Atmospheric Research 86 (2007) 61 – 75 www.elsevier.com/locate/atmos

Chemical composition of rainwater collected at a southwest site of Mexico City, Mexico A. Báez ⁎, R. Belmont, R. García, H. Padilla, M.C. Torres Laboratorio de Química Atmosférica, Centro de Ciencias de la Atmósfera, Universidad, Nacional Autónoma de México, Circuito Exterior, Ciudad Universitaria, México D.F. 04510, Mexico Accepted 5 March 2007

Abstract Measurements of the trace metals Cd, Cr, Mn, Ni, Pb, V and Al in soluble and insoluble rain fractions and SO−4 2, NO−3 , Cl−, HCO−3 , Ca2+, Mg2+, Na+, K+, NH+4 and H+ in soluble fractions were performed in rainwater collected at a southwest site of Mexico City during the rainy seasons of 2001 and 2002. Aluminum presented the highest volume-weighted mean concentration (VWMC) in both insoluble and soluble fractions. In the insoluble fractions, the VWM of the other trace metals decreased in the order Mn, Pb, Ni, V, Cr and Cd, and in the soluble fractions in the order Mn, V, Ni, Pb, Cd and Cr. Ammonium presented the higher VWMC, − − 2+ − + + 2+ followed by SO2− and K+. Air mass back trajectories were associated to the concentrations 4 , NO3 , HCO3 , Ca , Cl , H , Na , Mg + 2+ 2+ + , Ca , Mg , NH and H observed during each rainy day. Trace metal concentrations were not clearly of trace metals and of SO2− 4 4 related to wind direction. Enrichment factors related to the relative abundance of elements in crustal material were calculated using Mg as reference. The high enrichment factors (EFc) suggested that, in general, trace metals and major ions had an anthropogenic origin. Aluminum, K+, and Ca2+ were the only elements that had a significant crustal source. Factor analysis (Principal Component Analysis) with Varimax normalized rotation grouping the elements analyzed into three factors. Factor 1 indicated a crustal con+ tribution for Ca2+, K+, Mg2+ and anthropogenic sources for SO2− 4 , NH4 and V. Factor 2 indicated a high loading for Al, Ni and Mn, that indicate possible contribution of anthropogenic sources but with a significant crustal contribution for Al. Factor 3 indicated an anthropogenic origin for H+ and NO−3 . Pearson's correlations show that Al correlated with all the metals, including Ca2+ and Mg2+. The solubility of trace metals did not depend on rainwater pH. As it was expected, Al presented the highest wet deposition flux. © 2007 Elsevier B.V. All rights reserved. Keywords: Heavy metals; Wet precipitation; Major ions; Natural and anthropogenic sources; Enrichment factor; Mexico City

1. Introduction Precipitation chemistry has been exhaustively studied in urban and rural areas (Lee et al., 2000; Lara et al., 2001; Kulshrestha et al., 2003; Astel et al., 2004; Khare et al., 2004; Mouli et al., 2005) and some researchers ⁎ Corresponding author. Tel.: +52 55 5622 4071; fax: +52 55 5622 4050. E-mail address: [email protected] (A. Báez). 0169-8095/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.atmosres.2007.03.005

have included the study of trace metals (Tanner and Wong, 2000; Luo, 2001; Roy and Négrel, 2001; AlMomani et al., 2002; Hu and Balasubramanian, 2003; Al-Momani, 2003; Migliavacca et al., 2004). The study of trace metals in wet and dry precipitation has increased in the last decades because of their adverse environmental and human health effects. Some metals such as Pb, Cd and Hg, among others, accumulate in the biosphere and may be toxic to living systems (Galloway et al., 1982; Barrie et al., 1987). Anthropogenic sources

62

A. Báez et al. / Atmospheric Research 86 (2007) 61–75

have substantially increased trace metal concentrations in atmospheric deposition. In addition, acid precipitation favors the dissolution of many trace metals, which enhances their bioavailability. If the concentrations are too high, many of the trace metals can harm human health through the consumption of drinking water and/ or aquatic organisms. Trace metals from precipitation can also accumulate in surface waters and soils where they may cause harmful effects to aquatic life and forest ecosystems (Howard et al., 2004). Trace metals are deposited by rain, snow and dry fallout. Rainout and washout are the predominant processes of deposition by rain (Seinfeld and Pandis, 1998). Usually over 80% of wet deposited trace metals are dissolved in rainwater, reaching the vegetation canopy in the most favorable form for uptake (Valenta et al., 1986). Atmospheric transport and deposition processes are important in the global recycling of trace metals. Since the atmosphere of Mexico City is one of the most polluted cities in the world, it was considered important to analyze the trace metals Al, Cd, Cr, Pb, Ni, Mn, and V in soluble and insoluble rain fractions and the major ions SO4− 2, NO3−, Cl−, Ca2+, Mg2+, Na+, K+, NH4+ and alkalinity (HCO3−) only for the soluble fractions. Báez et al. (1980) determined the concentration of Pb, Cd and Cr in rainwater in several regions of Mexico. The results obtained in the Universidad Nacional Autónoma de Mexico (UNAM) in 1980 were 133,

0.88, and 4.54 μg l− 1 for Pb, Cd and Cr, respectively. Since the government enforced the reduction of lead in fuels, we considered that it was important to investigate present lead levels in the atmosphere after this reduction took place. The values obtained in this study, 22.7, 4.62 and 2.57 μg l− 1, in the same order, were noticeably lower than those obtained in 1980. However the 1980 results can not be compared with recent values because both the collection and the analytical methods were much different that the present techniques, for instance, rainwater sampling was made base on 30 days bulk precipitation, now it is made daily or in event bases collection, and the analytical instruments have had many improvements. 2. Materials and methods 2.1. Sampling site The Atmospheric Sciences building at UNAM campus, located in southern Mexico City at 2200 m above sea level (masl) at 19°19.57′ N latitude and 99°10.55′ W longitude (Fig. 1). Buildings surrounded by green areas with moderate to high traffic density characterize the UNAM campus. Mexico City is located on a volcanic high plateau. Volcanic rocks plus alluvial, fluvial and lacustrine deposits overlie limestones. A detailed description of the geological setting of Mexico City Basin

Fig. 1. Sampling site location.

