Science of the Total Environment 527–528 (2015) 335–343 Contents lists available at ScienceDirect Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv Long-term assessment of ecological risk from deposition of elemental pollutants in the vicinity of the industrial area of Puchuncaví-Ventanas, central Chile Soroush Salmanighabeshi a, M.Rosario Palomo-Marín a, Elena Bernalte a, Fernando Rueda-Holgado a, Conrado Miró-Rodríguez b, Ximena Fadic-Ruiz c, Víctor Vidal-Cortez c, Francisco Cereceda-Balic c, Eduardo Pinilla-Gil a,⁎ a b c Departamento de Química Analítica, Universidad de Extremadura, Avda. de Elvas, s/n, E-06006 Badajoz, Spain Departamento de Física Aplicada, Avda. de la Universidad s/n, E-10071 Cáceres, Spain Centro de Tecnologías Ambientales, Universidad Técnica Federico Santa María, Av. de España, 1680 Valparaíso, Chile H I G H L I G H T S • • • • 5 years monitoring campaigns on Chilean soils impacted by industrial activities Trace element profiles, comparison with impacted areas and soil quality standards Temporal evolution and source assignment by cluster analysis and PCA Ecological risk assessment indexes calculated and critically discussed a r t i c l e i n f o Article history: Received 7 December 2014 Received in revised form 27 March 2015 Accepted 4 May 2015 Available online xxxx Editor: D. Barcelo Keywords: Soil Metal Ecological risk assessment Industrial pollution Pollution indexes a b s t r a c t The present work investigates soil pollution by elemental contaminants and compares ecological risk indexes related to industrial activities for the case study of Puchuncaví-Ventanas: a relevant industrial zone located in central Chile. Selected elements (As, Pb, Cd, Ni, Hg, V, Mn, Zn, Sr, Sb, Cr, Co, Cu, K, and Ba) were analyzed during a long-term period (yearly sampling campaigns during 2007–2011), at 5 sampling stations representing different degrees of impact. PCA and cluster analysis allowed identifying a copper smelter and a coal-fired power plant complex as major pollution sources. Geoaccumulation index (Igeo), enrichment factor (EF), contamination factor (Cf), contamination degree (Cdeg), and integrated pollution index (IPI) are critically discussed for quantitative ecological risk assessment. Igeo, EF and Cf indexes are producing comparable environmental information, showing moderate to high pollution risks in the area that demands further monitoring and adoption of prevention and remediation measures. Capsule: Long term assessment of elemental pollution around an industrial area. New insight on ecological risk indexes for trace element pollution in soils, by critical comparison among them. © 2015 Elsevier B.V. All rights reserved. 1. Introduction Among the most significant soil pollutants, trace elements are relevant due to both acute and long-time toxic effects in the soils and related environmental media (Yaylah-Abanuz, 2011), since soil pollutants can enter the human body by ingestion, dermal contact, and inhalation, aside from the food chain pathway starting from plant uptake (MezaMontenegro et al., 2012). The potential ecological risk associated to soil contamination is a very controversial issue in recent years (MezaMontenegro et al., 2012), so ecological risk assessment aims to provide ⁎ Corresponding author. E-mail address: epinilla@unex.es (E. Pinilla-Gil). http://dx.doi.org/10.1016/j.scitotenv.2015.05.010 0048-9697/© 2015 Elsevier B.V. All rights reserved. information to measure and predict soil pollutant threats for humans and the environmental health (Fairbrother et al., 2007; Wei and Yang, 2010). Different indexes have been proposed to predict the environmental quality of soil and sediments (Caeiro et al., 2005). Punchuncaví-Ventanas, one of the main industrial areas of Chile, comprises a wide range of industrial factories and activities implying potential risks to human and environmental health. The most environmentally relevant factories in this area are the CODELCO División Ventanas copper refinery complex and the AES Gener coal-fired power plant complex (three operating units) facilities. For example, these activities are responsible for 68.1% (copper refinery) and 30.7% (coal-fired power plants) of SO2 emissions in the area, according to the local authority's reports (UNTEC, 2012). Aside from these main 336 S. Salmanighabeshi et al. / Science of the Total Environment 527–528 (2015) 335–343 sources, there is a wide set of less impacting industrial factories operating in the area, which are briefly described as follows. Puerto Ventanas is a provider of dock and port facilities. Cementos Melón is a producer of national construction materials which has three activity areas: cement, concrete aggregate, and mortars. Catamutún Import, sell coal, and steam division, which manages industrial systems from steam coal combustion. Panimex Química S.A. produces plastic and fumaric acid. Gasmar is a company dedicated to marketing of liquefied gas. Minera Montecarmelo is a plant for treating industrial wastes. Oxiquim is a maritime terminal. Cordex is a big factory for production of asphalt and is a fuel terminal also. Studies about environmental impact of industrial activities on the elemental soil levels in the Puchuncaví-Ventanas area are scarce. Ginocchio (2000) analyzed the effect of copper smelter on grassland in terms of physicochemical soil characteristics, plant species diversity and abundance, founding a significant impact of industrial activities on plant species regeneration capabilities. De Gregori et al. (2002) carried out the redox speciation of selenium present at ultratrace levels in rainwater collected at the area. The same group (De Gregori et al., 2003) conducted a work aimed to the monitoring of copper, arsenic and antimony levels in agricultural soils impacted and