PRELIMINARY AIR POLLUTION MONITORING IN SAN MIGUEL,BUENOS AIRES

1

L. A. FAGUNDEZ, V. L. FERNÁNDEZ, T. H. MARINO, I. MARTÍN, D. A. PERSANO,M. RIVAROLA y BENÍTEZ, I. V. SADAÑIOWSKI, J. CODNIA and A. ZALTS*Instituto de Ciencias, Universidad Nacional de General Sarmiento, San Miguel, Provincia de

Buenos Aires, Argentina(* author for correspondence, e-mail: azalts@ungs.edu.ar)

..,(Received 24 August, 1999; accepted 2 May, 2000)

Abstract. Passive diffusion samplers were employed in San Miguel (Buenos Aires MetropolitanArea) for a preliminary air pollution monitoring. The highest loads were observed in downtown,compared with an urban background site. Total suspended particulate matter (TSPM) varied from0.257 to 0.033 mg cm-2 month'<": dust was examined for particle nature and size distribution. Asimilar trend was observed for nitrogen dioxide (N02) and TSPM spatial distribution, suggesting thattraffic is the major pollution source. Sulphur dioxide (S02) values were low and rather homogeneous.Levels for the investigated pollutants are below EPA's guide line values. Geographic (flat area, nearto Rio de La Plata) and climatologic factors (rainfalls and variable wind directions) contribute todisperse pollutants.

Keywords: air monitoring, N02, particulate matter, passive sampler, S02

1. Introduction

!

Pollution consequences are enhanced in large urban areas, because of the highnumber of inhabitants and the location of main anthropogenic pollution sources.Climatological conditions determine turbulent particle diffusion and transporto Thelarger particles can be deposited by gravity, but the fine particle fraction presentsa higher atmospheric residence time and respirable particle size. The evidence ofdirect effects on public health is based primarily upon the results of epidemiolo-gical studies which have used PMlO as a measure of airborne particles (COMEAP,

. 1995). The results suggest that the chemical composition of the aerosol does' nothave a major impact on the human health, but probably the mass, the number ofparticles or even the surface area of the particles is the most important determinantof their toxicity (Harrison el al., 1999). Unlike particulate, gases do not tend tosettle out of the air. They are actually removed from the atmosphere by turbulentdifussion, adsorption, absorption, e.g. on particulate matter, washed out by rain orby chemical reactions. The gases are often very reactive and lead to multiple inter-actions during their residence time in the atmosphere. Between the most reactiveoxidized nitrogen (NOx) and sulphur forms, NO, N20, N02 and S02, respectively,are found as major pollutants. Further, NOx pollution contributes to ground level

~ Environmental Monitoring and Assessment 71: 61-70,2001.•• © 2001 Kluwer Academic Publishers. Printed in the Netherlands.

62 L. A. FAGUNDEZ ET AL.

ozone, an important component of photochemical smog, which affects many urbanareas throughout the world. Size and volume of particulate matter are also import-ant when one consider the reactions of gases at the surface of the particles or thereactions occurring within the particles themselves.Since 1992, air quality monitoring is performed in Buenos Aires city (Bogo et

al., 1998). Recent data about airbome particles in La Plata area have been published(Colombo et al., 1999). The lack of air quality data in the rest of Buenos Airesmetropolitan area makes it necessary to initiate diagnostic exploratory monitoringto determine present conditions and trends. Diffusion tube samplers were used formonitoring particulate matter, S02 and N02 concentrations in San Miguel duringOctober 1998. This method has been extensively used for monitoring atrnosphericpollutants both in urban and rural areas (Schultz, 1993; Ferm and Svanberg, 1998;Ayers et al., 1998; Lam el al., 1999). With this technique, the samples are takenby diffusion, e.g., by leaving a filter in a passive tube sampler exposed to ambi-ent air concentrations. After exposure, filters are analyzed by different techniques(optical rnicroscopy and image processing, spectrophotometric deterrninations orturbidimetric analysis). Passive diffusion sampling is one of the methods in whichmonitoring networks for diagnostic purposes may be based (UNEP/WHO, 1994).

