Applications of low-cost sensing technologies for air quality monitoring and exposure assessment: How far have they gone?

Development and Implementation of a Platform for Public Information on Air Quality, Sensor Measurements, and Citizen Science

The use of low-cost sensors for air quality measurements is expanding rapidly, with an associated rise in the number of citizens measuring air quality themselves. This has major implications for traditional air quality monitoring as performed by Environmental Protection Agencies. Here we reflect on the experiences of the Dutch Institute for Public Health and the Environment (RIVM) with the use of low-cost sensors, particularly NO2 and PM10/PM2.5-sensors, and related citizen science, over the last few years. Specifically, we discuss the Dutch Innovation Program for Environmental Monitoring, which comprises the development of a knowledge portal and sensor data portal, new calibration approaches for sensors, and modelling and assimilation techniques for incorporating these uncertain sensor data into air pollution models. Finally, we highlight some of the challenges that come with the use of low-cost sensors for air quality monitoring, and give some specific use-case examples. Our results show that low-cost sensors can be a valuable addition to traditional air quality monitoring, but so far, their use in official monitoring has been limited. More research is needed to establish robust calibration methods while ongoing work is also aimed at a better understanding of the public’s needs for air quality information to optimize the use of low-cost sensors.

Estimation of particulate matter and gaseous concentrations using low-cost sensors from broiler houses

Particulate and gaseous emissions from intensive poultry facilities are major public and environmental health concern. The present study was aimed at exploratively monitoring particulate matter (PM) and gaseous concentrations in controlled-environment facilities using low-cost sensors in Lahore, Pakistan. The indoors and outdoors of 18 broiler houses, grouped into three categories based on the age of birds: group I (1-20 days), group II (21-30 days) and group III (31-40 days), were examined. Low-cost sensors Dylos 1700 and Aeroqual 500 series with different gas sensor heads were used to monitor PM and different gases such as nitrogen dioxide (NO₂), hydrogen sulphide (H₂S), carbon dioxide (CO₂) and methane (CH₄), respectively. Overall, the mean PM and gaseous concentrations increased with the age and activity of birds as compared with the non-activity time of birds. Statistically significant differences were observed in all measured parameters among the groups. The negative correlation between indoor and outdoor environments for PM and gas concentrations at some broiler houses demonstrates the contribution of additional sources to emissions in outdoor environments. The findings contribute to our knowledge of temporal characteristics of particulate and gaseous concentrations from poultry facilities particularly in Pakistan and generally to the capability of using low-cost sensors to evaluate emissions from such facilities.

Distributed Multi-Scale Calibration of Low-Cost Ozone Sensors in Wireless Sensor Networks

New advances in sensor technologies and communications in wireless sensor networks have favored the introduction of low-cost sensors for monitoring air quality applications. In this article, we present the results of the European project H2020 CAPTOR, where three testbeds with sensors were deployed to capture tropospheric ozone concentrations. One of the biggest challenges was the calibration of the sensors, as the manufacturer provides them without calibrating. Throughout the paper, we show how short-term calibration using multiple linear regression produces good calibrated data, but instead produces biases in the calculated long-term concentrations. To mitigate the bias, we propose a linear correction based on Kriging estimation of the mean and standard deviation of the long-term ozone concentrations, thus correcting the bias presented by the sensors.

Long-term field comparison of multiple low-cost particulate matter sensors in an outdoor urban environment

Exposure to ambient particulate matter (PM) air pollution is a leading risk factor for morbidity and mortality, associated with up to 8.9 million deaths/year worldwide. Measurement of personal exposure to PM is hindered by poor spatial resolution of monitoring networks. Low-cost PM sensors may improve monitoring resolution in a cost-effective manner but there are doubts regarding data reliability. PM sensor boxes were constructed using four low-cost PM micro-sensor models. Three boxes were deployed at each of two schools in Southampton, UK, for around one year and sensor performance was analysed. Comparison of sensor readings with a nearby background station showed moderate to good correlation (0.61 < r < 0.88, p < 0.0001), but indicated that low-cost sensor performance varies with different PM sources and background concentrations, and to a lesser extent relative humidity and temperature. This may have implications for their potential use in different locations. Data also indicates that these sensors can track short-lived events of pollution, especially in conjunction with wind data. We conclude that, with appropriate consideration of potential confounding factors, low-cost PM sensors may be suitable for PM monitoring where reference-standard equipment is not available or feasible, and that they may be useful in studying spatially localised airborne PM concentrations.

