total dissolvable metal concentrations of Cu, Zn, As, Cd and Pb in the downstream sites of the main channel, especially for sites (R1–R4) close to the smelter, were considerably higher than those at the upstream site

M. Choi et al. / Science of the Total Environment 527–528 (2015) 384–392

387

Fig. 2. Metal concentrations in 1 N HCl leached fraction and enrichment factors for bulk samples in soils with distance from the Zn refinery. Upper (0–2 cm) and lower (15–20 cm) layer samples in soil column were depicted separately.

(R5); however, As concentrations in the lower reaches (R6–R9) were markedly higher than those in the upper reaches (R1–R4) (Fig. 4). High dissolved As concentrations were also detected in tri-1 and tri-3, while a high dissolved Zn concentration was found in tri-3. Cd concentrations decreased steadily towards the lake, though levels remained high in lake waters; however, Cu, Zn and Pb were detected at much higher concentrations in lake waters than in stream waters in the lower reaches. In summary, soils, waters and sediments at sites close to the smelter showed considerably higher concentrations of Cu, Zn, As, Cd and Pb than those at the remote sites (soil) and upstream sites (water and sediment). In addition, those metals in sediments and waters were enriched in the lake as well as in two tributaries (tri-1 and tri-3). Thus, it was evident that smelter effluents impacted on environmental media, but that the extent of impact seemed to be limited to within ~5 km. The tributaries, especially tri-1 and tri-3, could also contribute metals to the lake. However, the possibility of smelter impact on the lake cannot be neglected, since flooding events dump upstream materials into the lake, and metal transport, such as that seen in Cd concentrations in the stream waters, is a possibility (Fig. 4). 4.2. Temporal variations in metal concentrations and their sources The vertical distributions of Al and Fe in the Jujin core showed nearly constant values, with mean concentrations of 8.86 ± 0.54% and 4.22 ±

0.20%, respectively. Aluminum and Fe in the Domok core also showed little vertical variation, with mean concentrations of 9.37 ± 0.95% and 4.69 ± 0.48%, respectively, except for the section between 18 and 19 cm (Supplementary Fig. 1 and Table 4). However, organic carbon content in both cores showed great variability; values were high in the 8–15 cm interval (3.06–6.03%), and the lowest values at 38–43 cm, in the Jujin core (1.65–1.95%), on the meanwhile, the lowest values (1.41–1.46%) were found at 14–17 cm in the Domok core and it increased both upward and downward (Supplementary Fig. 1 and Table 4). Trace metal concentrations in sediments can vary depending upon physical properties (e.g., grain size, surface area) and chemical constituents (e.g., Fe & Mn oxyhydroxides, clay minerals, organic matter) as well as pollutant input (Horowitz, 1991). Therefore, the relatively high Al concentrations with low variability, as well as high and variable organic matter contents in both cores suggest that organic matter rather than grain size controls metal concentrations in the core sediments. Chromium (Cr), Co, Ni, Cu, As and Pb concentrations in the Jujin core were vertically homogeneous with an average concentration of 68 mg/kg, 19 mg/kg, 34 mg/kg, 41 mg/kg, 83 mg/kg and 68 mg/kg, respectively, and less than 10% variability. However, Zn and Cd concentrations showed minimum values at depths below 35 cm and increased upward, with the highest values in the 18–19- and 22–23-cm depths. Cr concentrations in the Domok core were vertically homogeneous with a mean of 66 mg/kg, similar to the Jujin core. However, Co and Ni

388

M. Choi et al. / Science of the Total Environment 527–528 (2015) 384–392

Fig. 3. Metal concentrations in 1 N HCl leached fraction and crustal enrichment factors for bulk samples in stream sediments with distance from upstream of Zn refinery to lake An-Dong. Enrichment factors for sediments in five tributaries were also depicted.

