ROADSIDE AIR POLLUTION MEASUREMENTS AND TRAFFIC VOLUME IN A US-MEXICO BORDER CITY: TIJUANA, B.C.

_______________

A Thesis Presented to the Faculty of San Diego State University _______________

In Partial Fulfillment of the Requirements for the Degrees Master of Public Health and Master of Arts in Latin American Studies _______________

by Edgar A. Rodriguez Spring 2011

iii

Copyright © 2011 by Edgar A. Rodriguez All Rights Reserved

iv

DEDICATION Este trabajo te lo dedico a ti, mi viejita, tú que tanto luchaste para que yo llegara a este momento. Tu sacrificio, consejos e insistencia no fueron en vano y los seguiré usando en el largo camino que aún me falta por recorrer. Espero que desde los cielos disfrutes este momento. Te extraño. . .

v

Racial injustice, war, urban blight, and environmental rape have a common denominator in our exploitative economic system. -Channing E. Phillips

vi

ABSTRACT OF THE THESIS Roadside Air Pollution Measurements and Traffic Volume in a US-Mexico Border City: Tijuana, B.C. by Edgar A. Rodriguez Master of Public Health and Master of Arts in Latin American Studies San Diego State University, 2011 Recent studies suggest that spending a significant amount of time in proximity to vehicular traffic has adverse effects on human health, especially to vulnerable populations such as the elderly and children. Concentrations of many traffic-related pollutants decrease rapidly with increasing distance from the road, underscoring the fact that fixed-site monitors for air pollutants in communities may not capture traffic-related exposures important to human health. Few traffic-related pollutant measurement studies have been conducted in developing nations where traffic-related pollution may be higher due to older vehicular fleets and other conditions. In the present study, roadside pollutants were measured at 55 sites in the city of Tijuana, Baja California, Mexico, including 18 in proximity to elementary schools. The relationship between traffic counts and air pollutants was analyzed by performing 20 min simultaneous roadside measurements of black carbon (BC), particle-bound polycyclic aromatic hydrocarbons (PAHs), fine particulate matter (PM2.5), ultrafine particle number (UFP), and carbon monoxide (CO). In addition, some 24 hr passive measurements of nitrogen oxides (NOx) were conducted. Traffic volume and flow was assessed by video recording the measurement session at each site and performing subsequent traffic counts. Most pollutant concentrations were highly variable across different areas of Tijuana. The median BC concentration was 4437 ng/m3 with the highest levels measured in the SE area of the city. The median concentrations for UFP, CO and NO2 were 30,265 particles/cc, 1.6 ppm, and 49 ppb, respectively, with the highest levels measured at the “5 y 10” intersection. The median counts for autos and trucks were 1758 ct/hr and 180 ct/hr. Ultrafine particles number concentrations were most highly correlated with traffic counts, as was BC to a lower extent. PM2.5 was least correlated with traffic counts. Carbon monoxide concentrations were highly correlated with both ultrafine particle number and BC concentration making CO a possible simple surrogate for near-traffic exposure measurements. At school-related sites, the median UFP number and BC concentrations were 3673 ng/m3 and 28,566 particles/cc. In addition, UFP number concentrations were significantly higher under moderate or intermittent traffic flow conditions. Finally, there was some evidence that low socioeconomic status (SES) areas had higher CO levels. However all other pollutants were not significantly different by SES status indicating that traffic-related pollution represents a hazard to many city residents. These findings suggest that traffic related-air pollution may affect public health in Tijuana, and that air quality interventions to reduce traffic-related emissions and policy measures to route traffic away from schools are warranted.

vii

TABLE OF CONTENTS PAGE ABSTRACT ............................................................................................................................. vi LIST OF TABLES .....................................................................................................................x LIST OF FIGURES ................................................................................................................. xi ACKNOWLEDGEMENTS .................................................................................................... xii CHAPTER 1

INTRODUCTION .........................................................................................................1  Background ..............................................................................................................1  Objectives ................................................................................................................4 

2

LITERATURE REVIEW ..............................................................................................5  Health Effects of Air Pollution ................................................................................5  Ambient Air Pollution....................................................................................... 5  Traffic Related Pollution................................................................................... 7  Metrics of Traffic Exposure .............................................................................. 8  Respiratory Effects...................................................................................... 8  Cardiovascular Effects ................................................................................ 9  Cancer and Traffic Exposures ................................................................... 10  Birth Outcomes and Traffic Exposures..................................................... 10  Roadway Pollution .................................................................................................11  Black Carbon .................................................................................................. 13  Polycyclic Aromatic Hydrocarbons ................................................................ 15  Particulate Matter ............................................................................................ 17  PM2.5 ......................................................................................................... 17  Ultra Fine Particles ................................................................................... 19  Carbon Monoxide ........................................................................................... 21  Nitrogen Oxides .............................................................................................. 22  Assessing Community Exposure to Roadside Air Pollution .......................... 24 

viii Factors Affecting Roadside Pollution Concentrations .................................... 24  Tijuana, B.C. ..........................................................................................................29  3

METHODOLOGY ......................................................................................................32  Design of the Investigation ....................................................................................32  Spatial Assessment.......................................................................................... 32  Temporal Assessment ..................................................................................... 33  School-Related Assessment ............................................................................ 34  Categorical Variables ...................................................................................... 35  Data Measurements ................................................................................................36  Pollutants and Meteorological Variables ........................................................ 36  Traffic Counts and Composition..................................................................... 36  Treatment ...............................................................................................................37  Instrumentation ............................................................................................... 38  Data Analysis .................................................................................................. 40  Quality Control ............................................................................................... 40 

4

RESULTS ....................................................................................................................43  Analyses Across All Sites ......................................................................................43  Descriptive Data.............................................................................................. 43  Pollutant Levels in Relation to Traffic Count ................................................. 44  Pollutant Associations ..................................................................................... 44  Relationship of Pollutants with Categories of Traffic Density and Flow ..............47  Temporal and Spatial Analyses .............................................................................47  Temporal and Spatial Analyses across All Sites............................................. 49  Temporal Variability at Multiple-Measured Sites .......................................... 50  Analyses at School-Related Sites...........................................................................52  Descriptive Data.............................................................................................. 52  Pollutant Levels in Relation to Traffic Count ................................................. 53  Association of Pollutants at School-Related Sites .......................................... 54  Pollutant and Traffic Comparisons with Socio-Economic Status Categories ..............................................................................................................54  Comparisons with Other Variables ........................................................................57 

ix 5

DISCUSSION ..............................................................................................................59  Pollution Levels near Roadways ............................................................................60  Pollution and Traffic ..............................................................................................61  Relating Traffic Pollution to Traffic Flow .............................................................63  Correlations between Pollutants ............................................................................63  Temporal and Spatial Variability ...........................................................................64  Relating Pollution to SES and Other Categories ...................................................66  Limitations .............................................................................................................67  Recommendations ..................................................................................................68 

6

CONCLUSIONS..........................................................................................................70 

REFERENCES ........................................................................................................................72  APPENDIX   A STUDY SITE MAP .....................................................................................................84  B STUDY SITE LIST AND SITE DESCRIPTION .......................................................86  C SCHOOLS LOCATIONS AND THEIR DESCRIPTIONS IN RELATION TO MEASUREMENT AND TRAFFIC REFERENCE SITES ..................................90  D DESCRIPTIVE RESULTS AT MULTIPLE-MEASUREMENT SITES ...................93  E DESCRIPTIVE RESULTS AT SINGLE-MEASUREMENT SITES.........................97  F NITROGEN OXIDES DESCRIPTIVE RESULTS AT SINGLE AND MULTIPLE-MEASUREMENT SITES ......................................................................99  G POLLUTANT TEMPORAL VARIABILITY ACROSS MONTHS THROUGHOUT STUDY .........................................................................................102  H DESCRIPTIVE RESULTS AT SCHOOL-RELATED SITES .................................105  I

POLLUTANT CONCENTRATIONS AND TRAFFIC MAPS ................................108 

x

LIST OF TABLES PAGE Table 1. Pollutants versus Traffic Counts Spearman Correlations (All Data) .........................45  Table 2. Spearman Correlations between Pollutants (All Data) ..............................................46  Table 3. Pollutant and Traffic Variability across Multiple-Measured Sites ............................51  Table 4. Pollutant versus Traffic Counts Spearman Correlations (School-Related Sites) ............................................................................................................................53  Table 5. Spearman Correlations between Pollutants (School-Related Sites) ..........................55  Table 6. Pollutant and Traffic Relationship with SES Categories (Overall and School-Related Sites) ...................................................................................................56  Table 7. Study Site List and Site Description ..........................................................................87  Table 8. Schools Locations and Their Descriptions in Relation to Measurement and Traffic Reference Sites ................................................................................................91  Table 9. Descriptive Results at Multiple-Measurement Sites ..................................................94  Table 10. Descriptive Results at Single-Measurement Sites ...................................................98  Table 11. Nitrogen Oxides Descriptive Results at Single-Measurement Sites ......................100  Table 12. Nitrogen Oxides Descriptive Results at Multiple-Measurement Sites ..................101  Table 13. Descriptive Results at School-Related Sites ..........................................................106 

xi

LIST OF FIGURES PAGE Figure 1. Measurement zones or sectors. .................................................................................33  Figure 2. Placement of sites for temporal analysis. .................................................................34  Figure 3. Placement of school-related measurement sites. ......................................................35  Figure 4. Association between UFPs and auto traffic counts. .................................................45  Figure 5. BC and UFP count median concentrations by truck and all vehicle traffic categories. ....................................................................................................................48  Figure 6. UFP count median concentrations by restricted traffic flow categories...................48  Figure 7. Variability of ultrafine particles across zones. .........................................................50  Figure 8. Study site map. .........................................................................................................85  Figure 9. Temporal variability of BC across months. ............................................................103  Figure 10. Temporal variability of PM2.5 across months. ......................................................103  Figure 11. Temporal variability of UFP count across months. ..............................................104  Figure 12. Study sites. ............................................................................................................109  Figure 13. Carbon monoxide (CO) median concentrations (ppm). .......................................110  Figure 14. Ultrafine particle (UFP) count median concentrations (pt/cc). ............................111  Figure 15. All vehicle traffic median counts (ct/hr). .............................................................112  Figure 16. Black carbon (BC) median concentrations (ng/m3)..............................................113  Figure 17. Truck traffic median counts (ct/hr).......................................................................114 

xii

ACKNOWLEDGEMENTS I would like to thank my advisor, Dr. Jenny Quintana, for her guidance and patience with me throughout this process. She did an incredible job keeping me focused when I wanted to go in so many directions. She is not only a great mentor, but she is also a very caring person and a professor that appreciates the work of her students. I am also very appreciative of all my professors for opening my eyes to so many new concepts and ideas that I will most definitely utilize and apply in my future endeavors. I also thank my colleagues in Tijuana for all their hard work and attention to this project that is important for all of us who live in the US-Mexico border region. I also would like to thank my family a long time friends for their support during the good and bad times, this meant so much to me and helped me stay the course. My friends in the program were also important to me, we kept each other honest and motivated throughout this long journey. Finally and most importantly, I want to thank a very special person who helped me believe in myself and ultimately supported my decision to obtain a Masters degree. Thank you Ceci for understanding me, your support and love meant so much to me. I will never forget you.

1

CHAPTER 1 INTRODUCTION BACKGROUND The ‘London Fog’ episode of 1952, where thousands of people died within a few weeks, along with similar events elsewhere in Europe and the United States marked the beginning of an era of investigation to mitigate the effects of harmful effects of air pollution. In the years following these events, researchers identified a correlation between high ambient concentrations of particulate matter and sulfur oxides, and acute increases in mortality (Pope et al., 2002). Later research indicated that regional air pollution did not only affect health in the short term, but also had long-term health effects on humans (World Health Organization [WHO], 2002) through prolonged exposures. Subsequently, with the improvement in pollution detection technologies and epidemiological methods, researchers were able to identify other ambient pollutants and toxic chemicals, as well their sources in the environment. The 1970 Clean Air Act in the United States and similar legislation in other developed countries, led to measures to reduce the concentrations of many air pollutants, especially those from stationary sources. In addition, the establishment of the criteria pollutants and the promulgation of standards for them helped mitigate the effects of outdoor air pollution. Despite all the air quality improvements in the United States and other developed nations over the last forty years, the occurrence of asthma and allergic diseases has continued to increase in the United States and the rest of the world (Peden, 2005). The list of health effects associated with air pollution has continued to grow and now includes cardiovascular disease and adverse birth outcomes (Ritz & Wilhelm, 2006; Samet 2006). In 2000, global burden of disease (GBD) estimates attributed 1.4% of total deaths and 0.8% of total disability adjusted life years (DALYs) to urban outdoor air pollution (WHO, 2009). In one of those studies, outdoor air pollution ranked 13th out of the 20 most important risk factors for global attributable mortality (Ezzati et al., 2002).

2 A considerable portion of this problem is produced by the transportation sector. The transportation sector is a major energy user across the world and is overwhelmingly petroleum based (Ortmeyer & Pillay, 2001). With globalization and political and economic pressures promoting a continued reliance on fossil fuels, outdoor air pollution remains a big problem, especially in urban areas where motor vehicles are the main transportation modes. In these settings, the number of vehicles, as well as the distances traveled by each vehicle has increased (Colvile, Hutchinson, Mindell, & Warren, 2001), leading to significant increases in emissions and thus a reduction in air quality (Lyons, Kenworthy, Moy, & dos Santos, 2003). In developed countries where vehicle emission controls, ongoing research, and regulations that are normally well enforced, poor urban air quality remains. High levels of pollution caused by the combination of a large vehicular fleet, topography of the land, and permissive meteorological conditions, are seen in cities like Los Angeles. At the same time, similar pollution levels are being measured in many cities of the developing world where the rapid urbanization experienced by these cities during the latter part of the 20th century (Gilbert, 1994) resulted in a substantial increase of their vehicle fleets (Faiz, Gautam, & Burki, 1995), which today remain largely outdated and poorly regulated. The Mexico City metropolitan area, a megacity of over 20 million inhabitants, is well known for its poor urban air quality. More recently, Mexican cities along the US-Mexico border region have sparked the interest of academia, policy makers, and many other stakeholders. In this region, a variety of environmental issues exist, principally from the rapid population growth sparked by many social, political, and economic drivers occurring in the last 150 years. In addition, because these issues affect people on both sides of the border, bi-national treaties and programs such as the La Paz agreement, the Integrated Border Environmental Plan, Border XXI, and the current Border 2012 program were created and implemented to improve environmental conditions in the border area. The city of Tijuana, Baja California, is a major player in these programs which is related to the city’s rapid demographic growth and various environmental issues. At the beginning of the 20th century Tijuana was merely a village of approximately 200 people (Ayuntamiento de Tijuana, 2011) but due to many social, political, and economic factors, the city is now inhabited by 1.6 million people (Instituto Nacional de Estadistica y Geografia [INEGI], 2011). This massive urbanization has also translated in a significant increase in the

3 city’s motor vehicle fleet. In fact, in less than 15 years, the total vehicle fleet has grown from approximately 363,000 to 737,000 units (Environmental Protection Agency [EPA], 2000; LT Consulting, personal communication, February 2, 2011). Traffic-related air pollution is a pressing issue in Tijuana, an issue that needs to be investigated in more detail. As was discussed earlier, much of the research of air pollution resulted in identifying adverse health effects associated with the short and long-term exposure to air pollutants, especially particulate matter and sulfur oxides. In light of these studies and the identification of new pollutants, criteria pollutants were established after 1970, and their levels assessed through fixed-site monitors scattered throughout the air basin. Much of the earlier epidemiological research was based on the results from these monitors (J. J. Kim et al., 2004). However, recent research has determined that traffic-related pollution levels next to busy roadways are much higher than at those fixed locations or at background urban locations (Levy, Bennett, Melly, & Spengler, 2003; Mukerjee, Smith, Johnson, Neas, & Stallings, 2009). Depending on variables such as wind direction, roadside pollution decreases with increasing distance from the roads (Y. F. Zhu, Hinds, Kim, & Sioutas, 2002b). This underscores the concern that fixed site monitors do not accurately reflect the air quality in the so-called near traffic microenvironments. Moreover, since the traffic composition and driving modes vary from location to location, pollution levels throughout an urban setting show spatial gradients. Furthermore, roadside pollution tends to vary on a temporal scale due to diurnal traffic patterns. In many cities of the world, traffic volume sharply increases during morning and evening rush hours with elevated pollution levels during these times. Therefore, motor vehicles influence the temporal and spatial patterns of pollutant concentrations (Baldauf et al., 2008). Roadside pollution has public health implications because many people live, work, or go to school in close proximity to busy roadways (Baldauf et al., 2008). The relationship between roadside or near traffic pollution and adverse health effects has been shown through epidemiological and toxicological studies. The focus of these investigations has been the elderly or children subpopulations because of their susceptibility to roadside pollution. Many different techniques have been used ranging from a combination of traffic metrics such as

4 road density, traffic volume, and indirect or direct pollution concentration measurements. Estimation of traffic metrics or pollution concentrations through land use regression (LUR) models has been used to determine exposure to traffic-related pollution (Dales et al., 2008; Gehring et al., 2009; Karr et al., 2009).

OBJECTIVES The purpose of this study was to assess the relationship between directly-measured roadside pollution concentrations and traffic counts in the city of Tijuana, Baja California, Mexico. In addition, through the same measurements from a subset of locations, a sub-objective was to assess the potential exposure that elementary school children have to roadside air pollution in Tijuana. Studies conducted in California, the United States and other parts of the world, have determined the general relationships between traffic density and roadside measures of air pollution. However, these relationships may not be valid for a border community such as the city of Tijuana, due to its poorly maintained vehicle fleet with inadequate emission controls. Despite the fact that there have been emissions inventory studies conducted in the past, including one last year, to our knowledge, this is the first simultaneous roadside pollution and traffic volume study conducted in Tijuana.

5

CHAPTER 2 LITERATURE REVIEW HEALTH EFFECTS OF AIR POLLUTION This section discusses background information on ambient air pollution followed by specific information and methodology used to study traffic-related pollution, and ends with the results of specific epidemiological studies that associate adverse health effects to trafficrelated pollution.

Ambient Air Pollution In the decades after the London fog episodes of the 1950s, numerous epidemiological studies were conducted in Europe and the United States to study the short-term effects of high particulate matter levels. Although using different methodologies, they all had a common result: that each 10 µg/m3 increase in PM10 was associated with an increased risk of an adverse health outcome (Schwartz, 1994). These outcomes included hospital admissions for vulnerable adults with cardiovascular diseases, asthma and chronic obstructive pulmonary disease (COPD), as well as daily all-cause mortalities (Valavanidis, Fiotakis, & Vlachogianni, 2008). Subsequent research placed an emphasis on long-term exposures to lower pollutant levels that had implications not only on frail individuals but also on healthy people, including children. One study (Pope et al., 2002) was one of the most important because of its comprehensiveness. From a sample of 1.2 million adults across the U.S., it associated long-term exposure to PM2.5 with a risk of all-cause, cardiopulmonary, and lung cancer mortality. More specifically, these risks (4, 6 and 8%, respectively) were associated with a 10 µg/m3 increase in PM2.5. Other North American and European studies associated same increase in PM2.5 with elevated risks for outcomes such as exacerbation of asthma, respiratory symptoms, reduced lung function, chronic bronchitis, and lung cancer (Valavanidis et al., 2008).

6 These results were somewhat surprising considering that most of the studies were conducted in countries that had established intervention campaigns and regulatory measures to reduce the health risk from air pollutants, including that from motor vehicles. However, since the vehicular fleets and number of miles driven were increasing, the persistence of adverse health outcomes made sense. Other studies during this time period, strongly associated motor vehicles with particulate matter, especially in the sub-micrometer range (Morawska, Thomas, Bofinger, Wainwright, & Neale, 1998). Thus, the obvious next step was to investigate the size and composition of particulate matter. Size is one of the most important characteristics of particulate matter since the fine and ultrafine fractions penetrate deeper into the lung, specifically in the acinar region where gas exchange takes place (Salma, Balashazy, Winkler-Heil, Hofmann, & Zaray, 2002). In addition, toxicological studies have investigated the association between the chemical composition of fine and ultrafine particulate matter and health effects. Toxic organic compounds and metals readily adsorb onto the porous surfaces of particulate matter which is composed mainly of carbonaceous chains (Valavanidis et al., 2008). Transition metals associated with PM have been shown to cause oxidative damage to cellular macromolecules via generation of hydroxyl radicals (Valavanidis, Salika, & Theodoropoulou, 2000). Also, organic compounds such as polycyclic aromatic hydrocarbons (PAHs) and nitro-PAH found in PM, which are associated with diesel engines, are highly mutagenic, with tumor promoter activity, and are attributed to the development of malignant neoplasms in animal lungs (Ohnishi & Kawanishi, 2002). Furthermore, the carbonaceous materials in PM contain free radicals that can initiate redox reactions with oxidative potential and possible cellular damage (Church & Pryor, 1985). The studies cited in the preceding paragraph, describe biological mechanisms of toxicity. For example, the association of PM exposure and the occurrence of respiratory infections, lung cancer, and chronic cardiopulmonary diseases appears to be caused by oxidative stress through the generation of reactive oxygen species (ROS). High levels of ROS from sources like PM can change the redox status of a cell, thus triggering a cascade of events associated with inflammation (Valavanidis et al., 2008). Conversely, damage at the chromosomal level appears to be caused by genotoxic mechanisms that involve oxidative damage to DNA (Healey et al., 2005).

7 Gaseous pollutants are also associated with negative health outcomes. However, because of the interventions and regulatory changes that have taken place in many countries, adverse health impacts have been significantly reduced in the last decades. Nonetheless, in many large urban cities like Los Angeles and Mexico City, ozone is a big problem, largely because its precursors continue to be emitted from motor vehicles. Ozone is one of the most ubiquitous and toxic pollutants found in the atmosphere. In addition to many respiratory system problems, under longer exposures ozone can impair the body’s ability to fight infection and it may injure tissue membranes through oxidative mechanisms (Godish, 2004). One of precursors of ozone, nitrogen dioxide (NO2), has been associated with increased incidence of acute respiratory diseases and also with decreased pulmonary function (Valavanidis et al., 2008). This occurs because of the low solubility of NO2 that readily enters the pulmonary region where it causes damage. Carbon monoxide (CO), a gas that has a greater affinity to hemoglobin than oxygen molecules, reduces the oxygen carrying capacity of red blood cells. As CO levels increase in human red blood cells, so will the severity of symptoms. From low to extreme exposures, the symptoms will be headache, drowsiness, coma, respiratory failure and death (Godish, 2004).

Traffic Related Pollution Most epidemiologic or exposure assessment studies have been conducted in Europe and the United States where through various measures, the risk of exposure to harmful levels of air pollution has been reduced. Until recently, research in this area has begun in a number of developing countries, including some in Latin America where human exposure to vehicular emissions is not only more frequent, but also potentially more dangerous than in developed countries due to the types and quality of fuels consumed in these countries. Additionally, because of their vulnerability, certain populations have been the focus of these studies. Children, for example, are a vulnerable population because their lungs are not fully developed, their breathing rates are much faster than those of adults, and also because they spend more time outdoors (Gilliland, 2009). The adult population, especially the elderly, can also be susceptible to the effects of traffic-related pollution.

8

Metrics of Traffic Exposure In many epidemiologic or exposure assessment studies, the relationship between traffic-related pollution and adverse health outcomes has been conducted utilizing different methods of measuring exposure. These methods include surrogate, modeling, and direct measure techniques (Pan-American Health Organization, [PAHO], 2004).Surrogate techniques are used as a proxy for exposure in studies with relatively large sample sizes. Examples of surrogate techniques are traffic density on the residential street, distance to roadway, total traffic within a certain radius, and distance-weighted traffic density. On the one hand, modeling techniques are more sophisticated ways of measuring exposure on broad and smaller spatial and temporal scales. There are two types of modeling techniques. One is the regression approach such as land use regression (LUR). These models characterize small scale variations in urban settings. They fit concentrations measured at multiple sites using statistical equations, land characteristics, traffic, and other data, to predict pollutant concentrations at specific receptor sites (Y. L. Huang & Batterman, 2000). On the other hand, in dispersion models, emission parameters are used to predict concentrations of pollutants on spatial and temporal scales. The last technique to measure exposure is by direct measurements using field instrumentation, personal samplers, and air quality monitoring stations. Direct measurements bypass the many complexities involved in estimating motor vehicle emissions and the subsequent transport and dispersions of pollutants (PAHO, 2004).This technique is practical and a relatively cheap way to measure pollution for a variety of study applications, such as exposure assessment studies. The following is a discussion of various studies that have used these exposure measures to associate traffic-related pollutants to adverse health effects.

