Models for human exposure to air pollution

Four models for human exposure to air pollution are discussed and compared. The simple microenvironment monitoring model measures pollutant concentrations at fixed locations, regarded as proxies for similar locations or microenvironments. Since this model does not require pollutant measurements on the individual level, it is easy to implement. However, the model can only be used to estimate the average exposure in a population, and it does not provide any estimate of the variability and distribution of individual exposures. The replicated microenvironment monitoring model provides some estimates of the variability and distribution. However, because of the possible discrepancy between the microenvironment concentration distribution and the individual concentration distribution, some adjustment might be necessary. Integrated personal monitoring allows direct estimation of the average exposure as well as the variability and distribution of individual exposures. Coupled with the appropriate time budget data, a regression analysis can be applied to estimate the contribution from each microenvironment type. However, possible collinearity problems might result in low precision in those estimates. Moreover, it might be difficult to adjust for a possible Hawthorne effect. Continuous personal monitoring has the advantage of recording exposure in each microenvironment type separately, allowing direct estimation of the average exposure as well as the variability and distribution of exposures in each microenvironment type. Moreover, it can also be conducted in conjunction with a two-stage sampling scheme, using information from a large data base on activity patterns, thereby making more efficient use of the monitoring data. It is also easier to adjust for a possible Hawthorne effect in this design.     

A review on the human health impact of airborne particulate matter

Particulate matter (PM) is a key indicator of air pollution brought into the air by a variety of natural and human activities. As it can be suspended over long time and travel over long distances in the atmosphere, it can cause a wide range of diseases that lead to a significant reduction of human life. The size of particles has been directly linked to their potential for causing health problems. Small particles of concern include “inhalable coarse particles” with a diameter of 2.5 to 10 μm and “fine particles” smaller than 2.5 μm in diameter. As the source–effect relationship of PM remains unclear, it is not easy to define such effects from individual sources such as long-range transport of pollution. Because of the potent role of PM and its associated pollutants, detailed knowledge of their human health impacts is of primary importance. This paper summarizes the basic evidence on the health effects of particulate matter. An in-depth analysis is provided to address the implications for policy-makers so that more stringent strategies can be implemented to reduce air pollution and its health effects.     

Personal monitoring of air pollution exposures

In industrial hygiene and health physics the goal has been to protect the health of the individual. Therefore monitoring the exposure people actually receive has been the principal concern. In regulating public exposures to air pollution, the focus has been much different. Recently, use of personal monitors and alternative means of estimating actual exposures has expanded rapidly. The role of personal monitors in epidemiology, exposure studies, and in supplementing the existing fixed station monitoring network for establishing trends and for regulatory purposes is discussed. The implications for air quality standards in recent findings of personal and indoor exposures is considered. New developments that are needed, and those that are not needed, are outlined.     

A study of effects of ambient urban air pollution using personal samplers; a preliminary report

Air pollution has been associated with an increased incidence of respiratory disease. However, significant differences may exist between air pollution levels measured at conventional fixed monitoring stations and actual levels inhaled by a subject. Furthermore, studies of effects of air pollution might best be done using asthmatics as study subjects, since they have irritable airways. This is a preliminary report of a study using a control and asthmatic group in which effects of air pollution are assessed by sympton and medication diaries and simple pulmonary function tests. Air pollution exposure is measured using a small portable sampler for particulates, SO₂ and NO₂; these samplers are carried by the subject (“personal”) and are situated inside and outside the home. Levels obtained are compared to data obtained from the same type of sampler located at a fixed monitoring station. Preliminary analysis of the data shows that the levels of pollutants are low and there are significant differences between the four air pollution measurements, with weak correlations among the various measurements. In this preliminary report, change in pulmonary function during the day correlates only with personal NO₂ measurements. This suggests the need for estimating air pollution exposure using “personal” samplers, when investigating health effects.     

