Determination of soluble ultra-trace metals and metalloids in rainwater and atmospheric deposition fluxes: A 2-year survey and assessment

Depósito de investigación de la Universidad de Sevilla https://idus.us.es/ “This is an Accepted Manuscript of an article published by Elsevier in: CHEMOSPHERE on 2013, available at: https://doi.org/10.1016/j.chemosphere.2013.02.044” 1 Determination of soluble ultra-trace metals and metalloids in rainwater and atmospheric deposition fluxes: A 2-year survey and assessment R. Montoya-Mayor, A.J. Fernández-Espinosa*, I. Seijo-Delgado, M. Ternero- Rodríguez Department of Analytical Chemistry. Faculty of Chemistry. University of Sevilla. C. Profesor García González 1, Campus of Reina Mercedes, E-41012 Sevilla Spain *Corresponding author. Tel: +34 95 455 4368. Fax: +34 954557168 E-mail address: [email protected] (A.J. Fernández-Espinosa). Abstract The present work investigates the relationships between composition of rainwater and dry deposition fluxes by trace metals and metalloids. A modification in automatic “wet-only” and “dry-only” samplers was applied, which allowed the collection and conservation of samples separately. ICP-MS technique was used for the determination of analytes in samples. A sequential acid treatment with nitric, hydrofluoric and finally perchloric acids was used to measure the total contents of metals and metalloids in coarse particles. Concentrations of soluble elements in rainwater were measured directly in filtered samples. Variation between periods of heavy and light rains was assessed. Almost all of the metals and metalloids studied in dry deposition showed important decreases in concentrations (40-92%) during periods of heavy rainfall. Most of these metals and metalloids presented their highest levels (25-40%) in heavy rainfall periods in rainwater samples. Sources were identified in both types of samples 2 collected using a new chemometric tool (SPCA). Urban traffic, surrounding contaminated soils and local anthropogenic sources were identified for rainwater samples. Natural and contaminated soils and general anthropogenic emissions were the sources identified for dry deposition fluxes. Keywords: Trace elements; Rainwater; Dry deposition fluxes; Sequential PCA 1. Introduction The chemical composition of atmospheric precipitation is closely related with the degree of air pollution in urban, industrial and rural ecosystems. It is well known that due to the dynamic nature of the atmosphere, metals in particle form can be deposited in areas far away from their original sources (Andreae and Rosenfeld, 2008). Trace metals in precipitation can also accumulate in surface water and soil, where they can cause harmful effects on aquatic life, forest ecosystems and plants (Barrie et al., 1987; Fernández and Rossini, 2006). The elimination of particles in the atmosphere can occur by: 1) aerosols can fall to the earth surface when particles are coarse, as dry deposition; 2) wet precipitation as rain or snow can remove coarse and fine particles, as wet deposition. In terms of wet deposition, it is accepted that there are two major wet scavenging processes: in-cloud scavenging and below-cloud scavenging (Slinn 1974; Charlson et al., 1983; Colin et al., 1987). Scavenging processes are determined by such particulate properties as particle size, gas solubility and the characteristics of precipitation. Other factors such as particle availability and vertical distribution could determine the relative significance of different removal processes during a single event (Lim et al., 1991). Besides, the potential capacity of dissolve toxic metals from atmospheric particulate matter was also studied employing speciation 3 schemes (Fernández et al., 2000; Fernández and Ternero, 2004b). Similar studies demonstrate