Fuel 90 (2011) 3509–3520

Contents lists available at ScienceDirect

Fuel

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

Air quality assessment in a highly industrialized area of Mexico: Concentrationsand sources of volatile organic compounds

Elizabeth Vega, Gabriela Sánchez-Reyna ⇑, Virginia Mora-Perdomo, Gustavo Sosa Iglesias, José Luis Arriaga,Teresa Limón-Sánchez, Sergio Escalona-Segura, Eugenio Gonzalez-AvalosInstituto Mexicano del Petróleo, Eje Central Lázaro 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

Article history:Received 24 August 2010Received in revised form 29 March 2011Accepted 31 March 2011Available online 24 April 2011

Keywords:VOCsAir qualityIndustrial areasHalocarbonsMexico

0016-2361/$ - see front matter � 2011 Elsevier Ltd. Adoi:10.1016/j.fuel.2011.03.050

⇑ Corresponding author. Tel.: +52 55 91757558.E-mail address: gsanchez@imp.mx (G. Sánchez-Re

Parallel to the economical benefits brought by the oil industry in Mexico, there have been some negativeenvironmental effects due to emission of pollutants to the atmosphere. Salamanca, a city located insideone of the most important industrial corridors of the country, has been frequently affected by elevatedconcentrations of sulfur dioxide and particle matter. However, little is known about volatile organic com-pounds (VOCs), which in this study are analyzed along with criteria pollutants and meteorologicalparameters during February–March 2003 at urban, suburban and rural sites. Although sulfur dioxideaverage levels were �0.017 ppm, a high concentration event (�0.600 ppm), attributable to emissionsfrom the oil refinery 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 constitutedby 40% alkanes, 13% aromatics, 11% olefins and 11% of halogenated. The most abundant species were pro-pane (167 ± 40 ppbC), n-butane (91 ± 23 ppbC), toluene (51 ± 10 ppbC) and i-pentane (44 ± 7 ppbC), thatare related to combustion processes. Freon-114, methyl bromide and 1,2-dichloroethane which are likelyemitted by application of pesticides, soil fumigation and fabrication of chemicals, showed high concen-trations (48 ± 10, 50 ± 10 and 32 ± 6 ppbC respectively) in the rural sites, highlighting the importanceof control measurements implementation for these species, as they represent a potential hazard for pub-lic health. Moreover, these halocarbons showed similar ratios regardless the monitoring site, suggestingsame source. Modeling results indicated that meteorological conditions generally transport air masses tothe northeast rural areas where the highest concentrations of ozone were calculated.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction

Over the last century, the oil industry has emerged as the pri-mary energy source [1]. Currently, the life style of human societiesdepends on energy (electricity generation, natural gas, crude oiland its refined products, coal, etc.); without it, societies as weknow them would collapse. Even though the oil industry has madeimportant contributions to the global economy, usually this hasbeen accompanied with negative environmental impacts from avariety of activities such as oil drilling, refinery, oil spillage, gasand flaring. Moreover, deterioration of the environment may notbe circumscribed to the local scale, it can reach regional and globalextent due to the emission of precursors of secondary pollutantsand chemical species that contribute enhancing global warmingand stratospheric ozone depletion [2–4]. Public health may alsobe affected if emissions contain toxic or carcinogenic species [5].In recent years concern over public health and environmental pro-tection has become a critical issue, this means that a growing

ll rights reserved.

yna).

amount of investment and effort is dedicated to reconcile the envi-ronment and development of countries.

The economy of Mexico strongly depends on oil industry; in2005 the primary distillation capacity (1540 MBD) ranked thecountry on the top 15 worldwide and 4 in Latin America. The crudeis processed in 6 oil refineries which mainly produce gasoline, die-sel, jet fuel, coal, asphalt, and lubricants. The third most importantrefinery is the Ing. Antonio M. Amor, which processes 197 MBD [6].The refinery and a variety of industries constitute one of the mostimportant industrial corridors of Mexico, known as the Bajío Indus-trial Corridor (BIC) located in the State of Guanajuato, in the centralarea of the country (Fig. 1). The BIC has nearly 465 industries, frommedium to large size, including Chemical, Power Generation, FoodProcessing, Textile and Metal-mechanic [7].

The Salamanca city, with 250,000 inhabitants, is located at20�3400900N latitude and 101�1105100W longitude, at 1720 m abovemean sea level [8]. The Salamanca municipality encompasses atotal area of 774 km2. The agriculture is now the second mostimportant economical activity, with a designated area of about80% of the municipality. The impact of agriculture on the environ-ment is important, especially for the use of fertilizers, pesticides,

GuanajuatoState

SalamancaCounty

(B)Zacatecas

Sn. Luis Potosí

Jalisco

Michoacán

Querétaro

León

Irapuato

Salamanca

Celaya

(A)

Fig. 1. Geographical location of: (A) Salamanca County and (B) The Bajío IndustrialCorridor.

3510 E. Vega et al. / Fuel 90 (2011) 3509–3520

deforestation in the uplands and post-harvesting burning. The big-gest impact to the atmosphere is represented by the post-harvest-ing burning, due to the emission of large amounts of ozoneprecursors and particulate matter [8].

According to the 2000 BIC Emissions Inventory, the emissions ofparticles (PM10), sulfur dioxide (SO2), carbon monoxide (CO), nitro-gen oxides (NOX) and hydrocarbons (HC) were 71,443, 112,480,1,650,772, 142,183 and 260,296 tons per year, respectively. It isestimated that PM10 are released to the atmosphere mainly bycommercial and service activities, SO2 and NOX by electricity gen-eration, and CO and hydrocarbons by vehicle exhaust. Salamancacontributes with 18% of PM10, 92% of SO2, 8% of CO and HC and14% of NOX of the total BIC emissions. Thus, Salamanca is by farthe major generator of SO2 emissions in the region; while other cit-ies release ozone precursors [8].

Parallel to the economical development of the BIC, there havebeen some adverse environmental impacts which have broughtthe attention of government agencies, civil and private associa-tions. As a result, since 2000 a Monitoring Network in Salamancaroutinely measures CO, SO2, NO2, O3, and PM10. According to localenvironmental authorities, the SO2 air quality standard, AQS,(0.13 ppm in a 24 h average, no more than once per year) was ex-ceeded 13%, 24% and 22% of days in 2000, 2001 and 2002 respec-tively in downtown Salamanca, mainly during winter. Theexceedences of other gaseous pollutants is less frequent, for in-stance, 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 gas-eous and particle contaminants that are not included in the local

monitoring network, such as the volatile organic compounds(VOCs). The negative effects of VOCs on the environment and pub-lic health are well documented. From the environmental point ofview, some VOCs (i.e. olefins and aromatics which are mainlyanthropogenic) are reactive species that break out the naturalequilibrium of generation–destruction of tropospheric ozone, thusthe concentrations of this compound and other photochemically-produced pollutants are frequently high in the urban environment.Besides, the reactive organic gases can partition into the aerosolphase generating secondary organic aerosols. Other importantgroup of VOCs is constituted by the halogenated species, whichare originated almost exclusively from anthropogenic emissionsdue to its usage as an industrial solvent and degreaser. Some ofthese compounds have been the focus of intensive research, suchas the chlorofluorocarbons due to their participation in the strato-spheric ozone depletion. In addition, many of the halogenated spe-cies represent a potential hazard to human health due to the toxicand/or carcinogenic effect [9–15].