Cost-Effective Control of Ground-Level Ozone Pollution in and around Beijing

Abstract

Ground level ozone pollution has become a significant air pollution problem in Beijing. Because of the complex way in which ozone is formed, it is difficult for policy makers to identify optimal control options on a cost-effective basis. This paper identifies and assesses a range of options for addressing this problem. We apply the Ambient Least Cost Model and compare the economic costs of control options, then recommend the most effective sequence to realize pollution control at the lowest cost. The study finds that installing of Stage II gasoline vapor recovery system at Beijing’s 1446 gasoline stations would be the most cost-effective option. Overall, options to reduce ozone pollution by cutting vehicular emissions are much more cost-effective than options to “clean up” coal-fired power plants.     

Exposure to long-term air pollution and road traffic noise in relation to cholesterol: A cross-sectional study

BackgroundExposure to traffic noise and air pollution have both been associated with cardiovascular disease, though the mechanisms behind are not yet clear.ObjectivesWe aimed to investigate whether the two exposures were associated with levels of cholesterol in a cross-sectional design.MethodsIn 1993–1997, 39,863 participants aged 50–64 year and living in the Greater Copenhagen area were enrolled in a population-based cohort study. For each participant, non-fasting total cholesterol was determined in whole blood samples on the day of enrolment. Residential addresses 5-years preceding enrolment were identified in a national register and road traffic noise (L_den) were modeled for all addresses. For air pollution, nitrogen dioxide (NO₂) was modeled at all addresses using a dispersion model and PM_2.5 was modeled at all enrolment addresses using a land-use regression model. Analyses were done using linear regression with adjustment for potential confounders as well as mutual adjustment for the three exposures.ResultsBaseline residential exposure to the interquartile range of road traffic noise, NO₂ and PM_2.5 was associated with a 0.58 mg/dl (95% confidence interval: − 0.09; 1.25), a 0.68 mg/dl (0.22; 1.16) and a 0.78 mg/dl (0.22; 1.34) higher level of total cholesterol in single pollutant models, respectively. In two pollutant models with adjustment for noise in air pollution models and vice versa, the association between air pollution and cholesterol remained for both air pollution variables (NO₂: 0.72 (0.11; 1.34); PM_2.5: 0.70 (0.12; 1.28) mg/dl), whereas there was no association for noise (− 0.08 mg/dl). In three-pollutant models (NO₂, PM_2.5 and road traffic noise), estimates for NO₂ and PM_2.5 were slightly diminished (NO₂: 0.58 (− 0.05; 1.22); PM_2.5: 0.57 (− 0.02; 1.17) mg/dl).ConclusionsAir pollution and possibly also road traffic noise may be associated with slightly higher levels of cholesterol, though associations for the two exposures were difficult to separate.     

Pollution biomarkers in two estuarine invertebrates,Nereis diversicolor and Scrobicularia plana,from a Marsh ecosystem in SW Spain

The polychaete worm Nereis diversicolor and the clam Scrobicularia plana were collected from several sites, affected by different types of contamination, in a littoral enclosure in the SW Spain (Caño Sancti-Petri and Rio San Pedro). N. diversicolor was present in 6 sampling sites whereas S. plana in 4 of them. The aim of our study was to relate several pollution biomarkers to chemical sources (metals and organic pollutants e.g. PCB, PAH) in these species, thereby confirming their adequacy as sentinels for this habitat. The biomarkers surveyed in the two invertebrates were the activities of the antioxidant enzyme catalase (CAT), the phase II detoxifying enzyme glutathione S-transferase (GST) and the neurotoxicity marker acetylcholinesterase (AChE). Metallothionein (MT) levels were measured as a biomarker of exposure to metals. The results suggested a different response in the two sediment-dwelling organisms, the sediment-eating polychaete and the water-filtering clam, probably as a consequence of different contamination exposures. The results also suggested that samples from the “Caño Sancti-Petri” were exposed to biologically active compounds that altered some of their biochemical responses. Of all the biomarkers tested, AChE was the most sensitive one and N. diversicolor the potentially most robust sentinel in this ecosystem. In this low to moderately polluted environment, the biochemical approach better reflected temporal trends than site-related differences although it was also able to detect punctual chemical insults.     