non-impacted by mining activities, from three regions in Chile including the Puchuncaví-Ventanas industrial area. The high concentrations measured in impacted soils from Puchuncaví-Ventanas (300 mg kg− 1 Cu, 34.5 mg kg−1 As and 5.3 mg kg−1 Sb) clearly showed the impact produced in this zone by the industrial and mining activities developed in their proximities. Ginocchio et al. (2004) reported Cu, Zn, Pb and Cd concentrations (among other parameters) in different soil layers around the area. The levels found in the 0–5 cm layer were 361.6 mg kg−1 for Cu, 157.8 mg kg−1 for Zn, 79.9 mg kg−1 for Pb and 0.8 mg kg−1 for Cd. Copper mobility in soil around the copper smelter was investigated by Neaman et al. (2009). The same author reported results about the effectiveness of lime and compost of in situ immobilization of trace elements in soil by using earthworms as bioindicators of toxicity (Neaman et al., 2009). These previous works in the area focused on a limited numbers of trace elements analyzed after single sampling campaigns. A more detailed evaluation of trace elements in soils, focused on the study of their distribution in particle size fractions has been recently published by Parra et al. (2014). In the present work, we describe the results of a systematic, long-term investigation about ecological risk from multi-element soil pollution at the Puchuncaví-Ventanas area, comprising spatial and temporal variability, source assignment and ecological risk index calculations. 2. Materials and methods 2.1. Sampling area and soil characteristics The Puchuncaví-Ventanas industrial area is located in a district (Fig. 1) belonging to the V Region of Chile, Valparaiso Province. The area is located in the Chilean mainland coast (34° 45′ S, 71° 29′ W), 58 km North from Valparaíso (Regional Capital Region V), 45 km North from Viña del Mar and 160 km North-West from Santiago de Chile. It comprises an area of 301 km2, with a population of 13,000. The main communication ways are the F-30E Road, Highway 5 North (via Catapilco), and Nogales-Puchuncaví way. Specific locations for soil sampling were selected in the area as depicted in Fig. 1. La Greda (LG) sampling point (32° 44′ 57″ S, 71° 28′ 30″ W) is located 1.69 km NE from the main emission sources in the area, whereas Los Maitenes (LM) sampling point (32° 45′ 41″ S, 71° 27′ 18″ W) is located 2.39 km E. These two locations are expected to be the most impacted points. Puchuncaví village (PU) sampling point (32° 43′ 17″ S, 71° 24′ 43″ W) is located 8 km NE from the industrial area and Valle Alegre (VA) sampling sites (32° 48′ 30″ S, 71° 26′ 10″ W) is 6.72 km SE around the Fig. 1. Punchuncaví-Ventanas industrial area and soil sampling points. 337 S. Salmanighabeshi et al. / Science of the Total Environment 527–528 (2015) 335–343 Table 1 Analytical results of certified reference material (2710a, Montana Soil). Elements Certified values (mg/kg) Measured concentration (mg/kg) Mean recovery (%) As* Cd Pb* V Mn* Co Cu* Zn* Sr Ti* Sb Ba 0.154 ± 0.010 12.3 ± 0.3 0.552 ± 0.003 82 ± 9 0.214 ± 0.006 5.99 ± 0.14 0.342 ± 0.005 0.418 ± 0.015 255 ± 7 0.311 ± 0.007 52.5 ± 1.6 792 ± 36 0.166 ± 0.039 12.4 ± 3.0 0.563 ± 0.160 78 ± 9 0.229 ± 0.033 6.52 ± 1.52 0.370 ± 0.101 0.492 ± 0.087 221 ± 3 0.353 ± 0.003 44.0 ± 0.4 736 ± 86 90 84 80 103 99 90 86 105 88 114 84 91 Data are in mg/kg except for * that are in %. industrial complex. Prevalent SW winds in the zone are expected to transfer atmospheric pollutants mainly to LG, LM and PU sampling points. Maitencillo (MA) sampling area (32° 36′ 5″ S, 71° 25′ 56″ W), 18 km N from the industrial area, was selected as a reference rural area, presumably not affected by the industrial activities. The sampling points were located in undisturbed soil areas surrounding small villages. The soils at the study area are classified as entisols. The topsoil (0–30 cm) is sandy loam (clay 13–18%, sand 65–74% and silt 13–17%), a coarse texture that implies low nutrient availability and limited water holding capacity (Ginocchio, 2000; Neaman et al., 2009). Soil organic matter ranges from 1.0 to 1.8, and soil pH ranges from 4.6 to 5.5 (Ginocchio, 2000; Neaman et al., 2009). 2.2. Reagents All chemicals used for the preparation of stock and standard solutions were of analytical grade. ICP multielemental standard solutions were obtained from PerkinElmer (Waltham, MA, USA). Working solutions were prepared by dilution with ultrapure water (resistivity N 12 MΩ) obtained from an Ultramatic system (Wasserlab, Spain). Dilute standards and real samples were adjusted to desired pH with sub-boiled HNO3 (69%) obtained from a quartz sub-boiling system (Kürner, Rosenheim, Germany). 1 g L− 1 Y(III) and In(III) standards (Parneac, Spain) were used as internal standards. HClO4 (70%) Suprapur® from Merck and HF (48%) Panreac Hiperpur were used for sample digestion and pH adjustment. Standard Reference Material 2710a Montana I Soil® from the U.S. National Institute of Standards and Technology (NIST) was used for accuracy testing. 