Performance of a Low-Cost Sensor Community Air Monitoring Network in Imperial County, CA.

Air monitoring networks developed by communities have potential to reduce exposures and affect environmental health policy, yet there have been few performance evaluations of networks of these sensors in the field. We developed a network of over 40 air sensors in Imperial County, CA, which is delivering real-time data to local communities on levels of particulate matter. We report here on the performance of the Network to date by comparing the low-cost sensor readings to regulatory monitors for 4 years of operation (2015-2018) on a network-wide basis. Annual mean levels of PM10 did not differ statistically from regulatory annual means, but did for PM2.5 for two out of the 4 years. R2s from ordinary least square regression results ranged from 0.16 to 0.67 for PM10, and increased each year of operation. Sensor variability was higher among the Network monitors than the regulatory monitors. The Network identified a larger number of pollution episodes and identified under-reporting by the regulatory monitors. The participatory approach of the project resulted in increased engagement from local and state agencies and increased local knowledge about air quality, data interpretation, and health impacts. Community air monitoring networks have the potential to provide real-time reliable data to local populations.

Measuring airflow through the portable high-efficiency air filtration (PHEAF) device to assess reliability of instrument and sample location.

The portable high-efficiency air filtration (PHEAF) device is an engineering control common to the environmental remediation industry. Damage to the high-efficiency particulate air (HEPA) filter (e.g., filtration media, gasket), improper installation of the filter into the mounting frame, or defects in the filtration housing affect the capture efficiency of the device. PHEAF devices operating at less than marketed efficiencies justify periodic leak testing of the PHEAF device, especially when the filtered air is exhausted into occupied spaces. A leak test is accomplished by injecting a known concentration of aerosol upstream of the HEPA filter and measuring the percentage of aerosol penetrating through the filtration system. The test protocol scripted for stationary systems (i.e., biological safety cabinets) states that upstream concentrations can be empirically determined using the aerosol photometer to measure particulate matter (PM) in the airstream. This practice requires a homogenous mixture of the aerosol challenge agent within the airstream. However, design of the PHEAF device does not include a validated induction point for the aerosol. Absent of an acceptable means to achieve a homogenous mixture for upstream measurement, the aerosol concentration is mathematically derived based on the measured air volume passing through the PHEAF equipment. In this study, intake volume and exhaust volume for each PHEAF device were measured by either the balometer or the hot wire anemometer. Variability of measurements was examined by instrument and sample location (intake vs. exhaust) to understand which combination would be most consistent for measuring airflow volume. From this study, the authors conclude that the balometer is preferred compared with the hot wire anemometer for measuring airflow through the PHEAF device. Exhaust measurement by balometer seems more reliable than intake measurements by hot wire anemometer. Implications: Although testing of PHEAF devices is recommended by various public health authorities, no nationally recognized test protocol has been published in the United States. In support of measuring the performance of the PHEAF device in a field setting, this study evaluated the hot wire anemometer and balometer techniques and sample locations (intake vs. exhaust) to reliably measure airflow through the PHEAF device. Since accuracy of the particle measurement is associated with airflow volume, it is essential to obtain a true airflow reading. This study suggests that the balometer was more consistent in measuring airflow through the PHEAF device.

Combining Community Engagement and Scientific Approaches in Next-Generation Monitor Siting: The Case of the Imperial County Community Air Network.

Air pollution continues to be a global public health threat, and the expanding availability of small, low-cost air sensors has led to increased interest in both personal and crowd-sourced air monitoring. However, to date, few low-cost air monitoring networks have been developed with the scientific rigor or continuity needed to conduct public health surveillance and inform policy. In Imperial County, California, near the U.S./Mexico border, we used a collaborative, community-engaged process to develop a community air monitoring network that attains the scientific rigor required for research, while also achieving community priorities. By engaging community residents in the project design, monitor siting processes, data dissemination, and other key activities, the resulting air monitoring network data are relevant, trusted, understandable, and used by community residents. Integration of spatial analysis and air monitoring best practices into the network development process ensures that the data are reliable and appropriate for use in research activities. This combined approach results in a community air monitoring network that is better able to inform community residents, support research activities, guide public policy, and improve public health. Here we detail the monitor siting process and outline the advantages and challenges of this approach.