Functional paraoxonase 1 variants modify the risk of Parkinson's disease due to organophosphate exposure

BackgroundWe previously demonstrated that carriers of the “slower metabolizer” MM genotype of paraoxonase (PON1) who were also exposed to ambient organophosphate (OP) pesticides at their residences were at increased risk of developing Parkinson's disease (PD). Here, with a larger sample size, we extend our previous investigation to consider additional sources of ambient exposure and examined two additional functional PON1 variants.MethodsFrom 2001 to 2011, we enrolled incident cases of idiopathic PD and population controls living in central California. We genotyped three well-known functional PON1 SNPs: two exonic polymorphisms (PON1_L55M and PON1_Q192R) and the promoter region variant (PON1_C-108T). Ambient exposures to diazinon, chlorpyrifos, and parathion at residential and workplace addresses were assessed using a validated geographic information system-based model incorporating records of agricultural pesticide applications in California.ResultsThe odds ratio (OR) for Caucasians exposed to OPs at either residential or workplace addresses varied by PON1 genotype; for exposed carriers of the “faster” metabolizer genotypes, ML or LL, we estimated lower odds ratios (range, 1.20–1.39) than for exposed carriers of the “slower” metabolizer genotype MM (range, 1.78–2.45) relative to unexposed carriers of the faster genotypes. We observed similarly increased ORs for exposure across PON1_Q192R genotypes, but no differences across PON1_C-108T genotypes. The largest ORs were estimated for exposed carriers of both PON1_192QQ and PON1_55MM (OR range, 2.84–3.57).ConclusionsSeveral functional PON1 variants may act together to modify PD risk for ambient OP exposures. While either PON1_L55M or PON1_Q192R may be sufficient to identify increased susceptibility, carriers of both slow metabolizer variants seem most susceptible to OP exposures.     

Short-term effects of air pollution on respiratory morbidity at Rio de Janeiro — PART I: Air Pollution Assessment

Exposure to air pollution has been related with the most varied adverse health outcomes. This study aims to assess the impact of air pollution on the emergency hospitalization for respiratory disease in Rio de Janeiro, Brazil. The study was divided in two parts: Part I specifically addressing the air pollution assessment and Part II addressing the health assessment. Accordingly, this Part I aims to: i) evaluate the concentrations of PM₁₀, SO₂ and CO at two sites in Rio de Janeiro and compare them; ii) analyse the concentrations observed according to the national and international standards; and iii) analyse the air pollutants behaviour, namely, annually, seasonally, daily and considering weekdays/weekends variations. The pollutant concentrations were measured at two different sites in Rio de Janeiro and the analysis was performed for the period between September 2000 and December 2005. Results showed that PM₁₀ concentrations in Rio de Janeiro exceeded the daily and annual standards imposed by the European Union, the Brazilian legislation and WHO guidelines. Regarding SO₂ and CO, concentrations were, generally, below both European and Brazilian standards. Nevertheless, considering WHO guidelines, SO₂ threshold for daily concentrations (20 μg m^− 3) was exceeded around 150 times. Behaviour assessment showed that the influence of traffic is a major factor affecting the air pollution in Rio de Janeiro.Considering the results achieved and the proven health effects of air pollution, strategies should be defined for its reduction, particularly concerning particulate matter, and consequently contribute to the protection of public health.     

Indoor air quality and minimum ventilation rate

In a 30-m³ test chamber the air pollutants caused by man were measured. Variables were the number of persons and the rate of air change. During 2-h test sessions the temperature, relative humidity, carbon dioxide, and intensity of odors were measured. The relationship between the perceived odor intensities and the concentrations of carbon dioxide-independent of the number of persons and the air change rate—was observed. At air change rates of 12–15 m³/person/h, the carbon dioxide concentration was not higher than 0.15% and the odor intensity was evaluated only as a “slight annoyance.” Higher ventilation rates are necessary if smoking and increased physical activities are done in the rooms.     