2.3. Soil sampling and homogenization A total of 121 surface soil samples were collected in the study area during five monitoring campaigns carried out in winter 2007, 2008, 2009, 2010 and 2011. These activities are in the frame of a large scale and long term ongoing environmental monitoring program in the area. Five samples were taken at randomly selected points around each designed sampling zone for each sampling campaign. For sampling, 1 m2 surface was delimited and the extraneous matter (stones, leafs, seeds or roots) was eliminated. Soil samples were collected to a depth of 5–10 cm by using a hand polypropylene drill. Approximately 3 kg of the samples was extracted with a plastic spade and transferred to a conditioned plastic container. The containers were immediately closed and transferred to the lab. According to ISO 11464, soil samples were appropriately dried in a stove by heating at 50 °C for 72 h and then sieved through a polypropylene 2 mm-mesh. Soil sub-samples were mechanically homogenized in a planetary mill and manually sieved through 0.2 mm using a stainless steel mesh. The samples were then stored in the fridge (4 °C) until the analysis. Table 2 Element concentrations in the soil collected at the sampling stations within the Puchuncaví-Ventanas industrial area. La Greda (LG), Los Maitenes (LM), Puchuncaví village (PU), Valle Alegre (VA) and reference station at Maitencillo (MA). All results are in mg/kg. n is the number of samples. Elements As Pb Cd Ni Hg V Mn Zn Sr Sb Cr Co Cu K Ba Max–min Mean Max–min Mean Max–min Mean Max–min Mean Max–min Mean Max–min Mean Max–min Mean Max–min Mean max–min Mean Max–min Mean Max–min Mean Max–min Mean Max–min Mean Max–min Mean Max–min Mean LG (n = 26) LM (n = 26) VA (n = 21) PU (n = 25) MA (n = 21) 93.26–22.8 50.84 165.84–24.56 84.8 2.83–0.52 1.34 92.62–3.62 20.02 3.81–0.0001 0.77 147.9–29.77 86.36 721.48–181.44 408.13 540.16–41.3 232.26 446.31–171.33 273.52 9.83–1.08 4.84 35.25–13.23 24.32 11.34–2.53 7.45 2872.75–452.77 1403.19 12769.35–1603.85 2499.11 464.93–57.14 301.88 110.56–9.56 56.98 210.27–16.8 89.06 2.84–0.45 1.3 85.12–2.43 16.76 1.38–0.0001 0.31 160.94–43.31 115.43 721.5–392.62 497.65 309.69–80 157.04 312.22–132.86 218.98 15.99–0.48 6.49 36.72–17.61 26.83 11.06–5.76 7.96 1786.77–237.19 771.29 14667.99–4381.55 9647.39 463.83–216.78 339.91 15.16–8.59 11.64 35.46–11.12 19.84 0.56–0.14 0.24 18.27–0.04 10.19 0.49–0.0001 0.08 185.55–76.28 135.72 1178.97–573.97 772.47 170.31–44.73 90.94 216.67–119.21 164.53 2.14–0.54 1.32 36.14–18.03 27.2 20.58–7.78 12.28 173.82–50.56 93.35 14109.01–3754.94 9725.81 429.06–280.5 343.36 49.48–8.7 27.51 94.09–6.41 45.57 1.46–0.05 0.55 79.42–1.54 14.17 1.34–0.0001 0.2 127.41–50.06 97.47 1014.12–325.6 569.19 161.93–33.28 100.43 218.86–142.27 175.97 7.28–0.39 3.33 50.76–16.25 33.38 14.4–7.04 9.08 629.91–40.37 284.07 10814.64–3512.23 7473.35 464.46–218.37 302.09 36.09–6.63 15.63 85.03–16.85 18.93 0.33–0.17 0.25 13.78–0.04 10.22 0.05–0.0001 0.01 170.57–94 137.91 864.91–576.77 711.06 254.6–63.11 123.2 301.43–222.92 242.25 1.84–0.68 1.32 42.2–28.5 34.72 11.26–8.61 9.58 51.74–32.22 42.34 11392.72–3768.36 8277.52 355.56–223.65 266.5 Hg results from Bernalte et al. (2014)are the same set of samples. 338 S. Salmanighabeshi et al. / Science of the Total Environment 527–528 (2015) 335–343 Table 3 Trace element concentration (mg/kg) in soil samples from Punchuncaví-Ventanas area compared with other case studies of polluted areas around the world, and reference soil quality standard guidelines. Copper smelting areas Coal-fired power plant areas Reference soil quality standards Huainan, Afsin-Elbistan, Canada (Canadian Elements Puchuncaví-Ventanas Port Kembla, Nsw, Copperbelt China (Tang Turkey (Çayır Council of Ministers soils (this study) Australia (Martley province, Zambia of the Environment, (Ettler et al., 2011) et al., 2013) et al., 2012) et al., 2004) 2007) Australia (Australian Department of Environment and Conservation, 2010) Netherlands (Ministry of Housing, Spatial Planning and the Environment, 2000) As Pb Cd Ni Hg V Mn Zn Sr Sb Cr Co Cu K Ba 20 1500 3 60 – 50 – 200 – – – – 100 – 300 55 530 12 210 10 – – 720 – 15 380 240 190 – 625 38.12 62.00qq 0.89 15.56 0.36 107.48 551.04 148.39 210.81 4.14 27.91 9.03 669.37 8539.64 320.91 4.1 29 – – – – – 63 – – 12 – 76 – – 3.1 15.3 – – – – – 34.4 – – – 56.8 1501 – – 12.77 33.79 0.64 26.14 0.01 53.14 384.43 39.95 – – 32.4 – 21.06 – – 2.4. Soil analysis The digestion protocol for the soil samples was based on a previously reported method (Palomo Marín et al., 2011). Briefly, 50 mg of soil or reference material samples were placed in Teflon digestion vessels (Savillex, USA). 2.5 mL HNO3 and 5 mL HF were added and the mix was left to react for some minutes. The vessels were then closed and heated to 90 °C in the stove for 8 h. After cooling, the vessels were opened, 2.5 mL HClO4 was added and the solution was evaporated to dryness on a plate at 200 °C. 1 mL HNO3 was then added and the solution was again evaporated to dryness. The samples were finally taken with 2.5 mL HNO3 and water to a total volume of 50 ml. 50 μL of a 10 mg L−1 In (III) and Y (III) solution was used as internal standards. Montana Soil Reference material (NIST code 2710a) was used to assess the accuracy of the experimental results. 2.5. Determination of elements by ICP-MS The digested samples were assayed on a PerkinElmer ELAN 9000 ICP-MS equipment by a standard protocol. The main instrumental parameters were as follows: RF power 1000 W, Ar plasma flow rate 1 L min−1, washing time 35 s, and 3 replicates per sample. Quantification was performed by internal standard calibration using PerkinElmer multi-element ICP-MS calibration standards. Blank samples were assayed and no significant concentrations of the studied elements were found. Concentrations of elements in the soil samples are expressed in dry soil weight terms. 