the importance of the soluble and mobile fraction regarding the total metal content (Davis et al., 2001; Joshi and Balasubramanian, 2010). Acid rain has harmful effects on aquatic organisms, mainly because it favors the solution of trace metals, increasing their bioavailability (Baez et al., 2007). Anthropogenic emissions of particles and gases come from mainly the industrial combustion of fossil fuels and agriculture (Galloway et al., 1982). Recently, a lot of information about the composition of precipitations has been gathered in America, Central and Northern Europe (Arsene et al., 2007; Celle-Jeanton et al., 2009) and other parts of the world, as China (Wei Hong et al., 2005), Japan (Hou et al., 2005), India (Momin et al., 2005), etc. In Spain, several authors have studied the chemistry of atmospheric precipitation (Alastuey et al., 1999; Encinas et al., 2004). Furthermore, the papers pertaining to rainwater and dry deposition composition are scarce when compared to those referring to Total Suspended Particles (TSP) (Fernández et al., 2004a). Therefore, the aim and novelty of this work is to study the relationship between atmospheric pollutants in wet deposition and in dry deposition samples in an urban location, covering a long period of more than two years; additionally other objective is to investigate the relationship between pollutants in periods of heavy rainfall and periods of light rainfall. Finally, concentrations measured will allow identifying sources of atmospheric pollutants using a recent novel chemometric tool on both types of samples. 2. Experimental 2.1. Instrumentation for analytical determinations Determination of heavy metals (Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Mo, Ag, Cd, Tl and Pb), light, alkali and alkaline-earth metals (Be, Al, Sr, Cs and Ba), actinides (Th and U) and 4 metalloids and nonmetals (B, As and Se and Sb) was carried out by Inductively Coupled Plasma-Mass Spectrometry (ICP-MS), and a X7 SERIES ICP-MS (Thermo Elemental, Pierce Biotechnology, Rockford, IL, USA) was used. ICP-MS is often used for the determination of environmental samples. Their operating conditions (Montoya et al., 2001) were optimized daily to give the highest sensitivity and S/N ratio. The complete determination of heavy metals and metalloids by ICPMS was carried out according to the 200.8 US EPA Method (USEPA, 1994) with some modifications relating to the tuning and mass calibration solution. There were some analytes where the number of samples over the detection limits were very low, <15%, (Ti, Fe, Ag, Sn, Hg, Tl, Cs, Be, Si, Se, Th and U) and they were excluded from statistical analysis, indicating only an informative value. 2.2. Methodology for chemical analysis The chemical composition determined in rainwater samples was obtained by analyzing concentrations of metals and metalloids. The general parameters pH, electrical conductivity (EC) and redox potential (E) were measured in situ during the collection of samples from the automatic sampler. Experimental pH values were converted to H+ concentrations for their statistical treatment. In addition, the concentration of rainwater solutes was calculated from a method based on the measurement of the pH-value and the electrical conductivity (Krämer et al 1996). Composition of soluble pollutants in rainwater was determined on direct samples only after vacuum filtration. The chemical composition determined in dust samples was obtained by acid treatment. For this, coarse particles were acid digested in two stages. Particles were digested with a mixture of 5.0:2.5 ml HF:HNO3 in a Teflon reactor (Savillex, Minnesota, USA; 60 ml) at 105°C in an oven over night. Later, 2.5 ml HClO4 was added to the residue and this was 5 dried and then re-dissolved with 2.5 ml HNO3 and a volume