Identification of fipronil metabolites by time-of-flight mass spectrometry for application in a human exposure study

Fipronil is a phenylpyrazole insecticide commonly used in residential and agricultural applications. To understand more about the potential risks for human exposure associated with fipronil, urine and serum from dosed Long Evans adult rats (5 and 10 mg/kg bw) were analyzed to identify metabolites as potential biomarkers for use in human biomonitoring studies. Urine from treated rats was found to contain seven unique metabolites, two of which had not been previously reported—M4 and M7 which were putatively identified as a nitroso compound and an imine, respectively. Fipronil sulfone was confirmed to be the primary metabolite in rat serum. The fipronil metabolites identified in the respective matrices were then evaluated in matched human urine (n = 84) and serum (n = 96) samples from volunteers with no known pesticide exposures. Although no fipronil or metabolites were detected in human urine, fipronil sulfone was present in the serum of approximately 25% of the individuals at concentrations ranging from 0.1 to 4 ng/mL. These results indicate that many fipronil metabolites are produced following exposures in rats and that fipronil sulfone is a useful biomarker in human serum. Furthermore, human exposure to fipronil may occur regularly and require more extensive characterization.     

Carbon monoxide levels in microenvironment types of four U.S. cities

Portable monitors were used to measure time-averaged personal exposure (10–30 min) to carbon monoxide. Data were collected from January through March 1981 in four cities where ambient carbon monoxide levels have been reported in excess of National Ambient Air Quality Standards: Stamford, CT; Los Angeles, CA; Phoenix, AZ; and Denver, CO. In each city, personal exposure were measured in three common microenvironment types (indoor, commuting, and residential driving) near fixed stations monitoring ambient levels of carbon monoxide. Measurements recorded at urban-residential fixed monitoring stations (excluding one station in Stamford) underrepresented the time-weighted mean of commuting and residential driving exposures by factors of 0.4 to 0.7. The highest mean commuting and residential driving exposures were found in Los Angeles (16.1 and 7.6 μL/L, respectively). Fixed monitoring stations in Los Angeles, Phoenix, and one station in Stamford overrepresented the time-weighted mean of indoor exposures by factors of 1.1 to 1.3. However, in Denver and another station in Stamford, urban stations underrepresented the mean of indoor exposures by factors of 0.4 to 0.8. The highest mean indoor exposure, 5.9 μL/L, was in Denver. In all four cities, regressing personal exposures on concurrent fixed-site concentrations for all recorded values and for values recorded during 8-h NAAQS exceedance time periods revealed no conclusive linear relationships.     

Relationships among personal,indoor and outdoor NO2 measurements

Using integrating NO₂ diffusion dosimeters, personal, indoor and outdoor exposures were measured for nine families in Topeka, Kansas. NO₂ exposures in homes that used gas for cooking were clearly different from those in homes that used electricity. The gas-cooking homes had indoor levels three times the outdoor levels. Members of the gas-cooking households had levels twice those of electric-cooking families and twice the outdoor levels. A linear model that includes outdoor concentrations and stove types explains 77% of the variance in observed NO₂ exposure. The differential NO₂ exposures in homes with and without gas stoves should be considered in epidemiologic studies of the health effects of air pollution.     