26.7 6.29 89.8 – – – 89.2 – – 70.5 – 59.5 – – – 600 22 50 – 130 – 360 – – 87 – 91 – 2000 – Geoaccumulation index (Igeo): It allows the estimation of contamination comparing preindustrial and recent metal concentrations in soils. It was originally proposed by Müller (1969) for sediments and then modified by Loska et al. (2004) for soil pollution estimation. It has been widely applied to several trace metal studies in Europe (Yaylah-Abanuz, 2011).This index is expressed in Eq. (1): Igeo ¼ log2 ðCn =1:5 Bn Þ ð1Þ where Cn is the measured concentration of element n in the soil sample, and Bn is the geochemical background value of element n in the upper Earth's crust (Wedepohl, 1995). The constant value 1.5 is included to consider the natural fluctuation of the concentration of a given substance. Igeo values are categorized in seven groups as follows: practically uncontaminated (Igeo ≤ 0); uncontaminated to moderately contaminated (0 b Igeo b 1); moderately contaminated (1 b Igeo b 2); moderately to heavily contaminated (2 b Igeo b 3); heavily contaminated (3 b Igeo b 4); heavily to extremely contaminated (4 b Igeo b 5) and extremely contaminated (Igeo N 5). – Enrichment factor (EF): This method is based on the standardization of an element concentration tested against a reference element (Yaylah-Abanuz, 2011). Fe, Sc, Ti, Al, Ca, or Mn is generally used as reference elements for calculation of EF (Quevauviller, 1989). Mn was selected as a reference element in this study. 2.6. Statistical analysis Principal Component Analysis (PCA) and Hierarchical Cluster Analysis (HCA) were performed by using SPSS 19.0 software and XLSTAT 2009.1.02 for Windows Excel. Hierarchical cluster analysis was performed by using Euclidean distance and the Ward agglomerative algorithm. To get a better insight into the latent structure of the data, the principal component extracted correlation matrix was subjected to a varimax orthogonal rotation. 2.7. Assessment of potential ecological risk In this study, the following indexes have been used as indicators for assessment of soil ecological risk: Fig. 2. Normalized concentration of trace and major elements during 2007 to 2011 sampling campaigns (refer to text for the normalization protocol used). S. Salmanighabeshi et al. / Science of the Total Environment 527–528 (2015) 335–343 339 Fig. 4. Indexes of geoaccumulation (Igeo) for the element assayed. The contamination factor (Cf) is calculated in Eq. (3): C f ¼ Cn ðsampleÞ=Cn ðbackgroundÞ Fig. 3. Cluster analysis. The value of the enrichment factor was calculated according to Eq. (2): EF ¼ ½Cn ðsampleÞ=Cre f ðsampleފ=½Bn ðbackgroundÞ=Bre f ðbackgroundފ ð2Þ where Cn (sample) is the concentration of element n in the sample, Cref (sample) is the concentration of the reference element (Mn) in the sample, Bn (background) is the concentration of element n in the upper Earth Crust (Wedepohl, 1995), and Bref (background) is the concentration of the reference element in the earth crust. Enrichment factor values are grouped in five categories (Sutherland, 2000): Deficiency to minimal enrichment (EF b 2); Moderate enrichment (2 b EF b 5); Significant enrichment (5 b EF b 20); Very high enrichment (20 b EF b 40) and extremely high enrichment (EF N 40). – Contamination factor (Cf), contamination degree (Cdeg), and integrated pollution index (IPI): these are related indexes. Cf was proposed by Håkanson (1980), as the ratio between the concentration of selected pollutants (PCBs, Hg, Cd, As, Cu, Cr, Zn, and Pb) in sediment samples and reference preindustrial concentration values. It has been later adapted to the study of soil pollution by using upper crust elemental concentrations (Wedepohl, 1995) as reference values. In our case, PCBs were not measured and for this reason we've replaced this parameter by Sb. Median lethal dose (LD50) of Sb (WHO, 2003) is in the same range of PCBs (Euro Chlor, 2002), so it can be assumed that both of them have a similar eco-toxic effect. Håkanson (1980) divided the Cf into four categories: Low contamination (Cf b 1); Moderate contamination (1 ≤ Cf b 3); Considerable contamination (3 ≤ Cf b 6) and very high contamination (Cf N 6). Contamination degree (Cdeg) is defined as the sum of the individual contamination factors in Eq. (4): Cdeg ¼ X Cf Factor 1 Factor 2 Factor 3 As Pb Cd Ni Hg V Mn Zn Sr Sb Cr Co Cu K Ba % Var 0.946 0.974 0.953 0.061 0.478 −0.049 −0.465 0.816 0.190 0.934 0.185 −0.212 0.903 −0.098 0.131 40 0.025 0.001 −0.124 −0.057 0.199 0.688 0.540 −0.034 −0.680 0.080 0.803 0.645 −0.214 0.053 0.116 22 0.008 −0.051 −0.085 −0.143 0.201 0.427 0.457 0.141 −0.012 −0.038 −0.058 0.568 −0.053 0.833 0.852 8 ð4Þ Cdeg is divided into four groups as Cf (Håkanson, 1980): Low contamination (Cdeg b 8); Moderate contamination (8 ≤ Cdeg b 16); Considerable contamination (16 ≤ Cdeg b 32) and very high contamination (Cdeg N 32). The integrated pollution index (IPI) is defined as the average of the contamination factors computed for each sample (Yaylah-Abanuz, 2011). IPI values are divided in four categories (Wei and Yang, 2010): Low pollution level (IPI ≤ 1); Moderate pollution level (1 b IPI ≤ 2); High pollution level (2 b IPI ≤ 5) and extremely high pollution level (IPI N 5). 3. Results and discussion 3.1. Analytical quality control NIST 2710a (Montana Soil) certified standard reference samples were digested and assayed using the analytical protocol for the real samples, to ensure the reliability of the results, and the results obtained Table 4 Matrix of loads for the principal component analysis (PCA). Significant values are in bold. Elements ð3Þ Fig. 5. Enrichment factors (EF) of the elements assayed. 340 S. Salmanighabeshi et al. / Science of the Total Environment 527–528 (2015) 335–343 Fig. 6. Correlation between EF and Igeo. are presented in Table 1. The percentage of recovery was between 80 and 120% for all the elements, showing that the analytical methodology is properly operating for the analysis of real soil samples. 