A. Báez et al. / Atmospheric Research 86 (2007) 61–75

is given by Flores-Estrella et al. (2007). Jazcilevich et al. (2000) concluded that two major changes have taken place in the basin, as regards to land use: a drastic growth in the urban area, coupled with a dramatic reduction of the lacustrine system. 2.2. Sampling Rainwater was collected on the roof of the Atmospheric Sciences building. Daily wet-only rainwater samples were collected from mid-May to the end of October in 2001 and 2002, the rainy season in Central Mexico (Jáuregui, 2000), with two automatic wet/dry precipitation collectors (Andersen, General Metal Works, Inc.) into standard high-density polyethylene (HDPE) buckets. Practically, all rainy days in Central Mexico occur from mid-May to mid-October. Rains are scarce or even absent during the rest of the year. Therefore, rain events during the dry season are too few to obtain meaningful statistics regarding seasonal effects. Prevailing winds during the rainy season come from the north–northeast and usually the clouds are of convective type (Jáuregui, 2000). One automatic collector was used to collect samples for the analysis of trace metals in soluble and insoluble fractions, and the other one for the analysis of pH and the major ions SO4− 2 , NO3− , Cl− , Ca 2+ , Mg2+ , Na+ , K + , NH4+ and alkalinity (HCO3− ) in soluble fractions only. The lid arms of both collectors were coated with Teflon and the original aluminum cover and plastic-covered pad was replaced with an aluminum lid covered with Teflon and Teflon pad, and the buckets were tightly sealed with polyethylene covers. Buckets used to collect rain samples for trace metal analysis, filtration flask and storage bottles for filtrated samples were washed and brushed, rinsed with distilled water, and soaked in 20% nitric acid (Merck) for 24 h, then, rinsed several times with deionized water (DI) and sealed into double plastic bags. All the material was again rinsed with DI water. The 0.4 μm polycarbonate membrane filters (47-mm in diameter) and a Gelman magnetic polysulfone filter funnel were leached with 300-ml of DI, 20% ultra pure nitric acid solution and with DI water 3 times. Buckets used to collect rain samples for major ions analyses, flask, and storage bottles were washed and brushed, rinsed with distilled water and then with DI water. When the amount of precipitation (sample volume) was not enough to analyze all the trace metals and major ions considered in this research, the rain samples were discarded, so, 81 rain samples were included for metals and 71 for major ions to perform the statistical analysis.

63

2.3. Sample treatment The following sample treatment was performed under a clean hood maintained at a positive pressure to limit the possibility of contaminants entering from the laboratory. After collection, rainwater samples for chemical analyses from major ions were filtered through Millipore 0.45-μm membrane filters leached with DI water before chemical analyses and stored at 4 °C. As NO3− were measured the filtered solution was not acidified. Rainwater samples collected for measurements of trace metals were filtered through a polycarbonate membrane filter, using the Gelman magnetic polysulfone filter funnel. Approximately, a 60-ml aliquot of filtrate was transferred into an HDPE bottle and acidified at pH 1.8 with 0.016 N HNO3 (Merck, Suprapure) and kept at 4 °C until analysis. On the other side, the polycarbonate membrane filters with insoluble particles were put into a nitric acid washed petri dish and also kept at 4 °C until sample treatment. Then the filters were subjected to an ultrasonic extraction procedure with 15 ml of a 3 M HNO3 solution. The resulting solution was diluted to 25 ml. 2.4. Chemical analysis Rainwater pH was measured within 24 h after collection of samples by using an Orion 960 autochemistry system. An YSI 3200 conductivity instrument was used for conductance measurements. Sulfates, NO3−, and Cl− were analyzed by non-suppressed ion chromatography, and ammonium was determined by suppressed chromatography using a Perkin Elmer instrument equipped with an isocratic LC pump 250 and a conductivity detector ConductoMonitor III Model. Na+, K+, Ca2+ and Mg2+ were analyzed by flame atomic absorption spectrometry with a GBC 932AA instrument at 589, 766.5, 422.7 and 282.5 nm, respectively. Hollow cathode lamps (Photron) for these ions were used. Calibration standards were prepared from certified standards of each metal. Alkalinity was determined using the Gran's titration method with an Orion 960 autochemistry system, and HCO3− concentrations were computed using the equation described by Stumm and Morgan (1981). The detection limits in μeq l− 1, were 4.58, 2.74, 1.13, 2.33, 0.074, 0.16, 0.5 and 0.13 for SO4− 2, NO3− , Cl− , NH4+, Na+, K+, Ca2+ and Mg2+, respectively. The precision and bias of the analysis for major ions and trace metals were determined from quality control check samples prepared in the laboratory. Ten replicate measurements of each element were made. The precision (standard deviation) was 0.48, 0.072, 0.04, 0.005, 0.014, 0.067,

64

A. Báez et al. / Atmospheric Research 86 (2007) 61–75

Table 1 Volume-weighted mean concentrations (VWMC), Standard Deviation of the VWMC (SDVWMC), Minimum (Min) and Maximum (Max) concentrations (μg l− 1) of trace metals in rainwater collected for the period 2001–2002 Metal

Al Cd Cr Mn Ni Pb V

VWMC ± SDVWMC

Minimum

Maximum

Soluble

Insoluble

Soluble + insoluble

Soluble

Insoluble

Soluble + insoluble

Soluble

Insoluble

Soluble + insoluble

15.3 ± 1.66 0.37 ± 0.04 0.26 ± 0.03 8.34 ± 1.29 2.98 ± 0.57 1.58 ± 0.28 4.78 ± 0.96

35.4 ± 4.67 0.04 ± 0.01 0.26 ± 0.02 1.30 ± 0.16 0.39 ± 0.06 0.90 ± 0.18 0.35 ± 0.06

50.7 ± 5.40 0.41 ± 0.04 0.52 ± 0.03 9.64 ± 1.35 3.37 ± 0.60 2.48 ± 0.33 5.13 ± 0.97

3.1 0.04 0.19 0.23 0.39 0.57 1.56

0.08 0.001 0.005 0.006 0.01 0.014 0.039

3.1 0.04 0.19 0.23 0.39 0.57 1.56

100 4.5 1.2 40.3 12.9 9.3 42.1

367 0.93 2.38 18.97 4.29 22.1 4.46

398 4.62 2.57 42.8 14.0 22.7 42.4

0.006 and 0.032 mg l− 1 for SO42−, Cl−, NO3−, Na+, K+ Ca2+, Mg2+ and NH4+, respectively. The bias, in the same order, was 0.2, 0.52, − 0.71, 6.3, − 3.1, 2, 2.7 and 8.9%. Graphite furnace absorption atomic spectrometry (GFAAS) was used to analyze trace metals from soluble and insoluble fractions. A GBC double beam atomic absorption spectrometer, model 932AA, coupled with a System 3000 graphite furnace accessory, a GF3000 graphite power supply and a PAL3000 furnace auto sampler, was used. A deuterium lamp, for background correction, pyrolytically coated graphite tubes, boosted discharge hollow cathode lamps (Photron Super lamp) for cadmium, nickel and lead analysis at 228.5, 232.0 and 217.0 nm respectively, and hollow cathode lamps (Photron) for Cr, Mn, V, Al, at 357.9, 279.5, 318.5, and 309.3 nm, respectively, were used. Calibration standards were prepared with the same acid concentration as the samples from certified standards of each metal (highpurity standards traceable to NIST). Cross check methods of standard additions were used. The analytical detection limits, in μg l− 1, were: 0.07, 0.38, 0.78, 1.14, 3.12, 0.46 and 6.2 for Cd, Cr, Ni, Pb, V, Mn, and Al, respectively. However, for insoluble particles, the final volume of the samples was 25 ml (see Sample treatment), and therefore, the detection limits, expressed in μg l− 1, became in the same order: 0.0018, 0.01, 0.02, 0.029, 0.078, 0.012, and 0.155. The precision (standard deviation) was 0.025, 0.22, 0.17, 0.27, 2, 0.27, and 1.54 μg l− 1 for Cd, Cr, Ni, Pb, V, Mn, and Al, respectively. The bias, in the same order, was − 0.31, 2.49, − 3.76, 0.21, −1.6, − 5.2, and 2.6%. 2.5. Quality control Blanks of all glass and plasticware were analyzed. Concentrations of blanks were below the detection limits of major ions and trace metals. The quality of analysis of each sample was checked for ion balance and specific conductance calculations,