Concentrations of polybrominated diphenyl ethers (PBDEs) in matched samples of human milk,dust and indoor air

Polybrominated diphenyl ethers (PBDEs) are lipophilic, persistent pollutants found worldwide in environmental and human samples. Exposure pathways for PBDEs remain unclear but may include food, air and dust. The aim of this study was to conduct an integrated assessment of PBDE exposure and human body burden using 10 matched samples of human milk, indoor air and dust collected in 2007–2008 in Brisbane, Australia. In addition, temporal analysis was investigated comparing the results of the current study with PBDE concentrations in human milk collected in 2002–2003 from the same region.PBDEs were detected in all matrices and the median concentrations of BDEs -47 and -209 in human milk, air and dust were: 4.2 and 0.3 ng/g lipid; 25 and 7.8 pg/m³; and 56 and 291 ng/g dust, respectively. Significant correlations were observed between the concentrations of BDE-99 in air and human milk (r = 0.661, p = 0.038) and BDE-153 in dust and BDE-183 in human milk (r = 0.697, p = 0.025). These correlations do not suggest causal relationships — there is no hypothesis that can be offered to explain why BDE-153 in dust and BDE-183 in milk are correlated. The fact that so few correlations were found in the data could be a function of the small sample size, or because additional factors, such as sources of exposure not considered or measured in the study, might be important in explaining exposure to PBDEs. There was a slight decrease in PBDE concentrations from 2002–2003 to 2007–2008 but this may be due to sampling and analytical differences. Overall, average PBDE concentrations from these individual samples were similar to results from pooled human milk collected in Brisbane in 2002–2003 indicating that pooling may be an efficient, cost-effective strategy of assessing PBDE concentrations on a population basis.The results of this study were used to estimate an infant's daily PBDE intake via inhalation, dust ingestion and human milk consumption. Differences in PBDE intake of individual congeners from the different matrices were observed. Specifically, as the level of bromination increased, the contribution of PBDE intake decreased via human milk and increased via dust. As the impacts of the ban of the lower brominated (penta- and octa-BDE) products become evident, an increased use of the higher brominated deca-BDE product may result in dust making a greater contribution to infant exposure than it does currently.To better understand human body burden, further research is required into the sources and exposure pathways of PBDEs and metabolic differences influencing an individual's response to exposure. In addition, temporal trend analysis is necessary with continued monitoring of PBDEs in the human population as well as in the suggested exposure matrices of food, dust and air.     

Air pollutant exposure and preterm and term small-for-gestational-age births in Detroit,Michigan: Long-term trends and associations

Studies in a number of countries have reported associations between exposure to ambient air pollutants and adverse birth outcomes, including low birth weight, preterm birth (PTB) and, less commonly, small for gestational age (SGA). Despite their growing number, the available studies have significant limitations, e.g., incomplete control of temporal trends in exposure, modest sample sizes, and a lack of information regarding individual risk factors such as smoking. No study has yet examined large numbers of susceptible individuals.We investigated the association between ambient air pollutant concentrations and term SGA and PTB outcomes among 164,905 singleton births in Detroit, Michigan occurring between 1990 and 2001. SO₂, CO, NO₂, O₃ and PM₁₀ exposures were used in single and multiple pollutant logistic regression models to estimate odds ratios (OR) for these outcomes, adjusted for the infant's sex and gestational age, the mother's race, age group, education level, smoking status and prenatal care, birth season, site of residence, and long-term exposure trends.Term SGA was associated with CO levels exceeding 0.75 ppm (OR = 1.14, 95% confidence interval = 1.02–1.27) and NO₂ exceeding 6.8 ppb (1.11, 1.03–1.21) exposures in the first month, and with PM₁₀ exceeding 35 μg/m³ (1.22, 1.03–1.46) and O₃ (1.11, 1.02–1.20) exposure in the third trimester. PTB was associated with SO₂ (1.07, 1.01–1.14) exposure in the last month, and with (hourly) O₃ exceeding 92 ppb (1.08, 1.02–1.14) exposure in the first month.Exposure to several air pollutants at modest concentrations was associated with adverse birth outcomes. This study, which included a large Black population, suggests the importance of the early period of pregnancy for associations between term SGA with CO and NO₂, and between O₃ with PTB; and the late pregnancy period for associations between term SGA and O₃ and PM₁₀, and between SO₂ with PTB. It also highlights the importance of accounting for individual risk factors such as maternal smoking, maternal race, and long-term trends in air pollutant levels and adverse birth outcomes in evaluating relationships between pollutant exposures and adverse birth outcomes.     