RESPIRATORY EFFECTS Various associations between roadside pollution with general respiratory symptoms have been established. For example, an increased risk of persistent wheeze and chronic phlegm as reported by male US veterans was associated with living within 50 meters of a major roadway (Garshick, Laden, Hart, & Caron, 2003). Elsewhere in the world, in a study performed in Quito, Ecuador, children ages six through 11 exposed to increasing traffic volumes were found to have significantly increasing levels of hemoglobin-bound carbon

9 monoxide (Estrella et al., 2005). In a study performed in Mexico City where both traffic volume and pollutant concentrations were used as measures of exposure, asthmatic children were found to have a greater incidence of cough, wheezing and bronchodilator use (Escamilla-Nunez et al., 2008). Other studies have found specific associations between asthma and roadside pollution. For instance, in a cross-sectional study performed in San Francisco, California, traffic-related pollution measured at elementary schools was found to be significantly associated with an increased risk of bronchitis and physician-diagnosed asthma in school children (J. J. Kim et al., 2004). In a study conducted in California, the investigators used models to estimate pollutant concentrations and found that these were associated with the development of asthma in kindergarten and first grade children (McConnell et al., 2010). Adverse respiratory effects relating to pulmonary function and lung development have also been linked to roadside pollution. A Canadian study used land regression models to estimate pollutant concentrations and road density in proximity to children’s homes and concluded that an increase in roadway density was associated with decreased lung function as measured by exhaled nitric oxide (Dales et al., 2008). This association was also found in a study performed in the border city of Ciudad Juarez, Mexico, where the traffic exposures of 6-12 year-old children were assessed (Holguin et al., 2007). A Dutch study used three measures of exposure to assess their relationship with decreased lung function in elementary school children. The results indicated that lung function was mostly associated with truck traffic volume, especially for children living less than 300 meters from roadways (Brunekreef et al., 1997). A cohort study conducted in Southern California measured the eight year effect of exposure to traffic on lung development in children. The results indicated that children living within 500 meters of a freeway had substantial deficits in their lung development, as measured by forced expiratory volume in one second (FEV-1) and maximum mid-expiratory rate (MMEF), compared to children who lived at least 1500 meters from the freeway (Gauderman et al., 2007).

CARDIOVASCULAR EFFECTS In many studies, adult cardiovascular complications have been linked to traffic-related pollution assessed thorough different exposure measures. In a study conducted

10 in Los Angeles, near home exposure to vehicle emissions was strongly associated with ambulatory blood pressure increases in a community of retirees (Delfino et al., 2010). Also, in a case-control study conducted near Boston that assessed long-term exposure to traffic related pollutants, the investigators found that a cumulative increase in levels of traffic near the home and living near a major roadway were associated with 4% and 5% increases in the odds of acute myocardial infarction (AMI) (Tonne et al., 2007). Finally, a Dutch study used regression models to calculate distance to major roads and regional background pollution to estimate their association with mortality cases of elderly subjects. In this study, cardiopulmonary mortality was associated with living near a major road and to a lesser extent with estimated background air pollution (Hoek, Brunekreef, Goldbohm, Fischer, & van den Brandt, 2002).

CANCER AND TRAFFIC EXPOSURES Some studies have associated traffic-related pollution with cancer. For example, a European study estimated that 5-7% of all lung cancers in European never-smokers and ex-smokers are attributable to high levels of traffic-related air pollution assessed by NO2 or proximity to heavy traffic roads (Vineis et al., 2007). Similarly, a Canadian study found evidence of an association between traffic-related NO2 concentrations and the incidence of post-menopausal breast cancer. More specifically, this study found an increased risk of approximately 25% for every increase of 5 ppb in exposure (Crouse, Goldberg, Ross, Chen, & Labreche, 2010). Many known carcinogens are emitted by vehicles, such as PAHs (reviewed later in this section).

BIRTH OUTCOMES AND TRAFFIC EXPOSURES Recent studies have also found associations between traffic pollution and adverse birth outcomes. For example, a study performed in Southern California found that exposure to increased levels of particulate matter, and possibly carbon monoxide, may contribute to the occurrence of preterm births (Ritz & Yu, 1999). Similarly, a study conducted in Spain concluded that pregnant women exposed to NO2 have an increased risk of having preterm births. This risk was statistically significant at NO2 levels above 46.2 µg/m3 throughout their entire pregnancy (Llop et al., 2010). Furthermore, a Japanese study evaluated the

11 relationship between proximity to roads and the occurrence of preterm birth classified by gestational age. The results indicated that living within 200 meters of a major road increased the risk of preterm births, especially before 28 weeks of gestation (Yorifuji et al., 2011). Adverse effects on newborn birthweight and other physical measurements have also been investigated in some studies. In a study conducted in Southern California, the investigators concluded that mothers who lived in the area with the highest levels of PM2.5 delivered smaller infants on average than their counterparts who lived in areas with lower levels of this pollutant (Parker, Woodruff, Basu, & Schoendorf, 2005). Similarly, a Spanish study that estimated NO2 pollution concluded that continuous exposures above 40 µg/m3 during the first trimester of pregnancy was associated with a decreased length, weight and head circumference of the newborn infant (Ballester et al., 2010). A study conducted in Vancouver, B.C., which used proximity to roads and pollutant concentration as measures of exposure, also found linkages to adverse birth outcomes. According to this study, living within 50 meters of a highway was associated with a 26% increase risk in small for gestational age (SGA) and an 11% risk increase in low birth weight (LBW) of newborn infants (Brauer et al., 2008). Spontaneous abortion (SAB) and preeclampsia have also been linked with traffic-related pollution. In a study conducted in California in which distance to road and annual average daily traffic (AADT) were used as measures of exposure, the investigators concluded that living within 50 meters of a road with an AADT of 15,200 or more, was significantly associated with SAB in African Americans and non-smokers (R. S. Green et al., 2009).

ROADWAY POLLUTION Roadway pollution is composed of a complex mixture of compounds and particles derived from different sources. More specifically, roadway pollution comes from motor vehicle emissions, particles from tire, brake, and engine wear, and also from deterioration from the road’s surface. In addition, fugitive particles from other distant sources can settle on roadways and later re-suspend in air by the action of passing vehicles or wind (Abu-Allaban, Gillies, Gertler, Clayton, & Proffitt, 2003).

12 Motor vehicles in the United States are the major source of volatile organic compounds, carbon monoxide, and nitrogen oxides, with similar values to those measured in Europe and Japan (Godish, 2004). Similar trends are likely occurring in urban setting throughout developing nations. Most fine particulate matter in an urban setting is produced by the combustion processes in cars and trucks (Abu-Allaban et al., 2003). Another byproduct of combustion in motor vehicle is carbon dioxide, a greenhouse house that has important implications at the global level. Motor vehicle emissions are complex. Although the majority of incompletely-burned pollutants come from engine exhaust systems, hydrocarbons are also emitted during evaporative emissions during non-operation of the vehicles (Godish, 2004). In addition, the rate of motor vehicle emissions depends on various factors such as vehicle characteristics and driving conditions. For example emissions from a vehicle are influenced by its age, size and weight, type of engine, maintenance, and whether or not it has an emissions control system (Pandian, Gokhale, & Ghoshal, 2009). Driving conditions will also have an influence on emissions rates. For example, emissions rates will vary when driving at different speeds such as slow in heavy traffic, moderately by deceleration and acceleration through city intersections, or uninterrupted in arterial roadways and highways. A Swedish study demonstrated that emissions were ten-fold higher while driving slowly through congested traffic conditions than when driving at a higher speed through a less congested area (Sjodin & Lenner, 1995). This may not be true for all components of emissions; Jayaratne, Wang, Heuff, Morawska, and Ferreira (2009) demonstrated that particle count emissions were higher with interrupted traffic flow at a pedestrian crossing, than with uninterrupted traffic flow. Additionally, the stress on an engine, or engine load, can influence the emissions rate from a vehicle. For example, Kean, Harley, and Kendall (2003) found that driving uphill produced higher levels of non-methane hydrocarbons (NMHCs), nitrogen oxides (NOx) and carbon monoxide (CO) than driving downhill. This presumably occurred because driving under high engine load conditions, as in the case of driving uphill, higher emissions from incomplete combustion occur due to a higher engine fuel-to-air mixture needed under higher engine load conditions. Finally, the type and quality of fuel will have an effect of emissions. Kirchstetter, Singer, Harley, Kendall, and Traverse (1999) demonstrated that a combination

13 of reformulated fuel and progressively newer vehicle fleet were associated with lower emission levels in California. In this study conducted in Tijuana, CO and NOx, two of the main traffic-related pollutants, were assessed. Other related pollutants such as volatile organic compounds (VOCs) and non-methane hydrocarbons (NMHCs) were not assessed due to the unavailability of the instrumentation for their analysis. In addition, we assessed particulate matter in the accumulation range (PM2.5) as well as ultrafine particulate matter (UFP). Black carbon (BC) and UV-channel particulate matter, an indicator of particle-bound polycyclic aromatic hydrocarbons (PAHs), were also assessed in this study. The following is a more detailed discussion of each of these pollutants.

Black Carbon Black carbon (BC) is a type of particulate matter. Its principal source is the incomplete combustion of biomass or fossil fuels but it can also come from naturally occurring soot, such as that from wildfires. As an absorber of solar radiation, it is a million times stronger than CO2 but with a much shorter lifetime (K. H. Kim, Sekiguchi, Kudo, & Sakamoto, 2011), thus it has a positive radiative forcing or warming effect on earth’s surface (Intergovernmental Panel on Climate Change, 2007). BC constitutes a high fraction of the particulate matter emitted from trucks (Lloyd & Cackette, 2001), and thus it is associated with diesel-powered engines. During the 1980s, heavy-duty diesel engines accounted for approximately 70% of total BC emissions from on-road vehicles in the Los Angeles area (Dreher & Harley, 1998). Particulate matter emissions, including BC, from diesel engines is often referred to as diesel particulate matter (DPM). The carbonaceous constituents in particulate matter exist in two basic forms: organic carbon (OC) and elemental carbon (EC). The former is unstable and thus volatile, and the latter is the non-volatile form that can be measured. For all intents and purposes, BC is the same thing as elemental carbon (EC). The difference lies in the fact that although both can absorb light in the visible range, EC is not reduced to CO2 when heated to 800°C in an inert atmosphere (Dutkiewicz, Alvi, Ghauri, Choudhary, & Husain, 2009). Additionally, a distinction between the two is made depending on the application and method of

14 measurement. Whereas EC is measured through thermal methods, BC is measured through optical techniques (Hitzenberger et al., 2006). BC or soot is the microcrystalline graphitic component of particulate matter that appears black when deposited on filter material, hence the name BC. Because there are no known secondary mechanisms for its production from airborne precursors, is assumed to only be formed by combustion processes. Also, because it is inert to transformation in the atmosphere, it makes it a good tracer for combustion emissions (Hansen, Rosen, & Novakov, 1984). The lifetime of BC or EC in the atmosphere is in the order of days to weeks depending on meteorology, thus transport over long distances could play a role in its regional distribution (Allen et al., 2009). However, BC is mostly associated with local sources such as those seen with motor vehicle emissions. As mentioned previously, BC is measured through optical methods. One is through the aethalometer, an instrument that measures the attenuation of light by a sample collected on a filter matrix. This instrument works under the principle that attenuation of light (defined as sigmaATN) passing through a filter material is directly related to the mass of the material being absorbed, this attenuation remains constant over time at the spot where the material is being collected (Snyder & Schauer, 2007). Establishment of the validity of optical measurement methods, such as the aethalometer, has been attempted by studies that compared them to thermal methods that measure elemental carbon (EC) concentration. Research has shown that BC and EC though very similar, are different in some ways. Both BC and EC represent a comparable fraction of carbonaceous aerosol, but have somewhat different thermal, optical, and chemical behavior (Venkatachari et al., 2006). In some inter-comparison studies, concentration results differed by a factor of two. These results may be a consequence of the different mixing states of the aerosol in the presence of other compounds, as well as on the effects on size distributions (Hitzenberger et al., 2006). In addition, the filter material onto which particles collect can also have light-scattering properties (Snyder & Schauer, 2007). This effect may produce an overestimation of light absorbed carbon as the attenuation cross section fluctuates in response to scattering effects (Hitzenberger, 1993).

15 Several studies have attempted to calculate correction factors for the light scattering effects of the filter material and apply them to thermal or thermo-optical instrumentation. However, limited success has been obtained and thus a standard reference method for measuring the stable portion of carbonaceous material has not been established (Snyder & Schauer, 2007). Despite these issues, BC measurement by optical methods is still considered good surrogate for EC, as measured by the more expensive thermal or thermo-optical instrumentation. Thus because of its ease of use and low maintenance characteristics, the aethalometer is today, one of the most widely used instruments to measure ambient BC or EC (Magee Scientific, 2005). This instrument, along with other collocated equipment, was used by Y. F. Zhu, Hinds, Kim, Shen, and Sioutas (2002a) and Y. F. Zhu et al. (2002b) during pollutant measurement campaigns at two important California freeways. The same type of instrument, made by Magee Scientific, was used by Westerdahl, Fruin, Sax, Fine, and Sioutas (2005) during a mobile measurement campaign of pollutants along a freeway and residential streets of Los Angeles. Similarly, in an epidemiological study, Maynard, Coull, Gryparis, and Schwartz (2007) utilized a Magee Scientific aethalometer to assess BC levels in a Boston community to study the association between traffic-related pollution and mortality.

Polycyclic Aromatic Hydrocarbons Polycyclic aromatic hydrocarbons (PAHs) are organic compounds having a benzene ring as their basic structural unit. The best known compound in this group is benzo-a-pyrene, a five-ringed PAH (Godish, 2004). These compounds are ubiquitous in the urban atmosphere and were one of the first atmospheric compounds to be classified as carcinogenic (Baek et al., 1991). Some sources of PAHs include gasoline and diesel motor vehicle exhaust, biomass burning of agricultural and forest fuels, coal combustion, cigarette and wood smoke, as well as other fossil fuels (Ohura, Noda, Amagai, & Fusaya, 2005; Polidori, Hu, Biswas, Delfino, & Sioutas, 2008). According to the classification of the International Agency for Research and Cancer (IARC), some PAHs are major mutagenic or carcinogenic agents. As a result, 16 of them have been listed as priority pollutants by the U.S. EPA (Gutierrez-Daban, Fernandez-Espinosa, Ternero-Rodriguez, & Fernandez-Alvarez, 2005).

16 Atmospheric PAHs are found in the gas phase, the particle phase, or partitioning between both. However, most PAHs, especially those with more than four rings, readily adsorb or condense onto combustion aerosols (Polidori et al., 2008). It is because of these properties that most high density PAHs, which happen to be the ones listed as priority pollutants, are associated with particulate matter or soot, especially that in the respirable or inhalable range. Moreover, the PAH content in urban air is considered a sample of the various sources with a small contribution from outside that area (Baek et al., 1991). In urban areas, PAH concentrations are directly affected by proximity to sources and the intensity of emissions (Motelay-Massei et al., 2005), such as that seen next to roads. It is estimated that many of the toxic components present in respirable particles, including some particle-bound PAHs, are emitted by diesel engines. In fact, looking at the IARC listing for motor vehicle emissions, a designation of probable carcinogen was given to diesel-related emissions while a designation of possible carcinogen was given to gasoline-related emissions (Yadav et al., 2010). Polycyclic aromatic hydrocarbons can be measured using a photoelectric aerosol sensor (PAS) that works under the principle of photoionization of particle-bound PAHs by a high-intensity UV light source. This instrument was employed by Levy et al. (2003) during a campaign that assessed the influence of vehicular traffic on particulate matter and PAH concentrations in an urban setting. Similarly, Westerdahl et al. (2005) used this instrument in the Los Angeles campaign previously mentioned. Another way to measure particle-bound PAHs is through the Magee Scientific aethalometer. In its dual or seven-wavelength channel version, this instrument can measure the concentration of particle bound PAHs through their attenuation of light at a wavelength of 370nm. Because of its portability and ease of use, this instrument is suitable to get an estimation of total particle-bound PAHs. However, this measurement is limited by the fact that the attenuation cross section of the many different types of particle-bound PAH compounds is highly variable. Thus quantification of specific PAH compounds is not possible (Magee Scientific, 2005).

17

Particulate Matter Particulate matter is a general term used to describe the mixture of solid particles and liquid droplets in air that vary in size, shape and density. It is typically composed of mixed chemicals that are strongly dependent on their source. Because of the public health implications, one of the most important ways to describe particulate matter is by size. Particle size influences health by affecting uptake and deposition in the human lung (Salma et al., 2002). It is based on aerodynamic equivalent diameters (AEDs), a term that refers to the diameter of a unit density sphere that would have the identical settling velocity as the particle (North American Research Strategy for Tropospheric Ozone, [NARSTO], 2004), thus it is a convenient way to study and manage particulate matter in air. Other ways to describe particulate matter are through characteristics such as origin, formation mechanisms, and chemical composition (NARSTO, 2004). All of these characteristics influence the behavior of particles which coupled with meteorological conditions will cause particles to constantly change size, sorb, collide, adhere to each other, or be removed from the atmosphere by depositional processes (Godish, 2004). Moreover, the size distribution of particles can be expressed in terms of particle number, mass, volume, and surface area. When both particle number and particle concentration from an urban setting are plotted in distribution plots, it is evident that the particles follow multimodal distributions. From these plots, particles are described in terms of two major fractions: coarse particles that have AEDs ranging from 2.5 to 10 µm (PM10) and fine particulate matter with AEDs less than 2.5 µm (PM2.5). In this study we assessed fine particulate matter in two sub-fractions or modes: the accumulation mode (0.1-2.5 µm) hereafter referred to as PM2.5, and the ultrafine mode (0.001-0.1 µm) hereafter referred to as ultrafine particles (UFPs).

PM2.5 Fine particulate matter is composed of carbonaceous material such as EC and OC, and other materials such as organic compounds, transition metals, reactive gases, particle-bound water, as well as inorganic ionic species such as sulfate, nitrate and ammonium (Valavanidis et al., 2008). While primary or directly emitted PM is formed from combustion and high temperature processes, secondary PM is formed by condensation,

18 coagulation or evaporation reactions converting gaseous precursors into solids. The typical atmospheric lifetime of PM2.5 is from days to weeks and can travel from hundredths to thousands of kilometers. The known removal processes of this PM fraction are wet and dry deposition (NARSTO, 2004). Fine particulate matter constitutes about one half of the particulate mass found in urban air (Godish, 2004) and approximately 20-50% of PM2.5 is composed of carbonaceous material (Brook et al., 2007). In the US, approximately 25-33% of PM2.5 comes from OC, derived from biogenic and anthropogenic sources. Conversely, approximately 5-14% of PM2.5 is composed of EC, which is mainly derived from the incomplete combustion of fossil fuels (Godish, 2004). Furthermore, emissions inventories, organic speciation measurements, and receptor model applications have shown that indeed a fraction of PM2.5 is derived from fossil fuel combustion processes, however, these analyses also indicated that other sources significantly contribute to PM2.5 (Brook et al., 2007), thus associating this pollutant with regional transport (Levy et al., 2003). There are several ways of measuring the mass of suspended particles in air. The traditional way is through filter-based gravimetric or integrated methods, such as the Harvard Impactor (HI), similar to the EPA’s Federal Reference Method (FMR) used to for regulatory compliance. However, integrated methods such as these cannot detect short-term increases in particle concentrations, thus continuous instruments are needed to capture these changes that have important public health implications. Another reason why continuous instruments are favored over filter-based integrated methods is that the latter underestimate particle mass due to loss of semi-volatile material (K. Zhu, Zhang, & Lioy, 2007). Different types of continuous instruments measure short-term exposures. One is the tempered element oscillating microbalance (TEOM) in which the mass accumulation on a tapered element is proportional to the change in its frequency of oscillation (D. C. Green, Fuller, & Baker, 2009). Another is the beta attenuation monitor (BAM) that measures the mass concentration of particulate matter via the attenuation of beta particles found in the collected material (C. H. Huang & Tai, 2008). The Mini-Vol tactical air sampler measures the impaction of particles via a particle size separator (Air Matrix Corporation, 2005). Another type of instrumentation measures light scattering in a stream of particles to infer PM concentration where the amount of scattered light is proportional to the volume

19 concentration of the aerosol (Y. Wu, Hao, Fu, Wang, & Tang, 2002). These nephelometers or suspended particle-measurement instruments include the Dusttrak aerosol monitor and the Personal Data Ram (pDR). Quintana et al. (2000) assessed the ability of the pDR in predicting gravimetric mass by collocating active and passive pDRs with PM2.5 and PM10 Harvard Impactors. The personal nephelometers correlated well with the Harvard Impactors, specially the PM2.5 impactor. Similarly, C. F. Wu et al. (2005) demonstrated a good correlation between personal nephelometer readings against a Harvard Impactor. In another study, Chakrabarti, Fine, Delfino, and Sioutas (2004) demonstrated that the gravimetric measurement calculated from the pDR’s online filter, correlated well against a federal reference method for PM2.5 mass. In all of these studies the overestimation of PM2.5 concentration effect, due to high humidity conditions, was demonstrated on the performance of the pDR. However, this discrepancy was overcome by applying an appropriate correction factor developed in previous studies. For their ease of use and practicality, nephelometers have been used by Y. Wu et al. (2002) to measure vertical and horizontal profiles of particulate matter in Macao, China, and by Vallejo et al. (2004) to measure the personal exposure to PM in Mexico City residents.

ULTRA FINE PARTICLES Ultrafine particles (UFPs) are composed of carbonaceous material, metals, low-volatility organic compounds, as well as sulfate ions. Like PM2.5, ultrafine particles are formed from combustion and high temperature processes, as well as by condensation and coagulation of gas-phase precursors. In addition, UFPs can be formed by homogeneous nucleation of low-vapor pressure compounds. The typical atmospheric lifetime of UFPs is from minutes to hours and can travel up to tens of kilometers. Moreover, ultrafine particles are removed by wet and dry deposition or grow into the accumulation mode (NARSTO, 2004). Due of their small size UFPs contribute little to the overall mass of particulate matter. In fact, while most of the particle mass emitted by engines is in the accumulation mode (> 0.05 µm), most of the particle number emitted by engines is in the sub-micrometer (< 0.05 µm) range (Kittelson, 1998). In most urban areas UFPs represent over 80% of all

20 particles (Morawska et al., 1998). This suggests that in terms of number concentration, vehicular traffic is the principal source of UFPs. In addition, gasoline emissions generate particles in the size range of 40-60 nm (Ristovski, Morawska, Bofinger, & Hitchins, 1998) and diesel engines generate particles in the size range 60-120 nm (Harris & Maricq, 2001). Toxicological studies have shown that UFPs can penetrate deep in the human lung and that the pulmonary deposition of these particles, as well as surface area available for reaction, is linked to adverse health effects (Oberdorster, 2010). This, coupled with the fact that numerous tunnel studies have demonstrated that vehicle emissions are the most important source of ultrafine particles (Imhof et al., 2006), has sparked an interest in the scientific community to investigate the effects of UFPs in addition to fine particulate matter. Sampling of UFPs is a difficult task. First, because of their small mass, they cannot be collected in sufficient amounts to be gravimetrically assessed, and second, also because of their small mass, they cannot be separated from larger particles by inertial impaction (Schwartz, Sarnat, & Coull, 2003). One way to measure UFPs is by the scanning mobility particle sizer (SMPS). This instrument consists of an electrostatic classifier that charges particles to a known charge distribution, which are then classified according to their ability to transverse an electric field, and are later counted by a the condensation particle counter component (Hitchins, Morawska, Wolff, & Gilbert, 2000). Another way to measure UFPs is by the condensation nuclei counter (CNC), an optical method that uses a pulse counter to count the number of particles passing through a laser beam. This instrument however, is used in atmospheric applications dealing with cloud formation. A more simple and practical instrument used to measure UFPs is the P-Trak condensation particle counter. This instrument is mostly used to measure occupational exposures but recently has been used in field exposure studies (Schwartz et al., 2003). The concentration of the specifically-sized particles depends on the amount of incident light scattered as detected by the P-Trak’s photometric sensing chamber. Hagler et al. (2009) demonstrated that the P-Trak counter represented the ultrafine particle size range well. Since this instrument has an operational particle size range of 20-1000 nm, well above the UFP high limit of 100 nm, it was necessary to investigate its functionality at the target UFP range. For this purpose, the research team compared results of the P-Trak counter with a collocated condensation particle counter (TSI Model 3494) and

21 scanning mobility particle sizer (TSI Model 3071). The Pearson correlation coefficients obtained for the 10-70 nm, 70-100 nm and > 100 nm size ranges were R = 0.7-0.9, R = 0.5-0.7 and R = 0.1, respectively. This demonstrated the appropriateness of the instrument to measure the ultrafine particle fraction. In another study, Matson, Ekberg, and Afshari (2004) compared a P-Trak counter with a Condensation Particle Counter (CPC). Their results revealed that aside from having a good correlation between the instruments across 16 different data series, that the difference in concentration values between the two instruments was less than 20%. They also concluded that the 1-min data averaging times (the lowest capable by the two instruments) was sufficient to capture short term variations of ambient UFP concentrations. The P-Trak counter has been used in a number of studies to assess UFP concentrations in urban settings (Levy et al., 2003; Reponen et al., 2003), and also personal exposure to sub-micrometer particles in the city of Taipei (Chan, Chuang, Shiao, & Lin, 2004).

