Fuel 90 (2011) 35093520

Contents lists available at ScienceDirect

Fuel
journal homepage: www.elsevier.com/locate/fuel

Air quality assessment in a highly industrialized area of Mexico: Concentrations and sources of volatile organic compounds
Elizabeth Vega, Gabriela Snchez-Reyna , Virginia Mora-Perdomo, Gustavo Sosa Iglesias, Jos Luis Arriaga, Teresa Limn-Snchez, Sergio Escalona-Segura, Eugenio Gonzalez-Avalos
Instituto Mexicano del Petrleo, Eje Central Lzaro Cardenas 152, Col. San Bartolo Atepehuacan, Distrito Federal C.P. 07730, Mexico

a r t i c l e

i n f o

a b s t r a c t
Parallel to the economical benets brought by the oil industry in Mexico, there have been some negative environmental effects due to emission of pollutants to the atmosphere. Salamanca, a city located inside one of the most important industrial corridors of the country, has been frequently affected by elevated concentrations of sulfur dioxide and particle matter. However, little is known about volatile organic compounds (VOCs), which in this study are analyzed along with criteria pollutants and meteorological parameters during FebruaryMarch 2003 at urban, suburban and rural sites. Although sulfur dioxide average levels were $0.017 ppm, a high concentration event ($0.600 ppm), attributable to emissions from the oil renery and the thermoelectric power plant, was observed at the urban site at night time. The VOCs concentration varied from 170 50 ppbC (rural) to 699 212 (urban) and were constituted by 40% alkanes, 13% aromatics, 11% olens and 11% of halogenated. The most abundant species were propane (167 40 ppbC), n-butane (91 23 ppbC), toluene (51 10 ppbC) and i-pentane (44 7 ppbC), that are related to combustion processes. Freon-114, methyl bromide and 1,2-dichloroethane which are likely emitted by application of pesticides, soil fumigation and fabrication of chemicals, showed high concentrations (48 10, 50 10 and 32 6 ppbC respectively) in the rural sites, highlighting the importance of control measurements implementation for these species, as they represent a potential hazard for public health. Moreover, these halocarbons showed similar ratios regardless the monitoring site, suggesting same source. Modeling results indicated that meteorological conditions generally transport air masses to the northeast rural areas where the highest concentrations of ozone were calculated. 2011 Elsevier Ltd. All rights reserved.

Article history: Received 24 August 2010 Received in revised form 29 March 2011 Accepted 31 March 2011 Available online 24 April 2011 Keywords: VOCs Air quality Industrial areas Halocarbons Mexico

1. Introduction Over the last century, the oil industry has emerged as the primary energy source [1]. Currently, the life style of human societies depends on energy (electricity generation, natural gas, crude oil and its rened products, coal, etc.); without it, societies as we know them would collapse. Even though the oil industry has made important contributions to the global economy, usually this has been accompanied with negative environmental impacts from a variety of activities such as oil drilling, renery, oil spillage, gas and aring. Moreover, deterioration of the environment may not be circumscribed to the local scale, it can reach regional and global extent due to the emission of precursors of secondary pollutants and chemical species that contribute enhancing global warming and stratospheric ozone depletion [24]. Public health may also be affected if emissions contain toxic or carcinogenic species [5]. In recent years concern over public health and environmental protection has become a critical issue, this means that a growing
Corresponding author. Tel.: +52 55 91757558.
E-mail address: gsanchez@imp.mx (G. Snchez-Reyna). 0016-2361/$ - see front matter 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.fuel.2011.03.050

amount of investment and effort is dedicated to reconcile the environment and development of countries. The economy of Mexico strongly depends on oil industry; in 2005 the primary distillation capacity (1540 MBD) ranked the country on the top 15 worldwide and 4 in Latin America. The crude is processed in 6 oil reneries which mainly produce gasoline, diesel, jet fuel, coal, asphalt, and lubricants. The third most important renery is the Ing. Antonio M. Amor, which processes 197 MBD [6]. The renery and a variety of industries constitute one of the most important industrial corridors of Mexico, known as the Bajo Industrial Corridor (BIC) located in the State of Guanajuato, in the central area of the country (Fig. 1). The BIC has nearly 465 industries, from medium to large size, including Chemical, Power Generation, Food Processing, Textile and Metal-mechanic [7]. The Salamanca city, with 250,000 inhabitants, is located at 20340 0900 N latitude and 101110 5100 W longitude, at 1720 m above mean sea level [8]. The Salamanca municipality encompasses a total area of 774 km2. The agriculture is now the second most important economical activity, with a designated area of about 80% of the municipality. The impact of agriculture on the environment is important, especially for the use of fertilizers, pesticides,

3510

E. Vega et al. / Fuel 90 (2011) 35093520

(A)
Salamanca County

Guanajuato State

Zacatecas

(B)
Len Jalisco Salamanca Irapuato

Sn. Luis Potos

Quertaro

Celaya

monitoring network, such as the volatile organic compounds (VOCs). The negative effects of VOCs on the environment and public health are well documented. From the environmental point of view, some VOCs (i.e. olens and aromatics which are mainly anthropogenic) are reactive species that break out the natural equilibrium of generationdestruction of tropospheric ozone, thus the concentrations of this compound and other photochemicallyproduced pollutants are frequently high in the urban environment. Besides, the reactive organic gases can partition into the aerosol phase generating secondary organic aerosols. Other important group of VOCs is constituted by the halogenated species, which are originated almost exclusively from anthropogenic emissions due to its usage as an industrial solvent and degreaser. Some of these compounds have been the focus of intensive research, such as the chlorouorocarbons due to their participation in the stratospheric ozone depletion. In addition, many of the halogenated species represent a potential hazard to human health due to the toxic and/or carcinogenic effect [915]. The public opinion on air quality deterioration in Salamanca, encouraged PEMEX (National Oil Company) to support an extensive 2-week monitoring eld study with the aim of augmenting the knowledge of sources, transport and fate of air pollutants in the region, therefore effective control measurements of atmospheric pollution can be designed. The main ndings of such campaign are presented in this work, particularly the chemical characterization, distribution and origin of VOCs, as well as the meteorological parameters that inuence the dilution and transport of pollutants. The later was also estimated by applying a 3 D air quality model.

Michoacn
Fig. 1. Geographical location of: (A) Salamanca County and (B) The Bajo Industrial Corridor.