of 50 ml with ultrapure water (Montoya et al., 2011) was made up. Acids and standard solutions for metals were taken from VRW (Darmstadt, Germany). Ultrapure water came from Waters-Millipore (Milli-Q grade, model Milli-Q Plus). Standard solutions for pH, EC and Eexp came from Crison® as well as the corresponding laboratory instruments. Additional care was taken in handling the samples to avoid contamination problems. Consequently, all reagents, materials and samples were handled within a vertical laminar airflow cabinet (Indelab, model IDL-48 V) that ensured air cleanliness standards of class 100 according to Federal standard 209E, and it was located in an isolated dark room for analytical weighing known as the balance room. 2.3. Measurement site and sampling In this paper, the results corresponding to over two years of sampling and analysis of rainwater and atmospheric deposition loads are presented. Samples were simultaneously collected in an urban-trafficked site close to the navigable Guadalquivir River (Andalusia, Spain). This study was undertaken in the city of Sevilla (38°12′–36°51′ N, 4°39′–6°32′ W, 141.3 km2), a city 60 km from the sea and only 8 m above sea level, the capital of Andalusia, in South West Spain. The collector was installed at 4 meters over the floor into the University Campus of Reina Mercedes. Sevilla is characterized by high temperatures and low wind speeds. Prevalent air currents come from the southwest–northeast direction. All this data is averaged from the 1971–2000 period (AEMET, 2012). The majority of rain episodes and Saharan inputs enter into the region of Andalusia through this valley from a westerly direction, mainly from the southwest. 6 Important episodes of ‘heat wave’ occur frequently in summer coming from Sahara and Sahel deserts. Rainwater samples (wet deposition) and bulk depositions fluxes (dry deposition) were collected with a TISCH ENVIRONMENTAL 78 100 APS (Tisch Environmental Inc., South Miami Ave. Cleves, Ohio. USA) automatic wet-only and dry-only sampler. This type of sampler is designed to collect atmospheric depositions in two separate buckets of 15 litres (Fig. 1): one of these, the rainfall bucket (“Wet” bucket), opens only during episodes of rain, detected by an electronic sensor, and the other, the dust fall bucket (“Dry” bucket) is always open during dry periods. This allows the collection of both wet deposition (rainwater) and dry deposition (dust depositions) separately. All samples were collected during a period of over 2 years, from September 2006 to January 2009. Samples of rainwater were collected dairy during each rain episode and accumulated later for a period of a week according to the recommendations of the OSPAR (OSPAR, 2007; European Community, 1998) and Barcelona Conventions (European Community, 1977; MAGRAMA, 2012a) for the coastal zones protection. Samples of dust depositions were collected every 28 days, once per month, according to the Federal methods of the USA (USEPA, 2008) and Regional laws of Spain (BOJA, 2006). This resulted in a total of 67 dairy samples collected and 39 weekly values accumulated for rainwater and a total of 22 samples collected of bulk samples, which are presented in this study. Also, for relationships with 22 monthly samples of dry deposition fluxes, 20 monthly values were accumulated for rainwater samples. According to these conventions, dry deposition fluxes were expressed as mg m-2 day-1 (dividing total monthly depositions by 28 days). Additionally to the weekly period – expressed as L m-2 week-1 – rainwater samples were also averaged over monthly periods to give an objective relationship with dry depositions, expressing it as L m-2 day-1, not by month. For metal and metalloid, concentrations in dry deposition fluxes and in rainwater were expressed 7 as µg m-2 day-1. Additionally concentrations of analytes in rainwater samples were also expressed as µg L-1 week-1 referring to a concentration in an aqueous basis for comparison with other world results. 