defined by Peden et al. (1986) as: Ion Percent Difference = [(Cations − Anions) / Cations + Anions)] × 100 and Conductance Percent Difference = [(Calculated Conductance − Measured Conductance) / (Measured Conductance)] × 100, respectively. Only 2 samples were excluded because they did not satisfy the criteria for ion balance and ion percent difference, therefore, 71 samples were used for statistical treatment. Values below the detection limit of trace metals and major ions were considered as one-half of the respective detection limit for statistical calculations. 3. Results and discussion The volume-weighted mean concentrations (VWMC), standard deviations of the VWMC (SDVWMC), minimum and maximum concentrations of trace metals in soluble and insoluble rain fractions and for major ions in soluble fractions, are shown in Tables 1 and 2, respectively, for the period 2001–2002 (whole sampling campaign). It is observed that Al presented the highest

Table 2 Volume weighted mean concentrations (VWMC), Standard Deviation of the VWMC (SDVWMC), Minimum (Min) and Maximum (Max) concentrations (μg l− 1) of ions in rainwater collected for the period 2001–2002 Ion

VWMC

SDVWMC

Minimum

Maximum

SO2− 4 −

61.94 9.56 42.62 7.00 2.16 26.44 2.46 92.35 8.34 27.36 5.08

6.42 1.67 4.74 1.56 0.75 5.92 0.52 8.20 1.79 6.69 5.74

10.63 2.25 5.71 0.03 0.08 0.25 0.07 28.57 0.12 0.00 4.11

311.88 136.62 202.86 73.74 16.83 437.90 23.69 413.57 77.62 485.70 6.92

Cl NO−3 Na+ K+ Ca2+ Mg2+ NH+4 H+ HCO−3 pH

A. Báez et al. / Atmospheric Research 86 (2007) 61–75

concentration in the insoluble fraction followed by Mn, Pb, Ni, V, Cr and Cd. In the soluble fractions, the concentrations of trace metals decreased in the order Al, Mn, V, Ni, Pb, Cd and Cr. Regarding major ions, NH4+ presented the higher VWMC, followed by SO42−, NO3− HCO3−, Ca2+, Cl−, Na+, Mg2+ and K+. Mexico City is located in a basin at 2200 masl, with predominant winds from the north–northeast, where most industries are located, in the rainy season from mid-May to the end of October (Jáuregui, 2000). These winds transport great amounts of air pollutants emitted by industries such as ferrous and non-ferrous smelter facilities, glass plants, bricks and ceramic factories, and thermoelectric power plants, to southern Mexico City where the Universidad Nacional Autónoma de México is located, resulting in the high concentrations of ionic species and trace metals observed in rainwater collected at this place. More than three million vehicles and around five thousand industries emit more than 57 ton day− 1 (20,686 ton year− 1) of particulate matter smaller than 10 μm (PM10), 18 ton day− 1 (6622 ton year− 1) of particulate matter PM2.5, 18 ton day− 1 (6646 ton year− 1), and 493 ton day− 1 (179,996 ton year− 1) of the precursors SO2 and NOx, respectively, and 17,514 ton year− 1 of NH3 (GDF, 2006a). Regarding trace metals, 4, 6, 37, 8 and 18 ton year− 1 of Cd, Cr, Mn, Ni and Pb, respectively, are emitted by chemical industries, steel manufacture, mobile sources and other sources (GDF, 2006b). The high concentrations of Al present in rainwater may be attributed to two sources a) the relative abundance in crustal or geological material (Vega et al., 2001; Al-Momani, 2003) and b) currently, most of the vehicle motors are made out of aluminum, then the wear of such motors emits substantial amounts of this metal to the atmosphere. More than 3 million of motor vehicles circulated in Mexico City, of which, more than 40% are new cars with aluminum motors (personal communication from auto-motor dealers). There are also several industries that process Al, so emissions of this metal are also expected from these industries. The scavenging of these pollutants is favored because of the convective nature of clouds in Mexico City during the rainy season. The VWMC of major ions in rainwater collected in UNAM, were compared with those reported in other urban areas (Table 3). The VWMC of NH4+ observed in this study was the highest, possibly because of the combined emissions of industries and mobile sources in the Metropolitan Zone of Mexico City and agricultural activities near Mexico City. Regarding SO42−, its VWMC is comparable with that observed in Seoul and higher than in Hong Kong. It is interesting to note that the VWMC of

65

Table 3 Comparison of VWMC of major ions (μeq l−1) in wet precipitation between Mexico City and other locations Element Ankara, Seoul, Turkey a South Korea b

Hong Piracicaba, Kong c Southeast Brazil d

Tirupati, This India e study

SO2− 4 Cl− NO−3 Na+ K+ Ca2+ Mg2+ NH+4 H+

48.6 37.6 18.9 31.8 2.2 15.3 7.8 – 26.3

127.96 33.91 40.84 33.08 33.89 150.66 55.51 20.37 0.34

a b c d e

52.1 41.1 35.5 23.0 3.55 132 19.7 66.7 0.068

70.9 18.2 29.9 10.5 3.5 34.9 6.9 66.4 21.3

18.7 7.0 16.6 2.7 2.9 5.3 2.3 17.1 33.0

61.94 9.56 42.62 7.00 2.16 26.44 2.46 92.35 8.34

Kaya and Tuncel (1997). Lee et al. (2000). Tanner and Wong (2000). Lara et al. (2001). Mouli et al. (2005).

SO42− and Ca2+ were much higher in Tirupati, India, due to the large contribution of soil dust. In relation to other sites the VWMC of Ca2+ was much lower than in Ankara because of the low rain amount in Ankara, comparable to that in Seoul; and higher than in Southeast Brazil and Hong Kong, obviously because rain amounts in these regions are noticeably higher than in the MZMC. Dust particles are the main sources of alkaline ions in the MZMC. The VWMC of NO3− is comparable to that in Ankara and Tirupati and higher than in Seoul, Hong Kong, and Southeast Brazil because of the numerous mobile sources and poor dispersion of air pollutants in the MZMC. The precursor gases of SO42− and NO3− are emitted mainly by mobile sources. Table 4 shows the comparison of trace metals in both soluble and insoluble fractions in wet precipitation among Mexico and other countries. From this table it is observed that Total Al (soluble plus insoluble) in Ankara was the highest value compared with the other countries. In general, the trace metal values measured in this country were higher than in Mexico with the exception of V, on the other hand, the Al reported in Singapore was lower than in Ankara but higher than in Mexico, Jordan and Brazil. With respect to Cd, Cr, Mn, Pb the values reported for Singapore were the highest compared with the other countries. Al and Mn were the only metals reported for Guaíba Brazil. The high values reported for Ankara Turkey and Singapore suggest that these countries have severe air pollution problems. The sources of major ions and trace metals are further discussed in the next section.