The public opinion on air quality deterioration in Salamanca,encouraged PEMEX (National Oil Company) to support an exten-sive 2-week monitoring field study with the aim of augmentingthe knowledge of sources, transport and fate of air pollutants inthe region, therefore effective control measurements of atmo-spheric pollution can be designed. The main findings of such cam-paign are presented in this work, particularly the chemicalcharacterization, distribution and origin of VOCs, as well as themeteorological parameters that influence the dilution and trans-port of pollutants. The later was also estimated by applying a 3 Dair quality model.

2. Field campaign

As an outcome of a collaborative effort, the Instituto Mexicanodel Petroleo (IMP), the Instituto de Ecología de Guanajuato, theInstituto de Investigaciones Científicas at the Universidad de Guan-ajuato, the Centro de Ciencias de la Atmósfera at the UniversidadNacional Autónoma de México, the Patronato de Salamanca, andthe Ing. Antonio M. Amor Oil Refinery, with the PEMEX sponsor-ship, carried out a field monitoring campaign, from February 21to March 9 2003. The main objectives were to chemically charac-terize the air pollution in the urban area of Salamanca in boththe particle and gas phases, and to assess the potential impact ofpollutants in the regional scale. The interested reader can consultVega et al. [16] for the particulate matter results found in thisregion.

The first week of measurements was focused on the character-ization of urban air quality; while the second week was designed toevaluate the regional impact of urban emissions. The monitoringsites 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 loca-tion, description, sampling period and measurements performed.

Three automated samplers (VOCCS-ANDERSEN and AVOCS-ANDERSEN models) with a Viton diaphragm pump were used tocollect VOCs (defined in this work as hydrocarbons from C2 toC12) in canisters over 12 h period (0600–1800 and 1800–0600 localtime) in the urban sites and 24 h period in the rural/boundary sites.A total of 80 canisters were analyzed in the Laboratory by cryo-genic pre-concentration/high-resolution GC technique, similar tothe TO-14A protocol [17].

Water Sep-Pak DNPH-Silica cartridges were used to trap car-bonyl species. Twelve samples were taken during the first weekof the campaign at Cruz Roja (CR) urban site from 0600 to 0900and from 1200 to 1500. The derivatives were eluted and analyzedby HPLC with UV photodiode array detector according to the TO-11A protocol [18]. Criteria pollutants were measured using a

(A)

(B)

-101.30 -101.25 -101.20 -101.15 -101.10 -101.05 -101.0020.50

20.55

20.60

20.65

20.70

CG

VAUS

CA

CRNA

DIF

RT

-101.6 -101.5 -101.4 -101.3 -101.2 -101.1 -101.0 -100.9 -100.8 -100.720.1

20.2

20.3

20.4

20.5

20.6

20.7

20.8

20.9

21.0

VS

PN

JR

SI

MI

SALSAL

Fig. 2. Monitoring sites location: (A) Urban Area: Cruz Roja (CR) and DIF; suburban: Cárdenas (CA), Cerro Gordo (CG), Universidad la Salle (US), and Valtierrilla (VA); theRefinery (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.

E. Vega et al. / Fuel 90 (2011) 3509–3520 3511

mobile Lab equipped with conventional analyzers (Monitor Labs).Methods used to determine these pollutants were NOM-034 –SEMARNAT – 1993 using dispersive spectroscopy for CO (detectionlimits from 0 to 50 ppm); NOM-037 – SEMARNAT – 1993 usingquimioluminescence for NOx (detection limits from 0 to0.50 ppm); USEPA – EQOA – 0193 – 091 using UV photometryfor O3 (detection limits from 0 to 1.0 ppm) and USEPA – EQSA –0193 – 092 using pulse fluorescence for SO2 with detection limitsfrom 0 to 1.0 ppm.

Along the field campaign, the surface meteorological parame-ters 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 fourin the boundary sites. The vertical thermodynamic profile variables(P, T, RH and horizontal wind vector) were measured using a Dig-icora II radiosonde system from Vaisala (Model SPS-220). Threeradiosondes were launched every day at 0800, 1200 and 1800 atSI, JR and VS sites. The information was used to determine the mix-

ing height based on potential temperature and specific humidityprofiles, and also as input for the mesoscale meteorological modelRAMS [19] and the 3D air quality model [20].

3. Results

3.1. Criteria pollutants

The average and standard deviation for 1 h average concentra-tion of criteria pollutants (CO, NO2, SO2 and O3) and nitrogen oxi-des (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 atmo-spheric concentrations of CO, NO2 and NOX were considerablyhigher at the urban sites compared with the suburban and ruralones. The suburban sites are surrounded by crop fields and un-paved small roads with little traffic of heavy-duty vehicles whichmay influence the measurements, however it is expected to beminimal.

Table 1Location and type of monitoring sites during the monitoring field Campaign in Salamanca, Mexico. February–March 2003.

Site name, code andcoordinates

Site type Site description Measurementperiod

Measurements

Cruz Roja (CR) 20.58�N,101.20�W

Urban Located in the main street of Salamanca with high vehicular (light and duty) traffic allday; 2 km west of the Refinery and the Power Plant

February 21–March 9 2003

Criteria pollutantsa,VOCs, carbonylspecies

DIF 20.56�N, 101.20�W Urban Located in a residential area with high vehicular traffic, mainly of gasoline-poweredvehicles

February 21–March 9 2003

Criteria pollutantsa

Cárdenas (CA) 20.63�N,101.22�W

Suburban The site is in a crop field near an unpaved road with little vehicular traffic February 22–28 2003

Criteria pollutantsa

Cerro Gordo (CG)20.59�N, 101.13�W

Suburban 7 km southwest of the Refinery and Power Plant, 200 m of a highway used mainly byheavy-duty vehicles. The site is surrounded by crop fields

February 21–28 2003

Criteria pollutantsa,VOCs

Valtierrilla (VA)20.56�N, 101.13�W

Suburban Located in a populated community with moderate transit of gasoline and dieselvehicles

February 22–28 2003

Criteria pollutantsa

Universidad La Salle(US) 20.55�N,101.23�W

Suburban Located inside the University, southwest Salamanca; site is surrounded by vegetation;400 m away from a four-lane road with light and duty-vehicles traffic

February 21–28 2003

Criteria pollutantsa,VOCs

Silao (SI) 20.59�N,101.42�W

Rural/boundary

The site was used for meteorological measurements, located northwest Salamanca March 2–92003

Radiosondes

Valle de Santiago (VS)20.35�N, 101.20�W

Rural/boundary

Located inside a natural protected area known as Siete Luminarias. The site is 300 mabove the Salamanca level, 1.5 km away from a road with little vehicular traffic

March 2–92003

Criteria pollutantsa,radiosondes

Pueblo Nuevo (PN)20.55�N, 101.35�W

Rural/boundary

The site is surrounded by crop fields near one unpaved road with little light-dutyvehicular traffic