2.4. Design of the modifications on the sampling engine To ensure the optimal preservation of samples during dairy rain events, some modifications were carried out to the automatic collector (Fig. 1): Firstly, a small-fridge was protected from the sun under the sampler. The rainwater was collected in bottles made of low density polyethylene (LDPE) with a 0.5-L capacity stored inside the small-fridge. The fridge power was adjusted to achieve a temperature between 3 and 5ºC. Secondly, a stainless steel funnel with a collection area of 0.063 m2 was inserted into the “Wet” bucket to collect rain drops and to channel rainwater into the small-fridge. The surface of funnel and all tubes were Teflon-coated, thus avoiding the contact between rainwater and the surface of the steel funnel, so as to prevent cross-contamination. Thirdly, a waste system was coupled to the sampling train to prevent the rain overflowing into the small-fridge during extreme rainfall events. All connections between different parts of the sampling train were made with Teflon tubes. The modified automatic sampler is shown in Fig. 3. 2.5. Sample pretreatment of rainwater and bulk deposition Rain drops wash atmospheric particles of different sizes (coarse and fine particles, see Fernández et al., 2004a) during their fall; these particles contain metals and metalloids that are dissolved, together dissolved gases, determining part of the composition of rainwater. After 8 each rain event, these particles are removed through filtering samples in the laboratory. Vacuum filtration was done with 47 mm  ultrapure quartz filters Whatman QM/A (Middlesex, UK), which have 0.45 mm film thickness and a particle retention for 0.3 µm of 99.95% -determined according to ASTM (American Society for Testing and Materials) D 2896-91-. The mean concentration of suspended solids in rainwater was extremely low, 57.0 mg L-1. After the in situ determinations and vacuum filtration, samples were divided in three portions for determinations by ICP-MS. For each replicate, an aliquot of concentrated Nitric acid (Suprapur quality, from Merck; Darmstadt, Germany) was added to 10 mL of the filtered sample until a concentration of 1% was achieved. In the case of bulk particles (only coarse particles) collected into the “Dry” bucket, the complete samples were carefully transferred from the bucket to vials using dry and clean nylon pencils. Vials were previously dried and weighed before the transfer and sealed hermetically after. The treatment of coarse particles was done in the balance room monitored at 50±5% R.H and 20±1ºC. After transfer coarse particles in each open vial were placed into a 96 L desiccator for a 48-h period continuously under 5% R.H. Conditions of low humidity inside the desiccator were achieved with a combination of Silica-gel and Phosphorus pentoxide, and using an indoor air extractor. After drying treatment vials were then weighed to a precision of 0.1 mg on an analytical balance (Sartorius LA 130 S-F). The bulk deposition concentration was expressed in mass of particles by surface area and day (mg m-2 day-1). 