Carbon Monoxide Carbon monoxide (CO) is a colorless, odorless, and tasteless gas. It is produced naturally from the oxidation of methane and non-methane hydrocarbons. The major anthropogenic sources of CO are combustion and industrial processes, as well as from biomass burning. Carbon monoxide is formed as an intermediate product in combustion oxidation of fossil fuels and biomass (Godish, 2004). If not enough oxygen is available for the reaction, CO will form from the incomplete combustion of the hydrocarbon precursor. In the decades prior to air quality regulation in the United States, CO emissions from vehicles used to be in the order of 90% of total CO emissions. Today, CO from on-road vehicle emissions accounts for approximately 59% of total CO emissions (EPA, 2011a, 2011b). Nonetheless, in urban areas most CO still comes from vehicle emissions. Carbon monoxide emissions are mostly associated with gasoline engines since they tend to operate under rich fuel-air mixtures resulting in the formation of intermediate CO. Contrastingly, diesel engines tend to operate under leaner fuel-air mixtures resulting in less CO emissions. Because CO is associated with vehicular emissions, its levels tend to vary on temporal and spatial scales. In addition, meteorological conditions seem to influence CO

22 levels. For example, higher CO levels are observed in the winter time when cold conditions favor poor combustion, and also when sink or removal processes are less efficient (Godish, 2004). Furthermore, CO levels tend to be even more elevated in cities located at higher elevations. Finally, CO also affects tropospheric concentrations of hydroxy-radicals and ozone, thus increasing the oxidizing potential of the atmosphere (Godish, 2004). In essence, CO serves as a low-level precursor for tropospheric ozone. Carbon monoxide is measured by optical remote sensing devices (ORS) that utilize a light beam in the visible or UV range projected over open paths to spatially measure gaseous pollutant concentrations in the intersected air column using optical absorption spectroscopy (Baldauf et al., 2008). The Q-Trak Plus monitor has mostly been used to measure indoor air quality. However, a number of studies have employed this instrument to measure outdoor CO concentrations. For example, Westerdahl et al. (2005) used the Q-Trak Plus to measure ambient CO concentrations from a mobile sampling platform in freeways and residential streets in Los Angeles. In a study performed in China, Zhao et al. (2004), utilized the Q-trak Plus to evaluate pedestrian exposure to CO and other pollutants on sidewalks along busy streets. This instrument has also been used in comparative studies of indoor and outdoor air. Wong, Sin, and Yeung (2002) measured CO levels in parked vehicles and compared them to outdoor air. These studies confirm the usefulness and reliability of the Q-Trak Plus monitor in measuring indoor and outdoor air quality.

Nitrogen Oxides The term nitrogen oxides (NOx) refers to the sum of nitric oxide (NO) and nitrogen dioxide (NO2) because they are readily converted from one to the other in air. Nitric oxide is a colorless, odorless, tasteless, and relatively nontoxic gas. It is produced naturally in soil through biological nitrification and de-nitrification processes. It is also produced anthropogenically during high-temperature combustion processes involving oxygen and nitrogen (Godish, 2004). In internal combustion engines, where high-temperature combustion and rapid cooling occur, significant NO emissions ensue, especially during the early morning rush hour. Once in the atmosphere and in the presence of tropospheric ozone and volatile organic compounds (VOCs), NO is readily oxidized to NO2 in photochemical reactions, with

23 levels peaking at mid-morning. Alternatively NO can be oxidized to NO2 in a slow reaction with oxygen. Because of its high oxidation rate, NO2 is relatively toxic and corrosive having a characteristic yellow to brownish color (Godish, 2004). The most important characteristic of NOx species is that they are the precursors of tropospheric ozone, another air pollutant that has significant public health implications. However, because of its relative toxicity, NO2 is considered the most dangerous roadside gaseous pollutant (Lam et al., 1999). In the United States, motor vehicle emissions comprise approximately 36% of total NOx emissions. Diesel engines are the highest sources of NOx because of the high temperature achieved in their turbo or super-charged designs (Godish, 2004). In addition, because of this design, NOx emissions cannot be controlled using the same types of catalytic converters that gasoline-powered vehicles use, thus requiring the need for distinct regulatory measures (Kittelson et al., 2004). Nitrogen dioxide is measured using passive diffusion tubes in which through molecular diffusion, NO2 moves across a chemical absorbing-matrix. The tubes are subsequently analyzed in the laboratory where the amount of absorbed material is used to calculate the average ambient NO2 concentration over the exposure period which, can be up to one month (Janssen, van Vliet, Aarts, Harssema, & Brunekreef, 2001). Nitrogen oxides can be measured in real time using a specialized chemiluminescence analyzer (Westerdahl et al., 2005). One more way of measuring ambient nitrogen oxides is with Ogawa passive samplers (Owaga & Company, Osaka, Japan). These simple devices contain specifically-coated pads that react with the gaseous pollutants. The exposed pads are then extracted in the laboratory and the average ambient NOx concentrations are calculated over the exposure periods through colorimetric techniques (Owaga & Company, 2006). Sather, Slonecker, Mathew, Daughtrey, and Williams (2007) collocated Owaga samplers with monitors at six federal reference method (FMR) locations in El Paso, Texas. Their results showed a high linear correlation (R = 0.95) and an absolute difference of only 1.2 ppb between methods. Additionally, the results showed that the collective annual arithmetic mean NO2 values were identical to the NO2 means from the FRM monitors. For their ease of use and affordability, Owaga passive samplers have been extensively used in studies that associate traffic-related pollution with adverse health outcomes

24 (Escamilla-Nunez et al., 2008; Holguin et al., 2007; J. J. Kim et al., 2004) and also for validation purposes such as to test the reciprocal accuracy of data from fixed site monitoring stations and the passive sampler (Escamilla-Nunez et al., 2008; Mukerjee et al., 2009; Singer, Hodgson, Hotchi, & Kim, 2004; Van Roosbroeck et al., 2007).

Assessing Community Exposure to Roadside Air Pollution The breadth of epidemiological evidence on the effects of pollutants from vehicle emissions, in particular next to roadways, has motivated academia, policy makers and, most importantly, communities to study the characteristics of traffic pollution. However, this is not an easy task considering that the gaseous and particle components of vehicular pollution are complex, not just in chemical and physical properties, but also because of the locations where they are emitted. Nevertheless, recent research has elucidated common facts about traffic-related pollutants, and ways of assessing their levels, that are applicable to various settings. The following is a discussion of factors that influence pollutant levels next to roadways as well as different methodologies used for assessing their levels. As a whole, all the studies relating to these topics increase the knowledge pool on the subject at hand. The ultimate goal of this knowledge is to protect public health through legislation and community programs that enforce the rules and educate people about the potential health risks of vehicular pollution. Much work on this subject has been done in developed countries; the challenge now is to continue the work in developing nations.

Factors Affecting Roadside Pollution Concentrations As previously mentioned, the mixture of pollutants from vehicle emissions is composed of different gaseous and particulate material, which is formed from primary and secondary processes. The size of the particulate matter material is important because of the respiratory deposition properties of differently-sized particles (Krudysz, Moore, Geller, Sioutas, & Froines, 2009). For example and as mentioned previously, gasoline engine exhaust is associated with particles in the 40-60 nm range (Ristovski et al., 1998), while diesel engine exhaust is associated with particles in the 60-120 nm range (Harris & Maricq, 2001).

25 Exhaust particles are emitted primarily as sub-micrometer aerosols from engines (Morawska et al., 1998). The process starts when upon rapid cooling of exhaust precursor gases, nucleation occurs, thus rapidly increasing UFP particle number (Kittelson et al., 2004). In fact, relative to background concentrations, vehicle exhaust emissions can increase UFP particle number concentrations by two orders of magnitude (Kumar, Fennell, Langley, & Britter, 2008). Under certain conditions, ultrafine particles undergo changes through condensation of low volatility products, evaporation of higher-volatility particle-bound species, dilution with clean air, and coagulation or enlargement of particles when they adhere to each other (Krudysz et al., 2009). With increasing distance from the road, which usually occurs within tens of meters form the road (Hitchins et al., 2000) bigger particles become the dominant species. It is then important that aside from measuring number concentration of particles, their size distribution is assessed. The transformation or evolution of UFPs can be influenced by wind direction and speed. For example, in two successive studies (Y. F. Zhu et al., 2002a, 2002b), pollution measurements were conducted at increasing distances from two major California highways, one dominated by gasoline-powered traffic and the other by diesel-powered traffic. The results of both studies were similar. Under downwind conditions, UFP number concentrations decreased exponential between 17 and 150 meters and at 300 meters, the concentrations were no different than those measured at an upwind location. Their results demonstrated that wind speed and direction are important in determining the characteristics of UFPs near freeways. In addition, the same exponential decrease was seen for CO and BC concentrations in both studies indicating that vehicle emissions were the major source of these three pollutants (Y. F. Zhu et al., 2002a). Conversely, Hitchins et al. (2000) measured the horizontal distribution of particles from three different directions, and roughly the same wind speed, in relation to sampling point next to two major roads. Under downwind conditions and at 100-150 meters from the roads, total particle number and PM2.5 concentrations were approximately half of their maximums observed at roadside. When the wind was blowing parallel to the road, concentrations were halved at 50-100 m. Under upwind conditions, total particle number concentrations did not change after 15 meters. However, when they measured the change in

26 particle size, no significant change in the size distribution was observed. This suggested that wind speed did not influence the transformation of particles. Y. Wu et al. (2002) measured PM2.5 concentration profiles at different vertical and horizontal distances from a major road, and observed that at a maximum vertical distance (79 meters), PM2.5 concentrations measured roadside had decayed. This result indicated the influence of vehicular traffic on particle matter concentrations. Conversely, for the horizontal measurements, they observed only a modest 10% decrease at the furthest distance (228 meters) in comparison to the roadside result. This indicated that dispersion of pollutants may be affected by buildings or other tall structures. This has implications for people at urban centers where pollution may be trapped by the area’s infrastructure. The concentrations of traffic-related particulate matter and gaseous pollutants can also be affected by temperature inversions and the topographical characteristics of a region. For example, in a Canadian study (J. Wallace, Corr, & Kanaroglou, 2010), people living in the valley portion of the city of Hamilton, Ontario, were found to be exposed to higher concentrations of NO2 and PM2.5 during temperature inversion periods. Contrastingly, measurements taken above the valley in other areas of the city were much lower during temperature inversions. In this situation, changes in wind direction helped dissipate pollution above the valley. Other studies have compared seasonal (i.e., winter versus summer, etc.) and diurnal (i.e., daytime versus evening) concentrations of traffic-related aerosol pollutants. Singh, Bowers, and Sioutas (2006) measured higher average total particle number during the winter season compared to the summer and spring at various urban sampling locations. They also measured higher total particle number in the evenings. Their conclusion was that the contribution of local vehicular emissions to ambient pollution was exacerbated in the cooler months. Conversely, they argued that the effect of long-range transport of particles was enhanced in the summer months. Similarly, Kuhn, Biswas, Fine, Geller, and Sioutas (2005) performed a series of roadside measurement at a major California freeway at different times of the day, and also during different seasons through an extended study. They found that PM2.5 concentrations were similar during both winter and summer seasons. Contrastingly, UFP mass concentrations were higher in the winter. The authors suggested that cooler conditions

27 favored particle formation by enhancing nucleation of vehicular exhaust. The same argument was used for the higher UFP number concentration observed in the evenings. The high degree of temporal variability shown in these and other studies indicates that aside from traffic volume, meteorological variables such as degree of solar radiation, atmospheric mix depth, humidity and temperature, contribute to the temporal variation in particle number concentration at a particular location (Singh et al., 2006). As has been presented, traffic-related pollution shows marked differences in concentration at different distances from the road. These differences are a function of the physical and chemical properties of the pollutants, as well as external factors related to meteorology. Therefore, traffic related pollution, especially UFPs, is characterized for having strong spatial and temporal gradients across urban areas. This is important because it helps predict or estimate exposure in unmonitored settings (Levy et al., 2003). Many studies have been conducted to measure these gradients at different locations or to evaluate conditions at certain “hot spot” microenvironments. One way to assess the spatial and temporal profiles of traffic-related pollution is through direct measurements conducted from a mobile platform along roadway segments. This approach is useful in identifying multiple trouble areas within an urban area. Westerdahl et al. (2005) used a non-fossil fuel-powered vehicle equipped with specialized instrumentation to continuously measure UFPs, CO, BC, PM2.5 and NOx along a California freeway as well as some residential neighborhoods. They found that UFP, NO and BC concentrations were up to 20 times higher along the freeway route than at the residential area, with the values increasing sharply in the presence of diesel trucks. On the other hand, PM2.5 showed little variation across the driven area. Similarly Weijers, Khlystov, Kos, and Erisman (2004) used a mobile platform to measure particle matter number concentrations along different urban and rural routes in the Netherlands. They found that particle number concentrations increased as the degree of urbanization increased. They also found that while the lowest values were seen at rural areas, the highest values correlated with high traffic locations. They also observed that particle mass was less variable than particle number, this validated the importance of particle number as an indicator if vehicle emissions.

28 Another way to assess the spatial and temporal variability and its relation to traffic is through direct measurements of pollutants and traffic volume over a wide range of locations. Levy et al. (2003) measured PAHs, UFPs and PM2.5 as well as traffic volume at nine roadside locations in a suburb of Boston, Massachusetts. They observed substantial variations in UFPs and PAHs across the sites with both pollutants correlating well against each other. Conversely, PM2.5 measurements showed less spatial variation across sites and little correlation with the other pollutants. Furthermore, when using traffic volume results, stratified in predefined density scores, as predictors of pollution at each location, their results showed good prediction of all scores for PAHs, prediction at the 100 and 200 meter radii for UFPs, and no prediction for PM2.5. Once “hot spots” or trouble traffic areas are identified in an urban setting, investigators can use this information to determine the potential exposure of people to traffic-related pollutants. Residential areas, work settings, or schools in close proximity to roadways are prime spots to measure traffic-related pollution. For instance, in a study conducted by Lena, Ochieng, Carter, Holguin-Veras, and Kinney (2002) in the South Bronx, NY, sidewalk measurements of PM2.5 and elemental carbon (EC), as well as simultaneous auto and truck traffic counts were conducted. Whereas the authors demonstrated little correlation of PM2.5 with either auto or truck traffic, local variations of traffic density, especially from trucks, were strongly associated with spatial variations of EC. This was important because heavy duty trucks readily circulate through this community. In another study performed by Patel et al. (2009) in New York, PM2.5 and BC measurements at three urban high schools located within 50 meters of varying-density roadways were conducted. In addition, they conducted measurements at two high schools located in suburban areas. The authors found significant urban to suburban differences in 24 hr PM2.5 concentrations. However, no significant differences in PM2.5 concentrations were found among the urban schools. For the BC measurements, the authors found that concentrations were 2-3 times higher at the urban schools compared to the suburban schools. In addition, the highest BC levels were seen at the schools adjacent to highways. As has been presented in this discussion, traffic-related pollution will vary considerably from location to location in an urban setting. Therefore the importance of assessing pollution levels at these microenvironments is essential to get a true estimate of

29 potential exposures to people spending a significant amount of time in close proximity a roadways. Unfortunately, current air pollution programs do not have the ability to assess the effects of traffic pollution on a localized subpopulation or geographic clustering (Levy et al., 2003). In the United States, as part of EPA’s Ambient Air Monitoring Program, pollution levels within an urban setting are routinely monitored through established monitoring stations or fixed-site monitors located in strategic urban areas, but not necessarily next to major roads. Aside from using the collected data for surveillance, trending, and regulatory compliance purposes, the data from these monitors is used to estimate background pollutant concentrations by various sources, including regional ones (Y. L. Huang & Batterman, 2000). Many countries have followed suit in establishing their own air pollution monitoring programs as well as setting their own standards. This is important for establishing a regulatory framework which, as in the case of the US and other developed countries, has been successful in mitigating the effects of air pollution. However, knowledge that has been acquired in relation to the local effects of vehicular pollution needs to be passed on to developing nations in terms of education and infrastructure, as well as research opportunities.

TIJUANA, B.C. The city of Tijuana, located in the Mexican state of Baja California, together with the city of San Diego in the US state of California, is part of the group of cities along the nearly 2000-mile stretch that constitutes the US-Mexico border region. The California-Baja California region in particular, is very dynamic with its expanding economies and increasing industrialization (Ganster, 1996). Consequently, this region has one of the fastest population growth rates along the border region. In fact, 44% of the total border population is concentrated in this area (Anderson & Gerber, 2008). In the San Diego-Tijuana area, the population is expected to be 5.7 million in 2020, an increase of 43% in just 20 years (Sweedler, Fertig, Collins, & Quintero Nunez, 2003). In its relatively short existence, Tijuana has experienced an explosive population growth, especially in the period after 1970 in which an unrelenting annual growth rate of 3.8% has doubled its population every 10 years (EPA, 2000; Sweedler et al., 2003).

30 Numerous reasons explain this explosive demographic growth. Starting in the 1920s, the working population grew with the demand for entertainment and services by people on the US side. Subsequently, the demand for workers in the United States during World War II resulted in the Bracero program which in turn attracted many people to the city hoping to cross the border to obtain employment. The establishment of the border industrialization program (BIP) during the 1960s increased the population growth rate by attracting workers from other areas of the country to work in the maquiladora industry. The implementation of the North American Free Trade Agreement (NAFTA) in 1994, fueled an already growing population in Tijuana. It increased investment and construction of infrastructure to cope with the demand for labor and services. These events sparked the economic growth in the city that, because of its close connections to the US economy, shielded it from Mexico’s stagnant national economy (Ganster, 1996). All of these factors have led to Tijuana’s extensive population, which in 2010 was measured at 1.6 million people (INEGI, 2011). Tijuana’s industrialization and rapid increase in population has led to numerous environmental issues, including poor air quality. The consumption of fuels by vehicles constitutes the principal source of emissions in the California-Baja California region (Quintero Nunez et al., 2006), especially in Tijuana. In fact, in less than 15 years, the vehicular fleet in the city has more than doubled from 363,000 to more than 737,000 total vehicles (EPA, 2000; LT Consulting, personal communication, February 2, 2011). In terms of technologies and infrastructure to control vehicle emissions, Tijuana fares much lower in comparison with San Diego, even with the fact that both cities are located within the same air basin. The chronic infrastructure deficit that plagues Mexican border cities like Tijuana is related to environmental problems including air quality (Ganster, 1996). Adding to this problem is the older, unregulated, and poorly-maintained vehicular fleet that continues to increase in Tijuana. Older and likely unregulated vehicles from the United States are frequently acquired by people in Tijuana. Moreover, a consequence of the maquiladora industry and the NAFTA agreement has been a significant increase of the heavy duty truck fleet (Ganster, 1996; Sweedler et al., 2003) that circulates daily through city roads transporting goods to the different manufacturing plants or across the border into the US.

31 The issues of air quality in Tijuana are further compounded by the climate of the region. Tijuana and San Diego belong to the same air basin. As a consequence, pollution from the US can affect the Tijuana. In this region, westerly winds are predominant, and in combination with Santa Ana winds, move pollution toward the eastern portion of Tijuana (EPA, 2000) where many of Tijuana’s working class reside. Furthermore, air quality is affected by temperature inversions that momentarily trap pollutants, including those generated by vehicle emissions. In conclusion, the city of Tijuana, despite its economic growth, faces many environmental challenges that include poor air quality generated by vehicle emissions as well as by industry. Such issues that occur not only in Tijuana but in all cities along both sides of the US-Mexico border, have prompted the mobilization of agencies and organizations on both sides to come up with bi-national solutions to mitigate the myriad of environmental issues in the border region. Starting with the 1993 La Paz agreement through the current Border 2012 program, the US Environmental Protection Agency and Mexico’s environmental agency SEMARNAT, in combination with organizations such as the Southwest Consortium for Environmental Research and Policy (SCERP) as well as other US and Mexican federal, state and local agencies, have made significant gains in documenting and mitigating environmental issues in the US-Mexico border region. However, further investigation to support the implementation of environmental policy is needed, especially as the population of the Tijuana-San Diego is projected to continue growing for years to come.

32

CHAPTER 3 METHODOLOGY DESIGN OF THE INVESTIGATION Fifty-five (55) total sites located in different areas across the city were selected for roadside measurements (see Appendix A, Figure 8). The roadways chosen varied in type (federal highways, arterial roads and local roads), number of lanes (1-4) and direction (one or two-way). Concurrent roadside pollutants and traffic volume measurements were conducted at all sites, except for one site (1000, away from a road) where only pollutants where measured. In addition, the data collected were used to assess the temporal (subset of 10 sites) and spatial (all sites) characteristics of pollutants. Furthermore, data collected at sites located in proximity to elementary schools (subset of 25 sites) were used to assess children’s potential exposure to roadside pollution. The comprehensive list of all sites is found in Appendix B, Table 7. In order to manage the sites across the city, they were grouped into artificially constructed zones: Zone 1 (NW) located near Playas de Tijuana; Zone 2 (SYBC) located near the San Ysidro border crossing; Zone 3 (Mid-City); Zone 4 (5 y 10) a high traffic area located near maquiladora workers residences; Zone 5 (OMBC) located near the Otay Mesa border crossing; Zone 6 (Cumbres) designated as a background area; Zone 7 (SW); Zone 8 (NE); and Zone 9 (SE). The zones, shown as shaded areas in Figure 1 are for illustrative purposes and are not meant to define boundaries. In addition, because the sites were selected arbitrarily, they are not representative of their respective zones. Nevertheless, the zones are useful in organizing the sites as well as for interpreting spatial variation.

Spatial Assessment From August 17, 2009 through June 1, 2010 and from November 16, 2011 through January 25, 2011, measurements (ranging from 1 to 20) were taken at all sites. For consistency, with the exception of the subset of samples used for temporal analysis,

33

Figure 1. Measurement zones or sectors. measurements at all other sites were conducted during mid-day (1000-1559) hours. These sites were selected by choosing scattered roadside locations in areas of varying traffic densities. However, because no official information on traffic density and characteristics was available prior to initiation of the study, selection of these locations was done arbitrarily (through Google Earth software, version 5.2.1 1588) by searching for different roads of seemingly varying traffic densities.

Temporal Assessment From August 2009 through June 2010, multiple (≥ 5) measurements were conducted at each of nine pre-determined roadside locations. An additional location (site 1000) measured several times later in the study was added to this group (see Figure 2). Measurements at these sites were not only conducted during different months, but were also conducted during mid-day (1000-1559) and evening (1600-2159) hours. In addition, measurements were conducted during different days of the week, including weekends in some instances. Selection of these sites was based on regional air movement where northwesterly wind direction is predominant during the day (EPA, 2000).

34

Figure 2. Placement of sites for temporal analysis.

School-Related Assessment From November 2010 through January 2011, two (2) measurements were conducted at each of 18 pre-determined roadside locations in proximity to schools (17 public elementary schools and one public university). The geographical location of these sites was obtained from an online government list of Tijuana schools (Appendix C, Table 8). In addition, five sites (100, 200, 300, 400 and 500), corresponding to each zone, were chosen at locations not in proximity to schools, but in areas of higher traffic density. The reason for having these sites was for them to serve as high traffic references during a given measurement session. A measurement session consisted of first taking measurements at a specific zone’s high traffic reference site and then taking measurements at the corresponding school-related sites. In some instances due to the high number of sites (zones 2 and 5), measurements had to be conducted over two days. Furthermore, to include a low-pollution reference, a non-traffic background site (1000) located within the campus of the Universidad Autónoma de Baja California (UABC) was measured during each measurement session. An additional site (2000) located in the Cumbres background zone, was included in this portion of the

35 study. Thus 25 total sites were assessed (at least two times) in this portion of the study (Figure 3).

Figure 3. Placement of school-related measurement sites.

Categorical Variables To verify the estimated traffic density at the selected sites, auto and truck traffic counts distributions were used to categorize traffic density at each location (Appendix B, Table 7). These categories were included in the analysis of pollutant concentrations versus traffic count. In addition, other categorical variables such as traffic flow, sampling orientation in relation to wind direction (downwind versus upwind), day of the week, month of the year, and type of weather. Moreover, in order to assess the relationship between pollution levels and socio-economic conditions, all 55 sites were placed into two simple socioeconomic status (SES) categories. These categories were based on site observations, as well as on poverty area designations (polygons) from the 2004 Habitat Program for Tijuana, B.C. (Diario Oficial de la Federación, 2003).