66

A. Báez et al. / Atmospheric Research 86 (2007) 61–75

Table 4 Comparison of VWMC of trace metals in soluble and insoluble fractions (μg l− 1) in wet precipitation between Mexico City and other locations Element

Al Cd Cr Mn Ni Pb V a b c d

Ankara, Turkey a Soluble

Insoluble

47 8.6 0.38 – 2.2 3.3 0.92

910 0.6 2.6 – 1.6 15.7 1.28

AlHashimya, Jordan b

Singapore c

18.44 0.33 1.62 2.78 3.86 3.37 3.54

386 11.8 6.2 48.6 – 26.9 –

Guaíba, Brazil d 16.9 – 1.4 – – –

This study Soluble

Insoluble

Soluble + insoluble

15.3 0.37 0.26 8.34 2.98 1.58 4.78

35.4 0.04 0.26 1.30 0.39 0.90 0.35

50.7 0.41 0.52 9.64 3.37 2.48 5.13

Kaya and Tuncel (1997). Al-Momani et al. (2002). Hu and Balasubramanian (2003). Migliavacca et al. (2005).

3.1. Back trajectory analysis The VWMC of major ions and trace metals in rainwater and number of samples for each sector are shown in Tables 5 and 6, the relation to wind directions obtained by air mass back trajectory analysis are also presented in these tables. The individual concentrations of SO42− (as a chemical tracer of anthropogenic pollution) of Al and Ca2+ (as chemical tracers of crustal particles) and trace metals, were associated with the corresponding air mass back trajectories calculated by the NOAA HYSPLIT model (Hybrid Single-Particle Lagrangian Integrated Trajectory Model) (Draxler and Rolph, 2003). Air mass back trajectories were calculated for 1000 and 3000 m above ground level (MAGL). NOAA trajectories were calculated for 2001 and 2002. Since the UNAM is sur-

rounded by intense anthropogenic emission sources, the air mass back trajectory sectors were subdivided into four equally-sized sectors, that is, 1 to 90, 90 to 180, 180 to 270 and 270 to 360°. Fig. 2 shows some examples of air mass back trajectories randomly selected during the rainy season. Fig. 3 shows one air mass back trajectory for each of the four studied sectors during the same season. In these tables, the variabilities of ions and trace metals in function of the four sectors are observed and corresponded to 1000 and 3000 MAGL trajectories. For instance, NH4+ was the most abundant ion in the four sectors in both back trajectories, followed by SO42−. With respect to trace metals, Al was the most abundant metal in all sectors following by Mn. The low sample number obtained in sectors 3 and 4, agrees with the synoptic meteorological conditions that

Table 5 Volume weighted mean concentrations of major ions (μeq l− 1) for four wind direction sectors in rainwater collected at the University of Mexico in 2001 and 2002 Ion

N SO2− 4 Cl− NO−3 Na+ K+ Ca2+ Mg2+ NH+4 H+

Air mass trajectory sector at 1000 m AGL a

Air mass trajectory sector at 3000 m AGL

N–E

E–S

S–W

W–N

N–E

E–S

S–W

W–N

Sector 1

Sector 2

Sector 3

Sector 4

Sector 1

Sector 2

Sector 3

Sector 4

1°–90°

90°–180°

180°–270°

270°–360°

1°–90°

90°–180°

180°–270°

270°–360°

24 54 9.8 34.1 5.11 3.3 32.6 3.0 87.9 8.96

27 51.7 10.3 55.3 7.38 1.93 35.8 3.3 110 6.91

7 73.5 6.14 53.9 3.13 1.2 20.03 2.70 90.2 11.79

13 84.5 8.66 43.4 13.7 1.65 24.6 2.83 96.2 7.33

22 55.4 7.88 37.3 4.63 1.94 28.9 2.95 97.2 8.71

33 52.6 11.6 51.7 7.18 1.76 32.4 3.0 101.3 8.02

4 63.5 7.54 64.4 3.97 1.52 20.8 2.87 101.2 15.71

12 77.5 8.14 41.5 11.9 4.34 37.6 3.53 94.8 5.33

N = sample number. a Meters above ground level.

A. Báez et al. / Atmospheric Research 86 (2007) 61–75

67

Table 6 Volume weighted mean concentrations of trace metals (soluble + insoluble fractions) in μg l− 1 for four wind direction sectors in rainwater collected at the University of Mexico in 2001 and 2002 Metal

N Al Cd Cr Mn Ni Pb V

Air mass trajectory sector at 1000 m AGL a

Air mass trajectory sector at 3000 m AGL a

N–E

E–S

S–W

W–N

N–E

E–S

S–W

W–N

Sector 1

Sector 2

Sector 3

Sector 4

Sector 1

Sector 2

Sector 3

Sector 4

1°–90°

90°–180°

180°–270°

270°–360°

1°–90°

90°–180°

180°–270°

270°–360°

29 62.0 0.37 0.43 11.5 4.56 2.52 4.03

31 39.2 0.56 0.50 8.51 1.98 2.49 4.34

7 76.2 0.34 0.72 9.04 5.43 2.03 2.22

14 47.1 0.68 0.71 9.79 3.56 2.57 11.4

27 57.1 0.46 0.43 10.4 3.20 3.19 5.87

37 53.4 0.54 0.57 8.74 2.48 2.23 3.61

4 49.8 0.21 0.53 9.26 6.26 2.39 1.70

13 37.2 0.50 0.56 11.3 4.98 2.00 8.25

N = sample number. a Meters above ground level.

prevail in Central Mexico during the rainy season; trade winds have a consistent component from the east, that is, winds blow between north and southeast most of the time during this season. Calcium in rainwater samples comes from windblown dust particles on sedimentary soils northeast of Mexico City. These particles neutralize acids in rainwater as it was suggested by the concentrations of H+ and Ca2+ which indicate a clear neutralization process. For example, at 1000 MAGL, the highest concentration of Ca2+ was observed for sector 2 (35.8 μeq l− 1) and the lowest concentration of H+ was also observed for this sector (6.9 μeq l− 1). Something similar was observed at 3000 MAGL (see Table 5). There was also a reasonably well correspondence with the physical characteristics of the areas located in the different sectors. Although sample number is too low to achieve any conclusions; it is nevertheless interesting to remark that the lowest concentrations of Ca2+ observed at both levels for sector 3, agree with the fact that wooded areas (which partially prevent dust particles from being blown by wind) predominate between south and west of UNAM, and the highest concentrations of this ion observed at both levels for sector 2, agree with the fact that industrial areas are situated between northwest and northeast of UNAM, although some wooded areas also lie between west and northwest of UNAM. It is important to mention the Ca2+ is mobilized in wooded areas by means of wind that transport Ca2+ particles from erode zones that have been deforested. Concerning trace metals, their concentrations were not clearly related to wind direction, as it can be seen in Table 6. This result was not expected because most of the sources of trace metals (except those of Al) are