2. Field campaign As an outcome of a collaborative effort, the Instituto Mexicano del Petroleo (IMP), the Instituto de Ecologa de Guanajuato, the Instituto de Investigaciones Cientcas at the Universidad de Guanajuato, the Centro de Ciencias de la Atmsfera at the Universidad Nacional Autnoma de Mxico, the Patronato de Salamanca, and the Ing. Antonio M. Amor Oil Renery, with the PEMEX sponsorship, carried out a eld monitoring campaign, from February 21 to March 9 2003. The main objectives were to chemically characterize the air pollution in the urban area of Salamanca in both the particle and gas phases, and to assess the potential impact of pollutants in the regional scale. The interested reader can consult Vega et al. [16] for the particulate matter results found in this region. The rst week of measurements was focused on the characterization of urban air quality; while the second week was designed to evaluate the regional impact of urban emissions. The monitoring sites of the urban domain (10 10 km) and of the regional domain (80 80 km) are shown in Fig. 2A and B. Table 1 shows site location, description, sampling period and measurements performed. Three automated samplers (VOCCS-ANDERSEN and AVOCSANDERSEN models) with a Viton diaphragm pump were used to collect VOCs (dened in this work as hydrocarbons from C2 to C12) in canisters over 12 h period (06001800 and 18000600 local time) in the urban sites and 24 h period in the rural/boundary sites. A total of 80 canisters were analyzed in the Laboratory by cryogenic pre-concentration/high-resolution GC technique, similar to the TO-14A protocol [17]. Water Sep-Pak DNPH-Silica cartridges were used to trap carbonyl species. Twelve samples were taken during the rst week of the campaign at Cruz Roja (CR) urban site from 0600 to 0900 and from 1200 to 1500. The derivatives were eluted and analyzed by HPLC with UV photodiode array detector according to the TO11A protocol [18]. Criteria pollutants were measured using a

deforestation in the uplands and post-harvesting burning. The biggest impact to the atmosphere is represented by the post-harvesting burning, due to the emission of large amounts of ozone precursors and particulate matter [8]. According to the 2000 BIC Emissions Inventory, the emissions of particles (PM10), sulfur dioxide (SO2), carbon monoxide (CO), nitrogen oxides (NOX) and hydrocarbons (HC) were 71,443, 112,480, 1,650,772, 142,183 and 260,296 tons per year, respectively. It is estimated that PM10 are released to the atmosphere mainly by commercial and service activities, SO2 and NOX by electricity generation, and CO and hydrocarbons by vehicle exhaust. Salamanca contributes with 18% of PM10, 92% of SO2, 8% of CO and HC and 14% of NOX of the total BIC emissions. Thus, Salamanca is by far the major generator of SO2 emissions in the region; while other cities release ozone precursors [8]. Parallel to the economical development of the BIC, there have been some adverse environmental impacts which have brought the attention of government agencies, civil and private associations. As a result, since 2000 a Monitoring Network in Salamanca routinely measures CO, SO2, NO2, O3, and PM10. According to local environmental authorities, the SO2 air quality standard, AQS, (0.13 ppm in a 24 h average, no more than once per year) was exceeded 13%, 24% and 22% of days in 2000, 2001 and 2002 respectively in downtown Salamanca, mainly during winter. The exceedences of other gaseous pollutants is less frequent, for instance, NO2 and CO have been practically below their AQS (0.21 ppm, 1 h average, and 11 ppm in 8 h average respectively) [8]. Although total mass of criteria pollutants is routinely measured, little is known in this highly industrialized region about the gaseous and particle contaminants that are not included in the local

E. Vega et al. / Fuel 90 (2011) 35093520

3511

(A)

20.70

20.65

CA
20.60

20.55

US

CR T NA DIF

CG

VA

20.50 -101.30

-101.25

-101.20

-101.15

-101.10

-101.05

-101.00

(B)
21.0

SI
20.9

20.8

20.7

MI
20.6

JR

PN
20.5

SAL SAL

20.4

VS
20.3

20.2

20.1 -101.6 -101.5 -101.4 -101.3 -101.2 -101.1 -101.0 -100.9 -100.8 -100.7

Fig. 2. Monitoring sites location: (A) Urban Area: Cruz Roja (CR) and DIF; suburban: Crdenas (CA), Cerro Gordo (CG), Universidad la Salle (US), and Valtierrilla (VA); the Renery (R) and the Power Plant (T). (B) Regional Area: Silao (SI), Mirandas (MI), Juventino Rosas (JR), Pueblo Nuevo (PN) and Valle de Santiago (VS), SAL is the Salamanca City.

mobile Lab equipped with conventional analyzers (Monitor Labs). Methods used to determine these pollutants were NOM-034 SEMARNAT 1993 using dispersive spectroscopy for CO (detection limits from 0 to 50 ppm); NOM-037 SEMARNAT 1993 using quimioluminescence for NOx (detection limits from 0 to 0.50 ppm); USEPA EQOA 0193 091 using UV photometry for O3 (detection limits from 0 to 1.0 ppm) and USEPA EQSA 0193 092 using pulse uorescence for SO2 with detection limits from 0 to 1.0 ppm. Along the eld campaign, the surface meteorological parameters temperature (T), relative humidity (RH), wind speed (WS), wind direction (WD), atmospheric pressure (P) and solar radiation (SR) were also measured at three sites in the urban area and at four in the boundary sites. The vertical thermodynamic prole variables (P, T, RH and horizontal wind vector) were measured using a Digicora II radiosonde system from Vaisala (Model SPS-220). Three radiosondes were launched every day at 0800, 1200 and 1800 at SI, JR and VS sites. The information was used to determine the mix-

ing height based on potential temperature and specic humidity proles, and also as input for the mesoscale meteorological model RAMS [19] and the 3D air quality model [20]. 3. Results 3.1. Criteria pollutants The average and standard deviation for 1 h average concentration of criteria pollutants (CO, NO2, SO2 and O3) and nitrogen oxides (NOX) at urban (CR and DIF), suburban (CG and US) and rural (MI, JR and VS) sites are shown in Table 2. As expected, the atmospheric concentrations of CO, NO2 and NOX were considerably higher at the urban sites compared with the suburban and rural ones. The suburban sites are surrounded by crop elds and unpaved small roads with little trafc of heavy-duty vehicles which may inuence the measurements, however it is expected to be minimal.

3512

E. Vega et al. / Fuel 90 (2011) 35093520

Table 1 Location and type of monitoring sites during the monitoring eld Campaign in Salamanca, Mexico. FebruaryMarch 2003. Site name, code and coordinates Cruz Roja (CR) 20.58N, 101.20W DIF 20.56N, 101.20W Crdenas (CA) 20.63N, 101.22W Cerro Gordo (CG) 20.59N, 101.13W Valtierrilla (VA) 20.56N, 101.13W Universidad La Salle (US) 20.55N, 101.23W Silao (SI) 20.59N, 101.42W Valle de Santiago (VS) 20.35N, 101.20W Pueblo Nuevo (PN) 20.55N, 101.35W Juventino Rosas (JR) 20.64N, 101.00W Mirandas (MI) 20.56N, 101.14W
a

Site type Urban

Site description Located in the main street of Salamanca with high vehicular (light and duty) trafc all day; 2 km west of the Renery and the Power Plant Located in a residential area with high vehicular trafc, mainly of gasoline-powered vehicles The site is in a crop eld near an unpaved road with little vehicular trafc 7 km southwest of the Renery and Power Plant, 200 m of a highway used mainly by heavy-duty vehicles. The site is surrounded by crop elds Located in a populated community with moderate transit of gasoline and diesel vehicles Located inside the University, southwest Salamanca; site is surrounded by vegetation; 400 m away from a four-lane road with light and duty-vehicles trafc The site was used for meteorological measurements, located northwest Salamanca Located inside a natural protected area known as Siete Luminarias. The site is 300 m above the Salamanca level, 1.5 km away from a road with little vehicular trafc The site is surrounded by crop elds near one unpaved road with little light-duty vehicular trafc 22 km northeast Salamanca, in a commercial area with buildings 46 m height with moderated vehicular trafc This monitoring site is located 6.5 km north Salamanca, in an wide-open area surrounded by crop elds

Measurement period February 21 March 9 2003 February 21 March 9 2003 February 22 28 2003 February 21 28 2003 February 22 28 2003 February 21 28 2003 March 2003 March 2003 March 2003 March 2003 March 2003 29 29 29 19 19

Measurements Criteria pollutantsa, VOCs, carbonyl species Criteria pollutantsa Criteria pollutantsa Criteria pollutantsa, VOCs Criteria pollutantsa Criteria pollutantsa, VOCs Radiosondes Criteria pollutantsa, radiosondes Criteria pollutantsa, radiosondes Criteria pollutantsa, VOCs, radiosondes Criteria pollutantsa, VOCs

Urban Suburban Suburban Suburban Suburban

Rural/ boundary Rural/ boundary Rural/ boundary Rural/ boundary Rural/ boundary

Criteria Pollutants were measured each minute and include O3, CO, NO2, SO2 and PM10.