36

DATA MEASUREMENTS This section details the parameters measured in this study, how they were treated, the instrumentation used to measure them, how the data obtained was analyzed, and finally how the instrumentation and data obtained were controlled for quality purposes.

Pollutants and Meteorological Variables The pollutants measured in the study were BC; UV-absorbing particulate matter (UVPM), an indicator of particle-bound polycyclic aromatic hydrocarbons (PAHs); ultrafine particulate matter (UFP), defined as having a particle size less than 0.1 micrometers; fine particulate matter (PM2.5); carbon monoxide (CO); and nitrogen oxides (NOx). Meteorological conditions were also measured within each microenvironment because these can influence pollutant levels. These included temperature and relative humidity (RH), as well as wind speed and direction. In addition, since wind direction has a strong influence on pollutant dispersion, the location of each site relative to wind direction (downwind versus upwind) was assessed at each site.

Traffic Counts and Composition Onsite manual traffic counts were not possible due to limited number of personnel. Therefore, during measurement sessions, traffic was video-recorded using the Flip Mino video camcorder (Cisco Systems, San Jose, CA). This hand held device is capable of recording up to sixty minutes of video. However, because it was impractical to download the video from every site during a measurement session, recording was limited to 10 minutes during each 20 minute measurement period. At a later time, and from each video record, cars and trucks were counted as separate categories. Personnel were instructed to count passenger cars, private cars and motorcycles as one category, and small and medium trucks, SUVs, vans, buses and heavy duty trucks into another category. Counting was performed every other minute during the first eight minutes of video and then the last two minutes to make a total of 6 minutes of actual traffic data. The values were then reported on a count per hour basis for analysis. Once all traffic counts were compiled, the distributions of both auto and truck counts were examined for stratification into three categories of traffic volume. This was done to

37 examine more closely the possible influence of traffic counts in pollutant levels. For autos, these categories were: ‘low’ (< 1330 vehicles per hour or vph); ‘moderate’ (1330-3331 vph); and ‘high’ (> 3331 vph). For trucks, the categories were: ‘low’ (< 200 vph); ‘moderate’ (200-465 vph); and ‘high’ (> 465vph; see Appendix B, Table 7).

TREATMENT With the exception of the NOx species measurements, all other pollutant measurements at each site were conducted roadside (0-5 meters), for periods of 20 minutes using an averaging period of one minute. These measurements were conducted using a university transport vehicle (2005 Ford Econoline XL) as a sampling platform at a height of approximately two meters above ground. For the NOx measurements, the sampling devices were secured to lamp posts or street signs, at a height of 3-5 meters above ground and passively measured for a period of 24 ± 1 hour. At the beginning of each measurement session and prior to leaving the laboratory, all the instruments were turned on. The sampling platform would then be driven to the first location, its engine shut off, and the instrumentation placed on the roof of the vehicle using cardboard receptacles with perforations appropriate for mounting the instruments and, if necessary, extending their components (e.g., P-Trak counter’s telescoping probe). In addition, due to the possible interference of CO measurements by the isopropyl alcohol employed with the P-Trak ultrafine particle counter, CO measurements were conducted ground level away from the rest of the instruments. Finally, while pollutants were being measured, video recording of traffic was conducted during a 10 minute window. After collecting measurements at the initial site, the instrumentation would be transported to the next site and the above procedure repeated until all of the sites scheduled during a measurement session (up to five sites) would be completed. At the end of each measurement session, the crew returned to the laboratory and proceeded to download all instrument data into a designated computer, convert it to an MS Excel readable format, and transmit it for review and processing.

38

Instrumentation At the beginning of each measurement session, all instruments were time-synchronized against a designated laptop computer into which their data was later downloaded. Measurements of BC and UVPM were made using the portable chassis dual-wavelength Aethalometer, Model AE42 (Magee Scientific Corporation, Berkeley, CA). This instrument provides real-time measurements of black or elemental carbon (EC) aerosol particles in a stream of air. As the sample is drawn at a determined flow rate (2 liter per minute [l/min] used in this study), it collects onto a section of quartz fiber filter tape. During collection, the instrument performs continuous optical measurements until the tape reaches a saturation point and has to advance (Magee Scientific, 2005). Fine particulate matter (PM2.5) was measured using the personal DataRam™ (pDR) Model 1200 (Thermo Electron Corporation, Franklyn, MA). The pDR-1200 is a nephelometric instrument that employs light-scattering technology to provide real-time mass concentration measurements. For the purposes of this study, a specific size-selective inlet cyclone was used to limit the measurement of particles with an aerodynamic size of 2.5 microns or less. Active operation of the pDR-1200 was conducted using a universal sampling pump Model 224-44XR (SKC, Eighty Four, PA). The pump had been previously calibrated before initiation of the study and it was operated at a flow rate of 4 l/min. Ultrafine particles (UFPs) were measured using the industrial hygiene P-Trak Ultrafine Particle Counter Model 8525 (TSI Incorporated, Shoreview, MN). This condensation particle counter employs isopropyl alcohol to enlarge sub-micrometer particles, thus making them easier to detect and count. As air is drawn into the unit, the particles serve as condensation sites for the alcohol, and then as the particles grow, an optical sensor counts them. Carbon Monoxide (CO) was measured using the Q-Trak Plus IAQ Monitor Model 8554 (TSI Incorporated, Shoreview, MN). For the purposes of this study, carbon dioxide, temperature and relative humidity (also measured by this instrument) were not considered. This instrument provides real-time concentration in ppm by way of sophisticated optical sensors. For this study, the Q-Trak monitor was operated under Log Modes 2 and 3. The

39 Q-trak Plus monitor has a sensor range of 0 to 500 ppm with an accuracy of ± 3% (or 3 ppm whichever is greater) per reading. Nitrogen oxides species (NOX and NO-NO2) were measured using passive samplers model PS-100 (Owaga Corporation, Osaka, Japan). The Owaga sampler contains two gas collection pads in its dual inlet configuration. The pads are 14.5 mm in diameter, and are specifically coated to react with a desired gas or gases. Analysis of the exposed pads is done in the laboratory to determine the average concentration of gas over the exposed period (Owaga & Company, 2006). Prior to deployment at each site, a sampler was assembled using NOX and NO2 pre-coated pads, stainless steel screen, diffuser cap, and sampler clip. Thereafter, the sampler was transported to the target location and attached to a secure location, such as a light pole, at a height of 2-5 meters above ground. The device was left to passively sample ambient air for a period of 24 ± 1 hour. A sampler was setup when the sampling crew would arrive at a particular location to measure other pollutants, and would return the following day to remove it. Extraction and chemical analysis of the exposed pads were conducted at the chemistry laboratory facilities of the Universidad Autónoma de Baja California, Tijuana campus. Reagent preparation for extraction and colorimetric analysis was done internally. The weights of NOx and NO2 were calculated using a standard curve prepared at the beginning a sampling period (three periods occurred during the entire study). Final concentrations of NO2 and NO were determined using conversions coefficients calculated based on temperature and RH conditions during the exposure periods. Temperature and relative humidity (RH) were measured using the HOBO U10 Data logger (Onset Computer Corporation, Bourne, MA). This instrument conveniently uses an internal sensor coupled with a data logger for the continuous recording of data. Wind speed and wind direction were measured using the Weather Wizard III station and WeatherLink Software (Version 5.8.3; Davis Instrument Corporation, Hayward, CA). This light weight and portable station conveniently measures microenvironment conditions. Its attached data logger can hold large amounts of information that can be easily downloaded at a later time.

40

Data Analysis Once a measurement session was completed, each instrument’s data was downloaded, converted to an MS Excel file, and submitted for review. After review and taking advantage of the synchronized time stamps across instruments, these individual files were combined into a single-session spreadsheet that contained all the instruments’ data for that session, the traffic counts, as well as coded variables, such as SES, sampling position, day of the week, for each site during the session. This session file was then analyzed for descriptive statistics using the Statistical Package for Social Sciences (SPSS; IBM, Arkmon, NY) version 19. Another MS Excel spreadsheet, referred to as the master file, was constructed using the descriptive information obtained from the SPSS analysis. The information was organized according to site identification number. As more measurement sessions occurred, the master file increased in content. Once all the study measurements were completed, the master file was analyzed once again using SPSS using median values for all continuous and categorical variables. Non-parametric Spearman bivariate correlations were used to assess the associations and significance between pollutants, and between individual pollutants and traffic counts. Furthermore, to compare the distributions and significance of continuous variables with categorical variables, two-sample independent t-tests by means of the Mann-Whitney U-test was used. In those instances where categorical variables contained more than three groups, the Kruskal Wallis test between pairs to examine the variance among groups was applied and if needed, it was followed by the Mann-Whitney U test. Significance was set as p < 0.05.

Quality Control In at least three occasions during the study, the pDR-pump assembly was operated under a laminar flow hood to calibrate its performance under low particle conditions. In addition, prior to each sampling session, the pDR-1200 was zeroed by connecting the green zeroing filter to the attached cyclone inlet and running the SKC pump. The SKC pump was calibrated at the beginning of each sampling session while attached to the pDR-1200 with a Bios Defender 510M primary calibrator (Bios International Corp., Butler, NJ) at a pump flow rate of 4.0 l/m.

41 Previous studies (Liu, Slaughter, & Larson, 2002; C. F. Wu et al., 2005) have reported baseline drift or unusually high values from operation of the pDR. A method to check for negative baseline drift in the passive version of the pDR (1000 model) reported by C. F. Wu et al. (2005), is used to compare the internal pDR-computed average (TWAC pDR) with the average of data points computed from the data file (TWAC manual). For the purposes of this study, an algorithm to verify if negative baseline drift [defined as (TWAC pDR–TWAC manual) ≤ 2.0 µg/m3] was written and applied to all session excel files. None of the data during measurement of target sites displayed negative drift. Short-term data outliers, defined as a concentration value 25 times greater than the previous or subsequent reading has been reported by C. F. Wu et al. (2005) for the passive pDR-1000 instrument. No such artifact was found after reviewing the pDR-1200 data from each measurement session. A detection limit of 1.6 µg/m3 has been reported for pDR nephelometers by L. A. Wallace et al. (2003). Therefore, an algorithm to check for any data below this value and replacing it with one half of it (0.80 µg/m3), was written and applied to all session excel files prior to analysis. Several PM2.5 mass concentration, as measured by this instrument, values below the limit of detection were observed during the study. The final and most important quality check for the pDR-1200 data was to ensure that the mass concentration readings were not overestimations due to ambient high relative humidity (RH). This effect is slight at RH > 60% but becomes large at RH > 95% (Chakrabarti et al., 2004). For the purposes of this study, we used the manufacturer’s recommendation of correcting data for RH greater than 70%. An algorithm was written to correct the one minute data points if they exceeded this parameter as measured from the collocated HOBO RH/Temp data logger. This algorithm was written based on the empirically-validated equation: pDR-1200corrected = pDR-1200 uncorrected/CF, where CF = 1 + 0.25 RH2 / (1-RH) (Chakrabarti et al., 2004). Since none of the RH values during measurements at target sites exceed 70%RH, no corrections were made by the algorithm. While reviewing the aethalometer data from measurement sessions, it was checked for data losses due to tape advancement after the filter loading reached a saturation point. This occurred in several instances because the unit could not be stopped and restarted between site measurements. For this reason, there were BC/UVPM gaps during many of the site measurements. In some cases, due to coincidence with tape advancement, no data was

42 obtained at a specific site. BC and UVPM data were also screened for sudden BC increases greater that 80 ng/m3 as well as sudden extreme negative values. Negative aethalometer values are displayed by the instrument in low concentration environments and can suddenly increase in the negative direction due to the programming of the unit. During this study, there were no sudden extreme changes in the data. However, there were some small negative values observed. For this reason and to recalibrate any value close to zero, we substituted any value less than 250 ng/m3 with 125 ng/m3. This algorithm was written and applied to all session MS Excel files prior to analysis.

43

CHAPTER 4 RESULTS ANALYSES ACROSS ALL SITES This section details the results obtained from data generated from all measured sites.

Descriptive Data The study took place from August 17, 2009, through January 26, 2011, with no measurements during the months of May 2010 and July through October 2010. Of the 55 sites selected, 33 were measured at least twice and up to 20 times (Appendix D, Table 9) and the other 22 were measured once (Appendix E, Table 10) during that time. BC/UVrelated particulate matter (BC/UVPM), fine particulate matter (PM2.5) and ultrafine particulate matter (UFP) measurements were intended at all of these sites. However, due to data losses from equipment malfunction, loss of power due to battery failure, or human error [BC/UVPM (17), PM2.5 (44) and UFPs (26)], measurements did not match the 210 expected total measurements for each pollutant. Nitrogen oxides (NOx) measurements were conducted at 22 sites. Most (17) of these were measured once, three were measured twice, and two were measured three times (Appendix F, Tables 11 and 12); thus the total measurements were 29. No data losses occurred from oxides NOx measurements. Lastly, CO measurements did not start until April 19, 2010, meaning that many of the chosen locations were not monitored for CO. Of those that were, 78 total measurements were taken, with no data losses reported. As observed in Appendix D, Table 9, the overall BC median concentration was 4437 ng/m3 with a large range from 293 to 94245 ng/m3 and the UVPM median concentration was 3769 ng/m3 with a range from 125 to 53792 ng/m3. The overall PM2.5 median concentration (uncorrected for scatter) was 26 µg/m3 with a range from one to 175 ng/m3. The overall median UFP concentration was 30265pt/cc with a range from 1314 to as high as 160193pt/cc. Additionally, the overall CO median concentration was 1.6 ppm with a range from 0.1 to 4.5 ppm. As for the nitrogen oxides species, the overall

44 NOx species median concentration was 285 ppb with a range from 120 to 697 ppb. The overall NO2 median concentration was 49 ppb with a range of 23 from 97 ppb. Finally, the overall NO median concentration was 216 ppb with a range of 93 from 610 ppb. With the exception of the background site (17 measurements) located inside the UABC campus where there was no vehicular traffic, video recording and subsequent counting of both, auto and truck traffic, occurred at the other 54 locations. From the total 193 measurements taken at these sites, 171 auto and truck counts were obtained. The remaining 22 traffic counts (autos and trucks) were inadvertently lost or not turned in for review, thus they were recorded as data losses. The overall median count for autos was 1758 ct/hr with a range of 0 to 4602 ct/hr, while the overall median count for trucks during the study was 180 ct/hr with a range of 0 to 912 ct/hr.

Pollutant Levels in Relation to Traffic Count The association between all pollutants and traffic counts across all seasons and sites is displayed in Table 1. BC correlated well with truck traffic counts (rho = 0.477, p < 0.001), and to a lesser extent, with auto traffic counts (rho = 0.377, p < 0.001). PM2.5 was not significantly correlated with traffic counts. UFP number concentrations had strong correlations with auto traffic counts (rho = 0.482, p < 0.001) and truck traffic counts (rho = 0.418, p < 0.001; Figure 4). In addition, CO was moderately correlated with auto traffic counts (rho = 0.348, p < 0.001) but not correlated with tuck traffic counts (rho = 0.156, not significant). Finally, NO2 had the strongest correlations with both auto and truck traffic counts of all pollutants (rho = 0.677, p < 0.001 and rho = 0.510, p < 0.001, respectively); albeit, the sample size for NO2 (n = 27) was much smaller than for the rest of the pollutants (n = 57-161).

Pollutant Associations The association between all pollutants is displayed in Table 2. As expected, BC and UVPM correlated very well with each other (rho = 0.985, p < 0.001), as these are from the same instrument. Carbon monoxide had strong correlations with BC (rho = 0.743, p < 0.001) and UFPs (rho = 0.736, p < 0.001), and a moderate correlation with PM2.5 (rho = 0.505, p < 0.001). Similarly, UFPs had a strong correlation with BC (rho = 0.636, p < 0.001) and a

45 Table 1. Pollutants versus Traffic Counts Spearman Correlations (All Data) Automobile Traffic Counts (ct/hour)

Truck Traffic Counts (ct/hour)

All Vehicle Traffic Counts (ct/hour)

Black Carbon 161

0.377***

0.477***

0.396***

UVPM

161

0.396***

0.485***

0.414***

PM2.5

128

0.075

0.026

0.079

UFPs

152

0.482***

0.418***

0.479***

CO

57

0.348**

0.156

0.328*

NOx

27

0.446*

0.278

0.469*

NO2

27

0.677***

0.510**

0.734***

NO

27

0.359

0.224

0.374

n

* p < 0.05 ** p < 0.01 *** p < 0.001

Figure 4. Association between UFPs and auto traffic counts.

rho

n

rho

n

rho

n

rho

n

rho

n

rho

n

rho

* p < 0.05 ** p < 0.01 *** p < 0.001

NO2

NOx

CO

UFPs

PM2.5

UVPM

Black Carbon 193

0.985***

UVPM

152

0.263**

152

0.272**

PM2.5

148

0.357***

169

0.645***

169

0.636***

UFPs

Table 2. Spearman Correlations between Pollutants (All Data)

68

0.734***

78

0.505***

71

0.736***

71

0.743***

CO

13

0.371

28

-0.013

22

0.038

27

-0.104

27

-0.076

NOx

29

0.672***

13

0.448

28

0.372

22

-0.103

27

0.300

27

0.305

NO2

0.519**

29

0.969***

13

0.272

28

-0.140

22

0.015

27

-0.185

27

-0.147

NO

46

47 moderate correlation with PM2.5 (rho = 0.357, p < 0.001). Finally, none of the NOx species were significantly correlated with the rest of the pollutants, however as expected, strong and significant correlations were observed among the three NOx species.

RELATIONSHIP OF POLLUTANTS WITH CATEGORIES OF TRAFFIC DENSITY AND FLOW To more closely examine the relationship between pollutant levels and traffic counts, we stratified traffic counts into three categories (see methods section). Moreover, during field measurements sessions we observed traffic flow at each site and recorded it according to predetermined categories. These categories were: low (defined as bumper-to-bumper heavy traffic flow); moderate (defined as intermittent flow, due to traffic signals and pedestrian crossings); and high (defined as continuous or free flowing traffic without major interruptions). The relationship between BC levels and categories of truck traffic, as well as the relationship between UFP and categories of all vehicle traffic are shown in Figure 5. Both BC and UFP levels increased as the traffic density increased. Contrastingly, only BC levels increased with increasing truck traffic flow. UFP levels increased up to a moderate or intermittent traffic flow and then decreased at during continuous traffic flow. Moreover, the concentrations of both pollutants and for both variables significantly increased from the low to moderate categories. Conversely, the changes from moderate to high categories in both variables and for both pollutants were not statistically significant. To explore the finding that UFP levels were higher (though not significantly) under moderate or intermittent traffic flow than under continuous traffic flow, we restricted the allvehicle traffic counts to a less variable range (1500-3000 ct/hr) and re-calculated UFP concentrations under the three categories for traffic flow. As seen in Figure 6, UFP levels were again the highest under moderate or intermittent flow conditions and significantly higher than under continuous traffic flow conditions (p = 0.028).

TEMPORAL AND SPATIAL ANALYSES This section details the spatial and temporal variability results obtained.

48

Figure 5. BC and UFP count median concentrations by truck and all vehicle traffic categories.

Figure 6. UFP count median concentrations by restricted traffic flow categories.

49

Temporal and Spatial Analyses across All Sites Pollutant level results from the entire data set were examined over the course of the different months during which the study took place. BC median concentrations showed some variability with modest increases in the fall and winter seasons. Similarly, UFP number median concentrations showed variability but with slightly higher levels in the fall and spring. On other hand, PM2.5 median concentrations did not appear to vary much across seasons. Appendix G, Figures 9-11 illustrate the temporal variability of pollutant levels throughout the study period. The following intra-month BC median concentrations were significantly different: levels in January were lower compared to November (p = 0.003) and September (p = 0.027); levels in March were lower compared to September (p = 0.027), October (p = 0.015), November (p = 0.001) and December (p = 0.005). Ultrafine particle number concentrations were significantly lower in March compared to January (p = 0.034), November (p = 0.021) and December (p = 0.003), as well as lower in September compared to December (p = 0.014) and lower in October compared to December (p = 0.030). There were no significant intra-month PM2.5 differences that had a valid sample size for analysis. In addition to month to month variability, we assessed the diurnal variability of BC/UVPM, PM2.5 and ultrafine particles at all site measurements. It is worth noting however, that most measurements (140-158) occurred during mid-day hours (1000-1559 hrs) with the rest (26-35) occurring during evening hours (1600-2159 hrs). Of significance in this analysis was that UFP number concentrations were higher during evening hours than at mid-day hours (n = 184, p = 0.004). The entire data set was used to assess spatial variability. In general, median BC and UVPM concentrations ranged from (1075-8253 ng/m3 and 732-8148 ng/m3) with the lowest and the highest concentrations measured at sites in the background and the SE zones, respectively for both pollutants. In turn, PM2.5 median concentrations ranged from 17 to 56 µg/m3 with the lowest and highest concentrations measured at sites located in the NE and mid-city zones, respectively. Ultrafine particle median number concentrations ranged from 9738 to 47,101 pt/cc with the lowest and highest concentrations measured at sites located in the background and 5 y 10 zones, respectively (Figure 7).

50

Figure 7. Variability of ultrafine particles across zones. Furthermore, CO median concentrations ranged from 0.3 to 3.2 ppm with the lowest and highest concentrations measured at sites located in the background and 5 y 10 zones, respectively. Finally, NO2 medium concentrations ranged from 27 to 92 ppb with the lowest concentrations measured at sites located in both the background and NW zones, and the highest levels measured at sites in the 5 y 10 zone. It is important to note however, that there were no NOx measurements taken at sites located in the NE and SE zones.

Temporal Variability at Multiple-Measured Sites A more robust analysis of pollutant variability across Tijuana was obtained by analyzing the coefficients of variance (CVs) among sites that were measured at least five times throughout the study. The CV values of these pollutants and their corresponding CV values for traffic density are shown in Table 3. As can be seen, there was considerable variability among the different pollutants with BC having the highest CVs across sites (median = 82, range 48-140), followed by UVPM (median = 81, range 42-102), PM2.5 (median = 78, range 35-94), and UFP (median = 66, range 36-86). Both auto and truck counts showed less variability across sites than pollutant levels across the same sites. Auto traffic count CVs ranged from 11 to 224 (median = 27). Noteworthy in these results was that a high CV value (224) caused by a single-day high auto

1297 34

2494 11

Auto Traffic Counts/hr

20

222

1247 40

*

29861 86

12.0 85

7099 102

7961 110

14

Site 6 Mean CV

30

299

1785 27

*

53898 82

35.4 67

12571 97

17144 121

15

Site 8 Mean CV

B

20 224

**

6687 83

40.0 35

590 101

605 87

5

Site 22 Mean CV

27

170

2327 19

*

37629 69

12.3 92

6037 77

6797 79

12

Site 35 Mean CV

23

362

2577 23

**

51523 68

33.4 63

9061 85

10774 99

15

Site 88 Mean CV

22

517

28

428

2424 12

2.0 27

3.4 46 2608 36

65772 47

56.6 81

13102 101

16951 140

20

Site 500 Mean CV

56683 46

25.3 80

6294 54

6780 54

15

Site 400 Mean CV

C

C

0.5 101

11837 51

23.3 95

950 47

1285 50

17

Site 1000 Mean CV

n represents the total number of measurements taken at each site; this value however may not be the same for all pollutants due to missing measurements from equipment failure or no measurements taken for a specific pollutant. B No truck traffic was observed at this site C No vehicle traffic at this background site * No measurements were taken for this pollutant ** Only one measurement at this site, thus no median and CV data could be calculated

41

*

*

CO

28

54126 64

42567 37

UFP

A

32.4 75

15.9 72

PM2.5

164

5735 42

3261 43

UVPM

129

6253 48

3834 50

Black Carbon

Truck Traffic Counts/hr

10

13

Site 5 Mean CV

NA

Site 2 Mean CV

Table 3. Pollutant and Traffic Variability across Multiple-Measured Sites

51

52 traffic count at site 22, a background location. Variability for truck counts ranged from 20 to 41 (median = 28).

ANALYSES AT SCHOOL-RELATED SITES This section details the results obtained at measured sites in proximity to schools.