concentrations were the highest at a depth of 36–37 cm, while As and Pb concentrations were at a maximum in the bottom sample. All metals showed nearly constant concentrations in upper samples (above 16–17 cm), similar to the Jujin core. In contrast, Cu, Zn and Cd concentrations were at their maximum in lower samples, decreased upward to a depth of 16–17 cm, but increased between this range and the top (Supplementary Fig. 1 and Table 4). Temporal variation of metal concentration by converting the depth into the deposition age is shown in Fig. 5. The bottom layer of the Domok core was deposited in 1980, 3 years after the completion of An-Dong dam. Since the sedimentation rate of the Jujin core was estimated to be twofold that of the Domok core, the bottom layer of the Jujin core was correlated with a depth of 20–21 cm in the Domok core. That is, the lowest concentrations of Zn, As, Cd and Pb at 14–21 cm in the Domok core coincided with the bottom two layers of the Jujin core. In addition, concentrations of metals in the Jujin core were similar to those in the upper part (0–21 cm) of the Domok core. The pattern of variation of enrichment factors for all metals was correlated with the temporal variation in metal concentrations, which may be related to the low vertical variability of sediment properties, such as grain size. The most striking feature was that metal concentrations were the highest in 1980 (Cu, Zn, As, Cd, Pb) or 1989 (Co, Ni), decreased to minimum values during 2000–2003 and remained constant (Co, Ni, As, Pb) or increased again (Cu, Zn, Cd) until the present. Furthermore, peak concentrations of Zn and Cd were in 2006 and 2008 (Jujin), and

2007 and 2010 (Domok). That is, contamination with five (Cu, Zn, As, Cd and Pb) metals in lake sediments was at a maximum in 1980, whereas the maximum Co and Ni levels were detected in 1989 and decreased until 2002. After that time, only three metals (Cu, Cd, Zn) were enriched in the lake sediments. In addition, the present enrichment of metal concentrations compared to the minimum value in 2002 was in the order Cd N Zn N Cu, compared to Zn Cd N Cu in early contamination (1980–1985). Thus, it may be hypothesized that different sources of metal contamination, contributing different metal species, could be responsible for contamination of lake sediments in the past and at present, and/or that the transport pathway to enrich metals in lake sediments may differ between those sources. Lastly, the strong correlation (r N 0.8 for Cu, Zn, Cd and Pb) between organic carbon and metal concentrations for the sediments deposited before 2002 in the Domok core and the lack of correlation between these parameters in the Jujin and upper Domok core sediments further suggest that different enrichment processes occurred before and after 2002. Possible sources of metals in this area are limited since it is mountainous and the human population is less than 1000 households; hence, mining and smelting activities are most likely the main contributors although the weathering of metal-enriched formations might be considered. In addition, ore wastes and mining waste accumulate after the abandonment of mines, if these were not properly managed, they would present another source of metal contamination of the lake through stream or groundwater pathways (Mighanetara et al., 2009;

M. Choi et al. / Science of the Total Environment 527–528 (2015) 384–392

389

Fig. 4. Total dissolvable metal concentrations of stream waters in the main channel and dissolved metals in tributaries along the distance from Zn refinery to the lake.

Park et al., 2012a, 2012b). Mining of Zn and Pb ores in this area has been undertaken since the 1960s, reached maximum production in 1977, and has been negligible since 2000 (Fig. 6a; KIGAM, 1994–2011). The scope of mining activities includes exploitation, ore dressing including crushing, grinding and washing to elevate the ore grade from lowgrade (2–3%) to concentrate (about 50%), and mining tailings; production rate is therefore directly connected to discharge of metal contaminated particles into the environment, especially through stream water (Dudka and Adriano, 1997). In addition, the aerial oxidation of sulfide ores acidifies waters and increases leaching of metals from ore/rock, potentially contributing to metal contamination of stream waters (Kuma, 2011; Behrooz and Borden, 2012; Soltani et al., 2014). The smelter produces nearly pure metals through a roasting and electrolysis process that involves leaching, purification, electrolysis, melting and casting; production of refined Zn from the smelter increased gradually from initiation (1970) to the present, with a large increase during the period 2003–2004 (Fig. 6a; KIGAM, 1994–2011). Although it is operated under regulatory control, the smelter may contribute metals to the environment through discharge of wastewaters into the main channel, emission into the atmosphere, and waste sludge piling; the rate of contamination should be proportional to smelter production rate. The decrease in Co, As and Ni concentrations in core sediments from 1980, and subsequent steady concentrations after 2000, correlate with the rate of Pb ore production (Figs. 5 and 6a). Cu, Zn, Cd and Pb