March 2–92003

Criteria pollutantsa,radiosondes

Juventino Rosas (JR)20.64�N, 101.00�W

Rural/boundary

22 km northeast Salamanca, in a commercial area with buildings 4–6 m height withmoderated vehicular traffic

March 1–92003

Criteria pollutantsa,VOCs, radiosondes

Mirandas (MI)20.56�N, 101.14�W

Rural/boundary

This monitoring site is located 6.5 km north Salamanca, in an wide-open areasurrounded by crop fields

March 1–92003

Criteria pollutantsa,VOCs

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

Table 2Basic 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 Suburban Rural

Average Maximum Minimum SD n Average Maximum Minimum SD n Average Maximum Minimum SD n

O3 0.022 0.094 0.001 0.018 699 0.021 0.072 0.001 0.017 353 0.034 0.114 0.001 0.020 525CO 2.200 19.500 0.960 1.150 699 0.158 1.890 0.001 0.240 347 0.038 0.570 0.001 0.067 525SO2 0.017 0.310 0.001 0.043 699 0.016 0.269 0.001 0.034 294 0.011 0.318 0.001 0.026 525NO2 0.022 0.064 0.002 0.013 699 0.014 0.047 0.002 0.008 353 0.010 0.046 0.001 0.008 525NOX 0.055 0.425 0.002 0.045 699 0.029 0.127 0.001 0.028 344 0.014 0.092 0.001 0.013 525

3512 E. Vega et al. / Fuel 90 (2011) 3509–3520

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

0.00

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0 3 6 9 12 15 18 21CR MI CG VS

O3

(ppm

)

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

mum of 0.052 which could be considered as background concen-trations. These results are within the range of ozone levelsreported at rural areas of Canada [21]. In addition, it is known thatareas of high altitude (the study region is 1720 m above sea level)could be influenced by inputs from free troposphere [21–23]. As itis discussed later, the peak of mixing height was observed every-day at 1800 reaching 3500 m above ground level, making possiblethis phenomenon, although more measurements would be neces-sary to reach a conclusion.

The highest ozone concentrations were measured at rural MIwhich is located 6.5 km north of the city. To investigate the influ-ence of surface wind direction and speed on the spatial distributionof ozone, an analysis was performed dividing the data-set into foursubsets according to surface wind direction: north (315–45�), east(45–135�), south (135–225�) and west (225–315�). As expected,the results reflected the influence of site location regarding mainpollution sources. The concentrations of ozone at MI were higherwhen the wind blew from the west and south, i.e., the site wasdownwind major pollution sources. These wind directions wereobserved �64% of the time, mostly from 1200 to 1900, which sug-gest that ozone levels at MI were strongly influenced by trans-ported ozone. Moreover, the correlation analysis indicated thatlevels observed at MI and CR are associated during this meteoro-logical condition (R2 = 0.72). The same analysis was performedusing data of the opposite wind direction (north), showing an evenstronger relationship (R2 = 0.91) between the urban and ruralozone levels. It has to be added that north winds generally blew

March 2, 2003

SO2

(ppm

)

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

00:00 03:00 06:00 09:00 12:00 15:00 18:00 21:00 24:00

E S-W

NNE S-SW

N SW N

erraticMI

JR

CR

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

E. Vega et al. / Fuel 90 (2011) 3509–3520 3513

from 0000 to 0700 and 2100 to 2300, thus measurements indicatebackground concentrations. The same analysis was performed forurban CR and rural VS; the data-set was filtered so only measure-ments taken when the rural site was located downwind Salaman-ca. A correlation coefficient of 0.55 was obtained between thesesites (the site is 25 km south), while similar coefficient was ob-tained for the analysis using only winds from the south(R2 = 0.57). In summary, the results suggest that transport of airmasses from downtown Salamanca and the elevation of the studyarea play an important role in the high ozone concentrations ob-served in the rural sites.

To evaluate the effect of wind direction on the concentration ofthe other pollutants, the average and standard deviation of concen-trations of criteria pollutants and nitrogen oxides were calculatedfor each of the wind sectors above described. At CR, carbon monox-ide and nitrogen dioxide registered higher levels when the windblew from the south or east, which is explained by the locationof this site in the northwest of Salamanca. On the other hand,ozone levels at CR were statistically lower when the wind blewfrom the north and higher when the wind blew from east. The con-centrations of SO2 showed higher values associated with east-winds (where the refinery is located) and lower for any other winddirection. At suburban CG, carbon monoxide and nitrogen dioxidewere higher when the wind blew from the east or south, as it wasenvisaged. The concentrations of pollutants were higher when thewind came from the south at the rural site, which is explained bythe site’s location (Table 3). Moreover, the observed results agreewith simulation results of transport of air masses, which are dis-cussed in Section 3.4.

As mentioned before, the largest sources of SO2 in the study re-gion are the Refinery and the Power Plant facilities, both located in-side the Salamanca city. Even though emissions are released to theatmosphere by elevated stacks (30–60 m above the ground), thenatural atmospheric processes such as turbulence, thermal inver-sions and high pressure systems, may increase the concentrationand residence time of pollutants. Fig. 4 exemplifies the typical dailyvariation of SO2 concentration. From 0000 to 0600, due to stableatmospheric conditions, the highest concentrations were observedat CR, located 2 km west of major emission sources (the other twosites, MI, and JR, are 6.5 km north and 22 km northeast, respec-tively). From 0600 to 0900 the wind blew from the east, thereforeCR site received directly the SO2 emissions. Around 0900, the sur-face wind direction blew from SSW, remaining under these condi-tions the rest of the day, as a result, concentrations at CR showed adecrement while at MI an increment was registered. It was foundthat short-term variability periods (e.g. at CR at 0900–0930 andat MI at 1200–1220 in Fig. 4) were associated to wind speed lowerthan 1.0 ms�1. According to RAMS modeling results, surface windsbefore 0900 were driven by cold air draining from the near moun-

Table 3Wind directional analysis for concentrations of criteria pollutants (ppm) du

Site Species Wind direction

North (315–45�) East (45

CR O3 0.016 ± 0.021 0.022 ±CO 1.946 ± 0.927 2.219 ±NO2 0.032 ± 0.010 0.037 ±SO2 0.020 ± 0.046 0.122 ±

CG O3 0.024 ± 0.021 0.012 ±CO 0.161 ± 0.229 0.383 ±NO2 0.016 ± 0.008 0.020 ±SO2 0.031 ± 0.046 0.010 ±

MI O3 0.016 ± 0.009 0.015 ±CO 0.074 ± 0.151 0.068 ±NO2 0.011 ± 0.005 0.012 ±SO2 0.004 ± 0.007 0.004 ±

tains towards lowest topographic levels in the basin, accumulatingair pollutants over the city. After 0900 and due to synoptic windforcing, the wind blew from SW in the whole area, sweeping pollu-tants out of the city. The estimated results were validated againstmeteorological measurements finding good correspondence be-tween observed and predicted values.