Descriptive Data The school-related portion of the study took place from November 16, 2010, through January 26, 2011. Of the 25 sites selected, 18 were in proximity to seventeen elementary schools and one public university (Appendix C, Table 8), five were the higher-traffic reference sites (100, 200, 300, 400 & 500), one (site 2000) was a site located in the Cumbres background zone where both pollutants and traffic were measured, and one was the traffic-less background site (1000) located inside the UABC campus. While all of 25 sites were intended to be measured for BC/UVPM, PM2.5, UFP and CO (see Appendix H, Table 13), the nitrogen oxides (see Appendix F, Tables 11 and 12) were intended to be measured only at the higher-traffic reference sites, as well as the background sites (1000 & 2000). Due to equipment malfunction or human error, there were a number of missing data measurements for BC/UVPM (7), PM2.5 (none), UFP (9), CO (none), and NOx (2). The missing data for NOx species measurements were from sites 100 and 200 where each site was measured only once. From these 25 sites, a total of 66 measurements were taken. The BC median concentration was 3673 ng/m3 with a range of 543-15312 ng/m3 and the UVPM median concentration was 3266 ng/m3 with a range of 399 to 13849 ng/m3. These ranges were lower than those observed across all sites. The PM2.5 median concentration was 32 µg/m3 with a range of 1-85 µg/m3. The UFP median concentration was 28566pt/cc with a range of 7057 to 100545pt/cc. The CO median concentration was 1.7 ppm with a range of 0.1 to 4.5 ppm. In regards to NOx, the median concentration was 272 ppb with a range of 128 to 697 ppb. The NO2 median concentration was 45 ppb with a range of 23 to 97 ppb. Finally, the NO median concentration was 208 ppb with a range of 96 to 610 ppb.

53 Not counting site 1000, which had no vehicular circulation, traffic density was assessed from all other location measurements (n = 50) with no traffic count data losses. The median auto traffic count was 1086 ct/hour with a range of 18 to 3912 ct/hour. The median truck traffic count was 75 ct/hour with a range of 0 to 912 ct/hour.

Pollutant Levels in Relation to Traffic Count The association between pollutants measured at school-related sites and traffic counts is summarized in Table 4. BC was well correlated with auto counts (rho = 0.446, p < 0.01) and with truck traffic counts (rho = 0.522, p < 0.001). PM2.5 in turn, had a moderate correlation with auto traffic counts (rho = 0.396, p < 0.01) but no significant correlation with truck traffic counts (rho = 0.274, not significant). The UFP counts had the strongest correlations with both auto and truck traffic counts (rho = 0.634, p < 0.001 and rho = 0.580, p < 0.001), respectively. Carbon monoxide was moderately correlated with auto traffic counts (rho = 0.350, p < 0.01) and not correlated with truck traffic counts (rho = 0.116, not significant).Of the nitrogen oxides species, only NO2 had a significant correlation, it was correlated with truck traffic counts (rho = 0.693, p < 0.05) but not significantly correlated with auto traffic counts (rho = 0.479, not significant). Table 4. Pollutant versus Traffic Counts Spearman Correlations (School-Related Sites) Automobile Traffic Counts (ct/hour)

Truck Traffic Counts (ct/hour)

All Vehicle Traffic Counts (ct/hour)

Black Carbon 49

0.446**

0.522***

0.454**

UVPM

49

0.452**

0.511***

0.457**

PM2.5

40

0.396**

0.274

0.380**

UFP

43

0.634***

0.580***

0.614***

CO

50

0.350*

0.116

0.330*

NOx

10

0.345

0.474

0.394

NO2

10

0.479

0.693*

0.588

NO

10

0.297

0.432

0.345

n

* p < 0.05 ** p < 0.01 *** p < 0.001

54

Association of Pollutants at School-Related Sites The Spearman correlations between pollutants during at the school-related sites are summarized in Table 5. As expected, BC and UVPM were highly correlated (rho = 0.980, p < 0.001), as they were measured by the same instrument. Ultrafine particle number concentrations had the strongest correlations with other pollutants, BC (rho = 0.834, p < 0.001) and PM2.5 (rho = 0.546, p < 0.001). Additionally, CO was strongly correlated with BC (rho = 0.747, p < 0.001), PM2.5 (rho = 0.505, p < 0.001) and UFP (rho = 0.728, p < 0.001). As seen in these results, PM2.5 was well correlated with other pollutants including BC (rho = 0.665, p < 0.001). On the other hand, the NOX species were not significantly correlated against other pollutants. Amongst each other, strong and significant correlations resulted, especially between the NOX and NO (rho = 0.979, p < 0.001).

POLLUTANT AND TRAFFIC COMPARISONS WITH SOCIO-ECONOMIC STATUS CATEGORIES For this analysis, both pollutant and traffic count data were analyzed using the entire data set for sites measured as well as for the school-related subgroup. The data are summarized in Table 6. The highest BC median concentration (4857 ng/m3) was observed at the overall low income sites (n = 76 for low SES and n = 117 for mid-high SES) followed by a median BC concentration (4670 ng/m3) at the school-related low income sites (n = 23-low and n = 36-Mid/High SES). In contrast, the highest PM2.5 median concentration (41 µg/m3) was observed at school-related mid-to-high income sites (n = 23-low and n = 43-Mid/High SES) followed by a median PM2.5 concentration (28 µg/m3) also at school-related but low income sites. The highest UFP median concentration (32,716pt/cc) was observed at school-related low income sites (n = 22-low and n = 35–Mid/High SES) followed by a median UFP concentration (31,705pt/cc) from the overall mid-to-high income site measurements (n = 67-low and n = 117-Mid/High SES).The highest CO median result (1.9 ppm) was observed at both the low income overall sites (n = 29) and the low income school-related sites (n = 23). Similarly, the same median CO concentration (1.4 ppm), was observed at both the mid-to-high income overall sites (n = 49) and the low income school-related sites (n = 43).

n

rho

n

rho

n

rho

n

rho

n

rho

n

rho

n

rho

0.05 ** p < 0.01 *** p < 0.001

NO2

NOx

CO

UFP

PM2.5

UVPM

Black Carbon 59

0.980***

UVPM

59

0.649***

59

0.665***

PM2.5

57

0.546***

51

0.862***

51

0.834***

UFP

57

0.728***

66

0.505***

59

0.751***

59

0.747***

CO

Table 5. Spearman Correlations between Pollutants (School-Related Sites)

12

0.294

11

-0.009

12

-0.021

10

0.176

10

0.103

NOx

12

0.797**

12

0.466

11

0.218

12

0.252

10

0.515

10

0.479

NO2

12

0.706*

12

0.979***

12

0.161

11

-0.209

12

-0.154

10

0.067

10

-0.006

NO

55

A

912 18-3210 42 0-912 4670 1134-14538 3855 664-13185 28 9-80 32716 11938-100545 1.9F 0.8-4.5 195 150-359 28 23-44 167 127-292

2211A 0-4350 195B 0-750 3673 293-94245 3304 125-51440 26 1-175 31705 1314-140216 1.4 0.1-4.5 286 128-697 57 27-97 216 96-610

1188 0-4602 162 0-912 4857 488-70485 4123 436-53792 24 1-80 28671 2680-160183 1.9E 0.8-4.5 245 120-398 42 23-78 201 93-334

p = < 0.001, B p = 0.019, C p = 0.032, D p = 0.014, E p = 0.009, F p = 0.008

NO (ppb)

NO2 (ppb)

NOx (ppb)

CO (ppm)

UFP (pt/cc)

PM2.5 (µg/m3)

UVPM (ng/m3)

BC (ng/m3)

Truck Traffic Cts

Auto Traffic Cts

Median Range

Median Range

Median Range

Low Income

Mid to High Income

Low Income

School-Related

All

All

School-Related

1758C 480-3912 108D 0-750 3067 543-15213 2466 399-13661 41 1-85 26263 7057-86671 1.4 0.1-4.5 301 128-697 59 27-97 236 96-610

Median Range

Mid to High Income

Table 6. Pollutant and Traffic Relationship with SES Categories (Overall and School-Related Sites)

56

57 Finally, the highest median concentrations for all three NOx species (301 ppb-NOX, 59 ppb-NO2 and 236 ppb-NO) were observed at the mid-to-high income school related sites (n = 4-low and n = 8-Mid/High SES) followed by an NOx median concentration (286 ppb-NOX, 57 ppb-NO2 and 216 ppb-NO) from the overall mid-to-high site measurements (n = 11-low and n = 17-Mid/High SES). Across all sites, the only significant difference of a pollutant by SES categories was that of CO (p = 0.009). Similarly, the only significant pollutant difference between SES categories measured at the school-related sites was that of CO levels (p = 0.008). In terms of traffic density measures, both auto and truck traffic counts were significantly higher in higher income than lower SES areas across all sites (p < 0.001, autos and p = 0.019, trucks). Similarly, both auto and truck traffic counts were significantly higher with higher income SES than lower SES areas at the school-related sites (p = 0.032, autos and p = 0.014, trucks).

COMPARISONS WITH OTHER VARIABLES For the entire data set, we examined the population distributions of independent continuous variables such as the median pollutant concentrations against other independent categorical variables. These categorical variables included, time of day measurement (e.g., mid-day and evening), day of the week, sampling position (e.g., downwind versus upwind), and regional weather. These analyses were also repeated for the school-related data set, with the exception of the time of day measurement since all measures in this portion were taken during mid-day hours. The comparisons of time-of-day measurements showed that during evening hours (1600-2159), pollutant levels were slightly higher than during mid-day hours (1000-1559). However, most of the comparisons of these two categories were not significant, with the exception of UFP levels (n = 184, p = 0.004). The day of the week comparison using the overall data set yielded a few significant pollutant differences (only one less than 0.01) across different days, however, because the sample sizes were too small, these results were not considered valid. Consequently, the analysis for the school-related sites, which were much less in number, was not possible. When pollutant levels were analyzed and compared using the wind direction categorical variable (downwind versus upwind), no significant results were observed across

58 all sites and in the school-related sites data set. Interestingly, the results for both data sets indicated that all pollutant levels were slightly higher at the sites measured upwind from traffic. This is opposite from studies that have found lower pollution levels from traffic under downwind conditions.

59

CHAPTER 5 DISCUSSION This study demonstrated that in the city of Tijuana, just like many other urban areas around the world, vehicular traffic influences the concentrations of certain pollutants in proximity to roadways. In addition, traffic-related pollution was very high in certain areas with levels comparable or higher than studies conducted throughout many urban areas in Europe and the United States. Moreover, the spatial heterogeneity of pollutants observed in Tijuana, validated the notion that vehicular traffic can pose a serious health risk to people living, working, or going to school in proximity to roadways. In order to study the relationship between roadside pollution and traffic volume, BC, UVPM (an indicator of PAHs), PM2.5, UFP, and NOx species (NOx, NO-NO2) were measured at 55 roadside locations, for a total of 210 measurements, scattered over 9 regions of the city. Simultaneously with pollutant measurements, we assessed traffic count and composition at 54 of those locations, for a total of 193 measurements. Furthermore, the roadways assessed included inter-city segments of two federal highways, arterial roads, and local residential roads, all with one to four traffic lanes. Additionally, while in some instances, traffic characteristics were assessed on both directions of a roadway, many of the assessments were only on one side, or instead, at intersections capturing the traffic from two different roadways. Our analyses of pollution levels and traffic volume indicated a strong influence of vehicular traffic on pollutants levels (Table 1). This is in line with studies that indicated that motor vehicle emissions have led to elevated levels of CO (Chen et al., 2009; Lau, Hung, & Cheung, 2009), NOX (Morawska, Jayaratne, Mengersen, Jamriska, & Thomas, 2002; Roorda-Knape et al., 1998; Wang et al., 2008), BC (Patel et al., 2009), and especially UFPs (Buonanno, Lall, & Stabile, 2009).

60

POLLUTION LEVELS NEAR ROADWAYS The pollution levels measured in this study are of particular interest when compared to those found in other studies. The maximum UFP number concentration (160,183 pt/cc, Appendices D, E, Tables 9, 10) measured in this study is higher than the average particle number concentration (150,000 pt/cc) measured by Y. F. Zhu et al. (2002b) from their roadside measurements at Interstate 405 in California, one of the busiest in the world. In addition, our median BC value (4437 ng/m3, Appendix D, Table 9) was comparable to their roadside average result (5400 ng/m3). Contrastingly, our median traffic volume (1980 ct/hr, Appendices D, E, Tables 9, 10) was much lower than theirs (14,000 ct/hr). What this indicates is that with less traffic volume, potential exposure to people in certain areas of Tijuana can be comparable to potential exposure for people spending a significant amount of time in proximity to a major California freeway. Other studies that measured ultrafine particle concentrations provide a reference point as to how Tijuana fares in terms of this pollutant. Our median UFP number concentration and traffic volume results were 30,265 pt/cc and 1980 ct/hr, respectively. Measurements conducted by Wang et al. (2008) at a Corpus Christi highway/avenue intersection yielded UFP number concentration and traffic volume values of 66,000 pt/cc and 5950 ct/hr, respectively. Also, in a study (Morawska et al., 2002) conducted next to a freeway in Brisbane, Australia, yielded results of 8800 pt/cc and 5900 ct/hr for the same parameters. Finally, Levy et al. (2003) measured UFPs at nine locations in a suburb of Boston, their results for the same parameters were 16,000 pt/cc and a traffic volume range up to 1150 ct/hr. To assess potential exposure to school children, data (Appendix H, Table 13) from a subset of 25 sites was analyzed and compared to the results from the overall data set. The median and maximum roadside pollutant concentrations measured at these sites were: 3673 ng/m3 and 15,213 ng/m3 for BC and 28,566 pt/cc and 100,545pt/cc for UFPs. In terms of traffic characteristics, the median and maximum counts for cars were 1086 ct/hr and 3912 ct/hr, respectively, and for trucks were 75 ct/hr and 912 ct/hr, respectively. As can be seen, pollutant concentrations were generally higher across all sites than those from the subset of school data. This is likely because sites selected for the school sub-study were not located in areas, such as the NE, that yielded the highest levels of pollution. Another possibility is that sites from the overall data set were measured over all

61 seasons, and included roads with very high traffic volumes. Thus, pollution levels may have been more prone to the effects of temperature inversions (J. Wallace et al., 2010) occurring at different times of the year. Comparing our school-related pollution results with other studies, demonstrates the potential exposure to traffic-related pollutants in Tijuana’s school children. However, it is important to point out that, in studies that assessed exposure on children, pollution measurements were conducted outside or at the playgrounds of schools. Contrastingly, in this study, measurements at roadways in proximity to the schools were conducted with approximate distances ranging from 30 to 480 meters (Appendix C, Table 8). Thus this study’s comparisons with other studies have to take into consideration the fact that the results may overestimate the children’s actual exposure. In a study conducted in the Bronx, NY, Patel et al. (2009) measured BC levels at five high schools. The first three schools were located within 50 meters of a large urban highway, an urban arterial road, and an urban street, respectively. The last two were located at two suburban background areas. Their BC results, in the order mentioned (2400, 2300, 1400, 660 and 730 ng/m3), were lower than our median BC value (3673 ng/m3) at all school-related roadside locations. Contrastingly, their mean values for car and tuck/bus volumes 6025 ct/hr and 626 ct/hr, respectively), were much higher than those measured in this study (1086 ct/hr and 75 ct/hr).

POLLUTION AND TRAFFIC The Spearman rank correlation results (Table 1) from all valid measurements throughout this study, corroborated the preliminary findings that vehicular traffic influences roadside pollution levels. Specifically, ultrafine particles were well correlated with both auto and truck counts (rho = 0.482, p < 0.001 and rho = 0.418, p < 0.001, respectively). These results are in line with the common knowledge that vehicular emissions are the main sources of sub-micrometer particles in urban areas (Hitchins et al., 2000; Morawska, Thomas, Gilbert, Greenaway, & Rijnders, 1999), as shown in a many studies. As expected, BC was well correlated with truck traffic counts (rho = 0.477, p < 0.001) and also, but to a lesser extent, with auto traffic counts (rho = 0.377, p < 0.001). These results support the findings by Y. F. Zhu et al. (2002a, 2002b) that BC levels were the

62 highest at the closest measurement points at two California freeways, especially the one with a higher percentage of heavy-duty traffic. Contrastingly, Westerdahl et al. (2005), from their mobile measurements at three segments of a Los Angeles freeway, found that BC concentrations sharply increased only as diesel truck traffic increased. In this study, PM2.5 was not correlated with either auto traffic counts (rho = 0.075, not correlated) or truck counts (rho = 0.026, not correlated). These results support the role of PM2.5 as a regional pollutant (Brook et al., 2007). In two exposure assessment studies (J. J. Kim et al., 2004; Van Roosbroeck et al., 2007), PM2.5 variability of measurements across distinct school locations was not statistically different, similar to our findings. In this study, CO was correlated with auto traffic counts (rho = 0.348, p < 0.01) but not correlated with track traffic count. These results are in line with the common knowledge that, due their leaner fuel-to-air ratios, diesel engines emit less CO in comparison to gasoline engines (Kittelson 1998). Conversely, mobile measurements of CO conducted by Westerdahl et al. (2005), showed that CO concentrations were correlated with truck and auto traffic alike. Nitrogen dioxide (NO2) in this study had the strongest correlation with auto traffic (rho = 0.677, p < 0.001) and to a lesser extent, with truck traffic (rho = 0.510, p < 0.01). These results are in line with those obtained by Morawska et al. (1999) where they concluded that a 50% increase in traffic flow, translated into a similar increase in NO2 concentrations. Similarly, Janssen et al. (2001) from their exposure assessment study concluded that NO2, as measured at sites in proximity to major roadways, can be used as direct measure of long-term exposure. The Spearman rank correlation results for the school-related subset in this study (Table 4) were generally more highly correlated for certain pollutants. Notable from these results is that PM2.5 was moderately correlated with auto traffic counts (rho = 0.396, p < 0.01). The reason for this is unknown but it is possible that transformation of ultrafine particles into the PM2.5 accumulation mode were occurring much faster around these sites. However, this is usually associated with diesel emissions (Lena et al., 2002), something not seen in these results. Another notable difference between correlation results in the subset and the entire data set was for NO2. As mentioned mention before, results from the entire data set showed that NO2 was correlated with vehicular traffic, especially automobiles. In

63 contrast, results from the subset yielded no correlation at all. The explanation for this was that in the school-related subset, only 10 measurements took place, thus making correlations difficult to detect.

RELATING TRAFFIC POLLUTION TO TRAFFIC FLOW Several studies have shown that vehicle emissions rates depend on the types of road intersections, vehicles, as well as on traffic characteristics such as driving modes (Pandian et al., 2009). To explore this topic, traffic flow as surrogate for different types of driving modes, such as acceleration, deceleration, idle and cruising was utilized. Thus in order to reflect these modes, the all-vehicle traffic counts were stratified into three traffic flow categories: low or stop-and-go, such as that seen at a vehicular border crossing; moderate or intermittent, associated with traffic signals or pedestrian crossings; and high or continuous, such as that seen in freeways or arterial roads without traffic lights. The initial analysis (Figure 5) in which pollution levels in each of these three categories were assessed, showed that ultrafine particle number concentrations were the highest (30,131 pt/cc) at the moderate/intermittent traffic flow category in comparison to the low/stop-and go (9,991 pt/cc) and high/continuous (27,786 pt/cc) traffic flow categories but there was no significant difference between the two. However, after restricting the all-vehicle traffic counts to a more stable, or less-variable, range (1500-3000 ct/hr) and re-calculating the data, the results this time confirmed significantly higher UFP levels during intermittent driving conditions than levels during continuous driving conditions (58,227 pt/cc vs. 27,496 pt/cc, p < 0.05). These results are in line with many studies that confirmed that during delay periods, vehicle emission rates are higher in comparison to when the vehicle is continuously in motion (Pandian et al., 2009).

CORRELATIONS BETWEEN POLLUTANTS In general, from the overall data set, the inter-correlations among UFPs, BC and CO were strong (UFP/BC, rho = 0.636, p < 0.001; UFP/CO, rho = 0.734, p < 0.001; and BC/CO, rho = 0.743, p < 0.001, Table 2). In comparison, the results from the school-related subset yielded slightly higher correlations, especially for UFPs (UFP/BC, rho = 0.834, p < 0.001; UFP/CO, rho = 0.728, p < 0.001; and BC/CO, rho = 0.747, p < 0.001, Table 5). This was expected based on the pollutants’ higher correlations with traffic counts for the school-related

64 sites. In general, these results confirm that vehicle emissions are the major source of these pollutants (Y. F. Zhu et al., 2002b). Additionally, for the entire dataset, the associations of PM2.5 with these pollutants were lower in magnitude (PM2.5/BC, rho = 0.272, p < 0.01; PM2.5/UFP, rho = 0.357, p < 0.001; & PM2.5/CO, rho = 0.505, p < 0.001). This indicates the role of PM2.5 as a regional pollutant , despite evidence of some influence of vehicle emissions, especially from diesel engines that may contribute to local variability of PM2.5 (Patel et al., 2009), as seen in the results. None of the NOx species were significantly correlated with the rest of the pollutants, except among each other, possibly due to low sample size. Of the pollutants, the correlation was highest between NO2 and CO, though not significant (Table 2). The association among BC, UFPs and CO was observed by Y. F. Zhu et al. (2002a, 2002b). In both of these studies, all three pollutants had the highest levels at the respective sites closest to the roadways, and then very similar decay curves as measurements were taken at increasing distances from the roadways. These results, which can be linked to our inter-correlation results, indicate that UFPs, BC and CO are good indicators of traffic-related pollution exposure. Finally, since CO had very strong correlations with both UFPs and BC, especially from the results across all measurements, it may indicate that CO could be used as a stand-alone surrogate measure for traffic related exposure. Inexpensive and easy-to-use personal monitors exist that could be used for traffic and health-related and exposure-related studies in the city of Tijuana.

TEMPORAL AND SPATIAL VARIABILITY Over the course of the different months in which measurements were conducted across Tijuana, it was observed that ultrafine particle number concentrations were slightly higher in the fall and winter seasons (Appendix G, Figures 9-11). This is in contrast with a Southern California study (Kuhn et al., 2005), that found that particle number counts, in the both the accumulation (PM2.5) and the ultrafine (UFP) modes, were similar during the winter and summer seasons. Contrastingly, they showed diurnal variability of measurements taken during daytime and evening hours, with higher particle number concentrations measured in the evenings. This is in line with our small number of measurements taken in the evenings

65 that yielded higher particle number concentrations than the ones measured during daytime hours. The temporal variability analysis (Table 3) from select sites measured multiple times during this study showed that pollutant levels varied temporally over the course of the study. BC showed that highest variability (CV range = 48-140) followed by PM2.5 (CV range = 35-94) and ending with UFPs (CV range = 36-86) showing the least variability of this group. Variability in traffic counts for both autos (CV range = 11-40) and trucks (CV range = 20-41) was relatively low in comparison to pollutant variability. Some of the variability likely arose from traffic variability between measurement points but the higher variability observed in the pollutants was probably caused by other factors specific to the region, such as changing wind speeds, inversion layers, or substantial differences in engines fuel-to-air ratios, which were no investigated in this study. Spatial variability across the zones defined in this study was assessed using all measurements taken throughout the study period. Median BC concentrations showed spatial gradients (1075 ng/m3-8253 ng/m3) with the highest value observed in the SE zone. The measurements taken at this zone included two segments of a federal highway, a heavy-duty truck route. Contrastingly, median PM2.5 concentrations showed little variability across zones (17 µg/m3-56 µg/m3), thus once again, supporting the knowledge that particulate matter the accumulation mode has a longer life time in the environment and can come from differ sources (NARSTO, 2004). Median UFP number concentrations (9,738 pt/cc-47,101 pt/cc, Figure 7), CO (0.3 ppm-3.2 ppm) and NO2 (27-92 ppb), all showed spatial variability across zones with the highest values observed in the 5 y 10 sector. This area features one of the busiest intersections of Tijuana characterized by high auto and truck traffic. For all of these pollutants, with the exception of PM2.5, measurements at the Cumbres-background zone and the other sites (2 and 1000) designated as background locations, median values were the lowest in the city. These results corroborate the knowledge that spatial variations in pollutant concentrations can be explained by vehicular traffic (Weijers et al., 2004).