Assessing exposure to outdoor air pollution for epidemiological studies: Model-based and personal sampling strategies

Epidemiologic studies have found air pollution to be causally linked to respiratory health including the exacerbation and development of childhood asthma. Accurately characterizing exposure is paramount in these studies to ensure valid estimates of health effects. Here, we provide a brief overview of the evolution of air pollution exposure assessment ranging from the use of ground-based, single-site air monitoring stations for population-level estimates to recent advances in spatiotemporal models, which use advanced machine learning algorithms and satellite-based data to accurately estimate individual-level daily exposures at high spatial resolutions. In addition, we review recent advances in sensor technology that enable the use of personal monitoring in epidemiologic studies, long-considered the "holy grail" of air pollution exposure assessment. Finally, we highlight key advantages and uses of each approach including the generalizability and public health relevance of air pollution models and the accuracy of personal monitors that are useful to guide personalized prevention strategies. Investigators and clinicians interested in the effects of air pollution on allergic disease and asthma should carefully consider the pros and cons of each approach to guide their application in research and practice.

Outlier detection and gap filling methodologies for low-cost air quality measurements

Air pollution is a major environmental health problem around the world, which needs to be monitored. In recent years, a new generation of low-cost air pollution sensors has emerged. Poor or unknown data quality, resulting from the intrinsic properties of the sensor as well as the lack of a consensus on data processing methodologies for these sensors, has, among other factors, prevented widespread adoption of these sensors. To contribute to the creation of this consensus, we reviewed the available methodologies for quality control, outlier detection and gap filling and applied two outlier detection methodologies and five gap filling methodologies to a case study (consisting of an 11-month long air quality data set from a low-cost sensor). We showed that erroneous data can be detected in a fully automated way, and that point and contextual outlier detection methodologies can be applied to low-cost air pollution data and yield meaningful results. The linear interpolation showed the best performance for gap filling for low-cost air pollution sensors. In conclusion, data cleaning procedures are important, and the presented methods can form part of a generalised data processing methodology for low-cost air pollution sensors.

Modeling, Monitoring, and Validating Green Roof and Green Facade Solutions with Semantic City Models Using Low Cost Sensors and Open Software Infrastructures

The usage of greenery systems as nature-based solutions to assist in urban cooling in summer time as well as urban warming in wintertime is considered a scientific validated approach in urban planning. The objective of this research is the investigation and quantification of the role of green roofs and green facade solutions concerning thermal behavior in buildings energy savings by using standardized semantic city models that allow the quantification of such measures on district and city scales. The implemented model uses standardized geospatial data based on the CityGML format, a semantic city model standard, for analysis and data storage. For storage of the thermal properties of the buildings, the behavior of its occupants as well as the sensor measurements the Energy ADE of the CityGML standard was used. A green roof/façades model was implemented to simulate the heat transfer in a building based on the heat balance principle of foliage, soil, and structural layers. This model allows analyzing the thermal influence of plant and substrate layers on the heat gains from incoming solar radiation into buildings and the heat losses. This implementation was validated for cooling solutions using monitoring data from real-time experiments during summer measurements at three locations in Germany. Results from this experiment correspond well with the findings of other relevant studies. A sensitivity analysis was conducted to test the impacts of climate, substrate and plants on the greenery layer performance.

Advancing environmental exposure assessment science to benefit society

Awareness of the human health impacts of exposure to air pollution is growing rapidly. For example, it has become evident that the adverse health effects of air pollution are more pronounced in disadvantaged populations. Policymakers in many jurisdictions have responded to this evidence by enacting initiatives that lead to lower concentrations of air pollutants, such as urban traffic restrictions. In this review, we focus on the interplay between advances in environmental exposure assessment and developments in policy. We highlight recent progress in the granular measurement of air pollutants and individual-level exposures, and how this has enabled focused local policy actions. Finally, we detail an illustrative study designed to link individual-level health-relevant exposures with economic, behavioral, biological, familial, and environmental variables.

Cyclists’ Multiple Environmental Urban Exposures—Comparing Subjective and Objective Measurements

Citizens in urban areas are exposed to multiple environmental stressors like noise, heat, and air pollution, with impact on human health. There is a great deal of evidence that connects human health, objective environmental exposure, and place of residence. However, little is known about subjective and objective multiple personal exposures while being mobile. To address this research gap, this paper presents results from a mixed-methods exploratory study with cyclists in the City of Leipzig, Germany. In the summer of 2017, cyclists (n = 66) wore a unique combination of sensors that measured particle number counts (PNC), noise, humidity, temperature, geolocation, and the subjective perception of each exposure on everyday routes for one week (n = 730). A smartphone application was developed to question participants about their perception of subjective exposure. The data were analyzed with three aims: (i) to compare the multiple exposure profiles of the cyclists, (ii) to contrast the objective data and subjective individual perception, and (iii) to examine the role of route decision-making and awareness of health impacts for healthier route choices. The results indicate distinct differences between the exposure profiles of cyclists. Over 80% of the cyclists underestimated their exposure to noise and air pollution. Except for heat, no significant associations between the objective and subjective data were found. This reveals an exposure awareness gap that needs to be considered in urban health planning and risk communication. It is argued that knowledge about health impacts and route characteristics plays a crucial role in decision-making about route choices. The paper concludes with suggestions to harness smart sensing for exposure mitigation and research in health geography.