2.6. Multivariate statistical technique (SPCA, Sequential Principal Component Analysis) On the basis of the results, basic and multivariate statistical analyses were applied to the analytical and meteorological data. The SPCA tool is a PCA technique modified by Fernández et al., 2004a. The SPCA method consists in identifying and characterizing outliers that appear 15 According to the SPCA procedure mentioned earlier, Table 3 shows the results corresponding to the Sequential PCA for rainwater samples. SPCA1 to SPCA6 include outliers, whereas SPCA7 is the final SPCA. Outliers have been interpreted and characterized before their elimination. Results obtained in SPCA1 showed one interfering case, se04. This case corresponds to an episode with high concentrations of variables (H+, B, Zn, Cd, Al and Cu) originating from urban traffic sources. In addition, this sample was characterized by its extremely low pH value (4.0) as an acid rain episode. PC1 (only one case) accounts for 27.2% of the variance of all cases; consequently, this proves that the first PCA – SPCA1 – is not valid to explain the variance and main sources of 67 samples. Therefore, according to the present methodology of SPCA, it is necessary to remove the outliers and repeat the PCA sequentially until no more outliers appear. After removing se04 and repeating SPCA, the outlier se34 (March 2007) appears in SPCA2 corresponding to the highest values of atmospheric pressure and evapotranspiration with southern winds of low speed, so, an anticyclonic period prior to spring rains. SPCA3 shows the case se62 (November 2007) corresponding to a dry period of more than 20 days with high wind speed and characterized by As and Ba (traffic emissions, Fernández and Ternero, 2004b). In SPCA4, the group Mo, redox potential and Sb (se91, October 2008) was originated mainly from emissions of intense traffic. In SPCA5, se47 (May 2007) was characterized by the highest values of rainfall, 85 L m-2, and a low value for electrical conductivity, 7.7 μS cm-1. The variables grouped V, rainfall-PP and Cu showed a mixed origin from both anthropogenic and natural sources. Finally, in SPCA6, se20 (November 2007) was characterized by Cr and high wind speeds (4.1 m s-1), showing a mixed origin, from foundries, soil or traffic. 16 Once all outliers were explained and removed, the last PCA – SPCA7 – could not reveal the presence of any interfering PC and 7 significant PCs explain up to 67.4% of the total variance in SPCA7 (Table 3). PC1 is formed by EC and Ba related to the traffic of the city (Fernández and Ternero, 2004b). Most of representative cases corresponded to April 2008, in days with winds coming from southern directions, with moderate and high wind speeds. PC2 is formed by V, Cu and Zn, related with soils polluted by a fertilizer factory of the city, confirmed by previous studies (Fernández 2001). PC3 and PC4 are represented by soils and also traffic emissions (Pb and Ba). The Ba2+–Sr2+ pair is frequently attributed to a terrestrial source, although Ba2+ together Cu and Pb is also attributed to vehicle traffic (Fernández and Ternero, 2004b). So, we can relate these PCs with the resuspended contaminated soils from the trafficked city. Samples of this group correspond to the summer-autumn period, characterized by high temperatures. PC5, PC6 and PC7 are formed by elements characterized by anthropogenic activities. The metals Cd, Cu, Ni and Zn are related to emission from foundries (Fernández and Ternero, 2004b). Mo comes from the making and use of alloys in the nearest the great national city’s shipyard and the plane factory of the industrial zone of the Port of Sevilla, with more than 25 enterprises conducting industrial activities. Dust depositions When we realized SPCA for dry deposition, we found 2 PCs that explained 89.3% of the variance. These PCs are very different to those from rainwater and no outliers appear. In SPCA for dust depositions there are only 2 PCs, while in SPCA for rainwater samples many different PCs appeared. As a consequence, there is no diversification of sources in coarse particle depositions, as shown in Table 4. 