Can car air filters be useful as a sampling medium for air pollution monitoring purposes?

Urban air quality and real human exposure to chemical environmental stressors is an issue of high scientific and political interest. In an effort to find innovative and inexpensive means for air quality monitoring, the ability of car engine air filters (CAFs) to act as efficient samplers collecting street level air, to which people are exposed to, was tested. In particular, in the case of taxis, air filters are replaced after regular distances, the itineraries are almost exclusively urban, cruising mode is similar and, thus, knowledge of the air flow can provide with an integrated city air sample. The present pilot study focused on polycyclic aromatic hydrocarbons (PAHs), the most important category of organic pollutants associated with traffic emissions. Concentrations of ΣPAHs in CAFs ranged between 650 and 2900 μg CAF^− 1, with benzo[b]fluoranthene, benzo[k]fluoranthene and indeno[123-cd]pyrene being the most abundant PAHs. Benzo[a]pyrene (BaP) ranged between 110 and 250 μg CAF^− 1, accounting regularly for 5–15% of the total carcinogenic PAHs. The CAF PAH loads were used to derive road-level atmospheric PAH concentrations from a standard formula relating to the CAF air flow. Important parameters/assumptions for these estimates are the cruising speed and the exposure duration of each CAF. Based on information obtained from the garage experts, an average ‘sampled air volume’ of 48,750 m³ per CAF was estimated, with uncertainty in this calculation estimated to be about a factor of 4 between the two extreme scenarios. Based on this air volume, ΣPAHs ranged between 13 and 56 ng m^− 3 and BaP between 2.1 and 5.0 ng m^− 3, suggesting that in-traffic BaP concentrations can be many times higher than the limit values set by the UK (0.25 ng m^− 3) and the European Union (1.0 ng m^− 3), or from active sampling stations normally cited on building roof tops or far from city centres.Notwithstanding the limitations of this approach, the very low cost, the continuous availability of very high amounts of “sample”, and the “retroactivity” render it very useful and complementary to existing passive sampling techniques. This approach yields estimated air concentrations that reflect the pollutant concentrations to which taxi drivers, pedestrians, cyclists and road-related professionals are exposed.     

Symptoms prevalence among office workers of a sealed versus a non-sealed building: Associations to indoor air quality

ObjectivesAn increasing number of complaints related to time spent in artificially ventilated buildings have been progressively reported and attributed, at least in part, to physical and chemical exposures in the office environment. The objective of this research was to investigate the association between the prevalence of work-related symptoms and the indoor air quality, comparing a sealed office building with a naturally ventilated one, considering, specially, the indoor concentration of TPM, TVOCs and the main individual VOCs.MethodsA cross-sectional study was performed to compare the prevalence of sick building syndrome (SBS) symptoms among 1736 office workers of a sealed office building and 950 of a non-sealed one, both in Rio de Janeiro's downtown. The prevalence of symptoms was obtained by a SBS standardized questionnaire. The IAQ of the buildings was evaluated through specific methods, to determine the temperature, humidity, particulate matter and volatile organic compound (VOC) concentrations.ResultsUpper airways and ophthalmic symptoms, tiredness and headache were highly prevalent in both buildings. Some symptoms were more prevalent in the sealed building: “eye dryness” 33.3% and 27.1% (p: 0.01); “runny nose” 37.3% and 31.3% (p: 0.03); “dry throat” 42% and 36% (p: 0.02); and “lethargy” 58.5% and 50.5% (p: 0.03) respectively. However, relative humidity and indoor total particulate matter (TPM) concentration as well as total volatile organic compounds (TVOCs) were paradoxically greater in the non-sealed building, in which aromatic compounds had higher concentration, especially benzene. The analysis between measured exposure levels and resulting symptoms showed no association among its prevalence and TPM, TVOCs, benzene or toluene concentration in none of the buildings.ConclusionsOther disregarded factors, like undetected VOCs, mites, molds and endotoxin concentrations, may be associated to the greater prevalence of symptoms in the sealed building.     