66

RELATING POLLUTION TO SES AND OTHER CATEGORIES Factors such as SES that often go unmeasured may lead to substantial confounding in results (Holguin et al., 2007). Thus it is necessary to include or control for socioeconomic status in an investigation. As seen in Table 6, BC and CO concentrations were the highest at low SES sites for both the overall data set (4857 ng/m3 and 1.9 ppm) and the school-related subset (4670 ng/m3 and 1.9 ppm). Contrastingly, NO2 levels were the highest at the mid-to-high SES sites for both data sets (57 ppb and 59 ppb, respectively). Ultrafine particle number concentrations were the highest at the mid-to high SES sites from the entire data set, and highest at the low SES sites in the school-related subset. For traffic count measurements, the highest counts were observed at the mid-to-high SES sites for both data sets. Carbon monoxide measurements were significantly higher in low SES sites than higher SES sites as obtained from the analyses from the overall and school related data sets (p = 0.009 and 0.008, respectively). Contrastingly, none of the other pollutants were significantly different as measured at sites with different SES categories. This indicates that the SES of areas in Tijuana appears to influence CO levels, but only slightly, and did not significantly influence other pollutants. This is in contrast to studies in the US that found that people living in low income areas may experience higher levels of vehicle emissions (Gunier, Hertz, Von Behren, & Reynolds, 2003; Korc, 1996). Thus, this result requires further investigation. It is important to mention that for the school-related measurements, with the exception of one (near site 117), the rest were public institutions, and thus this was a way of controlling for possible confounding effects of SES at these sites. In general, the results in this study are opposite of those many investigations that have found SES to be a better predictor of exposure. In a California study that investigated the frequency of asthma symptoms in low-income individuals, Meng, Wilhelm, Rull, English, and Ritz (2007), found that poorer individuals were twice as likely to develop asthma symptoms, and that traffic effects were stronger for asthmatics living in poverty. In the US, many people live in areas where their health is at risk due to environmental risks such as proximity to major roadways, thus explaining the above results. However, in

67 many parts of Europe, more affluent and educated people tend to live closer to roadways thus the relationship between SES and health outcomes is reversed. For example, an Italian study (Cesaroni et al., 2010) that categorized Rome residents in terms of their exposure to traffic pollutants, found that people living in mid-to-high SES areas or with the university degree were more likely to be exposed to traffic than people in low SES areas or without a degree. However, they found an opposite result for people living in the center of Rome. The results in this study indicate that traffic-related air pollution presents a potential hazard in multiple areas of the city. However, since the categorization of SES areas was based on observation and extrapolation from government information of marginalized zones, further investigation is needed to verify the results.

LIMITATIONS There were a number of limitations in the design of this study. For instance, the site selection did not rely on background information on the traffic and road characteristics of the roadways where measurements were conducted. Thus, the sites may not have been representative of all types of roadways in Tijuana. Similarly, there was no prior knowledge of the schools, and the roadways associated with them. Thus, the schools chosen may not have been representative of all elementary schools in Tijuana. Nonetheless, the results indicate that the areas in Tijuana with the highest levels of traffic-related pollution, as well as areas with little or no potential exposure to this pollution were captured. Another limitation in the study design is that, due to time and resource limitations, a vehicle fleet classification beyond differentiating between autos and trucks could not be done. In other words, a more sophisticated assessment, such as Gross Vehicle Weight Rating (GVWR) employed in emissions inventories, was not possible. Additionally, for simplicity, all small trucks, vans and SUVs were counted and included into the truck category; even if they might have been gasoline-powered. Thus, it is possible that pollution levels might have been slightly inaccurate due to an overestimation of trucks and underestimation of automobiles. Nevertheless, many exposure assessment studies have used a simple vehicle fleet categorization method and obtained valid results. Due to practical considerations, the measurements were limited to 20 minute windows at each site. This potentially decreased the reliability of the results, thus not

68 reflecting true conditions at the sites. However, we made every effort to reduce this issue by ensuring consistency in our sampling methodology so the any deficiencies would be uniformly present throughout the study. Measurements at a total of 55 sites across Tijuana were conducted. While this was good in terms of assessing the spatial characteristics of traffic-related pollution, it limited the number of measurements at each site. In fact, aside from the subset of sites used for the temporal analysis, most sites were only measured once or twice, including the 18 sites chosen in proximity to schools. This resulted in the analyses having lower statistical power. Finally, since very little significance was found in the analyses of the relationships between pollution levels and categorical variables, such as sampling position (for wind speed and direction effects) and socio-economic status, the methodology for these parameters may have been too simplistic. However, considering the nature of our study, these results, and all results in this study, can serve as reference point for future studies of this nature in the city of Tijuana.

RECOMMENDATIONS This is the first study in the city of Tijuana to evaluate roadside pollution levels as a function of traffic volume. Most of the knowledge of traffic-related pollution in the city comes from fixed-site monitors, or from emissions inventories that provide comprehensive but generalized information about the air quality conditions, as well as the various sources of pollutants. Thus the health impact of traffic-related pollution on people, especially vulnerable subpopulations, living, working or attending schools in close proximity to roadways, is not being assessed. It is recommended that this study be used as a pilot or guideline for future studies to measure the relationship between pollution levels and traffic volume utilizing more robust methodology. Because air pollution is not only influenced by vehicular traffic, other factors such as the climate of the region, conditions of the vehicular fleet, roadway characteristics, and traffic patterns need to be included as part of those investigations. In addition, areas of potential high exposure to auto and truck emissions were identified in this study (Appendix I, Figures 12-17). The areas surrounding the 5 y 10 intersection and the Otay Mesa border crossing, and the southeast region of Tijuana, are the

69 places of residence and employment of Tijuana’s working class. Not coincidentally, these areas are also the places with the highest volumes of autos and trucks, including commercial routes. Thus it is of utmost importance to perform studies at these specific areas. In this study pollution and traffic volume measurements at roadside locations in proximity to schools were conducted. While some of those schools were within 50 meters of from the measurement locations, many were located beyond 150 meters, a distance where studies have shown particulate matter pollution from the road significantly decays. Thus the measurements may not have captured actual potential exposures in children. Therefore, future studies of this nature need to be conducted at school locations, preferably school playgrounds where children have and increased risk of potential exposure to traffic pollution. In this study we provide a listing of schools, some of which are in proximity to vehicular traffic hot spots, where future studies can be conducted. The results of this study can also be used as a guideline for traffic pollution exposure assessment studies that, to the author’s best knowledge, are non-existent or few in Tijuana. One area of research in particular that could be investigated is personal exposure in relation to birth outcomes. This is a new area of investigation for traffic-related pollution exposure that, because of the implications to both mother and child, should be considered. Moreover, the results may be used as a reference for ongoing or future programs sponsored by local and state agencies or those dictated by the Border 2012 initiative. One in particular, is an emissions vehicular verification program (Periódico Oficial del Estado de Baja California, 2010) similar to California’s smog check program. Finally, studies like this one can serve as the foundation for policy and legislation to protect the health of Tijuana’s residents. With additional research, justification of anti-idling measures or zoning initiatives to restrict sitting of schools within a certain distance to major roadways, similar to the 500 feet setback bill introduced in California, can be obtained. Additionally, studies like this can be used to validate the need to re-route commercial trucking away from many Tijuana neighborhoods, including those where the working class reside.

70

CHAPTER 6 CONCLUSIONS In conclusion, levels of the traffic-related pollutants BC and ultrafine particles reached high levels in certain areas of Tijuana with high traffic density. Levels of PM2.5 and carbon monoxide were generally low. Traffic counts were most highly correlated with ultrafine particle concentrations and this was true for both auto counts and truck counts. BC concentrations were also correlated but to a lesser extent with both auto and truck counts and with a tendency to be more highly correlated with truck counts. In these results, the pollutants ultrafine particles, BC and carbon monoxide were all highly inter-correlated, indicating that these pollutants are good indicators of near-traffic exposure. In particular, the high correlation of carbon monoxide levels with the other pollutants suggests that measurements of this pollutant may be a surrogate for near–traffic exposures in future studies. Inexpensive and light weight measurement devices are available for carbon monoxide and could be used for personal exposure measurements and epidemiological studies. The pollutant PM2.5 was the least correlated with other pollutants and with traffic counts, emphasizing the regional nature of this regulated pollutant. Traffic flow appeared to affect roadside concentrations of ultrafine particles, with the highest levels observed during moderate and intermittent traffic flow conditions, rather than stopped or continuous flow. This finding is in line with a many studies that have identified different driving modes, such as acceleration, as pollutant emitters. Although in general, traffic-related pollution levels measured near schools in this study were lower than those at the busiest roadways, these levels present a target for exposure reduction to this vulnerable population. Children can be exposed while on the playground and inside the classrooms. Potential avenues for exposure reduction include anti-idling measures near schools and re-routing of traffic away from schools, especially heavy-duty trucks. Additional, future zoning ordinances can locate schools in areas of low traffic, such as the 500 feet buffer zone in the State of California.

71 Although significant spatial variability of pollution levels was observed across Tijuana, socio-economic status of the neighborhoods did not appear to influence pollution levels, with the possible exception of carbon monoxide. Other studies have reported a significant association between SES and traffic-related air pollution in communities. The results presented here indicate rather that traffic-related air pollution presents a potential hazard in multiple areas of the city. Taken as a whole, the findings of this study indicate that the population of Tijuana could experience significant exposures to traffic-related pollutants with associated public health risks. The city of Tijuana has recently undertaken a new program to link certification of vehicle emissions to car registration similar to the California vehicle smog check program. The findings of this study present a potential baseline measurement to assess the effectiveness of this vehicle certification program through additional measurements. The measurements presented in this study adds to the knowledge of environmental health conditions along the US-Mexico border by detailing near road air pollutant levels associated with traffic characteristics in the city of Tijuana. Due to the large and growing population of this region, reduction of traffic-related pollution should be a public health priority.

72

REFERENCES Abu-Allaban, M., Gillies, J. A., Gertler, A. W., Clayton, R., & Proffitt, D. (2003). Tailpipe, resuspended road dust, and brake-wear emission factors from on-road vehicles. Atmospheric Environment, 37(37), 5283-5293. Air Matrix Corporation. (2005). MiniVol TAS operation manual. Retrieved from htpp://www.airmetrics.com/downloads/MiniVol_TAS_Manual_Web.pdf Allen, A. G., McGonigle, A. J. S., Cardoso, A. A., Machado, C. M. D., Davison, B., Paterlini, W. C., . . . de Mellos Dias, C. (2009). Influence of sources and meteorology on surface concentrations of gases and aerosols in a coastal industrial complex. Journal of the Brazilian Chemical Society, 20(2), 214-221. Anderson, J. B., & Gerber, J. (2008). Fifty years of change on the U.S.-Mexico Border: Growth, development, and quality of life. Austin, TX: University of Texas Press. Ayuntamiento de Tijuana. (2011). Historia minima de Tijuana. Retrieved from http://www.tijuana.gob.mx/ciudad/CiudadHistoriaMinima.asp Baek, S. O., Field, R. A., Goldstone, M. E., Kirk, P. W., Lester, J. N., & Perry, R. (1991). A review of atmospheric polycyclic aromatic hydrocarbons-sources, fate and behavior. Water Air and Soil Pollution, 60(3-4), 279-300. Baldauf, R., Thoma, E., Hays, M., Shores, R., Kinsey, J., Gullett, B., . . . Bang, J. (2008). Traffic and meteorological impacts on near-road air quality: Summary of methods and trends from the raleigh near-road study. Journal of the Air & Waste Management Association, 58(7), 865-878. Ballester, F., Estarlich, M., Iniguez, C., Llop, S., Ramon, R., Esplugues, A., . . . Rebagliato M. (2010). Air pollution exposure during pregnancy and reduced birth size: A prospective birth cohort study in Valencia, Spain. Environmental Health, 9. doi:10.1186/1476-069X-9-6 Brauer, M., Lencar, C., Tamburic, L., Koehoorn, M., Demers, P., & Karr, C. (2008). A cohort study of traffic-related air pollution impacts on birth outcomes. Environmental Health Perspectives, 116(5), 680-686. Brook, J. R., Poirot, R. L., Dann, T. F., Lee, P. K. H., Lillyman, C. D., & Ip, T. (2007). Assessing sources of PM2.5 in cities influenced by regional transport. Journal of Toxicology and Environmental Health-Part a-Current Issues, 70(3-4), 191-199. Brunekreef, B., Janssen, N. A. H., deHartog, J., Harssema, H., Knape, M., & vanVliet, P. (1997). Air pollution from truck traffic and lung function in children living near motorways. Epidemiology, 8(3), 298-303. Buonanno, G., Lall, A. A., & Stabile, L. (2009). Temporal size distribution and concentration of particles near a major highway. Atmospheric Environment, 43(5), 1100-1105.

73 Cesaroni, G., Badaloni, C., Romano, V., Donato, E., Perucci, C. A., & Forastiere, F. (2010). Socioeconomic position and health status of people who live near busy roads: The Rome Longitudinal Study (RoLS). Environmental Health, 9, 1-12. Chakrabarti, B., Fine, P. M., Delfino, R., & Sioutas, C. (2004). Performance evaluation of the active-flow personal DataRAM PM2.5 mass monitor (Thermo Anderson pDR-1200) designed for continuous personal exposure measurements. Atmospheric Environment, 38(20), 3329-3340. Chan, C. C., Chuang, K. J., Shiao, G. M., & Lin, L. Y. (2004). Personal exposure to submicrometer particles and heart rate variability in human subjects. Environmental Health Perspectives, 112(10), 1063-1067. Chen, J. H., Wang, W., Zhang, J. Y., Liu, H. J., Ren, L. H., Liu, X. Y., . . . Wang, X. (2009). Characteristics of gaseous pollutants near a main traffic line in Beijing and its influencing factors. Atmospheric Research, 94(3), 470-480. Church, D. F., & Pryor, W. A. (1985). Free-radical chemistry of cigattete smoke and its toxicological implications. Environmental Health Perspectives, 64, 111-126. Colvile, R. N., Hutchinson, E. J., Mindell, J. S., & Warren, R. F. (2001). The transport sector as a source of air pollution. Atmospheric Environment, 35(9), 1537-1565. Crouse, D. L., Goldberg, M. S., Ross, N. A., Chen, H., & Labreche, F. (2010). Postmenopausal breast cancer is associated with exposure to traffic-related air pollution in Montreal, Canada: A case-control study. Environmental Health Perspectives, 118(11), 1578-1583. Dales, R., Wheeler, A., Mahmud, M., Frescura, A. M., Smith-Doiron, M., Nethery, E., . . . Liu, L. (2008). The influence of living near roadways on spirometry and exhaled nitric oxide in elementary schoolchildren. Environmental Health Perspectives, 116(10), 1423-1427. Delfino, R., Tjoa T., Gillen, D., Staimer N., Polidori, A., Arhami, M., Jamner, L., Sioutas, C., & Longhurst, J. (2010). Traffic-realted air pollution and blood pressure in elderly subjects with coronary artery disease. Epidemiology, 21, 396-1360. Diario Oficial de la Federación. (2003). Acuerdo por el se emiten y publican las reglas de operación del programa Hábitat, para el ejercicio fiscal 2003. Retrieved from http://www.elocal.gob.mx/work/sites/ELOCAL/resources/LocalContent/2986/1/habitat.pdf Dreher, D. B., & Harley, R. A. (1998). A fuel-based inventory for heavy-duty diesel truck emissions. Journal of the Air & Waste Management Association, 48(4), 352-358. Dutkiewicz, V. A., Alvi, S., Ghauri, B. M., Choudhary, M. I., & Husain, L. (2009). Black carbon aerosols in urban air in South Asia. Atmospheric Environment, 43(10), 1737-1744. Environmental Protection Agency, (EPA). (2000). Program to improve the air quality in Tijuana-Rosarito 2000-2005. Retrieved from http://www.epa.gov/usmexicoborder/infrastructure/tijuana-airplans/tijuanatosaritoeng .pdf

74 Environmental Protection Agency, (EPA). (2011a). The ambient air monitoring program. Retrieved from http://www.epa.gov/air/oaqps/qa/monprog.html Environmental Protection Agency, (EPA). (2011b). What are the six common pollutants? Carbon monoxide. Retrieved from http://www.epa.gov/airquality/carbonmonoxide Escamilla-Nunez, M. C., Barraza-Villarreal, A., Hernandez-Cadena, L., Moreno-Macias, H., Ramirez-Aguilar, M., Sienra-Monge, J. J., . . . del Rio-Navarro B. (2008). Traffic-related air pollution and respiratory symptoms among asthmatic children, resident in Mexico City: The EVA cohort study. Respiratory Research, 116, 832-838. Estrella, B., Estrella, R., Oviedo, J., Narvaez, X., Reyes, M. T., Gutierrez, M., & Naumova, E. N. (2005). Acute respiratory diseases and carboxyhemoglobin status in school children of Quito, Ecuador. Environmental Health Perspectives, 113(5), 607-611. Ezzati, M., Lopez, A. D., Rodgers, A., Vander Hoorn, S., Murray, C. J. L., & Comparative Risk Assessment. (2002). Selected major risk factors and global and regional burden of disease. Lancet, 360(9343), 1347-1360. Faiz, A., Gautam, S., & Burki, E. (1995). Air-pollution from motor-vehicles: Issues and options for Latinamerican countries. Science of the Total Environment, 169(1-3), 303-310. Ganster, P. (1996). Environmental issues of the California-Baja California border region border environment research reports no. 1. Retrieved from http://www.scerp.org/docs/berr1.html Garshick, E., Laden, F., Hart, J. E., & Caron, A. (2003). Residence near a major road and respiratory symptoms in US veterans. Epidemiology, 14(6), 728-736. Gauderman, W. J., Vora, H., McConnell, R., Berhane, K., Gilliland, F., Thomas, D., . . . Peters, J. (2007). Effect of exposure to traffic on lung development from 10 to 18 years of age: A cohort study. Lancet, 369(9561), 571-577. Gehring, U., Wijga, A. H., Brauer, M., Fischer, P., de Jongste, J. C., Kerkhof, M., . . . Brunekreef, B. (2010). Traffic-related air pollution and the development of asthma and allergies during the first 8 years of life. American Journal of Respiratory and Critical Care Medicine, 181(6), 596-603. Gilbert, A. (1994). The Latin American city (2nd ed.). London, UK: Latin American Bureau. Gilliland, F. D. (2009). Outdoor air pollution, genetic susceptibility, and asthma management: Opportunities for intervention to reduce the burden of asthma. Pediatrics, 123, S168-S173. Godish, T. (2004). Air quality (4th ed.). New York, NY: Lewis Publishers. Green, D. C., Fuller, G. W., & Baker, T. (2009). Development and validation of the volatile correction model for PM10: An empirical method for adjusting TEOM measurements for their loss of volatile particulate matter. Atmospheric Environment, 43(13), 2132-2141.

75 Green, R. S., Malig, B., Windham, G. C., Fenster, L., Ostro, B., & Swan, S. (2009). Residential exposure to traffic and spontaneous abortion. Environmental Health Perspectives, 117(12), 1939-1944. Gunier, R. B., Hertz, A., Von Behren, J., & Reynolds, P. (2003). Traffic density in California: Socioeconomic and ethnic differences among potentially exposed children. Journal of Exposure Analysis and Environmental Epidemiology, 13(3), 240-246. Gutierrez-Daban, A., Fernandez-Espinosa, A. J., Ternero-Rodriguez, M., & Fernandez-Alvarez, F. (2005). Particle-size distribution of polycyclic aromatic hydrocarbons in urban air in southern Spain. Analytical and Bioanalytical Chemistry, 381(3), 721-736. Hagler, G. S. W., Baldauf, R. W., Thoma, E. D., Long, T. R., Snow, R. F., Kinsey, J. S., . . . Gullett, B. K. (2009). Ultrafine particles near a major roadway in Raleigh, North Carolina: Downwind attenuation and correlation with traffic-related pollutants. Atmospheric Environment, 43(6), 1229-1234. Hansen, A. D. A., Rosen, H., & Novakov, T. (1984). The aethalometer an instrument for the real-time measurment of optical absortion by aerosol particles. Science of the Total Environment, 36(JUN), 191-196. Harris, S. J., & Maricq, M. M. (2001). Signature size distributions for diesel and gasoline engine exhaust particulate matter. Journal of Aerosol Science, 32(6), 749-764. Healey, K., Lingard, J. J. N., Tomlin, A., Hughes, A., White, K. L. M., Wild, C. P., & Routledge, M. N. (2005). Genotoxicity of size-fractionated samples of urban particulate matter. Environmental and Molecular Mutagenesis, 45(4), 380-387. Hitchins, J., Morawska, L., Wolff, R., & Gilbert, D. (2000). Concentrations of submicrometre particles from vehicle emissions near a major road. Atmospheric Environment, 34, 51-59. Hitzenberger, R. (1993). Absorbtion-measurments with an integrating plate phtomoeter calibration and error analysis. Aerosol Science and Technology, 18, 70-84. Hitzenberger, R., Petzold, A., Bauer, H., Ctyroky, P., Pouresmaeil, P., Laskus, L., & Puxbaum, H. (2006). Intercomparison of thermal and optical measurement methods for elemental carbon and black carbon at an urban location. Environmental Science & Technology, 40(20), 6377-6383. Hoek, G., Brunekreef, B., Goldbohm, S., Fischer, P., & van den Brandt, P. A. (2002). Association between mortality and indicators of traffic-related air pollution in the Netherlands: A cohort study. Lancet, 360(9341), 1203-1209. Holguin, F., Flores, S., Ross, Z., Cortez, M., Molina, M., Molina, L., . . . Romieu, I. (2007). Traffic-related exposures, airway function, inflammation, and respiratory symptoms in children. American Journal of Respiratory and Critical Care Medicine, 176(12), 1236-1242.

76 Huang, C. H., & Tai, C. Y. (2008). Relative humidity effect on PM2.5 readings recorded by collocated beta attenuation monitors. Environmental Engineering Science, 25(7), 1079-1089. Huang, Y. L., & Batterman, S. (2000). Residence location as a measure of environmental exposure: a review of air pollution epidemiology studies. Journal of Exposure Analysis and Environmental Epidemiology, 10, 66-85. Imhof, D., Weingartner, E., Prevot, A. S. H., Ordonez, C., Kurtenbach, R., Wiesen, P., . . . Baltensperger, U. (2006). Aerosol and NOx emission factors and submicron particle number size distributions in two road tunnels with different traffic regimes. Atmospheric Chemistry and Physics, 6, 2215-2230. Instituto Nacional de Estadistica y Geografia (INEGI). (2011). Mexico en cifras, Tijuana B.C. Retrieved from http://www.inegi.org.mx/sistemas/mexicocifras/default.aspx?ent=02 Intergovernmental Panel on Climate Change. (2007). Climate change 2007 synthesis report (AR4). Retrieved from http://www.ipcc.ch/publications_and_data/publications_ipcc_fourth_assessment_repo rt_synthesis_report.htm Janssen, N. A. H., van Vliet, P. H. N., Aarts, F., Harssema, H., & Brunekreef, B. (2001). Assessment of exposure to traffic related air pollution of children attending schools near motorways. Atmospheric Environment, 35(22), 3875-3884. Jayaratne, E. R., Wang, L., Heuff, D., Morawska, L., & Ferreira, L. (2009). Increase in particle number emissions from motor vehicles due to interruption of steady traffic flow. Transportation Research Part D-Transport and Environment, 14(7), 521-526. Karr, C. J., Rudra, C. B., Miller, K. A., Gould, T. R., Larson, T., Sathyanarayana, S., & Koenig, J. Q. (2009). Infant exposure to fine particulate matter and traffic and risk of hospitalization for RSV bronchiolitis in a region with lower ambient air pollution. Environmental Research, 109(3), 321-327. Kean, A. J., Harley, R. A., & Kendall, G. R. (2003). Effects of vehicle speed and engine load on motor vehicle emissions. Environmental Science & Technology, 37(17), 3739-3746. Kim, J. J., Smorodinsky, S., Lipsett, M., Singer, B. C., Hodgson, A. T., & Ostro, B. (2004). Traffic-related air pollution near busy roads: The East Bay children's respiratory health study. American Journal of Respiratory and Critical Care Medicine, 170(5), 520-526. Kim, K. H., Sekiguchi, K., Kudo, S., & Sakamoto, K. (2011). Characteristics of atmospheric elemental carbon (char and soot) in ultrafine and fine particles in a roadside environment, Japan. Aerosol and Air Quality Research, 11, 1-12. Kirchstetter, T. W., Singer, B. C., Harley, R. A., Kendall, G. R., & Traverse, M. (1999). Impact of California reformulated gasoline on motor vehicle emissions. Environmental Science & Technology, 33(2), 318-328.