17 PC1 is formed by Fe, Ti, Co, V, Al, Mn, Ni, As, B, Sr and Ba. Some elements such as As or Ni may have an anthropogenic origin, but most of the elements have a predominantly terrigenous origin, like Fe, Ti, Al, Mn, Sr and Ba. So we can relate this PC to the influence of soils in dry deposition. In contrast, PC2 is represented by Sb, Cd, Mo, Cr, Zn, Cu, Pb and As. Moreover, PC2 has elements that mainly have an anthropogenic origin. In dry depositions we found a very basic division of the variables studied; just a terrigenous PC and one anthropogenic PC, where the terrigenous PC is much more important. It must be taken into consideration that the bulk depositions were collected every 28 days. This is a long period of time with different weather conditions and emissions from different sources. Therefore, all these factors contribute to a mixture of pollutants, not occurring the diversification of sources of emission. 4. Conclusions For sampling, the modification and optimization of the original automatic sampler has allowed the optimal conservation of rainwater samples. Differences between periods of heavy and light rainfall were much more severe for metals and metalloids in dry deposition than in wet deposition. Dry deposition presented higher values for B (92%), Tl (49%), Th (44%) and U (41%) in periods of light rainfall than in periods of heavy rainfall. On the other hand, almost all of these metals and metalloids presented their highest values in rainy periods for rainwater samples, such as Cr (-90%), Co (-75%) and Cs (-72%). This way it was calculated the contribution of the rainwater to the scavenging of atmospheric coarse particles. Similar studies it would be interesting in the future for suspended particles, PM10, PM2.5 or ultrafine particles. The use of the advanced chemometric tools Sequential Principal Component Analysis (SPCA) have been very useful for the interpretation of the results in the complex environmental system studied. Thus, it has identified the mixture of a traffic source (25.6 % of the variance), 18 followed by contaminated soil (18.1 %) and anthropogenic sources (16.4 %) such as metallurgy and port activity. The terrestrial source is also present (6.9%), along with the combustion source (4.0 %). The chemometric study on dry depositions has shown no diversified sources as a result of the long period of sampling. Two principal components were only obtained: soils (69.3 % of the variance) and anthropogenic sources (20.0 %), however the two sources were good discriminated. Therefore anthropogenic pollutants contained in coarse particles probably could cause contamination in soils and plants in the city. This fact is another future objective to develop studies in rural and coastal sites where many crops constitute important economic activities for human consumption. Acknowledgments The authors would like to thank the Junta de Andalucía for its financial assistance in carrying out this research project (P05-RNM-1177) and for the generous grant provided to the first author. References Andreae MO, Rosenfeld D (2008) Earth-Sci Reviews 89:13–41 Alastuey A, Querol X, Chaves A, Ruiz CR, Carratala A, Lopez-Soler A (1999) Environ Pollut 106:359-367 Al-Momani, IF (2003) Atmos Environ 37:4507–4515 Arsene C, Olariu RJ, Mihalopoulos N (2007) Atmos Environ 41:9452-9467 Association of Analytical Communities, AOAC (2010) Guidelines for single laboratory validation of chemical methods for dietary supplements and botanicals. AOAC INTERNATIONAL. Gaithersburg, Maryland, USA. Baez A, Belmont R, García R, Padilla H, Torres MC (2007) Atmos Res 86:61–75 19 Barrie LA, Lindberg SE, Chan WH, Ross HB, Arimoto R, Church TM (1987) Atmos Environ 21:1133-1135 Celle-Jeanton H, Travi Y, Loÿe-Pilot MD, Huneau F, Bertrand G (2009) Atmos Res 91:118- 126 Charlson RJ, Vong R, Hegg DA (1983) J Geophys Res 88:1375-7 Chester R, Nimmo M, Keyse S (1996) Environ Sci Tech Lib 11:253-273 Colin JL, Jaffrezo JL, Pinart