3.2. Soil contamination levels Mean, maximum and minimum concentrations of elements in the soil samples collected at the studied sampling stations are shown in Table 2, and all individual results are included in the Supplementary material (Table S1). The variability of the elemental concentrations within the study area (including the sampling locations LG, LM, PU, and VA) was somewhat high. It has been reported that the dispersal of contaminants around smelters is highly dependent on the local situation, mainly on the prevailing wind direction (Ettler et al., 2011), so SW dominant winds in the study area have to be also considered for the interpretation of pollutants transfers. The global mean values including all samples were calculated as indicatives of the overall pollution status of the area (Table 3). The soil elemental concentration levels in the study area were first inspected by comparing with soil quality standards. There is not a general regulation about soil environmental quality applicable in Chile so we have selected soil quality standards values published by the Australian government (Australian Department of Environment and Conservation, 2010), Canada (Canadian Council of Ministers of the Environment, 2007) and The Netherlands (Ministry of Housing, Spatial Planning and the Environment, 2000) as reference values for assessment of the soil elemental levels found in this study (Table 3). About the samples collected at the most impacted areas of La Greda (LG) and Los Maitenes (LM), close to the pollution sources at the Puchuncaví-Ventanas area, the most significant finding is probably the very high concentration of copper found in the soils of the study area compared with Canadian, Australian and Dutch standards. The average Cu level at LG and LM is 7 and 4 times higher, respectively, than the least restrictive Dutch value. This result can be derived from the environmental impact of the large copper smelter operating in the area. As exceeds the Australian standard values but are in the range of standard values approved in Netherlands. Vanadium levels in LG and LM are between approved levels in Canada and Australia. Barium is in the range of Australian standard values but less than Canadian and Dutch reference values. Pb, Cd, Ni, Hg, Zn, Sb, Cr, and Co levels measured in LG and LM don't exceed the international soil quality standard values selected for comparison. So the sampling areas LG and LM can be preliminarily classified as significantly impacted by industrial activities, at least for some of the elements studied (especially copper). About the samples collected at the Table 5 Contamination factors and contamination degrees of metals of soil samples collected within the study area (Low Concentration = LC, Moderate Concentration = MC, Considerable Concentration = CC, Very High Concentration = VH). LG LM VA PU MA Value Contamination Value Contamination Value Contamination Value Contamination Value Contamination Cf (As) Cf (Pb) Cf (Cd) Cf (Hg) Cf (Zn) Cf (Sb) Cf (Cr) Cf (Cu) Cdeg 25.42 VH 28.49 VH 6.62 VH 13.75 VH 6.02 VH 4.99 CC 5.24 CC 1.38 MC 2.68 CC 0.87 LC 13.12 VH 12.72 VH 2.83 MC 5.39 CC 2.00 MC 14.76 VH 6.24 VH 1.60 MC 4.88 CC 2.38 MC 4.47 CC 3.02 MC 1.82 MC 1.93 MC 1.82 MC 15.62 VH 20.95 VH 5.2 CC 10.75 VH 2.88 MC 0.7 LC 0.77 LC 0.79 LC 0.95 LC 0.73 LC 98.13 VH 53.94 VH 7.82 VH 19.87 VH 3.04 CC 177.2 VH 131.36 VH 28.06 CC 60.2 VH 19.74 CC S. Salmanighabeshi et al. / Science of the Total Environment 527–528 (2015) 335–343 less impacted areas of Puchuncaví village (PU) and Valle Alegre (VA), most of the concentrations of the assayed element in the soil are under the limits established by the aforementioned international soil quality standards. Only As and V are within the quality standard ranges. About Cu, the mean value at the PU sampling point exceeds the reference values, showing the long distance impact of the copper smelter in the dominant wind direction (SW), but the VA values are in the range of Canadian and Australian standards, reflecting that this location is more isolated from wind transport from the emission sources. The reference sampling area at Maitencillo (MA) presents similar elemental soil concentrations with those measured in VA. Soil concentration of the elements in the study area was then compared with previously reported values found in comparable industrial areas around the world (Table 3), although direct comparison of the results is complicated due to the disparity of sampling conditions (sample number and sites) and digestion protocols. In our case, copper smelting and coal-firing are the two most impacting activities in the area. We have not found reported results about case studies with these industrial activities together, so we have compared our results with other studies with separated industrial activities. E.g. Nkana smelter in Zambia (Ettler et al., 2011) and Port Kembla smelter, NSW, in Australia (Martley et al., 2004) are two comparable industrial areas affected by copper smelting activities. The As mean value in the soils at our study area around Codelco copper smelter is significantly higher than those reported around Nkana and Port Kembla smelters. This effect is observed for all locations (LG, LM, PU and VA) even MA (Table 2). Higher Pb, Zn and Cr values are also observed in the soils of our study area compared with the reference smelters, whereas Cu soil level is within reported ranges. About reference coal-fired power plants, we have selected Huainan in China (Tang et al., 2013) and Afsin-Elbistan, in Turkey (Çayır et al., 2012). As, Pb, Hg, V, Mn, Zn and Cu concentration values at all sampling sites in our study area are higher than reported values in the reference coal-fired power plants areas, whereas Cd, Ni and Cr values are lower, even at the most affected sampling point at La Greda (LG). Our results were then compared with the results of previous soil monitoring campaigns conducted in the area. We have found similar mean values of As and Sb soil concentrations (38.12 mg kg− 1 and 341 Fig. 8. Integrated pollution index of the sampling locations within the study area. La Greda (LG), Los Maitenes (LM), Puchuncaví village (PU), Valle Alegre (VA) and Maitencillo (MA). 