Table 2 Basic statistics of concentrations of pollutants (ppm) in urban (CR and DIF), suburban (CG and US) and rural (JR, VS and MI) sites of Salamanca. Species Urban Average O3 CO SO2 NO2 NOX 0.022 2.200 0.017 0.022 0.055 Maximum 0.094 19.500 0.310 0.064 0.425 Minimum 0.001 0.960 0.001 0.002 0.002 SD 0.018 1.150 0.043 0.013 0.045 n 699 699 699 699 699 Suburban Average 0.021 0.158 0.016 0.014 0.029 Maximum 0.072 1.890 0.269 0.047 0.127 Minimum 0.001 0.001 0.001 0.002 0.001 SD 0.017 0.240 0.034 0.008 0.028 n 353 347 294 353 344 Rural Average 0.034 0.038 0.011 0.010 0.014 Maximum 0.114 0.570 0.318 0.046 0.092 Minimum 0.001 0.001 0.001 0.001 0.001 SD 0.020 0.067 0.026 0.008 0.013 n 525 525 525 525 525

Regarding ozone, it is noticeable that average levels in the urban sites were 0.022 ppm with maximum of 0.094 ppm, while the rural sites reached 0.031 ppm in average and 0.114 ppm maximum. The diurnal variations of this oxidant at CR (urban), suburban (CG) and rural (MI and VS) have been plotted in Fig. 3. For all sites, maximum levels took place from 1200 to 1500, being about 40% higher in the rural site (MI) than the urban and suburban ones. The monitoring site of VS located within a natural protected area, 25 km south of the city at an altitude 300 m above the level of Salamanca, registered ozone average concentration of 0.038 ppm and maxi-

0.08 0.07 0.06

O3 (ppm)

0.05 0.04 0.03 0.02 0.01 0.00 0 3 6 CR 9 MI 12 15 CG 18 VS 21

Fig. 3. Hourly average ozone concentration (ppm) at urban (CR), suburban (CG) and rural (MI and VS) sites of Salamanca during February 21March 8, 2003.

mum of 0.052 which could be considered as background concentrations. These results are within the range of ozone levels reported at rural areas of Canada [21]. In addition, it is known that areas of high altitude (the study region is 1720 m above sea level) could be inuenced by inputs from free troposphere [2123]. As it is discussed later, the peak of mixing height was observed everyday at 1800 reaching 3500 m above ground level, making possible this phenomenon, although more measurements would be necessary to reach a conclusion. The highest ozone concentrations were measured at rural MI which is located 6.5 km north of the city. To investigate the inuence of surface wind direction and speed on the spatial distribution of ozone, an analysis was performed dividing the data-set into four subsets according to surface wind direction: north (31545), east (45135), south (135225) and west (225315). As expected, the results reected the inuence of site location regarding main pollution sources. The concentrations of ozone at MI were higher when the wind blew from the west and south, i.e., the site was downwind major pollution sources. These wind directions were observed $64% of the time, mostly from 1200 to 1900, which suggest that ozone levels at MI were strongly inuenced by transported ozone. Moreover, the correlation analysis indicated that levels observed at MI and CR are associated during this meteorological condition (R2 = 0.72). The same analysis was performed using data of the opposite wind direction (north), showing an even stronger relationship (R2 = 0.91) between the urban and rural ozone levels. It has to be added that north winds generally blew

E. Vega et al. / Fuel 90 (2011) 35093520

3513

from 0000 to 0700 and 2100 to 2300, thus measurements indicate background concentrations. The same analysis was performed for urban CR and rural VS; the data-set was ltered so only measurements taken when the rural site was located downwind Salamanca. A correlation coefcient of 0.55 was obtained between these sites (the site is 25 km south), while similar coefcient was obtained for the analysis using only winds from the south (R2 = 0.57). In summary, the results suggest that transport of air masses from downtown Salamanca and the elevation of the study area play an important role in the high ozone concentrations observed in the rural sites. To evaluate the effect of wind direction on the concentration of the other pollutants, the average and standard deviation of concentrations of criteria pollutants and nitrogen oxides were calculated for each of the wind sectors above described. At CR, carbon monoxide and nitrogen dioxide registered higher levels when the wind blew from the south or east, which is explained by the location of this site in the northwest of Salamanca. On the other hand, ozone levels at CR were statistically lower when the wind blew from the north and higher when the wind blew from east. The concentrations of SO2 showed higher values associated with eastwinds (where the renery is located) and lower for any other wind direction. At suburban CG, carbon monoxide and nitrogen dioxide were higher when the wind blew from the east or south, as it was envisaged. The concentrations of pollutants were higher when the wind came from the south at the rural site, which is explained by the sites location (Table 3). Moreover, the observed results agree with simulation results of transport of air masses, which are discussed in Section 3.4. As mentioned before, the largest sources of SO2 in the study region are the Renery and the Power Plant facilities, both located inside the Salamanca city. Even though emissions are released to the atmosphere by elevated stacks (3060 m above the ground), the natural atmospheric processes such as turbulence, thermal inversions and high pressure systems, may increase the concentration and residence time of pollutants. Fig. 4 exemplies the typical daily variation of SO2 concentration. From 0000 to 0600, due to stable atmospheric conditions, the highest concentrations were observed at CR, located 2 km west of major emission sources (the other two sites, MI, and JR, are 6.5 km north and 22 km northeast, respectively). From 0600 to 0900 the wind blew from the east, therefore CR site received directly the SO2 emissions. Around 0900, the surface wind direction blew from SSW, remaining under these conditions the rest of the day, as a result, concentrations at CR showed a decrement while at MI an increment was registered. It was found that short-term variability periods (e.g. at CR at 09000930 and at MI at 12001220 in Fig. 4) were associated to wind speed lower than 1.0 ms1. According to RAMS modeling results, surface winds before 0900 were driven by cold air draining from the near moun-

0.40 0.35 0.30

CR JR MI

E N NNE S-SW SW

S-W N erratic

SO2 (ppm)

0.25 0.20 0.15 0.10 0.05 0.00 00:00 03:00 06:00 09:00 12:00 15:00 18:00 21:00 24:00

March 2, 2003
Fig. 4. Time series for SO2 concentrations at CR, MI and JR during March 2, 2003. Upper horizontal lines indicate the observed wind direction.