During the monitoring campaign, high SO2 concentrations(>0.60 ppm, 1 min sampling-time) were registered once at CR at0200; such high concentrations were associated with persistenteast wind direction (82–95�) and wind speed of 2.6 ms�1, whichwas higher than the average registered at this time of the night(1.3 ms�1). Moreover, concentrations above the average of toluene,m,p-xylenes, benzene, Freon-114 and methyl chloroform were alsoregistered during this event, suggesting emissions from industrialactivities.

The night-time high SO2 concentration events were frequentlyobserved in 1999–2002 and were associated with venting activitiesof the Refinery and/or Power Plant [8]. However since 2003 the airquality program began to operate in the region, tackling SO2 andparticle matter problems mostly [24]. The control measurementsinclude the usage of fuels with low sulfur content in the powerplant and an increment in the natural gas consumption thus is pos-sible that the first positive results were observed during the fieldcampaign.

3.2. Volatile organic compounds

3.2.1. ConcentrationsApproximately 200 chemical species of VOCs were identified

and quantified by GC analysis. A classification was made according

ring February–March 2003 in Salamanca, Mexico.

–135�) South (135–225�) West (225–315�)

0.021 0.018 ± 0.017 0.021 ± 0.0191.104 2.937 ± 1.604 2.201 ± 1.0940.010 0.034 ± 0.012 0.028 ± 0.0100.139 0.012 ± 0.034 0.008 ± 0.020

0.012 0.021 ± 0.018 0.020 ± 0.0190.518 0.264 ± 0.404 0.158 ± 0.2750.010 0.016 ± 0.009 0.014 ± 0.0080.019 0.030 ± 0.058 0.030 ± 0.049

0.010 0.029 ± 0.025 0.044 ± 0.0210.121 0.129 ± 0.149 0.093 ± 0.1360.006 0.017 ± 0.008 0.009 ± 0.0070.010 0.032 ± 0.063 0.014 ± 0.047

3514 E. Vega et al. / Fuel 90 (2011) 3509–3520

to the functional group, indicating that alkanes were the mostabundant group (�40% of the total mass), followed by aromatics(�13%), olefins (�11%) and halogenated compounds (�11%). Thetotal 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 locatedsouthwest of the city. The concentrations of VOCs taken from1800 to 0600 (nocturnal samples) were higher than those mea-sured in the morning, reaching 956 ± 120 ppbC at CR and554 ± 95 ppbC at US. Moreover, diurnal samples showed differentcomposition compared to nocturnal samples: the proportion of al-kanes was higher during the night at CR, while halogenated andthe unidentified groups showed higher percentages at night inthe suburban sites (Fig. 5A and B). The higher levels of VOCs duringthe night period were mainly driven by the increase of emissions ofpropane, n-butane, i-butane, Freon-114 and toluene. Velasco et al.[25] reported that VOCs levels for the Mexico City MetropolitanArea during 2003 fluctuated from 2473 in the industrial area to83 ppbC in the rural site for morning samples and from 1467 to81 ppbC for afternoon samples. A simple comparison between val-ues of VOCs measured in Mexico City and Salamanca, even thoughof the difference in sampling-time, indicated that concentrationsobserved in Salamanca were similar with some sites of MexicoCity. Regarding the distribution of VOCs by type, it was observedthe same order of abundance (i.e. alkanes > aromatics > olefins),although the proportion of olefins in Salamanca was twice the con-centration reported for Mexico City.

The highest concentrations of individual VOCs measured in theatmosphere of Salamanca during the sampling campaign are de-scribed in Table 4. At the urban and suburban sites, propane wasthe most abundant species, followed by n-butane, toluene and i-pentane. The first two species are attributable to the wide usageof LPG for cooking and heating; on the other hand, toluene and i-

39 38 36 41 46

20

8 16 109

19

12

16 1413

4

9

7 1011

1

1

2 43

927

21 16 13

7 5 3 5 6

0

20

40

60

80

100

CR CG MI JR USalkanes olefins aromatics halogenated

oxigenated unidentified HC2

%

(A)

5044

28

14

5

12

17

9

5

4

17

28

1

1 1

921 24

5 4 1

0

20

40

60

80

100

CR USalkanes olefins aromatics halogenated

oxigenated unidentified HC2

%

(B)

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

pentane may be emitted by both mobile and industrial sources.Propane and n-butane are reported as the most abundant speciesin the atmosphere of Mexico City [25], showing concentrationssimilar to those observed in this study.

The halogenated species such as Freon-114, methyl bromide,1,2-dichloroethane, 1,2-dichloropropane and vinyl chlorideshowed levels noticeably high at the suburban/rural sites, espe-cially during the night time. The concentrations of these speciesare markedly higher than those observed in an industrial area ofChina, where Freon-114 and methyl bromide are �16 and �18pptv [26,27]. Therefore, the results found in this study highlightthe importance of carrying out continuous monitoring of VOCs,so control measurements can be taken to reduce populationexposure.

Formaldehyde, acetaldehyde and acetone average concentra-tions were 3.75 ± 1.94, 2.45 ± 1.45 and 7.71 ± 6.38 ppb, respec-tively. Formaldehyde may have a significant influence on thelocal photochemistry, more than any other carbonyl species [28].However, concentrations of formaldehyde were low in comparisonto those measured in the southwest and downtown Mexico City(13.3 and 23.9 ppb respectively) [29,30].

3.2.2. Sources of VOCsAccording to abundance and spatial distribution of VOCs species

in Salamanca, the following categories can be identified as the ma-jor contributors:

(1) Vehicular Emissions (gasoline and diesel-powered vehicles).The emissions of vehicles powered by gasoline and diesel arehighly loaded with ethene, acetylene, propene, MTBE, n/i-pentane, 2,2-dimethylbutane, 2-methylpentane, benzene,toluene, and xylenes [31]. These species were observed inall monitoring sites in concentrations high enough to rankthem among the first 15 (see Table 4). MTBE average levelswere 4.79 ± 3.34 ppbC at the urban site (not shown in Table4), representing about 1.0% of the total VOCs. By comparison,a percentage lower than 2% is reported for samples collectedin Mexico City [25].

(2) LPG handling and leakage. This source is characterized byemissions rich in propane, i-butane, and n-butane [32]. Pro-pane was the major species at all sites (except rural MI).

(3) Refining processes. Emissions from refining processes arecomposed by a large variety of species, which depend onthe process itself. It is reported in the literature that ethane,propane, n/i-pentane, toluene and formaldehyde are emittedby refining oil processes and by fugitive emissions [33,34].

(4) Other sources: The high levels of halogen species such asmethyl bromide, methyl chloride, 1,2-dichloropropane anddichloroethane reveal the presence of sources related withagriculture activities (e.g. application of pesticides and soilfumigants) and with industrial solvent and degrease pro-cesses [35].