Pregnancy outcomes and ethanol cook stove intervention: A randomized-controlled trial in Ibadan, Nigeria.

BACKGROUND: Household air pollution (HAP) exposure has been linked to adverse pregnancy outcomes. OBJECTIVES: A randomized controlled trial was undertaken in Ibadan, Nigeria to determine the impact of cooking with ethanol on pregnancy outcomes. METHODS: Three-hundred-twenty-four pregnant women were randomized to either the control (continued cooking using kerosene/firewood stove, n=162) or intervention group (received ethanol stove, n=162). Primary outcome variables were birthweight, preterm delivery, intrauterine growth restriction (IUGR), and occurrence of miscarriage/stillbirth. RESULTS: Mean birthweights for ethanol and controls were 3076 and 2988g, respectively; the difference, 88g, (95% confidence interval: -18g to 194g), was not statistically significant (p=0.10). After adjusting for covariates, the difference reached significance (p=0.020). Rates of preterm delivery were 6.7% (ethanol) and 11.0% (control), (p=0.22). Number of miscarriages was 1(ethanol) vs. 4 (control) and stillbirths was 3 (ethanol) vs. 7 (control) (both non-significant). Average gestational age at delivery was significantly (p=0.015) higher in ethanol-users (39.2weeks) compared to controls (38.2weeks). Perinatal mortality (stillbirths and neonatal deaths) was twice as high in controls compared to ethanol-users (7.9% vs. 3.9%; p=0.045, after adjustment for covariates). We did not detect significant differences in exposure levels between the two treatment arms, perhaps due to large seasonal effects and high ambient air pollution levels. CONCLUSIONS: Transition from traditional biomass/kerosene fuel to ethanol reduced adverse pregnancy outcomes. However, the difference in birthweight was statistically significant only after covariate adjustment and the other significant differences were in tertiary endpoints. Our results are suggestive of a beneficial effect of ethanol use. Larger trials are required to validate these findings.

Capture efficiency of portable high-efficiency air filtration devices used during building construction activities.

The portable high-efficiency air filtration (PHEAF) device is used to control particulate matter (PM) generated from construction-type activities occurring within the built environment. Examples of activities where PHEAF devices are mobilized include building renovation, asbestos abatement, remediation of microbial contamination, and lead-based paint projects. Designed for use on short-term, temporary projects the PHEAF device captures airborne PM using a high-efficiency particulate air (HEPA) filter. This study sought to evaluate the capture efficiency of the PHEAF device in a field setting. An aerosol generator and photometer were used to measure particle penetration through 85 PHEAF devices. Average overall capture efficiency ranged from 41.78% to ≥99.97% with more than 88% of the tests failing to achieve 99.97% capture efficiency. Approximately 73% of the PHEAF device sample population failed to demonstrate HEPA performance criteria during any test round. A higher occurrence of PM concentrations measured around the perimeter of the filter suggested the presence of bypass leakage. While PHEAF devices were effective in capturing a significant quantity of aerosol test agent, these findings suggest that routine testing of the PHEAF device should be conducted to validate performance.

The Imperial County Community Air Monitoring Network: A Model for Community-based Environmental Monitoring for Public Health Action.

SUMMARY: The Imperial County Community Air Monitoring Network (the Network) is a collaborative group of community, academic, nongovernmental, and government partners designed to fill the need for more detailed data on particulate matter in an area that often exceeds air quality standards. The Network employs a community-based environmental monitoring process in which the community and researchers have specific, well-defined roles as part of an equitable partnership that also includes shared decision-making to determine study direction, plan research protocols, and conduct project activities. The Network is currently producing real-time particulate matter data from 40 low-cost sensors throughout Imperial County, one of the largest community-based air networks in the United States. Establishment of a community-led air network involves engaging community members to be citizen-scientists in the monitoring, siting, and data collection process. Attention to technical issues regarding instrument calibration and validation and electronic transfer and storage of data is also essential. Finally, continued community health improvements will be predicated on facilitating community ownership and sustainability of the network after research funds have been expended. https://doi.org/10.1289/EHP1772

Development and field validation of a community-engaged particulate matter air quality monitoring network in Imperial, California, USA.