concentrations varied with Zn ore production before 2002. However, since the increase in Cu, Zn and Cd concentrations after 2002 did not correspond to Zn and Pb ore mining, and increasing metal discharge from mining waste pilings was not expected, some other sources; e.g., the smelter, may be responsible for the increase in levels of these metals. The production rate of Zn products in the smelter increased rapidly from 2004, thus, we hypothesize that ore mining and smelting activities contributed to environmental contamination sequentially. Ore mining was responsible for metal enrichment in lake sediments before 2004, which resulted in metal enrichment in the lower part of the core; after this date smelting was the primary source of metal enrichment, as seen in the upper sediments. The specific impact of these two sources on lake sediments would be different, however, and their combined impact was modeled to explain the temporal variation in metal concentrations in the lake cores. If the contribution of metals to lake sediments is assumed to be described by the combination of production rates of mining ores (combined Zn and Pb ores) and smelter products, their total effective production rate can be described by considering their relative production rates, and then comparing this to metal concentrations in core sediments (Fig. 6b–d). For Cu, Zn and Cd, the relative contributions of the smelter as a proportion of the mining production rate were estimated at 3%, 7% and 12%, respectively, by comparison of the metal concentration profile of core sediments with annual combined production rates. Although some temporal peaks in

390

M. Choi et al. / Science of the Total Environment 527–528 (2015) 384–392

Fig. 5. Temporal variation of metal concentrations in the An-Dong lake core sediments. Deposition age was calculated using the depth and sedimentation rates (Jujin: 3.37 cm/yr, Domok: 1.53 cm/yr).

production rate were not reflected in metal concentrations, this was likely due to large errors in the estimation of sedimentation rate; overall, the temporal variation in metal concentrations in core sediments correlated well with combined mining and smelter production rates. However, the relative contribution of smelter production rate was estimated differently for each metal, and hence some explanation of the inconsistency between metals is required. Furthermore, even after 2000, when mining activities ceased and smelter production was limited, high metal concentrations with enrichment factors relative to UCC ranging from 1 to 2 (Co, Ni, Cu), 2–4 (Pb, Zn) and 20–40 (Cd, As) were found (Fig. 5). However, there are no data on background concentrations of metals in this area, hence the contribution from the smelter should be considered an additional input to background concentrations, which may be highly enriched. In addition, some large peaks (two samples per core) in Cu, Zn and Cd concentrations measured in sediments deposited after 2000 should be noted (Fig. 5). Some of these can be attributed to abrupt discharge of smelterrelated materials by summer flooding (Foulds et al., 2014); only three metals showed large peaks, their ratios (Cd/Zn, Cu/Zn) were similar in sediments found both above and below the period of operation of the smelter, and differed significantly from sediments affected by mining activities (e.g., 1980–1990 sediments). This may exclude the possibility of discharge of ore tailings during flood events. In addition, high metal enrichment factors found in sediments with relatively lower metal concentrations suggest that background metal concentrations in this area were originally high, due to natural weathering of metal-enriched formation, or continuous discharge of poorly managed ore tailings piled near the abandoned mines (Byeon et al., 2010; Park et al., 2012a,