As mentioned previously, burning of agriculture debris has beena frequent activity in the region that generates large amounts ofparticles and gases, mainly CO and organics. The presence of aceto-nitrile in ambient air indicates this source [36,37]. We have usedstyrene as surrogate species for acetonitrile, since the former hasalso been considered marker for biomass burning in the rural envi-ronments [25]. Average concentrations of styrene at the urban sitewere 2.0 ± 0.88 ppbC, while concentrations of 2.55 ± 2.27 and1.10 ± 0.80 ppbC were found at the rural sites MI and JR, respec-tively. Together, concentrations of styrene and characteristics ofMI site (surrounded by crop fields), suggest that this site is influ-enced by biomass burning emissions. However, no clear relation-

Table 4Concentration (ppbC) and standard deviation for the most abundant VOCs in Salamanca during February–March 2003.

CR CG US MI JR

06:00–18:00 h 18:00–06:00 h 06:00–18:00 h 18:00–06:00 h 06:00–18:00 h 18:00–06:00 h 00:00–23:00 h 00:00–23:00 h

1 Propane 81.1 ± 34.2 167.0 ± 40.3 31.6 ± 9.1 55.5 ± 20.6 16.0 ± 10.4 18.2 ± 8.6 18.7 ± 7.6 42.3 ± 13.42 i-Pentane 33.9 ± 15.4 43.6 ± 6.7 9.3 ± 3.3 31.8 ± 4.4 16.0 ± 4.6 18.0 ± 8.0 26.9 ± 28.7 14.0 ± 7.73 n-Butane 46.2 ± 19.7 90.8 ± 22.9 17.9 ± 4.4 26.7 ± 4.9 9.2 ± 6.0 9.0 ± 4.2 11.4 ± 4.4 20.5 ± 5.34 Toluene 37.7 ± 16.5 51.2 ± 10.5 11.6 ± 2.8 13.2 ± 1.5 4.9 ± 2.2 4.8 ± 3.0 6.9 ± 2.9 12.5 ± 5.35 m/p-Xylene 28.4 ± 13.0 37.6 ± 15.5 7.9 ± 2.2 8.4 ± 2.6 5.0 ± 1.4 8.3 ± 8.7 22.4 ± 36.1 10.8 ± 7.76 Ethane 23.6 ± 18.3 20.1 ± 7.1 5.6 ± 2.4 5.3 ± 3.1 2.9 ± 1.7 1.6 ± 1.1 2.1 ± 0.8 2.9 ± 1.67 Acethylene 9.1 ± 2.3 14.9 ± 8.4 6.7 ± 1.6 6.5 ± 2.5 4.0 ± 2.9 2.9 ± 2.2 4.7 ± 2.8 10.7 ± 11.18 Benzene 17.2 ± 7.1 19.5 ± 2.5 7.3 ± 1.7 7.2 ± 2.8 4.2 ± 2.5 3.3 ± 1.9 5.0 ± 3.4 6.9 ± 3.29 i-Butane 18.0 ± 8.4 33.1 ± 7.5 7.7 ± 2.3 11.1 ± 2.9 3.4 ± 2.2 3.3 ± 1.9 4.2 ± 2.5 10.8 ± 3.210 Ethylene 19.4 ± 13.6 17.2 ± 6.5 7.3 ± 3.6 7.6 ± 5.3 3.3 ± 2.4 2.4 ± 1.8 2.5 ± 1.1 3.4 ± 2.211 Propene 16.8 ± 10.0 18.3 ± 8.6 3.7 ± 1.0 4.4 ± 0.9 1.3 ± 0.6 1.4 ± 0.7 0.6 ± 0.7 2.3 ± 3.412 n-Pentane 13.6 ± 5.6 15.6 ± 3.8 6.6 ± 1.8 5.8 ± 1.7 1.6 ± 1.0 2.3 ± 1.3 6.1 ± 1.4 4.7 ± 1.913 2,3-Dimethylbutane 10.3 ± 6.0 14.6 ± 2.3 3.3 ± 1.7 4.3 ± 0.9 2.1 ± 0.9 2.9 ± 1.0 2.3 ± 1.1 5.5 ± 3.314 o-Xylene 10.9 ± 6.0 12.8 ± 2.3 3.1 ± 0.9 3.1 ± 0.8 1.4 ± 0.6 2.5 ± 2.4 7.3 ± 10.6 3.9 ± 2.415 3-Methylbutene 41.4 ± 71.8 18.4 ± 4.3 0.6 ± 0.5 0.5 ± 0.3 5.6 ± 3.4 0.4 ± 0.4 5.3 ± 9.8 1.9 ± 0.716 Freon-114 8.6 ± 2.7 14 ± 8 –a –a 7.1 ± 4.1 14.0 ± 4.2 8.9 ± 3.3 18.0 ± 1.717 Methyl bromide –a –a 9.0 ± 3.0 28.4 ± 3.9 –a 48.3 ± 9.7 –a –a

18 1,2-Dichloroethane 2.8 ± 3.0 2 ± 0.4 9.6 ± 3.1 31.8 ± 4.3 0.2 ± 0.1 49.9 ± 10.1 1.6 ± 1.2 2.3 ± 1.719 Methyl chlorine 3.6 ± 3.0 5.1 ± 6.4 1.3 ± 0.3 1.6 ± 0.7 1.0 ± 0.3 1.2 ± 0.5 1.0 ± 0.2 1.1 ± 0.220 1,2-Dichloropropane 1.1 ± 0.6 1.2 ± 0.2 7.0 ± 2.3 21.4 ± 3.7 –a 32.3 ± 5.9 –a –a

21 Vinyl chlorine 6.1 ± 2.7 7.7 ± 1.5 1.9 ± 1.8 1.7 ± 0.6 0.3 ± 0.2 0.4 ± 0.3 0.6 ± 0.2 1.5 ± 0.8

a Not available.

E. Vega et al. / Fuel 90 (2011) 3509–3520 3515

ship was found between styrene and other species indicative ofbiomass burning.

A series of scatter plots were constructed to examine the contri-bution of urban and industrial emissions in the region of study.Toluene and propene are emitted by both sources; while MTBE isa distinctive vehicular emission tracer. Fig. 6A shows the scatterplots for MTBE and toluene and Fig. 6B for MTBE and propene. Bothfigures show a subset of data where the two species correlate lin-early indicating the mobile source, while the second non-corre-lated subset indicates an industrial origin. The sites mostlyinfluenced by mobile sources were US, MI, JR and CR, while CGand in some days CR showed influence by industrial emissions.This result agrees with the assumption that pollutants are mainly

0

10

20

30

0 10 20 30 40 50 60 70 80 90

MTB

E (p

pbC

)

Toluene (ppbC)

Mobile

Industry

(A)R2=0.98

(

0

50

100

150

200

250

300

0 30 60 90 120 150

Prop

ane

(ppb

C)

n-butane (ppbC)

CR

CG

LS

MIJR

(C)R2=0.97

(

Fig. 6. Scatter plots of selected species and its relationship with ma

transported from the city towards northeast, and that CR and CGwere the sites with the higher Refinery and Power Plant emissionsimpact.

Propane and n-butane are known fingerprint species of leakagesand unburned LPG [32]. Fig. 6C shows the high correlation betweenthese species, which is an indication that LPG emissions were hom-ogenously emitted in the whole area. Mobile sources also contrib-ute to the levels of propane as corroborated by its correlation withi-butane, shown in Fig. 6D, particularly with data from US, MI andJR sites.