Testing the performance of sensors for ozone pollution monitoring in a citizen science approach

Tropospheric ozone (O₃) is an environmental pollutant of growing concern, especially in suburban and rural areas where the density of air quality monitoring stations is not high. In this type of areas citizen science strategies can be useful tools for awareness raising, but sensor technologies must be validated before sensor data are communicated to the public. In this work, the performance under field conditions of two custom-made types of ozone sensing devices, based on metal-oxide and electrochemical sensors, was tested. A large array of 132 metal-oxide (Sensortech MICS 2614) and 11 electrochemical (Alphasense) ozone sensors, built into 44 sensing devices, was co-located at reference stations in Italy (4 stations) and Spain (5). Mean R² between sensor and reference data was 0.88 (0.78-0.96) and 0.89 (0.73-0.96) for Captor (metal-oxide) and Raptor (electrochemical) nodes. The metal-oxide sensors showed an upper limit (approximately 170 μg/m³) implying that these sensors may be useful to communicate mean ozone concentrations but not peak episodes. The uncertainty of the nodes was 10% between 100 and 150 μg/m³ and 20% between 150 and 200 μg/m³, for Captors, and 10% for >100 μg/m³ for Raptors. Operating both types of nodes up to 5 months did not evidence any clear influence of drifts. The use of these sensors in citizen science can be a useful tool for awareness raising. However, significant data processing efforts are required to ensure high data quality, and thus machine learning strategies are advisable. Relative uncertainties should always be reported when communicating ozone concentration data from sensing nodes.

Fluctuation in time-resolved PM2.5 from rural households with solid fuel-associated internal emission sources

Indoor air contributes significantly to overall exposure, particularly for rural Chinese who often use solid fuels for cooking and/or heating. Unfortunately, overlooked rural indoor air leads to a critical knowledge gap. Simultaneous measurements in the kitchen, living room, and immediately outside of houses using six-channel particle counters were carried out in 18 biomass-burning rural and 3 non-biomass-burning urban households (as a comparison) in winter to characterize dynamic change patterns indoor air pollution and indoor-outdoor relationship. The rural households mainly used wood or crop residues for cooking and heating, while the urban households used pipelined natural gas for cooking and air conditioners for heating. In rural households with significant solid-fuel burning internal sources, the highest concentration was found in the kitchen (101 ± 56 μg/m³), with comparable levels in the living room (99 ± 46 μg/m³) and low levels in outdoor air (91 ± 39 μg/m³). A generally opposite direction of indoor-outdoor exchange was found between the rural and urban households. PM in kitchen air is smaller than that in living rooms and outdoors because solid fuel burning (mainly in rural households) and cooking oil heating (in rural and urban households). Indoor and outdoor PM concentration changed synchronously, with a slight delay in indoor air in urban households but a slight delay in outdoor air in rural households. Cooking, heating, and smoking elevated indoor PM significantly, but different from the cooking activity that produced peaks lasting for about 30 min, emissions from heating created a series of peaks due to frequent disturbance and fuel-feeding and had more significant impacts on the daily average concentration. Distinct indoor-outdoor relationships and dynamic change patterns between the two household categories w/o strong internal biomass burning sources imply that totally different model schemes are needed to quantitatively address indoor air pollution and inhalation exposure.

Development of an in-home, real-time air pollutant sensor platform and implications for community use

Air pollution exposure characterization has been shaped by many constraints. These include technologies that lead to insufficient coverage across space and/or time in order to characterize individual or community-level exposures with sufficient accuracy and precision. However, there is now capacity for continuous monitoring of many air pollutants using comparatively inexpensive, real-time sensors. Crucial questions remain regarding whether or not these sensors perform adequately for various potential end uses and whether performance varies over time or across ambient conditions. Performance scrutiny of sensors via lab- and field-testing and calibration across their lifetime is necessary for interpretation of data, and has important implications for end users including cost effectiveness and ease of use. We developed a comparatively lower-cost, portable, in-home air sampling platform and a guiding development and maintenance workflow that achieved our goal of characterizing some key indoor pollutants with high sensitivity and reasonable accuracy. Here we describe the process of selecting, validating, calibrating, and maintaining our platform - the Environmental Multi-pollutant Monitoring Assembly (EMMA) - over the course of our study to-date. We highlight necessary resources and consider implications for communities or researchers interested in developing such platforms, focusing on PM_2.5, NO, and NO₂ sensors. Our findings emphasize that lower-cost sensors should be deployed with caution, given financial and resource costs that greatly exceed sensor costs, but that selected community objectives could be supported at lesser cost and community-based participatory research strategies could be used for more wide-ranging goals.