2. Materials and Methods

2.1. STUDY AREA - E VIRONMENTAL CHARACTERISTICS

The Buenos Aires Metropolitan Area, a large urban area with more than 13 rnillioninhabitants, is located on the right shore of the Rio de la Plata, over a fiat terrain,covering a surface of about 200 krrr'. San Miguel is a suburban region, situatedat about 30 km NW of Buenos Aires city. Concentrations of pollutants in the at-mospheric boundary layer are determined by the ernission sources and are stronglyaffected by the meteorological conditions. In spring (September to December in thesouthem hemisphere), stable atrnospheric conditions are absent and the mean rnix-ing heights range between 715 and 814 m (Ulke and Mazzeo, 1999). The samplingwas carried out during October 1998. For this period, the mean temperature in SanMiguel was 16.9 °C, ranging from 22.4 °C (mean maximum value) and 11.7 °C(mean minimum value) with a relative humidity of 74%. It was one of the monthswith heavier rainfall (148.8 mm as a mean value) and stronger wind. The winddirection was variable.

2.2. SAMPLING AND SAMPLE ANALYSIS

In order to get a first diagnosis of pollutant levels (particulate matter, N02 andS02), a prelirninary field study was carried out during October 1998 at sevendifferent locations in San Miguel (Figure 1). The urban configuration of this areadeveloped on the railway distribution, leading to heavy concentration of different

PRELIMINARY AIR POLLUTIO 1MONITORING 63

•San Miguel

5S-1S' W

tN

10 km

Figure 1. Location of the study area.

commercial activities in the surroundings of the train station. Sampling sites werechoosen in order to provide different urban environrnents, covering from the mosttransited downtown, where the train station is located, to semi rural areas. Monit-oring stations were situated at places where people are exposed to pollution overlonger periods. In order to capture the highest values of traffic induced pollution,the samplers were installed in pairs at street level, at about 3 m above the ground.For different pollutants, passive diffusion samplers were exposed during 18 to 26days.

2.2.1. Total Suspended Particulate Matter (TSPM)The deposition samples were collected, using a sampling device (Figure 2) similarto Sigma-2 (Schultz, 1993). The samplers were designed to achieve rain- and wind-shielded partic1e sampling, preventing wet deposition almost completely. In spiteof the fact that windspeed and heavy rain can affect the partic1e size selectionefficiency of the sampler, acts its interior as a stilling chamber, minimizing partic1edeposition by turbulent forces. The employed sampler geometry provided a sat-isfactory size selective deposition. The maximum partic1e diameter obtained withthis sampler was around 100 iun. The samplers were made of PVc. Particulatematter was collected inside the samplers on highly transparent PVC collectingplates coated with a likewise transparent adhesive.

64 L. A. FAGUNDEZ ET AL.

E

2,5xO,6 CM

2

Figure 2. Sampling device for wind- and rain-shielded particuJate matter deposition measurement.1) Externa! cap with two windows. 2) Interior with two crossed windows. 3) Stilling chamber forrninimizing particle deposition by turbulent forces. 4) Collection plate coated with an adhesive.

The TSPM, expressed in mg cm-2 month"", was determined by the gravimetricmethod. The collection plates were weighed before and after dust exposure, undercomparable humidity and temperature conditions, in a digital balance having anaccuracy of 0.1 mg. Optical microscopy (Leitz microscope with a CCD camera,at 250 x magnification) was used for the analysis of particle deposition samples.Digitalized images of different sectors of collection plates, with a pixel size of1 ¡Lm and a spatial resolution of 3 iui: were obtained. Microscope photographsof dust particles were analysed, using the National Institute of Health image pro-cessing system NIH IMAGE. The microscopical particle analysis provides the totalnumber of particles for each sample, the dust covered area, and the mean opticaldensity of each particle. These data were used to calculate the total equivalent mass,particle size distribution and mas s concentrations of translucent and black particlefractions. Microscopic evaluation provided the discrimination of particles in twomain groups: black carbon (e.g., soot agglomerates and tire rubber fragments) andtranslucent particles, mainly due to minerals and biological matter.

PRELIMINARY AlR POLLUTION MONITORING 65

0,25

~f;e 0,20OE

<¡J

E0,15o

al.s"D 0,10(J).•...•o(J)

OÜ(f) 0,05(f)

CIl2

0,00

35

30 :oo..25

.9:.eO

20 cu"-e(J)

15 üeOO

10

5

°2 3 4 5 6 7

Sampling site

~ Particulate matter

Figure 3. Comparison between average TSPM (mg cm-2 month+l ), N02 (ppb) and 502 (ppb)values, determined at different sampling sites.