77 Kittelson, D. B. (1998). Engines and nanoparticles: A review. Journal of Aerosol Science, 29(5-6), 575-588 Kittelson, D. B., Watts, W. F., Johnson, J. P., Remerowki, M. L., Ische, E. E., Oberdorster, G., . . . Zhou, L. (2004). On-road exposure to highway aerosols. Inhalation Toxicology, 16, 31-39. Korc, M. E. (1996). A socioeconomic assessment of human exposure to ozone in the South Coast Air Basin of California. Journal of the Air & Waste Management Association, 46(6), 547-557. Krudysz, M., Moore, K., Geller, M., Sioutas, C., & Froines, J. (2009). Intra-community spatial variability of particulate matter size distributions in Southern California/Los Angeles. Atmospheric Chemistry and Physics, 9(3), 1061-1075. Kuhn, T., Biswas, S., Fine, P. M., Geller, M., & Sioutas, C. (2005). Physical and chemical characteristics and volatility of PM in the proximity of a light-duty vehicle freeway. Aerosol Science and Technology, 39(4), 347-357. Kumar, P., Fennell, P., Langley, D., & Britter, R. (2008). Pseudo-simultaneous measurements for the vertical variation of coarse, fine and ultrafine particles in an urban street canyon. Atmospheric Environment, 42(18), 4304-4319. Lam, G. C. K., Leung, D. Y. C., Niewiadomski, M., Pang, S. W., Lee, A. W. F., & Louie, P. K. K. (1999). Street-level concentrations of nitrogen dioxide and suspended particulate matter in Hong Kong. Atmospheric Environment, 33, 1-11. Lau, J., Hung, W. T., & Cheung, C. S. (2009). Interpretation of air quality in relation to monitoring station's surroundings. Atmospheric Environment, 43(4), 769-777. Lena, T. S., Ochieng, V., Carter, M., Holguin-Veras, J., & Kinney, P. L. (2002). Elemental carbon and PM2.5 levels in an urban community heavily impacted by truck traffic. Environmental Health Perspectives, 110(10), 1009-1015. Levy, J. I., Bennett, D. H., Melly, S. J., & Spengler, J. D. (2003). Influence of traffic patterns on particulate matter and polycyclic aromatic hydrocarbon concentrations in Roxbury, Massachusetts. Journal of Exposure Analysis and Environmental Epidemiology, 13(5), 364-371. Liu, L. J. S., Slaughter, J. C., & Larson, T. V. (2002). Comparison of light scattering devices and impactors for particulate measurements in indoor, outdoor, and personal environments. Environmental Science & Technology, 36(13), 2977-2986. Llop, S., Ballester, F., Estarlich, M., Esplugues, A., Rebagliato, M., & Iniguez, C. (2010). Preterm birth and exposure to air pollutants during pregnancy. Environmental Research, 110(8), 778-785. Lloyd, A. C., & Cackette, T. A. (2001). Diesel engines: Environmental impact and control. Journal of the Air & Waste Management Association, 51(6), 809-847. Lyons, T. J., Kenworthy, J. R., Moy, C., & dos Santos, F. (2003). An international urban air pollution model for the transportation sector. Transportation Research Part D-Transport and Environment, 8(3), 159-167.

78 Magee Scientific. (2005). The aethalometer. Retrieved from http://mageesci.com/support/downloads/Aethalometer_book_2005.07.03.pdf Matson, U., Ekberg, L. E., & Afshari, A. (2004). Measurement of ultrafine particles: A comparison of two handheld condensation particle counters. Aerosol Science and Technology, 38(5), 487-495. Maynard, D., Coull, B. A., Gryparis, A., & Schwartz, J. (2007). Mortality risk associated with short-term exposure to traffic particles and sulfates. Environmental Health Perspectives, 115(5), 751-755. McConnell, R., Islam, T., Shankardass, K., Jerrett, M., Lurmann, F., Gilliland, F., . . . Berhane, K. (2010). Childhood incident asthma and traffic-related air pollution at home and school. Environmental Health Perspectives, 118(7), 1021-1026. Meng, Y. Y., Wilhelm, M., Rull, R. P., English, P., & Ritz, B. (2007). Traffic and outdoor air pollution levels near residences and poorly controlled asthma in adults. Annals of Allergy Asthma & Immunology, 98(5), 455-463. Morawska, L., Jayaratne, E. R., Mengersen, K., Jamriska, M., & Thomas, S. (2002). Differences in airborne particle and gaseous concentrations in urban air between weekdays and weekends. Atmospheric Environment, 36(27), 4375-4383. Morawska, L., Thomas, S., Bofinger, N., Wainwright, D., & Neale, D. (1998). Comprehensive characterization of aerosols in a subtropical urban atmosphere: Particle size distribution and correlation with gaseous pollutants. Atmospheric Environment, 32(14-15), 2467-2478. Morawska, L., Thomas, S., Gilbert, D., Greenaway, C., & Rijnders, E. (1999). A study of the horizontal and vertical profile of submicrometer particles in relation to a busy road. Atmospheric Environment, 33(8), 1261-1274. Motelay-Massei, A., Harner, T., Shoeib, M., Diamond, M., Stern, G., & Rosenberg, B. (2005). Using passive air samplers to assess urban-rural trends for persistent organic pollutants and polycyclic aromatic hydrocarbons. Environmental Science & Technology, 39(15), 5763-5773. Mukerjee, S., Smith, L. A., Johnson, M. M., Neas, L. M., & Stallings, C. A. (2009). Spatial analysis and land use regression of VOCs and NO2 from school-based urban air monitoring in Detroit/Dearborn, USA. Science of the Total Environment, 407(16), 4642-4651. North American Research Strategy for Tropospheric Ozone, (NARSTO). (2004). Particulate matter science for policy makers, a NARSTO assessment. Cambridge, MA: Cambridge University Press. Oberdorster, G. (2010). Safety assessment for nanotechnology and nanomedicine: concepts of nanotoxicology. Journal of Internal Medicine, 267, 89-105. Ohnishi, S., & Kawanishi, S. (2002). Double base lesions of DNA by a metabolite of carcinogenic benzo a pyrene. Biochemical and Biophysical Research Communications, 290(2), 778-782.

79 Ohura, T., Noda, T., Amagai, T., & Fusaya, M. (2005). Prediction of personal exposure to PM2.5 and carcinogenic polycyclic aromatic hydrocarbons by their concentrations in residential microenvironments. Environmental Science & Technology, 39(15), 5592-5599. Ortmeyer, T. H., & Pillay, P. (2001). Trends in transportation sector technology energy use and greenhouse gas emissions. Proceedings of the IEEE, 89(12), 1837-1847. Owaga & Company. (2006). NO, NO2, NOx and SO2 sampling protocol using the Owaga sampler. Retrieved from http://www.ogawausa.com/pdfs/prono-noxno2so206.pdf Pan-American Health Organization, (PAHO). (2004). Virtual library of sustainable development and environmental health. Retrieved from http://www.bvsde.paho.org/bvsacd/cd27/honing.pdf Pandian, S., Gokhale, S., & Ghoshal, A. K. (2009). Evaluating effects of traffic and vehicle characteristics on vehicular emissions near traffic intersections. Transportation Research Part D-Transport and Environment, 14(3), 180-196. Parker, J. D., Woodruff, T. J., Basu, R., & Schoendorf, K. C. (2005). Air pollution and birth weight among term infants in California. Pediatrics, 115, 121-128. Patel, M. M., Chillrud, S. N., Correa, J. C., Feinberg, M., Hazi, Y., Deepti, K. C., . . . Kenney, P. L. (2009). Spatial and temporal variations in traffic-related particulate matter at New York City high schools. Atmospheric Environment, 43(32), 4975-4981. Peden, D. B. (2005). The epidemiology and genetics of asthma risk associated with air pollution. Journal of Allergy and Clinical Immunology, 115(2), 213-219 Periódico Oficial del Estado de Baja California. (2010, January 5). Número especial. H. XIX legislatura constitucional del estado de Baja California. Decreto no. 341. Retrieved from http://www.bajacalifornia.gob.mx/portal/gobierno/legislacion/periodico/2010/INDIC E-05-01-2010.pdf Polidori, A., Hu, S., Biswas, S., Delfino, R. J., & Sioutas, C. (2008). Real-time characterization of particle-bound polycyclic aromatic hydrocarbons in ambient aerosols and from motor-vehicle exhaust. Atmospheric Chemistry and Physics, 8(5), 1277-1291. Pope, C. A., Burnett, R. T., Thun, M. J., Calle, E. E., Krewski, D., Ito, K., & Thurston, G. D. (2002). Lung cancer, cardiopulmonary mortality, and long-term exposure to fine particulate air pollution. Journal of the American Medical Association, 287(9), 1132-1141. Quintana, P. J. E., Samimi, B. S., Kleinman, M. T., Liu, L. J., Soto, K., Warner, G. Y., & Bufalino, C. (2000). Evaluation of a real-time passive personal particle monitor in fixed site residential indoor and ambient measurements. Journal of Exposure Analysis and Environmental Epidemiology, 10(5), 437-445. Quintero Nunez, M., Reyna, M., Collins, K., Guzman, S., Powers, B., Mendoza, A. (2006). Issues related to air quality and health in the California-Baja California border region.

80 In R. Pumfrey (Ed.), SCERP monograph series #14 (vol. 14). San Diego, CA: San Diego State University Press. Reponen, T., Grinshpun, S. A., Trakumas, S., Martuzevicius, D., Wang, Z. M., LeMasters, G., . . . Biswas, P. (2003). Concentration gradient patterns of aerosol particles near interstate highways in the Greater Cincinnati airshed. Journal of Environmental Monitoring, 5(4), 557-562. Ristovski, Z. D., Morawska, L., Bofinger, N. D., & Hitchins, J. (1998). Submicrometer and supermicrometer particulate emission from spark ignition vehicles. Environmental Science & Technology, 32(24), 3845-3852. Ritz, B., & Wilhelm, M. (2006). Evidence for adverse reproductive effects of exposure to traffic-related air pollution: Are pregnant women at risk? Epidemiology, 17(6), S57-S57. Ritz, B., & Yu, F. (1999). The effect of ambient carbon monoxide on low birth weight among children born in southern California between 1989 and 1993. Environmental Health Perspectives, 107, 17-25. Roorda-Knape, M. C., Janssen, N. A. H., de Hartog, J. J., van Vliet, P. H. N., Harssema, H., & Brunekreef, B. (1998). Air pollution from traffic in city districts near major motorways. Atmospheric Environment, 32(11), 1921-1930. Salma, I., Balashazy, I., Winkler-Heil, R., Hofmann, W., & Zaray, G. (2002). Effect of particle mass size distribution on the deposition of aerosols in the human respiratory system. Journal of Aerosol Science, 33, 119-132. Samet, J. M. (2006). Lung cancer and inhaled carcinogens: What do we know and what do we need to know? Toxicologic Pathology, 34(7), 981-981. Sather, M. E., Slonecker, E. T., Mathew, J., Daughtrey, H., & Williams, D. D. (2007). Evaluation of Ogawa passive sampling devices as an alternative measurement method for the nitrogen dioxide annual standard in El Paso, Texas. Environmental Monitoring and Assessment, 124(1-3), 211-221. Schwartz, J. (1994). Air pollution and daily mortality: A review and meta-analysis. Environmental Research, 64, 36-52. Schwartz, J., Sarnat, J., & Coull, B. (2003). Estimating the effect of measurement error in pollution exposure on the results of regressions using ambient pollutants. Epidemiology, 14(5), S42-S42. Sjodin, A., & Lenner, M. (1995). On road-measurements of single vehicle pollutant emissions, speed and acceleration for large fleets of vehicles in different traffic environments. Science of the Total Environment, 169(1-3), 157-165. Singer, B. C., Hodgson, A. T., Hotchi, T., & Kim, J. J. (2004). Passive measurement of nitrogen oxides to assess traffic-related pollutant exposure for the East Bay Children's Respiratory Health Study. Atmospheric Environment, 38(3), 393-403. Singh, M., Bowers, K., & Sioutas, C. (2006). Seasonal and spatial trends in particle number concentrations and size distributions at the children's health study sites in Southern California. Journal of Exposure Science and Environmental Epidemiology, 16, 3-18.

81 Snyder, D. C., & Schauer, J. J. (2007). An inter-comparison of two black carbon aerosol instruments and a semi-continuous elemental carbon instrument in the urban environment. Aerosol Science and Technology, 41(5), 463-474. Sweedler, A., Fertig, M., Collins, K., & Quintero Nunez. M. (2003). Air quality in the California-Baja California border region. In A. Sweedler (Ed.), SCERP monograph series #6. San Diego, CA: San Diego State University Press. Tonne, C., Melly, S., Mittleman, M., Coull, B., Goldberg, R., & Schwartz, J. (2007). A case-control analysis of exposure to traffic and acute myocardial infarction. Environmental Health Perspectives, 115, 53-57. Valavanidis, A., Fiotakis, K., & Vlachogianni, T. (2008). Airborne particulate matter and human health: Toxicological assessment and importance of size and composition of particles for oxidative damage and carcinogenic mechanisms. Journal of Environmental Science and Health Part C-Environmental Carcinogenesis & Ecotoxicology Reviews, 26(4), 339-362. Valavanidis, A., Salika, A., & Theodoropoulou, A. (2000). Generation of hydroxyl radicals by urban suspended particulate air matter. The role of iron ions. Atmospheric Environment, 34(15), 2379-2386. Vallejo, M., Lerma, C., Infante, O., Hermosillo, A. G., Riojas-Rodriguez, H., & Cardenas, M. (2004). Personal exposure to particulate matter less than 2.5 mu m in Mexico City: A pilot study. Journal of Exposure Analysis and Environmental Epidemiology, 14(4), 323-329. Van Roosbroeck, S., Jacobs, J., Janssen, N. A. H., Oldenwening, M., Hoek, G., & Brunekreef, B. (2007). Long-term personal exposure to PM2.5, soot and NOx in children attending schools located near busy roads, a validation study. Atmospheric Environment, 41(16), 3381-3394. Venkatachari, P., Zhou, L. M., Hopke, P. K., Schwab, J. J., Demerjian, K. L., Weimer, S., et al (2006). An intercomparison of measurement methods for carbonaceous aerosol in the ambient air in New York City. Aerosol Science and Technology, 40(10), 788-795. Vineis, P., Hoek, G., Krzyzanowski, M., Vigna-Taglianti, F., Veglia, F., Airoldi, L., . . . Riboli, E. (2007). Lung cancers attributable to environmental tobacco smoke and air pollution in non-smokers in different European countries: a prospective study. Environmental Health, 6. Wallace, J., Corr, D., & Kanaroglou, P. (2010). Topographic and spatial impacts of temperature inversions on air quality using mobile air pollution surveys. Science of the Total Environment, 408(21), 5086-5096. Wallace, L. A., Mitchell, H., O'Connor, G. T., Neas, L., Li ppmann, M., Kattan, M., . . . Liu, L. J. (2003). Particle concentrations in inner-city homes of children with asthma: The effect of smoking, cooking, and outdoor pollution. Environmental Health Perspectives, 111(9), 1265-1272. Wang, Y. G., Zhu, Y. F., Salinas, R., Ramirez, D., Karnae, S., & John, K. (2008). Roadside measurements of ultrafine particles at a busy urban intersection. Journal of the Air & Waste Management Association, 58(11), 1449-1457.

82 Weijers, E. P., Khlystov, A. Y., Kos, G. P. A., & Erisman, J. W. (2004). Variability of particulate matter concentrations along roads and motorways determined by a moving measurement unit. Atmospheric Environment, 38(19), 2993-3002. Westerdahl, D., Fruin, S., Sax, T., Fine, P. M., & Sioutas, C. (2005). Mobile platform measurements of ultrafine particles and associated pollutant concentrations on freeways and residential streets in Los Angeles. Atmospheric Environment, 39(20), 3597-3610. Wong, Y. C., Sin, D. W. M., & Yeung, L. L. (2002). Assessment of the air quality in indoor car parks. Indoor and Built Environment, 11(3), 134-145. World Health Organization (WHO). (2002). The World Health Report 2002: Reducing risks, promoting healthy life. Retrieved from http://www.who.int/whr/2002/en/whr02_en.pdf World Health Organization (WHO). (2009). Global health risks: Mortality and burden of disease attributable to selected major risks. Retrieved from http://www.who.int/healthinfo/global_burden_disease/GlobalHealthRisks_report_full .pdf Wu, C. F., Delfino, R. J., Floro, J. N., Samimi, B. S., Quintana, P. J. E., Kleinman, M. T., & Lui, L. J. (2005). Evaluation and quality control of personal nephelometers in indoor, outdoor and personal environments. Journal of Exposure Analysis and Environmental Epidemiology, 15, 99-110. Wu, J., Ren, C. Z., Delfino, R. J., Chung, J., Wilhelm, M., & Ritz, B. (2009). Association between local traffic-generated air pollution and preeclampsia and preterm delivery in the south coast air basin of California. Environmental Health Perspectives, 117(11), 1773-1779. Wu, Y., Hao, J. M., Fu, L. X., Wang, Z. S., & Tang, U. (2002). Vertical and horizontal profiles of airborne particulate matter near major roads in Macao, China. Atmospheric Environment, 36(31), 4907-4918. Yadav, V. K., Prasad, S., Patel, D. K., Khan, A. H., Tripathi, M., & Shukla, Y. (2010). Identification of polycyclic aromatic hydrocarbons in unleaded petrol and diesel exhaust emission. Environmental Monitoring and Assessment, 168(1-4), 173-178 Yorifuji, T., Naruse, H., Kashima, S., Ohki, S., Murakoshi, T., Takao, S., . . . Doi, H. (2011). Residential proximity to major roads and preterm births. Epidemiology, 22, 74-80. Zhao, L. R., Wang, X. M., He, Q. S., Wang, H., Sheng, G. Y., Chan, L. Y., . . . Blake, D. R. (2004). Exposure to hazardous volatile organic compounds, PM10 and CO while walking along streets in urban Guangzhou, China. Atmospheric Environment, 38(36), 6177-6184. Zhu, K., Zhang, J. F., & Lioy, P. J. (2007). Evaluation and comparison of continuous fine particulate matter monitors for measurement of ambient aerosols. Journal of the Air & Waste Management Association, 57(12), 1499-1506.

83 Zhu, Y. F., Hinds, W. C., Kim, S., Shen, S., & Sioutas, C. (2002a). Study of ultrafine particles near a major highway with heavy-duty diesel traffic Atmospheric Environment, 36(2002), 4323-4335. Zhu, Y. F., Hinds, W. C., Kim, S., & Sioutas, C. (2002b). Concentration and size distribution of ultrafine particles near a major highway. Journal of the Air & Waste Management Association, 52(9), 1032-1042.

84

APPENDIX A STUDY SITE MAP

85

Figure 8. Study site map.

86

APPENDIX B STUDY SITE LIST AND SITE DESCRIPTION

City Sector

SYBC SYBC

5 y 10 NE

SE

NE

SW

SYBC SYBC SYBC SW Cumbres SYBC

NW-Playas OMBC NW-Playas NW-Playas

OMBC

NE OMBC

Site ID

2 3

5 6

8

9

12

13 14 15 16 17 19

22 23 24 26

27

28 35

N 32° 31' 7.2" W 116° 53' 21.4" N 32° 31' 47.5" O 116° 57' 54.2"

N 32° 32' 10.5" W 116° 54' 12.5"

N 32° 30' 47.9" O 117° 7' 22.3" N 32° 29' 20.2" W 116° 58' 51.8" N 32° 31' 59.5" W 117° 5' 24.3" N 32° 32' 5.3" W 117° 6' 21.7 "

N 32° 32' 25.31" W 116° 59' 49.95" N 32° 32' 28.3" W 117° 2' 27.5" N 32° 30' 47.7" W 117° 4' 14.9" N 32° 28' 34.3" W 117° 0.1' 37.9" N 32° 29' 16.9" W 117° 5' 25.9" N 32° 32' 1.09" W 117° 1' 7.05"

N 32° 28' 1.4" W 117° 1' 3.9"

N 32° 32' 32.4" W 116° 52' 42.2"

N 32° 27' 30.8" O 116° 50' 9.1"

N 32° 31' 1.4" W 116° 57' 41.9" N 32° 32' 11.6" O 116° 51' 25.6"

N 32° 32' 8.8" W 117° 1' 54.5” N 32° 3' 48.1" W 117° 1' 29.6"

Coordinates

Location

Mexico Federal 2 Highway (TecateTijuana) Mexico Federal 1 Highway (EnsenadaTijuana) Blvd. Cuauhtemoc Norte Via Internacional & Niños Heroes Ave. Libramento Sur Colinas St. & Punta Banderas St. Monte Illimani St. & Monte Elias St. Paseo Centernario Tijuana & Independencia St. Del Pacifico St. Aeropuerto Blvd. & Vicente Suarez St. El Mirador Blvd. & De la Bufadora St. Mexico Federal 1 Highway (TijuanaEnsenada) Hector Teran Teran Blvd. & 25 de Septiembre St. Turqueza Blvd. & Datil St. del Tecnologico Ave.

Corredor Tijuana 2000

Benito Juárez or Calle 2da Paseo de los Heroes Ave. & Francisco Javier Mina St. Aeropuerto Blvd. Corredor Tijuana 2000

Table 7. Study Site List and Site Description

City intersection In front of UABC campus

South of Manuel Capetillo St. City intersection On ramp (Veracruz St.) City intersection (roundabout) City intersection City intersection at Miguel Hidalgo Roundabout South of Parque Baja California Sur St. City intersection City intersection Off ramp leading to Playas de Tijuana Blvd City intersection

Off ramp leading to TJ bus station South of toll road to Mexico Federal 2D Highway North of Mexico Federal 2 Highway (Tecate-TJ) Site located just west of Benito Juarez Street Off ramp leading to Libramento Sur

One-way arterial leading to SYBC City intersection

Description

Low Mid-High

Low

Mid-High Low Mid-High Low

Low Low Low Low Low Mid-High

Low

Low

Low

Low Low

Mid-High Mid-High

SES level

(table continues)

L/L M/L

M/L

L/L M/M L/L H/L

L/L H/L M/L L/L L/L M/L

L/L

L/M

M/M

L/L L/M

Auto/Truck Traf. Density Category M/L M/L

87

SW OMBC OMBC

NWPlayas NWPlayas SW

36 37 40

41

SYBC SYBC SYBC SYBC

105 106

107 108 109 110

103

102

101

SE NE SYBC SE NWPlayas NWPlayas NWPlayas NWPlayas SYBC SYBC

48 49 51 88 100

47

46

City Sector

Site ID

N 32° 32’ 31.1” W 117° 1’ 13” N 32° 31’ 2.7” W 117° 1’ 25.8” N 32° 30’ 22.2” W 117° 2’ 9.3” N 32° 31’ 35.1” W 117° 2’ 34.9”

N 32° 32’ 22” W 116° 59’ 49” N 32° 32’ 28.8” W 117° 0’ 29.7”

N 32°31’ 23.2” W 117° 6’ 34.3”

N 32°31’ 53.9” W 117° 5’ 22.9”

N 32°31’ 37.7” W 117° 6’ 42.5”

N 32° 27' 39.2" W 116° 49' 25.4" N 32° 31' 26.1" W 116° 53' 35.8" N 32°32’ 32.0” W 117° 1’ 38.8” N 32° 29' 8.7" W 116° 51' 2.1" N 32° 29' 8.7" W 116° 51' 2

N 32° 29' 02" W 117° 0' 4.2"

N 32° 31' 36.5" W 117° 6' 34.4"

N 32° 30' 23.3" W 117° 7' 7.8"

N 32° 28' 41.3" W 116° 59' 33.7" N 32° 33' 5.9" W 116° 55' 5.6" N 32° 31' 58.86" W 116° 57' 2.2"

Coordinates

Table 7. (continued)

Cuauhtemoc Norte Blvd.& 19th St. Doceava or Felix Gomez St. & Garcia Naranjo St. Jose Maria Pino Suarez St. & 4th St. Juan de Dios Peza St. & Jalisco St. Los Fundadores Blvd. 9th or Ignacio ZaragozaSt. & Matamoros St.

Mirador Blvd. & Bahia de Sto. Domingo St. del Agua St. & de la Piedra St.

Bolaños Cacho St. & del Doctor Auvanel Vallejo St. Paseo de los Laureles Blvd. Roble St. Rampa Xicontencatl St. Corredor Tijuana 2000 Mexico Federal 1 Highway (EnsenadaTijuana) Paseo Pedregal Blvd.

Parque Mexico Sur St.

de las Penas St. & de la Luz St.

Aeropuerto Blvd. Via Internacional & Tercera St. del Tecnologico Ave.