As envisaged, the halocarbons species showed no relationshipwith compounds which are vehicular or LPG tracers. However,some halocarbons such as vinyl chlorine and 1,2-dicholorpropane

0

10

20

30

0 10 20 30 40

MTB

E (p

pbC

)

Propene (ppbC)

Mobile

Industry

B)R2=0.95

0

50

100

150

200

250

300

0 10 20 30 40 50 60

Prop

ane

(ppb

C)

i-butane (ppbC)

CR

CG

LS

MI

JR

D)R2=0.92

in local sources in Salamanca, Mexico. February–March, 2003.

Table 5Correlation coefficient (Pearson) for selected halocarbon species and VOCs inSalamanca during February–March 2003.

MTBE i-Butane Toluene

Freon-114 0.05 0.00 0.00Methyl bromide 0.01 0.20 0.341,2-Dicloroethane 0.03 0.22 0.24Methyl chlorine 0.09 0.18 0.211,2-Dicloropropane 0.04 0.37 0.71Diclorobenzene 0.00 0.04 0.06Vinyl chlorine 0.24 0.72 0.88

(a) March 5, 2003: 18 hrs

Wind Direction

Altit

ude

(m)

0

1000

2000

3000

4000

5000

0 60 120 180 240 300 360

VSSIJR

(b) March 9, 2003: 18 hrs

Altit

ude

(m)

0

1000

2000

3000

4000

5000

0 60 120 180 240 300 360

VSSIJR

3516 E. Vega et al. / Fuel 90 (2011) 3509–3520

showed moderate association with toluene, which is probably dueto the industrial origin of the later (Table 5). On the other hand,strong relationship was found between methyl bromide, 1,2-dichloropropane and 1,2-dicholoroethane (Fig. 7). Methyl bromideis mainly used as soil fumigant for the control of nematodes, fungiand weeds; it is also used in food-processing facilities (e.g. forextracting oils from nuts, seeds and flowers). 1,2-Dicholorporpaneis still used as pesticide, although this practice is prohibited inNorth America and Europe, it has also industrial use in the applica-tion of paint, lacquers and varnishes and for manufacture of manychlorinated compounds. 1,2-Dicholoroethane has agricultural andindustrial applications, as chemical intermediate in the vinyl chlo-ride monomer manufacture [38].

Finally, the concentrations of isoprene, the unique biogenicVOCs measured, showed concentrations �1.0 ppbC in the urbansites and lower than 1.0 ppbC in the suburban and rural sites. Itshigher level in the urban area may be the result of vehicular emis-sions rather than vegetation, which is scarce in the city.

Wind Direction

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

3.3. Meteorology

Before this work, no previous studies on local wind circulationhad been reported for this region. Meteorological measurementswere carried out to determine the occurrence and properties of lo-cal flow patterns, to examine the structure and evolution of themixing layer over the basin and to provide data for evaluationand testing numerical meteorological models.

Three important parameters were determined at JR, SI and VSsites to understand transport and dilution of pollutants in this re-gion: the mixing layer height (MH), the transport velocity (TW)and the ventilation index (VI) which is calculated as the productof MH and TW. In general, the VI values assured good ventilationconditions in the whole area in the afternoon. The highest mixinglayer height was observed at 1800, reaching mean values approx-imately of 3300 m above ground level.

The frequent meteorological conditions are exemplified withthe data from March 5: synoptic winds aloft had strong influence

R2 = 0.99

R2 = 0.99

0

10

20

30

40

50

60

70

0 10 20 30 40 50 60 70

ppbC

1,2 dochloroethane (ppbC)

1,2dichloropropane

methyl bromide

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

on the wind flow at lowest levels, wind direction above 3 km fromthe ground is similar for the three sites; below that level, winddirection changes slightly along the column, reaching its maximumdifference on the surface level, varying from S at JR to SSW at SI(Fig. 8a). This meteorological condition caused the transport of pol-lutants out of the city. On the contrary, on March 9 (Fig. 8b), winddirection below 1 km blew from NNE at JR and VS sites to ENE at SIsite, above that level wind changed constantly until it reached syn-optic winds (SSW), 5 km above the ground, causing transport of airmasses into the city [39].

The potential temperature evolution along the day is an impor-tant parameter that measures the dilution of pollutants. The fre-quent conditions observed at 0800 at all sites showed that thesurface-based inversion formed overnight was partially elimi-nated. At noon, the mixing layer height was above 1200 m at allsites, and at 1800 the mixing layer was approximately between3000 and 3500 m high. On March 9, the potential temperaturehad a different behavior, at 0800 at all sites presented surface-based thermal inversion, at noon the mixing layer was below500 m, and at 1800 the mixing layer height reached the highestvalues during the field campaign, above 4000 m.

3.4. Modeling of transport and dispersion of pollutants in the region

The Regional Atmospheric Model System (RAMS) and the CIT-SAPRC99 models were used to simulate the physical and chemicalprocesses which control dispersion, transport and formation ofpollutants in the atmosphere. The study region was set to an exten-sion of 140 � 140 km, divided into a grid of 74 � 74 tridimensionalcells with a resolution of 2 � 2 km. the Salamanca city was located

E. Vega et al. / Fuel 90 (2011) 3509–3520 3517

in the center of this region. The meteorological model RAMS wasused to calculate temperature, relative humidity, turbulence, windspeed and wind direction every 300 s for each cell. The topographyand land use were taken into account for an adequate simulation ofsurface wind pattern [19]. The photochemical model CIT-SAPRC99was used to solve the transport and mass conservation of 112chemical species using the SAPRC99 chemical mechanism [40].Modeling results were used as complement tool to improve theunderstanding of tropospheric ozone behavior throughout thestudy region where no measurements were available.

The CIT-SAPRC99 uses operator splitting technique to solvetransport and chemical processes separately in each time step. Ver-tical diffusion is modeled as a function of atmospheric stabilityclass, with the mixing height entered as input to the model. Themodel was modified to account for three-dimensional fields oftemperature and humidity as inputs, and to calculate reaction ratesin three dimensions on the basis of these parameters [20]. The CITmodel has been applied and tested extensively and reported else-where [41–44].

The CIT uses a terrain-following coordinate system and 15 ver-tical layers were defined for this work, with the top boundary ofthe modeling domain at 4628 m above the surface. The height ofeach layer became smaller as it approached to the surface (e.g.

(A) 0800.

(C) 1800.

Fig. 9. Surface wind field determined by RAMS model for the study regi

50 m for the lowest layer) for greater resolution. The wind fieldsobtained from RAMS model were interpolated to fit the CIT-SAP-RC99 grid domain, smoothed and filtered to improve mass consis-tency in the CIT-SAPRC99 model.

The results obtained from the RAMS model indicated two differ-ent dispersion conditions during the study period. The first andmore frequent condition showed circulation of winds during earlymorning and night that converged towards the Salamanca city i.e.mountain-valley circulation (Fig. 9A). From 1000, the circulationshifted towards the mountains (Fig. 9B) with a NW wind directionat 1800 (Fig. 9C). After this time, the wind speed decreased and thewind direction returned towards the city (Fig. 9D). The second lessfrequent condition indicated a constant wind direction from theNE, from 1000 to 2000 (not shown). Such meteorological conditionled to accumulation of emissions inside the valley and thereforethe increase of pollution levels.