supposed to be anthropogenic, but surprisingly, the air mass back trajectory analysis for both levels showed that the highest concentrations of only Cd, Mn, Pb and V were observed when winds blew from sectors 1 and 2, where lie most of urban and industrial pollution sources. The highest concentrations of Cr were observed for sector 2 at 3000 MAGL. The analysis at 3000 MAGL gave expected results because the highest concentrations of Al were observed for sector 1, where large extensions of barren soils lie. An attempt was made to associate trace metals and major ion concentrations with surface wind data obtained by two meteorological stations, however, the results were not congruent with what was expected, almost undoubtedly because surface winds in the MZMC use to be variable and weak. 3.1.1. Sources of anthropogenic species There are around 30,000 small, medium and heavy industries located in the northern part of the Metropolitan Zone of Mexico City. Among them, there are ferrous and non-ferrous smelters, glass producers, motor vehicle manufacturers (one smelter), a car assembly plant, lime, brick, ceramic, cement and tire factories. In addition there exist 2 natural gas-fired power plants, although with less significance regarding the other sources like particles re-suspension and intensive motor vehicles traffic. So, SO2, NOx, Cl−, NH3, particulate matter and heavy metals emissions are to be expected. Significant emissions, such as Cu, Cd, Pb, Mn, and Zn emanate from single sources like large smelters (Ross, 1986; Szefer and Szefer, 1986; Nriagu and Pacyna, 1989). Ni and Mn are mainly released from oil-fired furnaces and ferroalloys (Szefer and Szefer, 1986; Nriagu, 1989). Smelter processes emit vanadium,

68

A. Báez et al. / Atmospheric Research 86 (2007) 61–75

chromium and nickel. Vanadium, in the form of vanadium oxide, is a component of a special kind of steel that is used in the production of automobile parts, springs and ball bearings. It is also used to make rubber, plastics, ceramics and other chemicals (ATSDR, 1995). Several components of motor vehicles are Cd alloys,

and the manufacturing and disposal of Cd batteries and the wearing of vehicle tires are potential sources of Cd emissions (Scudlark et al., 1994; Mugica et al., 2002). The manufacture of lead batteries, the disposal of used batteries, and the production of lead tetraethyl as an anti-knock treatment all produce Pb emissions.

Fig. 2. Some air–mass back trajectories randomly selected observed during the rainy season for 2001 and 2002.

A. Báez et al. / Atmospheric Research 86 (2007) 61–75

69

Fig. 2 (continued).

SO42− and NO3−, suggests that these pollutants are probably produced by the same source of combustion processes using fuel oil with sulfur content that occur in industry and thermoelectric power plants. In addition to a secondary aerosol formation process, which is associated with their precursor gases SO2 and NOx. NOx is

also emitted by car emissions (Pio et al., 1996) and the domestic burning of gas in boilers and stoves. Regarding NH4+, there are many sources of ammonia in the MZMC the most important are cleaning items, industries, excrements and urine of domestic pets and agricultural activities. As it is widely known, NH3 reacts

70

A. Báez et al. / Atmospheric Research 86 (2007) 61–75

Fig. 3. Some air–mass back trajectories observed during the rainy season in 2001 and 2002 for the selected four sectors. Backward trajectories ending at: (a) 00 UTC 06 Sep 02 for sector 1, (b) 00 UTC 03 Jul 01 for sector 2, (c) 00 UTC 13 Jun 01 for sector 3, and (d) 00 UTC 04 Jun 02 for sector 4. Squares and triangles indicate 1000 and 3000 m above ground level, respectively.

EFcrust ¼ ðCx =CMg Þsample =ðCx =CMg Þcrust

significant crustal source, this element is referred to as a non-enriched element (NEE). If an average EFcrust value is N10, it is considered that a significant proportion of an element has a non-crustal source, this element is referred to as an anomalously enriched element (AEE) (Chester et al., 1999). Aluminum, K+, and Ca2+ were the only elements with a crustal source being considered as NEE. Cadmium, Pb, V, Ni, Mn, and Cr, in decreasing order, presented EFcrust values N 10 and are considered as AEEs indicating a non-crustal source. In general, Pb, V, Ni, Mn, and Cr are associated with fine particles (b 1 μm), which are generated as high-temperature combustion condensates

Where: (Cx/CMg)sample is the ratio of the concentrations of an element X and Mg in the rainwater sample and (CxCMg)crust is the concentration of the same element with respect to crustal material. The EFc's were calculated by using average concentrations from crustal elements from Mason (1966) and the average EFc's of elements from rainwater (Fig. 4). For trace metals, the soluble plus insoluble rainwater fractions were considered. By convention, an arbitrary average EFcrust value b 10 was chosen to indicate that an element in rainwater has a

Fig. 4. Average enrichment factors for Al, Cd, Cr, Mn, Ni, Pb, V, Na+, K+and Ca2+.

with H2SO4 and HNO3 in the atmosphere to yield (NH4)2SO4 and NH4NO3. 3.2. Enrichment of metals Crustal enrichment factors (EFcrust), of an element compared with relative abundance of that element in crustal material, were calculated to evaluate the contribution of non-crustal sources to elements concentrations in rainwater. EFc's were calculated using the following equation (Kaya and Tuncel, 1997):

A. Báez et al. / Atmospheric Research 86 (2007) 61–75 Table 7 Factor loading normal Varimax extraction Variable

Factor 1

Factor 2

Factor 3

Al Cd Cr Mn Ni Pb V SO2− 4 Cl− NO−3 Na+ K+ Ca2+ Mg2+ NH+4 H+ %Total variance

0.02 0.30 0.16 0.28 0.07 − 0.10 0.56 0.74 0.06 0.46 0.29 0.87 0.96 0.86 0.75 − 0.14 27.2

0.81 0.07 0.53 0.70 0.78 0.22 0.13 0.29 0.13 0.08 −0.29 0.12 0.08 −0.007 0.11 0.09 14.6

0.03 0.43 0.36 0.17 − 0.12 − 0.16 0.24 0.45 0.009 0.73 0.41 − 0.08 − 0.04 0.09 0.24 0.85 13.4

71

distance to the seashore and the high mountains that surround the MZMC. Al-Momani et al. (2002) have also stated that there is uncertainty in defining the composition of the crustal source material. Kaya and Tuncel (1997), Halstead et al. (2000) and Kim et al. (2000), found that Cd, Cu, Pb and Zn were enriched in rainwater relative to crustal material, thus a significant anthropogenic contribution of these elements was suggested. 3.3. Factor analysis

Principal components (bold numbers are significant at N0.5).

and injected into the boundary layer by smoke-stacks (Lindberg, 1982; Scudlark et al., 1994). The high enrichment factor of Na+, (the second highest, Fig. 4), is difficult to explain in terms of anthropogenic sources. Enrichment factors are based on a “standard crust composition”, thus they must be interpreted with caution in areas where the chemical composition of dust particles differs from that of the standard crust. In the case of this study, as far as we know, there are no anthropogenic sources in the MZMC capable of raising the enriching factor of sodium ion to the high value shown in Fig. 4. Obviously, sea spray is ruled out as a source because of the