2012b). Although they could not be differentiated in this study, it is possible that background metal concentrations would be relatively high even if they originated from natural weathering, since some metals (Co, Ni, As, Pb) showed nearly constant concentrations and enrichment factors after 2000. Lastly, since contamination by Co, Ni, As, Cu, Zn, Cd and Pb in lake sediments before 2000 can be explained by the temporal variation of mining activities, increases in the concentrations of only three metals (Cu, Zn, and Cd) after 2000 indicate that the smelter contributed to enrichment in lake sediments during this period. However, metal ratios (Cu/Zn and Cd/Zn) showed significantly different values before and after 2000. Comparing Cd/Zn and Cu/Zn ratios among lake sediments, stream waters and stream sediments from sites with the highest concentrations of metals close to the Zn smelter (e.g., R2 and R3 for stream waters, and SS2 and SS3 for stream sediments), two linear mixing trends with three metal sources are observed, as shown in Fig. 7. Samples deposited from 1980 to 2002, and also from 2000 to 2013, are illustrated. The end-members from the former mixing line may be attributed to ore mining and background, while those in the latter arise from smelting and background. Metal ratios (Cu/Zn and Cd/Zn) in both ore mining samples (4.90 ± 0.74 and 0.85 ± 0.03 for samples deposited 1980–1990, n = 8, respectively) and background (11.6 ± 0.57 and 1.00 ± 0.14 for samples deposited 1998–2002, n = 9, respectively) were well constrained and significantly differentiated, indicating preferential release in the order Zn N Cd N Cu from ore mining compared to background. In contrast, although the data comparing smelting to background seemed to be relatively scattered, presumably due to variable source characteristics, a significant linear trend (r = 0.73, p b 0.01) was established, resulting from controlled Zn release

M. Choi et al. / Science of the Total Environment 527–528 (2015) 384–392

391

Fig. 6. Annual production of ore mining (based on 50% purity) and smelter in the study area (a; KIGAM). Modeled contribution using combining ore mining and smelting production rates was fitted to temporal variation of metal concentration in lake sediments; Zn (b), Cd (c) and Cu (d).

and preferential release of Cd from Zn smelter waste, in combination with higher mobilities of Cd (Urszula and Edeltrauda, 2009; Eléonor et al., 2014). This can be seen in the high Cd/Zn ratio of stream waters.

Thus, preferential enrichment of Cd relative to Zn would be important in discriminating Zn smelter-related materials from ore mining wastes. In addition, temporal variability of controlled wastewater treatment at the smelter would cause variation in the metal ratios, but maintain high Cd/Zn and low Cu/Zn ratios. However, Cd/Zn ratios in stream sediments containing very high metal concentrations were similar to those in the background, though absolute concentrations in the background were much lower. Thus, the assumption that smelter-related materials cause very high concentrations of Cd, Zn, Cu and Pb in stream sediments close to the Zn smelter may be erroneous. Other evidence, such as the correlation between the temporal variability of metal enrichment in lake sediments and ore mining/Zn smelter production rates, different metal association patterns across different periods, and preferential enrichment of Cd relative to Zn for smelter-related materials could explain the temporal variability of metal concentrations in lake sediments. Therefore, other approaches such as metal isotope analysis (Cd, Zn, Pb) using the isotopic fractionation found in metallurgical processes (Gao et al., 2008; Shiel et al., 2010, 2012) and differentiation of the origin of Pb ores among domestic and imported ores (Mukai et al., 1993; Park et al., 2012b; Chae et al., 2014) should be employed to confirm the origin and transport pathways of metals. 5. Conclusion

Fig. 7. Mixing relationships using Cd/Zn and Cu/Zn ratios for core sediments, and stream waters and sediments in sites (R2 and R3 for waters and SS2 and SS3 for sediments) with the highest concentrations of metals near the Zn smelter.

This study investigated the synoptic features of metal enrichment in a range of environmental media – such as soil, stream waters, stream and lake sediments – and temporal variation of metal enrichment in lake sediments in a region of widespread abandoned mines and active