tains towards lowest topographic levels in the basin, accumulating air pollutants over the city. After 0900 and due to synoptic wind forcing, the wind blew from SW in the whole area, sweeping pollutants out of the city. The estimated results were validated against meteorological measurements nding good correspondence between observed and predicted values. During the monitoring campaign, high SO2 concentrations (>0.60 ppm, 1 min sampling-time) were registered once at CR at 0200; such high concentrations were associated with persistent east wind direction (8295) and wind speed of 2.6 ms1, which was higher than the average registered at this time of the night (1.3 ms1). Moreover, concentrations above the average of toluene, m,p-xylenes, benzene, Freon-114 and methyl chloroform were also registered during this event, suggesting emissions from industrial activities. The night-time high SO2 concentration events were frequently observed in 19992002 and were associated with venting activities of the Renery and/or Power Plant [8]. However since 2003 the air quality program began to operate in the region, tackling SO2 and particle matter problems mostly [24]. The control measurements include the usage of fuels with low sulfur content in the power plant and an increment in the natural gas consumption thus is possible that the rst positive results were observed during the eld campaign. 3.2. Volatile organic compounds 3.2.1. Concentrations Approximately 200 chemical species of VOCs were identied and quantied by GC analysis. A classication was made according

Table 3 Wind directional analysis for concentrations of criteria pollutants (ppm) during FebruaryMarch 2003 in Salamanca, Mexico. Site Species Wind direction North (31545) CR O3 CO NO2 SO2 O3 CO NO2 SO2 O3 CO NO2 SO2 0.016 0.021 1.946 0.927 0.032 0.010 0.020 0.046 0.024 0.021 0.161 0.229 0.016 0.008 0.031 0.046 0.016 0.009 0.074 0.151 0.011 0.005 0.004 0.007 East (45135) 0.022 0.021 2.219 1.104 0.037 0.010 0.122 0.139 0.012 0.012 0.383 0.518 0.020 0.010 0.010 0.019 0.015 0.010 0.068 0.121 0.012 0.006 0.004 0.010 South (135225) 0.018 0.017 2.937 1.604 0.034 0.012 0.012 0.034 0.021 0.018 0.264 0.404 0.016 0.009 0.030 0.058 0.029 0.025 0.129 0.149 0.017 0.008 0.032 0.063 West (225315) 0.021 0.019 2.201 1.094 0.028 0.010 0.008 0.020 0.020 0.019 0.158 0.275 0.014 0.008 0.030 0.049 0.044 0.021 0.093 0.136 0.009 0.007 0.014 0.047

CG

MI

3514

E. Vega et al. / Fuel 90 (2011) 35093520

to the functional group, indicating that alkanes were the most abundant group ($40% of the total mass), followed by aromatics ($13%), olens ($11%) and halogenated compounds ($11%). The total mass concentration for samples taken from 0600 to 1800 (diurnal samples) varied from 699 212 ppbC at the urban site (CR) to 170 50 ppbC at the suburban site (US), which is located southwest of the city. The concentrations of VOCs taken from 1800 to 0600 (nocturnal samples) were higher than those measured in the morning, reaching 956 120 ppbC at CR and 554 95 ppbC at US. Moreover, diurnal samples showed different composition compared to nocturnal samples: the proportion of alkanes was higher during the night at CR, while halogenated and the unidentied groups showed higher percentages at night in the suburban sites (Fig. 5A and B). The higher levels of VOCs during the night period were mainly driven by the increase of emissions of propane, n-butane, i-butane, Freon-114 and toluene. Velasco et al. [25] reported that VOCs levels for the Mexico City Metropolitan Area during 2003 uctuated from 2473 in the industrial area to 83 ppbC in the rural site for morning samples and from 1467 to 81 ppbC for afternoon samples. A simple comparison between values of VOCs measured in Mexico City and Salamanca, even though of the difference in sampling-time, indicated that concentrations observed in Salamanca were similar with some sites of Mexico City. Regarding the distribution of VOCs by type, it was observed the same order of abundance (i.e. alkanes > aromatics > olens), although the proportion of olens in Salamanca was twice the concentration reported for Mexico City. The highest concentrations of individual VOCs measured in the atmosphere of Salamanca during the sampling campaign are described in Table 4. At the urban and suburban sites, propane was the most abundant species, followed by n-butane, toluene and ipentane. The rst two species are attributable to the wide usage of LPG for cooking and heating; on the other hand, toluene and i-

pentane may be emitted by both mobile and industrial sources. Propane and n-butane are reported as the most abundant species in the atmosphere of Mexico City [25], showing concentrations similar to those observed in this study. The halogenated species such as Freon-114, methyl bromide, 1,2-dichloroethane, 1,2-dichloropropane and vinyl chloride showed levels noticeably high at the suburban/rural sites, especially during the night time. The concentrations of these species are markedly higher than those observed in an industrial area of China, where Freon-114 and methyl bromide are $16 and $18 pptv [26,27]. Therefore, the results found in this study highlight the importance of carrying out continuous monitoring of VOCs, so control measurements can be taken to reduce population exposure. Formaldehyde, acetaldehyde and acetone average concentrations were 3.75 1.94, 2.45 1.45 and 7.71 6.38 ppb, respectively. Formaldehyde may have a signicant inuence on the local photochemistry, more than any other carbonyl species [28]. However, concentrations of formaldehyde were low in comparison to those measured in the southwest and downtown Mexico City (13.3 and 23.9 ppb respectively) [29,30].

3.2.2. Sources of VOCs According to abundance and spatial distribution of VOCs species in Salamanca, the following categories can be identied as the major contributors: (1) Vehicular Emissions (gasoline and diesel-powered vehicles). The emissions of vehicles powered by gasoline and diesel are highly loaded with ethene, acetylene, propene, MTBE, n/ipentane, 2,2-dimethylbutane, 2-methylpentane, benzene, toluene, and xylenes [31]. These species were observed in all monitoring sites in concentrations high enough to rank them among the rst 15 (see Table 4). MTBE average levels were 4.79 3.34 ppbC at the urban site (not shown in Table 4), representing about 1.0% of the total VOCs. By comparison, a percentage lower than 2% is reported for samples collected in Mexico City [25]. (2) LPG handling and leakage. This source is characterized by emissions rich in propane, i-butane, and n-butane [32]. Propane was the major species at all sites (except rural MI). (3) Rening processes. Emissions from rening processes are composed by a large variety of species, which depend on the process itself. It is reported in the literature that ethane, propane, n/i-pentane, toluene and formaldehyde are emitted by rening oil processes and by fugitive emissions [33,34]. (4) Other sources: The high levels of halogen species such as methyl bromide, methyl chloride, 1,2-dichloropropane and dichloroethane reveal the presence of sources related with agriculture activities (e.g. application of pesticides and soil fumigants) and with industrial solvent and degrease processes [35]. As mentioned previously, burning of agriculture debris has been a frequent activity in the region that generates large amounts of particles and gases, mainly CO and organics. The presence of acetonitrile in ambient air indicates this source [36,37]. We have used styrene as surrogate species for acetonitrile, since the former has also been considered marker for biomass burning in the rural environments [25]. Average concentrations of styrene at the urban site were 2.0 0.88 ppbC, while concentrations of 2.55 2.27 and 1.10 0.80 ppbC were found at the rural sites MI and JR, respectively. Together, concentrations of styrene and characteristics of MI site (surrounded by crop elds), suggest that this site is inuenced by biomass burning emissions. However, no clear relation-

(A) 100
80 60

7 9 1 4 19

3 21

5 16 4 10 14 10

6 13 3 11 13 9

27 1 9 12 8 16 2 7 16

%
40 20 0

20

39

38

36

41

46

CR

CG
alkanes olefins oxigenated

MI
aromatics unidentified
4 21 1 17

JR
halogenated HC2
1 24 1 28

US

(B)

100 80 60

5 9 1 4 17

14

%
40 20 0
50

9 5 5 12 44 28

CR
alkanes olefins oxigenated aromatics unidentified

US
halogenated HC2

Fig. 5. Distribution of abundance of VOCs in Salamanca. (A) Diurnal samples (0600 1800); (B) nocturnal samples (18000600).