The results of coupled application of RAMS and CIT-SAPRC99models to estimate production and transport of ozone are illus-trated in Fig. 10A-D, which describe ozone concentration indicatedwith background colored from blue (0 ppb) to red (200 ppb), windfields (black arrows) and VOCs plumes (green zones) emitted fromLeón, Irapuato, Salamanca and Celaya cities (L, I, S and C yellow let-ters, respectively). Simulations were calculated for 0900, 1200,

(B) 1000.

(D) 2300.

on on March 4, 2003 at (A) 0800; (B) 1000; (C) 1800 and (D) 2300.

(A) 0900. (B) 1200.

(C) 1600. (D) 2100.

L

IS

C

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 duringMarch 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 figure legend, the reader is referred to the web version of this article.)

3518 E. Vega et al. / Fuel 90 (2011) 3509–3520

1600 and 2100. Fig. 10A shows that wind flow at 0900 was domi-nated by drainage flow of cold air descending from the mountains,transporting urban emissions to the lowest areas of the region. Athis time of the day, high atmospheric stability conditions and ther-mal inversion caused that emissions were trapped near the surfaceand minimum ozone production. At noon, VOCs were diluted dueto the rapid growth of the mixing height, so only a small plumeis observed at L. Ozone production is evident in those regionswhere precursors were transported to early in the morning, whileozone concentrations were minimal in the urban areas. Fig. 10Cshows that wind pattern at 1600 has changed, transporting airmasses from southwest to northeast. The ozone formation waswell developed in the whole region and the maximum ozone con-centrations were calculated in the NE of the model’s domain.Again, at night (Fig. 10D), cold air descending from the mountainsbegan 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

insignificant of NOX sources in that region and again VOCs concen-trations became significant at surface level in the urban zones.

The above described behavior was frequently observed in theregion, especially when weak synoptic conditions generated val-ley-mountain circulation. Nevertheless, it is not possible to deter-mine a general wind pattern due to the limited time of thestudy; it would be necessary to carry out the analysis for other sea-son of the year.

Finally, overall observed and modeled results of this surveyindicate environmental implications due to the exposition of pop-ulation to secondary pollutants and halogenated hydrocarbons,especially in the rural areas located northeast of the industrialand urban zones. As a consequence of this study, local and federalenvironmental authorities in collaboration with PEMEX (the Na-tional Oil Company) and the Comisión Federal de Electricidad (Fed-eral Electricity Bureau) developed the first Air Quality Program forthe region (2003–2006). Main achievements of this program werethe reduction in number of days with sulfur dioxide concentrations

E. Vega et al. / Fuel 90 (2011) 3509–3520 3519

above the standard, from 79 in 2003 to 34 in 2006. Later on, theprogram included definitive removal of methyl bromide usage foragricultural activities [24].

4. Conclusions

A 2-week field campaign of measurements of criteria pollutants,VOC and meteorological parameters was carried out during Febru-ary–March 2003 in the Salamanca city which is located inside oneof the most important industrial corridors of Mexico, to characterizeair quality in the urban and regional scales. Special emphasis was gi-ven to the chemical characterization of VOC to gain knowledgeabout levels, distribution and origin of these species. In spite of lim-ited period of time, this monitoring effort currently represents oneof the biggest studies on air quality realized in Mexico, without con-sidering the most recent studies carried out in the metropolitan areaof Mexico City, which allows investigating the air quality conditionin an area highly influenced by industrial emissions.

Regarding one of the most important pollutants, the SO2

showed a high concentration event at 0200 at the urban site, lo-cated 2 km west of the Refinery and the Power Plant. The levelsof this pollutant reached 0.60 ppm (1 min sampling-time). Histor-ically, extremely high levels of SO2 have been frequently observedfrom 0000 to 0300 in the city, attributed to venting operations ofthe Refinery and the Power Plant. However since 2003 controlmeasurements have been implemented to reduce levels of SO2

and particles in the atmosphere, such as the shifting to cleanerfuels (with lower content of S) in both the Power Plant and theRefinery. These measurements were already in execution by thetime of the field campaign, thus it is possible that the first positiveresults were observed.

The highest ozone levels were observed in the rural monitoringsites, which are explained by transport and possible inputs fromthe free troposphere, given the high altitude of the study region.

Total VOCs mass concentration fluctuated from 170 ppbC in therural area to 956 ppbC in the urban sites. The more abundant groupwas the alkanes followed by aromatics, olefins and halogenated.The total mass concentration of VOCs is comparable withvalues reported for some urban and rural sites of Mexico CityMetropolitan Area. However, the concentration of halogenatedspecies (particularly Freon-114, methyl bromide, 1,2-dichloroeth-ane and 1,2-dichloropropane) was notoriously higher in rural sitesof Salamanca compared with Mexico City and other highly pollutedareas of China. Formaldehyde and acetaldehyde showed averageconcentrations of 4.3 and 2.7 ppb, respectively which are lowerthan those reported for Mexico City (13.3 and 4.4 ppb respec-tively). Due to these low values, it is expected little contributionof carbonyl species to the photochemical activity in the region.

The origin of VOCs was inferred by examining the concentrationand spatial distribution of species, in conjunction with cross corre-lation analyses and relationship of marker species for specificsources. In the urban area, gasoline and diesel vehicular exhaustand marketing/leaking of LPG dominated the VOCs burden, whileemissions from industrial and agriculture activities were impor-tant in the rural area. The last two sources emit halogenated spe-cies that represent potential hazard to human health due to thetoxic and/or carcinogenic effect, thus the results of this study high-lighted the needed of monitoring and even implementing controlmeasurements for these species. As a consequence, federal and lo-cal environmental authorities removed usage of methyl bromide inagricultural activities since 2007.

Meteorological conditions generally favoured the transport ofpollutants outside Salamanca towards northeast, where the high-est concentration of ozone were observed. This result was in agree-ment with the modeling calculations.

Acknowledgements

This study was supported by Mexico national oil company(PEMEX) under contract IMP-ZC-001-2003. The authors are grate-ful to the authorities of Salamanca refinery by their help and sup-port provided during the study.

References

[1] Shahriar S, Topal E. When will fossil fuel reserves be diminished? Energy Policy2009;37:181–9.

[2] Interngovermental Panel on Climate Change (IPCC). Climate change 2001: thescientific basis. Contribution of the working group I to the third assessmentreport of the interngovermental panel on climate change. Technicalsummary. Cambridge International Press; 2001.

[3] Decker EH, Elliot S, Smith FA, Blake DR, Rowland FS. Energy and material flowthrough the urban environment. Ann Rev Energy Environ 2000;25:685–740.

[4] Molina LT, Kolb CE, de Foy B, Lamb BK, Brune WH, Jimenez JL. Air quality inNorth America’s most populous city – overview of the MCMA-2003 campaign.Atmos Chem Phys 2007;7:2447–73.

[5] Air Resources Board of California, USA. <http://www.arb.ca.gov/toxics/id/taclist.htm> [cited 22.01.11].