2.2.2. N02 and S02High-dose diffusion samplers, with a coarse net for avoiding turbulent diffusioninside the samplers, similar to those described by Ferm and Svanberg (1998), wereused for monitoring S02 and N02. The samplers were constructed by using com-mercial polypropilene vials with a threaded cap at one end. An impregnated filterwas placed in the solid cap. The filter acts as an almost perfect sink of the pollutant.A total of 21 samples (2 to 4 at each selected sampling place) were collected foreach gas.For analyzing N02, Whatrnan filters, impregnated with a solution containing

potassium iodide (10 g), sodium arsenite (1 g), ethylene glycol (5 g) and meth-anol (84 g), were prepared and exposed to air (Ferm, 1991). After exposure, thefilters were leached in distilled water and the NO;- concentrations were determinedby a spectrophotometric technique, after the aqueous solutions was mixed with adiazoting agent.For S02, the filters were impregnated with a 4% potasium carbonate aqueous

solution (Ferm, 1991). After exposure, the filters were treated with hydrogen per-oxide (0.03%) to ensure complete oxidation to sulphate. This ion amount wasanalysed adding BaCh to the leachates in acidic media. The turbidity due to theformation of BaS04 was employed to estimate S02 content in air.

66 L. A. FAGUNDEZ ET AL.

100 ¡Lm

Figure 4. Microphotograph of dust particles coUected at site 1 (San Miguel downtown, urban area).

3. Results and Discussion

Total particulate matter, collected by square centimeter per month (mg cm-2month "), and S02 and N02 concentrations (ppb) at each sampling site, are shownin Figure 3. The sites 1 and 2 correspond to the downtown, a highly traffic-impactedarea with mass values for TSPM, three times as large than those obtained in moreresidential areas (sites 3-6). Site 7, a less urbanized area with abundant vegeta-tion, show the lowest deposition. The same trend was observed for N02. Ambientlevels of S02 are low compared with levels reported in urban areas of the northemhemisphere. In all cases they were below 2.5 ppb, near the detection limit.

Photographs of dust particles obtained by optical microscopy are shown in Fig-ures 4 (site 1) and 5 (site 7), for urban and urban background areas, respectively.The coarse fraction of airbome particles, with a mean value at about 30 ¡Lm, ismostly ground resuspended particulate matter. On the other hand, the fine frac-tion is main anthropogenic, and may have direct effects on public health, as theseparticles are deposited deeply into the lungs (Miranda, 1996). Figure 6 shows theTSPM cumulative particle distribution curve at site 2, indicating that 90 and 50%of the particles present a diameter lesser than 30 and 10 ¡Lm, respectively.

Figure 7 shows a reasonable correlation (R2 = 0.96) between the total particlemass deterrnined by weighing the residue at each site and the equivalent mass,

PRELIMINARY AlR POLLUTION MONITORING 67

100 .Ilm

Figure 5. Microphotograph of dust partic1es collected at site 7 (urban background area),

1,0

0,8eo.•....:J..D

0,6.¡:;.•....Cf)

TITI 0,4(l).•....al•....Ol(l)

0,2.•....e

0,0

0,1

,:'J."""""'"~ ..

·:··r·'"":JÍj : : : :::

.",,¡.... ':'::~""" ....

. , .'.: .:-: :.:. .., .. ,..: , . ". II!I' . , " .

.' "~

10

Particle Diameter [um]Figure 6. TSPM cumulative partic1e size distribution curve at site 2 (downtown).

100

68 L. A. FAGUNDEZ ET AL.

Figure 7. Correlation between equivalent mass (calculated assuming a spherical particle shape andunit density) and total particle mas s (detennined by weighing).

1,0.---r-----,-----,------r----~~----r_----.__,

CI)~ 0,8;-~.-----_+------+_----_+------+_----_+------+_~otCUQ.

e 0,6-t---t------t-------t------t-------t------t-------t--t~o:::lCI)e~ 0,4-t---t------t-------t------t-------t------t-------t--t-cueO"O 0,2;---+------+------+------+------+------+------+-~m>-LL

I

2 73 4 '5

Sampling sites6

Figure 8. Fraction of translucent particles at different sampling sites.

PRELIMINARY AIR POLLUTIO T MONlTORlNG 69

calculated assuming a spherical particle shape and unit density (1 g cm>'), For allsampling sites, a fraction of about 80% corresponding to translucent particles wasobtained (Figure 8). The results are similar to those obtained at a suburban site inBerlín (Schultz, 1993). This fraction is mainly due to mineral or biological sources.The other fraction is black carbono In average, particle densities are higher than theassumed 1 g cm"? value, but lower than bulk densities of pure substances due topores, crack s and voids in the particles. As the density of most silicates is near 2 gcm-3, there is expected in Figure 7 a slope value greater than unity. Translucentdust level in the atrnosphere depends on the wind induced partic1e mobilizationfrom easily erodable surfaces, such as non paved roads, land filling areas, and otheruses of soil thateliminates its vegetation. Wind is effective to erode soil when itsspeed is greater than 5 m S-l.