E. Vega et al. / Fuel 90 (2011) 35093520 Table 4 Concentration (ppbC) and standard deviation for the most abundant VOCs in Salamanca during FebruaryMarch 2003. CR 06:0018:00 h 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
a

3515

CG 18:0006:00 h 167.0 40.3 43.6 6.7 90.8 22.9 51.2 10.5 37.6 15.5 20.1 7.1 14.9 8.4 19.5 2.5 33.1 7.5 17.2 6.5 18.3 8.6 15.6 3.8 14.6 2.3 12.8 2.3 18.4 4.3 14 8 a 2 0.4 5.1 6.4 1.2 0.2 7.7 1.5 06:0018:00 h 31.6 9.1 9.3 3.3 17.9 4.4 11.6 2.8 7.9 2.2 5.6 2.4 6.7 1.6 7.3 1.7 7.7 2.3 7.3 3.6 3.7 1.0 6.6 1.8 3.3 1.7 3.1 0.9 0.6 0.5 a 9.0 3.0 9.6 3.1 1.3 0.3 7.0 2.3 1.9 1.8 18:0006:00 h 55.5 20.6 31.8 4.4 26.7 4.9 13.2 1.5 8.4 2.6 5.3 3.1 6.5 2.5 7.2 2.8 11.1 2.9 7.6 5.3 4.4 0.9 5.8 1.7 4.3 0.9 3.1 0.8 0.5 0.3 a 28.4 3.9 31.8 4.3 1.6 0.7 21.4 3.7 1.7 0.6

US 06:0018:00 h 16.0 10.4 16.0 4.6 9.2 6.0 4.9 2.2 5.0 1.4 2.9 1.7 4.0 2.9 4.2 2.5 3.4 2.2 3.3 2.4 1.3 0.6 1.6 1.0 2.1 0.9 1.4 0.6 5.6 3.4 7.1 4.1 a 0.2 0.1 1.0 0.3 a 0.3 0.2 18:0006:00 h 18.2 8.6 18.0 8.0 9.0 4.2 4.8 3.0 8.3 8.7 1.6 1.1 2.9 2.2 3.3 1.9 3.3 1.9 2.4 1.8 1.4 0.7 2.3 1.3 2.9 1.0 2.5 2.4 0.4 0.4 14.0 4.2 48.3 9.7 49.9 10.1 1.2 0.5 32.3 5.9 0.4 0.3

MI 00:0023:00 h 18.7 7.6 26.9 28.7 11.4 4.4 6.9 2.9 22.4 36.1 2.1 0.8 4.7 2.8 5.0 3.4 4.2 2.5 2.5 1.1 0.6 0.7 6.1 1.4 2.3 1.1 7.3 10.6 5.3 9.8 8.9 3.3 a 1.6 1.2 1.0 0.2 a 0.6 0.2

JR 00:0023:00 h 42.3 13.4 14.0 7.7 20.5 5.3 12.5 5.3 10.8 7.7 2.9 1.6 10.7 11.1 6.9 3.2 10.8 3.2 3.4 2.2 2.3 3.4 4.7 1.9 5.5 3.3 3.9 2.4 1.9 0.7 18.0 1.7 a 2.3 1.7 1.1 0.2 a 1.5 0.8

Propane i-Pentane n-Butane Toluene m/p-Xylene Ethane Acethylene Benzene i-Butane Ethylene Propene n-Pentane 2,3-Dimethylbutane o-Xylene 3-Methylbutene Freon-114 Methyl bromide 1,2-Dichloroethane Methyl chlorine 1,2-Dichloropropane Vinyl chlorine

81.1 34.2 33.9 15.4 46.2 19.7 37.7 16.5 28.4 13.0 23.6 18.3 9.1 2.3 17.2 7.1 18.0 8.4 19.4 13.6 16.8 10.0 13.6 5.6 10.3 6.0 10.9 6.0 41.4 71.8 8.6 2.7 a 2.8 3.0 3.6 3.0 1.1 0.6 6.1 2.7

Not available.

ship was found between styrene and other species indicative of biomass burning. A series of scatter plots were constructed to examine the contribution of urban and industrial emissions in the region of study. Toluene and propene are emitted by both sources; while MTBE is a distinctive vehicular emission tracer. Fig. 6A shows the scatter plots for MTBE and toluene and Fig. 6B for MTBE and propene. Both gures show a subset of data where the two species correlate linearly indicating the mobile source, while the second non-correlated subset indicates an industrial origin. The sites mostly inuenced by mobile sources were US, MI, JR and CR, while CG and in some days CR showed inuence by industrial emissions. This result agrees with the assumption that pollutants are mainly

transported from the city towards northeast, and that CR and CG were the sites with the higher Renery and Power Plant emissions impact. Propane and n-butane are known ngerprint species of leakages and unburned LPG [32]. Fig. 6C shows the high correlation between these species, which is an indication that LPG emissions were homogenously emitted in the whole area. Mobile sources also contribute to the levels of propane as corroborated by its correlation with i-butane, shown in Fig. 6D, particularly with data from US, MI and JR sites. As envisaged, the halocarbons species showed no relationship with compounds which are vehicular or LPG tracers. However, some halocarbons such as vinyl chlorine and 1,2-dicholorpropane

(A)
MTBE (ppbC)

30

(B)
R2=0.98

30

R2=0.95

MTBE (ppbC)

20

Mobile Industry

20

Mobile Industry

10

10

0 0 10 20 30 40 50 60 70 80 90

0 0 10 20 30 40

Toluene (ppbC)

Propene (ppbC)

(C)
Propane (ppbC)

300 250 200 150 100 50 0 0 30 60 90 120 150 CR CG LS MI JR

(D)
R =0.97
2

300 250

R2=0.92

Propane (ppbC)

200 150 100 50 0 0 10 20 30 40 50 60 CR CG LS MI JR

n-butane (ppbC)

i-butane (ppbC)

Fig. 6. Scatter plots of selected species and its relationship with main local sources in Salamanca, Mexico. FebruaryMarch, 2003.