End-User Feedback on a Low-Cost Portable Air Quality Sensor System-Are We There Yet?

Low-cost sensors are a current trend in citizen science projects that focus on air quality. Until now, devices incorporating such sensors have been tested primarily for their technical capabilities and limitations, whereas their usability and acceptability amongst the public rarely goes beyond proof of concept, leaving user experience (UX) unstudied. The authors argue that UX should be taken into account to make sure that products and services are fit for purpose. Nineteen volunteers tested and evaluated a prototype device and provided feedback through semi-structured interviews and during focus group sessions. Their UX was then coded using mixed coding methods regarding device functionality and recommendations for future product development. The results indicate that UX can identify potentially problematic design aspects while giving deeper insights into user needs. For example, UX recognized that one of the most important aspects of user involvement and motivation was successful data harvesting, which frequently failed. This study recommends that future developers of low-cost portable air quality sensor systems prioritize reliable data transmission to minimize data loss. This will ensure an efficient and positive UX that supports user engagement in citizen science based research where collecting sensor-based data is the primary objective.

Investigations into factors affecting personal exposure to particles in urban microenvironments using low-cost sensors

Epidemiological studies have linked outdoor PM_2.5 concentrations to a range of health effects, although people spend most of the time indoors. To better understand how individuals' exposure vary as they move between different indoor and outdoor microenvironments, our study investigated personal PM_2.5 exposure and exposure intensity of 14 adult volunteers over one week (five weekdays and one weekend), using low-cost personal monitors, recording PM_2.5 concentrations in 5 min intervals. Further, the study evaluated community perception of air pollution exposure during the recruitment and engagement with the volunteers. We found that people with tertiary education across all ages had greater interest in participating, with younger people being interested regardless of the level of education. The derived exposures and exposure intensities differed between weekdays and the weekend due to larger variations in individuals' daily routines. In general, time spent at home and engaged in indoor activities was associated with the highest personal PM_2.5 exposure and exposure intensity on both, week and weekend days, implying the significance of both duration of the exposure and the indoor PM_2.5 concentrations. The results showed no relationship between personal exposures and indoor characteristics of home (ventilation, building age and cooktop), which are expected to be due to the study's small sample size. The observed PM_2.5 > 10 μg m^-3 were significantly higher for distances <50 m to the roads for both major and minor roads, and were observed in areas with <16% open space, which were also close to a major road.

Indoor Air Quality: A Bibliometric Study

What are the actual trends in Indoor Air Quality (IAQ), and in which direction is academic interest moving? Much progress has been made in identifying and mitigating indoor pollutants, due to both prevention campaigns (e.g., smoking bans) and greater control of product emissions. However, IAQ is still of interest and the future trends are unknown. In this study, a thorough bibliometric analysis was conducted on the scientific literature available on the Web of Science database with CiteSpace from 1990 until today. It was possible to identify past trends and current advances, both with the aim of introducing the IAQ topic to those encountering it for the first time and to examine the issues that are expected to be pertinent in the future.

Bridging the Data Gap From in vitro Toxicity Testing to Chemical Safety Assessment Through Computational Modeling

Chemical toxicity testing is moving steadily toward a human cell and organoid-based in vitro approach for reasons including scientific relevancy, efficiency, cost, and ethical rightfulness. Inferring human health risk from chemical exposure based on in vitro testing data is a challenging task, facing various data gaps along the way. This review identifies these gaps and makes a case for the in silico approach of computational dose-response and extrapolation modeling to address many of the challenges. Mathematical models that can mechanistically describe chemical toxicokinetics (TK) and toxicodynamics (TD), for both in vitro and in vivo conditions, are the founding pieces in this regard. Identifying toxicity pathways and in vitro point of departure (PoD) associated with adverse health outcomes requires an understanding of the molecular key events in the interacting transcriptome, proteome, and metabolome. Such an understanding will in turn help determine the sets of sensitive biomarkers to be measured in vitro and the scope of toxicity pathways to be modeled in silico. In vitro data reporting both pathway perturbation and chemical biokinetics in the culture medium serve to calibrate the toxicity pathway and virtual tissue models, which can then help predict PoDs in response to chemical dosimetry experienced by cells in vivo. Two types of in vitro to in vivo extrapolation (IVIVE) are needed. (1) For toxic effects involving systemic regulations, such as endocrine disruption, organism-level adverse outcome pathway (AOP) models are needed to extrapolate in vitro toxicity pathway perturbation to in vivo PoD. (2) Physiologically-based toxicokinetic (PBTK) modeling is needed to extrapolate in vitro PoD dose metrics into external doses for expected exposure scenarios. Linked PBTK and TD models can explore the parameter space to recapitulate human population variability in response to chemical insults. While challenges remain for applying these modeling tools to support in vitro toxicity testing, they open the door toward population-stratified and personalized risk assessment.