J, Roulette-Cadene S (1987) Atmos Environ 21:1147-57 Cuadros L, García A M, Bosque J M (1996) Anal Letters 29:1231–1239 Davis AP, Shokouhian M, Ni SB (2001) Chemosphere, 44:997–1009 Demircioglu E, Sofuoglu A, Odabasi M (2011) J Hazar Mat 186:328-335 Encinas D, Calzada I, Casado H (2004) Aerosol Sci Tech 38:685-691 European Community (1977) Decition 77/585/EC of 25 July 1977, relative to the Convention for the protection of the marine environment of the Mediterranean Sea (Barcelona Convention) European Community (1998) Decition 98/249/EC of 7 October 1997, relative to the Convention for the protection of the marine environment of the North-East Atlantic (Paris Convention) Fernández AJ, Ternero M, Barragán FJ, Jimenez JC (2000) Chemosphere 2: 123-136 Fernández Espinosa AJ, Rossini Oliva S (2006) Chemosphere 62: 1665-1672 Fernández AJ, Ternero M, Fernández F (2004a) Atmos Environ 38:873-886 Fernández AJ, Ternero M (2004b) Anal and Bioanal Chem 379:684-699 Fernández Espinosa AJ (2001) Fundación Focus-Abengoa. ISBN:84-472-0629-7. Sevilla, Spain. Fernández-Espinosa A J, Ternero-Rodríguez M (2004a) Idojárás 108:1-20 Galloway JN, Thornton JD, Norton SA (1982) Atmos Environ 16:1677-1700 Garnaud S, Mouchel JM, Chebbo G, Thevenot DR (1999) Sci Total Environ 235:235-245 Herut B, Starinsky A, Katz A, Rosenfeld D (2000) Atmos Environ 34:1281-1292 20 Hou H, Takamatsu T, Koshikawa MK, Hosomi M (2005) Atmos Environ 39:3583–3595 Joshi UM, Balasubramanian R (2010) Chemosphere 80: 310-318 Kaya G, Tuncel G (1997) Atmos Environ 31:3985-3998 Krämer M, Schüle M, Schütz L (1996) Atmos Environ 30:3291-3300 Lim B, Jickells TD, Davies TD (1991) Atmos Environ 25A: 745-762 Ministerio de Agricultura, Alimentación y Medio Ambiente, MAGRAMA (2012a) http://www.magrama.gob.es/es/costas/temas/proteccion-del-medio-marino/conveniosinternacionales/convenio_de_barcelona.aspx Momin GA, Ali K, Rao PSP, Safai PD, Chate DM, Praveen PS, Rodhe H, Granat L (2005) J Geophys Res Atmos 110:D08302/1-D08302/10 Montoya-Mayor R, Fernández-Espinosa AJ, Ternero-Rodríguez M (2011) Anal Bioanal Chem 399:2031-2041 Ospar Commission (Convention for the Protection of the marine Environment of the North- East Atlantic) (2007) Comprehensive Atmospheric Monitoring Programme. Pollutant deposits and air quality around the North Sea and The North-East Atlantic in 2005. ISBN 978-1- 905859-68-9. London, United Kingdom The Official Gazette of the Autonomous Community of Andalusia, BOJA (2006) Andalusian Decree 151/2006. US Environmental Protection Agency, USEPA (1994) Method 200.8. Determination of trace elements in waters and wastes by inductively coupled plasma-mass spectrometry. Environmental Monitoring Systems Laboratory. Office of Research and Development. Cincinnati, Ohio, USA. US Environmental Protection Agency, USEPA (2008) Quality Assurance Handbook for Air Pollution Measurement Systems, Volume IV-Meteorological Measurements, EPA/600/R- 94/0-38d 21 Wei Hong, Wang Jian-li, Li Xu-guang (2005). Xinan Shifan Daxue Xuebao, Ziran Kexueban 30:725–729 Wood R (1999) Trends in Anal Chem 18:624-632 Yi SM, Totten LA, Thota S (2006) Atmos Environ 40:6626–6637 22 Table 1 Recovery, linearity and sensitivity for the ICP-MS method Analyte Concentration obtained±SD (µg L -1 ) CRM concentration ±SD (µg L -1 ) Recovery (%) Linearity (%) LOD (µg L-1) LOQ (µg L-1) Be 20 ±0.10 19.2 ± 0.20 96.0 99.1 0.035 0.117 Al 120 ± 0.60 120.6 ± 1.33 100.5 97.4 0.409 1.363 Cr 20 ± 0.10 19.3 ± 0.09 96.5 97.8 0.042 0.140 Mn 40 ± 0.20 38.5 ± 0.45 96.3 98.7 0.035 0.117 Fe 100 ± 0.50 113.1 ± 2.38 113.1 95.7 2.872 9.573 Co 25 ± 0.13 24.6 ± 0.23 98.4 98.8 0.040 0.133 Ni 60 ± 0.30 53.8 ± 0.59 89.7 98.1 0.126 0.420 Cu 20 ± 0.10 18.5 ± 0.20 92.5 98.2 0.071 0.237 As 80 ± 0.40 80.6 ± 0.16 100.8 98.1 0.125 0.417 Se 10 ± 0.05 9.5 ± 0.19 95.0 98.2 0.463 