4.1 mg kg−1 respectively) than those reported by De Gregori et al. (2003) during a study of agricultural soils in the Puchuncaví-Ventanas area (34.5 and 5.3 mg kg−1 respectively). But we have found a mean Cu concentration in the soils (669.4 mg kg−1) that is more than double of the result reported by the 2003 study (300 mg kg−1). Our results for Cu are also higher than mean values of 361.6 mg kg− 1 reported by Ginocchio et al. (2004). These results may indicate a progressive enrichment of the soil due to persistent copper emissions in the area. Our study shows similar mean results for Zn (148.4 mg kg− 1), Pb (62.0 mg kg−1) and Cd (0.9 mg kg−1) soil concentrations respect to the results of Ginocchio et al. (2004): 157.8 mg kg−1 for Zn, 79.9 mg kg−1 for Pb and 0.8 mg kg−1 for Cd. The elemental soil concentration profile measured in our work is similar that the values reported by Parra et al. (2014) for the same study area, considering that the sampling timeframe is different (2007–2011 vs 2007–2009, respectively). As previously mentioned, our results were produced during a longterm study comprising yearly monitoring campaigns during the 2007–2011 period, so the results were examined to identify possible temporal evolution of the most significant elemental pollutants in the soils of the study area (Fig. 2). For a better visualization, elemental Fig. 7. Correlation between EF and Cf. 342 S. Salmanighabeshi et al. / Science of the Total Environment 527–528 (2015) 335–343 concentration results were normalized, by dividing average concentration of each element during a given year by the average concentration during the whole period (2007–2011). It's difficult to extract a general conclusion, about the overall temporal trend about elemental pollution in the soils of the study area. Most of the environmentally significant elements assayed show maximum soil concentration values in 2007 (except Hg that peaks in 2009). Then the values decrease in 2008 to peak again in 2009–2010, but reaching lower values than those measured in 2007. The values then decline in 2011 to the minimum values measured during the studied period. The observed evolution of soil concentrations of elemental pollutants can be related with improved environmental technologies adopted by the industries in recent years, resulting in less pollutant emission per production unit, combined to natural soil cleaning processes as wind resuspension and leaching by rain. Nevertheless, a permanent soil sampling and monitoring scheme seems essential to follow elemental concentration variability and confirm the observed trends in the medium and long term. 3.3. Cluster analysis and PCA In order to reveal relationships among elements to identify pollution sources, cluster and PCA multivariate statistical techniques were applied to the results. The results of the cluster analysis are shown in Fig. 3, and the results of the PCA analysis are presented in Table 4. PCA factor 1, with 40% of variance, comprises As, Pb, Cd, Hg, Zn, Sb, and Cu (bold figures in Table 4) with high loadings. Elements in this factor are also grouped in the cluster analysis (Fig. 3). The common origin of these elements is probably related to the industrial sources (Slavković et al., 2004). Our results confirm that the presence of a large copper smelter in the study area is probably a main source of these elements, as previously reported (Ginocchio et al., 2004; Parra et al., 2014). This is also supported by the fact that the concentration of these elements in the soil is inversely correlated with the distance to the main emission sources (Table 2). These factors are related to the sources of trace elements in the studied soil samples. PCA factor 2 (22% of the total variance) groups V, Mn, Cr, and Co, and these elements are also grouped by cluster analysis. The components in this factor are potentially derived from anthropogenic sources, namely traffic for V, Mn and Cr (Allen et al., 2001) and power plants for Co (ATSDR, 2004) but their ecological risk assessment indexes (see Section 3.4.2) are lower than critical values considered as a proof of significant enrichment (Yay et al., 2008). Moreover the soil levels of these elements are not so dependent of distance to the industrial pollution sources than the elements belonging to factor 1 (Table 2), indicating the influence of natural or diffuse anthropogenic sources (e.g. traffic in the area). PCA factor 3 (8% of variance) is composed of V, Mn, K and Ba; these elements being also grouped by the cluster analysis. This elemental profile suggests a common emission source related to traffic due to the presence of V and Mn (Allen et al., 2001), but V and Mn are also included in factor 2 so a mixing of sources cannot be excluded. K and V have been also assigned to a mineral–crustal fraction in a previous study (Parra et al., 2014). Sr and Ni appear together in the cluster analysis and they belong to factor 4 of the PCA (7% of the variance, data not shown), so a common origin is possible, but we have insufficient data to identify that source. 3.4.1. Geoaccumulation index (Igeo) About the index of geoaccumulation, the results of the calculations are presented in Fig. 4. Cu is the only element showing Igeo values above 5 (extreme contamination) in La Greda (LG) and Los Maitenes (LM), with Igeo values for Cu declining with distance from the copper refinery source but still high in Puchuncaví village (PU, Igeo value 3.7; heavy contamination). Moderate Cu contamination was detected in Valle Alegre (PU, Igeo value 2.3), and even in the reference area of Maitencillo (PU, Igeo value 1.0), suggesting a long-range effect of the copper smelter in the surrounding area. As, Cd, Hg and Sb show relatively high Igeo values at LG and LM (above 2, moderate to heavily polluted), with lower values at the less polluted sampling areas Puchuncaví village (PU) and Valle Alegre (VA), and mostly negative Igeo values (no contamination) at the Maitencillo (MA) reference area. Igeo values for Pb, V and Zn range from no contamination to moderate, whereas Ni, Mn, Sr, Cr and Co present negative Igeo values (no contamination). 