3516

E. Vega et al. / Fuel 90 (2011) 35093520

Table 5 Correlation coefcient (Pearson) for selected halocarbon species and VOCs in Salamanca during FebruaryMarch 2003. MTBE Freon-114 Methyl bromide 1,2-Dicloroethane Methyl chlorine 1,2-Dicloropropane Diclorobenzene Vinyl chlorine 0.05 0.01 0.03 0.09 0.04 0.00 0.24 i-Butane 0.00 0.20 0.22 0.18 0.37 0.04 0.72 Toluene 0.00 0.34 0.24 0.21 0.71 0.06 0.88

(a) March 5, 2003: 18 hrs

5000

4000

Altitude (m)

VS SI JR

3000

2000

1000

showed moderate association with toluene, which is probably due to the industrial origin of the later (Table 5). On the other hand, strong relationship was found between methyl bromide, 1,2dichloropropane and 1,2-dicholoroethane (Fig. 7). Methyl bromide is mainly used as soil fumigant for the control of nematodes, fungi and weeds; it is also used in food-processing facilities (e.g. for extracting oils from nuts, seeds and owers). 1,2-Dicholorporpane is still used as pesticide, although this practice is prohibited in North America and Europe, it has also industrial use in the application of paint, lacquers and varnishes and for manufacture of many chlorinated compounds. 1,2-Dicholoroethane has agricultural and industrial applications, as chemical intermediate in the vinyl chloride monomer manufacture [38]. Finally, the concentrations of isoprene, the unique biogenic VOCs measured, showed concentrations $1.0 ppbC in the urban sites and lower than 1.0 ppbC in the suburban and rural sites. Its higher level in the urban area may be the result of vehicular emissions rather than vegetation, which is scarce in the city. 3.3. Meteorology Before this work, no previous studies on local wind circulation had been reported for this region. Meteorological measurements were carried out to determine the occurrence and properties of local ow patterns, to examine the structure and evolution of the mixing layer over the basin and to provide data for evaluation and testing numerical meteorological models. Three important parameters were determined at JR, SI and VS sites to understand transport and dilution of pollutants in this region: the mixing layer height (MH), the transport velocity (TW) and the ventilation index (VI) which is calculated as the product of MH and TW. In general, the VI values assured good ventilation conditions in the whole area in the afternoon. The highest mixing layer height was observed at 1800, reaching mean values approximately of 3300 m above ground level. The frequent meteorological conditions are exemplied with the data from March 5: synoptic winds aloft had strong inuence

60

120

180

240

300

360

Wind Direction

(b) March 9, 2003: 18 hrs

5000

4000

VS SI JR

Altitude (m)

3000

2000

1000

60

120

180

240

300

360

Wind Direction
Fig. 8. Wind direction proles at Silao (SI), Juventino Rosas (JR) and Valle de Santiago (VS). (a) March 5, 2003; (b) March 9, 2003 both at 18:00.

70 60 50
1,2dichloropropane methyl bromide R2 = 0.99

ppbC

40 30 20 10 0 0 10 20 30 40 50 60 70
R2 = 0.99

on the wind ow at lowest levels, wind direction above 3 km from the ground is similar for the three sites; below that level, wind direction changes slightly along the column, reaching its maximum difference on the surface level, varying from S at JR to SSW at SI (Fig. 8a). This meteorological condition caused the transport of pollutants out of the city. On the contrary, on March 9 (Fig. 8b), wind direction below 1 km blew from NNE at JR and VS sites to ENE at SI site, above that level wind changed constantly until it reached synoptic winds (SSW), 5 km above the ground, causing transport of air masses into the city [39]. The potential temperature evolution along the day is an important parameter that measures the dilution of pollutants. The frequent conditions observed at 0800 at all sites showed that the surface-based inversion formed overnight was partially eliminated. At noon, the mixing layer height was above 1200 m at all sites, and at 1800 the mixing layer was approximately between 3000 and 3500 m high. On March 9, the potential temperature had a different behavior, at 0800 at all sites presented surfacebased thermal inversion, at noon the mixing layer was below 500 m, and at 1800 the mixing layer height reached the highest values during the eld campaign, above 4000 m. 3.4. Modeling of transport and dispersion of pollutants in the region The Regional Atmospheric Model System (RAMS) and the CITSAPRC99 models were used to simulate the physical and chemical processes which control dispersion, transport and formation of pollutants in the atmosphere. The study region was set to an extension of 140 140 km, divided into a grid of 74 74 tridimensional cells with a resolution of 2 2 km. the Salamanca city was located

1,2 dochloroethane (ppbC)

Fig. 7. Halogenated compounds in ambient air in Salamanca, Mexico. February March, 2003.

E. Vega et al. / Fuel 90 (2011) 35093520

3517

in the center of this region. The meteorological model RAMS was used to calculate temperature, relative humidity, turbulence, wind speed and wind direction every 300 s for each cell. The topography and land use were taken into account for an adequate simulation of surface wind pattern [19]. The photochemical model CIT-SAPRC99 was used to solve the transport and mass conservation of 112 chemical species using the SAPRC99 chemical mechanism [40]. Modeling results were used as complement tool to improve the understanding of tropospheric ozone behavior throughout the study region where no measurements were available. The CIT-SAPRC99 uses operator splitting technique to solve transport and chemical processes separately in each time step. Vertical diffusion is modeled as a function of atmospheric stability class, with the mixing height entered as input to the model. The model was modied to account for three-dimensional elds of temperature and humidity as inputs, and to calculate reaction rates in three dimensions on the basis of these parameters [20]. The CIT model has been applied and tested extensively and reported elsewhere [4144]. The CIT uses a terrain-following coordinate system and 15 vertical layers were dened for this work, with the top boundary of the modeling domain at 4628 m above the surface. The height of each layer became smaller as it approached to the surface (e.g.

50 m for the lowest layer) for greater resolution. The wind elds obtained from RAMS model were interpolated to t the CIT-SAPRC99 grid domain, smoothed and ltered to improve mass consistency in the CIT-SAPRC99 model. The results obtained from the RAMS model indicated two different dispersion conditions during the study period. The rst and more frequent condition showed circulation of winds during early morning and night that converged towards the Salamanca city i.e. mountain-valley circulation (Fig. 9A). From 1000, the circulation shifted towards the mountains (Fig. 9B) with a NW wind direction at 1800 (Fig. 9C). After this time, the wind speed decreased and the wind direction returned towards the city (Fig. 9D). The second less frequent condition indicated a constant wind direction from the NE, from 1000 to 2000 (not shown). Such meteorological condition led to accumulation of emissions inside the valley and therefore the increase of pollution levels. The results of coupled application of RAMS and CIT-SAPRC99 models to estimate production and transport of ozone are illustrated in Fig. 10A-D, which describe ozone concentration indicated with background colored from blue (0 ppb) to red (200 ppb), wind elds (black arrows) and VOCs plumes (green zones) emitted from Len, Irapuato, Salamanca and Celaya cities (L, I, S and C yellow letters, respectively). Simulations were calculated for 0900, 1200,

(A) 0800.

(B) 1000.

(C) 1800.

(D) 2300.

Fig. 9. Surface wind eld determined by RAMS model for the study region on March 4, 2003 at (A) 0800; (B) 1000; (C) 1800 and (D) 2300.

3518

E. Vega et al. / Fuel 90 (2011) 35093520

I S C

(A) 0900.