To assess relationships between concentrations of the studied elements, factor analysis (Principal Component Analysis) and Pearson's correlation were applied. Table 7 shows the factor loadings normalized with the Varimax rotation. The Varimax rotation is aimed to maximize the variances of the squared normalized factor loadings across variables for each factor and makes the interpretation easier. This is the most used method (StatSoft, Inc., 2003). From the principal component analysis only three factors were chosen that explained the 55.4% of the total variance. Factor 1 explained 27.2% of the total variance with high loadings for Ca2+, K+, Mg2+, moderate loadings for SO42− and NH4+ and a low loading for V. This factor indicates a soil crustal contribution for Ca2+, K+, Mg2+, with a significant crustal source for V, but mainly anthropogenic, as it is indicated by the enrichment factors. The loadings of SO42− and NH4+ suggest that these ions come from anthropogenic sources through their precursors SO2 and NOx. These ions significantly correlated because their

Table 8 Pearson correlations between the different total trace metals (soluble plus insoluble rain fractions) and ions at the University of Mexico for period 2001–2002 Cd

Cr

Mn

Al 0.187 0.462⁎ 0.516⁎ Cd 0.311⁎ 0.181 Cr 0.349⁎ Mn Ni Pb V SO2− 4 Cl− NO−3 Na+ K+ Ca2+ Mg2+ NH+4 ⁎Significant at p b 0.05.

SO2− 4

Ni

Pb

V

0.559⁎ 0.184 0.322⁎ 0.416⁎

0.227 0.166 0.104 0.156 0.271⁎

0.133 0.228 0.383⁎ 0.491⁎ 0.194 0.434⁎ 0.235⁎ 0.466 0.159 0.205 0.014 − 0.004 0.639⁎

Cl−

NO−3

0.316⁎ 0.132 0.323⁎ 0.052 0.094 0.066 0.121 0.119

0.137 −0.052 0.086 0.102 0.100 0.536⁎ 0.231 0.176 0.329⁎ 0.232 0.397⁎ 0.078 0.203 0.232 0.204 0.316⁎ 0.113 0.272⁎ 0.254⁎ 0.235⁎ 0.072 −0.119 0.156 0.121 0.086 0.041 −0.046 −0.062 −0.029 −0.056 0.349⁎ 0.172 0.307⁎ 0.584⁎ 0.336⁎ 0.659⁎ 0.200 0.579⁎ 0.735⁎ 0.567⁎ 0.105 0.041 0.084 0.118 0.206 0.378⁎ 0.313⁎ 0.411⁎ 0.523⁎ 0.196 0.138 0.425⁎ 0.825⁎ 0.819⁎ 0.787⁎

Na+

K+

Ca2+

Mg2+

NH+4

H+

0.177 0.462⁎ 0.321⁎ 0.359⁎ 0.106 0.152 0.510⁎ 0.749⁎ 0.080 0.676⁎ 0.267⁎ 0.480⁎ 0.730⁎ 0.552⁎

0.050 0.162 0.205 0.076 − 0.039 − 0.069 0.124 0.297⁎ 0.016 0.488⁎ 0.149 − 0.056 − 0.124 0.021 − 0.024

72

A. Báez et al. / Atmospheric Research 86 (2007) 61–75

caution must be taken in interpreting correlation coefficients in rainwater data since concentrations may vary because of selective scavenging of particles from the atmosphere (Kaya and Tuncel, 1997). It is noteworthy how H+ did not correlate with any of the metals, possibly as the result of the high concentration of alkaline particles in the atmosphere of Mexico City, which makes most rains alkaline (pH above 5). Therefore, alkaline rains present a different chemical composition, or in other words, the concentration of metals did not vary according to H+ concentration. The significant correlation of H+ with SO42− and NO3−, with the highest correlation between H+ and NO3−, and the significant and high correlation of SO42− and NO 3− with Ca 2+ , Mg 2+ , and NH 4+ , suggest that neutralization reactions occurred. Pb and Cd did not correlate, possibly due to the decrease of Pb levels in rainwater as the result of Pb reduction in fuels. Smelter emissions could be the main source of Pb. 3.5. Neutralization effect of alkaline ions

+ − 2− Fig. 5. Linear regressions of (NO−3 + SO2− 4 ) vs. NH4 , (NO3 + SO4 ) vs. 2+ + 2+ − 2− + Ca and (NO3 + SO4 ) vs. (H + Ca +NH4).

concentrations in rainwater varied in the same manner. Factor 2 indicates high loading for Al, Mn and Ni, and a low loading for Cr, explaining 14.6% of the total variance. This factor indicates a possible contribution of anthropogenic sources but a significant crustal contribution for Al. Also, the significant correlation between them suggests a common origin. Factor 3 explained 13.4% of the total variance with a high loading for H+ and NO3− . These results indicate a contribution of anthropogenic sources. The significant correlation between these ions, suggests that they can be attributed to reactions of this precursor NOx. Although Cd has a small load the high value found in the EFc indicate an important contribution from anthropogenic sources. 3.4. Correlations Table 8 shows the matrix of linear correlations among ions and trace metals (soluble plus insoluble rain fractions). Although correlation values between species are frequently used to infer the sources of pollutants,

The high ammonium ion concentration (92.3 μeq l− 1) seems to indicate that this ion was the main neutralizing compound, since Ca2+ concentration was much lower (26.4 μeq l− 1). The ratio H+/(SO42− + NO3−) would be united if only sulfuric and nitric strong acids were present (Kaya and Tuncel, 1997). Therefore, the ratio 0.08 observed at UNAM, clearly indicates that a considerable neutralization occurred. In Figs. 5a and b are shown the linear regression of (NO3− + SO42−) with NH4+ and Ca2+. The correlation coefficient between these variables was 0.701 and 0.475, respectively indicating that these ions are neutralizing rainwater acidity.

Table 9 Trace metal solubility in rainwater collected at the University of Mexico for the period 2001–2002 Metal Solubility (%) Whole data

Samples with pH b5

Samples with pH N5

Average Standard Average Standard Average Standard deviation deviation deviation Al Cd Cr Mn Ni Pb V

42.56 80.52 46.27 79.69 78.63 58.67 85.81

30.40 21.29 25.41 18.54 20.9 29.07 14.36

49.19 89.1 45.47 79.91 81.54 60.79 87.37

28.36 9.53 21.52 18.35 14.52 26.5 15.38

40.63 79.14 48.25 79.36 80.04 58.76 86.08

32.06 22.29 27.17 18.41 21.95 30.15 14.07

A. Báez et al. / Atmospheric Research 86 (2007) 61–75 Table 10 Wet-deposition fluxes in mg m− 2 period− 1 for trace metals in rainwater collected in the period 2001–2002 Metal

VWM

Loading a

1063 b Al Cd Cr Mn Ni Pb V a b

50.73 0.41 0.52 9.64 3.37 2.48 5.13

53.91 0.44 0.55 10.25 3.58 2.64 5.45

mg m− 2 period− 1. Amount of rainfall in mm.