The Imperial County Community Air Monitoring Network was developed as part of a community-engaged research study to provide real-time particulate matter (PM) air quality information at a high spatial resolution in Imperial County, California. The network augmented the few existing regulatory monitors and increased monitoring near susceptible populations. Monitors were both calibrated and field validated, a key component of evaluating the quality of the data produced by the community monitoring network. This paper examines the performance of a customized version of the low-cost Dylos optical particle counter used in the community air monitors compared with both PM2.5 and PM10 (particulate matter with aerodynamic diameters <2.5 and <10 µm, respectively) federal equivalent method (FEM) beta-attenuation monitors (BAMs) and federal reference method (FRM) gravimetric filters at a collocation site in the study area. A conversion equation was developed that estimates particle mass concentrations from the native Dylos particle counts, taking into account relative humidity. The R2 for converted hourly averaged Dylos mass measurements versus a PM2.5 BAM was 0.79 and that versus a PM10 BAM was 0.78. The performance of the conversion equation was evaluated at six other sites with collocated PM2.5 environmental beta-attenuation monitors (EBAMs) located throughout Imperial County. The agreement of the Dylos with the EBAMs was moderate to high (R2 = 0.35-0.81). IMPLICATIONS: The performance of low-cost air quality sensors in community networks is currently not well documented. This paper provides a methodology for quantifying the performance of a next-generation Dylos PM sensor used in the Imperial County Community Air Monitoring Network. This air quality network provides data at a much finer spatial and temporal resolution than has previously been possible with government monitoring efforts. Once calibrated and validated, these high-resolution data may provide more information on susceptible populations, assist in the identification of air pollution hotspots, and increase community awareness of air pollution.

Randomized Controlled Ethanol Cookstove Intervention and Blood Pressure in Pregnant Nigerian Women.

RATIONALE: Hypertension during pregnancy is a leading cause of maternal mortality. Exposure to household air pollution elevates blood pressure (BP). OBJECTIVES: To investigate the ability of a clean cookstove intervention to lower BP during pregnancy. METHODS: We conducted a randomized controlled trial in Nigeria. Pregnant women cooking with kerosene or firewood were randomly assigned to an ethanol arm (n = 162) or a control arm (n = 162). BP measurements were taken during six antenatal visits. In the primary analysis, we compared ethanol users with control subjects. In subgroup analyses, we compared baseline kerosene users assigned to the intervention with kerosene control subjects and compared baseline firewood users assigned to ethanol with firewood control subjects. MEASUREMENTS AND MAIN RESULTS: The change in diastolic blood pressure (DBP) over time was significantly different between ethanol users and control subjects (P = 0.040); systolic blood pressure (SBP) did not differ (P = 0.86). In subgroup analyses, there was no significant intervention effect for SBP; a significant difference for DBP (P = 0.031) existed among preintervention kerosene users. At the last visit, mean DBP was 2.8 mm Hg higher in control subjects than in ethanol users (3.6 mm Hg greater in control subjects than in ethanol users among preintervention kerosene users), and 6.4% of control subjects were hypertensive (SBP ≥140 and/or DBP ≥90 mm Hg) versus 1.9% of ethanol users (P = 0.051). Among preintervention kerosene users, 8.8% of control subjects were hypertensive compared with 1.8% of ethanol users (P = 0.029). CONCLUSIONS: To our knowledge, this is the first cookstove randomized controlled trial examining prenatal BP. Ethanol cookstoves have potential to reduce DBP and hypertension during pregnancy. Accordingly, clean cooking fuels may reduce adverse health impacts associated with household air pollution. Clinical trial registered with www.clinicaltrials.gov (NCT02394574).

An Overview of Occupational Risks From Climate Change.