Location

City Intesection (School-related) City Intesection (School-related) North of Rampa Sayula St. (School-related) City Intesection (School-related)

City Intesection (School-related) City Intesection (School-related)

City Intesection (School-related)

Blvd. and adjacent service street (Schoolrelated) City Intesection (School-related)

Infront of Tijuana Technological University West of Turqueza Blvd. End of street east of SENTRI lane of SYBC At entrance to Real de SF Community Traffic Reference Site (School-related)

City intersection

West of Mexico Federal 1 Highway1

West of Pacifico St. City intersection At interchange w/ Lazaro Cardenas & Via Juventud City intersection

Description

Low Mid-High Low Mid-High

Low Low

Mid-High

Mid-High

Mid-High

Mid-High Mid-High Low Mid-High Low

Low

Mid-High

Mid-High

Low Mid-High Mid-High

SES level

(table continues)

L/L L/L L/L L/L

M/L L/L

L/L

L/L

M/L

L/L L/L L/L M/M M/L

L/L

L/L

L/L

Auto/Trck Traf. Density Category H/M L/M H/M

88

Mid-City

Mid-City

Mid-City 5 y 10 5 y 10 OMBC

OMBC OMBC OMBC SYBC

Mid-City

5 y 10

OMBC

Bckgnd

Cumbres

111

112

113 114 115 116

117 118 119 200

300

400

500

1000

2000

N 32° 29’ 17.8” W 117° 5’ 29.1”

N 32° 31’ 56.9” W 116° 57’ 57.5”

N 32° 32’ 36” W 116° 56’ 19.6”

N 32°30´5.1’’ W 116° 57´53.3´´

N 32° 31’ 29” W 117° 0’ 33.5”

N 32° 32’ 48.3” W 116° 55’ 37” N 32° 31’ 15.8” W 116° 55’ 30.5” N 32° 31’ 28.6” W 116° 56’ 53.7” N 32° 32’ 20.6” W 117° 1’ 27.5”

N 32° 30’ 38.6” W 116° 59’ 2.3” N 32°29´32.1’’ W 116° 57´18.7´´ N 32°29´35.6’’ W 116° 58´17.5´´ N 32° 32’ 12.7” W 116° 55’ 41.2”

N 32° 31’ 41.9” W 116° 59’ 35.7”

N 32° 31’ 30.8” W 116° 59’ 57.7”

Coordinates

Monte Everest Ave.

Inside UABC campus

De las Bellas Artes Blvd.

Via Rapida Oriente Freeway & Calle Jose Vasconselos St. Lazaro Cardenas Blvd.

Jose Clemente Orozco St. & Diego Rivera St. PrimeraSt. & Carlos Bawver St. Aeropuerto Blvd. & Las Plazas St. Alfonso Reyes St.

Defensores de Baja California Blvd. & Angela Peralta St. el Tecnologico / Guadalupe Ramirez / Popotla or Primera Sur Almacen-Paseo de los Heroes Blvd Gustavo Diaz Ordaz Blvd. & Guadalajara St. Aeropuerto Blvd. & Durango St. Eje 2 Oriente Blvd.

Location

Three-way intersection (Schoolrelated) East of Las Rosas St. (School-related) City Intesection (School-related) City Intesection (School-related) Blvd. Lanes run under Highway 2 (School-related) City Intesection (School-related) City Intesection (School-related) City Intesection (School-related) East of Linea Internacional Blvd. (Ref. Site schools) Traffic Reference Site (Schoolrelated) At “5 y 10” sector (Traffic Ref. Site for schools) Near Otay Mesa BC (Traffic Ref. site for schools) Anchor Background Site/No traffic (school-related) Background school-related site

City Intesection (School-related)

Description

Not applicable L/L

M/M

M /H

H/L

L/L L/L M/M L/L

M/L M/M M/H L/L

L/L

Auto/Trck Traf. Density Category L/L

Low

Mid-High

Mid-High

Mid-High

Mid-High

Low Low Mid-High Mid-High

Mid-High Mid-High Low Mid-High

Mid-High

Mid-High

SES level

CITY SECTOR LEGEND: NW-Playas sector (Zone 1); San Isydro Border Crossing (SYBC) sector (Zone 2); Mid-City sector (Zone 3); “5 y 10” sector (Zone 4); Otay Mesa Border Crossing (OMBC) sector (Zone 5); Cumbres /Background sector (Zone 6); SW sector (Zone 7); NE sector (Zone 8); and SE sector (Zone 9) TRAFFIC DENSITY CATEGORIES: For Autos: Low (< 1330 vph); Moderate (1330-3331vph); and High (> 3331vph); For Trucks: Low (< 200vph); Moderate (200-465vph); and High (> 465vph)

City Sector

Site ID

Table 7. (continued)

89

90

APPENDIX C SCHOOLS LOCATIONS AND THEIR DESCRIPTIONS IN RELATION TO MEASUREMENT AND TRAFFIC REFERENCE SITES

Primaria Municipal Club Soroptimista

Escuela Primaria 12 de Octubre

Escuela Primaria General Lazaro Cardenas (Vesp.)

Escuela Primaria Don Andres Quintana Roo

Escuela Primaria Heroes de Granaditas

Escuela Primaria Fernando Montes De Oca

Escuela Primaria Francisco I. Madero

Escuela Primaria Alfonso Reyes

Escuela Primaria General Ignacio Zaragoza

Primaria Municipal Emma A. de Bustamante

Escuela Primaria Las Americas

Escuela Primaria Francisco Javier Mina

Escuela Primaria Valentín Gómez Farías

School Name / Location Parque México #1124 Deleg. Playas de Tijuana 32°31'38.17"N 117° 6'41.54"W Paseo Costa Del Pacifico #725 Deleg. Playas de Tijuana 32°31'54.10"N 117° 5'35.79"W Avenida Del Agua #1241 Deleg. Playas de Tijuana 32°31'22.00"N 117° 6'34.59"W Av. Félix Parra # 12951 Deleg. Mesa de Otay 32°32'20.13"N 116°59'53.64"W Calle 12 y M. Doria #11889 Deleg. Mesa de Otay 32°32'28.26"N 117° 0'33.26"W 4ta y Cañón Otay # 136 Deleg. Zona Centro 32°32'32.32"N 117° 1'10.31"W Avenida Fresnillo # 2782 Deleg. Zona Centro 32°31'2.61"N 117° 1'32.21"W Ave. Tepatitlán #44 Deleg. San Antonio de los Buenos 32°30'23.29"N 117° 2'12.62"W Calle Matamoros # 1750 Deleg. Zona Centro 32°31'33.63"N 117° 2'33.18"W Alberto Einstein S/N Deleg. Otay Mesa 32°31'21.69"N 117° 0'1.56"W del Tec y Guadalupe Ramírez Deleg. Mesa de Otay 32°31'44.07"N 116°59'35.57"W Callejon del Olvido # 204 Deleg. La Mesa 32°30'43.88"N 116°59'5.03"W Av. Guadalajara # 99 Deleg. La Mesa 32°29'23.57"N 116°57'24.58"W

Location and Coordinates

114

113

112

111

110

109

108

107

106

105

103

102

101

310

190

60

300

75

110

170

90

100

130

30

330

30

400

300

300

300

200

200

200

200

200

200

100

100

100

Mid-High

Mid-High

Mid-High

Mid-High

Mid-High

Low

Mid-High

Low

Low

Low

Mid-High

Mid-High

Mid-High

SES LEVEL

(table continues)

1500

2100

1600

900

2250

3850

2400

500

1500

2600

1700

800

1400

In proximity to In proximity to Approx. Traffic Approx. Site ID distance Ref site distance (m) ID (m)

Table 8. Schools Locations and Their Descriptions in Relation to Measurement and Traffic Reference Sites

91

Escuela Primaria Valentín Gómez Farías

Escuela Primaria Nicolas Bravo

Colegio Latinoamerica Iberoamericano (Kinder/Primaria)

Instituto Tecnológico de Tijuana

Primaria Municipal Club de Leones

School Name / Location

Table 8. (continued)

Av. Sonora S/N Deleg. La Mesa 32°29'31.32"N 116°58'12.42"W CalzadaTecnológico S/N Deleg. Centenario 32°32'9.08"N 116°55'29.09"W José Clemente Orozco # 22 Deleg. Centenario 32°32'48.43"N 116°55'36.46"W Venustiano Carranza y Murua Deleg. Centenario 32°31'21.02"N 116°55'20.22"W Avenida De Las Plazas S/N Deleg. Centenario 32°31'29.43"N 116°56'35.59"W

Location and Coordinates

119

118

117

116

115

480

300

25

140

200

500

500

500

500

400

2100

2800

1200

1600

1150

Mid-High

Low

Low

Mid-High

Low

In proximity to In proximity to SES Approx. Traffic Approx. Site ID distance Ref site distance LEVEL (m) ID (m)

92

93

APPENDIX D DESCRIPTIVE RESULTS AT MULTIPLE-MEASUREMENT SITES

Site ID

22

100

101

102

103

2

51

105

106

107

108

109

Zone

Playas

Playas

Playas

Playas

Playas

SYBC

SYBC

SYBC

SYBC

SYBC

SYBC

SYBC

2

2

2

2

2

2

13

2

2

2

2

5

n

440 293-1531 5507 3960-7054 1750 3054-2146 3919 3014-4823 2138 1102-3173 3371 1515-8270 3471 2587-4355 5272 3311-7232 4433 1134-7552 4396 4122-4670 6931 1468-12393 9233 3928-14538

Black Carbon (ng/m3) Median Range 388 125-1632 4935 4043-5827 1525 1245-1804 3167 2150-4184 2130 1099-3161 3127 1669-6260 2991 2108-3873 4921 3528-6314 4197 1122-7273 3675 3580-3769 5909 1145-10673 8726 3603-13849

Median Range

UVPM (ng/m3)

40 30-50 50.5 35-66 32 15-49 57.5 30-85 35.5 23-48 17 3-32 26 23-29 21 15-27 27.5 10-45 21.5 19-24 46.5 22-71 54 28-80

Median Range

PM2.5 (µg/m3)

Table 9. Descriptive Results at Multiple-Measurement Sites

*

43659 29663-57655 23665 11938-35391 25052 22618-27485 *

1894 1314-10625 70179 63495-76863 21108 14337-27879 38658 31455-45860 14397 11918-16876 41218 12930-59666 *

Median Range

UFP (pt/cc)

1.1 1.0-1.2 2.0 1.6-2.4 2.0 1.8-2.2 1.7 0.9-2.4 2.5 1.0-4.0 **

2.3 1.7-2.8 1.9 1.8-2.0 1.9 0.1-6.6 1.7 1.4-1.9 **

**

Median Range

CO (ppm)

0 0-102 3180 3150-3210 1605 1518-1692 915 798-1032 600 492-708 2520 1818-2862 30 30-60 1857 1692-2022 105 90-120 180 150-210 519 480-558 936 912-960

Traffic Density Autos (ct/hr) Median Range

(table continues)

126 120-132 12 0-60 6 0-12 9 0-18 42 0-60

78 0-240 6 0-12 51 42-60 15 0-30 120 60-192 ***

***

Traffic Density Trucks (ct/hr) Median Range

94

112

113

300

5

114

115

400

35

116

117

MidCity

MidCity

MidCity

5 y 10

5 y 10

5 y 10

OMBC

OMBC

OMBC

2

111

MidCity

MidCity

3

200

SYBC

2

3

12

15

2

2

10

3

2

3

3

110

SYBC

n

Site ID

Zone

Table 9. (continued)

3746 1843-4437 3642 3007-7776 3087 2727-3436 2225 1682-2433 9443 3673-15213 4201 2792-5610 6283 1579-12153 11317 11288-11346 9850 7784-11915 7021 1701-12841 5936 753-17831 8750 3121-11671 3062 1264-4861

Black Carbon (ng/m3) Median Range 3705 1586-4533 3068 2494-7451 2644 2594-2694 1622 1536-2485 8187 2713-13661 3792 2637-4947 6114 1381-9944 10007 9823-10191 8647 6767-10526 6282 1097-11798 5744 685-14721 10002 3170-10718 2597 1338-3856

Median Range

UVPM (ng/m3)

26 1-29 27 1-36 34 10-58 44 13-61 62.5 54-71 65 13-71 32 10-74 46 41-51 48 47-49 14 1-50 14 1-32 34 10-58 24 9-39

Median Range

PM2.5 (µg/m3)

52518 51025-54010 45348 16121-120431 76168 73821-78515 67923 35300-100545 46665 8081-86671 27041 13290-71110 37235 32713-55938 17847 12591-23103

*

16634 9991-23276

23944 16502-31385 29661 28566-53836 *

Median Range

UFP (pt/cc)

1.7 1.4-2.0 1.3 0.8-1.8

**

3.2 2.8-3.5 2.8 2.0-3.5 3.4 2.3-4.5

**

1.8 1.6-1.9 2.55 1.4-3.7 1.1 0.7-1.4 1.1 0-4.3 3.1 1.8-4.3 2.9 2.2-3.5

Median Range

CO (ppm)

1158 648-1188 843 828-858 702 30-948 2184 2088-2280 3570 3150-3912 1209 660-1842 2091 2022-2160 2715 2472-2958 2988 1158-4350 2478 1746-3120 612 558-960 111 72-150

Traffic Density Autos (ct/hr) Median Range

(table continues)

***

1110 648-1188 60 12-78 84 60-108 48 0-78 87 72-102 132 90-132 168 60-282 210 198-222 780 648-912 522 372-750 180 60-216 138 132-222

Traffic Density Trucks (ct/hr) Median Range

95

119

500

2000

1000

6

8

88

OMBC

OMBC

Cumbre s

Bkgrou nd

NE

SE

SE

15

15

14

17

3

20

2

2

n Median Range 4928 4063-5793 3762 3266-4258 8796 2834-51440 664 462-2927 1015 399-1844 4186 436-25127 9488 3416-53792 7837 2576-31757

5498 4982-6814 4786 3936-5635 9429 3210-94245 1134 494-3268 1401 543-2636 5159 488-32433 9201 4145-70485 7825 2895-44329

UVPM (ng/m3)

Black Carbon (ng/m3) Median Range

10.5 1-28 40.5 1-64 25 12-79

49 34-64 33.5 9-58 44.5 11-175 15 11-24 14 1-69

Median Range

PM2.5 (µg/m3)

17811 2680-67488 45955 6517-160183 34010 4671-96608

38186 30131-46241 56160 51716-60603 54903 18263-140216 13136 6820-44015 10417 4835-27975

Median Range

UFP (pt/cc)

**

**

**

1.5 1.2-1.7 1.4 1.3-1.4 1.8 1.5-2.9 2.0 1.9-2.0 0.3 0.1-1.6

Median Range

CO (ppm)

* Instrument failure during at one or both of the sampling sessions for this site, not median value or range could not be calculated ** Parameter not measured at this site *** Zero traffic was observed at this site during sampling sessions

118

OMBC

Footnotes:

Site ID

Zone

Table 9. (continued)

990 840-2100 1590 1302-2640 2580 1698-3480

N/A

780 480-1140 1770 1758-1782 2520 1890-2820 72 18-138

Traffic Density Autos (ct/hr) Median Range

228 138-312 309 180-420 366 240-492

N/A

***

420 228-690

180 132-228 240 192-288

Traffic Density Trucks (ct/hr) Median Range

96

97

APPENDIX E DESCRIPTIVE RESULTS AT SINGLE-MEASUREMENT SITES

16

36

47

9

28

49

48

SW

SW

SW

NE

NE

NE

SE

2248

1732

3409

3501

6840

1467

8844

1366

1300

4884

2175

3208

1962

709

2111

1494

3163

3261

6665

1168

8476

841

1398

1124

4532

3571

2000

2027

1612

3706

1484

1843

403

402

3268

12

*

*

16

*

*

*

*

17

34

*

1

*

45

24

1

48

4

40

40

51

14

MedianA

PM2.5 (µg/m3)

Median value taken from 1-minute averaged samples during a single measurement session

12

SW

1016

OMBC

40

27

37

OMBC

17

23

5 y 10

OMBC

19

Cumbres

4577

15

SYBC

2251 3417

SYBC

**Parameter not measured at this site

*Equipment malfunction, no value obtained

A

428 1985

13

3

SYBC

14

46

Playas

408

SYBC

41

Playas

3188

900

MedianA

MedianA

SYBC

24

26

Playas

Playas

Site ID

Zone

UVPM (ng/m3)

Black Carbon (ng/m3)

Table 10. Descriptive Results at Single-Measurement Sites

56586

17856

17856

42833

8581

29490

16374

10047

10744

11844

20331

24495

19085

2825

21126

27786

18024

27786

5442

2620

26158

11630

MedianA

UFP (pt/cc)

**

**

**

1.5

**

**

**

**

**

0.7

**

**

**

1.6

**

**

2.1

**

**

**

**

**

MedianA

CO (ppm)

588

162

1140

1272

0

3330

258

1080

270

4170

30

198

1548

1542

1638

4170

930

3180

48

108

4602

888

(ct/hr)

Mean Traffic Density Autos

78

0

162

270

0

360

18

162

30

240

348

1608

210

48

198

192

72

138

0

0

162

72

(ct/hr)

Mean Traffic Density Trucks

98

99

APPENDIX F NITROGEN OXIDES DESCRIPTIVE RESULTS AT SINGLE AND MULTIPLE-MEASUREMENT SITES

1551 1568

41

100

3

14

15

51

110

C

24

200

27

37

116

17

16

12

36

Playas

Playas

Playas

SYBC

SYBC

SYBC

SYBC

SYBC

SYBC

OMBC

OMBC

OMBC

Cumbres

SW

SW

SW

Measurement taken as part of school-related portion of the study

C

**Traffic counts at this site measured more than once, see Appendix C for results

*No truck traffic was observed at this site

Traffic density assessed during a 10 minute window at the beginning of passive sampler exposure period

67

64

42

27

48

42

42

67

61

41

58

79

70

23

27

62

Mean traffic count at this single measurement site

316

398

229

120

284

211

285

359

234

160

245

381

286

150

353

195

27

NO2 (ppb)

B

1512

1531

1558

1767

1663

1641

1613

1469

1755

1571

1535

1532

1582

1775

353

NOx (ppb)

A

C

22

Playas

1601

Site ID

Zone

Exposure Time (min)

248

334

188

93

236

170

243

292

173

119

187

302

216

127

326

133

326

NO (ppb)

Table 11. Nitrogen Oxides Descriptive Results at Single-Measurement Sites

3330

1080

258

270

**

30

**

**

**

**

1638

4170

3180

**

108

888

**

Traffic Density A,B Autos (ct/hr)

360

162

18

30

**

348

**

**

**

**

198

192

138

**

*

72

**

Traffic Density A,B Trucks (ct/hr)

100

75 34-91 92 87-97 49 33-61 30 27-34 28 24-32

***Non-traffic background site

*No truck traffic was observed at this site

Two of these measurements per site taken as part of school-related portion of the study

293 255-493 505 313-697 232 129-600 209 128-289 195 165-225

NO2 (ppb) Median Range

Traffic density assessed during a 10 minute window at the beginning of passive sampler exposure period

1599

1648

1583

1509

1572

NOx (ppb) Median Range

C

300 3 400 2 500 3 1000 2 2000 2

N

Mean Exposure Time (min)

A

Cumbres

Bckground

OMBC

5 y 10

Mid-City

Zone

Site ID

218 198-401 413 216-610 184 96-539 178 101-255 167 133-201

NO (ppb) Median Range

Table 12. Nitrogen Oxides Descriptive Results at Multiple-Measurement Sites

18-138

3570 3150-3912 2988 1158-4350 2520 1890-2820 ***

Traffic DensityA Autos (ct/hr) Median Range

*

132 90-132 522 372-750 420 228-690 ***

Traffic DensityA Trucks (ct/hr) Median Range

101

102

APPENDIX G POLLUTANT TEMPORAL VARIABILITY ACROSS MONTHS THROUGHOUT STUDY

103

Figure 9. Temporal variability of BC across months.

Figure 10. Temporal variability of PM2.5 across months.

104

Figure 11. Temporal variability of UFP count across months.

105

APPENDIX H DESCRIPTIVE RESULTS AT SCHOOL-RELATED SITES

2

2 3

3 2

100a

101

102

103

200a

105

106

107

108

109

110

300a

Playas

Playas

Playas

Playas

SYBC

SYBC

SYBC

SYBC

SYBC

SYBC

SYBC

Mid-City

2

2

2

2

2

2

2

n

Site ID

Zone

UVPM (ng/m3) Median Range 4935 4043-5827 1525 1245-1804 3167 2150-4184 2130 1099-3161 3068 2494-7451 4921 3528-6314 4197 1122-7273 3675 3580-3769 5909 1145-10673 8726 3603-13849 3705 1586-4533 3792 2637-4947

Black Carbon (ng/m3) Median Range 5507 3960-7054 1750 3054-2146 3919 3014-4823 2138 1102-3173 3642 3007-7776 5272 3311-7232 4433 1134-7552 4396 4122-4670 6931 1468-12393 9233 3928-14538 3746 1843-4437 4201 2792-5610

Table 13. Descriptive Results at School-Related Sites

65 13-71

50.5 35-66 32 15-49 57.5 30-85 35.5 23-48 27 1-36 21 15-27 27.5 10-45 21.5 19-24 46.5 22-71 54 28-80 26 1-29

Median Range

PM2.5 (µg/m3)

52518 51025-54010

23944 16502-31385

*

70179 63495-76863 21108 14337-27879 38658 31455-45860 14397 11918-16876 29661 28566-53836 43659 29663-57655 23665 11938-35391 25052 22618-27485 *

Median Range

UFP (pt/cc)

2.9 2.2-3.5

2.3 1.7-2.8 1.9 1.8-2.0 1.9 0.1-6.6 1.7 1.4-1.9 2.55 1.4-3.7 2.0 1.6-2.4 2.0 1.8-2.2 1.7 0.9-2.4 2.5 1.0-4.0 3.1 1.7-4.5 1.8 1.6-1.9

Median Range

CO (ppm)

132 90-132

78 0-240 6 0-12 51 42-60 15 0-30 60 12-78 126 120-132 12 0-60 6 0-12 9 0-18 42 0-60 60 12-78

Median Range

(table continues)

3570 3150-3912

3180 3150-3210 1605 1518-1692 915 798-1032 600 492-708 1158 648-1188 1857 1692-2022 105 90-120 180 150-210 519 480-558 936 912-960 1110 648-1188

Median Range

Traffic Density Traffic Density Autos (ct/hr) Trucks (ct/hr)

106

Site ID n

UVPM (ng/m3) Median Range

Black Carbon (ng/m3) Median Range

Median Range

PM2.5 (µg/m3)

111

2

Median Range

CO (ppm) Median Range

c

180 132-228 240 192-288 **

**

84 60-108 48 0-78 87 72-102 522 372-750 210 198-222 780 648-912 420 228-690 138 132-222

Median Range

Traffic Density Traffic Density Autos (ct/hr) Trucks (ct/hr)

1.1 843 0.7-1.4 828-858 16634 1.1 702 9991-23276 0-4.3 30-948 * 3.1 2184 1.8-4.3 2088-2280 46665 3.4 2988 8081-86671 2.3-4.5 1158-4350 76168 3.2 2091 73821-78515 2.8-3.5 2022-2160 67923 2.8 2715 35300-100545 2.0-3.5 2472-2958 54903 1.8 2520 18263-140216 1.5-2.9 1890-2820 37235 1.7 612 32713-55938 1.4-2.0 558-960 17847 1.3 111 12591-23103 0.8-1.8 72-150 38186 1.5 780 480-1140 30131-46241 1.2-1.7 56160 1.4 1770 51716-60603 1.3-1.4 1758-1782 13136 2.0 72 6820-44015 1.9-2.0 18-138 10417 0.3 c 4835-27975 0.1-1.6 c No traffic circulation at anchor background site

*

Median Range

UFP (pt/cc)

*Instrument failure during at least one measurement, median value or range could not be calculated ** Zero traffic was observed at this site during measurements

Mid-City

3087 2644 34 2727-3436 2594-2694 10-58 2225 1622 44 Mid-City 112 3 1682-2433 1536-2485 13-61 9443 8187 62.5 Mid-City 113 2 3673-15213 2713-13661 54-71 7021 6282 14 5 y 10 400a 15 1701-12841 1097-11798 1-50 11317 10007 46 5 y 10 114 2 11288-11346 9823-10191 41-51 9850 8647 48 5 y 10 115 2 7784-11915 6767-10526 47-49 9429 8796 44.5 OMBC 500a 14 3210-94245 2834-51440 11-175 8750 10002 34 OMBC 116 3 3121-11671 3170-10718 10-58 3062 2597 24 OMBC 117 2 1264-4861 1338-3856 9-39 5498 4928 49 OMBC 118 2 4982-6814 4063-5793 34-64 4786 3762 33.5 OMBC 119 2 3936-5635 3266-4258 9-58 Cumbres 2000 3 1134 664 15 494-3268 462-2927 11-24 11 1401 1015 14 Bckground 1000b 543-2636 399-1844 1-69 a High-traffic reference site at this zone b Anchor background site sampled at each session

Zone

Table 13. (continued)

107

108

APPENDIX I POLLUTANT CONCENTRATIONS AND TRAFFIC MAPS

Figure 12. Study sites.

109

Figure 13. Carbon monoxide (CO) median concentrations (ppm).

110

Figure 14. Ultrafine particle (UFP) count median concentrations (pt/cc).

111

Figure 15. All vehicle traffic median counts (ct/hr).

112

Figure 16. Black carbon (BC) median concentrations (ng/m3).

113

Figure 17. Truck traffic median counts (ct/hr).

114