3.4.2. Enrichment factor (EF) Enrichment factors (EF) values in the soils of the study area are shown in Fig. 5. Again, Cu is the only element showing an extremely high index value in the locations more affected by the industrial activities (113 in LG and 78 in LM). EF values for Cu decrease with the distance but they show a very high enrichment in Puchuncaví village (PU, EF value 36) and a significant enrichment in Valle Alegre (VA, EF value 15). Even the reference area soils at Maintencillo (MA) show a moderate Cu enrichment (EF 4). EF values for the rest of the studied elements are in accordance with Igeo values. We inspected the overall correlation between EF and Igeo for all the individual samples assayed, and the results for representative elements are shown in Fig. 6. The indexes showed a strong exponential correlation indicating that the environmental information provided by both of them is similar. 3.4.3. Contamination factor (Cf), contamination degree (Cdeg) and integrated pollution index (IPI) Contamination factor (Cf) values are presented in Table 5. The results at the most impacted locations of La Greda (LG) and Los Maitenes (LM) are mostly within the category of very high contamination (Cf N 6), especially for Cu and As. Cf values tend to decrease with distance to the pollution sources, but they are still indicating moderate (Cf 1–3) to considerable (Cf 3–6) pollution for some elements even at the reference area of Maitencillo. Contamination degrees (Cdeg, sum of individual Cfs) show very high but decreasing contamination in LG, LM and PU, whereas considerable contamination is assigned to Valle Alegre (VA) and Maintencillo (MA). It seems that the contamination factor and degree are more strict ecological impact indexes than Igeo or EF for the assessment of the study area. Cf shows a strong linear correlation with EF (Fig. 7) and a perfect match with Igeo, so we can conclude that the three indexes are producing comparable environmental information. Finally, the integrated pollution index (IPI, mean of Cf values), shown in Fig. 8, clearly shows the overall level of contamination in the soil sampling sites within the study area. IPIs are in the order LG N LM N PU N VA N MA. LG (IPI 20.41), PU (IPI 6.95), and LM (IPI 15.86) are classified as sites affected by extremely high level of pollution (IP N 5). VA (4.00) and MA (2.14) are classified as highly polluted sites (IPI 2–5), indicating the strict nature of the IPI (similar to Cdeg) compared to Igeo or EF. As previously discussed, SW wind direction can play an important role for the movement of pollutant from an industrial area to sampling sites since IPI level is higher in PU (8 km to the pollution sources) than in VA (6.72 km to the pollution sources). 3.4. Ecological risk assessment 4. Conclusions Ecological risk assessment derived from the presence of the investigated elements in Puchuncaví-Ventanas area soils was estimated by using the geoaccumulation index (Igeo), the enrichment factor (EF), and finally the contamination factor (Cf), contamination degree (Cdeg) and Integrated Pollution Index (IPI). A long term (2007–2011) soil monitoring campaign conducted around the industrial area of Puchuncaví-Ventanas in central Chile, based on the measurement of elemental concentration profiles, PCA and cluster statistics, and application of several quantitative risk S. Salmanighabeshi et al. / Science of the Total Environment 527–528 (2015) 335–343 assessment indexes (geoaccumulation index, enrichment factor, contamination factor, contamination degree and integrated pollution index), has revealed significant ecological impacts, more intense at locations close to the main pollution sources and under the influence of dominant winds but also noticeable at reference background locations more than 10 km away from the industrial area. A copper refinery and a coal-fired power plant complex have been identified as major pollution sources by the statistical analysis. The observed evolution of soil concentrations of elemental pollutants in the studied period, showing a decline from 2009 to 2011, can be related to improved environmental technologies adopted by the industries in recent years. A critical comparison among the ecological risk assessment indexed used showed a good correlation among them, so they give comparable environmental information. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.scitotenv.2015.05.010. Acknowledgments This work is supported by the Spanish Agency of International Cooperation for Development (AECID project A1/037813/11) the Spanish Ministry of Science and Innovation (project CTQ2014-52309-P), Gobierno de Extremadura and FEDER (GR15087). References Allen, A.G., Nemitz, E., Shi, J.P., Harrison, R.M., Greenwood, J.C., 2001. Size distributions of trace metals in atmospheric aerosols in the United Kingdom. Atmos. Environ. 35, 4581–4591. ATSDR (Agency for Toxic Substances & Disease Registry), 2004. Public health statement for cobalt. http://www.atsdr.cdc.gov/phs/phs.asp?id=371&tid=64 (Accessed June 2014). Australian Department of Environment and Conservation, 2010. Assessment level for soil, sediment and water. http://www.esdat.net/Environmental%20Standards/Australia/ WA/Assessment%20Levels%20-%202010.pdf (Accessed June 2014). Bernalte, E., Salmanighabeshi, S., Rueda-Holgado, F., Palomo-Marín, M.R., Marín-Sánchez, C., Cereceda-Balic, F., Pinilla-Gil, E., 2014. Mercury pollution assessment in soils affected by industrial emissions using miniaturized ultrasonic probe extraction and ICPMS. Int. J. Environ. Sci. Technol. http://dx.doi.org/10.1007/s13762-013-0461-3. Caeiro, S., Costa, M.H., Ramos, T.B., Fernandes, F., Silveria, N., Coimbra, A., Medeiros, G., Painho, M., 2005. Assessing heavy metal contamination in Sado Estuary sediment: an index analysis approach. Ecol. Indic. 