392

M. Choi et al. / Science of the Total Environment 527–528 (2015) 384–392

Zn smelter activity. High metal enrichment (Cu, Zn, Cd and Pb) was detected in all samples including soils, stream waters and sediments collected close to the Zn smelter, which dissipated for samples over 5 km downstream from the smelter. However, lake water and sediments also showed high concentrations of those metals as well as As. Considering the synoptic features of metal enrichment in the environmental media, it may be concluded that the effect of smelter production is limited in area and that mining piles near abandoned mines are the main source of metals in lake water and sediments. However, the temporal variation of metal enrichment established from two dated core sediments was consistent with the historical variation of ore mining and Zn smelter production rates. Furthermore, temporal variations in metal ratios implied a change in the sources of environmental metal concentrations. Thus, a model was developed to explain the temporal variation in metal enrichment. The model considered the combination of ore mining and Zn smelter production rates, which resulted in 3%, 12% and 7% relative contribution from Zn smelter compared to ore mining production rate for Cu, Cd and Zn, respectively. Based on Cd/Zn and Cu/Zn ratio pair diagrams, metal contamination from ore-mining sources decreased from 1980 to 2000, and increases in metal enrichment (Cu, Cd and Zn) are a result of smelting processes from 2000 to the present. However, some uncertainty arises from the assumption of the end-members of smelter-related materials; therefore, more direct evidence is required for the source of environmental metals, such as that obtained by analysis of stable metal isotopes and differentiation of the composition of domestic and imported ores (Pb). Acknowledgments This study was supported from the project of “Tracing metal sources in aquatic environments using stable metal isotopes” funded by the National Institute of Environmental Research. The authors thank to anonymous reviewers and editor for their constructive suggestions. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.scitotenv.2015.05.013. References Behrooz, M., Borden, R.C., 2012. Physical, hydrologic, and aqueous chemical characterization of the ore Knob tailings pile (Ashe County, North Carolina, USA). Mine Water Environ. 31, 3–15. Byeon, M., Lee, J., Park, J., Shin, S., Han, J., Kim, Y., 2010. Study on mercury concentrations of freshwater fish from Lake An-dong and its upper stream. Anal. Sci. Technol. 23 (5), 492–497 (In Korean with English abstract). Carroll, J., Lerche, I., 2003. Sedimentary Processes: Quantification Using Radionuclides. Elsevier 978-0-08-044300-3. Chae, J.S., Choi, M.S., Song, Y.H., Um, I.K., Kim, J.G., 2014. Source identification of heavy metal contamination using metal association and Pb isotopes in Ulsan Bay sediments, East Sea, Korea. Mar. Pollut. Bull. 88, 373–382. Cheng, H., Hu, Y., 2010. Lead (Pb) isotopic fingerprinting and its applications in lead pollution studies in China: a review. Environ. Pollut. 158, 1134–1146. Choi, M.S., Yi, H.-I., Yang, S.Y., Lee, C.-B., Cha, H.-J., 2007. Identification of Pb sources in Yellow Sea sediments using stable Pb isotope ratios. Mar. Chem. 107 (2), 255–274.