If H2SO4, HNO3, NH4+ and CaCO3 are the only species involved in acidity precipitation, it is expected a linear relation between (H+ + Ca2+ + NH4+) and (SO42− + NO3−). This relation is shown in Fig. 5c, the correlation coefficient between variables is 0.836 indicating that the acidity is caused by H2SO4 and HNO3 and neutralized mainly by NH4+ and CaCO3. The contribution of other anions and cations on the neutralization process and hence, on the acidity of the precipitation is not significant. Neutralizing factors (NF) were calculated to quantify their neutralization effect. The ratios of NH4+ and Ca2+ to NO3− + SO42− were used to calculate average NF. The obtained values were 0.99 and 0.27, respectively. These results indicate that NH4+ neutralizes a large fraction of rainwater acidity. 3.6. Solubility Table 9 shows the concentration of each metal in the soluble rain fraction expressed in terms of percentage of its concentration in the soluble plus insoluble rain fraction (percent solubility) for pH lower and higher than 5. It is also noteworthy that the percent solubility for all the metals is practically the same for both pH ranges and for the whole data (pH lower and higher than 5). The explanation to this could also be the high concentration of alkaline particles observed in Mexico City that neutralize an important fraction of hydrogen ions, and consequently, the H+ concentration (pH = 4.11) was too low to increase the concentration of metals in the soluble rain fraction significantly. Some authors have associated the pH of rain with the dissolution of metals contained in aerosols. The high acidity rainwater (pH values as low as 3) could make some trace metals highly soluble (Chester et al., 1997).

73

3.7. Wet deposition of trace metals Table 10 shows the wet-deposition fluxes (mg m− 2l for each metal. It is clear that Al presented the highest loading, as it was expected because it is an abundant component of terrestrial crust, and whose concentration in the atmosphere is increased by the high concentration of dust as a consequence of deforestation in areas that surround Mexico City. In this table, period refers to the rainy season (from mid-May to the end of October) of 2001 and 2002. 4. Conclusions The inconsistencies observed when it was attempted to couple ionic, and especially trace metal concentrations in rainwater, with air mass back trajectory analysis, occurred due to the complex topography of Mexico Valley and surrounding areas. Therefore, a detailed analysis of wind flow patterns by means of reliable soundings at various altitudes above ground level at various zones of Mexico Valley, would probably have been more useful to perform a high-resolution air mass back trajectory analysis, and consequently, the coupling of ionic and trace metal concentrations with air mass back trajectory analysis would have been improved. High concentrations of Mn, V and Ni were found in Mexico City in the period 2001–2002. Trace metals were emitted mainly by anthropogenic sources. This was supported by the EFc factor values, which showed a non-crustal origin of these metals. Aluminum presented the highest concentration in the soluble and insoluble rain fractions indicating that an important quantity of this metal had a crustal or geological origin. Magnesium was the only element with a significant crustal source being considered as a non-enriched element. Cadmium, Pb, V, Ni, Ca2+, Mn, Cr, Na+ and K+ were considered as anomalously enriched elements indicating a non-crustal source. The factor analysis (principal component analysis) indicated that Cr, Mn, Ni, V, Ca2+, K+, Mg2+, SO42− and NH4+ came from anthropogenic and soil sources, with a significant crustal source for Mg2+ and Al. On the contrary, Pb, Cd, H+ and NO3− came solely from anthropogenic sources. The high concentration of alkaline particles, usually present in the atmosphere of Mexico City, neutralizes an important fraction of H+ ion; therefore, this is possibly the reason why the solubility of trace metals did not depend on rainwater pH. The results of metal loadings suggest that it is necessary a continuous monitoring of metals deposition

74

A. Báez et al. / Atmospheric Research 86 (2007) 61–75

fluxes in a long-term to evaluate annual and seasonal atmospheric fluxes. Acknowledgements The authors gratefully acknowledge the NOAA Air Resources Laboratory (ARL) for the provision of the HYSPLIT transport and dispersion model and/or READY website (http://www.arl.noaa.gov/ready.html) used in this publication. The authors wish to thank Dr. Víctor Magaña Rueda and Gustavo Vázquez Cruz for providing the meteorological data of the “Red de Estaciones Meteorológicas PEMBU”, and Jorge Escalante, Wilfrido Gutiérrez, and Alfredo Rodríguez for the maintenance of the sampling equipment, and Calixto Cuevas for his assistance in the laboratory. References Al-Momani, I.F., 2003. Trace elements in atmospheric precipitation at Northern Jordan measured by ICP-MS: acidity and possible sources. Atmos. Environ. 37, 4507–4515. Al-Momani, I.F., Ya'qoub, A.-R.A., Al-Bataineh, B.M., 2002. Atmospheric deposition of major ions and trace metals near an industrial area, Jordan. J. Environ. Monit. 4, 985–989. Astel, A., Mazersiki, J., Polkowska, Z., Namieœnik, J., 2004. Application of PCA and time series analysis in studies of precipitation in Tricity (Poland). Adv. Environ. Res. 8, 337–349. ATSDR (Agency for Toxic Substances and Disease Registry), 1995. ToxFAQs for Vanadium and Compounds. U.S. Department of Health and Human Services, Washington, D.C. CAS # 7440-62-2. Báez, A.P., González, O., Solorio, F., Belmont, R., 1980. Determinación de plomo, cadmio y cromo en la precipitación pluvial en algunos lugares de la República Mexicana. Medio Ambiente 2, 35–46. Barrie, L.A., Lindberg, S.E., Chan, W.H., Ross, H.B., Arimoto, R., Church, T.M., 1987. On the concentration of trace metals in precipitation. Atmos. Environ. 21, 1133–1135. Chester, R., Nimmo, M., Corcoran, P.A., 1997. Rain water–aerosol trace metal relationships at Cap Ferrat: a costal site in the Western Mediterranean. Mar. Chem. 58, 293–312. Chester, R., Nimmo, M., Preston, M.R., 1999. The trace metal chemistry of atmospheric dry deposition samples collected at Cap Ferrat: a coastal site in the Western Mediterranean. Mar. Chem. 68, 15–30. Draxler, R.R., Rolph, G.D., 2003. HYSPLIT (HYbrid Single-Particle Lagrangian Integrated Trajectory) Model Access Via NOAA ARL READY Website. NOAA Air Resources Laboratory, Silver Spring, MD. http://www.arl.noaa.gov/ready/hysplit4.html. Flores-Estrella, H., Yussim, S., Lomnitz, C., 2007. Seismic response of the Mexico City Basin: a review of twenty years of research. Nat. Hazards 40, 357–372. Galloway, J.N., Thornton, J.D., Norton, S.A., Volchok, H.L., McLean, R.A., 1982. Trace metals in atmospheric deposition: a review and assessment. Atmos. Environ. 16, 1677–1700. GDF, 2006a. Inventario de emisiones de la Zona Metropolitana del Valle de México, 2004. Gobierno del Distrito Federal, Secretaría del Medio Ambiente, Dirección General de Gestión Ambiental del Aire, (8 de mayo de 2006). www.sma.df.gob.mx.