Changes in atmosphere and temperature are affecting multiple environmental indicators from extreme heat events to global air quality. Workers will be uniquely affected by climate change, and the occupational impacts of major shifts in atmospheric and weather conditions need greater attention. Climate change-related exposures most likely to differentially affect workers in the USA and globally include heat, ozone, polycyclic aromatic hydrocarbons, other chemicals, pathogenic microorganisms, vector-borne diseases, violence, and wildfires. Epidemiologic evidence documents a U-, J-, or V-shaped relationship between temperature and mortality. Whereas heat-related morbidity and mortality risks are most evident in agriculture, many other outdoor occupational sectors are also at risk, including construction, transportation, landscaping, firefighting, and other emergency response operations. The toxicity of chemicals change under hyperthermic conditions, particularly for pesticides and ozone. Combined with climate-related changes in chemical transport and distribution, these interactions represent unique health risks specifically to workers. Links between heat and interpersonal conflict including violence require attention because they pose threats to the safety of emergency medicine, peacekeeping and humanitarian relief, and public safety professionals. Recommendations for anticipating how US workers will be most susceptible to climate change include formal monitoring systems for agricultural workers; modeling scenarios focusing on occupational impacts of extreme climate events including floods, wildfires, and chemical spills; and national research agenda setting focusing on control and mitigation of occupational susceptibility to climate change.

Spatial and Temporal Distribution of Polycyclic Aromatic Hydrocarbons and Elemental Carbon in Bakersfield, California.

Despite increasing evidence that airborne polycyclic aromatic hydrocarbon (PAH) exposures contribute to adverse health outcomes for sensitive populations, limited data are available on short-term intraurban spatial distributions for use in epidemiologic research. Exposure assessments for airborne PAHs are uncommon because air sampling for PAHs is a labor-, equipment-, and time-intensive task. To address this gap we measured wintertime PAH concentrations during 2010-2011 in Bakersfield, California, USA, a major city in the Southern San Joaquin Valley. Specifically, 58 96-hour integrated PAH samples were collected during 4 time periods at 14 locations from November 2010 to January 2011; duplicates were collected at two sites. We also collected elemental carbon (EC) at the same 14 sites and analyzed the two time periods with the highest ambient PAH pollution. We used linear regression models to quantify the relationship between potential spatial and temporal predictors of PAH concentrations. We found that wintertime PAH concentrations in Bakersfield, CA, are best predicted by meteorological variables and traffic proximity. Our model explains a moderate amount of the variability in the data (R2=0.58), likely reflecting the major sources of PAHs in Bakersfield. We also observed that PAH concentrations were more spatially variable than EC concentrations. Comparing our data to historical monitoring data at one location in Bakersfield showed that the relatively low PAH concentrations during the 2010-2011 winter in Bakersfield is part of a long-term trend in decreasing PAH concentrations.

Integrating modelling and smart sensors for environmental and human health.

Sensors are becoming ubiquitous in everyday life, generating data at an unprecedented rate and scale. However, models that assess impacts of human activities on environmental and human health, have typically been developed in contexts where data scarcity is the norm. Models are essential tools to understand processes, identify relationships, associations and causality, formalize stakeholder mental models, and to quantify the effects of prevention and interventions. They can help to explain data, as well as inform the deployment and location of sensors by identifying hotspots and areas of interest where data collection may achieve the best results. We identify a paradigm shift in how the integration of models and sensors can contribute to harnessing 'Big Data' and, more importantly, make the vital step from 'Big Data' to 'Big Information'. In this paper, we illustrate current developments and identify key research needs using human and environmental health challenges as an example.

Assessing exposures to household air pollution in public health research and program evaluation.

Exposure to smoke from the use of solid fuels and inefficient stoves for cooking and heating is responsible for approximately 4 million premature deaths yearly. As increasing investments are made to tackle this important public health issue, there is a need for identifying and providing guidance on best practices for exposure and stove performance monitoring, particularly for public health research and evaluation studies. This paper, which builds upon the discussion at an expert consultation on exposure assessment convened by the Global Alliance for Clean Cookstoves, the Centers for Disease Control and Prevention, and PATH in late 2012, aims to provide general guidance on what to monitor, who and where to monitor, and how to monitor household air pollution exposures. In addition, we summarize information about commercially available monitoring equipment and the technical properties of these monitors most important for household air pollution exposure assessment. The target audience includes epidemiologists conducting health studies and program evaluators aiming to quantify changes in exposures to estimate the potential health benefits of cookstoves intervention projects.