[6] Secretaría de Energía. Prospectiva de petrolíferos 2006–2015. ISBN: 968-874-201-5 DF, Mexico; 2006.

[7] Instituto Nacional de Estadítica Geografía e Informática. XII Censo General dePoblación y Vivienda 2000. DF, México; 2001.

[8] SEMARNAT-IEG. Programa para mejorar la calidad del aire en la ciudad deSalamanca (PROAIRE). Periódico Oficial del Estado de Guanajuato No. 9, 6junio: México; 2003.

[9] Seinfeld JH, Pandis SN. Atmospheric chemistry and physics: from air pollutionto global change. New York (NY): John Wiley and Sons; 1998.

[10] Jacobson MZ. Atmospheric pollution, history, science, and regulation. UnitedKingdom: Cambridge University Press; 2002.

[11] Jenkin ME, Clemitshaw KC. Ozone and other secondary photochemicalpollutants: chemical processes governing their formation in the planetaryboundary layer. Atmos Environ 2000;34:2499–527.

[12] Atkinson R. Atmospheric chemistry of VOCs and NOX. Atmos Environ2000;34:2063–101.

[13] Atkinson R, Arey J. Gas phase tropospheric chemistry of biogenic organiccompounds: a review. Atmos Environ 2003;37:S197–119.

[14] Odum JR, Hoffmann T, Bowman F, Collins D, Flagan RC, Seinfeld JH. Gas/particle partitioning and secondary organic aerosol yields. Environ Sci Technol1996;30:2580–5.

[15] Molina MJ, Rowland FS. Stratospheric sink for chlorofluoromethanes:chlororine atom catalysed destruction of ozone. Nature 1974;249:810–2.

[16] Vega E, Ruiz H, Martinez-Villa G, Sosa G, Gonzalez-Avalos E, Reyes E, et al. Fineand coarse particulate matter characterization in a heavily industrialized cityin central Mexico during Winter 2003. J Air Waste Manage Assoc2007;57:620–33.

[17] Environmental Protection Agency. Compendium method TO-14a.Determination of volatile organic compounds (VOCs) in ambient air usingspecially prepared canisters with subsequent analysis by gas chromatography(EPA/625/R-96/010b). Compendium of methods for the determination of toxicorganic compounds in ambient air. 2nd ed. Cincinnati (OH); 1999.

[18] McCleanny WA, Plei JD, Evans GF. Canister based method for monitoring toxicVOCs in ambient air. J Air Waste Manage Assoc 1991;41:1308–18.

[19] Pielke RA. A comprehensive meteorological model system: RAMS. MeteorolAtmos Phys 1992;49:69–91.

[20] West JJ, Zavala MA, Molina LT, Molina MJ, San Martini F, McRae GJ, et al.Modeling ozone photochemistry and evaluation of hydrocarbon emissions inthe Mexico City metropolitan area. J Geophys Res 2004. doi:10.1029/2004JD004614.

[21] Vingarzan R. A review of surface ozone background levels and trends. AtmosEnviron 2004;38:3431–42.

[22] Kalabokas PD, Repapis CC. A climatological study of rural surface ozone inCentral Greece. Atmos Chem Phys 2004;4:1139–47.

[23] Kalabokas PD, Volz-Thomas A, Brioude J, Thouret V, Cammas JP, Repapis CC.Vertical ozone measurements in the troposphere over the EasternMediterranean and comparison with Central Europe. Atmos Chem Phys2007;7:3783–90.

[24] SEMARNAT/Gobierno del Estado de Guanajuato. Progama para mejorar lacalidad del aire en Salamanca 2007–2012. 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 volatileorganic compounds in the Valley of Mexico during the MCMA 2002 & 2003field campaigns. Atmos Chem Phys 2007;7:329–53.

[26] Chan CY, Tang JH, Li YS, Chan LY. Mixing ratios and sources of halocarbons inurban, semi-urban and rural sites of the Pearl River Delta, South China. AtmosEnviron 2006;40:7331–45.

[27] Barletta B, Meinardi S, Simpson IJ, Rowland FS, Chan C, Wang X, et al. Ambienthalocarbon mixing ratios in 45 Chinese cities. Atmos Environ2006;40:7706–19.

3520 E. Vega et al. / Fuel 90 (2011) 3509–3520

[28] Lei W, de Foy B, Zavala M, Volkamer R, Molina LT. Characterizing ozoneproduction in the Mexico City Metropolitan Area: a case study using achemical transport model. Atmos Chem Phys 2007;7:134766.

[29] Grutter M, Flores E, Andraca-Ayala G, Baez A. Formaldehyde levels indowntown Mexico City during 2003. Atmos Environ 2005;39:1027–34.

[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 apportionmentin Mexico City using the chemical mass balance receptor model. AtmosEnviron 2000;34:4121–9.

[32] Blake DR, Rowland FS. Urban leakage of liquefied petroleum gas and its impacton Mexico City Air Quality. Science 1995;269:953–6;Watson JG, Chow JC, Fujita EM. Review of volatile organic compounds sourceapportionment by chemical mass balance. Atmos Environ 2001;35:1567–84.

[33] Fujita EM. Hydrocarbon source apportionment for the 1996 Paso del NorteOzone Study. Sci Total Environ 2001;276:171–84.

[34] Environmental Protection Agency [Internet]. USA: SPECIATE4.2 data browser.<http://www.epa.gov/ttn/chief/software/speciate/index.html> [updated 2008November; cited 2001 January 21].

[35] Cox ML, Fraser PJ, Sturrock GA, Siems ST, Porter LW. Terrestrial sources and sinksof halomethanes near Cape Grim, Tasmania. Atmos Environ 2004;38:3839–52.

[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;9:4929–44.

[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;2009:9.3425–3442.

[38] International Program on Chemical Safety [Internet] Chemical SafetyInformation from Intergovernmental Organizations. <http://www.inchem.org/>[Updated 21.09.10; cited 21.01.11].

[39] Instituto Mexicano del Petróleo. Estudio local y regional del impacto de lasemisiones de PEMEX y CFE en la calidad del aire en Salamanca, Gto. FinalReport. PS-MAS-IF-F21026-02. DF, México; 2003.

[40] Carter WPL. Documentation of the SAPRC-99 chemical mechanism for VOCreactivity, final report to California Air Resources Board, Contract 92-329 and95-308, 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-generationmathematical model for urban air pollution: 1. Model formulation. AtmosEnviron 1982;16:679–96.

[42] Los Alamos National Laboratory and Instituto Mexicano del Petroleo. Mexicocity air quality research initiative (MARI), rep LA-12699, Los Alamos, N.M.;1994.

[43] Harley RA, Sawyer RF, Milford JB. Updated photochemical modeling forCalifornia’s south coast air basin: comparison of chemical mechanisms andmotor vehicle emission inventories. Environ Sci Technol 1997;31:2829–39.

[44] Marr LC, Noblet GS, Harley RA. Formation of photochemical air pollution incentral California: 2. Impact of revised emissions on Eulerian modelpredictions. Geophys Res 2002;107(D6):4048. doi:10.1029/2001JD00069.