(B) 1200.

(C) 1600.

(D) 2100.

Fig. 10. Hydrocarbon urban plumes (green) and ozone concentrations (from blue = 0 ppb to red = 200 ppb) modeled by coupled application of RAMS and CIT-SAPRC99 during March 4 2003 in Salamanca region. (A) 0900; (B) 1200; (C) 1600 and (D) 2100. Arrows indicate surface winds. Panel A shows the location of cities: L = Leon; I = Irapuato; S = Salamanca and C = Celaya. (For interpretation of the references to color in this gure legend, the reader is referred to the web version of this article.)

1600 and 2100. Fig. 10A shows that wind ow at 0900 was dominated by drainage ow of cold air descending from the mountains, transporting urban emissions to the lowest areas of the region. At his time of the day, high atmospheric stability conditions and thermal inversion caused that emissions were trapped near the surface and minimum ozone production. At noon, VOCs were diluted due to the rapid growth of the mixing height, so only a small plume is observed at L. Ozone production is evident in those regions where precursors were transported to early in the morning, while ozone concentrations were minimal in the urban areas. Fig. 10C shows that wind pattern at 1600 has changed, transporting air masses from southwest to northeast. The ozone formation was well developed in the whole region and the maximum ozone concentrations were calculated in the NE of the models domain. Again, at night (Fig. 10D), cold air descending from the mountains began to dominate the wind circulation in the region. As expected, the CIT-SAPRC99 predicted moderate ozone concentration (40 60 ppb) remaining at surface level in the NE rural zones due to

insignicant of NOX sources in that region and again VOCs concentrations became signicant at surface level in the urban zones. The above described behavior was frequently observed in the region, especially when weak synoptic conditions generated valley-mountain circulation. Nevertheless, it is not possible to determine a general wind pattern due to the limited time of the study; it would be necessary to carry out the analysis for other season of the year. Finally, overall observed and modeled results of this survey indicate environmental implications due to the exposition of population to secondary pollutants and halogenated hydrocarbons, especially in the rural areas located northeast of the industrial and urban zones. As a consequence of this study, local and federal environmental authorities in collaboration with PEMEX (the National Oil Company) and the Comisin Federal de Electricidad (Federal Electricity Bureau) developed the rst Air Quality Program for the region (20032006). Main achievements of this program were the reduction in number of days with sulfur dioxide concentrations

E. Vega et al. / Fuel 90 (2011) 35093520

3519

above the standard, from 79 in 2003 to 34 in 2006. Later on, the program included denitive removal of methyl bromide usage for agricultural activities [24].

Acknowledgements This study was supported by Mexico national oil company (PEMEX) under contract IMP-ZC-001-2003. The authors are grateful to the authorities of Salamanca renery by their help and support provided during the study.

4. Conclusions A 2-week eld campaign of measurements of criteria pollutants, VOC and meteorological parameters was carried out during FebruaryMarch 2003 in the Salamanca city which is located inside one of the most important industrial corridors of Mexico, to characterize air quality in the urban and regional scales. Special emphasis was given to the chemical characterization of VOC to gain knowledge about levels, distribution and origin of these species. In spite of limited period of time, this monitoring effort currently represents one of the biggest studies on air quality realized in Mexico, without considering the most recent studies carried out in the metropolitan area of Mexico City, which allows investigating the air quality condition in an area highly inuenced by industrial emissions. Regarding one of the most important pollutants, the SO2 showed a high concentration event at 0200 at the urban site, located 2 km west of the Renery and the Power Plant. The levels of this pollutant reached 0.60 ppm (1 min sampling-time). Historically, extremely high levels of SO2 have been frequently observed from 0000 to 0300 in the city, attributed to venting operations of the Renery and the Power Plant. However since 2003 control measurements have been implemented to reduce levels of SO2 and particles in the atmosphere, such as the shifting to cleaner fuels (with lower content of S) in both the Power Plant and the Renery. These measurements were already in execution by the time of the eld campaign, thus it is possible that the rst positive results were observed. The highest ozone levels were observed in the rural monitoring sites, which are explained by transport and possible inputs from the free troposphere, given the high altitude of the study region. Total VOCs mass concentration uctuated from 170 ppbC in the rural area to 956 ppbC in the urban sites. The more abundant group was the alkanes followed by aromatics, olens and halogenated. The total mass concentration of VOCs is comparable with values reported for some urban and rural sites of Mexico City Metropolitan Area. However, the concentration of halogenated species (particularly Freon-114, methyl bromide, 1,2-dichloroethane and 1,2-dichloropropane) was notoriously higher in rural sites of Salamanca compared with Mexico City and other highly polluted areas of China. Formaldehyde and acetaldehyde showed average concentrations of 4.3 and 2.7 ppb, respectively which are lower than those reported for Mexico City (13.3 and 4.4 ppb respectively). Due to these low values, it is expected little contribution of carbonyl species to the photochemical activity in the region. The origin of VOCs was inferred by examining the concentration and spatial distribution of species, in conjunction with cross correlation analyses and relationship of marker species for specic sources. In the urban area, gasoline and diesel vehicular exhaust and marketing/leaking of LPG dominated the VOCs burden, while emissions from industrial and agriculture activities were important in the rural area. The last two sources emit halogenated species that represent potential hazard to human health due to the toxic and/or carcinogenic effect, thus the results of this study highlighted the needed of monitoring and even implementing control measurements for these species. As a consequence, federal and local environmental authorities removed usage of methyl bromide in agricultural activities since 2007. Meteorological conditions generally favoured the transport of pollutants outside Salamanca towards northeast, where the highest concentration of ozone were observed. This result was in agreement with the modeling calculations.