Respiratory hospital admissions in young children living near metal smelters,pulp mills and oil refineries in two Canadian provinces

BackgroundIndustrial plants emit air pollutants like fine particles (PM2.5), sulfur dioxide (SO2) and nitrogen dioxide (NO2) that may affect the health of individuals living nearby.ObjectiveTo assess the effects of community exposure to air emissions of PM2.5, SO2, and NO2 from pulp mills, oil refineries, metal smelters, on respiratory hospital admissions in young children in Quebec (QC) and British Columbia (BC), Canada.MethodsWe assessed QC, BC and pooled associations between the following estimates of exposure and hospital admissions for asthma and bronchiolitis in children aged 2–4 years for the years 2002–2010: i) Crude emission exposures at the residential postal codes of children, calculated by multiplying estimated daily emissions of PM2.5, SO2, or NO2 from all nearby (< 7.5 km) pulp mills, oil refineries, metal smelters emitting yearly ≥ 50 t and their total emissions, by the percent of the day each postal code was downwind; ii) Daily levels of these pollutants at central ambient monitoring stations nearby the industries and the children's residences.ResultsSeventy-one major industries were selected between QC and BC, with a total of 2868 cases included in our analyses. More cases were exposed to emissions from major industries in QC than in BC (e.g. 2505 admissions near SO2 industrial emitters in QC vs 334 in BC), although air pollutant levels were similar. Odds ratios (ORs) for crude refinery and smelter emissions were positive in QC but more variable in BC. For example with PM2.5 in QC, ORs were 1.13 per 0.15 t/day (95% CI: 1.00–1.27) and 1.03 (95% CI: 0.99–1.07) for refinery and smelter emissions, respectively. Pooled results of QC and BC for crude total SO2 emissions from all sources indicated a 1% increase (0–3%) in odds of hospital admissions per 1.50 t/day increase in exposure. Associations with measured pollutant levels were only seen in BC, with SO2 and NO2.ConclusionHospital admissions for wheezing diseases in young children were associated with community exposure to industrial air pollutant emissions. Future work is needed to better assess the risk of exposure to complex mixture of air pollutants from multiple industrial sources.     

A framework for identifying distinct multipollutant profiles in air pollution data

BackgroundThe importance of describing, understanding and regulating multi-pollutant mixtures has been highlighted by the US National Academy of Science and the Environmental Protection Agency. Furthering our understanding of the health effects associated with exposure to mixtures of pollutants will lead to the development of new multi-pollutant National Air Quality Standards.ObjectivesIntroduce a framework within which diagnostic methods that are based on our understanding of air pollution mixtures are used to validate the distinct air pollutant mixtures identified using cluster analysis.MethodsSix years of daily gaseous and particulate air pollution data collected in Boston, MA were classified solely on their concentration profiles. Classification was performed using k-means partitioning and hierarchical clustering. Diagnostic strategies were developed to identify the most optimal clustering.ResultsThe optimal solution used k-means analysis and contained five distinct groups of days. Pollutant concentrations and elemental ratios were computed in order to characterize the differences between clusters. Time-series regression confirmed that the groups differed in their chemical compositions. The mean values of meteorological parameters were estimated for each group and air mass origin between clusters was examined using back-trajectory analysis. This allowed us to link the distinct physico-chemical characteristics of each cluster to characteristic weather patterns and show that different clusters were associated with distinct air mass origins.ConclusionsThis analysis yielded a solution that was robust to outlier points and interpretable based on chemical, physical and meteorological characteristics. This novel method provides an exciting tool with which to identify and further investigate multi-pollutant mixtures and link them directly to health effects studies.