Dudka, S., Adriano, D.C., 1997. Environmental impacts of metal ore mining and processing: a review. J. Environ. Qual. 26, 590–602. Eléonore, R., Corinne, C., Rémi, F., Laurent, D., Jérôme, V., Françoise, E.-P., 2014. Persisting impact of historical mining activity to metal (Pb, Zn, Cd, Tl, Hg) and metalloid (As, Sb) enrichment in sediments of the Gardon River, Southern France. Sci. Total Environ. 481, 509–521. Foulds, S.A., Brewer, P.A., Macklin, M.G., Haresign, W., Betson, R.E., Rassner, S.M.E., 2014. Flood-related contamination in catchments affected by historical metal mining: an unexpected and emerging hazard of climate change. Sci. Total Environ. 476–477, 165–180. Gao, B., Liu, Y., Sun, K., Liang, X., Peng, P., Sheng, G., Fu, J., 2008. Precise determination of cadmium and lead isotopic compositions in river sediments. Anal. Chim. Acta 612, 114–120. Horowitz, A.J., 1991. A Primer on Sediment-Trace Element Chemistry. 2nd ed. CRC Press (144 pp.). Jung, M.C., 2001. Heavy metal contamination of soils and waters in and around the Imcheon Au–Ag mine, Korea. Appl. Geochem. 16, 1369–1375. KIGAM, 1994–2011. Balance Status of Mineral Products. Korea Institute of Geosciences and Mineral Resources (http://mici.kigam.re.kr, In Korean). KMA, 2003–2012. Annual Climatological Report. Korea Meteorological Administration, Seoul, Korea. KRETC (Kyungpook Regional Environmental Technology Center), 2005. The Environmental Assessment and Proper Management of Downriver District Caused by Closed and Temporarily-Closed Mines in Upriver District of Nakdong River — The Research of Environmental Assessment of Mineral-Mixed Deposits in the Riverbed of Upriver District. Kuma, J.S., 2011. Numerical indices of the severity of acidic mine drainage: broadening the applicability of the Gray acid mine drainage index. Mine Water Environ. 30, 67–74. Lee, C.G., Chon, H.-T., Jung, M.C., 2001. Heavy metal contamination in the vicinity of the Daduk Au–Ag–Pb–Zn mine in Korea. Appl. Geochem. 16, 1377–1386. Mighanetara, K., Braungardt, C.B., Rieuwerts, J.S., Azizi, F., 2009. Contaminant fluxes from point and diffuse sources from abandoned mines in the River Tamar catchment, UK. J. Geochem. Explor. 100, 116–124. Monna, F., Clauer, N., Toulkeridis, T., Lancelot, J.R., 2000. Influence of anthropogenic activity on the lead isotope signature of Thau Lake sediments (southern France): origin and temporal evolution. Appl. Geochem. 15, 1291–1305. Mukai, H., Furuta, N., Fujii, T., Ambe, Y., Sakamoto, K., Hashimoto, T., 1993. Characterization of sources of lead in the urban air of the Asia using ratios of stable lead isotopes. Environ. Sci. Technol. 27, 1347–1356. Oehme, F.W., 1978. Toxicity of Heavy Metals in the Environment. 0-8247-6718-7 (970 pp.). Park, J.-J., Kim, K.J., Yoo, S.M., Kim, E.H., Seok, K.S., Huh, I.A., Kim, Y.H., 2012a. Distribution of heavy metals and mercury in sediment from the lake An-dong. Anal. Sci. Technol. 25 (6), 441–446 (In Korean with English abstract). Park, J.-J., Kim, K.J., Yoo, S.M., Kim, E.H., Seok, K.S., Shin, H.S., Kim, Y.H., 2012b. Tracking lead contamination sources of sediments in Lake Andong using lead isotopes. Anal. Sci. Technol. 25 (6), 429–434 (In Korean with English abstract). San Miguel, E.G., Bolìvar, J.P., Garcìa-Tenorio, 2004. Vertical distribution of Th-isotope ratios, 210Pb, 226Ra and 137Cs in sediment cores from an estuary affected by anthropogenic releases. Sci. Total Environ. 318, 143–157. 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. Shiel, A.E., Weis, D., Orians, K.J., 2012. Tracing cadmium, zinc and lead sources in bivalves from the coasts of western Canada and the USA using isotopes. Geochim. Cosmochim. Acta 76, 175–190. Soltani, N., Moore, F., Keshavarzi, B., Sharifi, R., 2014. Geochemistry of trace metals and rare earth elements in stream water, stream sediments and acid mine drainage from Darrehzar Copper Mine, Kerman, Iran. Water Qual. Expo. Health 6, 97–114. Song, Y.H., Choi, M.S., 2009. REE geochemistry of fine-grained sediments from major rivers around the Yellow Sea. Chem. Geol. 266 (3–4), 337–351. Taylor, S.R., McLennan, S.M., 1995. The geochemical evolution of the continental crust. Rev. Geophys. 33, 241–265. Urszula, A.-K., Edeltrauda, H.-R., 2009. Contaminated sediments as a potential source of Zn, Pb, and Cd for a river system in the historical metalliferous ore mining and smelting industry area of South Poland. J. Soils Sediments 9, 13–22. Vernet, J.P., 1991. Heavy Metals in the Environment. 0-444-89064-5 (405 pp.).