GDF, 2006b. Inventario de contaminantes tóxicos del aire en la ZMVM, 2004. Gobierno del Distrito Federal, Secretaría del Medio Ambiente, Dirección General de Gestión Ambiental del Aire, septiembre, 2006. Halstead, M.J.R., Cunninghame, R.G., Hunter, K.A., 2000. Wet deposition of trace metals to a remote site in Fiordland, New Zealand. Atmos. Environ. 34, 665–676. Howard, C.R., Brunette, R., Gürleyük, H., 2004. Determination of Arsenic, selenium, and various trace metals in rain waters. Proceedings of NADP 2004 Technical Committee Meeting and Scientific Symposium, September 21–23, 2004, Halifax, Nova Scotia, Canada. http://nadp.sws.uiuc.edu/lib/proceedings/ NADPpro2004.pdf. Hu, G.P., Balasubramanian, R., 2003. Wet deposition of trace metals in Singapore. Water Air Soil Pollut. 144, 285–300. Jáuregui, E., 2000. El clima de la Ciudad de México 1.4.1. In: y Valdés, Plaza, de C.V., S.A. (Eds.), Temas selectos de Geografía de México. Instituto de Geografía, UNAM, México. 131 pp. Jazcilevich, A., Fuentes, V., Jauregui, E., Luna, E., 2000. Simulated urban climate response to historical land use modification in the basin of Mexico. Clim. Change 44, 515–536. Kaya, G., Tuncel, G., 1997. Trace element and major ion composition of wet and dry deposition in Ankara, Turkey. Atmos. Environ. 31, 3985–3998. Khare, P., Goel, A., Patel, D., Behari, J., 2004. Chemical characterization of rainwater at a developing urban habitat of Northern India. Atmos. Res. 69, 135–145. Kim, G., Scudlark, R., Church, T.M., 2000. Atmospheric wet deposition of trace elements to Chesapeake and Delaware Bays. Atmos. Environ. 34, 3437–3444. Kulshrestha, U.C., Kulshrestha, M.J., Sekar, R., Sastry, G.S.R., Vairamani, M., 2003. Chemical characteristics of rainwater at an urban site of south-central India. Atmos. Environ. 37, 3019–3026. Lara, L.B.L.S., Artaxo, P., Martinelli, L.A., Victoria, R.L., Camargo, P.B., Krusche, A., Ayers, G.P., Ferraz, E.S.B., Ballester, M.V., 2001. Chemical composition of rainwater and anthropogenic influences in the Piracicaba River Basin, Southeast Brazil. Atmos. Environ. 35, 4937–4945. Lee, B.K., Hong, S.H., Lee, D.S., 2000. Chemical composition of precipitation and wet deposition of major ions on the Korean peninsula. Atmos. Environ. 34, 563–575. Lindberg, S.E., 1982. Factors influencing trace metal, sulfate and hydrogen ion concentrations in rain. Atmos. Environ. 16, 1701–1709. Luo, W., 2001. Wet-deposition fluxes of soluble chemical species and the elements in insoluble materials. Atmos. Environ. 35, 2963–2967. Mason, B., 1966. Principles of Geochemistry. John Wiley & Sons, Inc., New York. Migliavacca, D., Teixera, E.C., Pires, M., Fachel, J., 2004. Study of chemical elements in atmospheric precipitation in South, Brazil. Atmos. Environ. 38, 1641–1656. Migliavacca, D., Teixera, E.C., Wiegand, F., Machado, A.C.M., Sanchez, J., 2005. Atmospheric precipitation and chemical composition of an urban site, Guaíba hydrographic basin, Brazil. Atmos. Environ. 39, 1829–1844. Mouli, P.C., Mohan, S.V., Reddy, S.J., 2005. Rainwater chemistry at a regional representative urban site: influence of terrestrial sources on ionic composition. Atmos. Environ. 39, 999–1008. Mugica, V., Maubert, M., Torres, M., Muñoz, J., Rico, E., 2002. Temporal and spatial variations of metal content in TSP and PM10 in Mexico City during 1996–1998. Aerosol Sci. 33, 91–102. Nriagu, J.O., 1989. A global assessment of natural sources of atmospheric trace metals. Nature 338, 47–49.

A. Báez et al. / Atmospheric Research 86 (2007) 61–75 Nriagu, J.O., Pacyna, J.M., 1989. Quantitative assessment of worldwide contamination of air, water and soils by trace metals. Nature 333, 134–139. Peden, M.E., Bachman, S.R., Brennan, C.J., Demir, B., James, K.O., Kaiser, B.W., Lockard, J.M., Rothert, J.E., Sauer, J., Skowron, L.M., Slater, M.J., 1986. Development of Standard Methods for the Collection and Analysis of Precipitation. Contract CR810780-01. Illinois State Water Survey, Analytical Chemistry Unit, Champaign, IL. Pio, C.A., Castro, L.M., Cerqueira, M.A., Santos, I.M., Belchior, F., Salgueiro, M.I., 1996. Source assessment of particulate air pollutants measured at the southwest European coast. Atmos. Environ. 30 (19), 3309–3320. Ross, H.B., 1986. The importance of reducing sample contamination in routine monitoring of trace metals in atmospheric precipitation. Atmos. Environ. 20, 401–405. Roy, S., Négrel, P., 2001. A Pb isotope and trace element study of rainwater from the Massif Central (France). Sci. Total Environ. 277, 225–239. Scudlark, J.R., Conko, K.M., Church, T.M., 1994. Atmospheric wet deposition of trace elements to Chesapeake Bay: CBAD study year 1 results. Atmos. Environ. 28, 1487–1498.

75

Seinfeld, J.H., Pandis, S.N., 1998. Atmospheric Chemistry and Physics: From Air Pollution to Climate Change. John Wiley & Sons, Inc, New York. 1326 pp. StatSoft, Inc., 2003. Statistica (Data Analysis Software System), Version 6. www.statsoft.com2003. Stumm, W., Morgan, J.J., 1981. Aquatic chemistry: an introduction emphasizing chemical equilibria in natural waters, A WileyInterscience Publication, Second edition. John Wiley & Sons, New York, pp. 226–228. Szefer, P., Szefer, K., 1986. Some metals and their possible sources in rain water of the southern Baltic Coast, 1976 and 1978–1980. Sci. Total Environ. 57, 79–89. Tanner, P.A., Wong, A.Y.S., 2000. Soluble trace metals and major ionic species in the bulk deposition and atmosphere of Hong Kong. Water Air Soil Pollut. 122, 261–279. Valenta, P., Nguyen, V.D., Nornberg, H.W., 1986. Acid and heavy metal pollution by wet deposition. Sci. Total Environ. 55, 311–320. Vega, E., Mugica, V., Reyes, E., Sánchez, G., Chow, J.C., Watson, J.G., 2001. Chemical composition of fugitive emitters in México City. Atmos. Environ. 35, 4033–4039.