Biomarkers of secondhand smoke exposure in automobiles.

OBJECTIVES: The objectives of this study were: (1) to characterise the exposure of non-smokers exposed to secondhand smoke (SHS) in a vehicle using biomarkers, (2) to describe the time course of the biomarkers over 24 h, and (3) to examine the relationship between tobacco biomarkers and airborne concentrations of SHS markers. METHODS: Eight non-smokers were individually exposed to SHS in cars with fully open front windows and closed back windows over an hour from a smoker who smoked three cigarettes at 20 min intervals. The non-smokers sat in the back seat on the passenger side, while the smoker sat in the driver's seat. Plasma cotinine and urine cotinine, 3-hydroxycotinine (3HC) and 4-(methylnitrosoamino)-(3-pyridyl)-1-butanol (NNAL) were compared in samples taken at baseline (BL) and several time-points after exposure. Nicotine, particulate matter (PM2.5) and carbon monoxide (CO) were measured inside and outside the vehicle and ventilation rates in the cars were measured. RESULTS: Average plasma cotinine and the molar sum of urine cotinine and 3HC (COT+3HC) increased four-fold, urine cotinine increased six-fold and urine NNAL increased ∼27 times compared to BL biomarker levels. Plasma cotinine, urine COT+3HC and NNAL peaked at 4-8 h post-exposure while urine cotinine peaked within 4 h. Plasma cotinine was significantly correlated to PM2.5 (Spearman correlation rs=0.94) and CO (rs=0.76) but not to air nicotine. The correlations between urine biomarkers, cotinine, COT+3HC and NNAL, and air nicotine, PM2.5 and CO were moderate but non-significant (rs range = 0.31-0.60). CONCLUSIONS: Brief SHS exposure in cars resulted in substantial increases in levels of tobacco biomarkers in non-smokers. For optimal characterisation of SHS exposure, tobacco biomarkers should be measured within 4-8 h post-exposure. Additional studies are needed to better describe the relationship between tobacco biomarkers and environmental markers of SHS.

Particulate mass and polycyclic aromatic hydrocarbons exposure from secondhand smoke in the back seat of a vehicle.

CONTEXT: Exposure to secondhand smoke (SHS) has been reduced in the USA by banning smoking in public places. These restrictions have not had the same effect on children's exposure to SHS as adults suggesting that children are exposed to SHS in locations not covered by bans, such as private homes and cars. OBJECTIVES: Assess exposure to SHS in the backseat of a stationary vehicle where a child would sit, quantify exposures to fine particulates (PM2.5), polycyclic aromatic hydrocarbons (PAH), carbon monoxide (CO) and nicotine. Estimate the impact on a child's mean daily exposure to PM2.5. METHODS: SHS exposures in stationary vehicles with two different window configurations were monitored. A volunteer smoked three cigarettes in a one-hour period for twenty-two experiments. PM2.5, CO, nicotine and PAH where measured in the backseat of the vehicle. 16 PAH compounds were measured for in gas and particle phases as well as real-time particle phase concentrations. RESULTS: The mean PAH concentration, 1325.1 ng/m(3), was larger than concentrations measured in bars and restaurants were smoking is banned in many countries. We estimate that a child spending only ten minutes in the car with a smoker at the mean PM2.5 concentration measured in the first window configuration--1697 mg/m(3)--will cause a 30% increase to the daily mean PM2.5 personal average of a child. CONCLUSIONS: Estimates made using the measured data and previously reported PM2.5 daily mean concentrations for children in California showing that even short exposure periods are capable of creating large exposure to smoke.

A low-cost particle counter as a realtime fine-particle mass monitor.