The results indicated that there are higher levels of both, N02 and TSPM, atsampling sites located in downtown (Figure 3, sites 1 and 2) than in more res-idential areas. In spite of the fact that the correlation factor between particulatematter and N02 concent:rations is relatively weak (R2 = 0.68), a noticeable trend ispresent, showing higher TSPM for higher N02 levels. The suspended particulatematter concentrations on the street level would be mainly affected by mechanicalturbulence created by the moving vehicles and the wind. So, in the street canyonconfiguration prevailing at sites 1 and 2 - but absent at the other sites - a tra-verse wind across the street would create recirculation and increase the pol1utantconcentration, producing a lower correlation factor.

As far as gaseous pollutants are concerned, nitrogen dioxide is the most seriousair pollutant on the road side. Near1y half of all man-made NOx emissions comefrom cars and other forms of transport, the remainder coming mainly from powerstations and industrial boilers (Radojevic, 1998). In San Miguel, cars are the mostimportant pollution source. In spite of the benefitial catalytic converter action, redu-cing NOx emissions and converting them back to N2, they are not installed in oldermodels. Obsolescence of the local automobile fieet, which has líttle proportion ofcatalytic equipped units, and the intense circulation of diesel-powered buses andtrucks contribute to enhance atmospheric pollution problems. During 1999, specialfacilities were given to car owners to change older models (lO yr or more) fornew cars. TSPM and N02 concentrations are expected to decrease with this partialautomobile fieet replacement, but its benefical effects probably will be insignificantin poorer suburban regions.

In spite of the fact that the number of analysed samples is relatively small,the data presented in this diagnostic monitoring campaign, are important in orderto estimate the levels and obtain a prelíminary spatial distribution of pollutants.The lack of historical air quality data makes jt necessary to initiate monitoringcampaigns. The knowledge of air quality in suburban residential and industrialregions is a powerfull aid to the understanding of pollution - its origin, transportand fate - in large urban areas as Buenos Aires Metropolitan Area.

=

70 L. A. FAGUNDEZ ET AL.

Referénces

Ayers, G. P., Keywood, M. D., Gillett, R., Manins, P. c., Malfroy, H. and Bardsley, T.: 1998, Atmos.Environ. 32(20),3587.

Bogo, H., Garay, F., Santander Irarragorri, E., Negri, R. M., Aramendía, P. F., GordiUo, G., SanRomán, E. and Fernández Prini, R.: 1998, Gases y partículas en la atmósfera de la ciudad deBuenos Aires. Environmental Quality througb Science, 1st Reunión SETAC América Latina,Buenos Aires.

Colombo, J. C., Landoni, P. and Bilos, c.: 1999, Environ. Pollut. 104305.COMEAP: 1995, 'Healtb Effects of Non-Biological Particles', Committee on tbe Medical Effects of

Air Pollutants, Department of Healtb, U.K.Ferm, M.: 1991, 'A Sensitive Diffnsional Sampler', Ivl.Repon L91-172, IVL, Box 47086 - 40258,

Goteborg, Sweden.Ferm, M. and Svanberg, P.-A.: 1998, Atmos. Environ. 32(8), 1377.Francis, B. M.: 1994, Toxic Substances in (he Environment, John Wiley & Sons, New York, pp. 39-

42.Gidhagen, L.: 1998, Design and Operation of Air Quality Monitoring Networks, School of

Environmental Science and Technology, Buenos Aires:Harrison, R. M., Shi, J. P. and Jones, M. R.: 1999. Atmos. Environ. 33 1037.Lam, G. C. K., Leung, D. Y. C; Niewiadomski, M., Pang, S. w., Lee, A. W. F. and Louie, P. K. K.:

1999, Atmos. Environ. 33, 1.Miranda, J.: 1996, 'Studies of Atmospheric Aerosols in Large Urban Areas using PIXE: An

Overview', Nuclear Instruments and Methods in Physics Research B 109/110, 439-444.Radojevic, M.: 1998, 'Opportunity NOx', Chem. in Britain, 30-33.Schultz. E.: 1993, Atmos. Environ. 27A(8), 1241.U1ke, A. G. and Mazzeo, N. A.: 1998, Atmos. Environ. 32(9), 1615.UNEP/WHO: 1994, Passive and Active Sampling Methodologies for Measurement of Air Quality,