References
[1] Shahriar S, Topal E. When will fossil fuel reserves be diminished? Energy Policy 2009;37:1819. [2] Interngovermental Panel on Climate Change (IPCC). Climate change 2001: the scientic basis. Contribution of the working group I to the third assessment report of the interngovermental panel on climate change. Technical summary. Cambridge International Press; 2001. [3] Decker EH, Elliot S, Smith FA, Blake DR, Rowland FS. Energy and material ow through the urban environment. Ann Rev Energy Environ 2000;25:685740. [4] Molina LT, Kolb CE, de Foy B, Lamb BK, Brune WH, Jimenez JL. Air quality in North Americas most populous city overview of the MCMA-2003 campaign. Atmos Chem Phys 2007;7:244773. [5] Air Resources Board of California, USA. <http://www.arb.ca.gov/toxics/id/ taclist.htm> [cited 22.01.11]. [6] Secretara de Energa. Prospectiva de petrolferos 20062015. ISBN: 968-874201-5 DF, Mexico; 2006. [7] Instituto Nacional de Estadtica Geografa e Informtica. XII Censo General de Poblacin y Vivienda 2000. DF, Mxico; 2001. [8] SEMARNAT-IEG. Programa para mejorar la calidad del aire en la ciudad de Salamanca (PROAIRE). Peridico Ocial del Estado de Guanajuato No. 9, 6 junio: Mxico; 2003. [9] Seinfeld JH, Pandis SN. Atmospheric chemistry and physics: from air pollution to global change. New York (NY): John Wiley and Sons; 1998. [10] Jacobson MZ. Atmospheric pollution, history, science, and regulation. United Kingdom: Cambridge University Press; 2002. [11] Jenkin ME, Clemitshaw KC. Ozone and other secondary photochemical pollutants: chemical processes governing their formation in the planetary boundary layer. Atmos Environ 2000;34:2499527. [12] Atkinson R. Atmospheric chemistry of VOCs and NOX. Atmos Environ 2000;34:2063101. [13] Atkinson R, Arey J. Gas phase tropospheric chemistry of biogenic organic compounds: a review. Atmos Environ 2003;37:S197119. [14] Odum JR, Hoffmann T, Bowman F, Collins D, Flagan RC, Seinfeld JH. Gas/ particle partitioning and secondary organic aerosol yields. Environ Sci Technol 1996;30:25805. [15] Molina MJ, Rowland FS. Stratospheric sink for chlorouoromethanes: chlororine atom catalysed destruction of ozone. Nature 1974;249:8102. [16] Vega E, Ruiz H, Martinez-Villa G, Sosa G, Gonzalez-Avalos E, Reyes E, et al. Fine and coarse particulate matter characterization in a heavily industrialized city in central Mexico during Winter 2003. J Air Waste Manage Assoc 2007;57:62033. [17] Environmental Protection Agency. Compendium method TO-14a. Determination of volatile organic compounds (VOCs) in ambient air using specially prepared canisters with subsequent analysis by gas chromatography (EPA/625/R-96/010b). Compendium of methods for the determination of toxic organic compounds in ambient air. 2nd ed. Cincinnati (OH); 1999. [18] McCleanny WA, Plei JD, Evans GF. Canister based method for monitoring toxic VOCs in ambient air. J Air Waste Manage Assoc 1991;41:130818. [19] Pielke RA. A comprehensive meteorological model system: RAMS. Meteorol Atmos Phys 1992;49:6991. [20] West JJ, Zavala MA, Molina LT, Molina MJ, San Martini F, McRae GJ, et al. Modeling ozone photochemistry and evaluation of hydrocarbon emissions in the Mexico City metropolitan area. J Geophys Res 2004. doi:10.1029/ 2004JD004614. [21] Vingarzan R. A review of surface ozone background levels and trends. Atmos Environ 2004;38:343142. [22] Kalabokas PD, Repapis CC. A climatological study of rural surface ozone in Central Greece. Atmos Chem Phys 2004;4:113947. [23] Kalabokas PD, Volz-Thomas A, Brioude J, Thouret V, Cammas JP, Repapis CC. Vertical ozone measurements in the troposphere over the Eastern Mediterranean and comparison with Central Europe. Atmos Chem Phys 2007;7:378390. [24] SEMARNAT/Gobierno del Estado de Guanajuato. Progama para mejorar la calidad del aire en Salamanca 20072012. Mexico; 2007. [25] Velasco E, Lamb B, Westberg H, Allwine1 E, Sosa G, Arriaga-Colina JL, et al. Distribution, magnitudes, reactivities, ratios and diurnal patterns of volatile organic compounds in the Valley of Mexico during the MCMA 2002 & 2003 eld campaigns. Atmos Chem Phys 2007;7:32953. [26] Chan CY, Tang JH, Li YS, Chan LY. Mixing ratios and sources of halocarbons in urban, semi-urban and rural sites of the Pearl River Delta, South China. Atmos Environ 2006;40:733145. [27] Barletta B, Meinardi S, Simpson IJ, Rowland FS, Chan C, Wang X, et al. Ambient halocarbon mixing ratios in 45 Chinese cities. Atmos Environ 2006;40:770619.

3520

E. Vega et al. / Fuel 90 (2011) 35093520 [37] de Gouw JA, Welsh-Bon D, Warneke C, Kuster WC, Alexander L, Baker AK, et al. During the MILAGRO study. Atmos Chem Phys 2006;2009:9.34253442. [38] International Program on Chemical Safety [Internet] Chemical Safety Information from Intergovernmental Organizations. <http://www.inchem.org/> [Updated 21.09.10; cited 21.01.11]. [39] Instituto Mexicano del Petrleo. Estudio local y regional del impacto de las emisiones de PEMEX y CFE en la calidad del aire en Salamanca, Gto. Final Report. PS-MAS-IF-F21026-02. DF, Mxico; 2003. [40] Carter WPL. Documentation of the SAPRC-99 chemical mechanism for VOC reactivity, nal report to California Air Resources Board, Contract 92-329 and 95-308, Calif. Air Res. Board, Sacramento, Calif. USA. <http://www.cert.ucr. edu/~carter> [cited 19.03.11]. [41] McRae GJ, Goodin WR, Seinfeld JH. Development of a second-generation mathematical model for urban air pollution: 1. Model formulation. Atmos Environ 1982;16:67996. [42] Los Alamos National Laboratory and Instituto Mexicano del Petroleo. Mexico city air quality research initiative (MARI), rep LA-12699, Los Alamos, N.M.; 1994. [43] Harley RA, Sawyer RF, Milford JB. Updated photochemical modeling for Californias south coast air basin: comparison of chemical mechanisms and motor vehicle emission inventories. Environ Sci Technol 1997;31:282939. [44] Marr LC, Noblet GS, Harley RA. Formation of photochemical air pollution in central California: 2. Impact of revised emissions on Eulerian model predictions. Geophys Res 2002;107(D6):4048. doi:10.1029/2001JD00069.

[28] Lei W, de Foy B, Zavala M, Volkamer R, Molina LT. Characterizing ozone production in the Mexico City Metropolitan Area: a case study using a chemical transport model. Atmos Chem Phys 2007;7:134766. [29] Grutter M, Flores E, Andraca-Ayala G, Baez A. Formaldehyde levels in downtown Mexico City during 2003. Atmos Environ 2005;39:102734. [30] MIT-CAM. The MCMA Field Campaign; 2003. <http://mce2.org/fc03/ fc03.html> [cited 2010 August]. [31] Vega E, Mugica V, Carmona R, Valencia E. Hydrocarbon source apportionment in Mexico City using the chemical mass balance receptor model. Atmos Environ 2000;34:41219. [32] Blake DR, Rowland FS. Urban leakage of liqueed petroleum gas and its impact on Mexico City Air Quality. Science 1995;269:9536; Watson JG, Chow JC, Fujita EM. Review of volatile organic compounds source apportionment by chemical mass balance. Atmos Environ 2001;35:156784. [33] Fujita EM. Hydrocarbon source apportionment for the 1996 Paso del Norte Ozone Study. Sci Total Environ 2001;276:17184. [34] Environmental Protection Agency [Internet]. USA: SPECIATE4.2 data browser. <http://www.epa.gov/ttn/chief/software/speciate/index.html> [updated 2008 November; cited 2001 January 21]. [35] Cox ML, Fraser PJ, Sturrock GA, Siems ST, Porter LW. Terrestrial sources and sinks of halomethanes near Cape Grim, Tasmania. Atmos Environ 2004;38:383952. [36] Crounse JD, DeCarlo PF, Blake DR, Emmons LK, Campos TL, Apel EC, et al. Biomass burning and urban air pollution over the Central Mexican Plateau. Atmos Chem Phys 2009;9:492944.