1.543 Sr 250 ± 1.25 264 ± 1.85 105.6 99.4 0.035 0.117 Mo 100 ± 0.50 96.8 ± 1.07 96.8 99.3 0.062 0.207 Ag 2 ± 0.01 2 ± 0.03 100.0 99.5 0.052 0.173 Cd 10 ± 0.05 10.3 ± 0.17 103.0 99.4 0.053 0.177 Sb 10 ± 0.05 10.4 ± 0.07 104.0 98.1 0.048 0.160 Ba 50 ± 0.25 50 ± 0.04 100.0 99.5 0.083 0.277 Tl 10 ± 0.05 10 ± 0.01 100.0 99.5 0.045 0.150 Pb 40 ± 0.20 36.2 ± 0.29 90.5 99.1 0.135 0.450 U 10 ± 0.05 10.1 ± 0.18 101.0 99.8 0.053 0.177 SD standard deviation, CRM certified reference material 23 Table 2 Metal and metalloid concentrations in 22 monthly samples of bulk particles and 39 weekly samples of rainwater. Percentage %LH of variation between ELR and HER periods for bulk particles and rainwater *: Medians and minimum-maximum in brackets **: Rainfall expressed as µg L-1 week-1 according the international conventions ---: Not calculated because mean concentration was ≤LOD in rainwater Analyte Bulk (B)* (µg m-2 day-1) Rainwater (R)* (µg m-2 day-1) B/R-R/B (Ratio>1) Rainwater (µg L-1 week-1)** %LH in dry deposition %LH in Rainwater Al 1705 (1042-4666) 30.5 (≤LOD-853) 56 2.4 (≤LOD-67) 40 -30 Fe 939 (505-2352) 18.3 (≤LOD-36.5) 26 ≤2.87 28 --- Ti 173 (101-503) 2.3 (≤LOD-5.5) 32 ≤0.43 40 -67 Mn 29 (17-64) 14.0 (≤LOD-320) 2.1 1.1 (≤LOD -25) 37 -45 Zn 20 (8-75) 306 (≤LOD-3131) 15 24.0 (≤LOD-246) -38 59 Ba 10 (4-52) 153 (10.5-5116) 15 12.0 (0.825-402) -27 31 Cr 9 (4-22) 2.6 (≤LOD-776) 3.5 0.2 (≤LOD-61) 37 -90 Cu 9 (7-15) 12.7 (≤LOD-407) 1.4 1.0 (≤LOD-32) -11 23 Sr 7 (5-23) 43.3 (3.5-2800) 6.2 3.4 (0.277-220) 37 -45 Pb 6 (4-15) 1.8 (≤LOD-42.0) 3.4 ≤0.14 (≤LOD-3.3) -35 93 Ni 2 (1-4) 10.2 (≤LOD-331) 5.1 0.8 (≤LOD-26) 25 -65 B 0.8 (0.4-32) 34.4 (≤LOD-1044) 43 2.7 (≤LOD-82) 92 -56 Co 0.7 (0.4-2) 0.51 (≤LOD-99.3) 1.4 ≤0.04 (≤LOD-7.8) 31 -75 As 0.6 (0.3-1) 1.6 (≤LOD-127) 2.8 ≤0.13 (≤LOD-10) 30 99 Mo 0.5 (0.3-1) 0.76 (≤LOD-649) 1.5 ≤0.06 (≤LOD-51) -16 98 Sb 0.4 (0.3-0.9) 1.2 (≤LOD-35.6) 2.9 0.09 (≤LOD-2.8) -28 74 Th 0.3 (0.2-1) 0.89 (≤LOD-1.9) 3.0 ≤0.07 (≤LOD-0.15) 44 --- Cs 0.2 (0.08-0.4) 0.64 (≤LOD-54.7) 3.2 ≤0.05 (≤LOD-4.3) 40 -72 U 0.1 (0.06-0.3) 0.64 (≤LOD-1.3) 6.4 ≤0.05 (≤LOD-0.1) 41 --- Cd 0.1 (0.05-0.2) 0.64 (≤LOD-14.0) 6.4 ≤0.05 (≤LOD-1.1) 0.2 -64 Ag 0.06 (0.04-0.09) 0.64 (≤LOD-12.7) 11 ≤0.05 (≤LOD-1.0) 12 --- Be 0.05 (0.03-0.1) 0.51 (≤LOD-3.8) 10 ≤0.04 (≤LOD-0.3) 40 --- Tl 0.02 (0.01-0.06) 0.64 (≤LOD-1.3) 32 ≤0.05 (≤LOD-0.1) 49 --- 24 Table 3 Sequential PC Analysis (SPCA) of seven iterations for rainwater (19 variables x 67 cases) SPCAn PCn Variance (%)/ Cumulative variance (%) Sign Variables Cases Source SPCA1 PC1 27.2 / 79.1 + H+, B, Zn, Cd, Al, Cu se04 Acid rain SPCA2 PC1 27.5 / 76.6 + Co, Cu, Mn, V, Al, B, EC se34 High conc. SPCA3 PC6 5.3 / 80.2 + As / Ba se62 Traffic SPCA4 PC6 5.3 / 78.0 - Mo / E, Sb se91 Traffic SPCA5 PC8 4.3 / 77.6 + V, PP, Cu se47 Industrial- Plastics? SPCA6 PC11 3.3 / 78.4 + Cr se20 Unidentified SPCA7 PC1 25.6 / 25.6 + EC/ Ba se82, se 62 / se86 / se 83, se81, se60 Traffic PC2 10.4 / 36.1 + V / EC, Cu / Zn se80, se81 / se45, se87, se85, se47, se88 / se74, se50, se84 / se89 Contaminated soil PC3 8.1 / 44.2 + Pb, Sr, / Ba / Cu se60, se55 / se61 / se62, se58, se57 / se80 Soiltraffic PC4 6.9 / 51.1 + Mn, Al / Ba, B / Zn se80, se90 / se95 / se18, se26 / se20, se75, se62, se08 PC5 6.3 / 57.4 + Cd, Zn / Co, B se33 / se06, se80 / se47 / se50, se22, se09 Anthropogenic: Metallurgy and Port Activity PC6 5.6 / 62.9 - Mo, As / Sr, Cu se65 / se92 / se 74, se55, se77, se67 / se35 PC7 4.5 / 67.4 - Ni / Cr, Cu / Zn se10 / se80, se41, se43 / se38, se44, se 35, se 55, se 07 Variables are in decreasing order of loading above 0.5. Underlining indicates variables with the highest loading between PCs. Bold indicates cases with high scores (outliers). EC electrical conductivity, E redox potential, PP total rainfall SPCA1 to SPCA6 outliers, SPCA7 final PC analysis