Exposure to particles with aerodynamic diameters less than 2.5 µm is estimated to cause significant morbidity and mortality worldwide leading many countries to develop ambient air pollution standards and guidelines. At local scales, community and environmental justice groups are also concerned about PM 2.5 concentrations that may be elevated above regional concentrations typically measured by centrally located monitors and standards as well. In an attempt to develop a low cost, easy to use monitor we evaluated a low-cost optical particle counter, the Dylos™, as a fine particulate mass sensor. Modified into a system called the Berkeley Aerosol Information Recording System (BAIRS), we compared performance against standard commercial instruments in chambers using polystyrene latex spheres, ammonium sulphate, and woodsmoke and in an urban ambient setting. Overall we find that the limit of detection of the BAIRS is less than 1 µg m (-3) and the resolution is better than 1 µg m(-3) for PM 2.5. The BAIRS sizes small (<0.5 µm) particles, and is able to accurately estimate the mass concentration of particles of varying composition including organic, inorganic, and ambient particles. It is able to measure concentrations up to 10.0 mg m (-3). In an ambient roof-top test of the BAIRS and a more expensive commercially available light scattering particle monitor the BAIRS response tracked well with the commercial monitor and daily means were within 80% of each other. We conclude that with appropriate modification the system could be developed into an accurate low cost realtime particle mass monitor for use in a wide range of applications.

Estimating personal PM2.5 exposures using CO measurements in Guatemalan households cooking with wood fuel.

As a part of a longitudinal study in the highlands of Guatemala to elicit the chronic health effects of wood smoke from cooking, mean area and personal 48 h concentrations of 2.5 microm particulate matter (PM2.5) and carbon monoxide (CO) were measured every 3 months over 19 months. Monitoring was conducted in 63 households, 28 using traditional open wood fires and 35 using wood cookstoves with chimneys. The goal of this paper is to estimate personal exposure concentrations to PM2.5 using the measurements from CO diffusion tubes as a proxy. CO tubes are cheaper and easier to use than PM-monitoring devices, and can be worn by all family members, even infants. The relationship of PM2.5 and CO was determined by comparing measurements from both co-located instruments. CO measurements in ppm were corrected for temperature and pressure to mass concentrations. PM2.5 exposure was modeled with the following linear regression created using measured concentrations: PM2.5 (mg m(-3)) = 0.10 (0.093, 0.12) x CO (mg m(-3)) + 0.067 (0.0069, 0.13), R(2) = 0.76. No significant difference was found between the separate regressions for open fires and cookstoves. No significant improvement was obtained by applying a mixed statistical model. The equation was used to estimate personal exposures of PM2.5 using personal CO measurements from CO tubes worn by women, infants under 18 months, and children 48-72 months. Estimated 48 h mean personal PM2.5 concentrations for mother, infants, and children in open-fire homes were 0.27 +/- 0.02, 0.20 +/- 0.02, and 0.16 +/- 0.02 mg m(-3) respectively. In chimney-stove homes, mothers and children experienced PM2.5 personal concentrations of 0.22 +/- 0.03 and 0.14 +/- 0.03 mg m(-3), respectively.

SOA formation from partitioning and heterogeneous reactions: model study in the presence of inorganic species.

A predictive model for secondary organic aerosol (SOA) formation by both partitioning and heterogeneous reactions was developed for SOA created from ozonolysis of alpha-pinene in the presence of preexisting inorganic seed aerosols. SOA was created in a 2 m3 polytetrafluoroethylene film indoor chamber under darkness. Extensive sets of SOA experiments were conducted varying humidity, inorganic seed compositions comprising of ammonium sulfate and sulfuric acid, and amounts of inorganic seed mass. SOA mass was decoupled into partitioning (OM(P)) and heterogeneous aerosol production (OM(H)). The reaction rate constant for OM(H) production was subdivided into three categories (fast, medium, and slow) to consider different reactivity of organic products for the particle phase heterogeneous reactions. The influence of particle acidity on reaction rates was treated in a previous semiempirical model. Model OM(H) was developed with medium and strong acidic seed aerosols, and then extrapolated to OM(H) in weak acidic conditions, which are more relevant to atmospheric aerosols. To demonstrate the effects of preexisting glyoxal derivatives (e.g., glyoxal hydrate and dimer) on OM(H), SOA was created with a seed mixture comprising of aqueous glyoxal and inorganic species. Our results show that heterogeneous SOA formation was also influenced by preexisting reactive glyoxal derivatives.