5, 151–169. Canadian Council of Ministers of the Environment, 2007. Canadian soil quality guidelines for the protection of environmental and human health. http://www.esdat.net/ Environmental%20Standards/Canada/SOIL/rev_soil_summary_tbl_7.0_e.pdf (Accessed 11 June 2014). Çayır, A., Belivermiş, M., Kiliç, Ö., Coșkun, M., Coșkun, M., 2012. Heavy metal and radionuclide levels in soil around Afsin-Elbistan coal-fired thermal power plants, Turkey. Environ. Earth Sci. 67, 1183–1190. De Gregori, I., Lobos, M.G., Pinochet, H., 2002. Selenium and its redox speciation in rainwater from sites of Valparaiso region in Chile, impacted by mining activities of copper ores. Water Res. 36, 115–122. De Gregori, I., Fuentes, E., Rojas, M., Pinochet, H., Potin-Gautier, M., 2003. Monitoreo de los niveles de cobre, arsénico y antimonio en los suelos agrícolas afectadas y no afectadas por las actividades mineras, de tres regiones de Chile. J. Environ. Monit. 5, 287–295. Ettler, V., Mihaljevič, M., Kříbek, B., Majer, V., Šebek, O., 2011. Tracing the spatial distribution and mobility of metal/metalloid contaminants in Oxisols in the vicinity of the Nkana copper smelter, Copperbelt province, Zambia. Geoderma 164, 73–84. 343 Euro Chlor, 2002. Risk assessment for the marine environment OSPARCOM region — North Sea. PCBs (polychlorinated biphenyls), DDT (Di(para-chloro-phenyl)trichloroethane), dioxins. http://www.eurochlor.org/media/49366/8-11-4-16_ marine_ra_pcbs_ddt_dioxin.pdf (Accessed 11 June 2014). Fairbrother, A., Wenstel, R., Sappington, K., Wood, W., 2007. Framework for metals risk assessment. Ecotoxicol. Environ. Saf. 68, 145–227. Ginocchio, R., 2000. Effect of a copper smelter on a grassland community in the Punchuncaví Valley, Chile. Chemosphere 41, 15–23. Ginocchio, R., Carvallo, G., Toro, I., Bustamante, E., Silva, Y., Sepúlveda, N., 2004. Microspatial variation of soil metal pollution and plant recruitment near a copper smelter in central Chile. Environ. Pollut. 127, 343–352. Håkanson, L., 1980. An ecological risk index for aquatic pollution control. A sedimentological approach. Water Res. 14, 975–1001. Loska, K., Wiechula, D., Korus, I., 2004. Metal contamination of farming soils affected by industry. Environ. Int. 30, 159–165. Martley, E., Gulson, B.L., Pfeifer, H.R., 2004. Metal concentrations in soils around the copper smelter and surrounding industrial complex of Port Kembla, NSW, Australia. Sci. Total Environ. 325, 113–127. Meza-Montenegro, M.M., Gandolfi, A.J., Santana-Alcántar, M.E., Klmecki, W.T., AguilarApodaca, M.G., Del Río-Salas, R., De La O-Villanueva, M., Gómez, A., Mendivil-Quijada, H., Valencia, M., Merza-Figueroa, D., 2012. Metal in residential soils and cumulative risk assessment in Yaqui and Mayo agriculture valley, Northern Mexico. Sci. Total Environ. 433, 472–481. Ministry of Housing, Spatial Planning and the Environment (Netherlands), 2000. Circular on target values and intervention values for soil remediation. http://www.esdat.net/ Environmental%20Standards/Dutch/annexS_I2000Dutch%20Environmental% 20Standards.pdf (Accessed 11 June 2014). Müller, G., 1969. Index of geo-accumulation in sediments of the Rhine River. GeoJournal 2, 108–118. Neaman, A., Reyes, L., Trolard, F., Bourrié, G., Sauvé, S., 2009. Copper mobility in contaminated soils of the Puchuncaví valley, central Chile. Geoderma 150, 359–366. Palomo Marín, M.R., Pinilla Gill, E., Calvo Blázquez, L., Capelo Martinez, J.L., 2011. Determination of trace and major elemental profiles in street dust samples by fast miniaturized ultrasonic probe extraction and ICP-MS. Talanta 84, 840–845. Parra, S., Bravo, M.A., Quiroz, W., Moreno, T., Karanasiou, A., Font, O., Vidal, V., Cereceda, F., 2014. Distribution of trace elements in particle size fractions for contaminated soils by a copper smelting from different zones of the Puchuncaví Valley (Chile). Chemosphere 111, 513–521. Quevauviller, P., 1989. Impact of industrial and mine drainage wastes on the heavy metal distribution in the drainage basin and estuary of the Sado River (Portugal). Environ. Pollut. 59, 267–286. Slavković, L., Škrbić, B., Miljević, N., 2004. Principal component analysis of trace elements in industrial soils. Environ. Chem. Lett. 2, 105–108. Sutherland, R.A., 2000. Bed sediment-associated trace metals in an urban stream, Oahu, Hawaii. Environ. Geol. 39, 611–627. Tang, Q., Liu, G., Zhou, C., Zhang, H., Sun, R., 2013. Distribution of environmentally sensitive elements in residential soils near a coal-fired power plant: potential risks to ecology and children's health. Chemosphere 93, 2473–2479. UNTEC, 2012. Fundación para la transferencia tecnológica, Diagnóstico plan de gestión atmosférica-región de Valparaíso implementación de un modelo atmosférico. http://www.sinia.cl/1292/articles-52679_DignosticoPlanGestionAtmosfericoValpo. pdf (Accessed 11 June 2014). Water Sanitation Health (WHO), 2003. Background document for development of WHO Guidelines for drinking-water quality. Wedepohl, K.H., 1995. The composition of the continental crust. Geochim. Cosmochim. Acta 59, 1217–1232. Wei, B., Yang, L., 2010. A review of heavy metal contaminations in urban soils, urban road dusts and agricultural soils from China. Microchem. J. 94, 99–107. Yay, O.D., Alagha, O., Tuncel, G., 2008. Multivariate statistics to investigate metal contamination in surface soil. J. Environ. Manag. 86, 581–594. Yaylah-Abanuz, G., 2011. Heavy metal contamination of surface soil around Gebze industrial area, Turkey. Microchem. J. 99, 82–92.